ICOWOBAS 2015 - UNUD

5th International Conference

and Workshop on Basic and Applied Sciences

vii

ICOWOBAS 2015 October 16-17th, 2015

Surabaya Indonesia

FOREWORD BY CHAIRMAN OF 5TH ICOWOBAS 2015

Ladies and Gentleman,

International Conference and Workshops on the Basic and Applied Sciences (ICOWOBAS) is organized as impelementation of existing collaborations between Airlangga University, Universiti Teknologi Malaysia and Salahaddin University of Hawler, Erbil-Iraq in order to promote the development of sciences and their prospect of application in industry and medical devices. The program of this activity are the scientific program involves the presentation of the paper and poster in the area of chemistry, biology, physics, mathematics and their applications. It also conducts the workshop program that will be presented the current issues in optical instrumentation. Thus invited many participants as academic researchers, scientists, industrial professionals, government officers, students and other participants. The meeting intends to bring together researcher, scientists and scholars to exchange and share their experiences, new ideas, research novelties in related fields and discuss the practical challenges encounetered and the solutions adopted.

These proceedings hold the full papers presented at the 5th ICOWOBAS. The conference took place in Surabaya (Indonesia) at the Garuda Mukti Room, Kampus C Universitas Airlangga, October 16th - 17th, 2015, and the worskhop was conducted at the Faculty of Science and Technology, Airlangga University, October 15th, 2015.

Higlights of the conference include: Assoc. Prof. Dr. Takaya (Meijo University, Japan), Prof. Stephen Geoffrey Pyne (University of Wollongong, Australia), Prof. Zuhaimy Ismail (Universiti Teknologi Malyasia, Malaysia), Prof. Dr. Harith Ahmad (University of Malaya, Malaysia), Prof. Dr. Ismail Bin Mohd. (Universiti Malaysia Perlis, Perlis, Malaysia) and Prof. Dr. Retna Apsari (Universitas Airlangga, Indonesia) as keynote and invited speaker. In total, we received 152 abstarct submssions for oral and 29 posters, and more than 70 full paper submissions. The selected paper will be published on AIP Conference Proceedings (Scopus index). As the acceptance rates illustrates the competition is stiff, and the accepted submission reflected high rates of

Dr. Moh. Yasin



viii


Surabaya Indonesia

reviewer enthusiasm. By design, these are lighty reviewed and almost always accepted.

ICOWOBAS is a lot of work. We could not have done it without help from many people. We would especially like to thank: Scientific board of ICOWOBAS, for inviting us to chair the meeting; The Rector of Airlangga University, for permiting and supporting us to conduct the conference; Our collaeagues in the Faculty of Science and Technology, Airlangga University, for their support in the conference; the local committee, for organizing and handling the conference; the many reviewers, for providing professional reviews; our sponsor: Vitalong C and DGHE through Airlangga University funding.

Surabaya, October 16th - 17th, 2015

Surabaya, October 16th, 2015. Dr. Moh. Yasin, MSi



ix


Surabaya Indonesia

FOREWORD BY CO-CHAIR OF 5TH ICOWOBAS 2015

Salam,

It is an honor and greatest pleasure for me to welcome all participants and those supporting 5th International Conference and Workshop on Basic and Applied Sciences (5th ICOWOBAS).

The 5th International Conference and Workshop on Basic and Applied Sciences has been organized by the Faculty of Science and Technology, Universitas Airlangga (Surabaya, Indonesia), Salahaddin-Hawler University (Erbil, Iraqi Kurdistan) and Faculty of Science, Ibnu Sina Institute for Fundamental Science Studies Universiti Teknologi Malaysia (Johor Bahru, Malaysia). ICOWOBAS is a platform for the international community, especially among academics and scientists to present and exchange ideas and share insights of research findings and academic accomplishments. In addition, this conference provides the avenue to promote research, enhance skills in paper writing and oral presentation.

Finally, I would like to take this opportunity to express my utmost gratitude to the International Advisory committees, all the Reviewers and Organizing Committee for their relentless effort in ensuring the successful implementation of conferences. Last but not least, my sincere appreciation to Dr. Ahmed Anwar Dezaye, the President of SUH and all those involved in making this conference possible.

Thank you.

ErBil, October 16th, 2015. Dr. Hewa Y Abdullah

Dr. Hewa Y Abdullah



x


Surabaya Indonesia

TABLE OF CONTENT

VENUE, ORGANIZER, AND COMMITTEE......................................................................................... II

FOREWORD BY DEAN FACULTY OF SCIENCE AND TECHNOLOGY, AIRLANGGA UNIVERSITY ............ VI

FOREWORD BY CHAIRMAN OF 5TH ICOWOBAS 2015 .................................................................... VII

FOREWORD BY CO-CHAIR OF 5TH ICOWOBAS 2015 ....................................................................... IX

TABLE OF CONTENT ...................................................................................................................... X

KEYNOTE AND INVITED’S PAPERS ............................................................................................... 13

FUZZY LOGIC METHOD FOR SHORT TERM ELECTRICITY LOAD FORECASTING: A MALAYSIAN EXPERIENCE.............. 14 Zuhaimy Ismail* and Nuramirah Akrom ..................................................................................................... 14

ORAL PAPER OF ANALYTICAL AND FORENSIC CHEMISTRY (OAFC) ................................................ 28

DETERMINATION OF THIAMINE HYDROCHLORIDE USING FLOW INJECTION SPECTROPHOTOMETRIC METHOD ....... 29 Hijran Sanaan Jabbar

1,* and Azad Tawfiq Faizullah

2 ................................................................................... 29

ANALYSIS OF N-NITROSODIETHYLAMINE IN SALTED FISH USING CONE SHAPED MEMBRANE-LIQUID PHASE

MICROEXTRACTION-GAS CHROMATOGRAPHY-FLAME IONIZATION DETECTOR ................................................. 41 Usreg Sri Handajani*, Olivia Animiko & Yanuardi Raharjo ......................................................................... 41

PREPARATION AND PROPERTIES OF TETRABORATE ION SELECTIVE SENSOR BASED ON ZEOLITE........................... 51 Zuri Rismiarti

a*, Atikah

b, Chasan Bisri

b, Yuyun Irnawati

a ............................................................................. 51

ORAL PAPER OF PURE AND APPLIED MATHEMATICSPURE (OAMT) .............................................. 56

RULE-BASED AND CASE-BASED REASONING SYSTEM FOR PSYCHIATRIC PSYCHOSIS DISORDERS DIAGNOSIS .......... 57 Ause Labellapansa

1*, Sri Hartati

2 .............................................................................................................. 57

OPTIMIZATION OF JOB SHOP SCHEDULING PROBLEM USING MODIFIED GENETIC ALGORITHM ........................... 65 Eto Wuryanto

a*

and Dyah Herawatie

a,b ...................................................................................................... 65

EXPERT’S PROMOTION ELIBILITY SYSTEM USING BACKPROPAGATION ARTIFICIAL NEURAL

NETWORK ......................................................................................................................................... 70 Indah Werdiningsi

1*, Army Justitia, Rini Sumiati, and Nur Hesti ................................................................ 70

IMPLEMENTATION OF JOURNAL CLASSIFICATION INFORMATION RETRIEVAL SYSTEM WITH K-NEAREST NEIGHBOR . 77 Endah Purwanti*, Badrus Zaman ................................................................................................................ 77

ORAL PAPER OF BIODIVERSITY (OBDV) ....................................................................................... 82

BIOACTIVE COMPOUNDS OF THE MOSS HOMALIODENDRON FLABELLATUM (SMITH) FLEISCL ............................. 83 Junairiah

a*, Tri Nurhariyati

a, Suaibah

a, Ni’matuzahroh

a, Lilis Sulistyorini

b ................................................. 83

BIOSISTEMATIC OF BRYOPSIDA IN HOT SPRING AREA OF R. SOERYO CANGAR GRAND FOREST PARK, EAST JAVA ... 88 Hamidah

a*, Thin Soedarti

a, and Nathania Ernita Ekawati Edawua

a ........................................................... 88

DIVERSITY OF THE SCHMUTZDECKE LAYERS IN SLOW SAND FILTER (SSF) AND EFFECT ON EFFICIENCY

REMOVAL OF POLLUTANTS............................................................................................................... 94 Muchammad Tamyiz

1), Laily Noer Hamidah

2,* ............................................................................................ 94

FORTIFICATION CASHEW NUT LEAF (ANACARDIUM OCCIDENTALE L.) USING GLYSINE ON THE GROWTH

AND DEVELOPMENT OF WILD SILKWORM (CRICULA TRIFENESTRATA HELF.) ......................................... 100 Sulistyo Dwi Kartining Putro*, Jekti Prihatin, and Suratno ....................................................................... 100

ORAL PAPER OF BIOMEDICAL ENGINEERING (OBME) ................................................................ 103

‘LETSLEEP MASK’ (LED AND BINAURAL BEATS SLEEPING MASK), SLEEPING MASK AS SLEEP-WAKE CYCLE

REGULATIONS BASED ON BRAIN WAVE STIMULATION ............................................................................... 104



xi


Surabaya Indonesia

Azisya Amalia Karimasari1*, Amila Sofiah1, Priyanka Kusuma Wardhani

1, Andi Achmad Dzulfiqar

1, Choirul

Chabib2, Novia Dwi Asmaningtias

1 Prihartini Widiyanti

1 .......................................................................... 104

ORAL PAPER OF COMPUTATIONAL PHYSICS, CHEMISTRY & MATHEMATICS (OCPC) ................... 112

EVALUATION AND ANALYSIS OF SMART ANTENNAS BY ADAPTIVE BEAM-FORMING USING LEAST MEAN SQUARE

(LMS) ALGORITHM ............................................................................................................................. 113 Mudhaffer M. Ameen* and Avin Jawhar Ali* ........................................................................................... 113

ORAL PAPER OF ENVIRONMENTAL BIOCHEMISTRY AND BIOTECHNOLOGY (OEBB) ..................... 120

INDUCTION OF APOPTOSIS AND ANTIANGIOGENESIS EFFECTS OF PINOSTROBIN FROM KAEMPFERIA PANDURATA

ROXB AGAINST INDUCTION OF FIBROSARCOMA MICE RESULTS BENZOPIREN .............................................. 121 Adi Parwata

1*, Sukardiman

2, Mulja H. S.

3, Alit Widhiartini

4 ..................................................................... 121

THE ABILITY OF SWEET ORANGE PEEL’S(CITRUS SINENSIS) PECTIN AS BIOSORBENT OF HEAVY METAL CHROME (VI) ........................................................................................................................................................ 126

Natalia Widya Yuda Suryaningtyas, L. Indah M. Yulianti, and P. Kianto Atmodjo, ................................ 126

ORAL PAPER OF ENVIRONMENTAL AND GREEN CHEMISTRY (OEGC) .......................................... 131

INVESTIGATION OF FOODS CONTAINING BORAX IN SURABAYA WITH EXTRACT OF TURMERIC ........................... 132 Dwi Yulian Fahruddin Shah

1, M. Al Rizqi Dharma Fauzi

1* , Andre

Pratama

1, Nufida Dwi A.

2, Hery Suwito

1

.................................................................................................................................................................. 132

ORAL PAPER OF MICROBIAL BIOCHEMISTRY AND MOLECULAR BIOLOGY (OMBM)..................... 138

LIQUID ORGANIC FERTILIZER USED MICROBIAL FROM LOCAL INOCULANT...................................................... 139 Rita Tri Puspitasari*, Elfarisna,Yati Suryati, and Nosa T. Pradana ........................................................... 139

ORAL PAPER OF NATURAL PRODUCTS AND MEDICINAL CHEMISTRY (ONPM) ............................. 150

DETERMINATION OF 40K CONCENTRATION IN IMPORTED MILK POWDERS IN ERBIL CITY LOCAL MARKETS ......... 151 Ali H. Ahmed

* and Mohhammed I. Hussein ............................................................................................... 151

GANDARUSA (JUSTICIA GENDARUSSA BURM. F.) SHOOTS INDUCTION BY THE COMBINATION OF NAA AND BAP 157 Dwi Kusuma Wahyuni 1

a*, Bambang Prajoga 2

b, Cheli Maulana 2

a, Hery Purnobasuki 3

a , Hamidah 4

a

dan Noer Moehammadi 5a ........................................................................................................................ 167

SYNTHESIS OF Β-IONONE FROM CITRUS AURANTIFOLIA AS PRECURSOR OF VITAMIN A .................................... 173 Indah Rizki Ulya, Rissa Dwi Susanti, Meyta Restu Wigati, Maryam Putri Eradewi, Novia Eka Setyatama, Siti Mariyah Ulfa* ..................................................................................................................................... 173

POLYSACCHARIDE KRESTIN ACTIVITY OF CORIOLUS VERSICOLOR EXTRACT ON INTERLEUKIN-12 LEVEL OF MUS

MUCULUS EXPOSED TO MYCOBACTERIUM TUBERCULOSIS .......................................................................... 182 Sri Puji Astuti Wahyuningsih

1*, Sugiharto

1, Nurul Wiqoyah

2 & Zakiyatun Nuha

1 ..................................... 182

ORAL PAPER OF OPTICS AND PHOTONICS (OPHO) ..................................................................... 187

INVESTIGATION OF GRAPHENE SYNTHESISED BY ELECTROCHEMICAL EXFOLIATION AS PASSIVE SATURABLE

ABSORBER FOR PULSED LASER GENERATION ............................................................................................ 188 Fauzan Ahmad

1,2, Sulaiman Wadi Harun

3, Roslan Md Nor

4, Harith Ahmad

3 and Mohd Haniff Ibrahim

2 . 188

ORAL PAPER OF PURE AND APPLIED MATHEMATICS (OPMT) .................................................... 195

THE NUMERICAL SIMULATIONS OF INVERSE PROBLEMS ON THE PARAMETER ESTIMATION ............................... 196 Julan Hernadi

1,*......................................................................................................................................... 196

FORMULATING THE LINEAR MODEL OF GRAPH COLORING USING THE COMBINATION OF GOMORY CUTTING PLANE

METHOD AND BALAS ALGORITHM ......................................................................................................... 205 Sisca Octarina* , Merry Pusvita Sari and Eddy Roflin ................................................................................ 205

POSTER PAPER OF SEMINAR (PPOS) ......................................................................................... 213

ELECTROCHEMICAL DEGRADATION REMAZOL BLACK B USING NANOPOROUS CARBON PASTE ELECTRODES ....... 214 Nafila*, Muji Harsini, Pratiwi Pudjiastuti .................................................................................................. 214



xii


Surabaya Indonesia

BENZYLTRIETHYLAMMONIUM CHLORIDE AS CORROSION INHIBITOR ON ZINC WITH SOAKING METHOD ............. 220 Suyanto

1*, Djony Izak R

2, Ofi Rulytasari

1 .................................................................................................. 220

THE EFFECT OF MANGOSTEEN (GARCINIA MANGOSTANA L) PERICARP EXTRACT TOWARDS BODY WEIGHT AND

FASTING BLOOD CHOLESTEROL LEVEL OF DIABETIC MICE ........................................................................... 223 Saikhu Akhmad Husen

1* and Dwi Winarni

1) ............................................................................................. 223

THE ACTIVITY OF ANDONG LEAF SAPONIN (CORDYLINE TERMINALIS KUNTH.) AGAINST CHOLESTEROL AND DIPHENYL

PICRYL HIDRAZYL (DPPH) IN INVITRO ..................................................................................................... 228 Ni Wayan Bogoriani* ................................................................................................................................ 228

INTEGRATED MODEL OF ONE PARAMETER LOGISTIC MODEL AND RESPONSE TIME MODEL ............................. 234 Noer Hidayah

1*, Kumaidi

2, Badrun Kartowagiran

2 ................................................................................... 234

STUDY OF VARIATION ON THE CONCENTRATION OF DIGESTION METHOD TO HEAVY METAL DETERMINATION IN

RIVER SEDIMENT SAMPLES ................................................................................................................... 245 Yudhi Utomo*, Neena Zakia ...................................................................................................................... 245

NEGATIVE EFFECT OF CTAB IN SYNTHESIS OF ZNO/MSN PHOTOCATALYST FOR DECOLORIZATION OF

METHYL ORANGE ............................................................................................................................ 248 Jusoh, N.W.C.

1, Jalil, A.A.

1,2*, Triwahyono, S.

3 ........................................................................................ 248

INOVATION OF COLLAGEN BASED CORNEAL HYDROGEL WITH THE ADDITION OF GLYCOPOLYMER AS THE SOLUTION

FOR IRREVERSIBLE BLINDNESS BY CORNEAL ULCERS ................................................................................... 251

Disca Sandyakala Purnama*, 2)Hendita Nur Maulida, 3)Rara Setya Angtika, 4)Astriani

Hendayanti,5)Anandhika Muhammad Satria ............................................................................................ 251 PROPERTIES OF NANO HYDROXYAPATITE AND POLY (1,8 OCTANEDIOL-CO- CITRATE) (POC) FOR BIODEGRADABLE

BONE SCREW ...................................................................................................................................... 255 Fitriyatul Qulub*, Hendita Nur Maulida, Ewing Dian Setyadi, Systi Adi Rachmawati, Imroatus Sholikhah, Prihartini Widiyanti ................................................................................................................................... 255



13


Surabaya Indonesia

KEYNOTE AND INVITED’S PAPERS



14


Surabaya Indonesia

Fuzzy Logic Method for Short Term Electricity Load Forecasting: A

Malaysian Experience

Zuhaimy Ismail* and Nuramirah Akrom

1,2 Department of Mathematical Sciences, Faculty of Sciences, Universiti Teknologi Malaysia 81310, Skudai, Johor, Malaysia

* Corresponding author’s Email: [email protected]

Abstract. Forecasting electricity demand is a vital process since electricity is a hard-to-store resource. To accurately forecast electricity demand, this paper discusses a Malaysian experience in forecasting electricity load demand using fuzzy logic. This is a new concept in forecasting cycle based on fuzzy logic. The load data are influenced by the variances in day types such as the normal working day, the weekend and holidays as each of these day type has different load behavior. Rules were developed for each day factors, temperatures and load using current load, current temperature, previous load and previous temperature. Defuzzification is one of the procedures in fuzzy logic which is essential in producing the forecast load. Defuzzification methods such as Center of area (COA), Middle of maxima (MOM), Center of gravity (COG) were used and with simple statistics, the membership function of input and output is identified. The daily load data from Tenaga National Berhad (TNB) Malaysia were used as data sets for training and testing the performance of each defuzzification method. Using Mean Square Error (MSE) and average error, the results show that fuzzy model with center of area (COA) gave a better performance compared to other defuzzification methods.

INTRODUCTION

Forecasting the electricity load demand is an important task in power utility companies because accurate load forecasting results are economic, reliable and secure power system operation and planning. Accuracy in forecasting becomes important because a single percentage error may lead to gain or loss of millions of Ringgit. In Malaysia, Tenaga Nasional Berhad (TNB) and several Independent Power Producers (IPPs) are responsible to provide electric power. TNB is a government link company with core business is in the generation, transmission, and distribution of electricity. This company controls the grid system and has the capacity to produce about 14,000 MW of electricity. It managed and operated a complete transmission network, spanning the whole of Peninsular Malaysia. In handling such a large amount of electricity and its distribution, the company has to develop a very good planning and forecasting. In TNB, forecast are conducted by seasoned forecasters and uses methods such as regression analysis, time series methods and the hybrid methods with artificial intelligence techniques such as neural network, genetic algorithm and fuzzy logic (G.C. Liao, T.P. Tsao, 2004). In the last decade, the deterministic approach such as fuzzy forecasting has been widely studied for forecasting data which are of dynamic and non-linear.

In the last decades, various methods have been put forward for load forecasting. The methods can be categorized as; i) univariate method; ii) multivariate method; and iii) combined method (Chatfield, 2004). Univariate methods include exponential smoothing (Taylor et al., 2006),Box-Jenkins approach (Chen et al, 1995), nonparametric functional methods (Vilar et al. 2012), Kalman filters (Al-Hamadi and Soliman, 2004), Artificial neural network (ANN) (Badri et al, 2012) and Support Vector Machine (SVM) (Hu et al, 2014) are linear and non-linear models. Multivariate methods, that include lagged external variables such as Multivariate Adaptive Regression Splines (MARS) (Cheng & Cao, 2014), GARCH method (Sigauke and Chikobvu, 2011), Multivariate Non-parametric Regression (Sigauke and Chikobvu, 2010) and Nonlinear Autoregressive models with exogenous inputs (NAX) (Andalib and Atry, 2009) have produced great results for short term load forecasting. Most researchers used temperature (Matteo et al, 2013 and Liu, et al. 2014), wind generation (Cruz et al. 2011), special day effects (Kim, 2013) as their exogenous variables in multivariate methods. Bowerman et al. (2005) classify types of forecast in four time frames namely the Immediate forecasting, Short-term forecasting, Medium-term forecasting and Long-term forecasting. In this study, the types of forecast are classified to three time frames. The practices in TNB, forecast are conducted in all time frames described above. They run hourly load demand forecast for week-ahead, a month to four months ahead and five years ahead load demand forecast. For immediate forecasting, they used half-hourly load demand and forecast a day ahead. In this research, the types of forecast are confined to short term load demand where short term forecasting is limited to less than one month ahead. Short-term electricity forecasting is very important because the information from forecasting can help to build up cost effective risk management plans for any electric



15


Surabaya Indonesia

utility such as electricity price and electricity power plan, operation planning and real time planning (Fu-yuan et al., 2006 and Matteo et. al. 2013). Future demands are also required for the control and scheduling of power systems. The importance of demand forecasting needs to be emphasized at all level as the consequences of under or over forecasting the demand are serious and will affect all stakeholders in the electricity supply industry. If the load demand is under estimated, the result is serious since plant installation cannot easily be advanced, this will affect the economy, business, loss of time and image but if over estimated, the financial penalty for excess capacity and wasting of resources.

A large variety of statistical and artificial intelligent (AI) techniques have been developed for short term electricity load demand forecasting in order to minimize the prediction error. The different methodologies between researchers not only differ in term of factors included in the model but also the employed technique to forecast electricity load demand. Commonly used statistical forecasting model for load demand include the Box-Jenkins ARIMA (Razak et al., 2006; Saab et al., 2001; Taylor et al., 2006), exponential smoothing (Razak et al., 2006), linear regression and multiple linear regression (Ismail and Jamaluddin, 2009; Mirasgedis et al., 2006). These statistical methods are based on normality assumption for the interpretation of the test of significance and models have to have linear relationship between load demand and several factors that might affect the demand. If the data does not satisfy this assumption, the obtained results may be biased.

Fuzzy approach is widely applied in various fields. Ismail and Efendy (2011, 2014), Bagis (2008), Frantti and Mahonen (2009), Mastorocostas et al. (2000), Mori and Kobayashi (1996), Ozawa and Niimura (1999), Pandian et al. (2006), Ranaweera et al. (1996), Thang (2004) and Yang and Huang (1998) apply fuzzy approach in forecasting area. Research by Bagis (2008), Karakuzu (2007), Tarng et al. (2004) and Tsekouras (2003), applied in other fields and vehicle parking problem field by Esmin (2002) and operation of spillway gates (Bagis, 2003; Karaboga et al., 2008). A large variety of statistical and artificial intelligent (AI) techniques have been developed for short term electricity load demand forecasting in order to minimize the prediction error. A number of studies showed that the relationship between most factors and electricity load demand are nonlinear. For example, Pandian (2006) and Ranaweere (1996) develop fuzzy model for electricity load demand, In 2009, Ismail and Jamaluddin explore the used of artificial neural network in modeling electricity demand, while decision tree model was developed by Karapidakis (2007). On the other hand, Ranaweera (1996) reported that their fuzzy model is better than the ANN and Pandian (2006) showed that fuzzy approach is better than traditional time series models. Electricity load demand is influenced by many factors, such as weather, economic and social activities, and different load. By analysis of only historical load data, it is difficult to obtain accurate load demand for forecasting. The relation between load demand and the independent variables is complex and it is not always possible to fit the load curve using statistical models. The numerical aspects and uncertainties of this problem appear suitable for fuzzy methodologies (Momoh et al., 1995).

The rest of the paper is organized as follows. Section two discusses the data source and the different types of factors influencing short term load demand. The data are presented in the form of graphs of load verses half-hourly time in a day. This gives a description on the load profile of the days. The methodology is given in section three and it also described the proposed method of fuzzy logic in detail with some examples. The proposed defuzzification methods namely the COA, COG, MOM and LOM with simple statistics to identify the membership function of input and output is given in section four. Section five gives results of the empirical analysis of load demand by using daily load data from Tenaga National Berhad (TNB) Malaysia for training and testing, the performance of the fuzzy logic approach. The final section concludes the paper.

DATA DESCRIPTION AND ANALYSIS

Short-term load forecast (STLF) plays an important role in the management of electricity power system. It provides the load predictions for basic generation scheduling, assessing the security of the power systems at any time point and as timely dispatcher information. It is also used to drive the scheduling functions that determine the most economic commitment of generation sources consistent with reliability requirements, operational constraints and policies, environmental and equipment limitations; setting the spinning reserves maintenance scheduling and setting an optimal mix. There are some STLF articles in Malaysia‘s electricity load for instance

neural network approach (Mohamed et al., 2010; Ismail and Jamaludin, 2008) rule based approach and statistics (Ismail et al.; 2009). In Malaysia, the STLF is applied to hourly, daily, weekly and monthly load forecast. Methods employed at TNB is based on the experience of the forecaster where the forecasters have to develop an intuitive relationship between electrical load and weather parameters, time of the day, day of week, season and time lag response. Various factors are needed to derive the hourly, daily and weekly forecast load forecast. These factors included daily temperature, legal and religious holidays, seasonal effects and human behavior whether they will take a day off preceding and following the holidays as to take advantages of long break (Annual Report TNB, 2006). The weekly load profile based for Peninsular Malaysia end-user daily lifestyle may



16


Surabaya Indonesia

be describe quantitatively and in a subjective manner. The load profile based for Malaysia‘s end-user daily lifestyle in given in Figure 1.

FIGURE 1. Half-hourly load pattern of end-users for Peninsular Malaysia In Figure 1, it shows the different load values for different weekdays and weekends. The first group of line

graph contains five data sets from 27 February 2006 to 3 March 2006 (from Monday to Friday) present the weekday load pattern. The second group of line graph contains two data sets, 25 February and 4 March 2006 (Saturday) and the third group contains two data sets, 26 February and 5 March 2006 (Sunday). In general, it changes from weekdays to weekend. This graph also presents the different curves for public holidays and special event days. From these profiles, there are many differences in the pattern between weekdays and weekend due to various uncertainties which led us to explore the used of fuzzy logic approach to forecasting short-term load demand.

In general, the weekend electricity load demand was lower than weekday because of weekday is working day. Most all department, factory, university and schools are operating on weekday. Sunday load was the lowest load because most all building was closed and any activities such as machines, air conditioner and lamp were not used at the time. The fluctuation pattern of electricity demand can be as follows: The graph begins with a load demand from 12.30 am until 12.00 midnight. The load demand is rather low from 12.30am and begins to increase from 8am until 12.00pm. Most school, factory, office department and any building start activities at 8.00 am and some factory and private office department start operate on 9.00 am while the shopping complexes start operating at 10.00 am. Hence the pattern increases obviously during that time period. Human activities break for lunch between 12.00 pm until 2.00 pm and this is indicated in the reduction in load demand in that period. The activities resume after 2.00 pm until 5.00 pm and most office department will be closed after 5.00 pm. As a result, the fluctuation was decrease after 5.00 pm. The load somehow increase again after 5.00 pm and this is due to some factory starting the second shift, night shift for their workers. Many shops and shopping complexes stop their activities after 9.30 pm and all houses start to switch off their light to rest for the day, hence the load decreases after 9.30 pm.

SOLUTION APPROACH

In forecasting electricity demand with many uncertainties, fuzzy logic would be the most appropriate approach. The basic concept of fuzzy set theory was first introduced by Zadeh in 1965 (Zadeh, 1965). Fuzzy set theory can be considered as a generalization of the classical set theory. In classical set theory, an element of the universe either belongs to or does not belong to the set. Thus, the degree of association of an element is crisp. In a fuzzy set theory, the association of an element can be continuously varying. Mathematically, a fuzzy set is a mapping (known as membership function) from the universe of discourse to the closed interval {0,1}. The membership function is usually designed by taking into consideration the requirement and constraints of the problem. Fuzzy logic implements human experiences and preferences via membership functions and fuzzy rules. Due to the use of fuzzy variables, the system can be made understandable to a non expert operator. In this way, fuzzy logic can be used as a general methodology to incorporate knowledge, heuristics, or theory into controllers and decision makers. The analysis of fuzzy model was implemented using Matlab Ver-6.0 Fuzzy

7,000

8,000

9,000

10,000

11,000

12,000

13,000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47

Time in half hourly

Lo

ad

in

Meg

aw

att

s

25-Feb-06

26-Feb-06

27-Feb-06

28-Feb-06

1-Mar-06

2-Mar-06

3-Mar-06

4-Mar-06

5-Mar-06

Weekend

Weekday



17


Surabaya Indonesia

Toolbox. Some advantages of fuzzy set theory over conventional methods are as follows (Momoh et al., 1998; Bansal, 2003):

It is based on ordinary language and is conceptually easy to understand It resolves conflicting objectives by designing weights appropriate to the selected objectives It is tolerant of imprecise data and provides capability for handling ambiguity expressed in diagnostic

processes, which involves systems and causes It develops process control as a fuzzy relation between information about the conditions of the process

to be controlled. Integrating several fuzzy techniques with other techniques in electricity load demand forecasting produced

better results than individual forecast (Yang and Huang, 1998; Thang, 2004; Wang et al., 2005). Thang (2004) show that integrating hybrid method with fuzzy and neural network give better result than NN alone. Yang and Huang (1998) show that integrating fuzzy and autoregressive moving average with exogenous input variables (FARMAX) better than ARMAX and ANN. Wang et al. (2005) report that integrating Fuzzy-Rough Sets with ANN give better forecasting performance than ANN and FNN. In other study, Fu-yuan et al. (2006) propose Fuzzy Neural Network (FNN) based on combine PSO-BP training algorithm show better result than NN-BP.

A fuzzy set is fully defined by its membership functions. For most control applications, the sets that have to be defined are easily identifiable. However, for other application they have to be determined by knowledge acquisition from an expert or group of experts. Once the fuzzy sets have been establishes, one must consider their associated membership functions. How best to determine the membership function is the first question that has to be tackled. The successfulness of fuzzy application depends on a number of parameters such as fuzzy membership function (Esmin et al., 2002), fuzzy rule and also the significant input factors in the model.

However, the conventional fuzzy modeling has a drawback that fuzzy rules and the fuzzy membership function are determined by experienced knowledge or trial and error (Bagis, 2008; Mori and Kabayashi, 1996). Not only is this design method time consuming since it uses trial and error to find good fuzzy rule and membership function, it is also not guaranteed to find optimal or near optimal fuzzy rules and membership function (Bagis, 2008). That means, there is no standard method for transformation of the human knowledge or experience into the rule base of a fuzzy inference system, and no general procedure for choosing the optimal number of rules. There is a need for a good method for tuning the membership functions in order to minimize the output error measure or a maximize model performance. In this paper, we use simple statistic to define the membership functions.

used to drive the scheduling functions that determine the most economic commitment of generation sources consistent with reliability requirements, operational constraints and policies, environmental and equipment limitations; setting the spinning reserves maintenance scheduling and setting an optimal mix.

FUZZY MODEL: THE DATA There are two factors that we use to forecast next day electricity load which are temperature and load.

Temperature is important because demand of load is depending on temperature of the day. Normally when temperature is high, the demand will also high. The half-hourly temperature (oC) data obtained from The Climatologically & Hydrological Division, Malaysia Meteorological Department were data collected at the Subang Station. The half-hourly electricity load demand in Malaysia was provided by TNB in its annual demand report (2006). The data are the daily load demand data taken over a period of one week. Four factors were used to forecast the next day load. They are the current temperature (CurTemp), current load (CurLoad), previous temperature (PreTemp) and previous load (PreLoad). In this study, we used the load data from 7 March 2006 as date of previous date, 8 March 2006 as current date and the forecast load value for the 9 March 2006 for identify the rule. In other words the dates were used for training. For testing purpose, we applied load data on the 14 March 2006 as date of previous date, 15 March 2006 as current date and the forecast load value for the 16 March 2006.

Fuzzy Model: Membership Function

A number of membership functions have been proposed in the past few years, namely the triangular, trapezoidal and bell shape membership function. It is defined as a graph that defines how each point in the input space is mapped to the membership value [0,1]. The input value is often referred as the universe of discourse or universal set (u), which contain all the possible elements of concern in each particular application. The membership function chosen in this study is the triangular membership function for temperature, load input and



18


Surabaya Indonesia

load output. The triangular membership function µ(x), d1,d2,d3 :R[0,1] is specified by three parameter {d1,d2,d3} with d1<d2<d3 as (1):

30

3223

3

2112

1

10

)(

dxif

dxdifdd

xd

dxdifdd

dx

dxif

x

(1) It can also be expressed as in equation in (2):

]0),

23

3,

12

1max[min()(

dd

xd

dd

dxx

(2)

The parameter {d1,d2,d3} with d1<d2<d3 determine the x coordinates of three corners of the underlying

triangular membership function. Fig.ure 2 shows the triangular membership function with d1, d2 and d3 parameter.

FIGURE 2. Triangular membership function

Using simple statistical analysis approach on the load data and temperature for 7 and 8 March 2006 the

membership function was identified. The statistics used for load data were the minimum load, quartile 1, quartile 2 (median), quartile 3 and maximum load. The statistics used for temperature data were the minimum, quartile 2 (median) and maximum. The formula percentages of data for quartile as in (3), (4) and (5);

Total data ( n sample size) = 48 x 2 = 96

1

14

nQuartile

(3)

2 1

24

nQuartile

(4)

3 1

34

nQuartile

(5)

The load results;

µ

1

d1

d2 d3

x



19


Surabaya Indonesia

Min = 8477, Q1 = 9103.75, Q2 = 10922, Q3 = 11855 and Max = 12638. The load data are then classified into five different classes, namely the Very High, High, Medium, Low and

Very Low.

.*]5.12246(*,

2

3max

QHighVery

(6)

(7)

]5.11388,875.10012(

2

32,

2

21

QQQQMedium

(8)

]875.10012,375.8790(

2

21,

2

1

QQQMinLow

(9)

]375.8790(*,

2

1,*

QMinLowVery

(10) Figure 3 shows the triangular membership function for input and output load

FIGURE 3.Triangular membership function

The temperature results; Min = 24.3, Q2 = 26.6 and Max = 33.4. The temperature data are divided into three classes namely the Hot, Medium and Cool

(30, ]Hot

(25.45,30]Medium

( ,25.45]Cool

Figure 4 illustrate the temperature triangular membership function.

]5.12246,113885(*2

3max,

2

32

QQQHigh



20


Surabaya Indonesia

FIGURE 4. Triangular membership function for previous and current temperature

Fuzzy Model: Rules Development

One of the main characteristics of fuzzy model is the development of the rules. In this section, rules development which relates the fuzzy input and the required output are presented. Figure 5 shows the whole structure of fuzzy logic system included input, reasoning rules and also the proposed output

FIGURE 5. Fuzzy system structure The inference rules relate the input to the output and every rule represents a fuzzy relation. In this case, the

four inputs are the previous temperature, previous load, current temperature and current load. These inputs are then used to produce a single output of forecasted load. For illustration purposes, we used Tuesday (7 March 2006) for Previous Load; Wednesday (8March 2006) for Current Load, and Thursday (9 March 2006) for the forecasted day load. Table 1 shows the rules got after we list all the fuzzy parameters for half-hourly load.

TABLE (1). Daily rules for fuzzy load forecasting.

Previous Load

(PreLoad)

Previous Temperature (PreTemp)

Current Load

(CurLoad)

Current Temperature (CurTemp)

Forecast Load

Low Cool Low Cool Low

Very Low Cool Very Low Cool Low



21


Surabaya Indonesia

VeryLow Cool Very Low Cool Very Low

Low Cool Very Low Cool Very Low

Low Medium Very Low Medium Low

Low Medium Low Medium Low

Medium Medium Medium Medium Medium

High Medium High Hot High

High Hot High Hot High

Very High Hot Very High Hot Very High

Very High Hot Very High Hot High

Medium Hot Medium Hot Medium

Medium Medium Medium Hot Medium

Medium Cool Medium Medium Medium

High Cool High Medium High

Low Cool Low Medium Low Based on the rules given in Table 1, examples of rule-base model can be described as follows:

If Previous Load is MEDIUM Previous Temperature is MEDIUM And Current Load is MEDIUM Current Temperature is HOT Then Forecast Load is MEDIUM If Previous Load is VERY HIGH Previous Temperature is HOT And Current Load is VERY HIGH Current Temperature is HOT Then Forecast Load is VERY HIGH If Previous Load is LOW Previous Temperature is COOL And Current Load is LOW Current Temperature is MEDIUM Then Forecast Load is LOW

The fuzzy relation of these rules is shown in Figure 6:

FIGURE 6. Fuzzy relations function.



22


Surabaya Indonesia

COA COG

µ

i

Fuzzy Model: Defuzzification Methods

A wide variety of defuzzification methods have been described in the literature, with varying levels of

complexity. For instance, such as Center of area (COA), Center of gravity (COG), Indexed center of gravity (ICOG), Extended center of area (ECOA), Fuzzy Mean (FM), Mean of maxima (MeOM), Middle of maxima (MOM) and Last of maxima (LOM) (Leekwijek and Kerre, 1999).. COA,MOM and COG methods have been most frequently use in some papers (Broekhoven and Baets, 2006; Saade and Diab, 2000; Ruben, 2006; Leekwijek and Kerr, 1999).

Leekwijek and Kerre (1999), classify the 18 defuzzification methods into four based on a common basic; maxima methods and derivative; distribution methods and derivatives; area method; and miscellaneous method. In Wikipedia website, there list 20 defuzzification methods. In this paper, we generally categorized into the defuzzification methods into three; centroid method, area methods and maxima method. For experimental analysis, we choose COA for represent the area methods, COG represent the centroid methods and MOM and LOM represent the maxima methods.

Centroid Method

The Center of Gravity (COG) method calculates the position where the moment of the left and the right area are equal. Centre of Gravity is equivalent to Centroid of Area

∑

∑

where µx is the membership function with α and β borders

(11)

Area Method

In contrast from COG, the Center of Area (COA) method calculates the position where both areas are equal. It is also referred as to the Bisector area method

∑

∑

(12) Figure 7 illustrated the COG and COA.

FIGURE 7. Difference of COA and COG method



23


Surabaya Indonesia

Maxima Method

Maxima methods consider the fuzzy set with the degree of membership function reaching the maximum. Basically, there are three types of maxima method; First (FOM), Middle (MOM) and Last (LOM) of maxima method. The Fig.ure 8 shows the point of the three maxima method.

FIGURE 8. Difference of maxima method

EXPERIMENTAL RESULTS Experiments were conducted using all the prescribed rules described in the previous sections and the results

for the fuzzy forecasted load are given in Table 2. The fuzzy rule approach is designed to describe the input-output relationship of the actual problem in daily-life. The results are compared between COG, MOM, LOM and COA defuzzification methods.

From Table 2, it shows that the COA defuzzification method gives a lower error value with mean average percentage error equals to 1.645 while for MOM, LOM and COG was 2.58, 5.831 and 1.794 respectively.

µ

i

FOM MOM LOM



24


Surabaya Indonesia

TABLE 2. Conventional load and percentage error calculation between fuzzy forecasted and the actual load

Time Tuesday Wednesday Thursday

14-Mar-06 15-Mar-06 16-Mar-06 Center of gravity (COG)

Middle of maxima (MOM)

Last of maxima (LOM) Center of area (COA)

Hour PreLoad PreTemp CurLoad CurTemp Actual Load

Forecast Load APE

Forecast Load APE

Forecast Load APE

Forecast Load APE

0:30 9696 26.00 9748 24.75 9867 10100 2.361 9560 3.111 10400 5.402 10000 1.348 1:00 9506 25.70 9487 24.70 9730 9900 1.747 9480 2.569 10200 4.830 9840 1.131 1:30 9448 25.70 9272 24.45 9333 9720 4.147 9480 1.575 10200 9.290 9680 3.718 2:00 9126 25.70 9194 24.20 9429 9590 1.707 9480 0.541 10200 8.177 9600 1.814 2:30 9007 25.35 8853 24.20 9143 8990 1.673 9400 2.811 9920 8.498 9040 1.127 3:00 8821 25.00 8935 24.20 9061 9100 0.430 9400 3.741 9920 9.480 9200 1.534 3:30 8952 24.95 8716 24.15 8738 8820 0.938 8280 5.241 8640 1.122 8640 1.122 4:00 8774 24.90 8594 24.10 8731 8920 2.165 8,880 1.707 9920 13.618 8888 1.798 4:30 8707 24.85 8537 24.05 8795 8910 1.308 8,890 1.080 9760 10.972 8888 1.057 5:00 8620 24.80 8627 24.00 8549 8910 4.223 8,860 3.638 9520 11.358 8960 4.808 5:30 8513 24.60 8493 24.00 8675 8910 2.709 8830 1.787 9280 6.974 8960 3.285 6:00 8620 24.40 8695 24.00 8686 8910 2.579 8,850 1.888 9680 11.444 8888 2.326 6:30 8587 24.65 8687 24.10 8677 8910 2.685 8,850 1.994 9680 11.559 8888 2.432 7:00 9003 24.90 8884 24.20 8754 9050 3.381 9,320 6.466 9680 10.578 9120 4.181 7:30 8882 25.85 8771 24.70 8985 8930 0.612 8,920 0.723 10300 14.636 8960 0.278 8:00 9296 26.80 9313 25.20 9392 9800 4.344 9,480 0.937 10200 8.603 9680 3.066 8:30 10344 27.65 10190 26.35 10146 10300 1.518 10,800 6.446 11300 11.374 10400 2.503 9:00 11022 28.50 10833 27.50 10759 10600 1.478 10,800 0.381 11100 3.169 10600 1.478 9:30 11245 29.35 11352 28.75 11323 10900 3.736 10,700 5.502 11300 0.203 10800 4.619

10:00 11641 30.20 11508 30.00 11527 11200 2.837 11,800 2.368 12200 5.838 11400 1.102 10:30 11961 30.95 11749 30.90 11957 11800 1.313 11,800 1.313 12100 1.196 11800 1.313 11:00 12081 31.70 12008 31.80 12136 12100 0.297 11,800 2.769 12100 0.297 12000 1.121 11:30 12252 32.15 12165 32.20 12260 12400 1.142 11,800 3.752 12200 0.489 12300 0.326 12:00 12032 32.60 12044 32.60 12126 12200 0.610 11,800 2.688 12000 1.039 12000 1.039 12:30 12025 32.85 11994 33.40 12010 12100 0.749 11,800 1.749 12000 0.083 11900 0.916 13:00 12083 33.10 11841 34.20 11794 11800 0.051 11,800 0.051 12100 2.595 11800 0.051 13:30 12184 33.00 12030 34.50 12105 12200 0.785 11,800 2.520 12200 0.785 12100 0.041 14:00 12403 32.90 12161 34.80 12362 12400 0.307 12,400 0.307 13400 8.397 12400 0.307 14:30 12646 32.90 12360 34.75 12498 12400 0.784 12,400 0.784 13100 4.817 12400 0.784 15:00 12449 32.90 12345 34.70 12493 12400 0.744 12,300 1.545 13100 4.859 12400 0.744 15:30 12444 30.75 12387 32.90 12431 12400 0.249 12,300 1.054 13100 5.382 12400 0.249



25


Surabaya Indonesia

16:00 12492 28.60 12459 31.10 12516 12400 0.927 12,400 0.927 13500 7.862 12400 0.927 16:30 12492 26.75 12468 29.70 12407 11900 4.086 11,800 4.892 12500 0.750 11800 4.892 17:00 11924 24.90 11913 28.30 11922 11800 1.023 11,800 1.023 12000 0.654 11800 1.023 17:30 11496 25.20 11602 28.45 11438 11200 2.081 11,800 3.165 12200 6.662 11400 0.332 18:00 11077 25.50 11077 28.60 11139 10800 3.043 10,700 3.941 11400 2.343 10700 3.941 18:30 10898 25.50 10933 28.20 10800 10600 1.852 10,700 0.926 11400 5.556 10600 1.852 19:00 10694 25.50 10727 27.80 10782 10500 2.615 10,700 0.761 11400 5.732 10600 1.688 19:30 11184 25.45 11177 27.75 11102 10800 2.720 10,700 3.621 11400 2.684 10800 2.720 20:00 11510 25.40 11601 27.70 11368 11100 2.357 11,800 3.800 12200 7.319 11400 0.281 20:30 11603 25.40 11592 27.30 11309 11100 1.848 11,800 4.342 12200 7.879 11400 0.805 21:00 11272 25.40 11530 26.90 11142 11000 1.274 11,800 5.906 12300 10.393 11000 1.274 21:30 11200 25.20 11156 26.95 11101 10800 2.711 10,800 2.711 11300 1.793 10800 2.711 22:00 10824 25.00 11029 27.00 10839 10600 2.205 10,800 0.360 11200 3.331 10600 2.205 22:30 10552 25.05 10718 27.40 10657 10500 1.473 10,800 1.342 11200 5.095 10600 0.535 23:00 10452 25.10 10498 27.80 10541 10400 1.338 10,800 2.457 11200 6.252 10500 0.389 23:30 10253 24.95 10446 27.45 10361 10400 0.376 10,800 4.237 11200 8.098 10500 1.342

0:00 9813 24.80 10057 27.10 10041 10100 0.588 9400 6.384 10000 0.408 10000 0.408

Total

86.131

123.833

279.873

78.945

MAPE

1.794

2.580

5.831

1.645

Based on this MAPE values it is concluded that fuzzy approach using COA defuzzification method is more effective and gives a better forecast

accuracy. Figure 9 shows the comparison plot of forecast load from four different defuzzification methods with actual load.

FIGURE 9. Comparison defuzzification method

8000

9000

10000

11000

12000

13000

14000

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47

Load

(Meg

awat

t)

Actual Load COG MOM LOM COA



26


Surabaya Indonesia

CONCLUSION The major contribution of this paper is to propose the used of new approach to forecast the daily load

demand in a shorter period of time to half-hourly load demand. This has greatly benefited TNB in their management of the daily operation of electricity demand. The key features of the proposed methodology is the development of fuzzy logic approach to solving forecasting problems with uncertainty data such as day types, temperature and different load pattern. A new set of rules developed for this purpose has highlighted the importance of information in model building and this study has also demonstrated the application of triangular membership function and applied simple statistic to identify membership function parameter. As with any new approach for forecasting load demand, there are some areas for possible future improvement. In this study, comparison between four methods of defuzzification, shows that COA gave a better forecast values than MOM, LOM and COG (see Table 2). Further improvement in the forecast may be done using different new defuzzification method. This study finally show that fuzzy approach not only gives better forecasting performance but it has a simple procedure to handle the forecasting

ACKNOWLEDGMENTS This research was supported by the Ministry of Science, Technology and Innovation, Malaysia (MOSTI) under IRPA Grant Vote No. 4F014 and the Department of Mathematics, Faculty of Science, University Technology Malaysia. These supports are gratefully acknowledged. The authors would like to thank lecturers and friends for the helpful ideas and discussion.

REFERENCES 1. Abraham & B. Nath, A Neuro-Fuzzy Approach for Modelling Electricity Demand in Victoria. Applied Soft

Computing 1, pp. 127–138. (2001). 2. A. Esmin. et. al. ―Particle swarm optimization for fuzzy membership function optimization,‘‘ in IEEE

International Conference on System, Man and Cybernetics, 6-9 October 2002, (2002). 3. Bagis, Determining Fuzzy Membership Functions with Tabu Search—An Application to Control. Fuzzy

Sets and Systems 139, pp. 209-225. (2003). 4. Bagis, Fuzzy Rule Base Design using Tabu Search Algorithm for Nonlinear System Modeling. ISA

Transactions 47, pp. 32–44. (2008), 5. L. Bowerman. et. al. Forecasting Time Series, and Regression. An Applied. Approach. Fourth edition.

(Thomson Brooks/Cole, USA, 2005). 6. K. Ranaweera, et al. Fuzzy Logic for Short Term Load Forecasting. International Journal of Electrical

Power & Energy Systems 18(4), pp. 215-222. (1996). 7. Karaboga et al. Controlling Spillway Gates of Das by Using Fuzzy Logic Controller with Optimim Rule

Number. Applied Soft Computing 8, pp. 232-238. (2008). 8. S. Karapidakis, Machine Learning for Frequency Estimation of Power Systems. Applied Soft Computing,

7(1), pp. 105-114. (2007). 9. E. V. Broekhoven and B. D. Baets, Fast and Accurate Center of Gravity Defuzzification of Fuzzy System

Outputs Defined on Trapezoidal Fuzzy Partitions. Fuzzy Sets and Systems 157, pp. 904 – 918. (2006). 10. A. Razak et al. (2006, 28 - 30 August 2006). Forecasting the Maximum Demand of Electricity in Malaysia.

Paper presented at the International Conference on Energy and Environment 2006 (ICEE 2006), Universiti Tenaga Nasional, Malaysia.

11. Tsekourasa et al. A Simple Algorithm for Training Fuzzy Systems using Input–Output Data. Advances in Engineering Software 34, pp. 247-259. (2003).

12. G.W. Nurcahyo and S. M. Shamsuddin, Selection of Defuzzification Method to Obtain Crisp Value for Representing Uncertain Data in a Modified Sweep Algorithm. The Journal of Computer Science and Technology 3(2), pp. 22-28. (2003).

13. H. Fu-yuan, et al. A Modified Particle Swarm Algorithm Combined with Fuzzy Neural Network with Application to Financial Risk Early Warning. Paper presented at the Proceedings of the 2006 IEEE Asia-Pacific Conference on Services Computing (APSCC'06), (2006).

14. H. Mori and H. Kobayashi. Optimal Fuzzy Inference for Short -term Load Forecasting. IEEE Transactions on Power System 11(1), pp. 390 – 396.(1996).

15. H. Yang and C. Huang, A New Short-Term Load Forecasting Approach using Self –Organizing Fuzzy ARMAX Model. IEEE Transaction on Power System 13, pp. 217-225. (1998).



27


Surabaya Indonesia

16. J. A. Momoh. et al. Overview and Literature Survey of Fuzzy Set Theory in Power Systems”. IEEE Transactions on Power Systems 10(3), pp. 1676-1690. (1995).

17. J. J. Saade and H. B. Diab. Defuzzification Techniques for Fuzzy Controllers. IEEE Transactions on Systems, Man, and Cybernetic - Part B: Cybernetics 30(1), pg 223-229. (2000).

18. J. W. Taylor, Exponential Smoothing with a Damped Multiplicative Trend. International Journal of Forecasting 19, pp. 715-725. (2003).

19. J. W. Taylor et al. A Comparison of Univariate Methods for Forecasting Electricity Demand Up to a Day Ahead. International Journal Of Forecasting, 22(1), pp. 1-16. (2006)

20. K. F. Thang, MATLAB implementation of Neural & Neuro-Fuzzy Approaches for Short-Term Electricity Demand Forecasting, International Conference on Power System Technology. 21-24 November. Singapore. (2004).

21. K. Ozawa et al, Fuzzy time-series model of electric power consumption. Paper presented at the IEEE Canadian Conference on Electrical and Computer Engineering, (1999).

22. L. A. Zadeh, Fuzzy Sets. Information and Control 8, pp. 338-353. (1965). 23. M. S. Owayedh et al. "Identification of temperature and socialevents effects on weekly demand

behavior,"Proceedings of the Power Engineering Society Summer Meeting, 2000. IEEE , 2000, 4(2000), pp. 2397-2402

24. N. Mohamed et al. Comparing Forecasting Performances between Multilayer Feed Forward Neural Network and Recurrent Neural Network in Malaysia’s Load. Journal Of Interdisciplinary Mathematics 13(2), pp. 125-134. (2010).

25. N.O. Rubens, The Application of Fuzzy Logic to the Construction of the Ranking Function of Information Retrieval System. Computer Modeling And New Technologies 10(1), pp. 20-27. (2006).

26. P.A. Mastorocostas et al. A Hybrid Fuzzy Modeling Method for Short-Term Load Forecasting. Mathematics and Computers in Simulation 51, pp. 221-232. (2000).

27. R. C. Bansal, Bibliography on the Fuzzy Set Theory Applications in Power Systems (1994–2001). IEEE Transactions On Power Systems 18(4), pp. 1291-1299. (2003).

28. S. C. Pandian et al. Fuzzy Approach for Short Term Load Forcasting. Electric Power Systems Research, 76, pp. 541-548. (2006),

29. S. Mirasgedis, et al. Models for Mid-Term Electricity Demand Forecasting Incorporating Weather Influences. Energy 31(2-3), pp. 208-227. (2006).

30. S. Saab et al. Univariate Modeling and Forecasting of Energy Consumption: The Case Of Electricity In Lebanon. Energy 26(1), pp. 1-14. (2001).

31. T. E. Dielman , Applied Regression Analysis. (Fourth Edition, Thomson. USA, 2005). 32. T. Frantti and P. Mahonen, Fuzzy Logic-Based Forecasting Model. Engineering Applications of Artificial

Intelligence 14(2), pp. 189-201. (2001). 33. W. V. Leekwijck and E. E. Kerr, Defuzzification: Criteria and Classification. Fuzzy Sets and Systems,

108, pp. 159-178. (1999), 34. Y. S. Tarng et al. Genetic Synthesis of Fuzzy Logic Controllers in Turning. Fuzzy Sets and Systems 83, pp.

301-310. (1996). 35. Z. Ismail et al. Forecasting Peak Load Electricity Demand using Statistics and Rule Based Approach.

American Journal Of Applied Sciences 6 (8), pp. 1618-1625. (2009) 36. Z. Ismail and F. Jamaluddin (2009). ―Rule-based Forecasting: Malaysian Electricity Company Jouney‖.

The Journal of Business Forecasting, 28(1), (2009), pp. 23-33. 37. Z. Ismail and F. Jamaluddin. ―Rule-based Forecasting: Malaysian Electricity Company Journey‖. The

Journal of Business Forecasting, 28(1), (2009), pp. 23-33. 38. Z. Ismail and R. Efendy, ― Enrollment forecasting based on modified weight fuzzy time series‖. Journal of

Artificial Intelligence, 4(1), (2011), pp. 110-118. 39. Z. Pan et al. A neural-Fuzzy Syste for Forecasting. Paper presented at the The Thirtieth Annual Hawwaii

International Conference on system Sciences, Hawwaii, USA. (1997). 40. D. F. Matteo et. al. Electricity Demand Forecasting over Italy: Potential Benefits using Numerical Weather

Prediction Models. Electric Power System Research 104, pp. 71-79. (2013). 41. Z. Y. Wang et al. (2005, Nov. 29 2005-Dec. 2 2005). A new method for short-term load forecasting

integrating fuzzy-rough sets with artificial neural network. Paper presented at the The 7th International Power Engineering Conference IPEC 2005. (2005).

42. Ismail, Z., et al.,‖Application of Fuzzy Time Series Approach in Electric Load Forecasting‖, New Mathematics and Natural Computation‖, Volume 11, Issue 3, 28 November 2014.

https://uummail.uum.edu.my/owa/redir.aspx?C=6eb086caae084892be7ad8c460352e91&URL=http%3a%2f%2fwww.scopus.com%2frecord%2fdisplay.url%3feid%3d2-s2.0-77951005522%26origin%3dresultslist



28


Surabaya Indonesia

ORAL PAPER OF ANALYTICAL AND FORENSIC CHEMISTRY (OAFC)



29 Jabbar et al.


Surabaya Indonesia

Determination of Thiamine Hydrochloride Using Flow Injection

Spectrophotometric Method

Hijran Sanaan Jabbar1,* and Azad Tawfiq Faizullah2

1,2 Department of Chemistry, College of Science, Salahaddin University, Erbil, Iraq *Corresponding author’s Email: [email protected]

Abstract: A flow injection-spectrophotometric method for determination of thiamine hydrochloride (vitamin B1) has been developed. An intense yellow (421.5 nm) and orange (459.5 nm) colour is formed by coupling of vitamin B1 with diazotized 2-aminobenzoic acid (2-ABA) and 4-aminobenzoic acid (4-ABA) in alkaline medium respectively. Calibration graph with diazotized 2-ABA give linear range from 0.6 to 22μg/mL with correlation coefficient of 0.9983, while diazotized 4-ABA give linear range from 0.1 to 20μg/mL with correlation coefficient of 0.9985. Possible interferences that related to the determination of thiamine hydrochloride in pharmaceutical formulations were studied and the results showed that foreign species caused less than ±5.0% error. Therefore, the method was successfully applied for determination of thiamine hydrochloride in pharmaceutical formulations and a comparison take place between the results obtained with these two reagents.

INTRODUCTION Thiamine (vitamin B1) (Figure 1) is an essential nutrient for humans to prevent beriberi and it's necessary for

carbohydrate metabolism and for the maintenance of neural activity [1]. People usually absorb vitamin B1 from natural and fortified foods. When needed, the vitamin can also be obtained from various pharmaceutical preparations containing thiamine [2]. It is built up from a substituted pyrimidine ring and a substituted thiazole ring; both rings are coupled through a CH2 group. Salts are formed on the amino group of the pyrimidine ring and on the N atom of the thiazole ring [3].

NH2

N+

SOH

N

N

Cl-

Thiamine hydrochloride Figure 1. Structure of thiamine hydrochloride (vitamin B1).

Several analytical methods have been proposed for the determination of thiamine since its discovery and isolation, such as UV/Vis. spectrophotometry [4-9], spectrofluorometry [10-12], high performance liquid chromatography (HPLC) [13-17], chemiluminescence [18-22], gas chromatography (GC) [23], voltametry [24-26], potentiometry [27], amperometry [28] and polarography [29].

A flow-injection (FI) spectrophotometric procedure is proposed for the determination of thiamine hydrochloride in multivitamin preparations. Powdered sample was previous dissolved in 0.1 mol/L hydrochloric acid, and a volume of 250μL was injected directly into a carrier stream of 0.10% (w/v) potassium

hexacyanoferrate(III) in 0.5 mol/L sodium hydroxide at a flow rate of 2.46 ml/min. The thiochrome produced in the oxidation of thiamine hydrochloride by potassium hexacyanoferrate(III) in alkaline solution was directly measured at 369 nm [30].

In this work, a FI-spectrophotometric system is proposed for the determination of thiamine hydrochloride in pharmaceutical formulations. Two different diazotized compounds (2-Aminobenzoic acid and 4-aminobenzoic acid) used as reagent for the quantitative determination of vitamin B1 after coupling with thiamine to give an intense colour in alkaline medium. A comparison takes place between the results obtained with two reagents.



30 Jabbar et al.


Surabaya Indonesia

EXPERIMENTAL

Apparatus

The schematic diagram of the flow injection system used in this work is shown in Figure 2. It consists of a peristaltic pump (DESAGA Heidelberg, with 6 channels and variable speed up to 10 ml/min) to deliver flow streams. The silicon rubber pump tubes with (1.4 mm i.d) were used to transport the solutions. A rotary valve (Rheodyne U.S.A.) with variable sample volume were used to inject the sample into flowing carrier stream. The valve was made of a polytetra-flouroethylene (PTFE) with good resistance against the corrosion of chemicals. It contains grooves with internal diameter of (0.5 mm). A Y-shaped perspex piece was used to mix two streams of reagents. A JENWAY 6405 UV/Vis. spectrophotometer was used. The out-put was recorded by mean of x-t recorder (Type PM 825A PHILIPS – one line recorder) with various amplication factors and different chart speeds.

Figure 2. Schematic diagram of the FI-Spectrophotometric manifold used for the determination of vitamin B1.

Reagents

Distilled water was used in all preparations. The reagents used were of analytical grade. Sodium nitrite solution: A stock solution of 0.1 mol/L of NaNO2 (Fluka) prepared daily by dissolving

3.45g of sodium nitrite in a little of distilled water and then the volume completed to 500 ml by distilled water [31, 32]. Working solutions prepared by appropriate dilutions with distilled water.

Hydrochloric acid solution: 1.0 mol/L hydrochloric acid solution was prepared by diluting 85.92 ml of HCl (36%, Sp.gr. =1.18) (Gainland Chemical Company U.K. - GCC) to 1.0L with distilled water. Other solutions were prepared by serial dilution in 100 ml volumetric flasks.

2-Aminobenzoic acid and 4-Aminobenzoic acid solutions: 2-Aminobenzoic acid (anthranilic acid) (2-ABA) and 4-aminobenzoic acid (4-ABA) (M.Wt. 137.14) solutions (0.1 mol/L) were prepared by dissolving of 6.857 g of 2-Aminobenzoic acid (Fluka) or 4-aminobenzoic acid (Fluka) in 500 ml of distilled water. Working solutions were prepared by suitable dilution.

Potassium hydroxide solution: 1.0 mol/L KOH was prepared by dissolving 56.11g of potassium hydroxide (RIEDEL-DE HAEN) in a little of water; the volume was completed to 1.0L in a volumetric flask.

Stock solution of vitamin B1: Thiamine hydrochloride stock solution (1000 μg/mL) was prepared by dissolving 0.1g of thiamine hydrochloride (Fluka) in distilled water and diluted to 100 ml. Working standard solutions were prepared daily by appropriate dilution of the stock solution with distilled water.

Sample preparation

Tablets: Twenty tablets containing thiamine hydrochloride were accurately weighed, ground and powdered. An amount of this powder equivalent to about 0.10 g of thiamine hydrochloride was dissolved in distilled water. They were transferred into 100 ml volumetric flask, and diluted to the mark with distilled water. The content of the flask was shaken mechanically and then filterated. A 10 ml filterate further diluted with water to adjust the concentration to meet the requirement of the adopted experimental conditions.



31 Jabbar et al.


Surabaya Indonesia

Injections: The contents of 5 ampoules were mixed and an accurately measured volume equivalent to 200 mg of thiamine hydrochloride was transferred into a 100 ml calibrated flask. The volume was made up to the mark with distilled water. More dilute solutions were prepared by appropriate dilutions.

General procedure

The FIA system in Figure 2 was operated. Nitrous acid is produced by mixing of sodium nitrite and hydrochloric acid in the first mixing coil (L1) which is then merged with 2-aminobenzoic acid (or 4-aminobenzoic acid) and allowed to form diazonium salt in the second mixing coil (L2). Then 90μL of sample

solution (thiamine hydrochloride) injected to the diazonium salt stream to form the azo dye. Finally, this stream merged with potassium hydroxide solution (L4) to form stable yellow complex when 2-aminobenzoic acid or orange complex when 4-aminobenzoic acid used. The absorbances (A) at 421.5 nm for 2-aminobenzoic acid and at 459.5 nm for 4-aminobenzoic acid were monitored and the absorbance measured (mV). At least three injections were made for every sample solution. The concentration of the analyte was measured using the calibration curves of absorbance (mV) versus the concentration of vitamin B1 (μg/mL) obtained from reference solutions under the same working conditions.

RESULTS AND DISCUSSION The possible mechanism for the reaction of diazotized 2-aminobenzoic acid (or 4-aminobenzoic acid) with

vitamin B1 was explained in Figure 3. The first step is formation of the azo dye and the second step is coupling of diazonium salt with vitamin B1 in alkaline medium to form an intense colour complex which is soluble in water.

C O

OH

N N

N

N CH3

NH2

R

O

CHO N

N

N CH3

NH2

R

N

+

Thiaminehydrochloride

orange colour

max = 459.5nm

C O

OH

NH2

4-aminobenzoic acid

+ NO2- + H+

C O

OH

N N

+ H2O

R = C7H12ONSCl

1-

2-

Figure 3. Possible mechanism of the diazotization reaction.

Absorption spectra

When dilute aqueous solution of thiamine hydrochloride (5.0μg/mL) and diazotized 2-ABA or 4-ABA reagent solution are mixed in alkaline medium, an intense yellow or orange azo dye respectively forms immediately. The intense azo dyes formed show maximum absorption at 421.5 nm for 2-ABA and at 459.5 nm for 4-ABA (Figure 4), in contrast to the reagent blank, which gives almost zero absorption in the region (400 – 550 nm). The wavelength of maximum absorption at 421.5 nm for 2-ABA and at 459.5 nm for 4-ABA, characteristic of the azo dyes were used in all subsequent experiments.



32 Jabbar et al.


Surabaya Indonesia

Figure 4. Absorption spectra of 5.0 μg/mL thiamine hydrochloride, treated as described under general procedure and

measured against blank.

Optimization of experimental parameters

Effects of various parameters on the absorbance of the azo dye were studied and the reaction conditions were optimized. These optimizations started using 0.01 mol/L HCl, 0.005 mol/L NaNO2, 0.001 mol/L 2-aminobenzoic acid (or 4-aminobenzoic acid), 0.2 mol/L KOH, coils length (L1=20cm, L2=30cm, L3=0cm and L4=50cm), 2.0 ml/min flow rate, 90μL sample volume and 100μL volume of the flow-through cell at maximum wavelength 421.5 nm for 2-aminobenzoic acid and 459.5 nm for 4-aminobenzoic acid.

Chemical optimizations

Effect of hydrochloric acid concentration. The effect of hydrochloric acid concentration on the diazotization reaction was studied in the range of 0.002 – 0.04 mol/L hydrochloric acid. As shown in Figure 5 maximum diazotization was obtained with 0.028 and 0.013 mol/L hydrochloric acid for 2-ABA and 4-ABA respectively. Higher HCl concentration weakens the alkaline medium required for obtaining the colour of azo dye and at lower HCl concentration the production of nitrous acid is not enough to form intense colour of the azo dye. Therefore, these concentrations were considered for further studies.

Figure 5. Effect of hydrochloric acid concentration on the absorbance of 25μg/mL vitamin B1 using 2-ABA and 20μg/mL

vitamin B1 using 4-ABA.

0

20

40

60

80

100

120

140

160

0 0.01 0.02 0.03 0.04 0.05

Abs

orba

nce

(mV

)

Hydrochloric acid concentration (mol/L)

2-ABA4-ABA



33 Jabbar et al.


Surabaya Indonesia

Effect of NaNO2 concentration. The effect of sodium nitrite concentration was investigated by varying the nitrite concentration from 0.001 – 0.02 mol/L. The results are shown in Figure 6 which indicate that maximum absorbances were obtained at 0.012 and 0.009 mol/L nitrite for 2-ABA and 4-ABA respectively. Further increase of nitrite concentration produce unstable and noisy signals due to possibility of generating nitrogen gas. The excess of nitrous acid must be removed in the manual method and in the air-segmented continuous flow method, because of its slow reaction with diazotized compound and the azo dye produced. In designing the flow injection manifold, this step can be eliminated by keeping nitrite at low concentration and the total reaction time short enough, to minimize the destruction of azo dye [31]. Therefore, these concentrations were used in further studies. From the optimization of HCl and NaNO2 concentrations, we concluded that the ratio of HCl to NaNO2 is about (2:1) were agreed with that required for the stoichiometric equilibrium in the mechanism.

0

20

40

60

80

100

120

140

0 0.005 0.01 0.015 0.02 0.025

Abs

orba

nce

(mV

)

Sodium nitrite concentration (mol/L)

2-ABA

4-ABA

Figure 6. Effect of sodium nitrite concentration on the absorbance of 20μg/mL vitamin B1 using 2-ABA and 4-

ABA.

Effect of reagent concentration. The effect of 2-aminobenzoic acid and 4-aminobenzoic acid concentration upon the analytical response of the flow system was examined in the concentration range of

1.0×10–4

to 2.0×10–3

mol/L. Maximum response was obtained at (1.5×10–3

mol/L) 2-ABA and (1.0×10–3

mol/L) 4-ABA as shown in Figure 7. Therefore, these concentrations were employed in subsequent experiments.

Figure 7. Effect of 2-ABA and 4-ABA concentrations on the absorbance of 20μg/mL vitamin B1.

0

20

40

60

80

100

120

140

0 0.0005 0.001 0.0015 0.002 0.0025 0.003

Ab

sorb

ance

(m

V)

Reagent concentration (mol/L)

2-ABA

4-ABA



34 Jabbar et al.


Surabaya Indonesia

Effect of potassium hydroxide concentration. A stable yellow or orange complex was formed among mixing of thiamine hydrochloride with diazotized 2-aminobenzoic acid or 4-aminobenzoic acid respectively in basic medium, while in acidic or neutral medium these complexes are not formed. Therefore, some different bases (NaOH, KOH, Na2CO3) were tested to obtain maximum absorbance. Under the same conditions, it was found that the best results were obtained with potassium hydroxide. Therefore, potassium hydroxide was selected as the best reaction medium.

The effect of potassium hydroxide concentration was studied in the range of 0.01 – 0.4 mol/L KOH. Maximum absorbance was obtained at 0.15 and 0.1 mol/L KOH for 2-ABA and 4-ABA respectively (Figure 8). This indicates that at lower KOH concentration a complete coupling between diazotized 2-ABA or 4-ABA and vitamin B1 does not occur and an intense colour complex is not produced. Therefore, 0.15 and 0.1 mol/L were selected.

0

20

40

60

80

100

120

140

160

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45

Abs

orba

nce

(mV

)

Potassium hydroxide concentration (mol/l)

2-ABA4-ABA

Figure 8. Effect of KOH concentration on the absorbance of 20μg/mL vitamin B1 using 2-ABA and 4-ABA.

Physical optimizations

Length of the mixing coil. Using the optimized reactant concentrations, effect of different mixing coil

lengths for the nitrous acid formation, diazotization and coupling reactions illustrated in Figure 2 were also optimized. Their optimum values shown in Table 1 and Figs. 9 and 10.

Table 1. Optimum values of length of the mixing coils.

Coil Using 2-ABA Using 4-ABA

L1 20 cm 30 cm

L2 10 cm 10 cm

L3 0 cm 0 cm

L4 10 cm 10 cm



35 Jabbar et al.


Surabaya Indonesia

100

110

120

130

140

150

160

170

180

190

200

0 10 20 30 40 50 60 70

Abs

orba

nce

(mV

)

Coil length (cm)

L1L2L3L4

Figure 9. Effect of mixing coil length on the absorbance of 20μg/mL vitamin B1 using 2-ABA.

70

80

90

100

110

120

130

140

0 10 20 30 40 50 60 70

Abs

orba

nce

(mV

)

Coil length (cm)

L1L2L3L4

Figure 10. Effect of mixing coil length on the absorbance of 20μg/mL vitamin B1 using 4-ABA.

Effect of flow rate. The effect of flow rate was tested in the range of 0.5 – 3.0 ml/min in order to obtain a maximum absorbance signal. Figure 11 illustrated that the sensitivity increased up to 1.0 ml/min for 2-ABA and 1.5 ml/min for 4-ABA and decreased at higher flow rates, because the reactants were reached to the detector at a shorter time which is not sufficient for completing the reaction. Therefore, flow rates of 1.0 and 1.5 ml/min for 2-ABA and 4-ABA respectively were selected.



36 Jabbar et al.


Surabaya Indonesia

0

20

40

60

80

100

120

140

160

180

0.5 1 1.5 2 2.5 3 3.5 4

Abs

orba

nce

(mV

)

Flow rate (ml/min)

2-ABA4-ABA

Figure 11. Effect of flow rate on the absorbance of 20μg/mL vitamin B1 using 2-ABA and 4-ABA.

Effect of sample volume. The effect of the sample volume was examined between 50 - 175μL. The absorbance increased with increasing sample volume up to 150 μL, after this volume the signal remain constant

as shown in Figure 12. Increasing sample volume lead to an increase in the absorbance but increases the time for the peak to return to the base line and resulted in the formation of a broad peak. Therefore, 125 μL sample

volume was chosen for further work.

0

10

20

30

40

50

60

70

80

90

40 60 80 100 120 140 160 180

Abs

orba

nce

inte

nsit

y (m

V)

Sample volume (μL)

2ABA4ABA

Figure 12. Effect of sample volume on the absorbance of 5.0μg/mL vitamin B1.

The summaries of optimum chemical and physical conditions for the determination of vitamin B1 using FIA-

Spectrophotometric method were illustrated in Table 2 using diazotized 2-ABA and 4-ABA.



37 Jabbar et al.


Surabaya Indonesia

Table 2. Optimum chemical and physical conditions for the determination of thiamine hydrochloride.

Parameters Optimum value

Using 2-aminobenzoic acid Using 4-aminobenzoic acid Hydrochloric acid 0.028 mol/L 0.013 mol/L Sodium nitrate 0.012 mol/L 0.009 mol/L Diazotized compound 0.0015 mol/L 0.001 mol/L Potassium hydroxide 0.15 mol/L 0.10 mol/L Sample volume 125 μL 125 μL

Mixing coil length L1=20 cm, L2=10 cm L3=0cm, L4=10 cm

L1=30 cm, L2=10 cm L3=0cm, L4=10 cm

Flow rate 1.0 ml/min 1.5 ml/min Flow-through cell 100 μL 100 μL λ max 421.5 nm 459.5 nm

Calibration graph

Under optimal experimental conditions described in Table 2, the calibration curve was constructed by plotting the absorbance (mV) against the concentration of vitamin B1 (μg/mL) using diazotized 2-aminobenzoic acid and 4-aminobenzoic acid (Figure 13). The analytical parameters related to these calibrations are summarized in Table 3. Two different pure vitamin B1 concentrations are injected ten times to determine the accuracy and precision of the method using diazotized 2-ABA and 4-ABA as reagent. These results are shown in Table 4.

y = 14.591x - 7.9815r = 0.9983

y = 15.494x - 0.9791r = 0.9985

0

50

100

150

200

250

300

350

0 2 4 6 8 10 12 14 16 18 20 22 24

Ab

sorb

ance

(m

V)

Vitamin B1 concentration (µg/mL)

2ABA4ABA

Figure 13. Calibration graph for the determination of vitamin B1 using FI-Spectrophotometric method.

Table 3. Analytical characteristics for the determination of vitamin B1.

Diazotized compound Linear range

(μg/mL) D. L.

a

(μg/mL) R

b N

c

2-aminobenzoic acid 0.6 – 22 0.6 0.9983 19

4-aminobenzoic acid 0.1 – 20 0.1 0.9985 21 a Detection limit.

b Correlation coefficient.

c Number of measurements.



38 Jabbar et al.


Surabaya Indonesia

TABLE 4. Accuracy and precision of the method.

Reagent Conc. of vitamin B1 (μg/mL)

Mean of absorbance (mV) S.D. R.S.D% %E

2-ABA 6 79.75 0.50 0.62 0.21 14 191 0.81 0.42 -2.59

4-ABA 9.0 139.83 0.28 0.20 0.97 20 300.33 0.57 0.19 -2.76

S.D.: Standard Deviation; R.S.D.: Relative Standard Deviation; %E: Relative error

Interferences

Possible interferences that related to the determination of thiamine hydrochloride in pharmaceutical formulations were studied by injecting synthetic solutions containing pure drug solution and varying amounts of the interfering compounds. The selectivity of the proposed procedure was examined under the experimental conditions employed in Table 2. As shown in Table 5 no interferences in the flow injection spectrophotometric system procedure were observed up to 100-fold excess of the most interfering compounds. Magnesium streate interfered at concentrations above 10 fold excess. In addition relative errors in most cases were less than ±5.0%.

Table 5. Effect of interference on the absorbance of vitamin B1.

Interference Fold

excess

Using (16μg/mL) vit.B1 with 2-ABA

Using (15μg/mL) vit.B1 with 4-ABA

Absorbance (mV) %E Absorbance (mV) %E Glucose 100 229 1.56 227 -1.91

Galactose 100 228 1.12 228 -1.48 Lactose 100 230 2.01 225 -2.77 Fructose 100 227 0.67 228.5 -1.26 Starch 100 228 1.12 229 -1.04

Sucrose 100 226 0.23 227.33 -1.77 Magnesium streate 10 236 4.67 237 2.40

Mixture of all interferences

100 224.5 -0.43 229.5 -0.83

Application

The present FIA-spectrophotometric method for determination of thiamine hydrochloride was applied to various pharmaceutical preparations. A comparison take place between the results obtained with two different reagents (Table 6). In all cases the data were in good agreement with the labeled amounts. The results of proposed method using diazotized 2-ABA and 4-ABA are compared applying paired t-test and F-test [33, 34], it was found that all results are in agreement at the 95% confidence level and within an acceptable range of errors (Fexp. = 5.0139 < F-table = 19.00) and (texp.= 1.05376 < ttable= 4.30 for 95%.).

Table 6. Results of analysis of commercial drug formulations containing thiamine hydrochloride by the proposed methods and comparison between them.

Formulation Company Composition Amount

nominal (mg per tablet or ml)

Drugs found (mg) per tablet

or ml a %E

Diazotized 2-ABA

Diazotized 4-ABA

Vitamin B1 (tablet)

Darou Pakhsh Pharmaceutical Mfg. Co. - Iran

Thiamine hydrochloride

300 308.38 299.44 2.98

SAMAVIT (tablet)

Sammara drug industry - Iraq

(S.D.I.)


100 95.94 96.8 -0.88

Vitamin B1 (injection)

Hemofarm - Vrsac, Yougoslavie


100 100.35 98.95 1.42 a Average of three measurements (n=3).



39 Jabbar et al.


Surabaya Indonesia

CONCLUSIONS A simple and sensitive FIA-Spectrophotometric method for the determination of thiamine hydrochloride

(vitamin B1) in pure and pharmaceutical formulations has been devised, based on coupling of it with diazotized 2-aminobenzoic acid (anthranilic acid) or 4-aminobenzoic acid in alkaline medium. Low cost and reagent consumption are found among other interesting features of the procedure which make it applicable to drug analysis and quality control of pharmaceuticals. Batch method used for the determination of mole ratio of vitamin B1 to diazotized reagents (2-ABA and 4-ABA). The absorbance measured and plotted against the mole fraction of thiamin hydrochloride. The mole ratio of vitamin B1 to the reagent was found to be 1:1.

The proposed method is simple, precise, accurate, and has high analytical frequency (72 samples/h) when compared with spectrophotometric method that depends on the coupling with diazotized sulfanilic acid (3 samples/h) [6] or comparing with spectrofluoremetric methods (25 samples/h) [1], (30 samples/h)

[2] and (52

samples/h) [11]. The proposed method is comparable in sensitivity to many of the existing FIA-spectrophotometric methods that depend on the oxidation of thiamine hydrochloride by potassium hexacyanoferrate(III) in alkaline solution which give linear range between 2.5 – 50 μg/mL with RSD less than 1.0% [30]. The procedure is free from tedious steps like extraction or heating and involves least number of experiment variables, which is reflected in high precision.

ACKNOWLEDGEMENTS The researchers would like to thank Science College and Dr. Rebwar O. Hassan for their kind help.

REFERENCES 1. T. P. Ruiz, C. M. Lozano, A. Sanz and A. Guillen; J. Pharm. Biomed. Anal., 34, (2004) 551. 2. H. Zuh, H. Chen and Y. Zhou; Anal. Sci., 19, (2003) 289. 3. L. R. Engelking; "Textbook of Veterinary Physiological Chemistry", 3rd Edition, Elsevier Inc., USA

(2015), p. 260–264. 4. D. Melnick and H. Field; Upjohn Fellow in Clinical Research, 39, (1937) 505. 5. H. J. Prebluda and E. V. McCollum; J. Bio. Chem., 127, (1939) 495. 6. R. O. Hassan and Y. J. Azeez; Tikrit J. Pharm. Sci., 1 (2), (2005) 1. 7. M. Q. Al Abachi and H. Hadi; Journal of Pharmaceutical Analysis, 2 (5), (2012) 350-355. 8. B. S. Sroka; Journal of Analytical Chemistry, 68 (3), (2013) 218-222. 9- Khairia Mohammed Al‐Ahmary; European Journal of Chemistry, 5 (1), (2014) 81‐84. 10. J. M. Calatayud, C. G. Benito and D. G. Gimenez; J. Pharm. Biomed. Anal., 8 (8-12), 667 (1990). 11. J. L. Flores, M. L. F. Cordova and A. M. Diaz; Anal. Chim. Acta, 535, (2005) 161. 12. G. D. Tarigh and F. Shemirani; Talanta, 123, (2014) 71-77. 13. D. Bohrer, P. C. do Nascimento, A. G. Ramirez, J. K. A. Mendonca, L. M. de Carvalho and S. C. G.

Pomblum; Microchem. J., 78, (2004) 71. 14. F. Batifoulier, M. A. Verny, C. Besson, C. Demigne and C. Remesy; J. Chromatography B, 816, (2005)

67. 15. X. Tang, D. A. Cronin and N. P. Brunton; J. Food Composition and Analysis, 19 (8), (2006) 831. 16. R. S. J. Rodriguez, V. F. Ruiz, M. Camara and M. C. Sanchez-Mata; Journal of Cereal Science, 55 (3),

(2012) 293-299. 17. P. Jin, L. Xia, Z. Li, N. Che, D. Zou and X.in Hu; Journal of Pharmaceutical and Biomedical Analysis,

70, (2012) 151-157. 18. N. Grekas and A. C. Calokerinos; Talanta, 37 (11), (1990) 1043. 19. Z. Song and S. Hou; J. Pharm. Biomed. Anal., 28 (3-4), (2002) 683. 20. J. Du, Y. Li and J. Lu; Talanta, 57 (4), (2002) 661. 21. A. Wasielczuk, M. C. Icardo, J. V. G. Mateo and J. M. Calatayud; Anal. Lett., 37, (2004) 3205. 22. M. Kamruzzaman, A. Alam, S. H. Lee and T. D. Dang; Sensors and Actuators B: Chemical, 185, (2013)

301-308. 23. R. E. Echols, R. H. Miller and L. Thompson; J. Chromatography A, 347, (1985) 89. 24. E. Aboul-Kasim; J. Pharm. Biomed. Anal., 22 (6), (2000) 1047. 25. K. T. Rotko; Food Chemistry, 134 (2), (2012) 1239-1243. 26. B. Bas, M. Jakubowska and T. Gorski; Talanta, 84 (4), (2011) 1032-1037. 27. S. A. Halvatzis and M. T. Potamia; Anal. Chim. Acta, 227, (1989) 405. 28. E. Akyilmaza, I. Yasa and E. Dinckaya; Anal. Biochem., 354 (1), (2006) 78. 29. T. Vergara, D. Marin and J. Vera; Anal. Chim. Acta, 120, (1980) 347.



40 Jabbar et al.


Surabaya Indonesia

30. C. Aniceto, A. V. Pereira, C. O. Costa-Neto and O. F. Filho; Lab. Rob. Autom., 11 (1), (1999) 45. 31. M. A. Koupparis and P. I. Anagnostopoulou; Anal. Chim. Acta, 204, (1988) 271. 32. J. Fan, Y. Chen, S. Feng, C. Ye and J. Wang; Anal. Sci., 19, (2003) 419. 33. D. A. Skoog, D. M. West, F. J. Holler and S. R. Crouch; "Fundamentals of Analytical Chemistry" 8th

ed., Brooks/Cole-Thomson Learning, USA (2004) 146 and 158. 34. D. C. Harris; "Exploring Chemical Analysis", W. H. Freeman and Company, USA (2009) 86 and 88.



41 Handajani et al.


Surabaya Indonesia

Analysis of N-nitrosodiethylamine in Salted Fish using Cone Shaped

Membrane-Liquid Phase Microextraction-Gas Chromatography-Flame

Ionization Detector

Usreg Sri Handajani*, Olivia Animiko & Yanuardi Raharjo

Chemistry Department, Faculty of Science and Technology, Airlangga University *Corresponding author Email: [email protected]

Abstract. Cone shaped membrane-liquid phase microextraction (CSM-LPME) is a rapid sample preparation technique, environmental friendly, selective, sensitive and easy. In this research studied of CSM-LPME method was used to analyze N-nitrosodiethylamine (NDEA) in salted fish using gas chromatography-flame ionization detector. The optimized analytical parameter are solvent ethyl acetate, volume of organic solvent 60 µL, and agitation 360 rpm. The curve of linear standard generated for NDEA‘s concentration 0f 40-90 ppm of r = 0.9999, limit of detection of 0.47 ppm, average of recovery of 100.02%, precision between 0.36% to 1.21% and enrichment factor of 6667.73. Based on the results of the analysis, NDEA compounds in salted fish samples without the addition of KNO3 and with the addition of KNO3 cannot be detected by CSM-LPME technique.

INTRODUCTION Along with the development of science and technology, analytical methods has been developed to determine

the levels of a compound in foodstuffs by using a variety of instruments. However, the analysis just by using instruments can not include samples with very low concentrations. Therefore, sample preparation techniques were introduced. Microextraction is one of preparation sample techniques based on green chemistry principal. One of preparation sample based on microextraction is liquid-phase microextraction (LPME), where a membrane impregnated with an organic solvent is used to accommodate or protect microvolumes of acceptor solution [4].

Cone shaped membrane-liquid phase microextraction (CSM-LPME) is the development of LPME. CSM-LPME has been used for the extraction of pesticides and have advantages such as simple, efficient, does not require a long extraction time, small amount of organic solvents, has high sensitivity and selectivity [5]. So it can applied to to analysis of small concentrations of compounds in foods such as nitrosamines.

Nitrosamine compounds can be found in foods that have been preserved, such as sausages, corned beef and salted fish. Salted fish contain nitrosamines that can activate the Epstein-Barr virus where is a major cause of nasopharyng cancer. One of dangerous nitrosamines‘s derivative compounds is N-nitrosodiethylamine (NDEA).

Many sample preparation techniques that have been used for the analysis of nitrosamines, there are headspace-solid phase microextraction gas chromatography with thermal energy analyzer detection (HS-SPME-GC-TEA) [1], solid phase microextraction direct extraction device gas chromatography mass spectrometry (SPME-DED-GC-MS) [10] and solid-phase extraction-micellar electrokinetic chromatography (SPE-MEKC) [3]. However, in these sample preparation techniques, there are some deficiencies, such as the SPME technique, the fiber is damaged and can not be used in a solution that has a high salt content [9], and the SPE technique requires a long analysis time [8].

With the advantages of the CSM-LPME techniques expected to be used for sample preparation technique on determination of NDEA in salted fish samples in a fast, easy, sensitivity, accuracy and precision.

RESEARCH METHODS

Tools and Materials

Materials

The materials used in this study were methanol, toluene, ethyl acetate and n-hexane with a degree of purity of pro analysis, a standard solution nitrosodietilamin (NDEA 99.99%) and sample used in this study is the salted fish.

Tools



42 Handajani et al.


Surabaya Indonesia

The tools used in this study is a set of instrument GC-FID type Agilent 6890 Plus GC Version A.03.08 with HP-5 column (length 30 m; diameter 0.250 mm; and film 10.10 µm), membrane nylon Whatman 0.2 µm, microsyringe, micropipet, tube micropipet, mortar, filter paper, Buchner funnel, magnetic stirer, and glassware commonly used in the laboratory.

Work procedures

Preparation of working solution NDEA 500 ppm.

NDEA solution (1000 mg; 99.99%; 1 mL) put into volumetric flask 100 mL diluted with methanol up to the mark, thus obtained NDEA 10.000 ppm. Working solution 500 ppm made by taking 5,0 mL NDEA 10.000 ppm and put into volumetric flask 100 mL and diluted with methanol up to the mark.

Preparation of standard solution NDEA 50, 60, 70, 80 and 90 ppm.

Standard solution 50 ppm made by taking 500.0 μL working solution NDEA 500 ppm, and put into volumetric flask 5 mL and diluted with methanol up to the mark, then shaken until homogeneous. NDEA standard solution 60, 70, 80, and 90 ppm made by taking each 600.0, 700.0, 800.0 and 900.0 μL from working solution 500 ppm and put into volumetric flask 5 mL, and diluted with methanol up to the mark.

Preparation of standard curve NDEA without extraction.

3 µL standard solution NDEA 50, 60, 70, 80, and 90 ppm each injected directly into the GC-FID instrument and analyzed. The results of the analysis of a wide area of the chromatogram. From the analysis, a standard curve was made between the results of NDEA concentration and an area of the chromatogram.

Optimization of analytical parameters on NDEA extraction with Cone Shaped Membrane-Liquid Phase Microextraction (CSM-LPME).

This study uses the CSM-LPME techniques that optimize the following parameters: organic solvent, volume of organic solvent, and agitation. Following treatment optimization parameters described in outline: 20 mL extract solvent placed on the vial bottle. The membranes were immersed for 10 seconds in an organic solvent. After the organic solvent pervasive, the membrane was immediately placed on the bottle vial containing the extract solution, and then 80 µL of organic solvent pipetted into the membrane. Samples were stirred continuously with a magnetic stirrer at a speed of 360 rpm at room temperature (25°C). After 20 min of extraction, the solvent is pulled back into a microsyringe. Extraction results are injected into the GC-FID as much as 3 µL for analysis.

Organic solvent optimization.

Optimization of organic solvent, there are n-hexane, ethyl acetate and toluene. While other variables made permanent namely volume of organic solvent, extraction time, volume of extract solution and agitation according to the procedure 2.2.4. The extracted solution was analyzed by GC-FID instrument. Then made a graph of organic solvent to the chromatogram peak area.

Volume of organic solvent optimization.

The volume of organic solvent used is 40, 60, 80, and 100 µL. While other variables made permanent namely extraction time, volume of extract solution and agitation according to the procedure 2.2.4, organic solvent suitable with optimization 2.2.4.1. The extracted solution was analyzed by GC-FID instrument. Then made a graph of volume of organic solvent to the chromatogram peak area.

Agitation optimization

Agitation were performed 120, 240, 360, and 480 rpm. While other variables made permanent namely extraction time and volume of extract solution according to the procedure 2.2.4., organic solvent



43 Handajani et al.


Surabaya Indonesia

suitable with optimization 2.2.4.1 and volume of organic solvent suitable with optimization 2.2.4.2. The extracted solution was analyzed by GC-FID instrument. Then made a graph of agitation to the chromatogram peak area. After all the optimum parameters, these parameters can be used in the analysis of samples.

Preparation of standard curve NDEA results using the CSM-LPME extraction

5 NDEA standard solution with concentration 40, 50, 60, 70, 80, and 90 ppm each extracted using the analytical parameters have been optimized in the procedure 2.2.4. Then, the extracted solution was analyzed by GC-FID and made a graph of the concentration of the peak area of the chromatogram.

Sample preparation.

Salted fish samples cut into small pieces and weighed as much as 25 g, crushed and ground, added with methanol 50 mL until submerged and allowed to stand for ± 3 hours. Sample solution put into flask 250 mL, added with 25 mL TCA 8% and diluted with aquadem up to the mark. Then filtered using a Buchner funnel to separate the solids. Clear sample solution included in a closed glass bottle aluminum foil and stored for a maximum of 7 days prior to analysis.

Sample analysis

Sample solution preparation procedure results 2.2.6 extracted using analytical parameters have been optimized with CSM-LPME technique. Extraction results were analyzed using GC instrument with a FID detector. Nitrosodiethylamine compound levels were calculated using the linear regression equation obtained from the standard curve.

Spiking

Spiking intended to determine the effect of environment on the matrix in an analytical method. Spiking method is done by adding a standard solution with a certain concentration in the sample. Samples with standard solutions that have been added then extracted with CSM-LPME technique using procedures 2.2.4. Recovery obtained from the analysis of the sample compared to the recovery of a standard optimization results, so it can be known whether the matrix in the environment affect the analysis by comparing the recovery of standard optimization results.

RESULT AND DISCUSSION

Preparation of Standard Curve NDEA Without Extract

NDEA standard curve without extraction is made by measuring a standard solution with a concentration of NDEA 50, 60, 70, 80, and 90 ppm. Measurement data standard solution can be seen on Table 1 dan Figure 1.

Table 1 Measurement data standard solution NDEA without extraction

NDEA concentration (ppm)

The average area of the chromatogram (unit)

50 9.8257

60 14.6433

70 19.4692

80 24.3721

90 29.7152



44 Handajani et al.


Surabaya Indonesia

Figure 1 Standard curve NDEA without extraction

Optimization of Analytical Parameter

Organic solvent optimization.

In this study, organic solvents optimized such as toluene, n-hexane and ethyl acetate were obtained from the optimization. Data of organic solvents can be seen in Table 2 and in Figure 2. The selection of organic solvent used is based on the resulting chromatogram area after the organic solvent extract NDEA.

Based on the data obtained, the resulting chromatogram area of the organic solvent ethyl acetate is more optimal than the area of the chromatogram of n-hexane and toluene. This is consistent with research conducted [6] where ethyl acetate is optimum organic solvent to extract NDEA.

Table 2 Data of NDEA‘s area on various types of organic solvents

Type of organic solvent used The average area of the chromatogram (unit)

Toluene 35.47

n-hexane 37.30

Ethyl acetate 45.77

Figure 2 NDEA chromatogram area curves for 3 types of organic solvents

Volume of organic solvent optimization.

The volume of the organic solvent used are 40 µL, 60 µL, 80 µL and 100 µL. The data obtained from the optimization of the volume of organic solvent can be seen in Table 3 and Figure 3.

Based on Figure 3 is obtained optimum volumes of organic solvents are in volume of 60 µL. This is

indicated by the area of the greatest chromatogram. The more organic solvents are used, will result in dilution or

y = 0.4951x - 15.0504 R² = 0.9995

05

101520253035

40 50 60 70 80 90 100Lu

as a

rea

kro

mat

ogr

am

(sat

uan

) Konsentrasi NDEA (ppm)

0

10

20

30

40

50

toluene n-hexane ethyl acetateThe

ave

rage

are

a of

th

e ch

rom

ato

gra

m

(uni

t)

types of organic solvent



45 Handajani et al.


Surabaya Indonesia

dilution effect so that the analytes were extracted and the area slightly decreased chromatogram [2]. On the use of organic solvent volume less also support the principles of green chemistry.

Table 3 Data of NDEA‘s area at various volumes of organic solvents

Volume of organic solvent (µL) The average area of the chromatogram (unit)

40 µL 41.33

60 µL 48.73

80 µL 40.003

100 µL 31.93

Figure 3 Curve relationship between the area of the chromatogram NDEA by volume of organic solvent.

Agitation optimization.

In this study, the agitation optimized : 120 rpm, 240 rpm, 360 rpm and 480 rpm. The data obtained from the agitation optimization can be seen in Table 4 and Figure 4.

Based on Figure 4, obtained the maximum agitation of 360 rpm which can be seen from the chromatogram

area produced the greatest. Agitation under 360 rpm resulted in a less optimum chromatogram because the stirring speed is too small so that the mass transfer of the analyte from the donor phase to the acceptor phase is less than optimal. While the stirring speed above 360 rpm, the resulting chromatogram area is smaller due to back extraction of analytes from the donor phase to the acceptor phase with the equilibrium [5].

Table 4 Data of NDEA‘s area at various agitation

Agitation (rpm) The average area of the chromatogram (unit)

120 40.61

240 43.63

360 48.73

480 44.86

30

35

40

45

50

0 50 100 150The

ave

rage

are

a o

f th

e c

hro

mat

ogr

am

(un

it)

volume of organic solvent (µL)



46 Handajani et al.


Surabaya Indonesia

Figure 4 Curve relationship between the area of the chromatogram NDEA with agitation.

Preparation of Standard Curve NDEA with CSM-LPME technique

In this study, a standard curve NDEA with CSM-LPME technique measurements obtained from standard solution 40, 50, 60, 70, 80, and 90 ppm with the standard solution conditions that have been optimized analytical parameters. Measurement data with a standard solution NDEA CSM-LPME technique can be seen in Table 5 and Figure 5.

Figure 5 Standard curves NDEA using CSM-LPME

Table 5 Measurement data of NDEA standard solution by using CSM-LPME

NDEA concentration (ppm)

The average area of the chromatogram (unit)

40 12.17

50 24.56

60 36.60

70 48.73

80 61.14

90 73.76

Determination Validity of Method

Limit of detection.

In Table 6 it can be seen that the limit of detection of the analytical technique NDEA with CSM-LPME smaller at 0.47 ppm in comparison with no CSM-LPME is 1.16 ppm. This shows the CSM-LPME technique can analyze NDEA at smaller levels. Based on these data, it can be shown that the CSM-LPME technique increases the sensitivity of the instrument.

40

42

44

46

48

50

0 200 400 600

The

ave

rage

are

a o

f th

e c

hro

mat

ogr

am

(un

it)

Agitation(rpm)

y = 1.2281x - 36.9970 R² = 0.9999

0

20

40

60

80

0 50 100

The

ave

rage

are

a o

f th

e c

hro

mat

ogr

am

(un

it)

Concentration NDEA (ppm)



47 Handajani et al.


Surabaya Indonesia

Table 6 Data limit of detection of analysis NDEA without CSM-LPME and analysis NDEA with CSM-LPME

Method Standard curve regression equation

The correlation coefficient (r)

Limit of detection (ppm)

Without CSM-LPME

y = 0.4951x-15.0504 0.9995 1.16

With CSM-LPME y = 1.2281x - 36.9970 0.9999 0.47

Accuracy and precision (KV).

On the analysis NDEA with CSM-LPME technique, obtained accuracy with range 99.7% to 100.26%. This shows the recovery of NDEA as analyte with CSM-LPME technique approaching the truth NDEA concentrations. Meanwhile KV obtained from the analysis of the CSM-LPME NDEA has a range of 0.36 to 1.21%. This suggests that the NDEA analysis with CSM-LPME has a lower deviation for each outcome measure. Data recovery and precision analysis of NDEA with CSM-LPME can be seen in Table 7.

Table 7 Data recovery and precision analysis of NDEA without CSM-LPME

Description NDEA’s standard concentration (ppm)

40 50 60 70 80 90

KV (%) 0.48 1.21 1.17 0.67 0.97 0.36

R (%) 100.075 100.26 99.88 99.7 99.96 100.22

Enrichment factor.

NDEA concentration can be seen in comparison curve of NDEA‘s standard before extraction with CSM-LPME and after extraction with CSM-LPME in Figure 6.

Figure 6 Comparison curve of NDEA‘s standard before extraction with CSM-LPME and after extraction with CSM-LPME

Sample Analysis

In this research use 2 different salted fish sample, there are salted fish without added KNO3 (A) and with added KNO3 (B). Data analysis NDEA with CSM-LPME technique on sample A and sample B as in Table 8. Based on the data in Table 8, it was concluded that the NDEA in samples A and B are not detected by CSM-LPME technique.

0

20

40

60

80

0 20 40 60 80 100

Luas

are

a ra

ta-r

ata

kro

mat

ogr

am

(sat

uan

)

Konsentrasi NDEA (ppm)

sebelumekstraksi



48 Handajani et al.


Surabaya Indonesia

Table 8 Data analysis results of samples

Sample The area of the

chromatogram (unit) Chromatogram

A 0

B 0

Spiking

Data analysis spiking samples A and B can be seen in Table 9. Based on Table 9 the results of spiking the sample A showed there was no NDEA in the sample. The area average of chromatogram of spiking sample A 40.39 smaller than the area average of chromatogram standard solution 70 ppm after extraction with CSM-LPME, 48.73. Meanwhile spiking sample B‘s recovery was 142.37%. NDEA compounds were not detected in

the sample B is caused by a matrix. In fish contained composition such as water, minerals, protein and fat, meanwhile that precipitated so as not to interfere process of analysis only protein, with the addition of TCA. Nutrient composition of salted fish are shown in Table 10.

Table 9 Data analysis NDEA‘s spiking with CSM-LPME

Sample The average area of the chromatogram

(unit) R (%) Chromatogram

A 40.39 90.03

B 85.38 142.37



49 Handajani et al.


Surabaya Indonesia

Table 10 Nutrient composition of salted fish

Component Nutrient composition (%)

Water 43.85

Protein 28.44

Fat 4.73

Mineral (ash) 19.25

Fats contained in fish is one of the matrix on the extraction process NDEA with CSM-LPME. Fat soluble

perfect in esters, hydrocarbons, benzene, carbon disulfide, and halogen solvents. Ethyl acetate is the ester group of compounds, so that the fat can be dissolved in ethyl acetate. Therefore fat interfere the process of analyte transfer from donor phase to aceptor phase because the competition between NDEA and fat, so that the resulting of recovery 142.37%.

CONCLUSIONS AND SUGGESTION

Conclusions

Based on the research that has been done, it can be concluded as follows: 1. NDEA compound can be extracted by using the CSM-LPME technique and further analyzed by GC. 2. 3. Results of optimization of analytical parameters on NDEA analysis using CSM-LPME techniques, among

others: ethyl acetate as the organic solvent, the organic solvent volume 60 mL and agitation of 360 rpm. 4. 5. CSM-LPME technique can not detect the NDEA in salted fish samples.

Suggestion

1. Need for more research on the matrix that interfere the analysis of NDEA in salted fish samples, namely optimization of the matrix.

2. Require analysis of nitrosamine derivatives on a wide variety of processed food products that are carcinogenic with CSM-LPME technique.

REFERENCES 1. Andrade, R. Reyes, F.G.R. & Rath, S., A Method for the Determination of Volatile N-nitrosamine in Food

by HS-SPME-GC-TEA, Food Chemistry, 91, pp. 173-179, 2005. 2. Farajzadeh, M.A. & Khoshmaram, L., Air-Assisted Liquid-Liquid Microextraction-Gas Chromatography-

Flame Ionisation Detection: A Fast and Simple Method for the Assessment of Triazole Pesticides Residues in Surface Water, Cucumber, Tomato and Grape Juices Samples, Food Chemistry, 141, pp. 1881-1887, 2013.

3. Filho, P.J.S. Rios, A. Valcarcel, M. Zanin, K.D. & Caramao, E.B., Determination of Nitrosamines in Preserved Sausages by Solid Phase Extraction-Micellar Electrokinetic Chromatography, Journal of Chromatography A, 985, pp. 503-512, 2013.

4. Jonsson, J.A. & Mathiasson, L., Membrane Extraction in Analytical Chemistry, Journal of Separation Science, 24, pp. 495-507, 2001.

5. Sanagi, M.M. Heng, S.H. Ibrahim, W.A.W. & Naim, A.A., Determination of Pestisides in Water by Cone-Shaped Membrane Protected Liquid Phase Microextraction Prior to Micro-Liquid Chromatography, Journal of Chromatography A, 1152, pp. 215-219, 2007.

6. Shofiyah, A., Analisis Nitrosodietilamin (NDEA) dalam Ikan Asin dan Daging Kaleng dengan Teknik Kromatografi Gas melalui Headspace Single Drop Microextraction (HS-SDME), Skripsi, Departemen Kimia, Fakultas Sains dan Teknologi, Universitas Airlangga, Surabaya, 2012.

7. Singgih, W., Industri Pemindangan Ikan, Swadaya, Jakarta, 2000.



50 Handajani et al.


Surabaya Indonesia

8. Thiele, C. & Taufkirchen, S.A., Basics in Solid Phase Extraction (SPE), Euroanalysis, Belgrade Lunch Symposium, 2011.

9. Ulrich, S., Solid-Phase Microextraction in Biomedical Analysis, Journal of Chromatography A, 902, pp. 167-194, 2000.

10. Ventanas, S. Martin, D. Esteves M. & Ruiz, J., Analysis of Volatile Nitrosamines from a Model System using SPME-DED at Different Temperatures and Times of Extraction, Journal of Food Chemistry, 99, pp. 842-850, 2006.



51 Rismiarti et al.


Surabaya Indonesia

Preparation and Properties of Tetraborate Ion Selective Sensor Based

On Zeolite

Zuri Rismiartia*, Atikahb, Chasan Bisrib, Yuyun Irnawatia

a Study Program of Chemistry, Faculty of Science and Technology, University of Machung, Malang 65151Indonesia

bDepartment of Chemistry, University of Brawijaya, Malang 65145 Indonesia *Corresponding Author’s email: [email protected]

Abstract: The coated wire tetraborate ion selective elecrode based on zeolite as an active material was preparationed and characterized. Turen natural zeolite has been calcinated with NH4Cl 2M and heating at 550 oC. The membrane was made from a mixture of calcinated Turen natural zeolite, polyvinylchloride (PVC), activated carbon, dioctilphtalate (DOP) in tetrahydrofuran (THF) solvent and was coated on the surface of platinum wire as the electronic conductor. In this research, Na2B4O7 1x10-8 – 1x10-1 M solution were measured using Ag/AgCl reference electrode. The result showed ISE‘s membrane soaked in Na2B4O7 0.1 M for 40 minutes. It showed Nernstian reponse (the Nernst factor of 29.14 ± 0.083 mV/decade of concentration in a linear concentration range of 1x10-5 – 1x10-1 M, the detection limit of 3.24x10-6M (0.619 ppm tetraborate), response time of 50 seconds, life time of 64 days and have reproducibility with a coefficient of variation of 0.285%.

INTRODUCTION Tetraborate is a primary ion of borax or sodium tetraborate (Na2B4O7) compound. The dissosiation of borax

is tetraborate ion (B4O5(OH)42-) [1-2]. Borax is widely used as an additive (additive) foods as a preservative and

make foods more flexible and durable texture. In the body, borax can accumulate in the liver, brain, and in the renal tissue. When consumed continuously for long terms can cause cancer and level of 5-20 g/kg borax which may cause death in adults [3-4].

Therefore, the Ministry of Health declares that prohibits the use of borax in food. Otherwise, borax is often abused in foods such as meatballs, wet noodles, siomay, krupuk. These foods are usually produced by small and medium enterprises that are not listed in the POM so that the quality is not controlled. Because of that, monitoring borax in food is important so to keep consumers‘ healthy.

Several determination methods of tetraborate such as; UV-Vis spectrophotometric, titrimetric have been successfully developed. The methods have high precision and accuracy but they require expensive instrumentations as well as delicate, expert handling of the samples. Ion selective electrode (ISE) has been developed to analyze ion tetraborate. The advantages of ISE such as ease of preparation, low cost, simple procedure and instrumentation (pH/potentiometer), no requirement for sample destruction, and fast response [5]. ESI has two types, tube and coated wire type. The coated wire type has some advantages compared to that of tube type as it does not require special maintenance, can easily be stored without risk of losing the reference solution, and it also can be operated in a variety of positions (tube type only in an upright position) [6]. Ion selective electrode (ISE) borate has been developed using salt solids Ag3BO3, Ag2S, Cu2S, with concentration range of 1x10-1 - 1x10-6 M borate, Nernst factor of 31.2 ± 2 mV, response time of 20-30 seconds and life time of 2 years [7]. Coated Wire tetraborate ISE based chitosan has been developed showed Nernst factor of 29.18±0.12, concentration range of 1x10-5 to 1x10-1 M tetraborate, detection limit of 3,715x10-6 M (0.71 ppm tetraborate), the response time of 110 seconds, and a lifetime of 38 days [8].

Zeolite is an abundantly natural resource of Indonesia, and has affordable price. Zeolite has been successfully become an ion exchange, adsorption, catalyst and anion sensor for sulphate [9-10]. For anion sensors required a modification of the zeolite to form a positive site through a calcination process by heating above 500oC and be the formation of Lewis acid sites are positively charged site Si (Si +) [11]. The establishment of the site is used to bind the tetraborate ion.

In the present study, tetraborate sensor was constructed in coated wire type with the membrane prepared using zeolite ionophore, PVC carrier, DOP (diocthyl phatalate) plasticizer and activated carbon. The performance of the sensor was determined through the general ISE characters, including sensitivity (Nernst factor), linear concentration range, detection limit, response time, and life time.



52 Rismiarti et al.


Surabaya Indonesia

EXPERIMENT

Materials

Natural zeolt from Turen Malang Indonesia, dioctylphthalate (DOP) (Sigma Aldrich), , Polyvinylchloride (PVC) high molecular weight (60,000) (Sigma Aldrich), tetrahydrofurane (Sigma Aldrich), Na2B2O7 (Merck), HNO3 65% (v/v) (Merck), alcohol 96 % (v/v) (Merck), distilled water, teflon rod (Ø 1 cm), 99.9% platinum wire (Ø 0.5 mm), and jack coaxial cable RG-58.

Apparatus

The potentials occurred across the membrane electrodes were measured using a galvanic cell of the following type: Ag ǀ AgCl ǀ KCl (1M) ǁ sample solution (10

-8–10-1 M Na2B4O7) ǀ tetrabaorate membrane ǀ Pt

electrode. Potentiometer (Fisher Accumet Model 955), analytical balance (Adventurer models AR 2130).

Procedure

Preparation of Membrane Electrode

Prior to use zeolite as an ionophore, zeolit activation using NH4Cl. zeolites ground into powder and sieve 120-150 mesh. After that, zeolite powder washed with distilled water and dried in an oven at 105oC for 2 hours. Furthermore, the zeolite was soaked in a solution of 2M NH4Cl with a ratio of 10 g zeolite in 100 mL of solution and stirred for 24 hours. Then the zeolite is heated in a furnace at a temperature of 550oC for 4 hours. Membrane electrode was prepared by mixing various compositions of ionophores (zeolite powder), plasticizers (DOP), PVC and activated carbon followed by dissolution in THF (1:3 % w/v), and stirring for 3-4 hours until homogenous membrane solution was obtained.

Construction of ISE

The electrode body was constructed by dipping the platinum wire (diameter 0.5 mm, length of ± 5 cm) into the oxalate membrane with a thickness between 0.1-0.2 mm and allowed to stand overnight in the oven at 50 OC. The top of Pt wire was connected to RG-58 coaxial cable which was soldered to the jack RG 58 as connector to the potentiometer. Prior to use the electrode was initially conditioned by soaking it in 0.1 M of Na2B4O7 solution. This procedure is also purposed for doping chitosan with oxalate. The first conditioning time was approximately 12 hours, then 10–80 min for successive used.

Characterization of Tetraborate ISE

Characterization of tetraborate ISE includes the determination of the Nernst‘s factor, detection limit, response time, life time. The paramaters was determined by measuring the potential of tetraborate solution. Potential of the tetraborate solution was determined by measuring 10-8

─ 10-1 M of Na2B4O7 at room temperature (26 ─ 27 °C) using reference electrodes of Ag/AgCl. The electrode immersed to a constant depth of tetraborate solution and stirred at constant rate. The potential readings were recorded when they showed stable readings.


Performance of Tetraborate ISE

The performance of tetraborate ISE was determined by measuring the potential response of the optimized membrane tetraborate selective electrode to various concentrations of tetraborate ions. The ISE performance was shown by sensitivity (Nernst factor, linear concentration range, limit of detection), response time, life time. Nernst factor, the range of concentrations, and the detection limit was determined by the data were plotted as observed potential vs the logarithm of tetraborate concentration (M). The Nernst factor was determined from the slope of the resulting curve. Linear concentration range was determined from the straight line [12]. The curve gives the straight-line linear in the range of -log [tetraborate] 1-5 with the slope of 29.145. It means that the range of concentration measurements of ISE is 1x10-5 ─ 1x10-1 M tetraborate and the Nernst factor of 29.14 mV / decade concentration with regression of 0.998. Detection limit was determined by the intersection between the straight line (linear region) and curved lines (non-linear region). The point of intersection of the two lines are extrapolated to the x-axis in order to obtain the concentration of detection. Limit detection of tetraborate ISE



53 Rismiarti et al.


Surabaya Indonesia

based on zeolite of 3,240x10-6 M or 0.619 ppm tetraborate. Tetraborate ISE based on zeolite has reproducibility with a coefficient of variation of 0.285%

Soaking time

Effect of soaking time the membrane of the Nernst factor prices can be seen in Figure 1. Soaking serves to facilitate the occurrence of dissociation and ion binds tetraborate (B4O5(OH)4

2-) on the site of Al-O-Si+. The optimum soaking time for tetraborate-selective electrode in the solution of sodium tetraborate 0.1 M was 40 minutes. At this time, the necessity of water in membrane for dissociation process has been fulfilled. As a result, ion exchanger process was easily reached to give optimum sensitivity. At soaking time of 60 until 80 minutes, the sensitivity of the sensor to the tetraborate ion declined. The soaking time was too long so that excessive water allows into the pores of the sensor membrane so that ion exchanger process in the membrane-solution interfaces phase becomes hardly achieved and small potential response was observed [13]. Furthermore membrane permeability increased so that the sensitivity and selectivity of the sensor membrane to B4O5(OH)4

2- ion decreases. It was showed the response ISE was not Nernstian.

Figure 1. Soaking time

Response time

Response time was determined when a steady-state potential with less than 0.1 mV/min change was achieved by the lowest analyte concentrations within linear range of calibration curve [14]. The oxalate solution of 1x10-5 M showed steady state potential (response time) at 50 seconds; thus, the tetraborate ISE based on zeolite provides acceptable response time (less than 1 minutes).

The response time is affected by ion mobility. Based on Table 1 shows the higher the tetraborate ion concentration in analyte solution has the fastest response time because the more tetraborate ions in analyte solution which interacts with the site of zeolite-Al-O-Si+ as ionophore sensor membrane. As the resulted, ion exchanger process in the membrane-solution interfaces phase was easily reached. Beside of the concentration of the analyte solution, the response time is also influenced by the presence of interference that give the slower response time.

Table 1. Response time of Tetraborate ISE

p[B4O5(OH)42-]

Response time (seconds)

5 50 4 50

3 40

2 40

1 30

Life Time Life time determines the long-term stability of the electrode which gives reproducible response slope within

± 5 mV/decade. This study was conducted by periodically re-calibrating standard solutions and calculating the response slope (Nernst‘ factor) over the linear concentration range (1.0 × 10-5 -1.0 M). The Nernst‘ factor showed reproducible over a period of 64 days (Figure 2). The life time of tetraborate ISE based on zeolite

15

20

25

30

35

10 20 30 40 50 60 70 80

Ner

ns F

acto

r (m

V/d

ecad

e co

ncen

trat

ion)

Soaking time (minute)

The upper Nernst factor The theoretical Nernst factor The lower Nernst factor



54 Rismiarti et al.


Surabaya Indonesia

membrane is influenced by migration of plasticizer to the membrane surface as a result of increasing electrical resistivity of membrane. This condition would confine kinetic of tetraborate ion transport to the membrane. Consequently, both of homogeneity of membrane and physical stability decreased and the response was not Nernstian.

Figure 2. Life time

CONCLUSION The Tetraborate sensor outlined above allows simple, sensitive and fast analyzing time for determination of

tetraborate. This result suggests that the electrode is prospective to be used for monitoring of tetraborate in the food product regularly.

REFERENCES 1. Garrett, D.E, Handbook of Deposits, Processing, Properties, and Use Borates, Academic Press, 1998. 2. Emeleus, H.J. dan A.G. Sharpe., Advance in Inorganic Chemistry and Radiochemistry, Academic Press,

1982. 3. Dinas Peternakan Dan Kesehatan Hewan-Lampung, Bahaya Formalin Dan Boraks,

http://www.disnakkeswan-lampung.go.id, (24 Mei 2010) 4. Office of Prevention Pesticides and Toxic Substances, Report Of The Food Quality Protection Act

(FQPA) Tolerance Reassessment Eligibility Decision (TRED) for Boric Acid/Sodium Borate Salts, United States Environmental Protection Agency, http://www.epa.gov/oppsrrd1/ REDs/boric_acid_tred.pdf. Retrieved 2008-04-21, (10 Januari 2011).

5. Rawat, A., C. Sulekh dan S. Anjana., Highly Selective Potentiometric Oxalate Ion Sensors Based on Ni(II) Bis-(m-minoacetophenone)ethylenediamine, Chin. J. Chem., 28: pp 1140—1146, 2010.

6. Rismiarti, Z., Atikah, Sulistyarti, H., Construction and Characterization of Coated Wire Oxalate Ion Selective Electrode Based on Chitosan, J.Pure App. Chem.Res 3(1),pp. 19-26, 2014.

7. Somer, G., Sezer, S., Dogan, M., Kalayci, S., Sendil, O., Preparation and Properties of A New Solid State Borate Ion Selective Electrode and Its Application, Talanta, 85 (3),1461-5, 2011.

8. Afriansyah, A., Pembuatan dan Karakterisasi Elektroda Selektif Ion Tetraborat Tipe Kawat Terlapis Bermembran Kitosan, Skripsi, Jurusan Kimia, Universitas Brawijaya, Malang, 2011.

9. Widianti, T., Pengujian Kapasitas Tukar Kation Zeolit Sebagai Kapasitas Penukar Kation Alami Untuk Pengolahan Limbah Industri, ISSN 1907-7459, AMTeQ, 2006.

10. Huheey, J.E., A.K. Ellen, dan R.L. Keither., Inorganic Chemistry Principle Of Structure And Reactivity Fourth Edition, Hooper Collins College Publisher, 1993.

11. Fardiyah, Q., Atikah, Ningsih, R.D., Pemanfaatan Zeolit Teraktivasi Sebagai Bahan Aktif Sensor Potensiometri Ion Sulfat, Chem. Prog. 7(2),pp. 81-87, 2014.

12. Wang, J., Analytical Electrochemistry, Third Edition, John Willey& Sons, 2006.

20

25

30

35

0 10 20 30 40 50 60 70

Ner

nst

Fac

tor

(mV

/dec

ade

conc

entr

atio

n)

Time (days)

The upper Nernst factor The theoretical Nernst factor The lower Nernst factor

http://www.disnakkeswan-lampung.go.id/

http://www.epa.gov/oppsrrd1/REDs/boric_acid_tred.pdf

http://www.epa.gov/oppsrrd1/REDs/boric_acid_tred.pdf

http://www.epa.gov/oppsrrd1/%20REDs/boric_acid_tred.pdf.%20Retrieved%202008-04-21

http://www.epa.gov/oppsrrd1/%20REDs/boric_acid_tred.pdf.%20Retrieved%202008-04-21



55 Rismiarti et al.


Surabaya Indonesia

13. Kurniasih, D., Atikah dan H. Sulistyarti, The Coated-Wire Ion Selective Electrode (CWISE) of Chromate Using PVC-Membrane Based on Chitosan as A Carrier, J. Pure App. Chem. Res.,1 (1),pp 33‐40, 2012

14. Ardakani, M.M., M. Jalayer , H. Naeimi , A. Heidarnezhad, dan H.R. Zare, Highly Selective Oxalate-Membrane Electrode Based On 2,2_-[1,4-Butandiyle Bis(nitrilo Propylidine)]Bis-1-Naphtholato Copper(II), Biosens and Bioelectr, 21,pp. 1156–116, 2006



56


Surabaya Indonesia

ORAL PAPER OF PURE AND APPLIED MATHEMATICSPURE (OAMT)



Labellapansa et al. 57


Surabaya Indonesia

Rule-Based And Case-Based Reasoning System For Psychiatric

Psychosis Disorders Diagnosis

Ause Labellapansa 1*, Sri Hartati 2

1Kaharudin Nasution Street No.113, University of Islam Riau, Pekanbaru, Indonesia 2Kaliurang Street, Yogyakarta, University of Gadjah Mada, Indonesia

*Email: [email protected]

Abstract.According to The Global Burden of Disease conducted by Murray in collaboration with WHO and the World Bank predicts that mental illnesses will occupy the second position after cardiovascular disease in 2020. One of the mental illnesses is Psychosis disorder. Computer system that utilizes artificial intelligence methods to help professionals working in the field of psychiatric medicine becomes indispensable. This research uses a rule-based reasoning (RBR) to conduct an assessment consisting of psychosis Schizophrenia disorders, settled delusional disorders and acute transient psychotic disorders with Certainty Factor (CF) method to handle uncertainty and the use of case-based reasoning (CBR) for diagnosing types of Schizophrenia disorders. The process of diagnosis is done by entering the patients symptoms by paramedics. If the patient has a Schizophrenia disorder, then CBR is used to diagnose kinds of Schizophrenia. Each Schizophrenia new case will be calculated using Weighted Nearest Neighbor similarity method. Results of tests performed by experts have shown that the system is able to diagnose correctly. The test results using medical records show that the system is able to correctly identify the kind of Schizophrenia with high similarity criteria at 80% (0,8-1) and medium similarity criteria at 20% (0,6-0,79).

INTRODUCTION The term psychiatric disorders / mental disorders are the symptoms or behavior pattern that can be found in

clinical-related stress and in most cases associated with impaired function of a person [1]. Prevelence of psychiatric disorders according to the report by Indonesian National Health Policy Research (RISKESDAS) based on the survey of the Indonesian population aged ≥ 15 years was 11.6%, which means between about 19

million people there are around 2.2 million people who have a psychiatric disorder [1]. Research on mental illness "The Global Burden of Disease" performed by [2] collaborated with WHO and World Bank predicts that mental illness will occupy the second position after cardiovascular diseases in 2020. Five biggest psychiatric illnesses have most significant impact is unipolar depression, alcohol use, bipolar affective disorder, schizophrenia and obsessive compulsive disorder.

By knowing the facts above, causes a system that utilizes artificial intelligence methods to help working professionals in the field of psychiatric medicine is needed. Researchers feel the need to do research to develop a system that helps paramedics in upholding psychosis diagnosis of psychiatric disorder . The system that will be developed is the use of rule-based reasoning (RBR) and case-based reasoning (CBR) for psychosis disorder diagnosis. If patients diagnosed with disorders other than schizophrenia that is supposition disorders and acute and transient psychotic disorder, then the system will provide solutions for patients by using RBR and if patients diagnosed with schizophrenic disorder, the paramedic will do more digging to find out the symptoms of schizophrenia suffered by using CBR.

Rule-based reasoning is a form of knowledge base approach where knowledge will be represented using the form of IF-THEN rules. This form is used if there is a certain amount of knowledge in a particular domain which is completed sequentially. The general form of rule-based reasoning is antacedent IF THEN consequent which is the part that expresses antacedent situation or premise while consequence is a part that states a particular action or conclusion that is applied if a situation or premise is true [12].





Surabaya Indonesia

CBR is a technique to solve new problems by adapting solutions that have been used to solve previous problems. In CBR, the cases that already exist are stored in the base case / case base. If there is a new problem occurs, old cases which considered most similar to the new one will be retrieved. If the case retrieved considered similar with the new one, there will be a process of reusing old cases to solve problems in new cases. However, if the old case does not resolve a new case considerably, then process of revising where the case that has been revised will be retained for later use in solving upcoming new cases and stored in the case base [13].

LITERATURE REVIEW Researches in the psychiatric disorder domain have been done, among CBR in stress domain to

diagnose stress levels incurred by Begum et al. [3] by seeing the difference in the temperature of fingers as well as Ahmed et al. [4] which combines the signal time series data and patient document in seeing stress levels. Along with that, utilization and importance of ontology-based system for mental health interventions have also been shown by Coyle and Doherty [5]. Another study has been done by developing a group decision support system for amnesis, diagnosis and therapy of mental disorders in the diagnosis of non-psychosis disorders by Kusumadewi et al. [6]

Researches for psychosis disorder to diagnose Schizophrenia done by Razzouk et al. [7] by developing DSS involving three specialists, but this study experienced difficulties due to the absence of an agreement between one specialist with one another in looking at the symptoms related to schizophrenia.

Several other studies using CBR done by Tomar et al . [8] make a CBR clinical decision support system prototype for diagnosis of pulmonary disease due to work by using 127 cases for 14 kinds of chronic lung diseases due to work on 26 kinds of symptoms. Retrieval used is Nearest-Neighbor. Development of this GDSS will be able to help the pathologist to decide level of testing for sensitivity 95.3%.

Rismawan [9] used CBR to diagnose Ear-Nose-Throat by using 106 symptoms with 38 kinds of diseases which divided into 3 classes and 10 sub-classes. Indexing symptom is done by using the backpropogation method. Similarity method used is the Cosine Co-Efficient threshold value of 0.8. Gu et al. [10] applied CBR to reuse knowledge of dental medical record in making a new medical record. Retrieval is done by using several methods namely Fuzzy Mathematics method, Euclidian-Lagrangian Distance and weighting optimization with PULL&PUSH. Own revisions done by using rule-based acquired from experts in the field of dentistry. Cases collected up to 1000 cases and 100 cases for testing with the treshold value of 0.6. When number of cases reaches 100, one case found as the most similar case but if the weighting is applied, then two cases found similar. Mancasari [11] made an expert system to diagnose neurological diseases in children has adopted the weight difference between a disease and other diseases and percentage-confidence expert is also available in performing diagnosis.

RESEARCH METHODOLOGY

Steps in Diagnosing Phsycosis Disorers

Diagnosis use RBR consists of several steps which are knowledge acquisition from experts, representing knowledge to a decision tree, using certainty factor method and give diagnosis result and suggested solutions. On the other hand, diagnosis use CBR consists of several steps which are knowledge acquisition from medical records, concluding attributes that will be used in a case, representing a case, using Nearest neighbor formula to conclude the most similar case and provide suggested solution as diagnostic results.

Rule-Based Reasoning Psychosis Disorders Diagnose

Based on guidelines for PPDGJ III psychosis disease diagnosis, decision tree was made in which there are 17 kinds of symptoms and 3 kinds of psychotic illnesses which are Schizophrenia, delusional





Surabaya Indonesia

disorders and acute and transient psychotic disorders. Based on the decision tree, 36 rules of psychotic disorders diagnosis have been acquired.

Handling uncertainty with certainty factor (CF) methods

CF value will be in the range of 1 to -1 where the value of 1 indicates absolute trust and value of -1 states absolute distrust. In MYCIN threshold concept is also used. Threshold level of 0.2 is denoted by δ (delta). Equation (1) is used to determine the CF premise of a rule [12]

{

{ }

{ }

(1)

Furthermore Ignizio [12] calculated the CF output rule by using equation (2)

[ ] (2)

If more than one conclusion are reached for a disease, then combined CF calculation use equation (3)

{

(3)

Case-Based Reasoning (CBR) to Diagnosis Types of Schizophrenia

Representation Cases

In this study case representation is used by using frames. As for attributes consists of symptoms, patient data that contains patient's age and sex, the trust level, disorder names and case solution.

Retrieval and Similarity

Retrieval technique used is the weighted nearest neighbor by Kolodner [13] in equation (4) and Mancasari [11] in equation (5), so it will use equation (6) to compute the similarity of cases.

∑

∑ (4)

(5)

∑ ∑ (

| |

)

∑

(6)

Equation (7) is used to calculate local similarities if there is a symbolic features and Equation (8) is used to calculate numeric features [14].

{

(7)

(8)





Surabaya Indonesia

RESULTS AND DISCUSSION

System Architecture

Figure 1. Psychotic disorder diagnosis system architecture

Figure 1 is a system architecture for the diagnosis of psychotic disorder, where there are several components of the knowledge acquisition component to execute the acquisition of knowledge from expert users in making rules psychosis disease diagnosis, knowledge base components, consulting component, component inference engine with forward chaining method, the development components of the case base and the CBR component consisting of retrieval, reuse, revise and retain.

Case-Based Development Process

Stages in the CBR commenced to perform the manufacturing case-based Schizophrenia patients types. In data retrieval 95 cases have been successfully acquired. Those cases consist of 82 kinds of symptoms which 80 cases used as training data and 15 cases are used as test data.

Diagnostic Process

Diagnostic process begins with selecting the data of patients who will be diagnosed and continued with inputting symptoms felt by patient. As one of the diagnostic examples, user inputted psychotic symptoms by answering the question for G16 symptoms code with CF value of 1, G17 symptoms code with CF value of 1, G5 symptoms code with CF value of 0.8 and user answered No to the symptoms of other code with value of CF -1 user. So that selected rules are: For rule 1 CF Composite premise of Rule 1 (1;1;1) = Min (1;1;1) = 1 CF (R1) = CFCompositeR1xCFpremise (expert) = 1 x 0,9 = 0,9 For rule 5 CF Composite premise of Rule 5 (1;1;0,8)= Min (1;1;0,8) = 0.8 CF (R1) = CFCompositeR5xCFpremise (expert) = 0,8 x 0,9 = 0,72

Due to conclusion of Schizophrenia disease of > 1 rule with a value > 0, then calculations to figure CF conclusion by using equation (3) that in order to obtain the conclusion that CF patients are diagnosed with Schizophrenia disease with confidence level of 0,972 or 97.2%. The next step is to input the symptoms more detail to diagnose the type of Schizophrenia the patients have. User inputs those 3 symptoms in Figure 2 on 29 year old male





Surabaya Indonesia

Figure 2. Test case by experts

To see the new case calculations based on data found in Figure 2, then given an example calculation of the proximity of the new case with the old cases id cases 3 to 11, 15 and 46 with a description as follow

New case local similarities with case 11:

Age proximity,

= 0,675

Gender proximity = 1, new case and case 11 have the same gender (male). Symptoms proximity: Symptoms G013, G026 and G059 weighted at 1, due to both cases have the same symptom. Symptom G045 weighted at 0 because symptom does not appear on the new case.


Age proximity,

= 0,9

Gender proximity = 1, new case and case 15 have the same gender (male). Symptoms proximity: Symptoms G013, G026 and G059 weighted at1, due to both cases have the same symptom. Symptoms G019 and G021 weighted at 0 because symptom does not appear on the new case.


Age proximity,

= 0,95

Gender proximity = 1, new case and case 15 have the same gender (male) Symptom proximity G032 and G039 weighted at 0 because symptom does not appear on the new case

Global similarities: Based on the computation of local similarity, then global similarity is used for calculation using equation (6)

as follows:

%

= 0,935 x % = 93,5 %

%

= 0,98 x % = 98 %

%

= 0,975 x

% = 39 %

The case S15 is the most similar to the new case with 98% similarity

CBR Diagnostic System Disorders Types of Schizophrenia Testing





Surabaya Indonesia

Testing performed on the CBR system using similarity criteria in Table 1 which is divided into HS indicates that cases that appear to have a high similarity or have a very similar resemblance, MS indicates that cases that appear to have similar resemblance while LS indicates that the case appears less similar with the range from 0 to 1.

Table 1. Tests using data of cases on medical records

Tabel 1 No.

Tabel 2 Similarity values

Tabel 3 Similarity criteria

Tabel 4 1 Tabel 5 0.80 – 1.00

Tabel 6 High similarity (HS)

Tabel 7 2 Tabel 8 0.60 – 0.79

Tabel 9 Medium similarity (MS)

Tabel 10 3 Tabel 11 0.00 – 0.59

Tabel 12 Low similarity (LS)

Obtained from the test results that the diagnostic results through the system and the results of diagnosis by a physician are true overall. Figure 3 shows a graph of test cases where the horizontal line indicates the case 1 to case 15 which tested and vertical line show the similarity values ranging from 0 to 1. 3 (three) cases had similar similarity values/MS which is case 4, case 6 and case 8 while other cases have very similar similarity criteria.

Figure 3. Graphic of 15 cases test results

CONCLUSIONS AND RECOMMENDATIONS

Conclusions

This study has successfully created a system using a rule-based and case-based reasoning which can help early diagnosis of psychiatric disorders and psychosis diagnosis of diseases Schizophrenia. Early diagnosis of a psychotic disorder consists of Schizophrenia disorder, settled delusional disorder and acute psychotic disorder and transient. From the tests conducted by experts using the system, it showed that diagnosis results by the experts and diagnosis results by system give the same diagnostic conclusions. Schizophrenia Diagnosis is done by using case-based reasoning. From the tests conducted by the expert system diagnoses the type of Schizophrenia diagnosis system is obtained in accordance with the results of the expert diagnoses.

Recommendations

This research diagnoses psychotic disorder, so it will be better if the additional psychiatric disorder inputted to the system that will further accommodate the needs of paramedics in diagnosing psychiatric disorders

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Sim

ilari

ty v

alu

es

Case number





Surabaya Indonesia

NOMENCLATURE List the nomenclature = CF threshold level = Conclusion of CF C = CF rule 1 = CF rule 2

= Output CF of rule k

= Similarities function for attribute i between T case and S case

= The maximum value of f = The minimum value of f = Number of attributes in T that appears on attributes S = Number of attributes found in T = CF premise clause of i = CF rule k =

CF composites premise of rule k : Presentation series trust level to expert on a case-i in S

REFERENCES 1. RISKESDAS (Laporan Riset Kesehatan Dasar Nasional 2007 /National Health Research Report Year

2007), Agency for Health Research and Development Department of Health, Indonesia, 2008 2. Murray, L. J. C., The Global Burden of Disease: A Comprehensive Assesment of Mortality and

Disability From Diseases, Injuries, and Risk Factors In 1990 and Projected tp 2020, Harvard School of Public Health, World Health Organization, World Bank, Boston, 1996

3. Begum , S. M. U. , Ahmed , U. M. , and Funk , P., Case-Based Systems in Health Sciences - A Case Study in the Field of Stress Management, Wseas Transactions on Systems , Vol.8, 344-354, 2009

4. Ahmed , U. M. , Begum , S. , Funk , P. , Xiong , N. , and Scheele , V. , B. , Case-Based Reasoning for Diagnosis of Stress Using Enhanced Cosine and Fuzzy Similarity , Transactions on Case - Based Reasoning for Multimedia Data , Vol.1 , 3-19, 2008

5. Coyle, D., dan Doherty, G., Towards Ontologies for Technology in Mental Health Interventions, First International Workshop on Ontologies in Interactive Systems, Vol 1, 18-26, 2008

6. Kusumadewi, S , Hartati, S., Wardoyo, R dan Harjoko, A., Build and Design of Group Decision Support Systems for Amnesis, Diagnosis and Therapy of Mental Disorders, Journal Teknologi Industri Vol. XII No.1 , 7 – 18, 2008

7. Razzouk, D., Mari, J.J., Shirakawa. I., Wainer, J., dan Sigulem, D., Decision Support System for The Diagnosis of Schizophrenia Disorders, Brazilian Journal of Medical and Biological Research, Vol.39, 119-128, 2006

8. Tomar, S. P. P., Singh, R., Saxena, K. P., dan Sharma, J., Case Based Medical Diagnosis of Occupational Chronic Lung Diseases From Their Symptoms and Signs, International Journal of Biometrics and Bioinformatics (IJBB), Vol.5 , 216-224, 2011

9. Rismawan, T., Case-Based Reasoning untuk diagnosa penyakit THT, tesis, (Case-Based Reasoning for Ear-Nose-Throat disease diagnosis), Thesis, Universitas Gadjah Mada, Yogyakarta, 2012

10. Gu, D., Liang C. Y., Li, X. G., Yang, S., Zhang, P., Intelligent Technique for Knowledge Reuse of Dental Medical Records Based on Case-Based Reasoning. Journal Med. Syst, Vol.34, 213-222, 2010

11. Mancasari, A. U., Sistem Pakar Menggunakan Penalaran Berbasis Kasus Untuk Mendiagnosa Penyakit Saraf Pada Anak, skripsi (Expert System Using Case-Based Reasoning for Diagnosing Neurological Diseases On Children) skripsi, Department of Computer Science and Electronics, Univ. Gadjah Mada, Yogyakarta, 2012

12. Ignizio, P. J., Intorduction to Expert Systems: The Develpoment and Implementation of Rule-Based Expert Systems, Mc Graw-Hill, Inc., United States of America, 1991

13. Kolodner, J. L., An Introduction to Case-Based Reasoning, Artificial Intelligence Review, Vol.6, 3-34, 1992





Surabaya Indonesia

14. Shi, H., Xin, M., Dong, W., A Kind of Case Similarity Model Based on Case–Based Reasoning, International Conferences on Internet of Things, and Cyber, Physical and Social Computing IEEE, 453-457, 2011.



Wuryanto et al. 65


Surabaya Indonesia

Optimization of Job Shop Scheduling Problem Using Modified

Genetic Algorithm

Eto Wuryantoa* and Dyah Herawatiea,b

aDepartment of Mathemathics Faculty of Science and Technology bDepartment Technics Faculty of Vocational

Airlangga University, Surabaya 60115 Indonesia *Corresponding Author’s E-mail: [email protected]

Abstract. Job Shop Scheduling Problem (JSSP) is a timetabling/scheduling problem or a problem of time-based planning and combinatorial optimization. JSPP is a scheduling problem with j jobs and m machines, on each machine for each job has time and sequence. An job order on each machine with minimum makespan can be obtained using Genetic Algorithm that is one of the heuristic algorithms in combinatorial case. The main part of Genetic Algorithm (GA) is a principal of evolution. In the search process, it will generate a new solution applying genetic operator such as selection, crossover and mutation. This research had utilized Modified Genetic Algorithms (MGA) in JSSP problem that modify selection scheme, crossover, mutation and replacement strategy in GA. This method can avoid local optimal and discover optimal solutions efficiently. We used two data from Gen Cheng (1997) with 3 jobs 3 machines and benchmark data from Fang (1994) with 6 jobs 6 machines. By using those data, it is shown that the MGA can more efficiently determine the set of feasible regions than the GA.

INTRODUCTION Scheduling is a case of time planning and combinatorial optimization to minimize the needed time.

Scheduling issues can be found in various fields such as manufacturing and services. Job Shop Scheduling is one of a scheduling problem. In general, the issue of Job Shop scheduling can be described as follows: there j jobs and m machines, each job has a processing time and an order on each machine that aims to obtain a sequence of job-job on each machine to minimize the makespan (the time required to complete the entire job). Constraints for Job Shop Scheduling problems are: each job is scheduled for each machine only once, each job will be processed in each machine and each machine can only process one job at a time.

Genetic Algorithms (GA) is a methods that can be used to obtain a solution in Job Shop Scheduling problems. GA is a heuristic algorithm to solve combinatorial problems that inspired from the theory of evolution. In the process of evolution, individuals continuously change its genes to adapt to its environment. Only individuals who are strong can survive. This natural selection process modify genetic of individuals through the process of breeding. In the GA, this basic process has become the main concern, with the elementary thought: "How to get a better offspring". Operators used in GA are selection, mutation and crossover. While the modified genetic algorithm (MGA) is a scheme of GA selection, crossover, and strategies mutation that have modified. MGA aims to avoid local optimal and find the optimal solution efficiently (Modupe, et. al., 2014).

This research aimed to compare the effectiveness of the method of GA and MGA method in solving job shop scheduling problem. Indicators used are the value of minimum, maximum, mean and variance.

LITERATURE REVIEW GA is a group of methods to solve problems by using algorithm inspired by the processes of the new

evolutionary theory (Fang, 1994). The first step in the genetic algorithm is to describe the real problem into biological terminology. The Way of represention of the problem in the form of chromosomes is called encoding. There are several types of encoding such as encoding binary, permutations, value, and trees. Election of the right manner is depend on problems encountered. The operator used in GA are : (Mawaddah and Mahmudy, 2006)

Selection

The process selection will choose any individuals (cromosome) to be included in the activity of reproduction. The first step in the selection is calculation the value of fitness. The goal is to give greater

mailto:[email protected]



Wuryanto et al. 66


Surabaya Indonesia

reproductive opportunities for members of the population who have higher fitness value. Several methods of selection are roulette wheel, rankings and tournaments.

1. Crossover This crossover process produce two offspring from two parent chromosome selected. The resulting child

chromosome is a genes combination of the parent chromosome. 2. Mutation

This process is done by making changes to a gene or more than an individual. The purpose of the mutation is that individuals in increasingly diverse population. Mutations would be instrumental if the initial population there are few solutions that may be selected. Thus, the operation was very useful in maintaining the diversity of individuals in the population even though the mutation is not known what happened to the new individual.

Genetic Algorithm (GA)

A basic GA procedure has the following steps. 1. Define an objective/fitness function, and its variables. Set GA operations (such as population size,

parent/offspring ratio, selection method, number of crossovers and mutation rate). 2. Randomly generate the initial population. That is, obtain x1, x2, ..., xn. 3. Evaluate each chromosome in the initial population by the objective function. That is, compute

D(x1), D(x2), ..., D(xn). where D(xi), is the desirability function evaluated at the ith chromosome xi 4. Generate an offspring population, by GA operations (such as selection/mating, crossover, and

mutation). That is, generate a new set of n settings of the regressors, using the operations of selection, crossover, and mutation of x1 , ... , xn to obtain x1*, x2*, … xn*.

5. Evaluate each individual in the offspring population by the objective function. 6. Decide which individuals to include in the next population. This step is referred to as ―replacement‖

in that individuals from the current parent population are ―replaced‖ by a new population, whose

individuals come from the offspring and/or parent population. 7. If a stopping criterion is satisfied, then the procedure is halted. Otherwise, go to Step 4.

Modified Genetic Algorithm (MGA)

The MGA procedure is the same as that of GA, except that in the ith generation we add step D between steps 5 and 6 in the original GA procedure as follows: (Wan and Jeffrey, 2011)

D. Is the best chromosome in the offspring population also the best over the current parent population and does the best offspring have a D value > Dcutoff? That is, is max(D(xi )) > Dcutoff?

D-1. If no, directly go to Step 6. D-2. If yes, then define and implement local direction search with that chromosome as a starting point. The

local search will be ended when the objective function fails to increase. Find the chromosome with the largest desirability value and replace the best chromosome with the largest desirability value in the offspring population with this point. Then go to Step 6.

METHODS Operators used in this study are a partial schedule of exchange crossover, simple inverse mutation and

Roulette Wheel Selection for both GA and MGA. Encoding permutations use a representation that is based on the operation that proposed by Gen, Tsujimura, and Kubota (Gen and Cheng, 1997). This type of the coding permutation will directly produce an feasible individual because the generated genes randomly have sorted in order of operation and machine.

Determination of the performance of GA and MGA have used two groups of test data. The first data is a job shop scheduling with 3 jobs and 3 machines taken from Gen and Cheng (1997). The second data is the case of 6 jobs and 6 machines from the problem of benchmark Fisher Thompson of Fang (1994). All data can be seen in table 1 and 2.



Wuryanto et al. 67


Surabaya Indonesia

Table 1. Job Shop Scheduling problem with 3 jobs dan 3 machines

Tabel 13 Job Tabel 14 O0 Tabel 15 O1 Tabel 16 O2

Tabel 17 M Tabel 18 D Tabel 19 M Tabel 20 D Tabel 21 M Tabel 22 D

Tabel 23 0 Tabel 24 0 Tabel 25 3 Tabel 26 1 Tabel 27 3 Tabel 28 2 Tabel 29 2



O = Operation, M = Machine, D = Duration

Source : Gen dan Cheng, 1997

Table 2. Job Shop Scheduling problem with 6 job dan 6 machines

Job O0 O1 O2 O3 O4 O5 M D M D M D M D M D M D

0 2 1 0 3 1 6 3 7 5 3 4 6 1 1 8 2 5 4 10 5 10 0 10 3 4 2 2 5 3 4 5 8 0 9 1 1 4 7 3 1 5 0 5 2 5 3 3 4 8 5 9 4 2 9 1 3 4 5 5 4 0 3 3 1 5 1 3 3 3 5 9 0 10 4 4 2 1

O = Operation, M = Machine, D = Duration Source : Fang, 1994

Chromosome representation

The initial population of chromosomes are generated by generating genes randomly. On each chromosome was randomized a number of job as same as a machine number. This process is carried out as pop_size. By using chromosomes formed will be determined its operations by sorting operations based on the order of job. According to the first data set, if number 0, 1 and 2 respectively represent job 1, 2 and 3. an example of chromosomes generated is:

Job 1 2 0 0 2 2 1 0 1

The first Gen is job 2 with operation 1 because the first order on the job 2. The seventh gene is job 2 with

operation 2, the eighth gene is job 2 with operation 3. By the same way we can do on the job 1 and the job 3 such that is obtained :

Job 1 2 0 0 2 2 1 0 1 Operation 0 0 0 1 1 2 1 2 2

After we obtain the sequence of operations then we can determine machine by adjusting the jobs and

operations as shown in Table 3. So we have a matrice of JOM with J is job, O is operation and M is machine. For example, the first gene has J = 1 , O = 0 and M = 0 then we have code 100 where 1 is the second job, 0 is first operation, and 0 is first machine.

Table 3. matrice of JOM

Job 1 2 0 0 2 2 1 0 1 Operation 0 0 0 1 1 2 1 2 2 Machine 0 1 0 1 0 2 2 2 1



Wuryanto et al. 68


Surabaya Indonesia

To evaluate each chromosome, we will calculate the weight of each chromosome. The value of makespan is

able to utilize a machine Gantt chart. An example of the chromosome in table 2 can be ilustrated by machine Gantt chart as in Figure 1.

Fig. 1. Machine gantt chart

Code 100 and 201 is scheduled at the beginning of the schedule respectively in machine 0 and in machine 1,

next order 000 and 210 in machine 0. Put code 011 after 000 in machine 1, 222 after 210 in machine 2, 112 after 100 and 222 in machine 2, 022 after 011 and 112 in machine 2, at last 121 after 112 in machine 1, then obtained makespan 17.

RESULT AND DISCUSSION This study used two groups of test data: Data I in table 1 which consists of 3 jobs and 3 machines and Data II

in Table 2 contains 6 jobs and 6 machines. Both of these data are used to know the performance comparison between GA with MGA. Indicators used to examine the two methods are minimum value, maximum value, mean value and variance value. The calculation of these indicators is performed with the aid of a computer program written in the programming language PHP. We use the value of crossover probability (pc) 0.6, mutation probability (pm) 0.01 , Population size 15 individuals and 20 individuals and maximum generation 10, 30, 60. The output of the program can be seen in Table 3.

Table 4. Comparation result of GA and MGA

Data Pop. size

Nilai GA MGA

Maximum generation Maximum generation 10 30 60 10 30 60

Data I

15 Min, Max 11, 11 11, 11 11, 11 11, 11 11, 11 11, 11 Mean 11 11 11 11 11 11 Variance 0 0 0 0 0 0

20 Min, Max 11, 11 11, 11 11, 11 11, 11 11, 11 11, 11 Mean 11 11 11 11 11 11 Variance 0 0 0 0 0 0

Data II

15 Min, Max 61 , 69 55 , 69 58 , 65 60 , 69 58 , 63 58 , 60 Mean 64.6 62.1 61.1 64.4 60 59.3 Variance 7.3778 16.7667 5.4333 9.1555 2 0.45555

20 Min, Max 61 , 68 57 , 66 55 , 62 60 , 67 58 , 61 55 , 60 Mean 64.4 60.9 59.8 63.9 59.8 58.3 Variance 5.1555 9.8778 3.5111 6.76667 0.84444 2.23333

Results of the data I showed that both methods of GA and MGA have nearly the same value. All

combinations of population size and maximum generation give a same output for the minimum , maximum , mean and variance. While the data II, the mean value of both population size 15 and 20 have a downward trend proportionally to the increase in maximum generation value. But if we note the mean value of MGA is lower than the mean value of GA that can be looked at figure 2 and figure 3. Trend of variance of Data II is shown in figure 4 and figure 5. This trend shows that the mean and variance of MGA method is smaller than the GA method. But in computing practice MGA method requires a longer time and method of GA can find the optimal

1 2 3 4 5 6 7 8 9 10 11 12

m2

m1

m0 100 000 210

201 011 121

222 112 022

13 14 15 16 17



Wuryanto et al. 69


Surabaya Indonesia

makespan value in several trials. MGA method is more difficult to find the best makespan although in doing some experiments.

CONCLUSION By indicators of minimum value, maximum value, the mean and variance values, MGA method display a

better performance than the GA method. Data I with 3 jobs and 3 machines that was solved using php program for the value of the parameters, namely: pop_size = 15 and 20, pc = 0.6, pm = 0.01, the maximum generation = 10, 30 and 60 both methods GA and MGA give the makespan 11. On the other hand, the data II with 6 jobs and 6 machines with pop_size = 15 and 20, pc = 0.6, pm = 0.01, the maximum generation = 10, 30 and 60 the optimal makespan of methods GA is 55 while the method MGA is 58.

REFERENCES 1. Fang H-L., 1994, Genetic Algorithms in Timetabling and Scheduling, Department of Artificial Intelligence

University of Edinburgh. 2. Gen, M. and Cheng, R., 1997, Genetic Algorithms and Engineering Design, John Wiley and Sons, New

York 3. Mawaddah NK and Wayan Firdau Mahmudy, 2006. Optimasi Penjadwalan Ujian Menggunakn Algoritma

Genetika, Kursor, vol 2, no.2, pp.1-8. 4. Modupe, A. O., Omidiora E. Olusayo, and Olabiyisi S. Olatunde, 2014., Development of a University

Lecture Timetable using Modified Genetic Algorithms Approach, International Journal of Advanced Research in Computer Science and Software Engineering, Volume 4, Issue 9, September 2014.

5. Wan, Wen, and Jeffrey B. Birch, 2011. Using a modified genetic algorithm to find feasible regions of a desirability function. Quality and Reliability Eng. Int. 27(8): 1173-1182 (2011)

64.6

62.1 61.1

64.4

60 59.3

55

60

65

10 30 60

Me

an v

alu

e

Data II with Population size = 15

GA

MGA

64.4

60.9 59.8

63.9

59.8 58.3

55

60

65

10 30 60

Me

an v

alu

e


GA

MGA

7.3778

16.7667

5.4333 9.1555

2 0.45555 0

10

20

10 30 60

Var

ian

ce v

alu

e


GA

MGA5.1555

9.8778 3.5111

6.76667 0.84444 2.23333 0

10

20

10 30 60

Var

ian

ce v

alu

e


GA

MGA

Fig. 2. Mean value data II with pop_size 15 anrcd max_gen 10,30, 60

Fig. 3. Mean value data II with pop_size 20 and max_gen 10,30, 60

Fig. 5. Variance value data II with pop_size 20 and max_gen 10,30, 60

Fig. 4. Variance value data II with pop_size 15 and max_gen 10,30, 60

http://dblp.uni-trier.de/pers/hd/b/Birch:Jeffrey_B=

http://dblp.uni-trier.de/db/journals/qre/qre27.html#WanB11a



Werdiningsih et al. 70


Surabaya Indonesia

EXPERT’S PROMOTION ELIBILITY SYSTEM USING

BACKPROPAGATION ARTIFICIAL NEURAL NETWORK

Indah Werdiningsi1*, Army Justitia, Rini Sumiati, and Nur Hesti

Department of Mathematics, Airlangga University, Surabaya 60115 Indonesia *Corresponding Author’s E-mail: [email protected]

Abstract. Promotion can be interpreted as an acceptance of greater authority and responsibility than previous. Promotion is used to increase employee motivation to work in good habits as preferred by company. This research aims to create determining system of the expert‘s promotion. This study used back propagation artificial neural network. This method used 12 input layers, 12 hidden layers, learning rate 0.3. Binary sigmoid activation function used in hidden and output layer. The results show that the optimal accuracy achieved at 81.25% using 64 training data and 16 testing data. MSE value is obtained at 0.0023681619743600927 ≈ 0.0024 with threshold 0.15908713414381123. The results of experiments show that our method can be used to determinition expert‘s promotion accurately.

INTRODUCTION The company has several human resources that one of them are experts. To increase the maximum service

and quality, the company requires qualified experts as well. To ensure the experts working to meet the standards set by the company, the company must evaluate the performance of experts. Performance assessment is useful to review the past performance, check the capabilities of individual employees, quality improvement, compensation adjustments, placement decisions, the need for training and career planning[1]. It can be used to determine promotion. Promotion used to increase productivity and ensure the company's work in the company's success in reaching the target.

Since science is now developing rapidly, the artificial intelligence tried to imitate human intelligence. Artificial neural network is one of artificial intelligence methods[2]. Artificial neural networks are widely used to solve business and technical problems that involve prediction and also classification problems [3]. Artificial neural networks can train the network to recognize the pattern (historical data) which is used for training and responding the input patterns that is similar to the pattern used during training [4]. Backpropagation is one of the classic techniques used in artificial neural networks for data training [3]. By using backpropagation method, the result outputs are more accurate or closer accuracy in testing. This happens because during the training phase weights and biases are trained to adjust the pattern of historical data[5]. Several studies have been done using back propagation neural network method. These studies state that backpropagation can be used to predict the cleft lip [6], recognize Java letter patterns [7], and detect a disease, disorder, or cases that have past data[5].

Based on the problems already mentioned, it will be created a determining system of the expert‘s promotion using back propagation neural network. The system is expected to help the company determine professional promotion in order to improve the company's productivity.

DESIGN SYSTEM

Collecting Data and Information

Information and data used in the primary data form obtained directly from the company, namely PT. Fergaco Indonesia. The collected data were the criteria used in H2S expert‘s promotion process as well as an explanation

of each criterion related.

Processing Data

Each of the criteria that was used in decision support systems feasibility promotion experts drawn from the performance assessment criteria. These criterias include:

1. Technique ability 2. First aid ability 3. Presentation ability 4. Reporting ability 5. The behavior and decency of work 6. Compliance with regulations






Surabaya Indonesia

7. Company loyalty 8. Customer satisfaction 9. Leading team ability 10. Providing guidance and training ability 11. Marketing ability 12. Making work program and work evaluation.

Step-by-step Solution

Step-by-step to analyze the problems using neural network backpropagation were as follows: 1. Distribution training and testing data

The data has been obtained devided as training data and testing data. The amount of data that was obtained was 80 data. 64 data for training data and the remaining, 16 data, for testing data.

2. Design of network architecture Network architecture was composed of 3 layers, that is input layer, hidden layer and output layer. Before being processed in input layer, data must be normalized first. The design of network architecture can be seen in Figure 1.

3. Training Training aimed to obtain the appropriate weight to be used in testing. Step-by-step training were as follows : Step 0. Initializing weight with small random number Step 1. If the terminating condition is unfulfill, then do steps 2-9 Step 2. For each training data, steps 3-9 is done

X1 X2 X121

Z1 Z2 Zn1

Y1

V10

V11V22

Vn 12

V21V20

Vn 2

Vn 0Vn 1

V2 12

V12 V1 12

W10W11 W12

W1 n

...

... Input Layer

Hidden layer

Output layer

Figure 4 Proposed Backpropagation Architecture Design

Phase I : Feedforward Step 3. Each input unit (xi, i=1, 2, ...,n) receives the signal and forward it to the hidden units after it Step 4. Counting all output in hidden unit (zj, j=1, 2, ..., p) using activation function

n

ijiijj vxvnetz

10_

(1)

)_( jj netzfz

(2) Step 5. Counting all output in output unit (yk,, k=1,...,m) using activation function

p

jkjjkk wzwnety

10_

(3)

)_( kk netyfy (4) Phase II : Backpropagation Step 6. Calculating k output unit variable based on error in each output unit (yk, k=1,...,m)

)_(')( kkkk netyfyt (5)





Surabaya Indonesia

The k factor is the error unit used in the change of the layer weight below it Step 7. Calculating weight change rate wkj (which is used for changing weight wkj) with learning rate α

jkkj zw (6)

Step 8. Calculating j hidden unit variable based on error in each hidden unit (zj ,j=1,..., p)

m

kkjkj wnet

1

_ (7)

j hidden unit variable :

)_('_ jjj netzfnet (8) Calculating weight change rate vji (which is used for updating weight vji)

ijji xv (9) Phase III : Updating weight and bias Step 9. Updating all weight changes

kjkjkj wlamawbaruw )()( (10) Weight change lines leading to the hidden unit

jijiji vlamavbaruv )()( (11) Step 10. Checking termination condition

Data are normalized in range [0.1 ; 0.9] first, if its activation function is sigmoid biner. Data

normalization equation is described in equation 12 [3].

(12)

Meanwhile, if the activation function is bipolar sigmoid, then data were normalized in the range [-0.9 ; 0.9]. Data normalization equation is described in equation 13[8]

(13)

The weight is initialized using a random number between -0.5 to 0.5. Parameters consist of max epoch, target error, and learning rate (α). For termination condition is defined as iteration = maximum epoch or

MSE ≤ error target. Each epoch represents one calculation for all data on training data [9]. 4. Testing

Testing conducted after training data is performed. Testing data used to examine the data using weight obtained from final weight training.

RESULTS AND DISCUSSION Two stages in this study are collecting and data processing, and classification.

Collecting and data processing

This stage consists of collecting data in the company and interviewing. Expert‘s positions on PT. Fergaco

Indonesia are Junior 1, Junior 2, Junior 3, Junior 4, Intermediate 1, Intermediate 2, Intermediate 3, Intermediate 4, Senior 1, Senior 2, Senior 3 Senior 4, and Team Leader. Targets contained in the training data and data testing are decision obtained from previous data. Number 1 represents a decision to be promoted and number 0 represents a decision not to be promoted. The data which is used has 12 criteria for performance assessment. Each performance assessment criteria (qualitative) will be processed into input on the system and will be given the value of a scale of 1 to 6.

Classification

Before performing classification, must first design the network architecture. Input layer which are used are 12 single output layers Experiments using four scenarios. Testing are used to detemine the effect of learning rate

( ), the numbers of hidden layer, and activation function.

a. Test the effect of sigmoid biner activation function in the hidden layer and output layer.





Surabaya Indonesia

Based on the experiment, the most optimal accuracy reached when using 12 hidden layers and α = 0.3. The accuracy achieved at 81.25% and MSE value was 0.0023681619743600927 ≈ 0.0024. The results can be seen in

Figure 2.

b. Test the effect of sigmoid bipolar activation function in hidden layer and sigmoid biner activation function in output layer

Based on the experiment, the most optimal accuracy reached at 68.75% when using 12 hidden layers and α = 0.5. MSE value was 0.0022605871413535084 ≈ 0.0023. The results can be seen in Figure 3.

Figure 2. The result of accuracy result based on hidden layer and learning rate for biner sigmoid activation function

Figure 3. The result of accuracy result based on hidden layer and learning rate for sigmoid bipolar activation function in hidden layer and sigmoid biner activation function in output layer

c. Test the effect of sigmoid biner activation function in hidden layer and sigmoid bipolar activation function in

output layer Based on the experiment, the most optimal accuracy reached at 81.25% when using 6 hidden layers and α =

0.5. MSE value was 0.005239276489903087 ≈ 0.0054. The results can be seen in Figure 4. d. Test the effect of sigmoid bipolar activation function value in hidden layer and sigmoid bipolar activation

function in output layer Based on the experiment, the most optimal accuracy reached at 81.25% when using 6 hidden layers and

α = 0.5. MSE value was 0.0037260824944689466 ≈ 0.0037. The results can be seen in Figure 5.

56.25 56.25 56.25

81.25

62.25 56.25

0

20

40

60

80

100

0.3 0.5 0.7Acc

ura

cy (

in p

erc

en

t %

)

Learning Rate

The result of accuracy based on hidden layer and learning rate

6neuron…

50 56.25 56.25 56.25

68.75 56.25

0

20

40

60

80

0.3 0.5 0.7

Acc

ura

cy (

in p

erc

en

t %

)

Learning Rate


6 neuronsHidden…





Surabaya Indonesia

Figure 4. The result of accuracy result based on hidden layer and learning rate for sigmoid biner activation function in hidden layer and sigmoid bipolar activation function in output layer

Figure 5. The result of accuracy result based on hidden layer and learning rate for sigmoid bipolar activation function in hidden layer and sigmoid bipolar activation function in output layer

Based on 4 testing scenarios, binary activation function in the hidden layer and output layer produced the

most optimal accuracy. The accuracy obtained at 81.25% and MSE value was 0.0023681619743600927 ≈ 0.0024. The accuracy results at every activation function can be seen in Figure 6.

Threshold obtained in the training was 0.15908713414381123. It was used to determine the output of the system. If the output was greater than equal threshold, the output was 1 and otherwise, smaller than the threshold, the output was 0. Output 1 means that experts are concerned worthy promoted and 0 represents that not worth promoted. The most optimal accuracy results can be seen in Table 1.

50

81.25

62.5 62.5 56.25 56.25

0

20

40

60

80

100

0.3 0.5 0.7

Acc

ura

cy (

in p

erc

en

t %

)

Learning Rate


6neurons…

50

81.25

56.25 62.5

56.25 56.25

0

20

40

60

80

100

0.3 0.5 0.7Acc

ura

cy (

in p

erc

en

t %

)

Learning Rate


6neuron…





Surabaya Indonesia

Figure 6. The result of accuracy result based on activation function in hidden layer and output layer

Table 1. Output of Testing

CONCLUSION In this study, we have proposed backpropogation artificial neural network to expert‘s promotion elibility.

Two stages in this study are collecting and data processing, and classification. According to the experimental results, the proposed method is efficient for expert‘s promotion elibility. Design of network architecture is used is 12 neurons in input layer, 12 neurons in hidden layer, 1 neuron in output layer, activation function used is sigmoid biner in hidden and output layer. The correct elibiltion rate of proposed system is 81.25 % and MSE is 0.0023681619743600927 with learning rate is 0.3 and threshold is 0.15908713414381123.

REFERENCES 1. Rifai, Veithzal. 2005. Performance Appraisal Sistem yang Tepat untuk Menilai Kinerja Karyawan dan

Meningkatkan Daya Saing Perusahaan. Jakarta: PT RajaGrafindo Persada. 2. Russel, Stuart J., Peter Norvig,2010, "Artificial Intelligence, a modern approach―, Third Edition,

Prentice Hall, New Jersey. 3. Siang, Jong Jek. 2005. Jaringan Syaraf Tiruan dan Pemogramannya Menggunakan Matlab.

Yogyakarta: Andi Offset.

56.25 56.25 56.25

81.25 81.25

68.75

81.25

56.25

0

20

40

60

80

100

Biner - Biner Bipolar - BinerBiner - Bipolar Bipolar -Bipolar

Acc

ura

cy (

in p

erc

en

t %

)

Learning Rate


6 neuronsHidden…

Data Sistem Hasil

0.0 0.20288912205809836 0.0

0.0 0.21090672524988557 0.0 1.0 0.13187780214649733 1.0 1.0 0.13403948844736566 1.0 0.0 0.20189433609478197 0.0 0.0 0.17566213174087253 0.0 1.0 0.1598839018449736 0.0 1.0 0.15495231800244766 1.0 0.0 0.20311538982314517 0.0 1.0 0.13755258481442645 1.0 1.0 0.17035686145950926 0.0 0.0 0.1830250668579857 0.0 0.0 0.16302267065537737 0.0 1.0 0.1872851914024038 0.0 0.0 0.20288912205809836 0.0 0.0 0.21090672524988557 0.0





Surabaya Indonesia

4. Örkcü, H. Hasan. 2011. Comparing performances of Backpropagation and Genetic Algorithms in The Data Classification. Expert Systems with Applications. 38.

5. Wilson, D. Randall dan Martinez, Tony R. 2001. The Need for Small Learning Rates on Large Problems. Proceedings of the 2001 International Joint Conference on Neural Networks (IJCNN‘01). 115-119.

6. Puspita, Analia. 2007. Penggunaan Jaringan Saraf Tiruan Metode Backpropagation untuk Memprediksi Bibir Sumbing. Seminar Nasional Teknologi. 2007.

7. Nurmila, Nazla. 2010. Algoritma Backpropagation Neural Network untuk Pengenalan Pola Karakter Huruf Jawa. Jurnal Masyarakat Informatika. Volume I

8. Zhang, G., Pattuwo, B.E., dan Hu, M.Y. 1997. Forecasting with Artificial Neural Networks : The State of the Art. Elsevier International Journal of Forecasting. 14 (1998) 35-62.

9. Duda, Richard O. 2001. Pattern Classification. USA: John Wiley&Sons,inc.



Purwanti et al. 77


Surabaya Indonesia

Implementation of Journal Classification Information Retrieval System

with K-Nearest Neighbor

Endah Purwanti*, Badrus Zaman

Department of Mathematics Airlangga University, Surabaya 60115 Indonesia

*Corresponding Author’s E-mail: [email protected]

Abstract. With so many document available, from academic to non-academic causes undergraduate student felt difficult in finding which document was right for their literature. Hence, they need a system that provide them with simplicity to classififying based on suitable category. In this research will be expected to provide a system that could classified a document by it‘s suitable category.In this journal classification system was done by using 3 phases. The first phase was to collect all the data and information which is a collection of journal. The Second phases was to analyse the system which include, processing document with text mining, weighting every token with term frequency-inverse document frequency (TF-IDF), calculate the similarity for each document with csinus similarity, and classified the document with k-nearest neighbour classifier. Third phase was to implement the system in desktop using Netbeans and MySql. The fourth phase was to evaluate the system with comparing the F-Measure of the used k value. Based on trials with using 180 documents which consist of 40 train documents for each categories which are Physical Sciences and Engineering, Life Sciences, Health Sciences, dan Social Sciences and Humanities and 20 testing documents produces Recall 0.524 , Precision 0.501, and F-Measures 0.5193 at k value equals to 43.

INTRODUCTION Information Retrieval (IR) is the science of searching for information within relational databases,

documents, text, multimedia files, and the World Wide Web. The applications of IR are diverse; they include but not limited to extraction of information from large documents, searching in digital libraries, information filtering, spam filtering, object extraction from images, automatic summarization, document classification and clustering, and web searching. The breakthrough of the Internet and web search engines have urged scientists and large firms to create very large scale retrieval systems to keep pace with the exponential growth of online data.

With the explosive growth of the textual information from the electronic documents and World Wide Web, proper classification of such enormous amount of information into our needs is a critical step towards the business success. Recently, numerous research activities have been conducted in the field of document classification, particularly applying in spam filtering [1-3], emails categorization [4], website classification [5], formation of knowledge repositories [6], and ontology mapping [7].

However, it is time-consuming and labor intensive for a human to read over and correctly categorize an article manually. Attempts to address this challenge, automatic document classification studies are gaining more interests in text mining research recently. Consequently, an increasing number of approaches have been developed for accomplishing such purpose, including k-nearest-neighbor (KNN) classification [8], Naïve Bayes classification [9, 15], support vector machines (SVM) [10], decision tree (DT) [11], neural network (NN) [12], and maximum entropy [13].

Information retrieval system that is perfect is a system that is able to obtain documents relevant to the query is entered without displaying documents that are not relevant. However, the information retrieval system would not be perfect, this is due to irrelevant whether or not a document obtained in the system not all users consider the documents relevant to the query generated (Hiemstra, 2009). In this study, using a retrieval system with a supervised learning method of learning or learning that documents collected supervised. Supervised question is to provide a label or category in the class of training documents that became the basis for the classification of documents further or new documents. While the methods used in this research is the method of K-Nearest Neighbor (KNN) as the classification of documents. By using the cosine similarity, classification of documents will be more easily done because using the cosine similarity document will be processed by converting documents into vector shapes and by doing a comparison between the document vector can generate a value of similarity between documents. Then using the K-Nearest Neighbor method can classify the documents that have produced similarity value by looking at the value of the greatest similarity and the majority of the training document as a comparison that appears in the sphere of values of k previously been initialized by the user



Purwanti et al. 78


Surabaya Indonesia

Method

Information retrieval system contained data processing such as text mining. In text mining for document processing is done in four main stages. The first is a collection of words into the basic form of words collected by using stemming. The third is a feature selection, in this stage made a stoplist that contains a collection of words that are not relevant to the content of each document processed. Then do the stopword removal to eliminate the words contained in the stoplist. And the last is the pattern discovery. Pattern discovery is an important step in the processing of retrieval system, because in this stage determined the processing to be performed on the documents that have been collected.

The collected data is processed through four phases. First phase, processing of documents with text mining, start from tokenisasi, filtering, and stemming using the algorithm porter. Second, analyzing data from calculations using the cosine similarity method to determine the level of similarity between documents. Last phase, analyze the data from the classification results using k-Nearest Neighbor for further assessment based on the recall, precision and f-Measures to assess the success of the system.

Text Mining

Text mining is a knowledge-based process in which users interact and work with a set of documents by using multiple analysis tools. In the text mining process stage is divided into four main stages, namely text processing, text transformation, feature selection, and pattern discovery. Text preprocessing is tokenisasi stage which is the process of splitting the text into a form of words or commonly referred to as a token.Text stemming transformation is the stage that serves to transform words berimbuhan into basic shapes. In this process will use the algorithm porter. Because the porter algorithm is an algorithm that according to the document in English.

Feature selection is a step that aims to reduce the dimensions of a collection of text resulting from the transformation stage. In other words, delete the words that are not related to the content of the document or by using stopword removal. Pattern discovery is the stage of determination of test patterns to be processed. Supervised learning is a learning technique that uses a label or category class given on training data (training) which is then used as the basis for the classification of the new data.

Term Frequency dan Inverse Document Frequency

Term frequency is is a numerical statistic that is intended to reflect how important a word is to a document in a collection. Inverse document frequency (IDF) is an important measure weighting of a word in a document seen in the whole document globally. Function to calculate the value of TF can be seen in the following equation 2.1

(2.1)

f(dt): the emergence of the word t in the document d. Weighting IDF considers that the weight of a word will be even greater if the word is more and more frequently appear in the document, not in a lot of documents. Function to calculate the value of the IDF can be seen in the following equation 2.2

⁄ (2.2)

df(t): The number of documents that have the word t. While the function to calculate the TF-IDF according to the formula 2.3.

(2.3)

Cosine Similarity

Cosine similarity is the most common calculation used in the calculation of similarity between documents. If x and y are vectors documents, then.

| | | | (2.4)

Where, xy is the dot product by the equation

https://en.wikipedia.org/wiki/Document



Purwanti et al. 79


Surabaya Indonesia

∑

(2.5)

While || X || is the length of the vector x, where the equation of || x || is

| | √∑

√

(2.6)

K-Nearest Neighbor

K-Nearest Neighbor (KNN) is one method of vector space model that serves to classify an object [8]. KNN classification on the basis of existing boundary determination or scope locally. Basically KNN classification will be based on a hypothesis in which a test document d will have the same label category or categories of training documents that plays within the scope of k that surrounds the document d. The algorithm of k-Nearest Neighbor classification can be seen in Figure 2.2.

Figure 2.2 K-Nearest Neighbor algorithm

For the classification by using 1NN, the classification will be running not so strong. Classification of each

test document depends on the class of each training document, which could occur errors or atypical categories. Classification using KNN with k> 1 will be stronger. For the parameter k in KNN often chosen based on experience or knowledge of the classification problem faced. Preferably to use the k value that gives the results of odd to reduce the possibility of different categories in the same amount or the amount of the series [8].

RESULTS AND DISCUSSION User is prompted to conduct training on the training documents. Form to perform the processing of

training documents can be viewed according to Figure 3.1.

Figure 3.1. User Interface Clasified Query Document Clasified

After entering user training documents with the same amount for each category. Then the user can

perform the processing of new documents or document query to classify the document to be



Purwanti et al. 80


Surabaya Indonesia

categorized into one of four categories by making a comparison between a query document with training documents that have been processed previously.

Evaluation is done by looking at the system from the calculation of recall, precision and F-Measures against the results of trials that have been carried out by using 20 document testing (Fig. 3.2) and using the value of k = {37, 41, 43}. From the test results by using the k value showed an average value of recall, precision and F-measures listed in Figure 3.3.

Figure 3.2 Document Testing

Figure 3.3 Graph Average Value Recall, Precision and F-Measures Results Test System

CONCLUSION From the research conducted, it is concluded that in designing and building the system is the study of

literature against the cosine similarity and k-Nearest Neighbor, data collection and information such as documents training consists of 160 documents with the division of 40 documents for each category is Physical Sciences and Engineering , Life Science, Health Science, and Social Sciences and Humanities. Then the system design, desktop-based systems development, and testing and evaluation of the journal classification system sistem.Uji try performed using training documents a total of 160 documents and document testing as many as 20 documents.

From the test results showed that the system of testing used 20 documents can be classified according to the actual category. However, in some cases documents 4, 6, 9, 10, 12, 13, 14, and 16 errors in the classification. This is because in addition to the value of k is used also because of the number of tokens in the training documents that affect the value of cosine similarity test documents and ultimately affect the outcome of KNN classification. Evaluation system shows the level of success in classifying a document with the value of k = 43 which resulted in the recall value of 0539, precision of 0501, and the F-Measures at 0.5193.

REFERENCES 1. S.J. Delany, P. Cunningham, and L. Coyle, ―An assessment of case-based reasoning for spam

filtering‖, Artificial Intelligence Review Journal, Vol. 24, No. 3-4, 2005, pp. 359-378. 2. P. Cunningham, N. Nowlan, S.J. Delany, and M. Haahr, ―A case-based approach in spam filtering that

can track concept drift‖, In Proceedings: The ICCBR‟03 Workshop on Long-lived CBR Systems, Trondheim, Norway, 2003.

3. K. Wei, A naïve Bayes spam filter, Faculty of Computer Science, University of Berkely, 2003. 4. B. Kamens, Bayesian filtering: Beyond binary classification. Fog Creek Software, Inc., 2005. 5. M.I. Devi, R. Rajaram, and K. Selvakuberan, ―Generating best features for web page classification‖,

Webology, Vol. 5, No. 1, 2008, Article 52. 6. M. Hartley, D. Isa, V.P. Kallimani, and L.H. Lee, ―A domain knowledge preserving in process

engineering using self-organizing concept‖, In Proceedings: ICAIET 06. Sabah, Malaysia: Kota Kinabalu, 2006.

0.470

0.480

0.490

0.500

0.510

0.520

0.530

0.540

0.550

37 41 43

Nila

i Rec

al, P

reci

sio

n, F

-M

easu

res

k Value

Recall Precision



Purwanti et al. 81


Surabaya Indonesia

7. X. Su, A text categorization perspective for ontology mapping, Norway: Department of Computer and Information Science, Norweigian University of Science and Technology, 2002.

8. E.H. Han, G. Karypis, and V. Kumar, Text categorization using weight adjusted k-nearest neighbor classification, Department of Computer Science and Engineering, Army HPC Research Center, University of Minnesota, 1999.

9. A. McCallum, and K. Nigam, ―A comparison of event models for naïve Bayes text classification‖,

Journal of Machine Learning Research, Vol. 3, 2003, pp. 1265–1287. 10. S. Chakrabarti, S. Roy, and M.V. Soundalgekar, ―Fast and accurate text classification via multiple

linear discriminant projection‖, The VLDB Journal The International Journal on Very Large Data

Bases, 2003, pp. 170–185. International Journal of Software Engineering and Its Applications Vol. 5, No. 3, July, 2011.

11. J.R. Quinlan, C4.5: programs for machine learning, Morgan Kaufmann Publishers Inc., San Francisco, CA, 1993.

12. S. Wermter, ―Neural network agents for learning semantic text classification‖, Information Retrieval, Vol. 3, No. 2, 2004, pp. 87-103.

13. K. Nigam, J. Lafferty, and A. McCallum, ―Using maximum entropy for text classification‖, In Proceedings: IJCAI-99 Workshop on Machine Learning for Information Filtering, pp. 61–67, 1999.

14. S.L. Ting, W.H. Ip, Albert H.C. Tsang, Is Naïve Bayes a Good Classifier for Document Classification?, International Journal of Software Engineering and Its Applications Vol. 5, No. 3, July, 2011



82


Surabaya Indonesia

ORAL PAPER OF BIODIVERSITY (OBDV)



Junairiah et al. 83


Surabaya Indonesia

Bioactive Compounds of The Moss Homaliodendron flabellatum

(Smith) Fleiscl

Junairiaha*, Tri Nurhariyatia, Suaibaha, Ni’matuzahroha, Lilis Sulistyorinib

a Department of Biology,Faculty of Science and Technology, Airlangga University, Surabaya, Indonesia b Faculty of Public Health, Airlangga University, Surabaya, Indonesia


Abstract. The moss have potential as a medicinal plant. Homaliodendron flabellatum is one of the moss contained in the Cangar forest Batu. The objective of this research to identify bioactive compounds was contained in H. flabellatum. The moss was extracted with hexane, ethyl acetate and methanol, then was analyzed by Gass Chromatography Mass Spectra. The results showed that H. flabellatum hexane extract contains 17 compounds, ethyl acetate extract contains 22 compounds, while the methanol extract contains no compound. Each of the main components of hexane and ethyl acetate extraxt is hexadecanoic acid and 2-methyl propanoic acid.

INTRODUCTION Bryophyta is the second largest group in the plant kingdom. Bryophyta consists of 25,000 species. This plant

can be found in various habitats (Glime 2007 and Asakawa et al., 2013). Generally Bryophyta consist of mosses, liverworts and hornworts. This plant does not belong to flowering plants group, photoautotrof, and have gametophyte dominate phase. Sporophytes is produced from fertilization of egg and sperm cells. Egg cells are produced by archegonium, while sperm cells are produced by antheredium (Ardiles et al., 2009). Bryophyta can be distinguished from other plants because of their certain characteristics. These plants can live in a various ecosystems, grow in nutrient-poor conditions, and adapt quickly to changes in humidity (Slack, 2011).

Bryophyta have biological activities including insect-repelling, insecticidal, cytotoxic, phytotoxic, allergenic (dermatitis), neurotrophic, anti-obesity, muscle-relaxing, antibacterial, antifungal, antitumor, and anti-HIV properties (Saxena and Harinder, 2004; Asakawa, 2009). This is due to secondary metabolites contained in Bryophyta, such as benzoids, bibenzyles, bis-bibenzyles, fatty acids derivates, flavonoids, phenylpropanoids, S and N-containing compounds, and terpenoids (Sabovljevic et al., 2009; Zhu et al., 2006; Chun- Feng and Hong Xiang, 2009).

Homaliodendron flabellatum is one type of the moss, from Hypnales order, Neckeraceae family, and habitat is a tree. So far, there has been no research on the chemical compounds contained in H. flabellatum. The purpose of this study was to identify chemical compounds of the moss H. flabellatum by using Gass Chromatography Mass Spectra (GC-MS) method. The results hopefully could be used as a database for the development of moss potential as a medicinal plant, so as to improve public health

MATERIALS AND METHODS H. flabellatum was collected from the Cangar forest, Batu, Malang, East Java, Indonesia. Mosses was

washed with water, dried, and diced up into powder. H. flabellatum weighed and divided into three parts. Each part was extracted with 100 ml of n-hexane, ethylacetate and methanol solvent. Extraction was done by maceration method for three days. Each solvent was repeated four times. Extracts were weighed and then the chemical compounds were analyzed and identified by GC-MS. Identification of compounds of n-hexane, ethyl acetate, and methanol extract is carried out using GC-MS. Brand Agilent GC-MS is 6890 A. The samples are injected 1µl . Detector of GC is MSD Agilent 5973 inert. GC conditions include a flow rate of 1.3 ml / min. Temperature of injection is 2800C, the temperature of detector (Aux: 2800C), MS Source is 2300 C, MS quadrupole is 1500C. The stationary phase is 5% phenyl methyl ciloksan. Types of capillary column is HP 5. Size column is 30 μm x320 μm x0.25 μm. The mobile phase is helium ultra pure gas

RESULTS The moss were used in this study is H. flabellatum. To know more about this moss, while the description is

as follows. H. flabellatum had gametophyte with life form of an open turf which had branching. Moss gametophyte phase had length of 45-72 mm, the leaf was 1.64 mm long, 0.75 mm width, with dark green, pale



Junairiah et al. 84


Surabaya Indonesia

green, and yellowish green color. The shape of leaf was oval (Figure 1). Leaves apical was pointed and the edge was serrated. Leaf cell was shaped rhombus (Figure 2). Sporangium was attached to the trunk, seta was 4 mm long, and the capsule was 3.5 mm long and 1.3 mm width. Spore was spherical, brown colored, had uneven surface, and the diameter was 0.0125 mm (Figure 3).

. (a) (b)

Figure 1. (A) Phase gametophyte of H. flabellatum , (b) leaves of H. flabellatum (a) (b)

Figure 2. (a) Apical leaves of H. flabellatum , (b) leaf cell shape of H. Flabellatum (a) (b)

Figure 3. (a) Sporangium of H. flabellatum, (b) spores of H. flabellatum The result of H. flabellatum moss extraction can be seen in Table 1.

Table 1. Extraction H. flabellatum using n-hexane, ethylasetate, and methanol solvent.

No Solvent Volume Extract Weight (g) Extract Color 1 n-hexane 400 0.006 Colourless 2 ethylacetate 400 0.017 Yellow 3 methanol 400 0.128 Green

DISCUSSION Maceration method was a method often used for extraction process, because it was easy to do and did not

require expensive equipment. Margaretha et al. (2012) reported that the maceration method was used to isolate the bioactive compounds of propolis on Trigona spp. Dragici and Rapeanu (2011) reported that the maceration



Junairiah et al. 85


Surabaya Indonesia

was used to obtain polyphenols in red wine. GC-MS analysis of n-hexane and ethyl acetate extracts resulted in the form chromatogram profiles and compound identification. Chromatogram profile of H. flabellatum n-hexane and ethyl acetate extracts was presented on Figure 4 and 5. GC-MS analysis of H. flabellatum methanol extract was not identified, it was likely due to the chemical compounds could not be dissolved in methanol (polar solvent) but soluble in n-hexane and ethyl acetate or non-polar solvents

Figure 4. Chromatogram profiles of H. flabellatum n-hexane extract

Figure 5. Chromatogram Profile of H. flabellatum ethyl acetate extract

Identification of bioactive compounds of H. flabellatum n-hexane and ethyl acetate extract was presented in tables 2 and 3.

Table 2. Bioactive compounds of H. flabellatum n-hexane extract

Peak Retention Time Compound Area (%) 1 37,357 Propanoic acid 2,776 2 41,944 E-Heptadecanal 5,243 3 43,060 6,10,14-trimethyl-2-Pentadecanone 7,691 4 44,571 7,9-Di-tert-butyl-1-oxaspiro(4,5) deca-6,9- diene-2,8-dione 9,058 5 44,710 Hexadecanoic acid 15,979 6 45,593 Hexadecanoic acid 12,518 7 45,989 Cycloeicosane 8,523 8 46,035 Hexadecanoic acid 5,139 9 47, 693 2-methyl-Tricosane 1,934 10 48,019 2-methyl-Tricosane 3,719 11 48,476 Methylester Octadecanoic acid 1,070 12 49,623 1-Docosene 6,287 13 51,397 Hexadecanoic acid 2,049 14 52,955 1-Nonadecane 3,300 15 55,357 Dioctyl ester phtalic acid 1,436 16 56,039 Cycloeicosane 1,748 17 65,709 Diploptene 11,531



Junairiah et al. 86


Surabaya Indonesia

Table 3. Bioactive compounds of H. flabellatum ethyl acetate extract Peak Retention Time Compound Area (%)

1 37,434 Propanoic acid 66,549 2 37,636 Cyclohexanol 0,579 3 39,883 Heptadecane 0,338 4 41,944 1-octadecene 0,613 5 42,099 Octadecane 0,340 6 42,812 2- Hexadecene 0,552 7 42,920 Neophytadiene 3,995 8 43,067 2- Hexadecene 1,587 9 43,439 Neophytadiene 0,713 10 43,796 2-Hexadene-1-ol 1,244 11 45,423 Dibutyl phtalate 1,210 12 45,578 Hexadecanoic acid 1,675 13 45,981 1-Nonadecene 0,602 14 49,103 Drimenol 0,746 15 51,389 (Z)- Methyl.alpha.-Bromo-.beta-nitrocinnamate 1,002 16 52, 482 9-octadecenamide 0,350 17 54,163 Linoleic acid 1,568 18 54, 287 Linolenic acid 1,430 19 55,357 Bis 92-ethylhexyl)phtalate 1,480 20 56,573 Arachidonic acid 2,343 21 60,998 Solanesol 7,529 22 65,717 22- Imino-3-phenyl-4-thiazolidinone hydrochloride 3,555

N-hexane extract of H. flabellatum contained 17 compounds (Figure 6 and Table 2). The main components

contained in this extract was hexadecanoic acid. H. flabellatum ethyl acetate extract contained 22 compounds (Figure 7 and Table 3). The main component was propanoic acid. Hexadecanoic acid had common name of palmitic acid, with molecular formula of C6H32O2 and molecular weight of 256. N-hexadecanoic acid had biological activities as an antioxidant, hypocholesterolemic, nematicide, pesticide, lubricant, anti-androgenic flavor, hemolytic, and 5-alpha reductase inhibitors (Wankhede and Manik , 2015).

Hexadecanoic acid could also be found in the brown alga Padina pavonica and Homophysa triquetra. The active ingredients of fatty acids could be used as antibacterial agent against Escherichia coli, Pseudomonas aeruginosa, Salmonella typhimurium, Shigella boydii, Staphylococcus aureus, and Stretomyces antibioticus (El Shoubaky and El Rahman Salem, 2014). Neophytadiene molecular formula was C20H38, while the molecular weight was 278.51572 g/mol. Neophytadiene had potential as antibacterial, antioxidant, and prophylactic activities (Venkata et al., 2012). Bai et al. (2014) reported that chloroform extract of Acacia nilotica contained neophytadiene. Propanoic acid molecular formula was C3H6O2, while the molecular weight was 74.08 g/mol. Propanoic acid also was called ethanecarboxylic acid. Propanoic acid had potentials as antibacterial, antifungal, pesticides, and pharmaceuticals.

CONCLUSION The main components of n-hexane and ethyl acetate extracts were respectively hexadecanoic acid and

propanoic acid. Further research was required to evaluate the biological activities of both n-hexane and ethyl acetate extracts from H. flabellatum.

ACKNOWLEDGMENTS Thanks to DIPA DITLITABMAS-2015, Contract Number 519 / UN3 / 2015, March 26, 2015.

REFERENCES 1. Glime J.. Bryophyte Ecology. Volume5. Uses(Ebook sponsered by Michigan Technological University

and the International Association of Bryologists. 2007).URL:htt:// www.bryoecol.mtu.edu Viewed: May 15, 2011.

htt://

http://www.bryoecol.mtu.edu/



Junairiah et al. 87


Surabaya Indonesia

2. Asakawa Y, Ludwiczuk, Agnieszka, Nagashima, and Fumihiro. Chemical Constituents of Bryophytes (Springer, 2013). XVIII: 769-789

3. Ardiles VJ, Cuvertino, and F Osorio. Briofitas de los Bosques Templados de Chile; Guia de Campo (CORMA, Chile, 2009) 168pp.

4. Slack N. The Biological Value of Bryophytes as Indicators of Climate Change. In: Z Tuba, N.G. Slack and L.R. Stark (eds.), Bryophyte and Climate Change(Cambridge University Press, Cambridge, UK., 2011). , pp. 3-12

5. Saxena D and Harinder H. Uses of Bryophytes ( Resonance, 2004). 1: 56-65 6. Asakawa Y, Ludwiczuk A, Nagashima F, Toyota M, Hashimoto T, Tori M,.Fukuyama F, and

Harinantenaina L. Bryophytes: Bio and Chemical Diversity, Bioactivity and chemosystematics (Heterocycles,2009) 77: 99-150

7. Sabovljevic A, Sabovljevic M, and Jokovic N. In Vitro Culture and secondary Metabolite Isolation in Bryophytes. In: S.M. Jain and P.K. saxena (eds), Protocols for In Vitro cultures and Secondary Metabolite Analysis of aromatic and Medicinal Plants. Methods in Molecular Biology (Humana Press, New York, USA, 2009) pp. 117-128

8. Zhu R, Wang D, Xu L, Sh Ri, Wang J, and Zheng M. Antibacterial Activity in Extracts of Some Bryophytes from China and Mongolia (Journal of Botanical Laboratory, 2006) 100: 603-615.

9. Chun Feng X and Hong xiang L. Secondary Metabolites in Bryophytes: an Ecological aspect.(Chemistry and Biodiversity, 2009). 6: 303-312.

10. Margareth I, Suniati DF, Herda E, Mas‘ud ZA. Optimization and Comparative Study Different Extraction Methods of Biologically Active Components of Indonesian Propolis Trigona spp (Journal of natural Products, 2012). Vol 5. 233-242.

11. Dragici L and Rapeanu G. Evolution of polyphenols During The Maceration of The Red Grapes (Journal of Agroalimentary Processes and Technologies, 2011).17(2):169-172.

12. Wankhede TB and Manik SR. GC-MS Analysisof the Liverwort Plagiochasma appendiculatum Lehm. et Lindenb (International Journal of Chemical and physical Sciences, 2015). Vol 4: 372-376.

13. El Shoubaky GA and Rahman Salem El. Active Ingredients Fatty Acids as Antibacterial Agent from the brown algae Padina pavonica and Hormophysa triquetra (Journal of Coastal Medicine,2014).

14. Bai S, Seasotiya L, Malik A, Bharti P, Dalal S. GC-MS analysis of Chloroform Extract of Acacia nilotica L leaves (Journal of Pharmacognosy and Phytochemistry, 2014). 2(60:79-82.

15. Venkata RB, Samuel LA, Pardha Saradhi M, Narasimh RB, Naga vamsi krishna A, Sudhakar M, Radha Krishnan TM. Antibacterial, Antioxidant Activity, and GC-MS analysis of Eupatorium odoratum (Asian journal of Pharmaceuticals and clinical research, 2012).Vol 5. Suppl 2: 99-106



Hamidah et al. 88


Surabaya Indonesia

Biosistematic of Bryopsida in Hot Spring Area of R. Soeryo Cangar

Grand Forest Park, East Java

Hamidaha*, Thin Soedartia, and Nathania Ernita Ekawati Edawuaa

a Department of Biology, Airlangga University, Surabaya 60115Indonesia *Corresponding Author’s E-mail: hamidah [email protected]

Abstract. Moss is thalophyta. Study of moss biosystematic is not much, espcesialy in hot spring area of R.Soeryo Cangar Grand Forest Park, East Java. This study aimed to know the diversity of moss (Bryopsida class), and the filogenetic relationship of every genus in Bryopsida member. This was descriptive study and used random sampling method. Moss samples have been acquired and were identified until ―genus‖ using book ―How to

know the mosses and liverworts‖. And then those data were analyzed using SPSS 16.0. Result of this study was three genus, namely Leucobryum, Fissidens, and Hypnum. filogenetic relationship analyzed of class Bryopsida in hot spring area of R. Soeryo Cangar Grand Forest Park showed the genus Leucobryum was more closely related with the genus Fissidens was compared with the genus Hypnum.

INTRODUCTION Indonesia is a country that has the highest biodiversity in the world, after Brazil (Anonymous, 2009). Brazil

is one of the countries with the richest flora of any country in the world, with more than 56,000 species of plants, nearly 19% of world flora and 3,100 species of which comes from Bryophyta (Watson, E., 1981). Indonesia forest area is generally a tropical rain forest. Tropical rain forest is famous for its diversity of flora including species Bryophyta (mosses). Moss diversity research in several regions in Indonesia have been carried out among others in Sulawesi recorded 106 species and 607 species recorded in Borneo. Besides, some of the islands included in the archipelago of Sunda Kelapa had also reported the amount of moss leaves, which in Bali recorded 169 species, 152 kinds of Lombok, Sumbawa 44 species, 278 kinds of Flores, the surrounding been reported by Fleischer 1900-1908 amounted to 452 types (Kukwa & Martin.2012).

Moss is one group of low-growing vegetation and part of biodiversity that has not received much attention. There are 24,000 known species of Bryophyta, and all the moss plants require moist environmental conditions that enter into the life cycle of the plant. Bryophyta division is divided into three classes, namely liverworts (Hepatophyta) with 9000 species and 240 genera; hornwort (Anthocerotopyhta) only 500 species; and moss leaf (bryopsida) has 12000-14500 species and 670 genera (Judd et al, 1999). Bryophyta including one small part of the flora that has not been explored is also a proponent of diversity of flora. Moss plants is widespread and it is an interesting group of plants. They live above ground, rocks, wood, and sometimes in the water. Liverworts and lichens leaves that live alone are usually not attractive. But it can seem attractive even if it grows in groups. In general, plant species less adapted to the living conditions of the land, and most of the plants that live in moist environments and protected. Nonetheless, certain mosses especially true mosses (bryopsida), can survive the dry season. Growth is experiencing rejuvenation if water is available again (Tjitrosoepomo, 2009).

Bryophyta which is a low-level plants and one part of biodiversity that has not been widely studied. Mosses that live alone and no group would seem to appear unattractive, often regarded as dirty environmental causes. However, if carefully considered several types of moss plants look quite attractive, both of color and life bekelompok, such as those found in Forest Park R.Soeryo Cangar. Bryophyta which is a low-level plants and one part of biodiversity that has not been widely studied. Mosses that live alone and no group would seem to appear unattractive, often regarded as dirty environmental causes. However, if carefully considered several types of moss plants look quite attractive, both of color and life bekelompok, such as those found in Forest Park R.Soeryo Cangar.

Forest Park R. Soeryo in Cangar is one of the potential forest area for the habitat of plant diversity moss. R.Soeryo overall topography has a configuration varies between flat, hilly and mountains with an altitude of between 1,000-3,000 meters above sea level. Forest Park R.Soeryo including type C and D with an average rainfall of 2500-4500 mm per tahunmenurut Schmid and Ferguson climate classification. Temperatures range between 5ºC-10ºC (Anonymous, 2007). At Forest Park R. Soeryo also has a pool of hot water with a temperature of 30ºC - 40ºC (Anonymous, 2010 b). Where hot water baths have a carrying capacity of algae growth, which is different from other moss habitat. Moss growth is supported by the habitat moist, while in the hot water bath moss habitat dominated by tumbuhanperintis panas.Jenis temperature affect the most besarsifat-physical, chemical and biological soil (Prawito, 2009). Around the hot springs discovered Cangar many artificial caves the Japanese occupation in 1942-1945 (Anonymous, 2010 a).



Hamidah et al. 89


Surabaya Indonesia

This study aims to determine the diversity and kinship moss (Bryophyta). The results to be obtained from this study are expected to provide some information from knowledge about the diversity of moss (Bryophyta), as found in Forest Park R.Soeryo in Cangar. This research is useful for efforts to conserve natural resources.

MATERIALS AND METHODS Bryophyta placement sampling sites based on the difference in the condition of the four stations. Moss first

sampling done at the site of the first station which is an area that is not hot water flowed, Station II: the irrigation channels around a hot water bath, Station III: around stairs Counters 2. Station IV: on around the road to Goa Japan complex natural attractions Cangar hot spring.

The method in this study by means of moss samples that have been taken and identified in the laboratory using the identification of the book "How to know the mosses and liverworts", the data on the characteristics and scoring number then analyzed in SPSS 16.0 with Classify Hierarchical Cluster with average agglomerative linkage metode then obtained dendrogram.

RESULT Class diversity bryopsida Bryopsida successfully found in Cangar Tourism Thermal Baths Forest Park R.

Soeryo East Java as much as 3 genus Leucobryum, Hypnum, Fissidens.

Character morphological and habitat genus of bryopsida

Character morphological and habitat Leucobryum

This moss is commonly found in rocks and soil that is moist. The genus is rarely found growing in groups, sometimes found along with other leaves moss. This moss muscular and heavy boned. Gametophyte forms such as leaves grow dense and closely packed together. Color light green leaves are glossy, narrow and elongated, sometimes on the tip easily curved, tapered tip of the leaf, the base of which is blunt. Antheridium and archegonium not found. Sporofitnya shape, fused between the rods (Rod) or branches to one another, forming a unified root.

Figure 4.1 (A) morphology genus Leucobryum; (1) part of the gametophyte Leucobryum; (2) part of rhizoid Leucobryum.

The character of morphological and habitat Hypnum

Heavy stature and grow huddle, dark green, light green, dark green that sometimes. This plant has a stem (not the actual stem) creeping or climbing, Archegonium, Antheridium and sporagonium lateral or end branches arranged like a mat. Lanceolate leaf shape, its leaves are oval with a pointed tip and base obtuse. This plant lives in wet or damp ground, and in the area in hot water aliri, Hypnum included in terrestrial plants.

A

1

2



Hamidah et al. 90


Surabaya Indonesia

Figure 4.2 (1) morphology genus Hypnum; (a) capsules of Hypnum; (b) seta; (c) leaf.

The character of morphological and habitat Fissidens

The main character is owned gametophyte generation, centered on the leaves are composed of two rows (distichous) and each having duplicate leaf shaped like a boat on the side adaksialnya, called "vaginant lamina". The genus of this stature such as ferns, shoots straight or curved horizontal. Leaves flat, kosta; edges sometimes delimited. Lamina cells varied, subtle, berpapila, smooth or berpapila Seta; small capsule, short cylindrical, upright or hanging, beaked lid.

Figure 4.3 Morphology Fissidens. Picture: (a) rhizoid Fissidens; (b) the gametophyte.

Table 4.1. Morphological characters of the genus bryopsida

Morphological characters Genus

Leucobryum Hypnum Fissidens

Leaf Colour Light green Light green dark green Tapered Tapered Tapered Tapered Blunt Blunt round Blunt

round The base of

the egg Surface Smooth Smooth Smooth

Edge Score Score Score Reproduction

Sexual Antheridium No Exist No Archegonium No Exist No

Asexual Gemma No No No

a

b

c

1

A

B



Hamidah et al. 91


Surabaya Indonesia

Form gametophyte Leaf Leaf Leaf Sporophyte Exist Exist Exist Habitat

Tree Exist Exist Exist In soil Exist No Exist

Rocks Exist Exist Exist

Table 4.2 Environmental Conditions at four stations in the Thermal Baths Forest Park R. Soeryo Cangar East Java

Abiotic Factor Time

09.00

12.00

15.00

Average

Air temperature (°C) 19 18 18 18,4 Air humidity (%) 96 88 84 89,4

Soil moisture 7 8 3 6 Soil pH 6,5 6,2 5,8 6,17

Table 4.3 Distribution of Bryopsida (in Thermal Baths Forest Park R. Soeryo Cangar East Java, tick mark (√) indicates the

existence of the clan).

No. Name Highways / Genus Location Exploration

Habitat

I II III IV

1. Leucobryum √ Trees 2. Hypnum √ √ rocks 3. Fissidens √ √ Soil, rocks

DISCUSSION Assessment of the similarity of morphological characteristics of Bryopsida genus using dendrogram.

Analysis to determine the grouping of kinship between genus Bryopsida using SPSS 16.0, is based on thirteen kinds of characters. Each genus has similarities and differences in morphological characters. By using a matrix approach can be seen the results of correlation between the value of a single genus with other genus.

Table 4.4. Matrix approach to inter-genus of class Bryopsida.

Object

Correlation Values

1:G1 2:G2 3:G3 4:G4 5:G5

1:G1 1.000 0,296 0,806 -0,030 -0,592

2:G2 0,296 1.000 0,478 -0,243 -0,501

3:G3 0,806 0,478 1.000 -0,245 -0,681

Grouping on the basis of similarity of characteristics possessed by three genus of each class Bryopsida, using cluster analysis hierarachial classify. Morphological characteristics possessed by genus Bryopsida can be seen in Table 4.1 and 4.4 Grouping analysis performed on the morphological characteristics of each genus of each class, transformed in numeric form such as 0, 1, 2, and subsequently compiled in Appendix 1.

Morphological characteristics of the data that has been dinumerisasi and processed with SPSS, obtained coefficient groupings morphological characteristics in common with agglomerative methods (approaches merger) using average linkage cluster in Table 4.5 below.



Hamidah et al. 92


Surabaya Indonesia

Table 4.5 Grouping morphological characteristics based on average linkage.

Description: Figures shown in the column group 1 and group 2 shows the code of OTU were compared, Figures shown in column shows the similarity coefficient fenetik similarities of the two groups were compared and causing OTU 2

OTU compared to the clumped

ANALYSIS GROUPING G1 1 ─┬e─────────────── c G3 3 ─┘d ───────────────────────────────a G2 2 ───────────────── b

Figure 4.6 Dendrogram results genus penggelompokan Bryophyta in Forest Park R.Soeryo Cangar Description: G1=Leucobryum G2=Hypnum G3=Fissidens

Based dendogram in Figure 4.6 above, obtained two groups marked b and c with b value similarity (similarity) 38.2%. The first group (a) consisting of genus Hypnum, while the second group (b) consisting Leucobryum and Fissidens, which are separated by the character morfolgis leaves, shape gametophyte, asexual reproduction, and habitat in the trees

Dendogram in Figure 4.6 separates the kinship between the genus Bryopsida, based on the similarity of the morphological characteristics of each genus through several stages to obtain the approach so as to separate the kinship between the genus or grouping. Such separation causes the formation of groups with a level of similarity between the genus morphological characteristics set out in the scale of similarity (similarity) or the coefficient values listed in Table 4.6

Kinship between Bryopsida genus can be analyzed using morphological characteristics with descriptions of five genus followed fenetik analysis presented in the form dendogram. With the dendogram, the genus that has more resemblance and similarity great value. Penggelompokkan dendogram results indicate the formation of inter-genus. Besides genus Hypnum have a closeness value of 38.7% to the group d is Leucobryum and Fissidens. While the genus Leucobryum had an alliance with Fissidens worth 80.6% have the same character, among others, leaf color, leaf surface, leaf edge of the reproductive organs, the shape of gametophyte, sporophyte, and their habitats are found on trees and rocks.

REFERNCES 1. Anonimus. 2009. Manfaat Tumbuhan Lumut Sebagai Obat Alami. Artikel dari situs

http://www.google.com// (11 september 2010) 2. Anonimus. 2010 a. Obyek Wisata Alam. Artikel dari situs http://Guajepangcangar.go.id// (10 Oktober

2010) 3. Anonimus. 2010 b. Obyek Wisata Alam. Artikel dari situs http://Pemandiancangar.go.id//. (10 Oktober

2010) 4. Conard, H. S., 1979. How To Knows The Mosses and Liverworts. McGraw-Hill, United States of

America. Halaman 28, 230, 239 5. Judd,Campbell, Kellogg,and Stevens. 1999. Plant Systematic. Massachusetts, USA, Sinauer

Associates, Inc.Publishers.

Phase

The combination of group Similarity

coefficient Group 1 Group 2

1 1 3 0,806

2 4 5 0,720

3 1 2 0,387

http://www.google.com/

http://guajepangcangar.go.id/

http://pemandiancangar.go.id/



Hamidah et al. 93


Surabaya Indonesia

6. Kukwa & Martin.2012. Thirty Six Species Of The Lichen Genus Parmotrema (Lecanorales, Ascomycota) New to Bolivia. Polish Botanical Journal 57(1): 243–257. Bolivia

7. Tjitrosoepomo, Gembong. 2009. Taksonomi Tumbuhan. Yogyakarta : UGM Press 8. Watson, E. Vernon. 1981. British Mosses and Liverworts. Melbourn : Cambridge University Press

Australia

APPENDEX 1 Quantification of morphological characters Bryophyta

No characters Genus descreption Leucobryum Hypnum Fissidens

1 Color leaves 1 1 2 0= no leaf; 1= green 2= green dark

2 The end leaves 1 1 2 0= no leaf; 1= exist 2= green dark

3 The base leaves 1 1 2 0= no leaf; ;1= blunt; 2=ovate

4 Surface leaves 1 1 1 0= no leaf no leaf; 1= smooth 5 edge 1 1 1 0= no leaf; 1= average 6 anteredium 0 1 0 0= no leaf ;1= exist 7 archegonium 0 1 0 0= no leaf; 1= exist 8 Gemma 0 0 0 0= no leaf 1= exist 9 Form

gametophyte 1 1 1 0= thallus; 1= leaf

10 sporophytes 1 1 1 0= no; 1= exist



Tamyiz et al. 94


Surabaya Indonesia

DIVERSITY OF THE SCHMUTZDECKE LAYERS IN SLOW

SAND FILTER (SSF) AND EFFECT ON EFFICIENCY

REMOVAL OF POLLUTANTS

Muchammad Tamyiz1), Laily Noer Hamidah2,*

1,2Department of Environmental Engineering Nahdlatul Ulama Sidoarjo University, Sidoarjo 61218 Indonesia

*Corresponding author’s Email: [email protected]

Abstract.Exclusion of pollutants in water treatment using slow sand filter (SSF) mostly occur in the schmutzdecke layer. Biofilms are formed on the schmutzdecke layer have a diversity of microorganisms, so the process of elimination of pollutants occur as a result of biological activity. Visualization from diversity of microorganisms in the schmutzdecke layer that aged 30 day was performed using SEM (scanning electron microscopy), indicating the presence of various types of diatoms, bacteria shaped cocci and bacilli, as well as the extracellular matrix. To determine the effectiveness of SSF conducted water quality analysis in the form of COD, total N, and total P by using water from the outlet unit pre-sedimentation IPAM, Ngagel, Surabaya that have characteristics of turbidity between 16.5 to 174 NTU. Based on the research results obtained in laboratory scale the average efficiency of COD, total N, and total P respectively 47.55%, 39.80%, and 49%. Further, analysis showed that the efficiency of pollutant removal increases with enhancement the number of bacteria in the schmutzdecke layer.

INTRODUCTION Slow sand filters (SSF) is used in drinking water treatment effectively since 200 years ago (Weber and Dick,

1999; and Langenbach et al., 2010), because is able to remove carbon compounds, pathogenic bacteria, protozoan parasites and suspended solids (Schuler et al., 1991; Aslan and Cakici, 2007; and Elliott, et al., 2008).

The removal of contaminants in the raw water by using SSF, mostly occurs in the schmutzdecke layers (Joubert and Pillay, 2008). Schmutzdecke is a biofilm layer that forms on top of the sand media (Law, et al., 2001), it has a complex architecture, rich in mucopolysaccharides produced by microorganisms in situ and algaes that have died (Wotton and Hirabayashi, 1999). This layer is composed of biological material that is saturated making it suitable as a place of growth and development of bacteria in the sand with thick several centimeters of surface (Huisman and Wood, 1974). Based on previous studies, schmutzdecke layers consists of alluvial silt, organic waste, bacteria, algae, diatoms, zooplankton and biologically active compounds (Huisman and Wood, 1974; Samantha, et al., 2001; Joubert and Pillay, 2008). Types of microorganisms on schmutzdecke layers are one of the important parameters in supporting the performance of SSF (Eighmy, et al., 1992). This is due to biological activity that occurs in schmutzdecke through process of bioadsorption and biodegradation of organic compounds. (Ellis and Aydin, 1995; Hendel, et al., 2001; and Ho, et al., 2006).

Schmutzdecke layer composition can be shown through visualization using scanning electron microscopy (SEM). Visualization by SEM was used to see the development of a diversity of microorganisms during the process of forming a schmutzdecke layers of early stage (seeding) until become mature schmutzdecke layers (Samantha, et al., 2001; Joubert and Pillay, 2008). While for determine the number of microorganisms on schmutzdecke layers such as the number of bacteria can be calculated using the Total Plate Count (TPC).

Research on schmutzdecke layers are still very limited for countries with two seasons such as Indonesia. Therefore, the aim of this study was to get a diversity of microorganisms in schmutzdecke layer and examine the relationship between the number of bacteria in schmutzdecke layer towards removal of pollutants such as chemical oxygen demand (COD), total N and total P.

METHOD

Running SSF

Running SSF process carried out for 30 days with 14 days of acclimatization. The purpose of acclimatization is to grow the microorganisms are stable and can adapt to the raw water (Indriyati, 2003). SSF units that used have dimensions of length, width, and depth of each 60 cm (Figure 1).



Tamyiz et al. 95


Surabaya Indonesia

Figure1. Slow Sand Filter (SSF) Unit

Sand that used as a media has a diameter of 0.15 to 0.35 mm and the flow velocity (flow rate) of 0.2

m3/m2.hours. The raw water used comes from the outlet unit pre-sedimentation of PDAM Ngagel I. In this experiment, pretreatment with a Roughing Filter (RF), due to research carried out during the rainy season so that the river water has a fairly high turbidity, reaching 259 NTU. RF types of units used is Vertical Roughing Filter (VRF) 4 series with stream downflow, has a length and width dimensions of each 30 cm and depth of 50 cm of media.

SEM Analysis

Sampling of Schmutzdecke performed on the 30th day surgery unit on SSF. Schmutzdecke samples taken after the water in the reactor is emptied. Sampling process conducted with scrapper aseptically (Lazarova and Manem, 1995). Further samples of aerated for the drying process, this is because about 97% of the composition of the biofilm is composed of water (Samantha, 2001; Joubert and Pillay, 2008), so that the drying process when the SEM sample preparation needs to be considered in order not to damage the desired visualization. Further, sample is coating process using gold (Au) to change the conductivity of a sample to be observed. Because the basic principle of the SEM is an image created based on the detection of new electrons (secondary electrons) or reflected electrons emerging from the sample surface when the sample surface is scanned with an electron beam. Therefore, the sample must be conditioned conductive.

Analysis of the number of bacteria

Sampling of schmutzdecke carried out every 2 days for 15 days after the acclimatization process is complete. Samples were taken from the surface of the sand (1 cm from the surface). Analysis of the number of bacteria carried out using total plate count (TPC). The process inoculation of sample on NA media (Nutrient Agar) carried out by Pour Plate Technique. This technique give benefit easily observed, a separate colony, and there is no competition between the bacteria in decision O2 because its location scattered (Anyleite, 2013).

Analysis of water quality of COD, N, and P

The next parameter being analyzed is the value of COD, total N and total P. Sampling was carried out every 2 days for 15 days. Samples were taken from the inlet and outlet units of SSF to see efficiency of removal from the three parameters.


Result of SEM Analysis

SEM analysis results of the samples schmutzdecke that aged 30 days showed the presence of extracellular matrix and the microorganisms that grow on the surface of the sand on the SSF (Figure 2).



Tamyiz et al. 96


Surabaya Indonesia

Figure 2. The result of SEM analysis in sand (single media), (a) the species of diatom (Navicula sp.) -Mag: 5000x, (b) the fossil diatom - Mag: 3000X, (c) the form of Cocci and Bassil bacteria and - Mag: 5000x, (d) cocci bacteria is enveloped

extracellular matrix - Mag: 5000x In Figure 2, diatom Navicula sp. has been found species that attached to sand grains. The Navicula sp. is a

type of diatoms, which are crucial for the survival of the natural ecology. Nearly a quarter of O2 in nature produced by this species, so this type diatoms serve as a key species in the food chain for aquatic organisms. In addition, the presence of cocci bacteria has been found with a diameter of 3 to 3.7 μm and Bassil bacteria

with a length of ± 4.5 μm. Cocci bacteria seen more dominating compared Bassil bacteria. It has also been

proved by Joubertand Pillay (2008), using SEM visualization for schmutzdecke that aged 2 weeks dominated by Bassil bacteria with various lengths from 1.5 to 4.1 μm. Based on previous research, the results of staining gram

of 13 isolates of bacteria found in the samples schmutzdecke aged 15 days, found 11 isolates of Bassil bacteria (Hamidah, et al., 2013), after that, the Bassil bacteria is in line with the development of schmutzdecke replaced by coccus bacteria such as in Figure 2.

According to Joubert and Pillay (2008), until four week, schmutzdecke dominated by bacteria with a variety of sizes, the next step on week 5 and 6 are replaced by diatoms with different varieties, and at eight week schmutzdecke already apparent, only a layer of biofilm increasingly thickened covering grain of sand. The diversity of microorganisms in schmutzdecke operations continue to increase over time, due to be carried away by the raw water that is trapped and embedded in the biofilm. Increasing of schmutzdecke age, the extracellular matrix also thickened, besides there is accumulation of remnants of metabolic products of cells and cell death causes schmutzdecke visualization is not clearly. The extracellular matrix is the result of excretion from bacterial cells that are concentrated are used to trap nutrients for the growing population of microorganisms and helps prevent the escape of cells from the surface (Anonymous, 2011)

Result of TPC Analysis

The number of bacteria in the schmutzdecke sample is calculated as the value of TPC (total plate count) as shown in Table 1 below.

Table1. Results of the analysis of the schmutzdecke bacteria number in the SSF units

Days Number of bacteria per mL of sample (CFU/mL)

1 6,0 x 108

3 5,6 x 108

5 7,1 x 108

7 7,7 x 108

9 23,0 x 108



Tamyiz et al. 97


Surabaya Indonesia

13 17,0 x 108

15 5,5 x 108 Average 10,0 x 108

Table 1 shows the average number of bacteria in the SSF units of 1.0 x 109 CFU/mL. Schmutzdecke tends to

increase the number of bacteria in the filter operates. This type of bacteria is calculated as the value of TPC is aerobic bacteria. According to Pell and Nyberg (1989) in Bahgat et al, (1999), the population of aerobic bacteria thrives better on the upper surface of the sand. O2 is used by aerobic bacteria as an electron acceptor in the process of cell respiration to oxidize substrates and to produce energy, then, this energy will be used for cell growth (Anonymous, 2012).

Organic matter carried by the raw water is oxidized by bacteria to produce energy for metabolic processes (dissimilation), and other organic matter is converted into new cells for growth (assimilation) (Huisman and Wood, 1974). So the number of bacteria also increased along with the development schmutzdecke layer, until a point where the numbers are decreasing due to the death of the cell where the new organism replaces the old organism.

Relationship between Bacteria Number with Pollutants Removal

The pollutants such as COD, nitrogen (N) and phosphorus (P) carried by the raw water is an essential nutrient for microorganisms. Efficiency of COD, N, and P in the SSF units can be seen in Table 2.

Table 2. Efficiency of COD, N, and P in the SSF units

Days Efficiency of pollutants removal (%)

COD N total P total

1 60,53 41,33 44,08 3 90,91 38,46 67,74 5 66,67 32,77 48,03 7 6,25 47,62 57,51 9 75,00 39,39 85,75

13 11,76 60,00 30,84 15 21,74 12,90 47,51

Average 47,55 39,80 49,00 SSF units have efficiency of removal of COD with average of 47.55%, the pollutants can removed by

biological and chemical processes. Biological processes occur because schmutzdecke layer that can remove organic materials, transforming organic compounds and kill pathogens, and produce safe drink water from contaminant of microbiology (Campos et al,. 2002). Chemical processes occur through adsorption events. Adsorption is a reduction in smaller particles and suspended particles such as colloidal particles and dissolved particles (Huisman and Wood, 1974), this adsorption process occurs as a result of differences in surface charge of media with suspended particles and colloids around (Huisman and Wood, 1974).

According to Huisman and Wood, (1974), at normal pH, filter media has a negative charge, while the inorganic material has a positive charge. Inorganic materials in the raw water adsorbed on the filter media. As for the organic material has a negative charge. Therefore, at the beginning of the operation has not been a reduction in organic matter through the process of adsorption. However, after the filtration process runs and a lot of positively charged particles are trapped in the surface of the media, the organic material that negatively charged will adsorbed too.

In addition, the secondary adsorption processes also occur on the surface of the sand, bacteria that have drawn negative charge and adsorbed by the grains of sand that are positively charged. Therefore, bacteria and viruses carried by raw water contact with the surface of the sand grains. The presence of electrostatic forces that occur between the media of sand and bacteria (including viruses) there is a process of attraction so that the coliform bacteria in the water can be removal (Huisman and Wood, 1974).

Furthermore, the removal of N and P involved in the process of nitrification and denitrification. Nitrification occurs at depths below 30-40 cm (Huisman, 1974). Nitrification and denitrification on the SSF units occur due to bacterial nitrification and denitrification. According Nakhla and Farooq (2003), the number of denitrifying bacteria found to be more than the number of bacterial nitrification. At the bottom of the filter, increasing nitrification activity is proportional to the length of the filter operation. This is in contrast to the nitrification activity on the surface of the filter, where the slower nitrification process. Therefore, the operating time of the filter affects the amount of nitrification bacteria and activities. So, it can be said that the estimated removal of N does not only take place on the surface of the sand, but also in the depth of a particular sand. Similarly, the



Tamyiz et al. 98


Surabaya Indonesia

removal of the P element not only occur under aerobic conditions, but also can occur in anoxic-anaerobic conditions. In this condition, nitrate resulting from nitrification is used as the electron acceptor replace O2 in the removal of the element P. organism that serves as the accumulation of phosphorus (PAOs/phosphorus accumulating organisms), such as Acinetobacter, E. coli. Therefore, the removal of phosphorus is not only happening on the surface of the filter bed, but also at a certain depth (Ahn, et al., 2003). Furthermore, the removal of the three elements of pollutants (COD, N, P) if the number of bacteria associated with schmutzdecke can be seen in Figure 2.

Figure 3. The relationship between the number of bacteria with removal efficiency of pollutants (COD, N, and P)

Figure 3 shows the relationship between the number of bacteria with efficiency of COD, N, P in the SSF

with a flow rate of 0.2 m3/m2.hour. Based on the figure 3 can be seen that the graph pattern efficiency of removal of COD, N, P comparable to the number of bacteria. When removal efficiency increases, the amount of bacteria on schmutzdecke also increased. This is due to COD is used by bacteria as a carbon source for growth in generating new cells and energy for cellular processes, so the number of bacteria in schmutzdecke will increase. In addition, nitrogen (N) and phosphorus (P) is important element for determining cell growth activity of microorganisms.

Based on the graph N and P, indicate some point where the elimination of N and P are not comparable to the number of bacteria, this is because the ratio of COD: N: P fluctuated. According to Wulan, et al., (2012) in order to achieve optimal growth of bacteria, the three nutrients (COD, N, P) must be in the proper ratio. C:N ratio is low (the content of N element is relatively high), will increase tthe emission of nitrogen as ammonium which may hinder the proliferation of bacteria. While the C: N ratio is high (the content of N element is relatively low) will cause the degradation process is slower because nitrogen will be a limiting factor (growth-rate limiting factor) (Alexander, 1994, in Wulan, et al., 2012). According to some literature suggests that the ratio of C: N: P optimum is 100: 10: 1.

CONCLUSION Based on SEM visualization, a schmutzdecke layer that aged 30 days found several species such as diatoms,

Bassil and Cocci bacteria, extracellular matrix, and dominated by cocci bacteria. The number of bacteria in the schmutzdecke layer also increased along with the filter operation. Increasing the number of bacteria was followed by an increase in pollutant removal efficiency of COD, N, P carried by raw water. This is because COD, N, P is used by the bacteria in the metabolism process for the formation of energy and cell growth.

REFERENCES 1. Ahn, K. H., Song, K. G., Cho, E., Cho, J., Yun, H., Lee, S., & Kim, S.,Enhanced Biological

Phosphorus and Nitrogen Removal Using A Sequencing Anoxic/Anaerobic Membrane Bioreactor (SAM) Process, Journal of Desalination, 157, pp. 345-352, 2003.

2. Anonymous,Biofilm,http://id.wikipedia.org/wiki/Biofilm,2011, (30Maret 2012).

0 2 4 6 8 10 12 14 160

5

10

15

20

25

30

Amount of Bactery COD N P

Day's

Am

ount

of B

acte

ry (1

0^8

CFU

/ML)

0

10

20

30

40

50

60

70

80

90

100

Rem

oval Efficiency (%

)

http://id.wikipedia.org/wiki/Biofilm



Tamyiz et al. 99


Surabaya Indonesia

3. Anonymous,Organisme Aerobik. http://id.wikipedia.org/wiki/Organisme aerobik, 2012,(30Maret 2013).

4. Aslan, S. & Cakici, H, Biological Denitrification of Drinking Water in A Slow Sand Filter. Journal of Hazardous Materials, 148, pp. 253–258, 2007.

5. Bahgat, M., Dewedar, A., & Zayed, A.,Sand-Filters Used for Wastewater Treatment: Buildup and Distribution of Microorganisms,Journal of Water Research, 33(8), pp. 1949-1955, 1999.

6. Campos, L. C., Su, M.F.J., Graham, N.J.D., & Smith, S.R., Biomass Development in Slow Sand Filter, Journal of Water Research, 36, pp. 4543-4551, 2002.

7. Eighmy, T.T., Collins, M.R., Spanos, K.,&Fenstermachert, J.Microbial Populations, Activities and Carbon Metabolism in Slow Sand Filters. Journal of Water Research, 26(10), pp. 1319-1328, 1992.

8. Elliott, M.A., Stauber, C.A., Koksal, F., DiGiano, F.A., & Sobsey, M.D. Reductions of E. coli, Echovirus Type 12 and Bacteriophages in an Intermittently Operated Household-Scale Slow Sand Filter,Journal of Water Research, 42, pp. 2662-2670, 2008.

9. Ellis, K.V. & Aydin, M.E.,Penetration Of Solids and Biological Activity Into Slow Sand Filters,Journal of Water Research, 29(5), pp. 1333-1341, 1995.

10. Hamidah, L.N., Fitriani, N., & Trihadiningrum, Y., Pengaruh Media Tumbuh terhadap Komunitas Bakteri schmutzdecke pada slow sand filter untuk pengolahan air minum, Prosiding Seminar Nasional Pascasarjana XI. ITS, Surabaya, 2013.

11. Hendel, B., Marxsen, J., Fiebig, D., & Preu, G.,Extracellular Enzyme Activities During Slow Sand Filtration in A Water Recharge Plant, Journal of Water Research, 35(10), pp. 2484–2488, 2001.

12. Ho, L., Meyn, T., Keegan, A., Hoefel, D., Brookes, J., Saint, C.P., & Newcombe, G.,Bacterial Degradation of Microcystin Toxins within A Biologically Active Sand Filter, Journal of Water Research, 40, pp. 768-774, 2006.

13. Huisman, L. & Wood, W. E.,Slow Sand FiltrationHandbook, World Health Organiation, Geneva, Swistzerland, 1974.

14. Indriyati, Proses Pembenihan (Seeding) and Aklimatisasi Pada Reaktor Tipe Fixed Bed. Jurnal Teknik Lingkungan,P3TL-BPPT, 2, pp.54-60, 2003.

15. Joubert, E.D. & Pillay, B., Visualisation of the Microbial Colonisation of A Slow Sand Filter Using an Environmental Scanning Electron Microscope. Journal of Biotechnology, 11(2), pp. 1-7, 2008.

16. Joubert, E.D. & Pillay, B., Visualisation of the Microbial Colonisation of A Slow Sand Filter Using an Environmental Scanning Electron Microscope,Journal of Biotechnology, 11(2), pp. 1-7, 2008.

17. Langenbach, K., Kuschk, P., Horn, H., & Kastner, M., Modeling of Slow Sand Filtration for Disinfection of Secondary Clarifier Effluent, Journal of Water Research, 44, pp. 159-166, 2010.

18. Law, S.P., Melvin, M.M.A., & Lamb, A.J., Visualisation of The Establishment of A Heterotrophic Biofilm Within the Schmutzdecke of A Slow Sand Filter Using Scanning Electron Microscopy, Journal of Biofilm, 6(1),pp. 1-17, 2001.

19. Lazarova, V., & Manem, J.,Biofilm Characterization and Activity Analysis in Water and Wastewater Treatment,Journal of Water Research, 29(10), pp. 2227-2245, 1995.

20. Nakhla, G. & Farooq, S. Simultaneous Nitrification–Denitrification in Slow Sand Filters, Journal of Hazardous Materials, B96, pp. 291–303, 2003.

21. Schuler, P.F., Ghosh, M.M., & Cropalan, P., Slow Sand and Diatomaceous Earth Filtration of Cysts and Other Particulates. Journal of Water Research, 25(8), pp. 995-1005, 1991.

22. Weber, S.M.L.,& Dick, R.I.,Bacterivory by A Chrysophyte in Slow Sand Filters, Journal Of Water Research, 33(3), pp. 631-638, 1999.

23. Wotton, R.S. & Hirabayashi, K., Midge Larvae (Diptera: Chironomidae) as Engineers in Slow Sand Filter Beds, Journal of Water Research, 33(6), pp.1509-1515, 1999.

24. Wulan, P., Gozan, M., Arby, B., &Achmad, B., PenentuanRasio Optimum C:N:PSebagaiNutrisipada Proses BiodegradasiBenzena-Toluena and Scale Up KolomBioregenerator. TugasakhirTeknik Kimia, FakultasTeknik, Universitas Indonesia, 2012.

http://id.wikipedia.org/wiki/Organisme%20aerobik



Putro et al. 100


Surabaya Indonesia

FORTIFICATION CASHEW NUT LEAF (Anacardium occidentale

L.) USING GLYSINE ON THE GROWTH and DEVELOPMENT

OF WILD SILKWORM (Cricula trifenestrata Helf.)

Sulistyo Dwi Kartining Putro*, Jekti Prihatin, and Suratno

Jurusan Pendidikan MIPA, Fakultas Keguruan dan Ilmu Pendidikan, Universitas Jember Jln. Kalimantan 37, Jember 68121

*Corresponding Author : E-mail: [email protected]

Abstract : Wild silkworm is a non-mulberry silkworms that produce different types of silk are very varied kind. Cricula trifenestrata Helf is one wild silkworms that produce golden silkworm. Requires a specific feed silkworms that determine population growth and cocoon production. Nutritional needs silkworms can change at any time, depending on the growth, reproduction, diapause, or displacement. Glycine as a non-essential amino acids required in the process of metabolism larval C. trifenestrata Helf. serves to structural requirements, as enzymes, receptors, for transport and storage needs. The process of adding glycine larvae feed on the leaves of C. trifenestrata Helf silkworms. using the techniques of fortification. The results of the research and analysis of the value showed that, there is the effect of glycine on the weight parameter instar larvae 4, length 4 instar larvae, cocoon weight, cocoon shell weight, and long days to grow. As for the weight parameters and long larval instar larval instar 5 5, the results obtained have not demonstrated the influence of glycine on Cricula trifenestrata helf.

INTRODUCTION Wild silkworm is a non-mulberry silkworms that produce different types of silk very varied kind. Cricula

trifenestrata helf is one wild silkworm which produces golden yellow silk [1]. But its presence in society known as pests. So far C. trifenestrata helf. Used by collecting cocoon of nature by a small part of society. One important means of cultivating silkworms are groceries (food) [2]. Silkworms require specific feed that determine population growth and cocoon produced.

Silkworm nutritional needs can change at any time, depending on the growth, reproduction, diapause, or displacement. Glycine as one of the non-essential amino acid needed in the process of metabolism of larvae of C. trifenestrata helf. serves to its structural needs, as enzymes, receptors, for transport and storage needs [3]. The process of adding glycine larvae feed on the leaves silkworms C. trifenestrata helf. fortification technique. Where fortification technique is the addition of the substance in the food which does not change the structure and form of food, and also not add weight to the food. In this study, the efforts are underway to augment the needs of amino acids in silkworm can be done fortification engineering group adding protein amino acid glycine in the feed with the aim of improving the enzyme in the gene for the enzyme fibroin thus adding time working on jouvenile hormone and can promote the development and optimal growth of C. Trifenestrata helf.

EXPERIMENTAL This study is a research laboratory using a completely randomized design (CRD) with 4 treatments and 3

repetitions in each treatment. This research was conducted at the Laboratory of Biology Education Building III FKIP Jember University for three months, ie January to March 2014. The sample used was a wild silkworm larvae Criculla trifenestrata helf. obtained from cocoons superior drawn from the host tree Cashew (Anacardium occidentale L.). Glycine treatment using a fortification with four treatments. P0 (control treatment with feeding fortified with distilled water), P1 (treatment by feeding fortified with glycine concentration of 250 ppm), P2 (treatment by feeding fortified with glycine concentration of 500 ppm), and P3 (Treatment with feeding fortified with glycine concentration of 750 ppm). To determine the influence of fortification of glycine to the growth and development of the wild silkworm Cricula trifenestrata helf. ANOVA test with 95% significance level (p <0.05). If the result is significant then tested lanju namely Duncan test with significance level of 95% by using SPSS version 17 to determine differences between treatments.

RESULT AND DISCUSSIONS In this research, measurement of multiple parameters of growth and development Cricula trifenestrata helf.

For the calculation of the growth parameters of weight and length during the larval stage, the resulting cocoon



Putro et al. 101


Surabaya Indonesia

weight, and long wingspan when imago. Glycine fortification result data on the growth of Cricula trifenestrata helf. are presented in Table 1.

Tabel 1. Glycine fortification result of the growth Cricula trifenestrata helf.

Variable Treatment

Control 250 ppm 500 ppm 750 ppm weight of Instar 4 (g) 1,10 1,05 1,08 1,11 weight of instar 5 (g) 1,09 1,19 1,28 1,32 long of instar 4 (cm) 5,17 3,92 4,30 5,67 Long of instar 5 (cm) 7,11 4,83 5,20 6,87 Weight of kokon (g) 0,10 0,08 0,09 0,10 Cocon shell weight (g) 0,06 0,05 0,06 0,06 Wingspan (cm) 6,14 6,35 6,39 6,30

Explanation: larvae were maintained for each treatment = 30 individuals Whereas observations show the development parameters on how long it takes from larva, pupa, until the

imago. Based on the research results, there is the effect of glycine on the weight parameter fourth instar larvae, long larval instar 4, cocoon weight, cocoon shell weight, and long days to grow. The weight parameter instar larval instar larvae length 5 and 5, the results obtained have not shown any effect of glycine on Cricula trifenestrata helf. Data from Glycine fortification on the development Cricula trifenestrata helf. presented in Table 2.

Tabel 2. Glycine fortification results on the development of Cricula trifenestrata helf.

Variable Treatment Control 250 ppm 500 ppm 750 ppm

Old instar 4 (day) 5,07 4,12 4,10 4,13 Old instar 5 (day) 5,06 4,11 4,10 4,05 Oldes pupa (day) 6,56 6,44 6,44 6,38 Oldes imago (day) 6,81 5,22 5,19 5,23 Mortalitas (%) 46,67 46,67 46,67 53,33

Explanation: larvae were maintained for each treatment = 30 individuals Cricula wild silkworm trifenestrata helf. young people will grow up healthy if eaten leaves contain lots of

water, soft and brownish green. While the adult instar silkworms require older leaves and not too young [4]. In the larval stage, silkworms Cricula trifenestrata helf. has a unique metabolic system to produce a large number of cocoon proteins, by utilizing nitrogen from food sources. Silkworm larvae Cricula trifenestrata helf. will digest and absorb about two-thirds of the nitrogen content and leaf protein in the feed is consumed [5]. Besides insects feed should be available, acceptable, can be ingested, can be assimilated, and it contains all the nutritional requirements should also contain allelochemichals functioning influence the behavior of insect feeding either stimulants or attractants [6].

The presence of allelochemichals will attract insects to consume leaves of good quality feed. Leaves are good quality feed that contains nutrients that are good, if good nutrient content will affect both the larval development [7]. Based on the results of this research is that in general fortification is done on the leaves feed jambe cashew (Anacardium occidentale) has a good influence on the growth and development of the Cricula trifenestrata helf. among others: severe long-instar 4th instar 4, cocoon weight, cocoon shell weight, old growth instar 4, 5 instar growth long, long imago, mortality, and the number of eggs produced. Several other parameters are less influential on the fortification glycine performed on feed jambe leaves cashew (Anacardium occidentale), among others: heavy larval instar 5, 5 instar larvae length, wingspan, long pupa, and imago number of good quality.

Heavy growth of larvae and larval length influenced by nutrients derived from leaf feeding. The leaves feed rich in nutrients needed by the larvae will affect accretion formation and cell number rapidly increased in larvae Cricula trifenestrata helf. Glycine content of fortified feed on leaves of cashew leaves, resulting in additional compounds that cause appetite allelochemichals the larvae climb. If the protein amino acid glycine metabolic needs are exceeded, the amino acids would be intermediate amfibolik catabolized to be used as an energy source as Subtract the biosynthesis of carbohydrates and lipids [8].

High content of amino acids able to form carbohydrates and lipids, wherein the compound is able to increase the weight and length growth significantly larvae can grow rapidly. Insects have outer frame which does not allow growth to enlarge the body (body size) is quite fathomable overcome by the process of molting (moulting). Insects pradewasa just out of the egg develops through a series of skin turnover, and increased in size after each molt. Cricula development trifenestrata helf. characterized by changes in body shape and size of



Putro et al. 102


Surabaya Indonesia

larvae were even greater, by way of ecdysis. Ecdysis process occurs due to the increase in the size and shape of larvae [9].

Data obtained from several growth parameters Cricula trifenestrata helf. with fortification glycine, showed that the higher the concentration of the treatment given, the higher the results obtained. The greater the concentration of glycine is given, the less time is used by Cricula trifenestrata helf to grow and develop. However, the old parameters pupa growth, the concentration of glycine treatment did not affect the long development of the pupa Cricula trifenestrata helf. this is because, during the pupal larvae undergo a resting period (diapause). In this process pupa becomes very passive, and do not eat, if the environmental conditions are adequate and suitable to be turned into a pupa imago, but if inadequate environmental conditions will remain a pupa. So that the long process of the development process of the pupa is also influenced by some external factors that can affect it.

CONCLUSION Fortification glycine in the feed leaves Criculla trifenestrata helf effect on growth and development with the

highest yield at the concentration of 750 ppm, for average larval weight (10.1053 g), length instar (5.6 cm), cocoon weight (0.0968 g), cocoon shell weight (0.0638 g), old growth instar (4 days), long imago (5 days), and the number of eggs produced (1.4 point), while the highest mortality in the control treatment by (5.42% ), and the wingspan of the highest results in the treatment of 500 ppm (5.13 cm). Glycine concentration which can affect the growth and development of C. trifenestrata helf. the maximum is 750 ppm treatment. Based on the analysis for the overall parameters of the observed treatment using glycine concentration of 750 ppm and petumbuhan development occurs sooner, and the results obtained cocoon is increasing, and lower mortality.

REFERENCES 1. Situmorang, J. 1996. ―An Attempt to Produce Attacus atlas (Lepidoptera: Saturnidae) Using

Baringtonia Leaves as Plant Fodder‖. Dalam Int.J. Wild Silkmoth and Silk 2. Japan: The Japan Society

for Wild Silkmoth, 55-57. 2. Guntoro, S. 1994. Budidaya Ulat Sutera. Kanisius. Yogyakarta: Medika Pintar 3. Widhayasa, B. 2012. Ilmu Hama Tanaman : Pengaruh Protein bagi pertumbuhan dan Perkembangan

Serangga. Malang: Fakultas pertanian Universitas Brawijaya 4. Jolly, M.S., Sik Sen, T.N. Solwaker and G. K. Prasad 1979. Non Muberry Silk. Rome: Food and

Agriculture Organization of The United Nation Rome 5. Sukardina, D. 2003. Pengaruh Pemeliharaan secara Gregarius dengan Pakan Alami Daun Jambu Mete

terhadap Perkembangan Ulat Sutera Emas Criculla trifenestrata Helf. SKRIPSI tidak diterbitkan. Jember: Universitas Jember Press

6. Wuliandari, J. R. Dan Situmorang, J. 2005. Pengaruh Pakan dan Tempat Pemeliharaan yang Berbeda terhadap Masa Perkembangan Larva Attacus attlas (L.) (Lepidoptera: Saturniidae). Universitas Gajah Mada Press: Yogyakarta.

7. Karimah, R. 2003. Pengaruh Perbedaan Jenis Daun Pakan terhadap Perkembangan Ulat Sutera Liar Cricula trifenestrata Helf. (Lepidoptera: Saturniidae). SKRIPSI tidak diterbitkan. Jember: Universitas Jember Press

8. Murray R.K., Darly, K. G. M.D, Peter, A., Victor W. R. 2003. BIOKIMIA (Harper) (Edisi Terjemahan: Dr. Andy Hartanto) Edisi 25. Jakarta: Penerbit Kedokteran EGC

9. Slanksky., F., Jr. And J.M. Scriber. 1985. Food Consumption and Utilization. In G.A. Kerkut and L.I. Gilbert (eds.). Comprehensive Insect Physiology Biochemistry & Pharamocology Vol.4 Regulation: Digestion Nutrition Exretion. Oxford: Pergamon Press

10. Prihatin, Jekti. 2010. Sains Polusi, Dampak Hujan Asam terhadap Budidaya Ulat Sutera. Malang : UMM Press.



103


Surabaya Indonesia

ORAL PAPER OF BIOMEDICAL ENGINEERING (OBME)



Karimasari et al. 104


Surabaya Indonesia

‘LETSLEEP Mask’ (LED and Binaural Beats Sleeping

Mask), Sleeping Mask as Sleep-Wake Cycle

Regulations Based on Brain Wave Stimulation

Azisya Amalia Karimasari1*, Amila Sofiah1, Priyanka Kusuma Wardhani1, Andi Achmad

Dzulfiqar1, Choirul Chabib2, Novia Dwi Asmaningtias1 Prihartini Widiyanti1

1 Biomedical Engineering Study Program, Faculty of Science and Technology, Universitas Airlangga, Surabaya, Indonesia

2 Otomatization Information System, Faculty of Science and Technology, Universitas Airlangga, Surabaya, Indonesia

*Coresponding author: email [email protected]

Abstract. Depression has reached 17-27 % in Indonesia. About 5-10% people in the world have sleep dis.order. Furthermore, one symptom of depression is sleep disorder. Sleep disorder is related to disruption of neurotransmitter serotonin and the hormone regulation. In this study, brain wave stimulation is able to manage sleep quality and duration. LETSLEEP Mask is a sleeping Mask that consist to two parts, they are sleeping Mask and the controller. Sleeping Mask has two kind of LED (light emitting diode), blue and yellow; and earphone with delta and gamma wave of binaural beats. The material of sleeping Mask is PLA (polylactic acid) in 3D printing. Second part, Arduino Uno is a microcontroller to control the lights of LED and sound of binaural beats. Arduino Uno also controls the mechanism of alarm for sleep-wake cycle. Device works to the first 30 minutes for blue LED and delta binaural beats and the last 30 minutes for yellow LED and gamma binaural beats. User can set the time that will make he/she wake up for the alarm time. Blue LED and delta wave of binaural beats are able to stimulate the delta brainwave. This stimulation can trigger a deep sleep. While, yellow LED and gamma binaural beats are capable of stimulating the gamma brain waves and trigger awakening. The research methode is causal quantitative with experimental primary data collection. There is two test in this research, sleep and awakening test. A delta wave stimulation is shown in lead C3-C4 during the sleep test. The delta wave in stimulation condition is 23,42 times than no stimulation. Thus, the system is able to trigger a deep sleep faster. In the awakening test, the stimulation using gamma binaural beats and yellow light shows delta brainwave amplitude in lead C3-C4 is low and alpha brainwave amplitude in lead O1-O2 is higher. Thus, this system is able to stimulate alpha wave faster than no stimulation. The conclusion is LETSLEEP Mask can regulate sleep-wake cycle for alarm device and sleep quality excalation.

INTRODUCTION Depression is a fairly serious health problem in the society. The depression case rate in Indonesia of up to

17-27% was recorded, and 5-10% worldwide at the same time. World Health Organization (WHO) even states that depression will be the second global disease burden after ischemic heart disease [1]. One of the symptoms of depression is sleep disorder. The etiology of depression that supports its relation with sleep disorder is the disruption of neurotransmitter serotonin and Cortical-Hypothalmic-Pituitary-Adrenal Corticol Axis (CHPA) hormone regulation [2]. This is certainly related to the lack of balance between sleep quality and the all day long activities.

However, in a number of particular cases, professionals with more hectic activities are forced to sleep for a shorter duration, but they have to be fresh in the following day. Entrepreneurs, doctors, technicians, or other officers sleep for only four hours in a day on average. The loud ring of the alarm from mobile phone or other





Surabaya Indonesia

alarm clocks causes a shock to the body and causes the forced sudden awakening to the body before the right time to wake. This is the reason that causes the workers do not feel fresh when they wake up.

Based on the case above, the idea to offer the solution to improve the sleep quality and freshen somebody‘s awakening process from the short sleep duration was generated. This instrument was called ‗LETSLEEP MASK

(Automatic Sleeping Mask)‘, which utilized the visible light radiation of particular wavelength and binaural

beats concept using sound frequency difference. Delta wave as the sleep quality manager and gamma wave as the sleep duration manager (alarm) were the waves that became the focus in the brain wave stimulation method.

This research was intended to produce LETSLEEP Mask as the instrument functioning as alarm that improved somebody‘s sleep quality. Therefore, this instrument was expected to be the alternative solution in

managing sleep duration and quality for poor sleep quality, sleep disorder, even depression, and able to help the Government‘s Indonesia Sehat 2025 Program (Healthy Indonesia 2025 Program) by improving the Indonesian‘s

sleep quality and duration management.

LITERATURE REVIEW

Visible Light Wave

Visible light wavelength is related to the frequency and the energy of the light. The higher the frequency, the shorter the wavelength and the bigger the energy radiated.

Visible light consists of red, orange, yellow, green, blue, and violet spectrums. In this research, the visible light utilized in LETSLEEP MASK sleeping mask is blue light with wavelength of 450-500 nm and yellow light with wavelength of 570-590 nm. The previous study states that stimuli in the form of certain light can stimulate human‘s brainwave.

Binaural beats

Binaural beats is the result of neuron superposition originated from the left and right ears [4]. The binaural beats effect works when the sound of two tones is perceived each to only one ear, using headphones. The auditory nerve transmits the information to recticular system that controls the concentration and conscious-sness called diffuse activating system. If the internal and external stimuli are compatible with the informa-tion, the recticular system will transform the brain wavelength activity to be adjusted with binaural beats stimulation from the frequency difference. In the previous study [5], the subject that heard binaural beats with delta and theta patterns was able to produce percentage change on the amplitude number from EEG data analysis.

Brainwave

The consciousness phase of human is closely re-lated to the brain activity. This activity can be mea-sured by using EEG to see the brainwave frequency produced. Based on the frequency, brainwave is classified into delta, theta, alpha, beta, and gamma.

The wave that has the lowest frequency is delta wave, which is 0.5-4 Hz. This brainwave is produced when somebody experiences a deep sleep. Meanwhile, the frequency that has the highest frequency is gamma wave. Gamma wave occurs when somebody experiences a quite high brain activity.

Figure 1. Brainwave during EEG screening[6]

Sleep Cycle

Sleep is an unconscious condition that is relatively more responsive to internal stimuli. The difference between sleep and other unconscious conditions is that the sleep cycle is predictable and the person who sleeps is less responsive to the external stimuli. Sleep‘s function is to restore the body organs (restorative).





Surabaya Indonesia

Table 1. The frequency and amplitude of sleep wave

According to Rechtschaffen and Kales [7] the stages of sleep are divided into Non Rapid Eye Movement

(NREM) and Rapid Eye Movement (REM). Based on the EEG (electroencephalogram), EOG (electrooculogram), and EMG (electromyogram) screenings, five sleep pattern stages were found, four of which are in NREM stages (deep sleep) and the rest is in REM (shallow sleep).

Electroencephalograph (EEG)

EEG is a non-invasive measuring method that represents electrical signal of brain activity gained by placing a number of electrodes on the scalp in the brain area. Thus, the brain electrical signal is obtained to be processed and analyzed. EEG is an important instrument for clinical practitioners to diagnose, monitor, and overcome nerve disorders or diseases.

Currently, portable EEGs using some required channels are available, one of which is EEG SMT. EEG SMT is an EEG device produced by OLIMEX Ltd., that has two channels called CH1+, CH1-, CH2+, and CH2- and one ground. In EEG SMT, each epoch produces 256 data per second, so the number is bigger and more precise.

RESEARCH METHOD The research method employed was quantitative causal experimental method with primary and secondary

data collection. This research took five months in Medical Instrumentation Laboratory in Faculty of Science and Technology of Airlangga University, ITD (Institute of Tropical Desease), and RSUD Dr. Soetomo Surabaya.

The research procedures conducted is presented in FIGURE 2.

Figure 2. The flowchart of research procedures

Meanwhile, the block diagram of LETSLEEP Mask instrument is presented in FIGURE 3.

Literature Review

Materials preparation

Construction designing of

ASLEEP Mask

Hardware development

Software development EEG testing EEG data

analysis Revision of

ASLEEP Mask





Surabaya Indonesia

Figure 3. Block diagram


LETSLEEP MASK PROTOTYPE The 'LETSLEEP MASK' prototype consists of several designs, first of which was the hardware assembly

that included the LED lamps and headphones fitting on the mask. LED light exposure over the user's eyes used yellow light as the gamma wave stimulator and blue light as the delta wave stimulator. The second step was creating the program using Arduino Uno, which integrated LED light with binaural beats delta and gamma.

Both designs were then tested by using EEG to obtain brainwave after they were completed. The last stage was packaging using two layers of cotton. The second layer reduced the intensity of the LEDs. The 'LETSLEEP MASK' prototype is presented in FIGURE 4.

Figure 4. LETSLEEP Mask Prototype

Construction designing of

LED and headphone

Program designing of

microcontroller and binaural

beats

Electroencephalograph testing

PSQI instrument

testing

Setting designing of

light intensity and frequency

ASLEEP Mask effectiveness

testing Data analysis





Surabaya Indonesia

Signal Data Collection

The data of brainwave signal were obtained by using EEG SMT. The electrodes placement was based on 10-20 electrode placement rule. In this study, it was recommended to have a bald-haired subject to reduce noise. The electrodes were placed on the C3-C4 lead (central) and O1-O2 lead (occipital). They could obtain sleep-wake brainwave and sleep stages [8].

LETSLEEP Mask Testing Method

Signal Analysis

The testing of 'LETSLEEP MASK' was conducted in several stages as presented in FIGURE 5. The output signal from the EEG was read using Open Vibe software in time domain. It had to be converted

into frequency domain to analyze the brainwave frequency using Matlab 7.6.

FIGURE 5. Testing on the user; (a) control condition, (b) using LED, (c) using binaural-beats, and (d) using both LED and

binaural-beat

(a) (b)

Figure 6. EEG signal in frequency domain: (a) channel 1 (C3-C4 lead) and (b) channel 2 (O1-O2 lead)

Testing using PSQI (The Pittsburgh Sleep Quality Index)

The sleep quality was tested using PSQI (The Pittsburgh Sleep Quality Index) shown in TABLE 2.

TABEL 2. The results of PSQI test

No Criteria Subject 1 Subject 2

Skor Note Skor Note 1. Subjective sleep quality 0 very good 1 fairly good

2. The duration required for sleeping

2 31-60 minutes 0 < 15 minutes

3. Sleep duration 2 5-6 hours 3 < 5 hours

4. The efficiency of sleeping habits

0 very efficient 0 fery efficient

5. Sleep disorders 1 < once a week 2 1-2 times a

week 6. Medication of sleep 0 never 1 < once a week

7. Activity disruptions in the daytime

1 rarely happens 1 rarely happens

PSQI global score 6 Sleep disorder occurred 8 Sleep disorder

occurred

TABEL 3. The results of lulling test





Surabaya Indonesia

No Treatments

Results Channel 1 (C3-C4) Channel 2 (O1-O2)

Dominant waves

Amplitude (μV)

Dominant waves

Amplitude (μV)

1. No treatment Delta & Alfa 38 delta 120 2. Blue LED Delta 450 delta 56

3. Blue LED + Binaural beats Delta

Delta 890 delta 26,5

TABEL 4. The results of awakening test

No Treatments

Results Channel 1 (C3-C4) Channel 2 (O1-O2)

Dominant waves

Amplitude (μV)

Dominant waves

Amplitude (μV)

1. Tanpa Perlakuan

Delta 405 delta 54

2. Binaural beats Gamma

Delta 108 delta 98

3. LED Kuning+ Binaural beats Gamma

Delta 26,5 delta 155

Testing using EEG (Electroencephalograph)

EEG signal was tested to prove that ‗LETSLEEP MASK' could lull and wake user. The dominant frequency

of the brainwave when the user was asleep was observed. The sleep indicator could be observed from the dominant frequency between 0-4 Hz which was known as delta brainwave. Therefore, the focus of the observation was the amplitude of the delta wave in C3-C4 lead (channel 1) and O1-O2 lead (channel 2). Results of EEG recording was shown in Table 3 and 4.

Data Analysis of LETSLEEP Mask Testing

Testing by using PSQI

PSQI was indicated by seven components. The first indicator showed subjective sleep quality. The second indicator showed the time it took to fall asleep (sleep latency), the third and the fourth indicators showed the sleep duration and habitual sleep efficiency, the fifth indicator showed sleep disorders (sleep disturbance), the sixth indicator showed the medications to sleep (use of sleeping medication), and the seventh indicator showed activities disruption during the daytime (daytime dysfunction). It was found that the subjects had Global PSQI scores of 6 and 8, which means they were experiencing sleep disorders.

Brain signal testing using EEG (Electroencephalograph) The testing on the C3-C4 lead was conducted to analyze delta wave during sleep, so it showed that the

higher the occurring amplitude the deeper the person slept. Meanwhile, the testing on the O1-O2 lead was done to see the occurrence of delta wave too, but it was different from the C3-C4 lead. O1-O2 lead could show alpha wave (initiate awareness) so deep sleep stage could show delta wave amplitude decreases and vice versa.

The results showed that for the person that sleeps without any treatment (normal), the brain wave was the dominated by delta and alpha waves (sleepiness). Meanwhile, when the person was given one treatment in the form of a blue LED (stimulus delta brain waves) on channel 1 (C3-C4), the amplitude of the brainwave increased and in channel 2 (O1-O2) the amplitude decreased, which means a person slept deeper. When two stimuli were given to the subject at the same time, the amplitude of brain's delta wave on the C3-C4 lead increased (nearly twice higher) and on the O1-O2 decreased. It showed that deep sleep might happen by using two stimulation in the form of blue light and the sound of binaural beats. More details can be seen in FIGURE 7 and FIGURE 8 below.





Surabaya Indonesia

Figure 7. The graph of the results of delta wave amplitude during lulling

In the awakening test results, it could be seen on channel 1 (C3-C4) that without treatment obtained the

amplitude of delta wave was quite high and channel 2 (O1-O2) was still fairly small, suggesting that the person was still in deep sleep stage, treatment stimulus using binaural beats alone could reduce the amplitude of the delta wave on channel 1 and increase the amplitude in channel 2, which means the sleeping conditions were getting closer to the waking state, while using two stimulations of yellow LED and binaural beats gamma obtained the amplitude of the low frequency delta waves on channel 1 and higher on channel 2 that shows a person will start to wake up.

Figure 8. The graph of the results of delta wave amplitude during lulling awakening

CONCLUSION 'LETSLEEP MASK' is an instrument that can function as an alarm based on high amplitude generated in the

process of sleeping and awakening. 'LETSLEEP MASK' also improves sleep quality with irregular waves that pass through the delta-theta-alpha-beta-gamma phase, and vice versa.

ACKNOWLEDGEMENTS We, the researcher, address a deep gratitude to our parents who have given full support at every step. Thanks

to Drs. Eko Supeno, M.Si; Drs. Siswanto, M.Si; and Dr. Moh. Yasin, M.Si who have support ed our research and scientific publications. Thanks to all lecturers who took time to listen to our complains, the subject as the primary data that we collect, our friends in Department of Physics, and people who have helped the process of these scientific publications.

050

100150200250300350400450

Tanpa Perlakuan Binaural Beats Gamma LED Kuning+ BinauralBeats GammaA

mpl

itud

e of

del

ta w

ave

(μV

)

Treatment

The Results of Testing (Awakening)

Channel 1 (C3-C4) Channel 2 (O1-O2)

0100200300400500600700800900

1000

Tanpa Perlakuan LED Biru LED Biru + BinauralBeats Delta

The

am

plit

ude

of d

elta

wav

e (μ

V)

Treatment

The Results of Testing (Lulling)

Channel 1 (C3-C4) Channel 2 (O1-O2)





Surabaya Indonesia

REFERENCE 1. Kartika, Unoviana. 2013. Empat Alternatif Pengobatan Depresi, Berita Kompas.com, Oktober 2013 2. Amir N. 2004. Aspek Neurobiologi Molekuler Depresi. JIWA. XXXVII:2 3. Paulinasari, Wenny. 2010. Pengaruh Gelombang Cahaya Terhadap Sinyal Otak Terkait Fungsi

Penglihatan. Skripsi S1 Fisika: FST Universitas Airlangga. 4. E. Ozimek, Sound and Its Perception. 2002. Aspects of Physical and Psychoacoustics. PWN:

Warszawa. 5. Owens, J. E; Atwater, F. H. EEG Corellates of an Induced Altered Sttate of Conciousness: ―Mind

Awake/Body Asleep‖ 6. Khandpur, R. S. 1986. Handbook of Biomedical Instrumentation. New York: McGraw-Hill. 7. Rechtschaffen A, Kales A. 1968. A Manual of Standardized Terminology Techniques and Scoring

System for Sleep Stages of Human Subjects. NIH Publication 204. Washington, DC: U.S. Government Printing Office, Departement of Health Education and Welfare

8. Tarokh, L., M. A. Carskadon, P. Achermann. 2010. Developmental Changes in Brain Connectivity Assessed Using The Sleep EEG, Journals of Neuroscience. Elsevier Ltd.



112


Surabaya Indonesia

ORAL PAPER OF COMPUTATIONAL PHYSICS, CHEMISTRY & MATHEMATICS (OCPC)



Ameen et al. 113


Surabaya Indonesia

Evaluation and Analysis of Smart Antennas by Adaptive Beam-forming

Using Least Mean Square (LMS) Algorithm

Mudhaffer M. Ameen* and Avin Jawhar Ali*

Physics Department, College of Education, Salahaddin University, Erbil – Kurdistan Region- IRAQ *Corresponding Author’s E-mail: [email protected]

Abstract. The smart antennas system are widely used for wireless communications, because they have ability to increase the number of simultaneous users, and satisfy the requirements for expanding frequency reuse and for combating propagation problems and interference. Smart antenna performs two main functions such as: Direction of Arrival estimation (DOA) and adaptive beam-forming. Using beam-forming algorithm, Smart antenna is able to form main beam towards desired user and null in the direction of interfering signals. In this work, overview of smart antenna system concerning parameters, performance and analysis have been studied. A uniform linear array in which the elements are fed with a current of equal magnitude with uniform progressive phase shift along with line of a system of equal spaced elements (straight line array). A communication system which contains a smart antenna system to show in the improvement the radiation pattern and performance of linear and planar array smart antenna have been modeled. The investigation of the smart antenna was done by considering multiple directions of arrivals of signals, beam pattern, radiation pattern by varying the number of elements in the array and the spacing between the sensor elements. The evaluations and results are obtained using MATLAB program which indicates that the smart antenna performs over conventional single element antenna and it provides away of better performance. The simulations were done confirmed that smart antenna systems analyzed and evaluated using Least Mean Square (LMS) beam-forming algorithm are able to adjust their pattern and to enhance desired signals. Moreover, to reduce interference, the (LMS) beam-forming algorithm is tested using the following parameters:

1. The number of elements, and . 2. Equidistance element spacing and .

The results show that the smart antennas radiation patterns are related to the number of elements in the array. Whenever, the number of elements increases accordingly good directing ability increases. Whereas, the optimum value of the beam pattern, and radiation pattern occur at the number of elements, and the equidistance element spacing, ⁄ .

INTRODUCTION A smart antenna system combines multiple antenna elements with a signal processing capability to optimize

its radiation and/or reception pattern automatically in response to the signal environment [1]. An antenna in telecommunication system is a port in which radio frequency (RF) energy is combined from

the transmitter to the outside world for transmission purposes, and in converse, to the receiver from the outside world for reception purposes [1, 2, 3]. Smart antennas have a special characteristic of purifying that capable for obtaining energy from a certain direction and at the same time block others from different ways. In addition, smart antenna increases the range and capacity of systems with aid to minimize the interfering and diminishing. Furthermore, they can raise signal to noise ratio, decrease multipath and other interference signals by directing the main beam to the user and formation of nulls in the route of the interference signals [4]. With a mean of internal feedback control, smart antenna capable to alter the radiation pattern during the antenna‘s operation [5]. Some people believe that smart antenna systems are smart antenna, but the fact is antenna by itself is not smart; it is the digital signal processing capability alongside with antennas which make them smart [2,3].

The smart antenna consists of (i) array antenna (ii) complex weights and (iii) adaptive signal processor as shown in Fig. (1). The array antenna comprises of a Uniform Linear Array (ULA) or Uniform Circular Array (UCA) of antenna element. The individual antenna elements are assumed to be identical, with omnidirectional patterns in the azimuth plane. The signal received at the different antenna elements are multiplied with the complex weights and then summed up. The complex weights are continuously adjusted by the adaptive signal processor which uses all available information such as pilot or training sequences or knowledge of the properties of the signal to calculate the weights. This is done so that the main beam tracks the desired user and/or nulls are placed in the direction of interferers and/or side lobes towards other users are minimized. It should be noted that the term "smart" refers to the whole antenna system and not just the array antenna alone [5, 3].



Ameen et al. 114


Surabaya Indonesia

Figure 1: Block diagram of a smart antenna system [5]

EXPERIMENTAL The ―adaptive arrays‖ was first coined by Van Atta in 1959 to describe a self-phase array. Self-phase array

reflect all incident signal back in the direction of arrival by using phase conjugation. Self-phase arrays are instantaneously adaptive arrays since they essentially reflect the incident signal in a similar fashion to the classic corner reflector [6]. In 2014, Dungriyal. et al. have implemented a new and effective array beam-forming algorithm called Matrix Inverse Robust Least Mean Square (MIR-LMS) algorithm which combines the individual good aspect of Sample Matrix Inverse (SMI), Normalized LMS (NLMS) and Robust LMS (R-LMS) algorithm [7].

RESULTS AND DISSCUSSION

The Least Mean-Square (LMS) Algorithm:

The LMS algorithm has become one of the most popular adaptive signal processing techniques adopted in many applications, including antenna array beam-forming [8].It was invented by Stanford,s et al. in 1959 [9]. The least Mean Square (LMS) based algorithms offer a relatively simple adaptive array beam-forming solution. However, the performance of these algorithms often depends on the actual step size adaptation process. Also, since these algorithms make use of LMS processing, their operations are influenced by the characteristics of the input signals [8].

The Least Mean Square (LMS) algorithm is an adaptive algorithm, which uses a gradient-based method of steepest decent. It uses the estimates of the gradient vector from the available data. It incorporates an iterative procedure that makes successive corrections to the weight vector in the direction of the negative of the gradient vector which eventually leads to the minimum mean square error compared to other algorithms. LMS algorithm is relatively simple it does not require correlation function calculation nor does it require matrix inversions. Consider a Uniform Linear Array (ULA) with N isotropic elements, which forms the integral part of the adaptive beam-forming system as shown in the Fig.(2).The output of the antenna array x(t) is, [8].

∑

S(t) denotes the desired signal arriving at angle and denotes interfering signals arriving at angle of incidences respectively. and represents the steering vectors for the desired signal and interfering signals respectively. Therefore it is required to construct the desired signal from the received signal amid the interfering signal and additional noise n(t). From the method of steepest descent, the weight vector equation is given by [10]

⁄ [ { } ] (2) [ ] (3) (4)

Where μ is the step-size parameter and controls the convergence characteristics of the LMS algorithm e2(n) is the mean square error between the beam-former output y(n) and the reference signal which is given by[10]



Ameen et al. 115


Surabaya Indonesia

[ ] (5)

Figure 2: Adaptive Beam-forming System [10].

The gradient vector in the above weight update equation can be computed as [10]. { } (6)

In the method of steepest descent the biggest problem is the computation involved in finding the values r and R matrices in real time. The LMS algorithm on the other hand simplifies this by using the instantaneous values of covariance matrices r and R instead of their actual values [10].

(7) (8)

Therefore, the weight update can be given by the following equation, [10] [ ] (9) (10)

The LMS algorithm is initiated with an arbitrary value w(0) for the weight vector at n=0. The successive corrections of the weight vector eventually leads to the minimum value of the mean squared error. Therefore the LMS algorithm can be summarized in following equations [10].

Output: (11) Error: (12) Weight: = (13)

The LMS algorithm initiated with some arbitrary value for the weight vector is seen to converge and stay stable for ⁄ where is the largest eigenvalue of the correlation matrix R. The convergence of the algorithm is inversely proportional to the eigenvalues spread of the correlation matrix R. When the Eigen values of R are widespread, convergence may be slow. The Eigen value spread of the correlation matrix is estimated by computing the ratio of the largest Eigen value to the smallest Eigen value of the matrix. If μ is chosen to be very small then the algorithm converges very slowly. A large value of μ may lead

to a faster convergence but may be less stable around the minimum value. An upper bound for μ based on several approximations as (μ 1 / 3tace R) [10].

Linear Beam Pattern for a Linear Array Smart Antennas:

The linear beam pattern in ( ) was computed for a smart antenna array by using number elements and 8 with interelement spacing . The array beam patterns are shown in Fig. (3), for and 8 at .

Whereas, Fig. (4) shows the linear beam pattern in ( ) versus in degree for number of elements, 8 at with inter-element spacing and .

It is obvious from the figures that the maximum power of beam pattern will occur at the given values of .



Ameen et al. 116


Surabaya Indonesia

Figure 3: Beam pattern of ( and ) linear array with spacing of using LMS algorithm

Number of data sample

-70

-60

-50

-40

-30

-20

-10

0

10

-100 -50 0 50 100

Po

we

r (d

B)

θ (degrees)

Linear beam pattern (N=7, d=0.5λ, θ=45

-60

-50

-40

-30

-20

-10

0

10

-100 -50 0 50 100

Po

wer

(d

B)

θ (degrees)

linear beam pattern (N=8, d=0.5λ, θ=45)

-80

-60

-40

-20

0

20

-100 -50 0 50 100

Mag

nit

ud

e (d

B)

θ (degrees)

Linear beam pattern (N=8, d=0.75λ, θ=75



Ameen et al. 117


Surabaya Indonesia

Figure 4: Beam pattern of number of elements ( 8) linear array with spacing of ( ) using LMS algorithm Number of data sample

Normalized Radiation Pattern for Linear Array of Smart Antennas:

The radiation pattern of smart antenna is nothing but a graph which shows the variation in actual field strength of electromagnetic field at all points which are at equal distance from the antenna. Obviously the graph of radiation pattern will be three-dimensional and hence cannot completely be represented.

In fact, the graphical representation of radiation of a smart antenna as a function of direction is given the name radiation pattern of the antenna. If the radiation from the antenna is expressed terms of field strength (volt/ meter), the radiation pattern is called as ―the field strength pattern‖. If, on other hand, the radiation in a given direction is expressed in terms of ―power per unit solid angle‖, then the resulting pattern is called ―power

pattern‖. Since the radiation pattern is three dimensional figures and hence the coordinate system usually used for the

same is the spherical coordinate . The antenna is assumed to be located at origin of the spherical coordinate system and the field strength is specified at points on the spherical surface of radius . The shape of the radiation pattern does not depend on the radius provided .

The direction of field strength for the radiation field is always tangential to the spherical surface of imaginary sphere of radius and for vertical antenna electric field strength is in the direction and for the horizontal loop in the direction .

A complete radiation pattern is in a three-dimensional solid figure and gives the radiation for all angles of the and . However, to represent the radiation pattern on a plain paper (that is in two dimensions) a cross-section through three dimensional patterns is taken. Cross-section generally taken are in horizontal plane (when ) and in a vertical plane (when constant). Thus the two dimensional pattern so obtained from three-dimensional pattern by cutting with a horizontal and vertical planes are respectively, known as ―Horizontal pattern‖ and ―vertical pattern‖.

The normalized radiation patterns for a linear smart antenna are shown in Fig. (5) for linear array element and 8, respectively at optimum inter-element spacing, . Whereas, each figure represents LMS algorithms for as:

It is obvious from the figures that the main beam of a large array can be resolved because of its narrower beam-width, the signals-of- interest (SOIs) will be more accurately recognized therefore, it allows the smart antenna system to reject whose signal-not- of-interest (SNOIs). Again, looking at the figures, the requirements of the beam-formed of a linear array can be chosen at a maximum of interest (SOIs) at a specific angle. However, the requirements of null signal noise of interest (SNOIs) at other angle can be rejected for different values of and .

To get the best radiation pattern one must maximize signal of interest and minimized signal noise of interest, therefore, LMS algorithm was used to plot radiation pattern for number of elements of smart antenna, and 8 with interspacing element, .

Next, one can observe that as the number of elements of smart antenna array increases, and then it has the following effects on the radiation patterns:

The width of the main lobe decreases that is, it becomes narrower. In fact, this is crucial for the applications of smart antennas since a single narrow beam is required to track a mobile or cluster of mobile operation.

-80

-60

-40

-20

0

20

-100 -50 0 50 100

Mag

nit

ud

e (d

B)

θ (degrees)

Linear beam pattern (N=8, d=λ, θ=75



Ameen et al. 118


Surabaya Indonesia

The number of side-lobes decreases. In an addition to the level of the first and subsequent side-lobes decreases in comparing with the main lobe. Whereas side-lobes represent power radiated or received in potentially unwanted directions. So in a wireless communications system, side-lobes will contribute to the level of interference spread in the cell or sector by a transmitter as well as the level of interference as seen by a receiver when antenna arrays are used.

The number of nulls in the pattern increases. However, in interference cancellation applications, the directions of these nulls as well as the null depths have to be optimized. All these affects are shown in Figs. (3) and (4).

Moreover, increasing the inter-element spacing , the number of sidelobes increases too in comparing with the main lobes. Whereas, the beam width of the main lobe decreases accordingly which means that the directivity and gain increase toward desired direction as shown clearly in Figs. (5) and (6).

Figure 5: Normalized radiation pattern of a five-element ( and ) linear array with interelement spacing

using LMS algorithm and number of data sample

Figure 6: Normalized radiation pattern of an eight-element ( ) linear array with element spacing and

respectively using LMS algorithm with number of data sample

CONCLUSION The power of beam pattern for linear smart antennas depends upon the number of element arrays as well as

the inter-element spacing. Moreover, its value increases with increasing the number of element arrays and inter-element spacing as well which is shown in Figs. (3) and (4). The radiation pattern value for linear array of smart antennas depends upon number of element arrays and inter-element array spacing. When the number of element arrays increases the main lobe of radiation pattern becomes narrower and the number of nulls in the radiation pattern increases accordingly. Moreover, the number of side lobes increases too by increasing the inter-element spacing as compared with main lobes. All these effects are shown in Figs. (5) and (6). The beam width of the main lobe decreases accordingly which means that the directivity and gain toward desired direction increases. This is clear in Figs. (5) and (6).

REFERENCES 1. Boucher, N. J., The Paging Technology Handbook, John Wiley and Sons, 2nd ed., 1992. 2. [arsamba, M. C., and Talwar, M., Evaluation of Smart Antenna for 3G Network: A Survey, History,

Vol. 10, No. 21, pp. 24-32, 2014. 3. Balannis, C. A., Introduction to Smart Antennas, 1st ed., Morgan and Claypool Publisher, 2007.



Ameen et al. 119


Surabaya Indonesia

4. [4] Jain, M., and Sharma, B. S., Implementation of Wireless Communication Using Adaptive Beam-forming of Smart Antenna, Journal of Electronics and Communication Engineering, Vol. 7, No. 5, pp. 65-71, 2013.

5. Durrani, S., Investigations into Smart Antennas for CDMA Wireless Systems, PhD. Thesis, The University of Queensland, Brisbane, Australia, 2004.

6. Rodríguez – Estrello, C. B., and Pérez, F. A. C., An Insight into the Use of Smart Antennas in Mobile Cellular Networks: INTECH Open Access Publisher, 2011.

7. Balannis, C. A., Antennas Theory: Analysis and Design, 3rd ed., Johan Wiley and Sons Publisher, 2004.

8. Warudkar, S. N., Chincholkar, Y., and Kawitkar, R., Convergence Performance of LMS & Combined LMS-LMS Beam-forming Algorithm, International Journal of Science, Engineering and Technology Research, Vol. 2, No. 9, pp. 1664-1667, 2013.

9. Dungriyal, K., Ananad, S., and Kumar, D. S., Performance of MIR-LMS Algorithm For Adaptive Beam Forming in Smart Antenna, International Journal of Innovative Science, Engineering & Technology, Vol. 1, No. 5, pp. 21-27, 2014.

10. Surendra, L., Shameem, S., Khan, and H., Performance Comparison of LMS, SMI and RLS Adaptive Beam-forming Algorithms For Smart Antennas, International Journal of Computer Science and Technology, Vol. 3, No. 2, pp. 973-977, 2012.



120


Surabaya Indonesia

ORAL PAPER OF ENVIRONMENTAL BIOCHEMISTRY AND BIOTECHNOLOGY

(OEBB)



Parwata et al. 121


Surabaya Indonesia

Induction of Apoptosis and Antiangiogenesis Effects

of Pinostrobin from Kaempferia pandurata Roxb against

Induction of Fibrosarcoma Mice Results Benzopiren

Adi Parwata1*, Sukardiman2, Mulja H. S.3, Alit Widhiartini4

1) Nature Materials Study Group, Laboratory of Organic Chemistry, Department of Chemistry, Faculty Mathematics and Natural Sciences, Udayana University, Denpasar, Bali.

2,3) Laboratory of Phytochemistry and Pharmacognosy Faculty Pharmacy,Airlangga University, Surabaya 4) Section of Pharmacy, Faculty of Medicine, University of Udayana, Jalan PB Sudirman, Denpasar

5) Laboratory of Biophysic, Departemen of Physic,Faculty Mathematics and Natural Sciences, Udayana University, Denpasar, Bali.


Abstract. Induction of apoptosis and antiangiogenesis effects of Pinostrobin from Kaempferia pandurata Roxb against Fibrosarkoma mice results benzopiren induction has been done. Examination or surgery begins taking tissue fibrosarcoma in mice infected and weigh fibrosarcoma obtained . Fibrosarcoma tissues were then stored in containers that have contained 10 % formalin . Weighing results showed that the concentration of pinostrobin oral 80 mg / kg can inhibit the growth of fibrosarcoma with a gram weight of 68.62 % and a cancer drug ( control + ) there is resistance 95.95 % compared to the negative control which is only given orally CMC - Na, this means pinostrobin potentially be developed as a cancer chemotherapy. Furthermore done shooting patohistologi tissue fibrosarcoma with HE staining with a light microscope with 400x magnification. The results obtained showed many chromatin (polikromatin) which prove the damage caused by having fibrosarcoma cells. Immunohistochemical assay showed oral pinostrobin concentration 80 mg / kg body weight can increase the expression of p53 to apoptosis induction could take place and the decreased expression of VEGF angiogenesis which proves the existence of barriers .

INTRODUCTION Rhizome plants of Kaempferia pandurata Roxb as a traditional medicine in Indonesia in general is widely

used as a dry cough , cancer sores , irritation of the colon , stomach bloated , difficult urination in children , inflammation of mucous membranes in the mouth of the womb , dysentery , and tumor / cancer. Rhizomes of Kaempferia pandurata Roxb was extracted with n-hexane produce levels of flavonoid components pinostrobin relatively large , namely 2.5 % and as much as 20 grams / 800 grams of powder or alpinetin approximately 1 % (Oka, 2000) . The results of the study as it turns out pinostrobin have anti-oxidant activity and relaxes smooth muscle . One method to test materials - materials that are cytotoxic are the toxicity tests on larvae shrimp ( Artemia salina L. ) . This method is often used for initial screening of the active compounds as anticancer agent in the extract of the plant because they are easy , inexpensive , fast and reliable outcome (Oka, 2000)

The content of rhizome pinostrobin in Kaempferia pandurata Roxb is big enough then pinostrobin isolation as a pure substance can be done fairly quickly. Pinostrobin the polarity of the structure is reduced as a result of intra-molecular hydrogen bonding between the carbonyl group at C-4 with a hydroxy group at C-5 then extraction can be performed with less polar solvents such as chloroform and n-hexane (Oka, 1998)

It can be seen in the structure below :

O

O

CH3O

OH

Pinostrobin

O

CH3O

HO

O

Alpinetin

OHO

PinocembrinOH O

Based on the structure then pinostrobin can be identified by UV-Vis spectroscopy to look at max as a

characteristic of flavonoids consists of two absorption bands are bands I (325 nm) and II bands (287 nm) and the bathochromic shift when coupled shear reaction AlCl3 at 20 -26 nm to indicate the presence of-OH substituent at C-5 position. Identification by IR spectroscopy to look at group functions, Proton and Carbon NMR spectroscopy to look at the type and amount of the H and C atoms of Pinostrobin and mass spectroscopy to look at relative molecular mass (Mr) and fragmentation - fragmentation of 5-hydroxy-7-methoxy flavanones or pinostrobin (Markam KR, 1988;. Silverstein, Bassler and Morrill, 1981)



Parwata et al. 122


Surabaya Indonesia

Pinostrobin flavonoid compounds have been isolated from the rhizome of Kaempferia pandurata Roxb with levels of ± 2.5%, and it is known that pinostrobin compounds have activity inhibiting the growth of human breast cancer cells and inhibit the enzyme activity of DNA Topoisomerase I (Oka, 1998). DNA Topoisomerase enzymes have an important function in intracellular processes, which play a role in the replication process and the proliferation of cancer cells. By being such that the enzyme activity by DNA Topoisomerase inhibitor compounds, the process of bonding between the enzyme with the longer DNA of cancer cells. So that will be formed Protein Linked DNA Breaks (PLDB), resulting in cancer cell DNA fragmentation and the subsequent effect on the cells in the process, especially the process of cell replication, which ended with the death of cancer cells (Lowe Scoot.W., Copero gerard Enrique and Evan, 2004; Sukardiman et al., 2006).

The presence of DNA damage and subsequent cell cancer affects the cells in the process, especially the process of cell replication ends with the death of the cancer cell apoptosis (Sukardiman et al., 2006). Thus compounds suspected pinostrobin have cancer chemopreventive activity and therapeutic in vivo in mice results Fibrosarkoma cancer induction Benzopirena. The presence of DNA damage allegedly cells can activate the p53 tumor suppressor gene can induce apoptosis in addition, can also affect the cell cycle by affecting the cell cycle inhibitor p27 and cyclin-dependent enzymes barriers kinase (C. Whibley, 2009).

The presence of p53 activation and stabilization products by pinostrobin then be expected to inhibit angiogenesis through the mechanism of the effect of down-regulation or decreased expression of VEGF (Vascular Ephidermal Growth Factor) and decreased expression of Cyclooxigense-2 (COX-2) and matrix metalloproteinase - 9 (MMP-9) were also involved in the metastasis of cancer cells. Angiogenesis is the formation of blood capillaries of solid cancer cells that have a size larger than 1-2 mm, the result would be the development and growth of cancer and may eventually spread throughout the network (C. Whibley, 2009; Samiasih S., 2010; Chrestella J., 2009).

MATERIALS AND METHODS The material used is the rhizome powder Kaempferia pandurata Roxb were obtained from Badung Market ,

Denpasar, Bali. Mice used were male mice, aged 2 months and weigh 20-25 grams. The chemicals used are technical hexane, n-hexane (pa), methanol (pa), ethanol (pa), chloroform (pa), ethyl acetate (pa) distilled water, silica gel and TLC plates with silica gel GF Aluminum-254 from E Merck), primary and secondary antibodies p53 and VEGF 10% formalin buffer, PBS, DAB, H2O2 3% and Destilate Water (DW). The tools used are Light Microscopy, polilysin preparations and glass objects.

Induction of cancer cell Fibrosarkoma Induction in Mice Results Benzopirena

Fourty male mice tails undergo a process of adaptation studies, all mice were subsequently induced benzopiren get as much as 0.3 mg / 0.2 mL in oleum olivarium subcutaneous injection in the scapular region 5 times, every other day. Subsequently, the whole atmosphere and maintained in mice the same diet for two months / form of cancer in the nape area, after reaching the cancer volume ± 100 mm3, mice with cancer were randomly divided into 3 groups. Group I is the negative control, only given CMC-Na. Group II was given pinostrobin dose of 80 mg / kg; Group III was given cyclophosphamide at a dose of 13.33 mg / kg. All test materials are given in intraperitonial (oral) and administered daily for 14 days. After the mice were sacrificed and done taking the cancerous tissue and then do the weighing of cancer.

Examination of expression of p53 and VEGF expression in Immunohisto-Chemistry

At first conducted clearing and rehydration in stages against both cancer slice preparations both the control group and the treatment group were stored in 10 % formadehid solution , then washed 3 times with destilate water ( DW ) , respectively 5 minutes then drained fluid around the network . After that the tissue pieces circled with a pen and then immersed in 3 % H2O2 in DW 15 minutes at room temperature . Then the preparations were rinsed 3 times with DW and 3 times with PBS each 5 minutes.

The next step preparations were incubated with 10 % normal goat serum 30 min at room temperature and then rinsed 3 times with PBS each 5 minutes . Then the preparations were incubated with primary and then rinsed 3 times with PBS each 5 minutes , after which it was incubated with the secondary for 30 min at room temperature . l avidin and biotin were incubated in 1 ml of PBS . l 10 At the same time , 10 Furthermore, preparations were rinsed 3 times with PBS each 5 minutes and then incubated with a mixture of avidin and biotin 30 min at room temperature and then rinsed 3 times with PBS 5 min each and incubated with DAB



Parwata et al. 123


Surabaya Indonesia

solution . The next step performed counterstaining with hematoxylin and then dehydrated , then covered with a cover glass preparations . Mixture further examined by microscope to see the expression of p53 and VEGF expression . Penbacaan done with 400x magnification are seeing positive and negative .


Surgery Fibrosarkoma

Prior to the treatment of one of the infected mice were sacrificed to prove positive fibrosarcoma fibrosarcoma to see whether histopathological tissue fibrosarcoma with HE staining. The results obtained surgically weighs = 3.4756 grams and examination with magnification 400x hitopatologi seen many chromatin (polikromatin) which prove the damage suffered as a result of fibrosarcoma cells as shown by the following figure 1 :

Figure 1. Histopathology Fibrosarkoma Cells in Mice Results Induction benzopiren

After treatment for 14 days later mice were sacrificed her cancer tissue were taken for subsequent

examination as severe cancer who received treatment , making preparations for the examination polilysin p53 , COX , VEGF and paraffin blocks . The results also show that the concentration of pinostrobin oral 80 mg / kg could inhibit the growth of fibrosarcoma with a gram weight of 68.62 % and the cancer drug ( control + ) there is resistance 95.95 % compared to the negative control which is only given orally CMC - Na , this means pinostrobin potentially be developed as a cancer chemotherapeutic agent as shown in the following table 1 :

Table 1 . Weighing Weight Fibrosarkoma Mice Results of Treatment Results Jenis Perlakuan Berat Fibrosarkoma (gram )

I II III IV rata-rata Without treatment 3,4756 3,4557 3,4656 CMC-Na [ Control (-)] 4,2765 4,1665 4,1095 4,1383 4,1727 Pinostrobin 80 mm/kg WW 1,0665 1,0605 1,1198 1,1027 1,0874 Cyclophosphamide [Control (+)] 0,1709 0,1021 0,2553 0,1021 0,1576

Results painting , inspection and reading of p53 and VEGF expression can be seen in the following table 2 and figure 2 :

Table 2 . The reading results with p53 and VEGF Immunohistochemical Method

No Jenis Sampel p53 VEGF

Positif negatif Positif Negative 1. CMC-Na (Control negatif/C-) 287 122 342 161 2. Pinostrobin 80 mg/kgWW 299 107 299 107 3. Cancer Drug 265 250 270 237



Parwata et al. 124


Surabaya Indonesia

Histologi VEGF Histologi VEGF Histologi VEGF Control (-) Pinostrobin 80 mg/kgBB Control (+)

Figure 2. Results Immunohistochemistry of VEGF treatment results

Based on the pictures and the table above shows the results of treatment with VEGF expression Pinostrobin

80 mg / kg reduced the number of blood vessels compared to the number of blood vessels in the negative control ( no treatment ) . This proves that Pinostrobin can inhibit angiogenesis through decrease in the number of blood vessels ( Pratiwi D.2009; Puspita N.2009 ) .

CONCLUSION

Based on the above results it can be concluded is : 1. Isolated compounds obtained are pinostrobin or 5-hydroxy-7-methoxy flavanones 2. Pinistrobin potential to be developed as a cancer chemotherapy because of fibrosarcoma weight

decreased more than 50 % and in the statistical analysis turns pinostrobin have a significant effect (p < 0.05) in inhibiting cancer growth

3. Histopathology examination fibrosarcoma cells by HE staining in get that there are a lot of chromatin ( polikromatin ) on fibrosarcoma as evidence of damage to normal cells due to fibrosarcoma

4. Oral pinostrobin concentration 80 mg / kg body weight can reduce 68.62 % by weight of fibrosarcoma and the cancer drug ( control + ) decreased 95.95 % and this concentration can increase the expression of p53 so that apoptosis can take place and decreased the expression of VEGF signaling can be inhibited angiogenesis or as antiangiogenesis, this means pinostrobin potentially be developed as a cancer chemotherapeutic agent .

ACKNOWLEDGEMENTS The author would like to thank profusely to all parties, especially Prof.Dr.Sukardiman, Drs , Apt . MS and

Dr.rer.nat . Mulja Hadi Santosa , Apt for all the guidance and permission to use the lab Phytochemistry and Pharmacognosy Faculty . Pharmacy Airlangga University, and Mr. Eko Adiputra for his help in the maintenance and surgery in the Lab mice . Animal Fac. Pharmacy Airlangga University during the study undertaken . We did not forget to thank profusely to the Head and staff LPPM Unud over Grant Scheme funding Character Letter Agreement Contract Number :175.79/UN14.2/PNL.01.03.00/ 2013 , May 16, 2013 date given in the research fund this .

REFERENCES 1. L.Bail, L.Aubourg and G. Habrioux, Effect of pinostrobin on estrogen metabolism and estrogen

reseptor transactivivation, Cancer Lett, Aug. 1(2000),37-44. 2. S.Banerjee, B.R.Carlos and B.A. Bharat, Suppression of 7,12-Dimethylbenz(a) anthracene-induced

Mammary Carcinogenesis in Rats by Resveratrol , Carcinogenesis, 62 (2002), 4945-4954. 3. J.Chrestella, Gambaran Immunoekspresi Matrix Metalloproteinase- 9 (MMP-9) pada Lesi-Lesi

Prakanker dan Karsinoma Serviks Invasif. Thesis, (2009), Medan Universitas Sumatra Utara. 4. J.Cumming, J.F. Smyth, DNA Topoisomerase I and II as target of Rational Design of New Anticancer

Drugs, Ann Oncology, Aug, 3(7) (1993), 533-534. 5. M.Jianguo, Karin, A.Reed, James and M.Gallo, Cells Designed to Deliver Anticancer Drug by

Apoptosis,Cancer Research, 62(2002),1382-7. 6. D.Manahan and R.A.Wienberg, The Hallmarks of Cancer, Cell, 100 (2002), 57-70. 7. S.Nagata, Apoptosis by Death Factor, Cell. 88 (1997), 355 – 365.



Parwata et al. 125


Surabaya Indonesia

8. Oka Adi Parwata, Isolasi, Identifikasi Senyawa Pinostrobin pada Rimpang Temu Kunci (Kaempferia pandurata Roxb) dan Standarisasi Ekstrak Etanol berdasarkan Kadar Pinostrobinnya dengan KLT Densitometri.Thesis (1998) PPS Unair, Surabaya.

9. Oka Adi Parwata, Uji Toksisitas Senyawa Pinostrobin pada Rimpang Temu Kunci (Kaempferia pandurata Roxb), DIK/DIKS (2000), Lemlit Unud, Bali.

10. S.Pai, D. Kaustubh and M.Debabrata, Central Role of p53 Regulation of Vascular Permeability Factor/Vascular Endothelial Growth Factor (VPF/VEGF) Expression in Mammary Carinoma, Cancer Research, 61(2001), 6952-6957.

11. D.Pratiwi,Ekstrak Etanol Jeruk Nipis (Citrus auranttiifolia (C) Swingle) meningkatkan ekspresi p53 pada sel payudara tikus galur spague Dawley terinduksi 7,12-dimetilbenzene-[A] Antrasena, Cancer Chemoprevention Research Center (2009) Fakultas Farmasi, UGM, Yogyakarta.

12. N.Puspita, 2009, Ekstrak etanol kulit jeruk mandarin (Citrus reticulata ) meningkatkan ekspresi faktor VEGF pada sel kanker kolon WiDr, Cancer Chemoprevention Research Center, (2009) Fakultas Farmasi, UGM, Yogyakarta.

13. R.Ravi, M.Bijoyesh, M.B.Zaver, H.S. Carrie, A.Dmitri, Regulation of tumor angiogenesis by p53-induced degradation of hypoxia-inducible factor 1α, Gene and Development, 14(2000).

14. S.Samiasih, Perbedaan Ekspresi VEGF Sel Adenokarsinoma Kolorektal Tikus Sprague Dawley Dengan dan Tanpa Pemberian Ekstrak Phyllantusniruri, Thesis, (2010), Semarang, UNDIP.

15. Sukardimana, Hadi Poerwono., Sofia Mubarika., Sismindari. Penapisan Senyawa Antikanker dari Tanaman Obat Indonesia dengan Molekul Target Enzim DNA Topoisomerase, Laporan Penelitian Domestic Collaborative Research Grant (2000).

16. Sukardiman, Noor Cholies Zaini, Sismindari ,2006. Induksi Apoptosis dan Peningkatan Ekspresi p53, Bax serta Aktivasi Enzim Caspase Sel Kanker Payudara Manusia oleh Pinostrobin dari Kaempferia pandurata Roxb, Laporan Penelitian Hibah Bersaing Tahun I(2006), Lembaga Penelitian Universitas Airlangga.

17. S.Wittmann , P.Bali, S.Donapaty, R.Nimmanapalli, F.Guo, H.Yamaguchi, M.Huang, R.Jove, H.G.Wang and K.Bhalla, Flavopiridol down-regulates antiapoptotic proteins and sensitizes human breast cancer cells to epothilone B-induced apoptosis. Cancer Res, 63(2003), 93−99.

18. C.Whibley, p53 polymorphisms cancer implications, Nature RevIew Cancer, 9 (2009), 95-107.



Atmodjo et al. 126


Surabaya Indonesia

The Ability of Sweet Orange Peel’s(Citrus sinensis) Pectin as

Biosorbent Of Heavy Metal Chrome (VI)

Natalia Widya Yuda Suryaningtyas, L. Indah M. Yulianti, and P. Kianto Atmodjo,

Biotechnology Faculty, Atma Jaya Yogyakarta University Jl. Babarsari No.44, Yogyakarta,

Corresponding Author’s Email: [email protected]

Abstract.Use In treating waste that contain heavy metal chromecheaply and safely is a challengefor biotechnology environment. A methods that is thriving is biosorption, using pectin of sweet orange peel as a biosorbent. Pectin contains of group such as carboxylate that can interact with Cr (VI). The aim of this research is to know the ability mass of pectin of sweet orange peel and the time estimate contact to metal‘s decrease chrome. The highest

absorption reaches 51% optimally using pectin 1 gram with optimally contact time estimate 2 hours.

INTRODUCTION Chromium, with its great economic importance in industrial use is one of the major metal pollutants and, in

the last few decades, the amount of chromium in aquatic and terrestrial ecosystems has increased as a consequence of human activities. The discharge of effluents by a variety of industries such as leather tanning, textile dyeing, electroplating, pigment manufacturing, refineries, wood preservative treatment, and steel fabrication constitutes one of the major causes of water pollution by chromium compounds, gaining great significance to detoxify them 1.

Chromium is hazardous pollutant for ecosystem, especially in hexavalent form is very toxic, has high solubility and mobility, teratogenicity, mutagenicity and carcinogenicity to living system related with its oxiding power 2,3. Chromium ion is most carcinogenic in the form of CrO2

-4, which enters the body cell by sulfate uptake pathway and is ultimately reduced to Cr(III) through a Cr (IV)-glutathione intermediate species. The hexavalent latter complex then binds with the DNA to produce a kinetically inert and potentially damaging lesion and can cause abnormal phenotype 4,5.

To reduce Chromium from environmental have been done by various technology and methods. However, these methods and technology have high operating costs and problems in the disposal of the residual metal sludges. Remediation of soil and water contaminated of heavy metals was important caused soil and water as medium for food producing 1,4. Different metal removal methods (ion exchange, chemical reduction and precipitation, reverse osmosis, phytoremediation, bioremediation, etc.) have been tested for detoxification of Chromium-laden wastewaters in the recent years. Conventional methods for heavy metals remediation consist of physical and chemical process but these applications were costly and less effective. However, these methods consume high amounts of energy and large quantities of chemical reagents which are not economically feasible. Furthermore, the resultant metal-containing chemical sludge can be a potential source of metal pollution 3,5.

On the other hand, biological methods such as biosorption techniques could use inexpensive sorbent materials as a feasible alternative for Chromium removal that features high efficiency, low operating costs with no adverse effects on the environment 3,4 . The remediation technologies is the using living organisms such as microorganisms, plant or organic material, because they have ability to reduce Cr(VI) into non toxic form, Cr(III). Beside, The living material can adsorb chromium, so that it is called bioadsorption or biosorbtion 6. For example maple of sewdust, brown seaweed biomass 7, and rice husks 8 . In recent years, several types of agricultural wastes such as sugarcane bagasse, grainless stalk of corn, etc. have been applied with the aim of removing Chromium from wastewaters 9.

Agro-waste materials can be used as biosorbents of heavy metals in aqueous solution 10,11. However, it is necessary to further study the contribution of agro-waste materials components (i.e. hemicelluloses, cellulose, pectin and lignin) to the heavy metal ions removal from aqueous solution to better understand the biosorption mechanism, and also based on the biosorbents main components, to predict their potential to remove heavy metals8,12. Orange or Citrus peel is easy to obtained as agro, domestic and food industry waste or litter 13.14 . This study investigates the removal of chromium by pectin from ―Pacitan Sweet Citrus‖ (Citrus sinensis) peels, based on time contact and weight pectin.



Atmodjo et al. 127


Surabaya Indonesia

MATERIAL AND METHODS

Location and Time Research

This research was done in Food Biotechnology Laboratory Atma Jaya Yogyakarta University and Laboratory of Chemistry, Water and Enviroment Comodity Test, Center for Leather, Rubber and Plastics Yogyakarta, on August untill December 2014.

Material and Instrumentation

The materials used in this study is Pacitan sweet orange peel. Chemicals used for the extraction of pectin is 95% ethanol, hydrochloric acid (HCl), water distillate and silver nitrate (AgNO3) to test the chloride ion as well as chemicals for analysis ie NaOH, ethanol, NaCl, HCl and phenol red, potassium dichromate (K2Cr2O7).

The instruments used in this study are Erlenmeyer, pans, measuring cups, thermometer, pH meter, filter paper, filter, propipet, pipette measuring, stirring, stove, oven, knives, cutting boards, basins, test tubes, blender, paper labels, spectrophotometer UV-Vis spectrophotometer (Shimadzu UV-1601PC), shaker (VAN-200) centrifuge (Hettich EBA 3S), the balance of electricity (AND GF-200) and some other tools are required to process and analysis.

Preparation Pectin

According to 14, 15, Peel of citrus fruits weighed as much as 100 g, plus 1 liter of water and crushed using a blender for 1.5 minutes so that it becomes a slurry, then datur pH to 1.5 using hydrochloric acid 0.1 N / then heated at 80C for 40 minutes, while stirring. Then this slurry was filtrated and cooled. The filtrate was heated back up to stay half, then cooled. Furthermore, 96% alcohol plus filtrate which has been acidified, with a ratio of 1: 1.5. Left for 8-12 hours to complete separation and precipitation. The precipitate/sediment is filtered to separate the alcohol and water. The precipitate was washed with alcohol 95% to remove the acid (the test AgNO3 was done foe chek it). If chlorides still there will be formed a white precipitate (AgCl), and the washing repeated 2-3 times until free of chloride. Then, it was dried at a temperature of 40 degrees Celsius, then crushed and sieved using a sieve of 60 mesh. It was was stored as pectin powder.

Measurements of methoxyl pectin sweet orange

Pectin powder much as 0.5 gram of wetted 5 ml of 96% ethanol and dissolved in 100 ml of distilled water in which there is free carbonate 1 gram NaCl. Solution was added 25 ml of 0.25 N NaOH, shaken and left for 30 minutes at room temperature in a closed state. Solution was added 25 ml of HCl 0.25 N, ditambakan indicator phenol and then titrated using a few drops of 0.1 N NaOH until the color changes to red yellow.

Methoxyl concentration (%) = ((mL x N NaOH )x 31 x 100 ) : Sample weight (mg)

Preparation of Chromium Stock Solution

Potassium dichromate (K2Cr2O7) P.A. 564 mg dissolved 0.5 liter of distilled water to obtain a solution of 400 ppm Cr concentration as a stock solution. Then this solution will be diluted to obtain the required concentration for the study is 10 ppm.

Testing of Biosorption Ability 3,10

Pectin powder of 0.5 g, 1 g and 1.5 g put in 50 ml of 10 ppm Cr. The solution was homogenized by shaker with 2000 rpm for 1, 2 and 3 hours. Then the solution was centrifuged at 3000 rpm for 15 minutes to separate the supernatant with sorbent. The Cr content of Supernatant was measured using UV-vis spectrophotometer at a wavelength of 543 nm. Absorbance values obtained are within the range of values a calibration curve of standard solution so that the concentration of metal in solution can be calculated using the equation of the regression line. Pectin absorption of the metal calculated as pectin persentation absorption capacity of the metal is determined by equation8:

% Absorption: [(Lb-Ls) / Lb] x 100% Lb = concentration of the blank solution (mg / L);



Atmodjo et al. 128


Surabaya Indonesia

Ls = concentration after equilibirium (mg / L)

Data Analysis

Data were analyzed using ANOVA variation. When there is a real difference it will be Duncan's Multiple Range Test (DMRT) with a 95% confidence level and to see the weight of pectin and long relationship remediation to decrease the effectiveness of chromium (Cr) correlation and regression analysis


Pectin absorption of the metal Chromium

Pectin is one compound found in plant cell walls mainland. Pectin is a polymer of D-galacturonic acid that linked by 1,4 glycosidic bonds and more contained in the middle lamella cell wall plant 9. Pectin makes cells hard and strong. Beside, Pectin is used in the food industry, pharmaceuticals and cosmetics. At these industries pectin used primarily as a gelling10. However, the structure of the pectin component contain many active groups, the pectin can also be used as a source biosorbent 13,14 . Pectin absorption test results are in the table below.

Table 1. Pectin absorption of the metal Chromium

Time of contact (hour)

pectin weight (gram)

0,5 1 1,5

1 30 % 33 % 36 %

2 35 % 44 % 51 %

3 38 % 49 % 43 %

In table 1 above shows the increase absorption porsentase as the length of time and the addition of pectin

contact between the sorbent and the metal ion. The absorption occurs at a contact time of 2 hours and 1.5 grams of pectin giving the percentage value of 51%. The longer the metal is contacted with the cell surface of pectin as a sorbent, the more the cell surface of pectin as a sorbent, the more the surface of cells that become activated and do the binding to the metal. Metal binding by pectin because of the active group which has a free electron pair of the metal cations so that the metal cations can be attracted and bind to form a complex and metal pectin 1.

After the metal absorption capability by pectin as a sorbent be decreased to the optimal limit. At optimal conditions indicate there is no longer the surface of cells that can become activated and form a bond with metal, because the cell surface has been at saturation point. Pectin will bind Cr (VI) optimally to the longer it takes. Metal can be bound by pectin (sorbent) specifically. Adsorption occurs on the surface of the sorbent has not reached saturation point3. Each sorbent has the ability to bind metal ions to the maximum, but once the maximum limit has been exceeded, then the sorbent surface becomes too saturated to hold adsorb metal ions.

The effect of weight variation and contact time of Pectin on Reducing Chromium

Knowing the amount of pectin and duration of contact against the absorption of pectin to the metal becomes very important because it provides information on the wastewater treatment system. Therefore, testing is done until the time of contact for 3 hours due to optimal absorption time is 2 hours earlier study 13,14,15. The test results with the treatment of weight variation and contact time are shown in Table 2.

Table 2. Levels of Cr (VI) to the addition of pectin and long treatment time. Duartion Contact (hour)

Pektin (gram) Mean 0,5 1 1,5

0 11,87a 11,87a 11,87a 11,87A 1 8,32b 7,93bc 7,63bc 7,95B 2 7,70bc 6,61d 5,73e 6,68C 3 7,38c 5,99e 6,70d 6,68C

Mean 8,82X 8,10Y 7,96Y



Atmodjo et al. 129


Surabaya Indonesia

Note: each number followed by the same letters in the same column showed no significant difference.

Based on the average results obtained that the longer the contact time decreasing the concentration of Cr (VI) and the more pectin, getting down the metal ion concentration of Cr (VI) as well. This is due to the active groups on the pectin is a carboxyl group which is considered as the active component in biosorption process that would attract and bind metal forming complex compounds with metal and insoluble in water3,5. Thus the addition 1 gram of pectin powder with a contact time 3 hours or the addition of 1.5 grams of pectin with long contact time of 2 hours into a 50 ml sample solution can absorb heavy metals is higher than other treatments. This can be taken into consideration in the process of sewage treatment. When a high waste volume, selectable alternative to the use of pectin in large quantities (1.5 grams) with a shorter time (2 hours), whereas when the volume of waste less, can be selected by using the alternative provision of pectin in small amounts (1 gram) with time longer (3 hours).

The duration time contact of 2 hours and pectin weight of 1 gram, is considered the most optimal, because it is able to reduce levels of heavy metals Cr (VI). Based on the correlation and regression analysis with SPSS, that weight variation and hours together have a relationship to the concentration of (R) 0.868 and a significant difference between the variation of the weight and the length of time remediation with Cr concentration of 0.753. The results are consistent with studies of reference 13,14,15, sweet citrus pectin has a higher absorption is 51% of compared to the reference using banana peel, rice hasks, and seaweed as biosorbent which has a 45% absorption 5,6,7,8. These results also support the idea that Citrus peels or agrowaste (agriculture litter) as the most promising biosorbent due to high metal uptake in conjunction with physical stability . Uptake was rapid with equilibrium reached after 30–80 min depending on the particle size (0.18–0.9 mm). The sorption kinetics followed a second-order model. Sorption equilibrium isotherms could be described by the Langmuir model in some cases, whereas in others an S-shaped isotherm was observed, that did not follow the Langmuir isotherm model. The uptake of heavy metal by biosorbents was dependent on chemical structure of pectin and increased with biosorbent concentration10.

Mechanisme Heavy Metal Biosorpsi By Pectin

Heavy metals biosorption processes with adsorbent is a complex biological process and the mechanism can vary depending on the source of raw material3. When based on the metabolism of the cell, then the mechanism can be divided into adsorption depending on the metabolism of cells and which are not depending on cell metabolism. When raw materials biosorpstion is from agricultural waste8, the possible mechanism is that not depending on cell metabolism. Mechanism biosorption on these materials generally based on the chemical interaction between ion physics metals with functional groups that exist in the cell surface. Such interactions can be electrostatic interaction, ion exchange or chelating complex formation2. .Meanwhile, the process of biosorption can be separted into two main processes i.e. ion adsorption on the surface of cells, and bioaccumulation cell adsorbent (4).

CONCLUSION The ability Pacitan sweet orange peel pectin as biosorbent heavy metal Cr (VI) from aqueous solution is

evident, rapid and favorable.

REFERENCES 1. Kurniasari, L., Riwayanti, I., Suwardiyono. Pektin Sebagai Alternatif Bahan Baku Biosorben

Logam Berat. Momentum 8 (1) : 1-5. 2012 2. Ahalya, N., Ramachandra, T. V. and Kanamadi, R. D. Biosorption of Heavy Metal. Research Journal

of Chemical and Environment 7 (4), 71-79. 2003 3. Garcia-Reyes R. B and Rangel-Mendez†,* J. R. Contribution of agro-waste material main

components (hemicelluloses, cellulose, and lignin) to the removal of chromium (III) from aqueous solution. Journal of Chemical Technology and Biotechnology . 84(10):1533–1538. 2009

4. Bellú S, Luis S, González J, and García S. Thermodynamic and Dynamic of Chromium Biosorption by Pectic and Lignocellulocic Biowastes. Journal of Water Resource and Protection, 2:888-897. 2010

5. Ashraf, MA., Maah, MJ., and Yusoff, I., Study of Banana peel (Musa sapientum) as a Cationic Biosorben. Journal of Agriculture & Environment Science. 8(1): 7-17. 2010

6. Sudiarta, I. W. and Suarya, P., Biosorption and Desorption chromium Ion on Eucheuma spinosum Biosorbent, Young Lecturer Research Report. DP2M DIKTI. 2007

http://onlinelibrary.wiley.com/doi/10.1002/jctb.v84:10/issuetoc



Atmodjo et al. 130


Surabaya Indonesia

7. Yun, Y-S., Park, D., Park, J. M., and Volesky, B., Biosorption of Trivalent Chromium on The Brown Seaweed Biomass, Environ. Sci. Technol., 35 : 4353-4358. 2001

8. Purwaningsih, D., Adsorpsi Multi Logam Ag(I), Pb(II), Cr(III), Cu(II), dan Ni (II) Pada Hibrida Etilendiamino-Silika Dari Abu Sekam Padi. Jurnal Penelitian Saintek. 14. 59-76. 2009

9. Kupchick, L. A., Kartel, N.T., Bogdanov, E.S., Begdanova, O. V., and Kupchick, M. P.. Chemical Modification of Pectin to Improve It’s sorption properties. Russian Journal of Apllied chemistry. 79 (3). 457. 2005

10. Retno Wulandari, The Use of Na-pectin as Cadmium Adsorbent in aqoues solution. Thesis S2 Gadjah Mada. University. 45 p. 2011

11. Wilats, J. William. G.T., Paul K. and Jorn D.M.. Pectin: New Insights Into and Old Polymer Are Starting To Gel. Journal of Trends in Food Science & Technology. 97-104. 2006

12. Silke S, and Santosh B. Pl. Pectin-rich fruit wastes as biosorbents for heavy metal removal: Equilibrium and kinetics. Bioresource Technology. 99 (6): 1896–1903, 2008

13. Wong, W.W., Abbas F.M.A., Liong, M.T., Azhar, M.E. Modification of Durian Rind Pectin for Improving Biosorbent Ability. International Food Research Journal 15 (3), 363-365.2008.

14. Wong. W W, Abbas F.M. AlKarkhi, and Azhar M. E, Comparing biosorbent ability of modified citrus and durian rind pectin. Carbohydrate Polymers. 79(3):584–589. 2010

15. Fitriani, Vina. Pectin Extraction and Characterization of Several Types of Skin Orange Lemon. Scription. Agriculture Technology Faculty. Bogor Agriculture Institute. Bogor. 2003

http://www.sciencedirect.com/science/article/pii/S0960852407003537


http://www.sciencedirect.com/science/journal/09608524

http://www.sciencedirect.com/science/journal/09608524/99/6




http://www.sciencedirect.com/science/journal/01448617

http://www.sciencedirect.com/science/journal/01448617/79/3



131


Surabaya Indonesia

ORAL PAPER OF ENVIRONMENTAL AND GREEN CHEMISTRY (OEGC)



Shah et al. 132


Surabaya Indonesia

Investigation of Foods Containing Borax in Surabaya with Extract of

Turmeric

Dwi Yulian Fahruddin Shah1, M. Al Rizqi Dharma Fauzi1* , Andre

Pratama1, Nufida Dwi A.2, Hery Suwito1

1Department of Chemistry, Universitas Airlangga 2Department of Physic, Universitas Airlangga

*Corresponding author’s Email: [email protected]

Abstract. Indonesia, a developing country with its beauty and harmony, is threaten by borax contained in many foods. It is proven by many news and articles discussing about borax in foods and its effect of our body. Borax, also known as ―bleng,‖ has a wide variety of uses such as: detergents and cosmetics. However, if it is consumed in a significant dose, it will cause severe symptoms or death. We invented a simple detector of borax from the extract of turmeric which is combined with chemicals which is not dangerous to environment in order to achieve the green chemistry needs. This detector will be a reachable and portable detector which is very useful for people in order to prevent their self from the fatal effect of borax. We also investigated many foods in several area of Surabaya to prove its efficiency and accuracy in detecting borax. The backwards philosophy herein is that the continuous studying of the detection of borax in simple

INTRODUCTION Food is one of our primary necessity. It supplies energy to human body so we can do our activity well. There

are many kinds of food and it contain many kind of nutrient too. They exist in many additional flavor, so people will more interested to taste and more durable is thay want to keep the food for a while. But in this era, many people take advantageous of this condition. Some of them add a dangerous matter on it to take a result like what they want. It will threaten human life. One of the most familiar matter is sodium tetraborate, or well-known as borax or bleng.

Borax is used as a food preservation. But actually it is used to disinfection. Usually it is used as a material for detergent, wood preservation. Borax that is containing poisonous matter that threat human health. If it already consumed by human, it will be absorbed by intestines and pile up in liver, brain, intestines, kidney. At last it can be a cancer for human who consume it continuously.

There are many effects that are caused by borax. If it is consumed in low doses, it can decrease eat passion, digestion disturbance, inhalation disturbance, and nerves system disturbance. If it is consumed in high doses, it will make us queasy, diarrhea, headache, kidney disorder, cancer, and even death. Death is happened when an adult consumes 15-25 gr borax and a child consume 5-6 gr borax.

Borax can be detected by turmeric that is already extracted and processed through several treatments. In past researches, we have already make test kit of borax using the extract of turmeric (Fauzi et al., 2013). Because of the spreading of borax in many area in Surabaya, we want to examine and compare the spreading of food that is containing borax in every area in Surabaya. The purpose of this paper is to investigate the real situation of foods containing borax in order to educate people about the dangerous of borax. It is a big hope for the society of Surabaya can be more careful to consume foods and beverages in their daily life.

Turmeric and Its Main Compound

Turmeric is very useful natural source for detecting borax in a simple way. Turmeric is mainly composed of curcumin with molecular formula C₁₂H₂O₆. Curcumin is a kind of natural dye in turmeric that gives a yellow appearance of it when turmeric is cut or scraped. When curcumin is reacted with borax, it will produce a reddish and darker product. This change of color is what will become the standard in investigating foods that are thought to contain borax should not be used in food.



Shah et al. 133


Surabaya Indonesia

Figure 1. A piece of turmeric

Sodium Tetraborate (Borax)

This material is widely used as an anti-fungal, wood preservative and antiseptic in cosmetics (Svehla, G., 1996). Borax is a chemical compound with the formula Na2B4O7.10H2O. It is existed in a white crystalline, odorless and stable solid at normal temperature and pressure (Shah, 2005).

From various studies that have been conducted data showed that boric acid compound is used in some types of food among them is tofu, noodles and meatballs in order to make a good texture of them. Effects of borax that given to food can improve the structure and texture of food. For example when borax given to meatballs and tofu will make meatballs and tofu are very resilient and durable, while the crackers containing borax if fried inflate and padded and has a nice texture and crisp. Worse, the food has been given borax with no or unspoiled, it is difficult to distinguish if only the senses, but have to do a special test of borax in laboratory.

(a) (b)

FIGURE 2. (a) Chemical Structure of Borax and (b) the appearance of Borax



Shah et al. 134


Surabaya Indonesia

MATERIAL AND METHODS Foods used as the target are tofu, meatball, crackers, and noodles. All targets were brought from north, east,

west, south, and central representative area of Surabaya. Borax detector used in this paper was supplied by UD Tundrea. Research about detection boraks of the several foods which spread out in Surabaya area is use qualitative research methods which the method of this study is a research method that produces descriptive data about the words spoken or written, and behavior that can be observed from those studied (Taylor and Bogdan, 1984). The reason of use qualitative research methods in studying this problem are because the author wants to understand in detail and deep with with the dynamics of environment and social life as well to describe and interpret the social dynamics of the whole in accordance with reality (what actually) without simplifying them intoa many variable.

Preparation of Sample

Foods brought in five representative areas in Surabaya were cut into small pieces. A piece of it was located on a surface of a white plate. The sample was given by a single drop of borax detector to determine the change of color. Replication was done fifteen times for each samples.

FIGURE 3. A cracker and its pieces

According to Sturge‘s Rule:

where k is the number of class intervals and n is the number of observations in the set, if there are five areas, then fifteen samples are required in order to fulfill the minimum requirement of descriptive qualitative statistics.

RESULTS AND DISCUSSION Surabaya is safe enough from the illegal circulation of borax. It is shown by TABLE 1 that from five kind of

targets in separated area, less than three targets are identified containing borax.



Shah et al. 135


Surabaya Indonesia

Table 1. Results of Investigation Area Tofu Meatball Cracker Noodle

Central

1. Negative 2. Negative 3. Negative 4. Negative 5. Negative 6. Negative 7. Negative 8. Negative 9. Negative 10. Negative 11. Negative 12. Negative 13. Negative 14. Negative 15. Negative

1. Positive 2. Negative 3. Negative 4. Negative 5. Negative 6. Negative 7. Negative 8. Negative 9. Negative 10. Negative 11. Negative 12. Negative 13. Negative 14. Negative 15. Negative

1. Negative 2. Negative 3. Negative 4. Positive 5. Negative 6. Negative 7. Negative 8. Negative 9. Negative 10. Negative 11. Negative 12. Negative 13. Negative 14. Negative 15. Negative


North


1. Positive 2. Negative 3. Negative 4. Negative 5. Positive 6. Negative 7. Negative 8. Negative 9. Negative 10. Negative 11. Negative 12. Negative 13. Negative 14. Negative 15. Negative

1. Positive 2. Positive 3. Positive 4. Negative 5. Negative 6. Negative 7. Negative 8. Negative 9. Negative 10. Negative 11. Negative 12. Negative 13. Negative 14. Negative 15. Negative


West


1. Negative 2. Negative 3. Positive 4. Negative 5. Negative 6. Negative 7. Negative 8. Negative 9. Negative 10. Negative 11. Negative 12. Negative 13. Negative 14. Negative 15. Negative

1. Positive 2. Negative 3. Negative 4. Negative 5. Negative 6. Negative 7. Negative 8. Negative 9. Negative 10. Negative 11. Negative 12. Negative 13. Negative 14. Negative 15. Negative


South

1. Negative 2. Negative 3. Negative 4. Negative 5. Negative 6. Negative 7. Negative 8. Negative 9. Negative 10. Negative 11. Negative 12. Negative

1. Negative 2. Negative 3. Negative 4. Positive 5. Positive 6. Negative 7. Negative 8. Negative 9. Negative 10. Negative 11. Negative 12. Negative

1. Negative 2. Negative 3. Negative 4. Positive 5. Positive 6. Negative 7. Negative 8. Negative 9. Negative 10. Negative 11. Negative 12. Negative

1. Negative 2. Negative 3. Negative 4. Negative 5. Negative 6. Negative 7. Negative 8. Negative 9. Negative 10. Negative 11. Negative 12. Negative



Shah et al. 136


Surabaya Indonesia

13. Negative 14. Negative 15. Negative




East





Surabaya is safe enough from the illegal circulation of borax. It is shown by Table 1 that from five kind of

targets in separated area, less than three targets are identified containing borax. This descriptive qualitative data showed that in statistics, the safest area where borax did not circulated was the east area of Surabaya. We need to be more careful to choose foods in the rest of areas because there are foods containing borax. In conclusion, Tofu is the safest kind of foods that is not containing borax.

(a) (b)

Figure 4. (a) Investigated Meatballs and (b) Tofu



Shah et al. 137


Surabaya Indonesia

CONCLUSION The safest food from the dangerous of Borax is Tofu and the safest area in Surabaya to prevent us from it is

the East Area. Borax is not the only dangerous additive substance used by the people when making their food. There are formalin, synthetic dyes, and so on. This paper is only shown the investigation of foods containing borax and hopefully for the next investigation, those another additive substances can be also investigated.

ACKNOWLEDGEMENT We thanked to UD Tundrea for their simple detector of borax and Universitas Airlangga that facilitated

literatures to take the statistical theories.

REFERENCES 1. Fauzi, Muhammad, 2015, Simple Detector of Borax: A Big Step to Reduce Poverty in Indonesia,

Proceeding on the 3rd Asia-Pacific Student Forum. 2. Hidayat, Dandik, 2011, Uji kandungan boraks, Universitas Negeri Jember. 3. Svehla, G., 1996, Vogel‘s qualitative inorganic analysis, Prentice Hall 4. Triatama, Joni, 2014, Identifikasi kandunagn boraks pada keripik yang dijual dipasar bebas kota kuala

kapuas kalimantan tengah. Universitas Muhammadiyah Palangkaraya



138


Surabaya Indonesia

ORAL PAPER OF MICROBIAL BIOCHEMISTRY AND MOLECULAR BIOLOGY

(OMBM)



Puspitasari et al. 139


Surabaya Indonesia

Liquid Organic Fertilizer used Microbial from Local Inoculant

Rita Tri Puspitasari*, Elfarisna,Yati Suryati, and Nosa T. Pradana

Study Program Agrotechnology, Faculty of Agriculture University of Muhammadiyah Jakarta

*Corresponding Author’s Email: [email protected]

Abstract. Wastewater of rice can be used as organic fertilizer. The research that have been done to Dendrobium, Brassica rapa, Lactuca sativa, Allium fistulosum, Amaranthus, Cucumis sativus resulted that it is not significantly different with inorganic fertilizer. The problem arising is the smell from fermented wastewater of rice. In order to eliminate the odor, the option was add EM4, which has been tested on Phaleonopsis and Glycine max. EM4 is a non-local products, meanwhile Indonesia also has a lot of local diversity of microorganisms. As a source of inoculants, is used wastewater of rice, kombucha, tapai yeast, and yogurt. From observations, it is obtained that there was three spesies of bacteria from wastewater of rice, yeast and yogurt. As the results of this research is three spesies of bacteria and two spesies of yeast that can live either in the wastewater of rice and does not stinks. The results of morphological identification and SSU rRNA sequencing is three kind of bacteria which are Burkholderia metallica, Burkholderia seminalis, and Gluconacetobacter saccharivorans. There are two kinds of yeast from the sequence analysis of ITS rDNA area: Trichosporonasahii, and Pichia kudriavzevii. Researchs which have been done on plants Polianthes tuberosa, and Vigna radiata showed not significant result with inorganic fertilizer.

INTRODUCTION Application of a balanced organic fertilizer (suitable) have been known in improving the structure, nutrients,

and biology of soil. Besides, it can improve efficiency of fertilizers use[1]. Products become more environmental friendly and to some extent can reduce the negative impacts of chemicals that are harmful to humans and the environment [2]. Organic fertilizer is good to be used for long term due to its feature in loosen up the soil and improve soil ability to store water, so that the soil fertility maintained. While the synthetic chemical fertilizers even though has rapid reaction effect, for long term use it will harden the soil and reduce soil fertility [3].

Increasing of population lead to a quite alarming environmental problems due to waste disposal, ranging from industrial waste (manufactory) to domestic waste. Domestic waste became the 60-80% waste problem in big cities. Evidence from several studies in Faculty of Agriculture UMJ indicate that the waste water of rice as one of domestic waste which have been stored for 2 weeks in some types of plants can substitute chemical fertilizers/inorganic. The plants include: Dendrobium, Brassica rapa, Lactuca sativa, Allium fistulosum, Amaranthus, Cucumis sativus[4][5][6][7][8][9]. But the odor becomes an obstacle after wastewater of rice stored for 2 weeks and reached pH 4.5 [4]. Suryati [10][11] and Elfarisna et al, [12] in their research using commodity Phalaenopsis sp. and Glycine max. EM4 can eliminate odor factors of fermented wastewater of rice and reduce the fermentation time.

Local Microorganisms is the microorganism that is used as a starter in the production of solid and liquid organic fertilizer. Laboratory of Agriculture Faculty UMJ have found five microorganisms, that is three bacteries and two yeasts. Three bacteria from wastewater of rice, kombucha, and tape yeast. Two yeasts from wastewater of rice in previous studies.

The use of waste water of rice to mung bean has not been done. Mung bean is one commodity of beans which is widely consumed by people. In 2013, an area of 82 hectares with productivity reaches 24.98 kw/ha [13]. Indonesia imported mung bean to meet domestic needs. Import of mung bean also increased quite drastically in March 2014 compared to the previous month. In February 2014, imports of mung bean was recorded at 6.27 thousand tons. Import mung bean show rapid increase to 13.96 thousand tons in March 2014. Total imports of mung bean during the first three months of 2014 recorded 23.45 thousand tons [14].

Mung bean (Vigna radiata L.) belong to the family of legumes (Fabaceae) has benefit as high nabati protein foodstuff. Along with the increasing population growth and the increasingly diversity of products made from mung bean, the demand of mung bean will continue to increase. Based on data from Ministry of Aqriculture, Directorate of Assorted Nuts and Tubers, the development of mung bean harvested area in 2008-2012 has decreased by an average of 278 627 ha per year (0.10%) and the production increase by an average of 311 658 tons per year (1.72%). Therefore, intensification is necessary for this case.






Surabaya Indonesia

One of the ornamental flowers which can be tested and used as a business line is tuberose (Polianthes tuberose L.) [15]. Tuberose plants has advantages in color, fragrance, and the unique flower thread, with a production period of two to three years.

One of tuberose farmers in Salabintana, Sukabumi - West Java, have seized this opportunity since the 80s. According to the farmer, grow tuberose is relatively easy, not need much fund, and the planting does not require complicated installation[15]. Tuberose planting area in 2013 was 1,081,325 m2 with production centers in East Java (603 130 m2), West Java (191 707 m2), Central Java (174 616 m2), and Banten (49 630 m2) [17].

The use of fertilizers on tuberose planting relatively large and quite expensive in order to achieve the desired quality. Addition of inoculant on waste water of rice is expected to meet the needs of fertilizer with the desired qualifications in the market.

EXPERIMENTAL METHODS The research conducted from May 2014 until June 2015 at the Experiment farm and Laboratory of

Agriculture Faculty of University Muhammadiyah Jakarta, as well as in Pesantren Indonesia Foundation (YPI) Al-Zaytun, Indramayu-Indonesia. Nutrient analysis conducted in the Soil Research Institute, Bogor. The morphology and DNA isolates conducted at the Center of Excellence Indigenous Biological Resources - Genome Studies (CoE IBR-GS), University of Indonesia.

Bacterial strains maintained and nurtured in media Mann Ragosa Sharp Agar (MRSA) with wastewater of rice solvent, and the yeast also maintained on media Patato Dextrose Agar (PDA) wastewater of rice solvent. All of that were incubated at room temperature + 27o C. Application of the strains in the two types of plants, that is Mung bean and Tuberose.

Research on mung bean, the used design was a Randomized Complete Block Design (RCBD), with eleven 11 treatment wastewater of rice which containing yeasts and bacteria. Each treatment repeated four times as follows:

M0: Control (urea 0.25g, TSP 0.5g and KCL 0.25g) M1: 2 Yeast (L, A2) + 1 Bacteria (A1) M2: 1 Yeast (L) + 2 Bacteria (A1, R) M3: 1 Yeast (A2)+ 2 Bacteria (A1, K) M4: 2 Yeast (L, A2)+ 1 Bacteria (R) M5: 2 Yeast (L, A2) + 1 bacteria (K) M6: 1 Yeast(L) + 2 Bacteria (R, K) M7: 1 Yeast (A2)+ 2 Bacteria (L, R) M8: 2 Yeast (L, A2) + 2 Bacteria (L,K) M9: 1 Yeast (L) + 3 Bacteria (L,K, R) M10: 2 Yeast (L, A2) + 3 Bacteria (L,K, R) The observed parameters: Plant height, Branches amount, blooming time, pod amount, The percentage of

full pods, seed weight per plant, and weight of 100 seeds. Research on Polianthes tuberosa L., experiment design used was randomized Complete Block Design

(RCBD) with five treatments and repeated five times, as follows: P0: control / 100% inorganic fertilizers (NPK Grower 5g, Saprodap 5g) P1: inoculum I (3 bacteria + 2 yeast ) + 50% control P2: inoculum I (3 bacteria + 2 yeast ) + 25% control P3: inoculum II (2 bacteria + 2 yeast ) + 50% control P4: inoculum II (2 bacteria + 2 yeast ) + 25% control The observed parameters: stalk lenght, stalk diameter, florets amount per panicle, Harvest Time, yields. Note : L, A1, A2 source isolates from wastewater of rice K source isolates from Kombucha, R source isolates from tape yeast Inoculant I ( all microbes), Inoculant II ( non R)

RESULTS AND DISCUSSION Based on the research, insulation obtain seven yeast, from wastewater of rice (4 species), tapai yeast (1

species), and kombucha (1 species). The research resulted selected three types of bacteria, and 2 types of yeast which could live well in the wastewater of rice and does not causing smell in 2014, but has not been well identified.

In 2015, based on morphology of microorganism and electrophoresis, it is known that isolated inoculants consists of: two yeasts (A2 and L) and 3 bacteria (A1, K, and R). They are shown from the results of electrophoresis (Figure 1) and microbial morphology (Figure 2, 3, 4, 5 and 6) as follows:





Surabaya Indonesia

Figure 5. Results of electrophoresis to 5 Microbes

Results of the DNA Identification

DNA isolation and amplification (PCR) for area Internal transcribed spacer (ITS) rDNA was performed on two yeast isolates (isolates code REYN-A2 and REYN-L). Sequences of area ITS rDNA for isolates is performed using forward primer ITS1.

DNA isolation and area amplification (PCR) of the 16S ribosomal RNA gene or Small Subunit (SSU) rRNA gene has been conducted on three bacterial isolates (code REYN-A1, REYN-K, and REYN-R). Execution of PCR area of 16S rRNA or SSU rRNA gene from three isolates using primers 27F (Forward) and primer 1494R (Reverse).

Sequencing area SSU rRNA gene from three isolates using primers 1494R (Reverse), the resulted sequence data edited manually and sent to the DNA Database Genbank through BLAST homology search program, the website http: //www.ncbi.nlm.nih. gov / Blast.cgi, in order to search for sequence similarity. The closest species isolates can be seen in Table 1.

Sequence analysis of area ITS rDNA from the yeast isolates (isolates code REYN-L) placed these isolates into the Fungi (Kingdom); Ascomycota (Phylum); Saccharomycetes (Class); Saccharomycetales (Order); Pichiaceae (Family); Pichia (Genus); Pichia kudriavzevii (Species). Pichia kudriavzevii (synonym Issatchenkia orientals Kudryavtsev). Issatchenkia orientalis once has been isolated from the berries juice, fruit flies, fermented foods, and sea water [18].

Table 1 Name of identified species

No Code

Isolate

Primers used in

sequencing Species Name

Source of Isolates

1 REYN-L ITS1(F) Pichia kudriavzevii (89%) Wastewater

of rice

2 REYN-A2 ITS1(F) Trichosporon asahii (98%) Wastewater

of rice

3 REYN-A1 1494R Burkholderia metallica (98%) Wastewater

of rice

4 REYN-R 1494R Burkholderia seminalis (93%) Tape Yeast

5 REYN-K 1494R Gluconacetobacter saccharivorans Kombucha





Surabaya Indonesia

(93%)

Description: Code R = Rita Code Y = Yati Code E = Elfarisna Code N = Nosa A1 / K / R / L / A2 = Coat types Sequence analysis of area ITS rDNA from the yeast isolates (isolates code REYN-A2) placed the isolates

into the Fungi (Kingdom); Basidiomycota (Phylum); Tremellomycetes (Class); Tremellales (Order); Tremellaceae (Family); Trichosporon (Genus); Trichosporonasahii (Species). Trichosporon asahii Akagi ex Sugita, Nishikawa & Shinoda. This species has been isolated from the tofu, shrimp, feces, blood [18].

Analysis of sequence area SSU rRNA gene from bacterial isolates REYN-A1 and REYN-R placed the isolates into the Bacteria (Kingdom), Proteobacteria (Phylum), Betaproteobacteria (Class), Burkholderiales (Order), Burkholderiaceae (Family), Burkholderia (Genus), Burkholderia metallica (code isolates REYN-A1 ) and Burkholderia seminalis (isolates code REYN-R) (Species). Burkholderia are Gram-negative, which is the species that can live at a temperature of 30 ° C, 37 ° C and 40 ° C [19].

Morphology of each species of microbes found:

Yeast

Figure 6. Pichiakudriavzevii

Figure 7. Trichosporonasahii

Bacteria

Figure 8. Burkholderiametallica





Surabaya Indonesia

Figure 9. Burkholderiaseminalis

Figure 10. Gluconacetobactersaccharivorans

Analysis of sequence area SSU rRNA gene of bacteria isolates Reyn-K placed the isolates into Bacteria

(Kingdom), Proteobacteria (Phylum), Alphaproteobacteria (Class), Rhodospirillales (Order), Acetobacteraceae (Family), Gluconacetobacter (Genus), Gluconacetobacter saccharivorans (Species). Gluconacetobacter is Gram-negative, which has the optimum temperature for growth 30 ° C and optimum pH 2.5 to 6[19].

Results of Wastewater of Rice Nutrient Analysis

As the result of nutrient analysis, in Table 2, it is visible that the degradation of 5 microbial has a higher or equal value to the 9 items, namely N (total), P2O5, K2O, Mg, Mn, Zn, Co, Ag, and Mo with 4 microbes and the addition of EM4. The result of the application to mung beans in Table 2, it is seen that the treatment with a combination of 5 microbes (M10) show the best results compare to other treatments though not nonsignificantly different from inorganic fertilizers, or a combination of other microbes. While the results of nutrients analysis, the wastewater of rice fermented for 2 weeks contains nutrients NH4 14.09 ppm, NO3 194.18 ppm, P 114.6 ppm, K 60 ppm, Ca 13.4 ppm, Mg 40.9 ppm, Fe 0.07 ppm, Al 0.27 ppm and Mn 0.23 ppm [4], which is in Table 2 have been converted to percentage (%). Results of nutrient analysis from fermentation for 2 weeks showed N, P, K and Mg were in higher amount.

The degree of acidity (pH) in the 2 weeks fermentation is higher than 4.5. It is caused by the yeast that has been run out of starch and sugar, so that it will utilize the acetate into alcohol in the longer fermentation process. In this process of 2 weeks fermentation, dispersive isolation test is done, which there isn‘t seen any microbes growing on the medium agar wastewater of rice. So that the high level of N, P, K and Mg caused by the microbial cells autolysis which degrade wastewater of rice[4]. On the four day fermentation process, microbia could still be isolated back on dispersive isolation.

Table 2. Results of Nutrient Analysis

The contents of Wastewater of Rice Fermentation

The 4 day fermentation process Natural fermentation

for 2 weeks [4] 5 microbes 4 microbes + EM4

pH:H2O (1:5) 3,5 3,5 3,4 4,5 Walkley& Black

------------ % ---------

C – Organik 0,54 0,55 0,16





Surabaya Indonesia

N (total) 0,12 0,09 0,12 0,21 - Organik 0,05 0,03 0,00 - - NH4 0,01 0,01 0,03 - - NO3 0,06 0,05 0,08 - Total HNO3 P2O5 0,05 0,04 0,03 0,12 K2O 0,04 0,04 0,01 0,06 Ca 0,01 0,01 0,03 0,01 Mg 0,02 0,01 0,01 0,04 --------- ppm --------- S 21 29 29 - Fe 32 48 15 0.07 Mn 8 1 1 0,23 Cu 1 1 1 - Zn 17 6 1 - B 6 1 1 - Al 5 2 11 0,77 Pb 0,9 1,0 1,0 - Cd Td Td Td - Co 2 1 1 - Ag 0,0 2 2 - Mo 0,9 0,5 0,5 - Hg 0,02 0,00 0,01 - La 0,0 0,0 0,0 - Ce 0,0 0,0 0,0 - Extra material 0,0 0,0 0,0

Applications on Plants

Mung bean

Treatment on mung bean with various combinations of microbes obtained result as shown in Table 3. The results from analysis of variance application of wastewater of rice inoculated with 50 ml MOL towards the height of mung bean plants showed no significant effect on plant 2 to 7 weeks after planting (WAP). At the age of 7 WAP, plant height was not significantly different between treatment with high plant shown in M0 ie as a control treatment with 0.25g urea, 0.5g TSP and 0.25g KCl (44.20 cm) and on the treatment M8 with wastewater of rice added with MOL from 2 yeasts and added with 2 bacteria from waste water of rice and kombucha.

The contents of wastewater of rice will initially be in cloudy color. Muddy wastewater color indicates that the outermost layer of rice also eroded. In the epidermis, one of the main elements needed in plants is element P. The element P increases the sugar phosphate which contributed in the dark phase reaction of photosynthesis, which will increase the rate of photosynthesis so generated fotosintat will be allocated to the growth of the plant height. Furthermore, element P contribute in the ATP forming, with more P is absorbed by the plant so that it will increase the ATP formation which is required by the crop as energy source in cell division and therefore affected in plant height increase [20].

The result from analysis of variance, the addition of wastewater of rice inoculated with MOL as much as 50 ml on mung bean plant branches amunt had no significant effect on plants aged 4 to 6 WAP. At the age of 7 WAP, branch amount of the plant was not significantly different between treatment with the branch amount that indicated in treatment M0 as control with urea 0.25g, TSP 0.5g and KCl 0.25g and to treatment M7 with wastewater of rice yeast provided with MOL 2 yeast added with 2 Bacteria from wastewater of rice, and tape yeast (9.50 item).

A lot of nitrogen nutrients contained in wastewater of rice is very useful among others to improving plant growth, producing chlorophyll, increasing protein levels and accelerating leaves growth of plants. Nitrogen is an energy source for microorganisms in soil that play an important role in weathering or decomposition process of organic material. As many as 78% of the air volume containing nitrogen which is needed in photosynthesis





Surabaya Indonesia

process. Deficiency of nitrogen nutrient could cause leaves growth distortion and plants become stunted. Nitrogen generally should be converted into ammonia or nitrates to facilitate absorption [21].

Table 3 Growth and Production of Mungbean Treatment

Plant height

at 7 WAP

Branch amount

at 7 WAP

Blooming time (days)

Pods amou

nt

Percentage of

full pods (%)

Grain weight

per Plant

(g)

Yields (kg/ha)

Weights of 100

grain (g)

M0 44.20 9.50 41.25 37.38 88.99 19.55 4,887.5 6.45

M1 37.58 8.88 41.88 33.88 85.86 18.19 4,547.5 6.60

M2 42.13 9.13 40.63 27.63 90.35 13.15 3,287.5 6.01

M3 33.43 9.00 41.88 34.00 87.17 15.39 3,847.5 7.09

M4 43.13 9.38 40.50 32.13 85.86 15.96 3,990.0 6.70

M5 36.08 9.38 39.38 31.25 94.36 17.35 4,337.5 6.89

M6 43.30 9.13 40.75 31.88 94.23 17.63 4,340.0 6.92

M7 40.40 9.50 40.25 34.38 89.38 18.16 4,540.0 6.38

M8 44.20 9.38 41.38 35.00 87.41 17.93 4,482.5 6.85

M9 40.13 8.75 41.88 36.50 84.30 18.02 4,505.0 6.66

M10 34.05 9.38 39.38 38.00 88.84 19.30 4,757.5 6.82

Observations of blooming time conducted when the plants first bring blooming flower out in each experiment plant trials at each treatment. Based on analysis of variance which have been treated mung bean plant with wastewater of rice which inoculated with 50 ml MOL towards blooming time of mung bean had no significant effect to the plant. In Table 3, the treatment M5 which is wastewater of rice which added with MOL 1 Yeast plus 2 Bacteria from wastewater of rice and kombucha, and treatment M10 which is wastewater of rice added with MOL 2 yeast plus 3 Bacteria from wastewater of rice, tape yeast and kombucha has rapid blooming time, that is 39.38 DAP (Days After Planting) but not significantly different to other treatments.

MOL solution containing micro and macro nutrients and also microbes. The presence of microbes in the MOL solution has potential as organic materials decomposer, growth stimulants, and plant disease and pest control agents. Therefore, a MOL solution has multifunctional use [22]. Wastewater of rice contains elements phosphor so it will stimulate the formation of flowers, fruits and seeds as well as being able to make seeds become pithy. The elements K can improve seed size and quality so that obtained yields are also increase [23].

Results from analysis of variance addition wastewater of rice inoculated with MOL 50 ml towards Amount of pithy pods in mung bean plant had no significant effect on the plant. In Table 3, after further tested, pods amount-of plant were not significantly different between treatments, with big amount of pods indicated in the treatment M10 which is wastewater of rice added with MOL 2 Yeast plus 3 Bacteria from wastewater of rice, kombucha and tape yeast (38.00 piece).

Wastewater of rice contains one of phospor compounds which useful in improving yields. Therefore, added wastewater of rice to plant have an important role in improving optimal result. Process of fruit ripening affected by the phosphor rate absorbed from the soil. The reason is because of phospor deficiency would reducing harvest quality besides phospor also help in asimilation and respiration process [bukhari, 2013].

Observation of pithy pods amount per plant conducted at the same time with pods amount observation. Based on the results analysis of variance, added wastewater of rice inoculated with MOL 50 ml towards the total percentage of pithy pods per plant had no significant effect to plant. After further tested, based on the data in Table 3, high percentage of pithy pods per plant on treatment M5 which is wastewater of rice added with MOL 1 Yeast plus 2 Bacteria from wastewater of rice and kombucha (94.36%), but not significantly different to other treatments.

Nutrients contained in MOL are the elements needed by plants in large quantities, especially in the formation and plant cells process. Element phosphorus for plants is useful to increasing the quality in grain crop and affecting the formation of the cell nucleus. Element Phosphorus for plants contribute in accelerate the growth of roots on the seeds, and also strengthen and accelerate growth in mature plants [21].





Surabaya Indonesia

Observation of seed weight per plant conducted during the harvesting time for all experiment plant. Based on the analysis of variance results, added wastewater of rice inoculated with MOL 50 ml towards the seed weight per plant had no significant effect on the plant. In Table 4, seed weight per plant were high in M0 (control), which is 0.25g urea, TSP 0.5g and 0.25g KCl (19.55 g) but not significantly different to other treatments. Seed weight per plant shown in Table 4 has been calculated per hectare in the treatment M0 (control), ie 4,887 kg. For treatment of added wastewater of rice obtaining heavy seed in treatment M10, which is wastewater of rice added MOL 2 Yeast plus 3 Bacteria (wastewater of rice, kombucha and tape yeast), ie 4,757.5 kg. Results of seed weight per plant which resulted in different description to variety swallow as yields dry weight.

Inorganic fertilizers contains elements potassium which could be derived from various types of minerals, and residual microorganisms. Potassium contribute in improving the grain quality [21].

Observations to seed weight per plant conducted per treatment through weigh 100 grains of mung bean. Based on the analysis of variance results of added wastewater of rice inoculated with MOL 50 ml towards the weight of 100 grains had no significant effect to the plant. After further tested, based on Table 4, the weight of 100 grains are high in treatment M3, which is 2 Yeast plus1 Bacteria from kombucha (7.09 g) but not significantly different to other treatments.

Result of MOL proved through completion of the containing nutrients. Although the nutrient in small amounts, but all the needs of micro and macro nutrients for plants could be met [21]. Element Phosphorus for plants is very important as an energy source in a variety of metabolic activities. One role of phosphorus is to boost the growth of plant shoots and roots, enhance the activity of other nutrients such as nitrogen and potassium balanced for the plants needs. Phosphorus help in accelerates N fixation through boodsting the blooming and forming seed and fruit, and also accelerate ripening of pods [24].

Based on analysis of wastewater of rice liquid fertilizer, can be seen that the contents of N nutrients is higher than the other nutrients. In the 5 yeast, the highest level of nutrient N obtained is 0.12%, whilst in the 4 yeast nutrients obtained is 0.09%.

There were 11 treatment in this research, the results obtained from this study is the best treatment is M0 as a control which using 0.25g urea, 0.5g TSP and 0.25g KCl. It also found in the treatment M10 with 2 yeast plus 3 bacteria from wastewater of rice, tape yeast, and kombucha. In the treatment M0, the highest result are parameters of plant height, branch amount, and seed weight. However, in the treatment M10, the highest result of treatment M10 are parameters pithy pods percentage per plant and blooming time but not significantly different from the other treatments.

Application of Microbia to Polianthes tuberosa flower

Analysis of variance results showed that fertilization does not provide a significant effect on the blooming time and stalks length of tuberose crop. From DAP test, the results showed that treatment P3 provide the most rapid blooming time, that is 84.25 days. While the longest time of blooming time occured in treatment P4, that is 95.50 days. Flower forming involves a phase changes from vegetative growth into reproductive growth. This transition is triggered by environmental gudances such as day length, and internal signals, such as hormones [25].

Table 4. Response of fertilization to crop growth and production of tuberose (Polianthes tuberosa L.)

Based on the results of DAP Test, the tuberose flower stem length showed that treatment P4 produce the

stem length as long as113.75 cm, while the short stalk length found in in treatment P1, which is 105.06 cm. Cell division in the meristems increaseing the potential for growth, however the cell expansion, especially cell elongation, which is responsible for the size enhancement [25].

Treatment Flowering time (days)

Length of stalk (cm)

Stalk diameter

(cm)

Number of florets per

panicle

Harvest time

(days)

Amount of harvest

P1 95,17 105,06 0,644 29,39 108,89 2,33

P2 87,60 108,70 0,661 32,10 102,40 2,00

P3 84,25 112,33 0,740 36,33 101,02 3,00 P4 95,50 113,75 0,760 32,50 116,50 3,00 P5 91,50 108,56 0,712 34,50 108,42 2,33





Surabaya Indonesia

Results from analysis of variance showed that fertilization does not provide significant effect on stalks diameter and number of florets per panicle on tuberose plant. Based on results of DAP Test, tuberose plant stalk diameter, treatment P4 resulted the biggest stalks diameter, which is 0.760 cm. Application of organic fertilizer or decayed organic material will increase the production of root branches. Stem growth also influenced by nutrients provided by the plant roots [26].

Based on results of DAP Test, as for observation of florets amount per panicle, treatment P3 provides the most amount of florets per panicle, that is 36.33 florets. While the least number of florets per panicle, that is 29.39 florets in treatment P1. Blooming proces basically is the interaction from two major influence factors, external and internal factors. External factors are influenced by the environment, such as day length (photoperiod), temperature, humidity, water and nutrients availability, while the internal factors are influenced by plant genetics itself [27].

Results from analysis of variance showed that fertilization does not provide a significant impact on the harvesting time and yields. Based on DAP Test results, observations indicate the fastest harvesting time is on treatment P3, ie101.02 days. The longest harvest time is in treatment P4, that is 116.50 days. Crop entering the reproductive phase after achieving the genetic character called size effect and endogenous timing. Effect size is a certain size associated with the ability of plants in regulate absorption, supply, and the allocation of food. Endogenous timing is a certain age which are genetically related to readiness for blooming [27].

Based on DAP Test results, Table 3 shows that the highest yields produced is from treatment P3 and P4, which is 3 times harvest. While the least yields is on treatment P2, which is 2 times harvesting. In order to achieving maximum results, application of organic fertilizer should be complemented with inorganic fertilizer so that the two are complementary, so that it would be forming nutrient rich farmland, with loose or crumb structure, and blackish brown coloured soil[28].

The study on Polianthestuberosa plants, research with a ratio of 4 microbes and 5 microbes shows that the 4 microbes and inorganic fertilizers showed no significant difference between treatments. In Agronomy, 4 microbes better than other treatments, while 5 Microbial better than others economically as it is saving fertilizer up to 75% compared to controls (inorganic fertilizers 100%) [12].

In terms of agronomy, the study showed that treatment P3 more profitable than other treatments. Treatment P3 gives tuberose plant blooming time faster, more number of florets, faster harvesting time, and the most yields, although not significantly different compared to other treatments.

Production periode of tuberose crop is 2 years. Table 5 shows the production cost per year on the treatment 100% Inorganic (P5 / Control) is Rp 2,157 billions, while in treatment 50% Inorganic (P3) is Rp 1,393 billion, and in the treatment 25% Inorganic (P4) is Rp 0.902 billion. Therefore the comparison of production cost per year for the three treatments showed that treatment 25% Inorganic (P4) is more economical.

Table 5 shows that the net income per 2 years per hectare in the treatment 100% Inorganic (P0 / Control) is Rp 7.805 billions, while in the treatment 50% Inorganic (P3) is Rp 11.958 billions, and in the treatment 25% Inorganic (P4), Rp 12.449 billions. Therefore comparison of net income per two years for the three treatments showed that 25% Inorganic (P4) more profitable.

Table 5 Summary of Tuberose Flower Farm Analysis (Production Period 2 Year with a land area of 1 hectare) in September 2014

No Treatment Production

cost (Rp) (billions)

Production cost / year (Rp) (billions)

Net income / year (Rp) (billions)

BEP (month)

1

100% Anorganik (P0/Control) 4,313 2,157 7,805 6,63

2 50% Anorganik (P3) 2,786 1,392 11,958 2,8 3 25% Anorganik (P4) 1,804 0, 902 12,449 1,74

Table 5 above shows that the BEP (Break Event Point) of treatment 100% Inorganic (P5 / Control) is 6.63

months, while the treatment of 50% Inorganic (P3) is 2.8 months, and the treatment of 25% Inorganic (P4 ) is 1.74 months. Therefore the BEP comparison for three treatments showed that 25% Inorganic (P4) is faster in reaching the break even point (BEP).

CONCLUSIONS Results of morphological identification, and sequence analysis of SSU rRNA from the three bacteria is

Burkholderia metallica, Burkholderia seminalis, and Gluconacetobacter saccharivorans.





Surabaya Indonesia

Results of morphological identification and sequence analysis of ITS rDNA to both yeasts is Trichosporan asahii, and Pichia kudriavzevii.

The results of research on mung bean showed that treatment M10 with 2 yeasts and 3 bacteria better than other treatments, though no significantly different.

On agronomy view, study on tuberose plant showed that treatment with the microbes mentioned above without Burkholderia seminalis with inorganic fertilizers 50% control tends to be better than other treatments.

On economy view, study on tuberose plant showed that treatment the microbes mentioned above without Burkholderia seminalis with inorganic fertilizers 25% control tends to be better than other treatments.

REFERENCE 1. Hadwani, Mayuri. 2014. Integrated Nutrient Management in Ratoon Tuberose. LAP LAMBERT

Academi Publishing. Germany. 2. Simanungkalit, R.D.M. 2006. Pupuk Organik dan Pupuk Hayati. Balai Besar Litbang Sumber Daya

Lahan Pertanian Badan Penelitian dan Pengembangan Pertanian, (Online), (http://www.academia.edu/3077297/PUPUK_ORGANIK_DAN_PUPUK_HAYATI, diakses 25 Desember 2014).

3. Susetya, Darma. 2013. Panduan Lengkap Membuat Pupuk Organik. PustakaBaru Press. Yogyakarta. 4. Puspitasari, R.T. 2003. Fermentasi Alamiah Limbah Cucian Beras sebagai Pupuk Hayati Anggrek

Dendrobium sp. pada Fase Vegetatif. Prosiding Simposium Nasional dan Kongres PERAGI VIII. Bandar Lampung.

5. Elfarisna. 2003. Penggunaan Air Limbah Cucian Beras sebagai Pupuk Organik Anggrek Dendrobium sp.pada Fase Generatif. Jurnal Penelitian Universitas Muhammadiyah Jakarta Vol.9 No.1. UMJ. Jakarta.

6. Mucharam, I. 2004. Pengaruh Dosis Air Limbah Cucian Beras Sebagai Pupuk Organik pada Tanaman Selada (Lactuca sativa L). Skripsi. Fakultas Pertanian. Universitas Muhammadiyah Jakarta.

7. Hermawan, D. 2004. Pengaruh Pupuk Cair dan Media Tanam Terhadap Pertumbuhan dan Produksi Tanaman Selada (Lactuca sativa L). Skripsi. Fakultas Pertanian. Universitas Muhammadiyah Jakarta.

8. Irwansyah. 2004. Pengaruh Air Limbah Cucian Beras Terhadap Pertumbuhan dan Produksi Tanaman Bayam (Amarantus tricolor L). Skripsi. Fakultas Pertanian. Universitas Muhammadiyah Jakarta.

9. Sofar, 2005. Pengaruh Air Limbah Cucian Beras sebagai Pupuk Organik terhadap Pertumbuhan dan Produksi Tanaman Bawang Daun (Allium fistulatum L). Skripsi. Fakultas Pertanian Universitas Muhammadiyah Jakarta.

10. Suryati, Yati. 2003. Penerapan Teknologi Efektif Mikroorganisme Pada Air Limbah Cucian Beras Sebagai Pupuk Anggrek (Phalaenopsis sp.). Prosiding Simposium Nasional dan Kongres Peragi VIII. Bandar Lampung.

11. Suryati, Yati. 2010. Peluang Pemanfaatan Air Limbah Cucian Beras Sebagai Pupuk Organik. Prosiding Seminar Nasional Pertanian Indonesia Menuju Millenium Development Goals (MDGs).Yogyakarta.

12. Elfarisna, RT. Puspitasari, Suryati dan Pradana NT. 2014. Penambahan Mikroorganisme Lokal pada Air Cucian Beras sebagai Pupuk Organik . Jurnal FMIPA Volume 12- No. 2. Universitas Terbuka. Tangerang.

13. Anggoro,U.K. 2014. Laporan Tahunan Direktorat Jendral Tanaman Pangan. 14.https://www.google.co.id/url?sa=t&rct=j&q=&esrc=s&source=web&cd=4&cad=rja&uact=8&ved=0CD

QQFjAD&url=http%3A%2F%2Ftanamanpangan.pertanian.go.id%2Ffiles%2FLAPTAH_DIRJEN%25202013.pdf&ei=306VVbDKIoOhugTTv4L4DA&usg=AFQjCNFomMy7ZmoEzayV_CNsbtR3gHAKog&bvm=bv.96952980,d.c2E. [29Juni 2015].

15. Setiaji, H. 2014. RI juga Impor Kacang Hijau, Salah Satunya dari Etiopia. http//m.detik.com/finance/read/2014/05/downloads/detikFinanc[13 Mei 2015].

16. Deni. 2012. Bisnis Bunga Potong Sedap Malam Cukup Menjanjikan, (Online), 17. (http://denicage.blogspot.com/2012/10/bisnis-bunga-potong-sedap-malam-cukup-menjanjikan.html,

diakses 25 Februari 2015). 18. Ucup. 2014. Wawancara Produksi Bunga Sedap Malam. Pasar Bunga Rawa Belong. Jakarta. 19. Central Bureau of Statistics. 2014. Statistik Tanaman Hias. Badan Pusat Statistik. Jakarta. 20. Kurtzman& Fell. 1998.the monograph "The yeasts, A Taxonomic Study". pp. 222-223& 857-858. 21. Anonim. Bergey's Manual of Systematic Bacteriology, Volume 2, 2nd ed., 2005, p. 575& 72-73. 22. Hasibuan R., Nurbaiti, and Adrian. 2011. Pertumbuhan dan Hasil Kacang Hijau (Vigna radiata L.)

Varietas 129 pada Beberapa Dosis Batuan Fosfat di Medium Gambut. Universitas Riau. Riau.





Surabaya Indonesia

23. Mulyono. 2014. Membuat MOL dan Kompos dari Sampah Rumah Tangga.Agromedia Pustaka. Jakarta.

24. Suwahyono. 2014. Cara Cepat Buat Kompos dari Limbah. PenebarSwadaya. Jakarta. 25. Esrita. 2009. Respon Tanaman Kacang Hijau (Vigna radiata L.) terhadap Pupuk Organik Lengkap.

Universitas Jambi. Jambi. 26. Istiqomah, N. 2012. Efektivitas Pemberian Air Cucian Beras Coklat terhadap Produtivitas Tanaman

Kacang Hijau (Phaseolus radiatus L.) pada Lahan Rawa Lebak. 33, pp.99–108. Kalimantan. 27. Campbell, NA dan JB Reece. 2008. Biologi, Edisi Kedelapan, Jilid 2. Penerbit Erlangga. Jakarta. 28. Goldsworthy, PR dan NM Fisher. 1992. Fisiologi Tanaman Budidaya Tropik. GadjahMada University

Press. Yogyakarta. 29. Elisa. 2014. Kualitas dan Produksi Bunga, (Online), (http://elisa1.ugm.ac.id/II-

kualitas%20dan%20prod-bunga, diakses 27 Maret 2015).

30. Lingga, Pinus dan Marsono. 2013. Petunjuk Penggunaan Pupuk. Penebar Swadaya. Jakarta.



150


Surabaya Indonesia

ORAL PAPER OF NATURAL PRODUCTS AND MEDICINAL CHEMISTRY (ONPM)



Ahmed et al. 151


Surabaya Indonesia

Determination of 40K Concentration in Imported Milk Powders in

Erbil City Local Markets

Ali H. Ahmed* and Mohhammed I. Hussein

1Iraq/Kurdistan Region/ Salahaddin University-Erbil, College of Science, Physics Dept. *Corresponding author’sEmail: [email protected]

Abstract.The control and monitoring of radioactive elements in foodstuffs is fundamental for human health maintenance. This work presents procedures to measure the 40K radioactivity levels in the powdered milk samples. The amounts of radioactivity in the fourteen powdered milk samples have been investigated; samples were collected from the local markets of Erbil governorate which are imported from different countries. The measurements were performed utilizing the NaI(Tl) scintillation detector with an active area of 2×2 inches and have energy resolution 11.3% at the 662keV Cs-137 line. The results allowed the quantification of 40K, were the average activity was 288.54±11.54 Bq/kg. The lower and higher level of detection was, respectively, 107.52±7.18 and 438.25±14.50 Bq/kg. The results obtained for the milk samples were compared to data found in the other countries and to the limits established by the Kurdistan Region-Iraq to assure its safety to human consumption.

INTRODUCTION Measurements of radioactivity in environment and in foodstuffs are extremely important for controlling

radiation levels to which mankind is direct or indirectly exposed. Another important fact is that, importation of contaminated food from any region that suffered a nuclear accident can be indirectly affect people health around the world [1]. Radionuclides are found in air, water and soil, and additionally in us, being that we are products of our environment. Every day, we ingest/inhale nuclides in the air we breathe, in the food we eat and the water we drink [2, 3, 4]. Foodstuffs are known to contain natural and man made radionuclides that after ingestion, contribute to an effective internal dose.

Knowledge of radioactivity levels in human diet is of particular concern for the estimation of possible radiological hazards to human health [5]. Potassium is a soft, silver-white metal, an important constituent of soil; it is widely distributed in nature and is present in all plant and animal tissues. Potassium-40 have half-life t1/2=1.28×109 years (long life isotope), also is a naturally occurring radioactive isotope of potassium. Two stable (nonradioactive) isotopes of potassium exist, potassium-39 and potassium-41. Potassium-39 comprises most (about 93%) of naturally occurring potassium, and potassium-41 accounts for essentially all the rest. Radioactive postassium-40 comprises a very small fraction (about 0.012%) of naturally occurring potassium [6]. Potassium is an essential element for human life. We usually obtain potassium from vegetables, meat or fruits. The recommended daily intake amount of potassium should be between 1.56 to 5.86 g [7]. Milk is a sensitive indicator of many contaminants due to fallout of radionuclides or other contaminants because of the feeding of cows on grass, a plant with a high absorbency of various dangerous substances against producers due to political corruption, lack of education and knowledge [8]. Among different kinds of foodstuffs milk is a reliable indicator of the general population intake of certain radionuclides, since fresh milk is consumed by a large segment of the population, and contains several biologically significant radionuclides [9]. Milk is considered one of the main nutrient materials and it‘s high nutrition value is due to its contents of main metal elements. The percentage of potassium in milk is estimated to be 14% it may differ in different types of the same nutrition [10].

EXPERIMENTAL METHOD Fourteen different samples of powdered milk were collected from local markets in Erbil governorates. The

types of powdered milk are collected from different imported country. All samples were dry milk (powdered milk). The gamma ray spectrometer have been used for measuring energy spectra for each sample. The spectrometer consists of NaI(Tl) scintillation detector with an active area 2×2 inches and have energy resolution 11.3% at the 662keV Cs-137 line. An amount of 1000 g of KCl was transferred to a Marinelli beaker and placed in front of the scintillation detector (surrounded by a lead shield designed to remove the effect of the background variation) to be examined by the gamma ray spectrometer for a period of four hours (14400 sec), and with an operating voltage of 960 Volt to detect the 40K isotope in the KCl.

A computer assisted science system (MAC-CASSY) was used in this study. The background spectrum of the laboratory environment have been determined by using an empty Marinelli beaker putted on the detector with the same period and operating voltage used for determination of samples energy spectra. By the same way



Ahmed et al. 152


Surabaya Indonesia

(500gm) for each sample has been used for counting and consequently obtaining the net sample spectra by subtracting the background spectrum. Each sample of powdered milk contains an amount of natural potassium. Table 1 illustrates natural potassium, description, employee, and country.

Table 1. Natural Potassium (K) content per 100 gms in each studied sample

DATA ANALYSIS

Sample Spectra and 40K Radionuclide Identification

The gamma-ray spectrum of each studied sample were taken by the NaI(Tl) detector. We can determine the activity concentration of 40K. The detector was calibrated using the Ra-226 standard source; which has six γ-ray emitters ranged from 186 to 1120 keV, Fig.(1). The detector is surrounded by a lead shield to reduce the background of the system.

Fig.1. Ra-226 Energy calibration

Net Area Calculation

The area above the background represents the total counts between the vertical lines minus the trapezoidal

area below the straight line. If the total counts are T and the endpoints of the straight line are 1B and 2B , then the net area is given by [11]:

Sample Code Description Employee Country Natural Potassium

(K) content in mgm (per 100 gms)

S1 Nactalia 1 Infant France 665 S2 Nido Adult United Arab Emirates - S3 Incolac Adult Belgium - S4 Lait 1 Infant France 460 S5 Al-Mudhish Adult Oman - S6 Nactalia 2 Infant France 770 S7 Guigoz 1 Infant Philippines 578 S8 Guigoz 2 Infant Philippines 645 S9 Green Farms Adult Newzealand 1565

S10 Dielac 1 Infant Newzealand 661 S11 Dielac 2 Infant Newzealand 843 S12 Biomil 1 Infant Belgium 240 S13 Premier Adult Ireland - S14 Anchor Adult Newzealand 1260



Ahmed et al. 153


Surabaya Indonesia

)(

221 BB

nTAnet

(1)

Where n is the number of channels between 1B and 2B , T is the gross area or total counts, and )(

2 21 BBn

is the background. The net area netA is also called area under the peak; Figure (2).

.

Fig.2. Net Area determination.

The K-40 Content 40K content expressed in percent (%), has been investigated as follow: The 40K activities is deduced

from KCl. The Activity of 40K from (1000 gm) KCL is expressed according to [12]:

2/1

2lnt

NA

(2)

Where

2010.69.9..1000

elAol

rNM

gmN

(3)

olM= 74.55gm/mol (Molecular weight of KCl), AN = 6.022×1023 mol-1 (Avogadro number), elr

=0.012

% (Relative content of K-40), 2/1t 1.28×109 y (Half-life of K-40). Then A=16.7kBq. The calibration factor is

given by [13]

atek R

AK

(4)

and count rate ateR=Net Area/Counting Time=Anet/21600=4.379S-1 then,

KK=3813.6 Bq.s

From eq. (5) kate KRA . the activity can be calculated.

Activity concentration of the K-40 of standard samples (1000gm) is equal to activity/523gm, where 523gm is the portion of potassium in 1000 g of KCl. The activity concentration for each sample is expressed in the same steps, by dividing the samples activity by the 500gm. The ratio between the specific activities of the measured samples to specific activity of the standard sample is the potassium content of the sample as given in ref. [12]



Ahmed et al. 154


Surabaya Indonesia


The radioactivity in powdered milk collected from different imported country has been studied in order to find 40K activity. The Activity concentration of 40K of the samples under investigation in Bq/Kg where determined from the 1461 KeV photopeaks within the gamma spectra. The Figures (3 and 4) show gamma-spectra of the highest and lowest activity concentration powdered milk samples.

Fig.3: The net gamma ray spectrum of highest sample.

Fig.4.The net gamma ray spectrum of lowest sample.

Potassium-40 activity concentrations in the powdered milk ranged from 107.52±7.18 to 438.25±14.50 Bq.kg-1 ,(average 288.54±11.54 Bq kg-1), Table (2) shows the highest and lowest value of activity concentrations of 40K which may be referred to the eat of grass from animal. The Incolac type of powdered milk sample show the highest value of 40K activity (438.25±14.50 Bq.kg-1) and percentage content among all the studied samples. (Lait) shows the lowest value of 40K activity (1107.52±7.18).



Ahmed et al. 155


Surabaya Indonesia

Table (2): Activity Concentration and 40K content per cent in each studied sample

The activity concentration in above table can be described in the following figure

Fig. 5 The activity concentration of 40K within the studied samples.

Table 3 show the Comparison of the evaluated mean activity concentration with those of other Countries.

Table (3): Comparison of the evaluated mean activity concentration with those of other countries.

0

100

200

300

400

500

S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14

Act

ivit

y C

once

ntra

tion

(B

q/K

g)

Sample Code

Sample Code

Activity Concentration (Bqkg-1)

40K content percent (%)

S1 144.96±8.34 0.45 S2 425.77±14.29 1.33 S3 438.25±14.50 1.369 S4 107.52±7.18 0.336 S5 405.61±13.95 1.26 S6 279.3±11.57 0.87 S7 231.84±10.54 0.724 S8 175.2±9.17 0.54 S9 350.89±12.97 1.096 S10 230.4±10.51 0.72 S11 276.01±11.51 0.862 S12 218.88±10.24 0.684 S13 317.77±12.35 0.99 S14 437.29±14.48 1.36

Country Average Activity Concentration of K-40

(Bq.kg-1) Kurdistan, Iraq, 2012 288.54

Iran [5], 2007 475.67

Saudi Arabia [13], 1995 353

Brazil [1], 2004 482

60 Europeans Laboratories

[14], 2008 540



Ahmed et al. 156


Surabaya Indonesia

Therefore, we can see significant differences between milk powder from different countries depending on the dominant farming methods.

CONCLUSIONS

The 40K radionuclide has been identified in milk powder sample spectra. The values of activity concentration of 40K are lower than of other studied countries. As the potassium is rough uniformly distributed in the body, follows intake in foods, and its concentration in the body is under homeostatic control, it is less dangerous for human health than 137Cs. In general terms, it can be concluded that the implemented technique presented good results when compared with other literature data. Also it can be concluded that milks here analyzed are safe for human consumption because their radioactivity levels are lower than the maximum levels permitted.

REFERENCES1. MELQUIADES F.L, APPOLONI C.R, Ciênc. Natural Radiation Levels in Powdered Milk Samples 2. Tecnol. Aliment., Campinas, 24(4): 501-504 (2004). 3. O‘Connor, C., Currivan L., Cunningham N., Kelleher K., Lewis M., Long S., McGinnity P., 4. Smith V., McMahon C., Radiation Doses Received by the Irish Population, Radiological 5. Protection Institute of Ireland, (2014). 6. Alias M., Hamzah Z., Saat A., Omar M., and Wood A., An assessment of absorbed dose and radiation 7. hazard index from natural radioactivity, The Malaysian Journal of Analytical Sciences, Vol. 12, No. 1, 8. (2008). 9. Saleh H., Abu Shayeb M., Natural radioactivity distribution of southern part of Jordan 10. (Ma'an) Soil, J. Annals of Nuclear Energy 65 184–189 (2014). 11. Report of the United Nations Scientific Committee on the Effects of Atomic Radiation, 12. Sixtieth session ). 13. Argonne National Laboratory, EVSHuman, Health Fact Sheet, August (2005). 14. Radioactivity of tobacco leaves and Radiation Dose induced from smoking ,international 15. Journal of environmental research and public health, 6, 558-567 2009. 16. Pietrzak-Fiećko R., Smoczyński S., Evaluation of Cs-137 Content in Powdered Cow Milk from Four 17. Regions of Poland, Polish J. of Environ. Stud. Vol. 18, No. 4, 745-748 (2009). 18. Afshari N., Abbasisiar F., Abdolmaleki P., and Nejad M., Determination of 40K concentration in milk

samples consumed in Tehran-Iran and estimation of its annual effective dose, Int. J. Radiation. Res., 7 (3): 159-164, 2009

19. Hassan H. I., and Mheemeed A. Kh., Transfer of K40 from soil to plants in an agricultural field and 20. its EDE from milk ingestion Damascus University Journal for Basic Sciences Vol. 24, No. 2, (2008). 21. Canberra 81-82, Canberra Industries, Inc., USA, (1982). 22. Leybold Physics Leaflets, Leybold Didactic GmbH, P6.5.5.4., (2001). 23. Abdul-Fattah A., and Abdul-Majid S., The fourth Saudi Engineering Confrence, 24. Nov.1995.Volume V 25. Wätjen U., Spasova Y., Altzitzoglou T., Emteborg H. and Pommé S. JRC Scientific and 26. technical reports EUR 23270 EN - 2008 27.

http://www.ijrr.com/files/site1/user_files_fad21f/admin-A-10-1-332-5f0382d.pdf

http://www.ijrr.com/files/site1/user_files_fad21f/admin-A-10-1-332-5f0382d.pdf



Wahyuni et al. 157


Surabaya Indonesia

Beans (Phaseolus Vulgaris L.) Extract as Anti Dyslipidemia:

Decrease of total cholesterol, malondialdehyde, low density

lipoprotein, and increase of high density lipoprotein on rat wistar

Sri Wahjuni*

Chemistry Departement, Faculty of Math and Science Udayana University *Corresponding author’s Email: [email protected]>

ABSTRACT

Abstract. Excessive fat consumption leads to increase of blood cholesterol level more than normal condition which is called as Dyslipidemia. Dyslipidemia is an abnormal lipoprotein metabolism, usually associated with lipoprotein over production or deficiency. Dyslipidemia is also often described as an event of hyperlipidemia and also as a risk factor for cardiovascular disease. This is an experimental study for accessing the ability of beans extract as an anti dislipidemia. The anti dislipidemia were marked by decrease of total cholesterol, malondialdehid, Low Densyty Lipoprotein levels, and increase HDL of Wistar rat induced dislipisemia with high cholesterol diet for 16 week. This study applied randomized posttest only control group design. The samples were 24 Wistar Rat, randomized into 6 treatment: negative control group diet standard (Treatment group 1), positive control group with diet high cholesterol (Treatment cholesterol group 2), diet high cholesterol with beans ( Phaseolus vulgaris) 50 kg/bw (Treatment group 3),diet high cholesterol with beans (Phaseolus vulgaris L) Steenis) 100 kg/bw (Tretment group 4), diet high cholesterol with beans (Phaseolus vulgaris L) 150 kg/bw (Treatment group 5). diet drug simvastatin 0,18 mg/day /200 gram BB (Treatment group 6), After 16 weeks treatment, blood of rats were driven for total cholesterol and MDA assays. LDL and HDL . All of data analyzed by Anova to obtain the treatment different toward control by statistically with significance at α=0.05. The result shows that extracts of beans (Phaseolus vulgaris L) in a dose of 150 mg/kg bw decrease total cholesterol of 23.88 %, MDA of 70.60 % LDL 38.09 % also increase the HDL of 59,63%, Extract beans (Phaseolus vulgaris L) has an ability to prevent cardiovascular disease. This is caused by the present of phytosterol in the beans extract which has been phyto-chemistry tested and analyzed by GC-MS such as stigmasterol substance.

INTRODUCTION

Nowdays many Indonesian people have the wrong diet, trend to like fast food that contains nutrients that are not balanced. In general, this fast food containing saturated fat and high salt with low fiber content. Excessive consumption of saturated fat, low carbohydrate, and less fiber from the dailydietisafactorcausingincreasedbloodcholesterol.High blood cholesterol (dyslipidemia) is not only experienced by obese people, but it can also occur in people who are thin and can afflict people who are young. Various walks of life, have to try to live a healthy lifestyle in order to keep cholesterol in the blood remained normal. In the body there is a normal cholesterol level is 160-200 mg (LIPI, 2009). Dilipidemia is a situation where an increase in blood cholesterol levels that exceed normal circumstances, accompanied by increased levels of total cholesterol, triglycerides, and LDL cholesterol. Dilipidemia occur when total cholesterol concentration reaches ≥ 240 mg / dl, LDL ≥ 160 mg / dl, and triglycerides ≥ 150 mg / dl (Montgomery, 1993)

High blood cholesterol (dyslipidemia) is not only experienced by obese people, but it can also occur in

people who are thin and can afflict people who are young. Various walks of life, have to try to live a healthy lifestyle in order to keep cholesterol in the blood remained normal. In the body there is a normal cholesterol level is 160-200 mg (LIPI, 2009). Dilipidemia is a situation where an increase in blood cholesterol levels that exceed normal circumstances, accompanied by increased levels of total-cholesterol, triglycerides,and LDLcholesterol.Fat derived from food will undergo pencernaaan processes in the intestine into free fatty acids, triglycerides, phospholipids and cholesterol, then absorbed into the form of chylomicrons. Breakdown of




Wahyuni et al. 158


Surabaya Indonesia

chylomicrons circulate to the rest of the liver and separated into cholesterol. Most cholesterol is discharged into the bile as bile acids and in part again together with triglycerides to be allied with a particular protein (apoprotein) and form very low density lipoprotein (VLDL). VLDL very low density lipoproteins are further broken down by the enzyme lipoprotein LDL can not be the last 2-6 hours and immediately be converted to LDL. This process is called mechanism of dyslipidemia (Soeharto, 2004).

Demage Lipids in LDL-cholesterol in the blood due to dyslipidemia produce various products of decomposition are relatively stable, especially aldehyde reactive α, β-unsaturated, such as malondialdehyde (MDA), 4-hydroxy-2-nonenal (HNE), heksanal, and 2-propenal (acrolein) (Uchida, 2003 Carani.et.al, 2004). Malodialdehid a (marker) one of the products of decomposition of per oksidadasi acids plural unsaturated fats (polyunsaturated fatty acids; PUFA). Several studies medicinal plants tadisional that have been carried out include research conducted by Ratni (2013), gedi (Abelmoschus manihot) shows the effect of hypolipidemic that can lower cholesterol levels in male rats Wistar with the content of flavonoids and steroid in gedi (Abelmoschus manihot) , Utariningsih (2007) determines the hair decoction of maize (Zea mays) is effective in lowering cholesterol levels of mice, rats decreased levels of cholesterol using -sitosterol compounds contained in the hair of corn (Zea mays).

According to experts the researchers herbal, black cumin (Nigella sativa) play a role in reducing the excess cholesterol in the blood by active substance thymoquinone, omega 9, omega 3, omega 6, saponins and phytosterols (Edi Junaedi, 2011).Subekti (2006) using leaf katuk (Sauropus adrogynus L. Merr) to produce low cholesterol Japanese quail. Leaves katuk (Sauropus adrogynus L. Merr) phytosterol-containing compounds that can lower cholesterol in Japanese quail products. Wahyu (2013), using n-hexane extract of pomegranate fruit flesh is white (Punica granatum) which can lower blood cholesterol levels in rats (Rattus norvegicus L.) with a content of phytosterols in the flesh white pomegranate (Punica granatum). Based on research conducted by Jannah, et. al. (2013), the fruit extract of beans (Phaseolus vulgaris L.) contain phytosterol compounds. Phytosterols are natural plant sterols that contains a minimum of 70% -sitosterol. According Bonsdoff-Nikander (2005), a compound phytosterols can lower cholesterol through competition between cholesterol and phytosterols in a mixture of micelles, as well as by binding cholesterol in the digestive tract. There has been no research on the effectiveness of the test pieces of bean (Phaseolus vulgaris L.) in lowering cholesterol levels, so as to investigate the ability of ethanol extract of the fruit of beans (Phaseolus vulgaris L.) as an anti-hypercholesterolemia through reductions in total cholesterol, Low Density Lipoprotein (LDL) as well as an increase High Density Lipoprotein (HDL) in Wistar rats dyslipidemia.


Results of the analysis by GC-MS techniques to extract ethanol kentak fruit beans (Phaseolus vulgaris L.) obtained 19 peak compounds were detected , shown in Figure 1 with a retention time (tR), peak area, and different molecular weights in Table 1.



Wahyuni et al. 159


Surabaya Indonesia

Figure 1 Chromatograms of ethanol extractnof beans (Phaseolus vulgaris L.)

Table 1 Table1. Substance detected in ethanol extract of beans

Peak Conpound tR m/z %area 1 Propane, 1-(1-ethoxydthoxy)-(CAS) 1-Ethoxy 3,07 45,00 6,63% 2 Ethanol, 2,2-diethoxy- 3,25 47,00 37,24% 3 Ethane, 1,1-diethoxy-(CAS) 1,1-Diethoxyethane 3,40 45,00 22,92% 4 Butane, 1,1-diethoxy-3-methyl- 3,57 103,00 2,02% 5 1-ethoxy-1-pentoxy-ethane 3,83 73,00 2,68% 6 1,1,3-Triethoxybutane 5,67 73,00 1,21% 7 2-Pentanone, 5,5-diethoxy- 6,49 103,00 1,43% 8 Ethanol, 2,2-diethoxy-(CAS) Glycolaldehyde 6,93 103,00 2,45% 9 1(2H)-Pyridinecarboxaldehyde, 3,4-dihydro-5-(2-piperidinyl) 10,83 110,00 0,78%

10 Propanoic acid, 2-methyl-, 1-(1,1-dimethylethy)-2-methyl-1,3-propanediyl ester (CAS) 12,38 71,00 1,96%

11 1-(+)-Ascorbic acid 2,6-dihexadecanoate 16,41 73,00 5,16% 12 Hexadecanoic acid, ethyl ester (CAS) Ethyl palmitate 16,75 88,00 2,03% 13 9,12-Octadecadienoic acid (Z,Z)- 18,05 67,00 4,88% 14 2H-Pyran, 2-(2-heptadecynyloxy)tetrahydro-(CAS) 18,11 55,00 3,99% 15 Octadecanoic acid 18,32 73,00 1,78% 16 9,12,15-Octadecatrienoic acid, ethyl ester, (Z,Z,Z)- 18,38 79,00 0,56% 17 9-Octadecenamide, (Z)-(CAS) OLEOAMIDE 20,06 59,00 0,61% 18 Hexadecanoic acid, 2-hydroxy-1-(hydrocymethyl)ethyl ester (CAS) 21,29 98,00 0,67% 19 Stigmasterol 28,89 55,00 0,96%

In the ethanol extract of the fruit of beans (Phaseolus vulgaris L.) contain phytosterol compounds shown in the retention time of 28.89 minutes as stigmasterol. Results of mass spectra of compounds stigmasterol the ethanol extract of the fruit of beans (Phaseolus vulgaris L.) is shown in Fig.2, beheading pattern table shown in Table 2 and the results fragmentation stigmasterol compounds shown in Figure 3.



Wahyuni et al. 160


Surabaya Indonesia

Figure 2. Spektra Massa Stigmasterol

Table 2 Pattern beheading

No m/z The possibility that the missing fragment

1. 412 M+ - 2. 394 M+-18 -H2O 3. 351 (M+-18)-43 -C3H7

Figure 3. Fragmentation stigmasterol

Stigmasterol compound was detected at a retention time of 28.89 minutes and abundance



Wahyuni et al. 161


Surabaya Indonesia

Of 0.96% showed the presence of molecular ions (M +) at m / z 412 were shot by high-energy electrons (70ev) resulting fragment peaks at m / z 394 and releases H2O m / z 351 C3H7 release. The molecular weight compounds l stigmasterol even showed that the compound does not contain components of atoms N

DYSLIPIDEMIA

Decrease of total Cholesterol level

Measurement of total cholesterol was conducted to determine the amount of cholesterol in the body. Results of the determination of total cholesterol shown in Table 3.

Table 3 Table average level of total cholesterol

Group Average level of total-cholest (mg/dL) Persentation decrease

K1 62,50 - K2 88,75 - P1 69,55 21,63% P2 68,50 22,81% P3 67,55 23,88% KS 69,85 21,29%

Figure 4. Average level of total cholesterol

Description:

K1 = negative control group; K2 = positive control group P1 = group of ethanol extract of beans fruit 50 mg / kg bw P2 = group of ethanol extract of fruit bean dose of 100 mg / kg bw P3 = group of ethanol extract of fruit bean dose 150 mg / kg bw KS = group of medication simvastatin dose of 0.18 mg / kg bw Results of this study showed total cholesterol levels in the blood of Wistar rats hypercholesterolemia decreased affected by the ethanol extract of the fruit of beans (Phaseolus vulgaris L.) with dose variation is a dose of 50 mg / kg BW, 100 mg / kg bw and 150 mg / Kg BB.

Analysis of the results followed by one-way ANOVA test and Post Hoc Study by Tukey's test / HSD to know the group that has the same effect or different from the others. ANOVA analysis results which show the

0

20

40

60

80

100

K1 K2 P1 P2 P3 KS

Average level of total-cholest (mg/dL)



Wahyuni et al. 162


Surabaya Indonesia

value of P <0.05 so H1 H¬0 accepted and rejected, which means there is a significant difference in total cholesterol levels of hypercholesterolemic control group to the treatment group after administration of ethanol extract of the fruit of beans (Phaseolus vulgaris L.) with a dose of 50 mg / kg BW, 100 mg / kg BW, 150 mg / kg body weight and drug delivery simvastatinatadoseof0.18mg/kgBw. Tukey test results / HSD showed that the difference in total cholesterol levels were significantly occurred in the control group and all treatment groups with the value of the most significant differences in the treatment group of fruit ethanol extract of beans (Phaseolus vulgaris L.) with a dose of 150 mg / kg bw. This shows that the ethanol extract of the fruit of beans (Phaseolus vulgaris L.) with dose variation has an influence in the determination of total cholesterol Wistar

This is due to the content of phytosterols in fruit ethanol extract of beans (Phaseolus vulgaris L.) were detected as having significant influence stigmasterol.Antidislipidemia effect of phytosterols depends on the amount of plant sterols and stanols that are used by beberapan doses (Jones, 2000). This research has the best results for lowering total cholesterol levels at doses of ethanol extract of fruit bean (Phaseolus vulgaris L.) 150 mg / kg bw. Drug simvastatin is used to lower the total cholesterol level has a value almost equal to the reduction in the ethanol extract of the fruit of beans (Phaseolus vulgaris L.) dose of 50 mg / Kg. Tangible results are also seen in the levels of total cholesterol medication simvastatin group approached the values of total cholesterol ethanol extract of the fruit group snaps dose 150 mg / kg bw.

Decrease Low Density Lipoprotein (LDL)

Measurement of Low Density Lipoprotein cholesterol (LDL) is performed to determine the risk of disease atherosclerosis. Atherosclerosis is a disease of blood vessel constriction by too much cholesterol amounts in the blood that form deposits on the walls of blood vessels (Almatsier, 2004). Cholesterol forming plaque in the blood vessel wall is cholesterol Low Density Lipoprotein (LDL). This is due to atherogenic LDL cholesterol has properties that memudahkanya attached to the inner wall of blood vessels (Heslet, 1996). LDL cholesterol is one of the causes of hypercholesterolemia. Eating a high fat can cause the concentration of cholesterol in the body increases and decreases the synthesis and activity of LDL receptors, so that the levels of LDL cholesterol in the body is increased (Murray, 2006). LDL cholesterol determination results are shown in Table 4

Table 4 Table Average of LDL-cholesterol

Group Average of LDL-cholesterol (mg/dL) Persentation decrease

K1 28,50 - K2 50,60 - P1 33,80 33,20% P2 32,00 36,75% P3 31,33 38,09% KS 35,00 30,83%

Figure 5. Average level of LDL-cholesterol

0

10

20

30

40

50

60

K1 K2 P1 P2 P3 KS

Average of LDL-cholesterol (mg/dL)



Wahyuni et al. 163


Surabaya Indonesia

Tukey test results / HSD in table4 above showed significant differences in levels of LDL occurs in controk group and all treatment groups with significant value palig difference in the treatment group of ethanol extract of beans (Phaseolus vulgaris L.) with a dose of 150 mg / kg bw. This shows that the ethanol extract of the fruit of beans (Phaseolus vulgaris L.) with dose variation has an influence in determining the levels of LDL Hypercholesterolemia Wistar rats. Decreased levels of LDL Wistar rats caused by the ethanol extract of the fruit of beans (Phaseolus vulgaris L.) contains phytosterols such as stigmasterol compounds that can lower blood cholesterol levels by inhibiting HMG CoA reductase will decrease the synthesis of cholesterol LDL (Price, 2006). In addition phytosterol is a sterol group also has antioxidant properties (Wang, 2002), which acts as a reducing LDL in the body (Radhika, 2011). Antioxidants can increase acid secretion empendu (Lamson, 2000).

In the heart, phytosterols and cholesterol that is secreted into bile by the transporter ABCG5 and ABCG8. The ability of phytosterols faster experience into the digestive secretion of the cholesterol, so more phytosterol enters the blood circulation and prevents the buildup of LDL cholesterol back to the blood circulation (Bonsdoff-Nikander, 2005). Phytosterols can lower LDL cholesterol levels the most effective at a dose of 150 mg / kg bw. Decrease in LDL cholesterol levels are influenced by a decrease in total cholesterol levels, as seen in total cholesterol levels in this study experienced a decrease as well as LDL cholesterol levels

Increase Kolesterol High Density Lipoprotein (HDL)

Measurement of High Density Lipoprotein cholesterol (HDL) was conducted to determine the transport of excess cholesterol in extrahepatic tissue and cell cleanser (scavenger cells) that will be brought back to the heart. In the extrahepatic tissues and cell cleanser (scavenger cells), HDL cholesterol will interact with the enzyme lecithin cholesterol acyl transferase release remnant and VLDL cholesterol to the liver and then excreted into the bile (Bahri, 2004).HDL cholesterol levels in the blood are high very beneficial in reducing the risk of atherosclerosis, because the function of HDL cholesterol transports cholesterol from extrahepatic tissues to the liver so menncegah occurrence of calcification (Hartayo 2008). HDL cholesterol assay results are shown in Table 5. Table 5 : Average of HDL-cholesterol (mg/dL)

Group Average of HDL-cholesterol (mg/dL) Persentation Increase

K1 26,60 - K2 17,15 - P1 23,90 39,35% P2 22,85 33,23% P3 27,38 59,63% KS 24,10 40,52%

Figure 4. Average level of HDL-cholesterol

0

5

10

15

20

25

30

K1 K2 P1 P2 P3 KS

Average of HDL-cholesterol (mg/dL)



Wahyuni et al. 164


Surabaya Indonesia

HDL cholesterol levels in Table 5 shows that the Wistar rat blood hypercholesterolemia increased influenced oleh.pemberian fruit ethanol extract of beans (Phaseolus vulgaris L.) with dose variation is a dose of 50 mg / kg BW, 100 mg / kg bw and 150 mg / kg. Analysis of the results followed by one-way ANOVA test and Post Hoc Study by Tukey's test / HSD to mengaetahui group that has the same effect or different from the others. ANOVA analysis results which show the value of P <0.05 so H1 H¬0 accepted and rejected, which means there is a significant difference in HDL levels kelommpok control hypercholesterolemia with treatment group after administration of ethanol extract of the fruit of beans (Phaseolus vulgaris L.) with dose 50 mg / kg BW, 100 mg / kg BW, 150 mg / kg body weight and drug delivery simvastatin at a dose of 0.18 mg / kg bw. Tukey test results / HSD showed that significant differences in levels of HDL occurs in the control group and all treatment groups with the value of the most significant differences in the treatment group of fruit ethanol extract of beans (Phaseolus vulgaris L.) with a dose of 150 mg / kg bw.

This shows that the ethanol extract of the fruit of beans (Phaseolus vulgaris L), with dose variation has an influence in the determination of HDL levels Wistar rat Hypercholesterolemia. In the group of dyslipidemia Wistar rats were given ethanol extract fruit perlakuaan beans (Phaseolus vulgaris L.) with a dose of 150 mg / Kg and perlakuaan medication simvastatin at a dose of 0.18 mg / kg bw overdosed, because it had higher levels of HDL cholesterol on average exceeded HDL cholesterol level of negative control group. This is due to the rich content of phytosterols in fruit ethanol extract of beans (Phaseolus vulgaris L.) that increase HDL cholesterol levels. HDL cholesterol levels increased significantly had the best results at a dose of 150mg/kgbw.

Decrease Malondialdehid level (MDA)

HDL cholesterol levels in the blood are high very beneficial in reducing the risk of atherosclerosis, because the function of HDL cholesterol transports cholesterol from extrahepatic tissues to the liver so menncegah occurrence of calcification (Hartayo 2008). HDL cholesterol assay results are shown in Table 5.

Table 5 : Average of MDA (( μM )

Group Average of MDA ( μM ) Persentation Increase

K1 26,60 -

K2 17,15 -

P1 23,90 39,35%

P2 22,85 33,23%

P3 27,38 59,63%

KS 24,10 40,52%

Figure 4. Average level of Malondialdehid

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

9,000

K1 K2 P1 P2 P3 KS

Average of malondialdehid (μM)



Wahyuni et al. 165


Surabaya Indonesia

In Table 6 at a dose of 150 mg / kg showed a decrease MDA 70.60%, with analysis test results followed by one way ANOVA and Post Hoc Study by Tukey's test / HSD to mengaetahui group that has the same effect or different from the others. ANOVA analysis results which show the value of P <0.05 so H1 H¬0 accepted and rejected, which means there is a significant difference in HDL levels kelommpok control hypercholesterolemia with treatment group after administration of ethanol extract of the fruit of beans (Phaseolus vulgaris L.) with dose 50 mg / kg BW, 100 mg / kg BW, 150 mg / kg body weight and drug delivery simvastatin at a dose of 0.18 mg / kg Bw. Decrease in MDA levels in the blood may prevent decrease in membrane fluidity and cell damage. Based on the research that has been conducted in animals suggest that eating unsaturated fatty acids cause a buildup chain unsaturated fatty acids in the body. Increased toxic products from the synthesis of unsaturated fatty acids (lipid peroxidation) generating malondialdehyde.

CONCLUSSION 1. Consumtion of the ethanol extract of the fruit of beans (Phaseolus vulgaris L.) can reduce total

cholesterol in dyslipidemia Wistar rats at a dose of 150 mg / kg BW of 23.88%. 2. Consumtion of the ethanol extract of the fruit of beans (Phaseolus vulgaris L.) can reduce levels of Low

Density Lipoprotein (LDL) on dyslipidemia Wistar rats at a dose of 150 mg / kg BW 38.09%. 3. Consumtion of the ethanol extract of the fruit of beans (Phaseolus vulgaris L.) can increase the levels of

High Density Lipoprotein on dyslipidemia Wistar rats at a dose of 150 mg / BW 59.63%. 4. extract the ethanol extract of the fruit of beans (Phaseolus vulgaris L.) can decrease malondialdehyde in

dyslipidemia Wistar rats at a dose of 150 mg / BW amounted to 70.60%. 5. The ethanol extract pieces of bean (Phaseolus vulgaris L.) contain phytosterol compounds based analysis

Gas Chromatography - Mass Spectrometry (GC-MS)

REFERENCES 1. Almatsier S. 2009. Prinsip dasar ilmu gizi. Jakarta: gramedia cetakan IX. 2. Bahri, A. 2004. Dislipidemia Sebagai Faktor Resiko Penyakit Jantung Koroner. e-USU Repositor.

Fakultas Kedokteran Universitas Sumatera Utara. http://www.library.usu.ac.id/download/fk/gizibahri3. pdf. Diakses tanggal 17 mei 2015.

3. Bonsdorff-Nikander, A.V. 2005. Studies on Cholesterol-Lowering Microcrystalline Phytosterol Suspension in Oil. Helsinki : Faculty of Pharmacy of the University of Helsinki.

4. Carini.,M,.Aldini,Fcino,R.M.,2004. Mass Spectrometry for Detection of 4-hydroxy-trans-2-nonenal (HNE) Adducts with peptides and Proteins [ Review].Mass.Spectrum Rev,23 :281-305.

5. Hartoyo, A., Dahrulsyah, N., Sripalupi dan Nugroho P. 2008. Pengaruh Fraksi Karbohidrat Kacang Komak (Lablab Purpureus (L) Sweet). Jurnal teknologi dan industri pangan, 19: 25-31.

6. Heslet, L. 1996. Kolesterol. Terjemahan Anton Adiwijoto. Jakarta : PT. Kesaint Blanc Indah. 7. Jannah, H., Sudarma, I.M., Andayani, Y. 2013. Analisis senyawa fitosterol dalam ekstrak buah buncis

(Phaseolus vulgaris L.). Chem, Prog, Vol. 6, No.2. 8. Jones P.J., Raeini-Sarjaz M., Ntanios F.Y., Vanstone C.A., Feng J.Y., Parsons W.E. 2000. Modulation

of plasma lipid levels and cholesterol kinetics by phytosterol versus phytostanol esters. J. Lipid Res., 2000, 41, 697–705.

9. Lamson., Davis., dan Matthew B. 2000. Antioxidants and cancer III. Quercetin, Alternative Medicine, Review Journal 5(3).

10. Montgomery, R. 1993. Biokimia Suatu Pendekatan Berorientasi Kasus. Yogyakarta: Gadjah Mada University Press.

11. Murray, R.K. 2006. Biokimia harper. Jakarta : buku kedokteran EGC. 12. Price dan Wilson. 2006. Patofisiologi Konsep Klinis Proses-Proses Penyakit. Volume 1. Edisi 6.

Jakarta: EGC. 13. Radhika, S., Smila K.H. dan Muthezhilan R. 2011. Antidiabetic and Hypolipidemic Activity of Punica

granatum Linn on Alloxan Induced Rats. World Journal of Medical Sciences 6 (4): 178-182, 2011. 14. Ranti, G.C., Fatimawali, Wehantouw, F. 2013. Uji efektivitas ekstrak flavonoid dan steroid dari gedi

(Abelmoschus manihot) sebagai anti obesitas dan hipolipedemik pada tikus putih jantan galur wistar. Jurnal Ilmiah Farmasi – UNSRAT, Vol.2 No. 02: 2302-249.

15. Soeharto, I. 2004. Jantung Koroner dan Serangan Jantung. Jakarta: Gramedia Pustaka Utama.



Wahyuni et al. 166


Surabaya Indonesia

16. Subekti, S., Piliang, W.G., Manalu, W., Murdiati, T.B. 2006. Penggunaan Tepung Daun Katuk dan Ekstrak daun Katuk (Sauropus androgynus L. Merr) sebagai Substitusi Ransum yang Dapat Menghasilkan produk Puyuh Jepang Rendah Kolesterol. JITV Vol. 11 No.4 th 2006.

17. Uchida.,K.2003. 4-Hydroxy-2-nonenal a Product and Mediator of Oxidative Stress [ Review]. ProgLipid Res 42 : 318 – 43.

18. Utariningsih, D., Novita, W.R., Sari, R.P., Wati, E.M., Arifin, A.S., 2007. Dekok rambut jagung (Zea mays) efektif dalam menurunkan kadar kolesterol tikus putih (Rattus norvegicus). Malang : Universitas Muhammadiyah.

19. Wang T., Hicks K.B., Moreau R., Antioxidant activity of phytosterols, oryzanol, and other phytosterol conjugates. J. Am. Oil Chem. Soc., 2002, 79, 1201–1206.



Wahyuni et al. 167


Surabaya Indonesia

Gandarusa (Justicia gendarussa Burm. f.) Shoots Induction by The

Combination of NAA and BAP

Dwi Kusuma Wahyuni 1a*, Bambang Prajoga 2b, Cheli Maulana 2a, Hery Purnobasuki 3a , Hamidah 4a dan Noer Moehammadi 5a

a Department of Biology, Airlangga University, Surabaya 60115 Indonesia bDepartment of Phytochemistry and Pharmacognocy, Airlangga University, Surabaya 60115 Indonesia

*Dwi Kusuma Wahyuni’s E-mail: [email protected]

Abstract. The aim of this research were to determine the effect of BAP and NAA on gandarusa (Justicia gendarussa Burm. f.) shoot induction. Stem explant cultured on MS medium with additional BAP (0; 1; 2; 3; and 4 ppm) and NAA (0; 0,2; 0,4; and 0,8 ppm). Observation was conducted for 6 weeks to examine explants growth and development, the shoot number, shoot length, leave number, root number, and callus existence. The data were collected qualitatively (stem explant growth and development) and quantitatively (shoot formation time, shoot number, shoot length, leave number, root number, and callus existence). Results showed that there were varieties data on gandarusa nodal explant growth and development. Most of shoots were developed to be a complete plant (shoot, leaf and root).

INTRODUCTION Gandarusa (Justicia gedarussa Burm. f.) is Acanthaceae member. Gandarusa habitat is lowland up to a height

of 500 m above sea level [Handayani, 1]). Gandarusa can be found in Java, Ambon and Ternate [Permawati, 2]. Gandarusa was used traditionally by the people of Papua as a male contraceptive drug [Prajogo et al., 3].

Gandarusa has many benefits, therefore, it needed a method to produce seedlings. One kind of vegetative propagation of plants is in vitro culture [Kartikasari et al., 4]. We can produce many seedlings in a short time by in vitro culture. The success of in vitro culture is determined by many factors, including the suitability of the medium used, hormone, and the source of explant.

Hormone used is a group of auxin and cytokinin [Samudin, 5]. Cytokinin used is Benzyl Amino Purin (BAP) commercially, because it is stable, inexpensive, easy to obtain and more effective. High concentrations of BAP would encourage cells to form shoot [Manurung, 6]. Cytokinin stimulates the growth of the shoots [Wetherell, 7]. Auxin is compounds that affected cell growth. The role of auxin is to stimulate cell division found in plant shoots and causes the growth of new shoots [Andaryani, 8].

Shoot formation depends on a proportion of cytokinin and auxin in the media [Kartikasari et al., 4]. Therefore, this study was conducted to determine the effect of various combinations of BAP and Nafthalene Acetic Acid (NAA) to the gandarusa nodal explant on MS medium. This study is an early stage of a series of studies designed to produce gadarusa plantlet in large quantities and high quality.

MATERIAL AND METHOD

Media and explants

This research was conducted at the Laboratory of Plant Physiology, Faculty of Science and Technology, Airlangga University, Surabaya. The explants used were gandarusa (Justicia gendarussa Burm.f.) nodes, obtained from the Institute of Materia Medica Malang. Murashige and Skoog (MS) (Murashige and Skoog, 1962), were added sucrose 3 %, agar 8g / l , BAP ( 0 ; 1 ; 2 ; 3 ; 4 mg / l ) and NAA ( 0 , 0.2 ; 0.4 ; 0.6 ; 0.8 mg / l ).

Explant Sterilisation

Explants were washed by water and rinsed with running water, then sterilized for 10 minutes in a 10 % solution of sodium hipoclorit ―bayclin‖, and then rinsed with sterilized distilled water 3 to 5 times (this step was done in Laminair Airflow Cabined (LAF)).



Wahyuni et al. 168


Surabaya Indonesia

Explants planting

Sterilized stem which were already taken its nodes were to be planted in the MS solid media. The explants were observed every day until the formation of shoots from the nodes. The shoot formation was characterized by the appearance of small bumps on the light green nodes. Shoot number, shoot length, leave number, root number and callus existence were observed visually, while the length of shoots measured using calipers. These observations carried out at sixth weeks after planting. Each treatment was repeated 3 times.

Data analysis

Data from the morphological characters observation such as the shoot formation time, shoot number, shoot length, leave number and root number were analyzed descriptively.

RESULTS

Shoot Development of Justicia gendarussa Burm. f.

In the early step of shoot formation is growing small protrusions was growing, namely apical meristem and the color was light green (Figure 1). The emergence of shoot from explants showed that the explant regeneration was successfully and would grow up to be plantlets.

Figure 1. Early shoot formation in nodal stem explants Justicia gendarussa Burm , f . tn : bud , nd : nodes , bt : internodes, kl : callus ( bar = 1 cm )

The shoot formation time, shoot number, shoot length, leave number, root number, and callus growth

The shoot formation time data show that the treatment NAA 0.6 ppm was fastest (5th days), while the longest shoot formation time was NAA 0.2 ppm and BAP 2 ppm (16.6th days) (Table 1). The shoot number data show that the treatment NAA 0.2 ppm and BAP 1 ppm ; NAA 0.2 ppm and BAP 2 ppm ; NAA 0.8 ppm and BAP 2 ppm; and NAA 0.2 ppm and BAP 4 ppm were binggest (1.67), while the shoot was not grow on treatment BAP 2 ppm; NAA 0.8 ppm and BAP 1 ppm, and NAA 0.2 ppm and BAP 3 ppm (Table 1). The shoot length data show that the treatment NAA 0.8ppm was longest (8 cm) (Table 1). The leaves number data show that the treatment NAA 0,8ppm was highest (8.33) (Table 1). The root number data show that the treatment NAA 0.8ppm was highest (10) (Table 1). The callus growth data show that callus was presence in allmost treatment, while there was not callus on the treatment NAA 0.6ppm (Table 1).

The nodal explants Justicia gendarussa Burm, f. development at the 6th weeks

The last observations (6th weeks) show there was variation of nodal explant development. Most nodal explant developped into plantlet (stem, leave, and root) (Figure 2 a, e, f, g, i, and l), the nodal explant developped into roots and callus (Figure 2 b, c, j, r, s, x and y), the nodal explants developped into callus (Figure 2 p and q), the nodal explant developped into shoot (Figure 2 d, h, and u), and the nodal explant grew into shoot and callus (Figure 2 m, n, o, t, v, and w).



Wahyuni et al. 169


Surabaya Indonesia

DISCUSSION The right hormone type and appropriate concentrations will stimulate the growth of cultured explants

[George & Sherrington, 10]. In this study, the hormone may provide different impact on development of the explants into shoots, and even formed callus. This is in accordance with the opinion of Andaryani [8] that the hormone can accelerate the shoot formation. Shoots formation indicates the success of regenerating explants.

The shoot formation time data show that the treatment NAA 0,6 ppm and BAP 1 ppm is fastest (5th days). Table 1 shows that the the treatment of without BAP and BAP 1 ppm form shoots faster than the treatment BAP 2 ppm, 3 ppm and 4 ppm. This is contrast to the opinion of Manurung [6] that the shoot formation more rapidly by increasing concentrations of BAP used. This distinction is possible that gandarusa nodal explant has endogenous cytokinin, so it is necessary that a high concentration of endogenous hormone to accelerate the shoot growth.

Table 1 Average shoot formation time, shoot number, shoot length, leave number, root number, and the percentage of callus formation from Justicia gendarussa Burm . f . explants stems

Treat-ment (ppm)

Shoot Formation Time (days)

Shoot number

Shoot lenght (cm)

Leave Number

Root number

Callus (%)

N0B0 7 1 1.83 5.33 1.33 0.33 N0,2B0 7 0.67 0.1 1 3.33 0.33 N0,4B0 9 1 0.67 3.33 5.67 0.67 N0,6B0 6 0.33 0.13 2 6.67 0 N0,8B0 8.3 1.33 8 8.33 10 0.67 N0B1 7 1 0.67 5.33 0 0.67 N0,2B1 8 1.67 0.43 3 3.33 0.33 N0,4B1 16 0.67 0.1 1.67 0 0.67 N0,6B1 5 1 0.73 3.33 0.67 1 N0,8B1 8 0.33 0.067 0.67 3.33 0.67 N0B2 0 0 0 0 0 0.67 N0,2B2 16.67 1.67 0.5 7.33 0.33 0.67 N0,4B2 14 2 0.73 2.67 0.67 1 N0,6B2 11 0.67 0.6 2.67 0.67 0.67 N0,8B2 13.67 1.67 1.4 5 3.67 1 N0B3 19 0.67 0.267 2.67 0 1 N0,2B3 0 0 0 0 0.66 0.67 N0,4B3 17 0.33 0.1 0 4.33 1 N0,6B3 16 0.67 0.2 1 0.33 1 N0,8B3 15 0.33 0.1 1.33 3.67 1 N0B4 15.5 1.33 0.56 2.67 0 0.33 N0,2B4 15 1.67 0.63 2 0 0.33 N0,4B4 12.5 1 0.36 4 0 0.67 N0,6B4 13.5 1 0.13 1.33 0 1 N0,8B4 12 0.67 0.03 1 4.67 1

Noted: N=NAA, B=BAP



Wahyuni et al. 170


Surabaya Indonesia

Figure 2. The development of Justicia gendarussa Burm f. nodal explant at 6th weeks, Bar = 1 cm .

a. N0B0

b. N0,2B0

c. N0,4B0

d. N0,6B0

e. N0,8B0

f. N0B1

g. N0,2B1

h. N0,4B1

i. N0,6B1

j. N0,8B1

k. N0B2

l. N0,2B2

m. N0,4B2

n. N0,6B2

o. N0,8B2

p. N0B3

q. N0,2B3

r. N0,4B3

s. N0,6B3

t. N0,8B3

u. N0B4 v. N0,2B4 w. N0,4B4 x. N0,6B4

y. N0,8B4

Base on observations in media containing BAP, BAP 2 ppm can produce shoot more than BAP 3 or 4 ppm. Shoot number is an important factor to produce plantlet. George and Sherrington [10] also stated that the BAP is cytokinin which plays a role in the formation and multiplication of shoots. The effect of BAP is stronger than the other cytokinin. In this case the BAP 2 ppm is the optimum concentration to produce shoot from gandarusa nodal explants.

The combination of NAA 0.4 ppm and BAP 2ppm is an optimum concentration for the shoot formation from gandarusa nodal explants. Similarly, in study conducted by Manurung [6], the explants were cultured in media containing BAP produce shoots in large quantities, so the media containing BAP produce shoots more than in absence of BAP. In addition Hariyanti et a l. (2004) in Andaryani [8] reported that exogenous auxin influence to inhibit shoot formation.

Shoot length was observed to determine shoot development and growth until the end of the study. The shoot length was measured at sixth weeks after planting. The results showed NAA 0.8 ppm is the best for shoots length parameter (8 cm) (Figure 2 part e). The shoot length of treatment of NAA 0.8 ppm and BAP 1ppm, NAA 0,8ppm and BAP 2ppm, NAA 0,8 ppm and BAP 3ppm, and NAA 0,8 ppm and BAP 4 ppm are 0 to 1.4



Wahyuni et al. 171


Surabaya Indonesia

cm. These results indicate that the higher concentration of BAP does not affect to the speed of shoot growth. Bhojwani and Razdan (1983) in Manurung [6] stated that the higher the concentration of cytokinin, more shoots number obtained, but the growth of each shoot will be hampered.

The shoots formed will grow every day and there are leaves formation. The high leaves number indicates the shoot growth is better. Growth is affected by the nitrogen in the media. The highest leaves number obtained at treatment of NAA 0.8 ppm (8.33). The difference leaves number is not only influenced by the combination of hormone is given the culture medium, but also influenced by endogenous hormone. Endogenous hormone were different in each explant. They will affect the explant growth [Kartikasari et al.,4]. Abidin (1990) in Kartikasari et al. [4] suggest that hormone at a certain level can block the action of endogenous hormones and interfere cell growth and development.

Roots number can be optimized by the absorption of many nutrients contained in the culture medium. Roots is white, thin and sturdy. The root adventive of Jatropha is characteristic by yellowish-white, without roots, thin and sturdy [Andaryani, 8]. The NAA 0.8 ppm is the best for root formation. On media containing more NAA, it produces more root number. According to Samudin [5] research that higher NAA produce roots in large quantities on roots. Similarly, according to Manurung [6] the higher BAP tends to suppress the roots growth because it can inhibit roots initiation and growth.

NAA and BAP affected different growth of nodal explant for parameter of shoots formation time, shoot length, number shoot, leave number, and root number. The results of this study, callus is present. One indicator of growth in vitro culture is the appearance of callus on the explants. In general, callus formed due to wounding or also called tissue wound coverings. Callus formation which is close the wound as a result of the proliferation of stem cells or tissue explants [Dodds and Roberts, 11]. Callus in this study grew in all combinations of concentrations of NAA and BAP.

Callus from nodal explant is white willow green, white and white to yellowish (Figure 2 parts r and t). The good callus for plant regeneration is a green color. According to Fatmawati [12] callus color indicating the presence of chlorophyll in tissue. White callus may indicate that the callus is still in good shape. Besides color, the texture of the callus is one of the markers used to assess the quality of the callus. Texture callus has formed callus friabel texture. This is consistent with the statement that the callus friabel or crumb considered good because it allows the separation into single cells in suspension cultures, besides that it will improve the aeration of oxygen between cells.

Based on the result it is known that the buds can emerge and grow in gandarusa nodal explants. Shoot formation indicates the success of regeneration of nodal explant inoculated on media [Andaryani, 8].

Combination concentration of NAA and BAP N0,4B2 is the best combination for shoot formation, the highest of shoot number (2 shoot). The NAA 0,8 ppm is the best concentration for shoots length (8 cm), leaves number (8.33 leaves), and root number (10 roots).

CONCLUSION There is various development on nodal explant of Justicia gendarussa Burm.f. by BAP and NAA treatment.

The treatment NAA 0.6 ppm is fastest shoot growth (5th day), the treatment NAA 0.2 ppm and BAP 1 ppm ; NAA 0.2 ppm and BAP 2 ppm ; NAA 0.8 ppm and BAP 2 ppm; and NAA 0.2 ppm and BAP 4 ppm are binggest shoot number (1.67), the treatment NAA 0.8ppm is longest shoot (8 cm), the treatment NAA 0,8ppm is highest leaves number (8.33), the treatment NAA 0.8ppm is highest root number (10) and the callus is presence in allmost treatment, while there is not callus on the treatment NAA 0.6ppm.

AKNOWLEDMENT This work was supported by the Ministry of Research, Tecnology and Higher Education, Republic of

Indonesia, under BOPTN funding, 2015. Contract No. 519/UN3/2015

REFERENCES

1. Handayani, L., Pil kontrasepsi laki-laki dengan bahan dasar gandarusa, Majalah Kedokteran Indonesia, 57, pp.279-284, 2007. (Journal)

2. Permawati, M. Karakterisasi ekstrak air daun gandarusa (Justicia gendarussa Burm.f.) dan pengaruhnya terhadap kadar asam urat plasma tikus putih jantan yang diinduksi kalium oksonat. Fakultas Matematika dan Ilmu Pengetahuan Alam, Universitas Indonesia, Depok. 2008. (Thesis)



Wahyuni et al. 172


Surabaya Indonesia

3. Prajogo, B.E.W., Farida,I., Alifia,P.F., dan Jusak, Efek fase air gandarusa (Justicia gendarussa Burm.f.) pada fungsi hati dan fungsi ginjal kelinci jantan (uji toksisitas fase air daun gandarusa sebagai bahan kontrasepsi pria), Veterinaria Medika, 1 (3), 2008. (Journal)

4. Kartikasari, P., Hidayat, M.T., dan Ratnasari, E., Pengaruh Zat Pengatur Tumbuh 2,4-D (2,4-Dichlorophenoxyacetic acid) dan Kinetin (6-Furfurylaminopurine) untuk Pertumbuhan Tunas Eksplan Pucuk Tanaman Jabon (Anthocephalus cadamba Miq. ex Roxb.) secara In Vitro. Jurnal Lentera Bio, 2(1), pp.75-80, 2013. (Journal)

5. Samudin, S., Pengaruh kombinasi auksin sitokinin terhadap pertumbuhan buah naga, Media Litbang Sulteng, 2(1), pp.62-66, 2009. (Journal)

6. Manurung, L.Y.S., Pengaruh auksin (2,4-D) dan sitokinin (BAP) dalam kultur in vitro buah Makasar (Brucea javanica [L.] Merr.), Fakultas Kehutanan, Institut Pertanian Bogor, 2007. (Thesis)

7. Wetherell, D.F., Pengantar Propagasi Tanaman Secara In Vitro. Terjemahan dari Introduction to In vitro Propagation. Penerjemah: Koesoemardiyah. IKIP Semarang Press. Semarang, 1982. (Book)

8. Andaryani, S., Kajian penggunaan berbagai konsentrasi BAP dan 2,4-D terhadap induksi kalus jarak pagar (Jatropha curcas L.) secara in-vitro, Jurusan Agronomi, Fakultas Pertanian, Universitas Sebelas Maret Surakarta, 2010. (Thesis)

9. Murashige,T. & Skoog, F., A revised medium for rapid growth and bioassays with tobacco tissue culture, Physiologi Plantarum, 15, pp.473-497, 1962. (Journal)

10. George, E.F., & Sherrington, P.D., Plant Propagation by Tissue Culture, Exegenetics Limited, England, 1984. (Book)

11. Dodd, J.H., & Robert, L.W., Experiments in Plant Tissue Culture, Cambridge University Pres. England, 1985. (Book)

12. Fatmawati A., Kajian Konsentrasi BAP dan 2,4-D terhadap Induksi Kalus Tanaman Artemisia annua L. secara In Vitro, Fakultas Pertanian UNS, Surakarta, 2008. (Thesis)



Ulya et al. 173


Surabaya Indonesia

Synthesis of β-Ionone from Citrus Aurantifolia as Precursor of

Vitamin A

Indah Rizki Ulya, Rissa Dwi Susanti, Meyta Restu Wigati, Maryam Putri Eradewi, Novia Eka Setyatama, Siti Mariyah Ulfa*

Department of Chemistry, Faculty of Science, Brawijaya University, Malang65145 Indonesia *Corresponding Author’s E-mail: [email protected]

Abstract. Vitamin A is an essential compound for the maintenance of the immune system. Precursor of vitamin A is -ionone, can synthesised from citral and acetone with condensation reaction, then followed by cyclization reaction. In this research, citral is prepared from steam distillation of fresh lime peel Citrus aurantifolia during 2,5 hours. From this process, 0,54 g of crude essential oil is obtained, and containing 10,25% of citral A and 6,11% citral B (ratio area in GCMS). Condensation reaction of citral and acetone is carried out in autoclave using 1,522 g (0,01 mol) of citral A, 1,16 g acetone (0,02 mol), and 0,05 g CaO as catalyst in methanol 5 mL at 80oC for 4, 6, 8, and 10 hours. Reaction for 10 h give pseudoionone (PI) as major product (43.25% yield). Cyclization of PI into β-ionone is perform by acid catalytic reaction using H2SO4 as catalyst. Ratio of pseudoionone : H2SO4is 1:1; 1:5; 1:7, and 1:9 (mol ratio) and reacted at -5oC by stirring the mixture for thirty minutes.The reaction in 1:1 ratio didn‘t give any β-ionone. Increasing mol ratio of H2SO4 (1:7) leads to formation -ionone 100%. Ratio 1:9 gives isomerization products as α-ionone (0.46% yield).

INTRODUCTION Regulation of the Minister of Industry of the Republic of Indonesia Number: 87 / MIND / PER / 12/2013 on

the implementation of the Indonesian National Standard (SNI) cooking oil palm mention that the palm cooking oil shall be added vitamin A known as fortification of vitamin A. Vitamin A is added is the minimum level of 40 IU (units of vitamin). The ministerial regulation on vitamin A fortification is excluded because many of Indonesia's population are unknowingly deficient in vitamin A, especially for pregnant women. Vitamin A deficiency in pregnant women will have a direct impact on the baby.

During this time, vitamin A used in Indonesia is imported from Germany. If the need for vitamin A fortification is also pursued by import, the cost required to reach 200 billion / day or 72 trillion / year. Synthetically, vitamin A, which the German BASF production produced from citral A synthesized from L-menthol and D-menthol [1]. Naturally, citral A can be obtained from natural materials that are widely available in Indonesia, for example, lemon containing 7,74% citral A [2], lime contains citral B amounted to 9.88% and citral A 12.26 % [3]. Other plant species is Citrus latifolia tanaka (9.7%) [4]; lemongrass; Cymbopogon flexuosus; Cymbopogon martini; Cymbopogon winterianus. Among some of the plants on which could potentially generate enough citral is the citrus crop.

Orange is one of the tropical plant that has spread throughout citrus Indonesia.Species containing citral A sizeable is lime. Lime is a plant that can live and bear fruit throughout the year. The area of Kuningan, West Java, lime is cultivated and processed into syrup for easy consumption. With the scale of production of 5 tonnes / day resulted in the emergence of the magnitude of the problem of waste orange peel[9]. According Nallely [3], in lime peel contained citral A as much as 12.26%. Therefore, vitamin A synthesis problem can be overcome by isolation citral of waste orange peel. Innovation is offered through the synthesis of b-ionon which are precursors of vitamin A by the condensation reaction and cyclization of citral A using selective catalyst. Production masse of b-ionon, the availability of vitamin A in the country will be fulfilled, thereby reducing the cost of exports and solve the problem of vitamin A deficiency.

EXPERIMENTAL The tools used are Spectrophotometer UV-Vis double beam Shimadzu 1600, spectrophotometer FT-IR

Shimadzu 8400S, Gas Chromatography Mass Spectroscopy (GC-MS): GC2010 MSQP 2010S Shimadzu, NMR Spectroscopy, rotary evaporator IKA, set steam distillation apparatus, a set of glasses, Autoclave and analytical balance Ohaus Precision Advanced. The materials used are 95-98% H2SO4, n-hexane, acetone, CaO, citral A, diethyl ether, and lime.

Isolation Citral A of Lime peel




Ulya et al. 174


Surabaya Indonesia

200 gram lime peel sliced into pieces. Pieces of lime peel purified by steam distillation for 5 hours. Essential oil and water was transferred in to separating funnel. Diethyl ether was added to the distillate for liquid-liquid extraction. Extraction results obtained organic phase is a solution of diethyl ether-volatile oil. Na2SO4 anhydrous was added to organic phase. Liquid phase pipetted and incorporated into the vial bottle, and then diethyl ether evaporated at room temperature for one day, then the pure essential oils found in the analysis by GC-MS to determine percent yield and structure of the isolated compounds based on their molecular weight.

Synthesis Pseudoionone from citral A and acetone with catalyst CaO

Aldol condensation reaction is carried out by reacting acetone and citral A with mole ratio of 1:1. 1 mole of acetone is reacted with 1 mole of citral A, and 0.0538 g CaO of mass quantities of acetone and citral A[10]. The reaction carried out in a stainless steel reactor at a temperature of 60-90°C with variations in time of 4 hours, 6 hours, 8 hours, and 10 hours. Pseudoionon analyzed using UV-Vis spectrophotometry and FTIR spectrophotometry to determine the functional groups.

Synthesis β-ionone with variation of concentration sulfatic acid

1 mole of solution Pseudoionon is added concentration sulfatic acid at variations moles with ratio of 1: 5; 1: 7; and 1: 9 sulfatic acid. Sulfatic acid added in to pseudoionone little by little while stirring to maintain the temperature -4ºC. Once the addition is complete (30 minutes), the mixture allowed to warm to 0ºC, and poured into 100 mL of ice water. A layer of liquid phase is separated and extracted with 2x20 ml of n-hexane. The organic layer combined and added 4 M solution of sodium hidroxyde to neutral. After drying and evaporation of the solvent, the product obtained -ionon. The products obtained were analyzed by IR spectrophotometry and GC-MS. IR spectrophotometry is used to determine the functional groups in the molecule. GC-MS was used to determine polafragmentasi compounds and the percent purity of the resulting product.

RESULTS AND DISCUSSIONS

Isolation citral A of lime peel

Isolation citral A with steam distillation using fresh lime peel and lime peel dried as in Table 1. The results obtained showed that the essential oil from the distillation of fresh orange peel more than dried orange peel. The resulting yield of as much as 0.27% of 200 gram of lime peel. Form essential oil obtained in the form of a colorless liquid and distinctive smell of orange oil and citral A standard.

Tabel 1. Yield essential oil from isolation citral A of lime peel No Waktu pengeringan kulit jeruk nipis Rendemen 1 2 minggu 0% 2 3 hari 0% 3 Semalam 0% 4 Tanpa Pengeringan 0,27%

Characterization of citral A

UV-Vis Spectrophotometer

UV-Vis spectrophotometry test to compare the maximum wavelength (max) and essential oil isolated citral A standard fit Fig 1. UV-Vis analysis results showed max = 235.5 nm for essential oils isolation and max = 238.5 nm for citral A. This suggests that the essential oil compounds are citral A. insulation calculation results max citral A theoretical use calculations based on Silverstein et al [11], max = 257 nm which is close to the value citral A result of isolation.



Ulya et al. 175


Surabaya Indonesia

Fig. 1. Spectrum UV-Vis of Essential Oil

FTIR Spectrophotometer

FTIR spectrophotometry test to determine the specific functional groups on essential oil isolated. IR spectrum of essential oils and citral insulation standards presented in Fig. 2 with identification of groups as shown in Table 2 below:

Tabel 2. Identification of functional groups on the volatile oil and citral A

Gugus fungsi Bilangan gelombang (cm-1)

Citral A Minyak atsiri isolasi

Aldehid 2900-2700 2725 Karbonil (C=O) 1650 1650

CH2 bending 1450 1379

CH3 bending - 1677-1664

C=C terisolasi 1500 1438 C-H sp2 - 3072 C-H sp3 2968 2921

Fig. 2. Spectrum FT-IR of Essential Oil

0

0.5

1

1.5

2

200 400 600 800

Ab

s

Wavelength (nm)

minyak atsiri

citral A

5007501000125015001750200025003000350040001/cm

10

20

30

40

50

60

70

80

90

100

%T

Smoothcitral A

minyak atsiri



Ulya et al. 176


Surabaya Indonesia

1H-NMR 1H-NMR analysis to determine the structure of compounds contained in the isolated lime peel, then the

resulting spectrum compared to a spectrum citral A standard. Based on the results of 1H-NMR analysis as shown in fig. 3, it is known that the chemical shift between citral A standard (black color spectrum) and citral isolated (red color spectrum) is the same. So it may be indicated that the essential oil which is isolated from the process of steam distillation citral-containing compounds.

Fig. 3. Spectrum NMR of Essential Oil

GC-MS

GC-MS analysis is used to separate the constituent components of essential oils based on the affinity of each compound to the polysiloxane column. The structure of the compound is determined from the resulting fragmentation pattern of Mass Spectroscopy (MS) according to Fig. 4. Based on data from the NIST library, known citral THAT A is a component of the No. 11 peak with a retention time of 18 minutes and the extent of the area of 10.25%. A purity citral obtained was lower than the results Nallely [3], which states in lime peel contained citral A as much as 12.26%. This is because of the time difference steam distillation performed.

Fig. 4. Spectrum GC-MS of Essential Oil

Synthesis Pseudoionone from citral A and acetone with catalyst CaO

Synthesis pseudoionone from citral A and acetone by aldol condensation reaction. This reaction is carried out by reacting 1.522 g of citral; 1.16 g of acetone and 0.05 g CaO with variations in reaction time 4 hours, 6 hours, 8 hours, and 10 hours at a temperature of 60-90ºC. The reaction results are presented in Table 3.



Ulya et al. 177


Surabaya Indonesia

Table 3. The yield of the products produced in the synthesis pseudoionon

No Waktu yang digunakan saat reaksi Rendemen 1 4 jam 98,14 % 2 6 jam 37,55 % 3 8 jam 46,84 % 4 10 jam 64,31 %

From Table 3, it is known the longer the reaction time the greater the yield obtained. Deviations is the

reaction time of 4 hours produced yield of 98.14%. This is because there are many unreacted citral forming pseudoionon on reaction time 4 hours, according to the results of GC-MS characterization.

Characterization of Pseudoionone

UV-Vis Spectrophotometer

UV-Vis spectrophotometer test to identify the maximum wavelength pseudoionone synthesized. Measurements were performed in ethanol, produced a spectrum as shown in Fig. 5. UV-Vis analysis results showed that there are three maximum wavelength, there are 205.8 nm, 236.6 nm and 292.85 nm. 205.8 nm wavelength identified as citral A, which has not undergone a condensation reaction, while the wavelength 236.6 nm and 292.85 nm showed the product pseudoionone. Based on theoretical calculations obtained wavelength pseudoionone calculated theoretically is 281 nm.

Fig. 5. Spectrum UV-Vis of Pseudoionone

FTIR Spectrophotometer

FTIR spektrofotemetri test to characterize the results to see what groups are contained it. The reaction between citral A, acetone and CaO catalyst in general is as follows:

0

0.2

0.4

0.6

0.8

1

1.2

1.4

200 250 300 350 400 450

Ab

sorb

ance

Wavelength (nm)

4 jam

6 jam

8 jam

10 jam

citral A



Ulya et al. 178


Surabaya Indonesia

Based on the results of FTIR analysis, which identified functional groups of pseudoionone as shown in Fig. 6 with identification of groups as shown in Table 4. The presence of a primary alcohol groups is indicated methanol remaining due to this reaction using methanol.

Table 6 Table 4. Characterization FTIR of pseudoionone

Gugus fungsi Bilangan gelombang (cm-1)

Citral A Pseudoionon

-OH - 3500-3200 Aldehid 2900-2700

Karbonil (C=O) 1650 1650

CH2 bending 1450 -

Alkohol primer - 1100

C=C terisolasi 1500 1456 C-O - 1650

C-H sp3 2927 2921

Table 7 Fig. 6. Spectrum FT-IR of Pseudoionone with variation time of synthesis

1H-NMR 1H-NMR analysis to determine the structure of compounds contained in the results of the synthesis of citral

A with acetone fit Fig. 7. The analysis shows the existence of a shift or chemical additions from peak to peak pseudoionon citral A, there is a new peak at 3.2 ppm which indicates the addition of hydrogen to the intensity of the 4 hydrogen, with the addition of this hirogen may indicate the formation of pseudoionon.

5007501000125015001750200025003000350040001/cm

10

20

30

40

50

60

70

80

90

100

%T

citral Aps 10 jamps 8 jam

Citral A



Ulya et al. 179


Surabaya Indonesia

Table 8 Fig. 7. Spectrum NMR of Pseudoionone

GC-MS

GC-MS analysis to determine the purity of the product pseudoionon synthesized in accordance with Fig. 8. Based on data from the NIST library, pseudoionon is known that the peak of No. 9 with a retention time of 22.9 minutes and the extent of 48.07%, which means the area pseudoionon on product purity is 48.07%, with% 43.25% yield obtained against The theoretical mass.

Fig. 8. Spectrum GC-MS of Pseudoionone

Synthesis β-ionone with variations concentration of Sulfic Acid

Synthesis of β-ionone are using an acid catalyst through pseudoionone cyclization reaction with sulfic acid.

Variations ratios of concentrations pseudoionone and sulfic acid are presented in Table 5. β-ionon resulting brownish red liquid.

Table 5. Cyclization reaction pseudoionon into β-ionon

No Pseudoionon : H2SO4 Massa Hasil 1 1:5 (0,05 mol : 0,25 mol) 0,39 gram 2 1:7 (0,05 mol : 0,35 mol) 0,04 gram 3 1:9 (0,05 mol : 0,45 mol) 0,78 gram

Characterization of β-ionone

Characterization of β-ionon synthesized using chromotagraphy Gas Mass Spectrophotometer (GC-MS) to determine the purity of β-ionon generated. Based on the data in the appropriate NIST library Fig.9 which is the result of β-ionon characterization of the results of the synthesis on the concentration ratio pseudoionon : sulfic acid 1: 5 or pseudoionon 0.05 mol and sulfic acid 0:25 mol peak No. 11 with a retention time of 21.609 minutes and cotton area of 85.88%, which means β-ionon purity is 85.88% with yield 33.85%% of the theoretical mass.



Ulya et al. 180


Surabaya Indonesia

For the synthesis of the concentration ratio pseudoionon : sulfic acid 1: 7 or pseudoionon 0.05 mol and 0.35 mol sulfic acid in conformity with Fig. 10 shows peaks only one that has particularly cotton area of 100% with a retention time of 21.614 minutes, which means purity β-ionon is 100%. As for the results of the synthesis on the concentration ratio pseudoionon : sulfic acid 1: 9 or 0.05 mol pseudoionon and 0.45 mol sulfic acid according to Fig. 11 shows the No. 3 peak with a retention time of 21.597 minutes and the extent of 97.99% area, which means purity β-ionon stretcher is 97.99%, the ratio of the concentration of sulfic acid is used a lot, so that its H atom excess that eventually form new compounds isomerization be shown at the top No. 2 with a retention time of 20.7 minutes with an area of this peak area 1:10% meurut NIST library are compounds α-ionon. From the analyst GC-MS showed that the optimum condition that is the ratio of the concentration of sulfic acid sulfic acid pseudoionon with 1: 7, because the comparison produces β-ionon with a purity of 100%.

Fig. 9 Spectrum GC-MS β-ionon with ratio concentration Pseudoionone : Sulfic Acid 1:5



CONCLUSION From this research was found 0.54 gram of essential oils isolated from 200 gram of lime peel waste

containing 10.25% citral A. CaO catalyst activity to form enolate ion in acetone which is then reacted with a citral be evidenced pseudoionon and the result is pseudoionon with most of the reaction for 10 hours is 1.73 gram containing pseudoionon as much as 48.07%. The optimum conditions sulfic acid is at concentration pseudoionon comparison with sulfic acid 1: 7, because the comparison is produced 100% β-ionon.

REFERENCE 1. Jespersen,J. 2009.Citrus, rose and lavender – nature-identical fragrances from the chemicallaboratory

[Internet]. Germany: BASF The Chemical Company. Tersedia dalam:<www.basf.de/science_around_us> [Diakses 17 September 2014].

http://www.basf.de/science_around_us

http://www.basf.de/science_around_us



Ulya et al. 181


Surabaya Indonesia

2. Jomaa, S., Rahmo A., Alnori, Ahmad S., Chatty, and Mohammad E. 2012. The CytotoxicEffect of Essential Oil of Syrian Citrus limon Peel on Human Colorectal CarcinomaCell Line (Lim1863). Middle East Journal of Cancer. 3 (1). pp.15-21.

3. Nallely E., Sandoval-M., Abraham G., Elizabeth E., Elvira G., Laura A., and María del R.2012. Chemical Composition of Hexane Extract of Citrus aurantifolia and AntiMycobacterium tuberculosis. Molecules.17. pp.11173-11184.

4. Atti-Santos, A.C., Marcelo R., Luciana A.S., Eduardo C. and Patrick M. 2005. Extraction ofEssential Oils from Lime (Citrus latifolia Tanaka) by Hydrodistillation andSupercritical CarbonDioxide. Brazilian Archives of Biology and Technology. 48.pp.155-160.

5. Gunawan, Iwan Reny Yuniawati, dan Rani Andriani Koswara. 2010. Sentra Bisnis JawaBarat. TRANSMEDIA : Jakarta.

6. Noda, C., Alt G., Werneck R., Henriques C., and Monteiro J. (1998). Aldol Condensation Of Citral With Acetone On Basic Solid Catalysts. Braz. J. Chem. Eng. Vol. 15 No. 2 São Paulo June.

7. Silverstein, Robert M., Francis X. W., dan David J.K., 2005, Spectrometric Identification of Organic Compounds Seventh Edition, John Wiley & Son, Inc, USA.



Wahyuningsih et al. 182


Surabaya Indonesia

Polysaccharide Krestin Activity of Coriolus versicolor Extract on

Interleukin-12 Level of Mus muculus Exposed to Mycobacterium

tuberculosis

Sri Puji Astuti Wahyuningsih1*, Sugiharto1, Nurul Wiqoyah2 & Zakiyatun Nuha1

1Depart. of Biology, Faculty of Science and Technology, Airlangga University, Surabaya 60115 Indonesia 2 Microbiology Laboratoty, Faculty of Medicine, Airlangga University, Surabaya 60115 Indonesia

Corresponding author’s Email: [email protected] [email protected]

Abstract. This study aimed to determine the effect of polysaccharide krestin (PSK) of Coriolus versicolor‘s extract on IL-12 levels in the Mus musculus‘s blood serum that had been exposed by Mycobacterium tuberculosis. This study used a completely randomized design were divided into six treatment groups. They were K (control group without treatment), K+ (positive control by providing PSK), K- (negative control with exposed M. tuberculosis), P1 (treatment using PSK before exposured by M. tuberculosis, P2 (treatment using PSK after exposured by M. tuberculosis), and P3 (treatment using PSK before and after exposured by M. tuberculosis). PSK given concentration was 200 mg/kg bw, while the number of bacteria for exposure was 0.25 Mc Farland with double exposure. Each treatment there were four replicates. Blood serum of mice were isolated and measured levels of IL-12 by ELISA kit. Data analysis used Brown Forsythe test. The results showed that the highest levels of IL-4 was K-. PSK administration in the treatment of P1, P2, and P3 showed the levels of IL-12 not significant with the K and K +. Conclusion of the study was the treatment using PSK had no effect on the level of IL-12 in Mus muculus exposed by M. tuberculosis.

INTRODUCTION Tuberculosis or TB is an infectious bacterial disease caused by Mycobacterium tuberculosis, which most

commonly affects the lungs. It is transmitted from person to person via droplets from the throat and lungs of people with the active respiratory disease. The symptoms of active TB of the lung are coughing, sometimes with sputum or blood, chest pains, weakness, weight loss, fever and night sweats [Departemen Kesehatan RI, 7].

In healthy people, infection with M. tuberculosis Often causes no symptoms, since the person's immune system acts to "wall off" the bacteria. Tuberculosis can be treated with antibiotics for 6 months in a row. It is important to look for other alternative materials that can be used to enhance the immune response to tuberculosis.

Medicinal mushrooms have an established history of use in traditional oriental therapies. Modern clinical practice in Japan, China, Korea, and other Asian countries continues to rely on mushroom. Mushrooms effects have been demonstrated for many including extracts of species from C. versicolor [Ooi & Liu, 14].

It is well established that many mushroom-extracted compounds are commonly used as immunomodulators or as Biological Response Modifiers (BRM). The basic strategy underlying immunomodulation is to identify aspects of the host response that can be enhanced or suppressed in such a way as to augment or complement a desired immune response. Whether certain compounds enhance or suppress immune responses depends on a number of factors, including dose, route of administration, timing of administration of the compound, mechanism of action, and site of activity. Knowledge of the specific components of cytokine networks and signaling pathways and their role in the regulation of immune responses is important in designing strategies to augment these responses.

C. versicolor extract can increase the number of leukocytes, macrophages and spleen weight [Wahyuningsih, 20], the provision of PSK from C. versicolor increase the number of immunocompetent cells, increase immune response non-specific and specific due to infection with M. tuberculosis (Wahyuningsih et al., 21). The active compounds contained in mushrooms is β-glucan [Guggenheim, 9; Moradali et al., 13]. β-Glucan is known to stimulate the formation of pro-inflammatory mediators such as complement

components, interleukin 1 (IL-1), tumor necrosis factor (TNF-α), interleukin 2 (IL-2) and eicosanoids [Yu et al., 27]. β-Glucan increases the production of IL-2, which stimulates the differentiation of B cells activated [Vetvicka et al., 24]. In this study, administration of PSK done in three different time ie





Surabaya Indonesia

before exposure, after exposure and before and after exposure to M. tuberculosis. Giving PSK before exposure to M. tuberculosis can serve as a preventative that will encourage the formation of antibodies so the body does not easily infected. Giving PSK after exposure was conducted to determine the effectiveness of the PSK in eliminating acid-resistant bacteria. Giving PSK before and after exposure to M. tuberculosis as a preventive and curative efforts that encourage the formation of antibodies [Wahyuningsih et al., 22]. Pietro [15], β-glucan is effective for the prevention and treatment of diseases related to the resilience of the body's immune system. Β-glucan compounds of PSK related to the primary receptor of the immune system that is Dectin-1, Toll-

Like Receptor 2/6 (TLR-2/6) and Complement Receptor 3 (CR3). Dectin-1 known to play a role as a co-stimulator of T cells and is the receptor that is expressed by macrophages [Willment, 23]. In addition Dectin-1 can work with TLRs to induce and increase the production of cytokines [Gantner et al., 8). Interleukin-12 binding to its receptor activates the tyrosine kinase 2 (Tyk2) and janus kinase 2 (JAK2). This causes the phosphorylation of tyrosine residues of the signal transducer and activator of transcription 3 and 4 (STAT3 and STAT4). The occurrence of tyrosine phosphorylation process is responsible for the formation of STAT4 / STAT4 homodimer and STAT3 / STAT4 heterodimers. Both dimers are experiencing translocation into the nucleus and binds to the IFN-γ gene promoter [Thierfelder et al., 19]. Promoter activation in the IFN-γ gene resulted in the synthesis of proteins that express IFN-γ protein.

Along with the growing understanding of the body's immune response in the face of infection, the more developed also research into components that can affect the immune response. If PSK can increase the levels of IL-12 in the blood, the immune system can increase too. If it is proven that extracts of C. versicolor can be used as a natural immunostimulatory to inhibit the growth of M. tuberculosis and reduce the number of TB patients. Active TB disease is estimated to have reached 1.9 billion, and it can prevent transmission because only people with active TB can transmit the disease.

MATERIALS AND METHODS

Preparation research

Thirty female mice aged 8-10 weeks strain Balb/C acclimatized for one week and are grouped into six, namely K (only given distilled water), K- (exposure to M. tuberculosis only), K + (given PSK only), P1 (PSK was given before exposure to M. tuberculosis), P2 (PSK administered after exposure to M. tuberculosis), P3 (PSK given before and after exposure to M. tuberculosis). Feed and water provided ad libitum.

Isolation and Measurement Levels PSK

Coriolus versicolor were collected from Kediri, Tulungagung and Surabaya. Fungi identified, wind dried and pulverized to form a coarse powder. Then it was extracted by the method of Cui & Cristi [5] and Cui et al. [6] as modified by Wahyuningsih et al. [20] PSK levels determined by the method of phenol-sulfuric acid assay. PSK dose used was 200 mg/Kg body weigth.

Treatment research

Giving PSK done for 7 days. Exposure to M. tuberculosis performed twice with an interval of two weeks via intraperitoneal. The number of bacteria is 0.5 Mc Farland.

Isolation of serum

Blood was drawn through intracardia. Blood left in a tilted position at room temperature for three hours to form two phases, namely the colored translucent top and the bottom is red. Then, it centrifuged 3000 rpm, 10 min at 4 ° C. Serum or supernatant portion was taken.

Immunosrbent-Like Enzyme Assay (ELISA)

Detection of IL-12 used to a sandwich ELISA (Koma Biotech ELISA Kit IL-12). Levels of IL-12 was read by ELISA reader at a wavelength of 450 nm.





Surabaya Indonesia

Data analysis

Data were tested with the Kolmogorov-Smirnov test and homogeneity of variance. Then analyzed by Brown Forsythe at the level of 5% (α = 0.05).

RESULT AND DISSCUSSION Levels of Interleukin 12 in the blood serum is a picture of an immune response as a result of

exposure to M. tuberculosis were shown in Table 1. The results of the ELISA is a yellow color reaction. This color is obtained after a complete bond between the antigen, primary antibody and enzyme-labeled secondary antibody.

The results showed that average levels of IL-12 in the control group (K) was 25.91 ± 11.08 (pg / mL). Positive control group (K +) was 26.37 ± 11.49 (pg / mL). Negative control group (K) was 37.25 ± 27.73 (pg / mL). P1 group was 31.85 ± 16.22 (pg / mL). P2 group was 31.43 ± 10.47 (pg / mL). P3 group was 26.07 ± 10.99 (pg / mL). Kolmogorov-Smirnov test showed a significance level of 0.068. This indicates that the data were normally distributed. Homogeneity test showed a significance level of 0,010. This indicates that the data has a variance that is not homogeneous. Data were analyzed by Brown Forsythe test to determine the effect of treatment in each group. The test results showed a significance level of 0.828. This shows that there was no effect of feeding time on levels of IL-12 in mice blood serum as a result of exposure to M. tuberculosis.

Table 1 Levels of IL-12 after administration of PSK from C. versicolor in mice exposed to M. tuberculosis

In the study used an indicator to determine the levels of cytokine immune response to infection with M. tuberculosis and also the effect of PSK. Cytokines are small proteins as a mediator and regulator of immunity, inflammation and hematopoesis. The cytokines produced in response to the stimulus of the immune system. Cytokines works by binding to specific membrane receptors, which then carry signals into the cell via second messengers (tyrosine kinase), to alter gene expression (Judarwanto, 11). Cytokines produced in the body when the immune response as a result of infectious microorganisms or foreign substances that enter the body (Romagnani, 16). Type of cytokine that is used as an indicator in this study is the production of Th1 cytokines, namely IL-12. Interleukin is a cytokine that acts against leukocytes, while interferon is a type of cytokine that plays a role due to interference virus (Lewis, 12).

Tuberculosis is a disease caused by intracellular bacterial pathogens, namely M. tuberculosis (Higuchi et al., 10). One type of cytokine that plays an important role in the defense against intracellular pathogens such as M. tuberculosis is IL-12. APC cells produce IL-12 when stimulated by lypotechoat acid and peptidoglycan. Two components are bacterial cell wall constituent. Wall of M. tuberculosis can induce macrophages to produce IL-12 (Yoshida & Koide, 26).

In this study, M. tuberculosis acts as an antigen. The bacteria will induce the secretion of IL-12. PSK act as immunostimulatory which is expected to enhance the immune response due to infection with M. tuberculosis. Polysaccharides krestin from C. versicolor extract known to contain β-glucan compounds. The compounds may be associated with the primary receptor of the immune system such as Dectin-1, Toll-Like Receptor 2/6 (TLR-2/6) and Complement Receptor 3 (CR3). Dectin-1 acts as a co-stimulator of T cells and is the receptor that is expressed by macrophages (Willment, 23).

In Table 1 it can be seen that the levels of IL-12 between the control and treatment groups showed no significant difference. However, if compared to K, then the levels of IL-12 on K +, K-, P1, P2, and P3 showed an increase. Levels of IL-12 highest detected in the K group treated only exposure to M. tuberculosis. It shows

Treatment Levels of IFN-γ on replay to ..... Mean ± SD

(pg/mL) 1 2 3 4

K 24.90 30.75 27.32 20.68 25.91a ±11.08

K+ 30.44 31.07 23.28 20.68 26.37a ±11.49

K- 81.88 26.90 19.84 20.36 37.25a ±27.73

P1 33.57 48.90 23.04 21.89 31.85a ±16.22

P2 24.77 32.05 45.97 22.93 31.43a ±10.47

P3 21.22 26.90 29.66 26.48 26.07a ±10.99





Surabaya Indonesia

that there are components of antigens on the cell wall of bacteria that can enhance the immune response. According to Crick et al. (4), TB bacterial cell wall composed of mannose-capped lipoarabinomannan (ManLAM), lipomannan (LM), phosphatidyl-myo-inositol mannosides (PIMs), arabinomannan, mannan and Manno-glycoproteins. It is a strong antigen to induce an immune response. Through TLRs, M. tuberculosis cell wall components associated with lipoprotein so as to induce the production of IL-12 (Brightbill et al., 2).

PSK was known to induce the body's immune response, but the strength of induction was weaker when compared to components of M. tuberculosis antigens. It can be seen from Table 1, which showed that levels of IL-12 to K + is higher when compared with K. Levels of IL-12 in the group P1, P2, and P3 showed lower than K-. This was probably related to the role of β-glucan. Glucan can bind directly to specific receptors of immune cells causing immunomodulating effects (Vos et al., 25). This means it can increase or decrease the immune response. But the mechanism of action of β-glucan as immunomodulators, especially in M. tuberculosis infection is not known.

Statistically, the provision PSK does not affect the levels of IL-12. Most likely behind it was the existence of mechanisms mediated cascade of other cytokines stimulated due to the antigen of M. tuberculosis. Another possibility was happening related to the nature of cytokine that a type of cytokine can be produced by various types of cells and can cause different effects through various channels (Soeroso, 17). Various types of cytokine sometimes also have the same function for one type of cells (Abbas et al., 1). The redundancy properties that make it difficult to know the mechanism of action of one single type of cytokine. β-Glucan contained in PSK of C. versicolor known to play a role as an immunomodulator that acts on the

immune system of specific and non-specific. In the non-specific immune system, β-glucan helps to recognize the antigen and respond faster. Dectin-1 is a receptor type II transmembrane protein, can bind to β-1,3 and β-1,6 glucan that can do the initiation and regulation of the non-specific immune response (Brown et al., 3). Dectin-1 is expressed by immune cells that play a role in non-specific immune system and have been found on macrophages, neutrophils, and dendritic cells (Taylor et al., 18). In the specific immune response, Dectin-1 is able to activate T cells and induces the secretion of cytokines (Taylor et al., 18). According to Gantner et al. (8), Dectin-1 is able to induce and increase the secretion of cytokines through TLRs. In addition, the β-glucan known to be effective for the prevention and treatment of diseases associated with immune system resistance (Pietro, 15).

CONCLUSIONS The conclusion from this study is the provision PSK dose of 50 mg/kg body weight had no effect on levels

of IL-12 in mice blood serum as a result of exposure to M. tuberculosis. Further studies, using other indicators that can be known immune mechanisms of TB patients by administration of PSK from C. versicolor extracts.

REFERENCES [1] Abbas, A.K., Lichtman, A.H., & Pober, J.S., Cytokine in cellular and molecular immunology, WB

Sounders Co, Philladelphia,1994. [2] Brightbill, H.D., Libraty, D.H., Krutzik, S.R., Yang, R.B., Belisle, J.T., Bleharski, J.R., Maitland, M.,

Norgard, M.V., Plevy, S. E., Smale, S.T., Brennan, P. J., Bloom, B.R., Godowski, P.J., & Modlin, R.L., Host defense mechanisms triggered by microbial lipoproteins through toll-like receptors, Science, 285:732–736, 1999.

[3] Brown, G.D., Herre, J., Williams, D.L., Willment, J.A., Marshall, A.S., & Gordon, S., Dectin-1 mediates the biological effects of beta-glucans, The journal of experimental medicine, 197(9): 1119-1124, 2003.

[4] Crick, D.C., Brennan, P.J., & McNeil, M.R., The cell wall of Mycobacterium tuberculosis, Lippincott Williams and Wilkins, Philadelphia, 2003.

[5] Cui, J. & Chisti, Y., Polysaccharopeptides of Coriolus versicolor: physiological activity, uses, and production, Biotechnology advances, 21 (2):109-122, 2003,.

[6] Cui, J., Goh, K.K.T., & Archer, R., Characterisation and bioactivity of protein-bound polysaccharides from submerged-culture fermentation of Coriolus versicolor Wr-74 and ATCC-20545 strains, Journal Industrial Microbiology Biotechnology, 34: 393–402, 2007.

[7] Departemen Kesehatan Republik Indonesia, TBC, Pusat Komunikasi Publik Setjen Depkes RI, Jakarta, 2008.

[8] Gantner, B.N., Simmons, R.M., Canavera, S.J., Akira, S., & Underhill, D.M., Collaborative induction of inflammatory responses by dectin-1 and Toll-like receptor 2, Journal of experimental medicine, 197:1107–1117, 2003.





Surabaya Indonesia

[9] Guggenheim, A.G., Wright, K.M., & Zwickey, H.L., Immune modulation from five major mushrooms: Application to integrative oncology, Integrative medicine, 13: 32, 2014.

[10] Higuchi, K., Sekia, Y., & Harada, N., Characterization of M. Tuberculosis-derived IL-12-inducing materialby alveolar macropages, Vaccine, 22: 724-734, 2003.

[11]Judarwanto, W., Imunologi dasar: Sitokin dan aspek klinisnya, http://allergycliniconline.com, 2012.

[12] Lewis, R., Cytokine, http://www.the-scientist.com (14 Januari 2015). [13] Moradali, M.F., Mostafavi, H., Ghods, S., & Hedjaroude, G.A., Immunomodulating and anticancer

agents in the realm of macromycetes fungi (macrofungi). Int immunopharmacology. 7: 701-724, 2007,. [14] Ooi, V.E & F. Liu, Immunomodulation and Anticancer activity of Polysaccharide - Protein Complexes,

National Library of Medicen, CurrMedChem, http://www.ncbi.nlm.nih.gov/pubmed/10702635 (4 Mei 2009)

[15] Pietro, P. , Composition for preventif and/or treatment of lipid metabolism disorders and allergic form. http://freepatent online.com. (20 April 2014).

[15] Romagnani, S., Lymphokine production by human T cells in disease states. Annu review immunology, 12: 227-257, 1994.

[17] Soeroso, A., Sitokin, Jurnal oftalmologi Indonesia, 5: 171-180, 2007. [18] Taylor, P.R., Brown, G.D, Harre, J., Williams, D.S., Willment, J.A., & Gordon, S., The role of SIGNR I

and the beta-glucan receptor (dectin-I) in the nonopsonic recognition of yeast by specific macropages, Journal immunology, 172(2): 1157-1162, 2004.

[19] Thierfelder WE, van Deursen JM, Yamamoto K, Tripp RA, Sarawar SR, Carson RT, Sangster MY, Vignali DA, Doherty PC, Grosveld GC, & Ihle JN, Requirement for Stat4 in interleukin-12-mediated responses of natural killer and T cells, Nature, 382: 171, 1996.

[20]Wahyuningsih, S.P.A., Pemanfaatan ekstrak jamur Coriolus versicolor sebagai imunomodulator respon imun non-spesifik pada tikus putih akibat infeksi M. tuberculosis, Lembaga Penelitian Universitas Airlangga, Penelitian DIPA Universitas Airlangga, Surabaya, 2006.

[21] Wahyuningsih, S.P.A., Darmanto, W., & Wiqoyah, N., Bioaktivitas polisakarida krestin dari ekstrak jamur Coriolus versicolor sebagai imunomodulator respon imun akibat infeksi Myobacterium tuberculosis, Lembaga Penelitian Universitas Airlangga, Hibah Penelitian Strategis Nasional, Surabaya, 2009.

[22] Wahyuningsih, S.P.A., Winarni, D., & Masitha, A.N., Aktivitas polisakarida krestin dari ekstrak Coriolus versicolor terhadap peningkatan antibodi Mus musculus akibat paparan Mycobacterium tuberculosis, Berkala penelitian hayati, 17: 177-183, 2012.

[23] Willment, J.A., The humanbeta-glucan receptor is widely expressed and functionally equivalent to murine Dectin-1 on primary cells, Eur journal immunology, 5: 1539–1547, 2005.

[24] Vetvicka, V., Kiyomi, T., Rosemade, M., Paulin, B., Bill, K., & Gary, O. Orally-Administered Yeast β-1,3 glucan Prophylactically Protects Against Anthrax Infection and Cancer in Mice, Journal American Nutraceutical Assosiation, 5: 2, , 2002.

[25] Vos, A., M'Rabet, L., Stahl, B., Boehm, G., & Garssen, J., Immune-modulatory effects and potential working mechanisms of orally applied nondigestible carbohydrates, Crit review immunology, 27: 97-140, 2007.

[26] Yoshida, A., & Koide, Y., Arabinofuranosyl-terminated and mannosylated lipoarabinomannans from Mycobacterium tuberculosis induce different levels of interleukin-12 expression in mouse macrophages, Infect immun, 65: 1953–1955, 1997.

[27] Yu, S., Weaver, V., Martin, K., & Cantorna, M.T., The effects of whole mushrooms during inflammation, BMC Immunology, 10:12, 2009.

http://allergycliniconline.com/

http://www.the-scientist.com/

http://www.ncbi.nlm.nih.gov/pubmed/10702635%20(4

http://www.ncbi.nlm.nih.gov/pubmed/8700208?dopt=Abstract

http://www.ncbi.nlm.nih.gov/pubmed/8700208?dopt=Abstract



187


Surabaya Indonesia

ORAL PAPER OF OPTICS AND PHOTONICS (OPHO)



Ahmad et al. 188


Surabaya Indonesia

Investigation Of Graphene Synthesised By Electrochemical Exfoliation

As Passive Saturable Absorber For Pulsed Laser Generation

Fauzan Ahmad1,2, Sulaiman Wadi Harun3, Roslan Md Nor4, Harith Ahmad3 and Mohd Haniff Ibrahim2

1Department of Electronic Systems Engineering, Malaysia Japan International Institute of Technology (MJIIT), Universiti Teknologi Malaysia, 54100 Kuala Lumpur, Malaysia

2Lightwave Communication Research Group, Faculty of Electrical Engineering, Universiti Teknologi Malaysia, Skudai, Johor, Malaysia

3Photonic Research Centre, University of Malaya, 50603 Kuala Lumpur, Malaysia 4Department of Physics, University of Malaya, 50603 Kuala Lumpur, Malaysia Corresponding author’s Email: [email protected] and [email protected]

Abstract. We demonstrate a Q-switched Erbium-doped fiber laser (EDFL) using a graphene synthesised by electrochemical exfoliation of graphite at room temperature in 1% sodium dodecyl sulphate ( SDS) aqueous solution. Graphene flakes obtained are mixed with polyethylene oxide (PEO) as the host polymer to produce free standing composite thin film which act as passive Q-switcher in EDFL in ring cavity. The EDFL operates at around 1561 nm with the pulse train repetition rate ranged from 16 to 44 kHz, the average output from 0.07 to 0.20 mW, maximum pulse energy of 6.2 nJ and the shortest pulse duration of 3.0 μs.

INTRODUCTION The use of graphene and carbon nanotubes (CNTs) as passive saturable absorber (SA) for ultrafast laser

generation in Q-switched and mode-locked regime have been entensively reported [1-5]. The compatibility of the material to act as passive saturable absorber is due to their intrinsic properties such as ultrafast second and third order nonlinearities for CNTs [2] and ultrafast carrier relaxation and ultra-broad-band resonate nonlinear optical response for graphene [6]

Compare to CNTs which is diameter dependence, graphene works in wider spectral range without any chirality and diameter dependence [7]. Graphene is a flat monolayer of carbon atom tightly packed a into two dimensional (2-D) honeycomb lattice. It can be stacked to form 3D graphite, rolled to form 1D nanotube and wrapped to form 0D fullerenes [8,9]. Nair et al. [10] has demonstrated that despite being only one atom thick, graphene absorb a significant (πα=2.3%) fraction of incident white light due to its unique electronic structure. The optical absorption is also found to be frequency independent and proportional to the number of layers [11-12]. Compared to mode-locking, Q-switching requires less control of cavity parameter and is more efficient in term of cost, operation and implementation [3].

Graphene synthesis for enabling Q-switching in fiber laser has been demonstrated using chemical vapor deposition (CVD) [13, 14], using graphene oxide (GO) [5,15,16] and reduced graphene oxide (rGO) [17,18]. CVD synthesis requires high temperature up to 1000C and requires complex processing. To produce graphene oxide, graphite is react with mixture of potassium permanganate (KMnO4) and concentrated sulfuric acid (H2SO4) which is known as the Hummer method [19] and due to the chemical reaction, the exfoliated GO is highly decorated with oxygen-containing groups and can be (partly) reduced to graphene-like sheets by removing the oxygen-containing groups to obtain rGO [20]. There are several other reduction methods to obtain rGO from GO such as chemical reduction [21], thermal reduction [22] and photothermal reduction [23].

Recently, Wei et al. [24] reported the fabrication of an energy storage devices using graphene synthesised by electrochemical exfoliation. Electrochemical exfoliation is regarded as a green approached among the chemical methods where a specific voltage is applied between the two graphite rods in an ionic liquid and the exfoliation of the graphite anode is occurs after several hours [25-27].

In this paper, we demonstrate a Q-switched Erbium-doped fiber laser (EDFL) using graphene based SA synthesized by electrochemical exfoliation. Graphene flakes are mixed with PEO as the host polymer and free standing composite thin film is formed by drying the graphene polymer composite in petri dishes at 56oC. The SA is integrated in the EDFL by sandwiching the thin film graphene based SA between two fiber connectors and the performance of the laser is investigated.





Ahmad et al. 189


Surabaya Indonesia

MATERIAL AND METHODS In this work, graphene was produced using the electrochemical exfoliation process. A constant voltage ~ 20

V was applied to two electrodes (graphite rods) 1 cm apart, placed in an electrolysis cell filled with electrolyte (1% SDS in deionized water) as shown in Figure 1. According to Wie et al. [24] and Lu et al.[25], the process occurs during the electrochemical exfoliation is production of hydroxyl and oxygen radicals due to electrolysis of the water at the electrode, then oxygen radicals start corroding the graphite anode, followed by intercalation of anionic surfactant and finally graphene sheets is creates in the solution. In this work black sediments (graphene) started to peel off from the anode after several minutes. The exfoliation continued for 2 hours to obtain stable graphene suspension in the SDS solution.

Figure 1 Electrochemical exfoliation of graphene The suspension was centrifuged at 3000 rpm for 30 min to remove large agglomerates. The supernatant

portion of the suspension was decanted. The concentration of the produced graphene was estimated from the weight of the suspension used. To fabricate the composite, 1 g of polyethylene oxide (PEO) (Mw = 1 000 000 gmol-1) was dissolved in 120 ml of deionized water. Graphene solution (obtained from electrochemical exfoliation) was mixed with PEO solution at ratios of 1g: 4ml of graphene:PEO respectively. Then the solution was dried in petri dishes at 56oC to obtain free standing films with 50 µm thickness.

The fiber laser used in the experiment is schematically shown in Fig. 2. It has a ring configuration and is Q-switched by the graphene polymer composite based SA. The SA is fabricated by cutting a small part of the earlier prepared film (2×2 mm2) and sandwiching it in between two FC/PC fiber connectors, after depositing index-matching gel onto the fiber ends. The insertion loss of the SA is measured to be around 3 dB at 1550 nm. The gain medium is a 3 m long Erbium-doped fiber (EDF) with a 1480 nm laser diode is used to pump the EDF through a 1480/1550 wavelength division multiplexer (WDM). An isolator is incorporated in the laser cavity to ensure the unidirectional propagation. The output of the laser is tapped out from the cavity via a 95/5 fiber coupler. On the other hand, the 5% port of the Coupler is connected to the 1×2 3dB coupler and this serves as the output of the graphene based SA, which becomes the input for the 3 dB coupler. The 3dB coupler splits the input into two equal outputs and each output is then connected to an optical spectrum analyser and an oscilloscope respectively. The optical spectrum analyser (Yokgawa, AQ6370B) is used for the spectral analysis of the Q-switched EDFL with a spectral resolution of 0.02 nm whereas the oscilloscope (Tektronix,TDS 3052C) is used to observe the output pulse train of the Q-switched operation in the form of electrical signal via a 460 kHz bandwidth photodetector (Thor lab, PDA50B-EC).



Ahmad et al. 190


Surabaya Indonesia

Figure 2 Experimental setup. (WDM: 1480/1550 nm wavelength division multiplexer; EDF: Erbium-doped fibre; OSA: optical spectrum analyser; OSC: Oscilloscope.

EXPERIMENTAL RESULTS

Figure 3 Optical spectrum of the

Q-switched fiber laser Figure 3 shows a typical optical spectrum analyzer trace at the maximum pump power of 77 mW which

operates at around 1561 nm. Further experimentation reveals that stable Q-switched pulse train appeared when the pump power is at 29 mW. Figure 4 shows the pulse train for pump powers of 29 mW, 69 mW, and 77 mW, respectively. The repetition rate is observed to be proportional to the pump power, and it was discovered that the repetition rate is increases from 16 kHz to 44 kHz when the pump power is increased from 29 mW to 77 mW. This behaviour is characteristic of Q-switching operation, where the pulse repetition rate changes with the pump power [28,29]. It was also found that the average pump power increases from 0.07 mW to 0.20 mW with increasing pump power, as illustrated in Figure 5. Figure 6 depicts the relation of the of the pulse width and pulse energy with increasing pump power. From Figure 6, it can be seen that the shortest pulse width obtained with the proposed graphene exfoliation was 3 μs and comparable with others using graphene synthesisby CVD

[13], graphene oxide [5, 15,6 ] and CNTs [30,31], corresponding to the maximum repetition rate at a pump power of 77 mW.



Ahmad et al. 191


Surabaya Indonesia

Figure 4 Train of pulses at different pump power

Pump power : 29 mW Repetition rate: 16 Khz





Ahmad et al. 192


Surabaya Indonesia

Figure 5 Pulse repetition rate and average output power versus pump power

Figure 6 Pulse width and pulse energy versus pump power

CONCLUSION In conclusion, we have demonstrated a passively Q-switched Erbium doped fiber laser around 1.5 μm. The

graphene saturable absorber was fabricated by electrochemical exfoliation of graphite rod in electrolyte in room temperature and then mixed with polymer to form a free standing film with thickness of 50 µm. The thin film is then attached to the end of fiber ferrule with the aid of index matching gel and connected with another clean fiber ferrule via FC connector. The pump threshold of Q-switching operation of the fiber laser was about 29 mW. When pump power was increased from 29 to 77 mW, the pulse train repetition rate ranged from 16 to 44 kHz, and the average output power increased from 0.07 to 0.20 mW. At the pump power of 77 mW, we achieved the maximum pulse energy of 4.6 nJ and the shortest pulse duration of 3.0 μs.

ACKNOWLEDGEMENT The authors thank Universiti Teknologi Malaysia (UTM) and Ministry of Higher Education (MOHE)

Malaysia for supporting this research work under Research University Grant (RUG) Scheme grant no: 05J60, Fundamental Research Grant Scheme (FRGS) grant no: 4F317, and Photonics Research Centre, University of Malaya.

REFERENCES [1] Martinez, A., Fuse, K., Xu, B., & Yamashita, S., Optical deposition of graphene and carbon nanotubes

in a fiber ferrule for passive mode-locked lasing, Optics Express 18(22), 23054-23061, 2010. [2] Hasan, T., Sun, Z., Wang, F., Bonaccorso, F., Tan, P. H., Rozhin, A. G., & Ferrari, A. C., Nanotube–

Polymer Composites for Ultrafast Photonics, Adv. Materials 21, 3874,2009.



Ahmad et al. 193


Surabaya Indonesia

[3] Popa, D., Sun, Z., Hasan, T., Torrisi, F., Wang, F., & Ferrari, A. C., Graphene Q-switched, tunable fiber laser, Appl. Phys. Lett. 98, 073106, 2011.

[4] Liu, H.H., Chow, K.K., Yamashita, S., & Set, S.Y., Carbon-nanotube-based passively Q-switched fiber laser for high energy pulse generation, Optics & Laser Technology 45, 713–716 (2013)

[5] Luo, Z. Q., Zhou, M., Weng, J., Huang, G. M., Xu, H. Y., Ye, C. C. & Cai, Z. P., Graphene-based passively Q-switched dual-wavelength erbium-doped fiber laser,Opt. Lett. 35, 3709,2010.

[6] Bonaccorso, F., Sun, Z., Hasan, T., & Ferrari, A. C., Graphene Photonics and Optoelectronics, Nature Photonics, 4, 611,2010.

[7] Sun, Z., Hasan, T., Torrisi, F., Popa, D., Privitera, G., Wang, F., Bonaccorso, F., Basko, D. M. & A. C. Ferrari, Graphene Mode-Locked Ultrafast Laser, ACS NANO, 4, 803,2010.

[8] Allen, M. J., Tung, V. C., & Kaner, R. B., Honeycomb Carbon: A Review of Graphene, Chemical Reviews 110(1), 132-145, 2010.

[9] Geim, A. K., Graphene: Status and Prospects, Science, 324(5934), 1530-1534, 2009. [10] Nair, R. R., Blake, P., Grigorenko, A. N., Novoselov, K. S., Booth, T. J., Stauber, T., Peres,

N.M.R., & Geim, A. K., Fine Structure Constant Defines Visual Transparency of Graphene, Science 320(5881), 1308, 2008.

[11] Kuzmenko, A. B., van Heumen, E., Carbone, F., & van der Marel, D., Universal Optical Conductance of Graphite, Physical Review Letters, 100(11), 117401, 2008.

[12] Yamashita, S., A Tutorial on Nonlinear Photonic Applications of Carbon Nanotube and Graphene, Journal of Lightwave Technology, 30(4), 427-447, 2012.

[13] Zhang, L. Q., Zhuo, Z., Wang, J. X., & Wang, Y. Z., Passively Q-switched Fiber Laser Based on Graphene Saturable Absorber, Laser Physics, 22(2), 433–436, 2012.

[14] Wei, L., Zhou, D. P., Fen, H. Y., & Liu, W. K., Graphene-based Q-switch erbium doped fiber laser with wide pulse repetition rate range, IEEE Photon. Technol. Lett., 24(4), 309-311,2012.

[15] Ahmad, H., Muhammad, F.D., Zulkifli, M.Z., & Harun, S.W., Graphene-Oxide-Based Saturable Absorber for All-Fiber Q-Switching With a Simple Optical Deposition Technique, IEEE Photonics Journal, 4 (6), 2205-2213, 2012.

[16] Yap, Y. K., Huang, N. M., Harun, S. W., & and Ahmad, H.,Graphene Oxide-Based Q-Switched Erbium-Doped Fiber Laser, CHIN. PHYS. LETT. 30(2), 024208, 2013.

[17] Jiang, M., Ren, Z., Zhang, Y., Lu, B., Wan, L., & Bai, J., Graphene-based passively Q-switched diode-side-pumped Nd:YAG solid laser, Optics Communications 284, 5353–5356 (2011)

[18] Sobon, G., Sotor, J., Jagiello, J., Kozinski, R., Librant, K., Zdrojek, M., Lipinska, L., & Abramski, K. M., Linearly polarized, Q-switched Er-doped fiber laser based on reduced graphene oxide saturable absorber, Applied Physics Letters. 101(24), 241106 - 241106-4, 2012.

[19] Hummers, W. S., & Offeman, R. E., Preparation of graphitic oxide, J. Amer. Chem. Soc., 80( 6), 1339, 1958.

[20] Pei, S.F., & Cheng, H.-M., The reduction of graphene oxide, Carbon, 50( 9) 3210–3228, 2012. [21] Liu, J. B., Fu, S. H., Yuan, B., Li, Y. L., & Deng, Z. X., Toward a universal ―adhesive nanosheet‖

for the assembly of multiple nanoparticles based on a protein-induced reduction/decoration of graphene oxide, J. Am. Chem. Soc., 132, 7279-728,2010,

[22] Lin, Z., Yao, Y., Li, Z., Liu, Y., Li, Z. & Wong, C.-P., Solvent-assisted thermal reduction of graphite oxide, J. Phys. Chem. C, 114,14819-14825, 2010.

[23] Cote, L. J., Cruz-Silva R., and Huang, J. X., Flash reduction and patterning of graphite oxide and its polymer composite, J. Am. Chem. Soc., 131, 11027-11032,2009.

[24 ] Wei, D., Grande, L., Chundi, V., White, R., Bower, C., Andrew, P. & Ryhanen, T., Graphene from electrochemical exfoliation and its direct applications in enhanced energy storage devices, Chem. Commun., 48, 1239–1241, 2012.

[25] Lu, J., Yang, J., Wang, J., Lim, A., Wang, S., & Loh, K. P., One-Pot Synthesis of Fluorescent Carbon Nanoribbons, Nanoparticles, and Graphene by the Exfoliation of Graphite in Ionic Liquids, ACS Nano, 3, 2367, 2009.

[26] Su, C. Y., Lu, A. Y., Xu, Y., Chen, F. R., Khlobystov, A. N., & Li, L. J., High-Quality Thin Graphene Films from Fast Electrochemical Exfoliation, ACS Nano, 5, 2332–2339, 2011.

[27] Wang, J., Manga, K. K., Bao, Q. & Loh, K. P., High-Yield Synthesis of Few-Layer Graphene Flakes through Electrochemical Expansion of Graphite in Propylene Carbonate Electrolyte, J. Am. Chem. Soc., 133, 8888–8891, 2011.

[28] Kurkov, A. S., Sadovnikova, Y. E., Marakulin, A. V., & Sholokhov, E. M., All fiber Er-Tm Q-switched laser, Laser Phys. Lett. 7, 795, 2010.

[29] Sholokhov, E. M., Marakulin, A. V., Kurkov, A. S., and Tsvetkov, V. B., All-fiber Q-switched holmium laser, Laser Phys. Lett. 8, 382, 2011.



Ahmad et al. 194


Surabaya Indonesia

[30] Song, Y.-W., Q-switched fiber lasers with carbon nanotubes hosted in ceramics, App. Optics, 51(3),2012.

[31] Zhou, D.-P., Wei, L., Dong, B., & Liu, W.-K., Tunable Passively Q-switched Erbium-Doped Fiber Laser With Carbon Nanotubes as a Saturable Absorber, IEEE Photonics Technology Letters, 22(1), 9-11 2010.



195


Surabaya Indonesia

ORAL PAPER OF PURE AND APPLIED MATHEMATICS (OPMT)



Hernadi et al. 196


Surabaya Indonesia

The Numerical Simulations of Inverse Problems on the Parameter

Estimation

Julan Hernadi1,*

1)Department of Mathematics Education, Muhammadiyah University of Ponorogo, East Java Indonesia. *Corresponding author’s Email: [email protected]

Abstract. The problem of determination an estimator through sample data is a part of the inverse problems. Generally, the inverse problem has no solution in ordinary sense since the observation data have been contaminated by noises. The minimizer of the least square functional is usually taken as the solution of the inverse problem. The optimal design on parameter estimation uses the Fisher information matrix (FIM) as a tool for optimal criteria that minimizes some cost functional over set of FIM's. The consideration is based on the Cramer-Rao lower bound inequality can only be attained by the inverse of FIM. This article demonstrates how to implement the inverse problems in connection with parameter estimation numerically where the set of noises is generated independently from one trial to another. The numerical simulation is applied to a distributed parameter system of parabolic equation to find the optimal sensor locations for the parameter. The simulation is also carried out to a model of dynamical system to obtain the optimal time for measurements. Correspondingly, two algorithms are composed to do numerical realization by computer. The results from numerical experiments are confirmed to theoretical background. In particular, the accuracy of estimators are compared to the prior supposed nominal parameters and the variance of estimators are contrasted with the lower bound of Cramer-Rao inequality through the functional value acting on FIM's. The numerical results are also confirmed to the premise used in the parameter estimation that the information content on the parameter may vary considerably from one time measurement to another.

INTRODUCTION In general, two approaches are often used to study the real world phenomena, namely through their

mathematical representation and by making replication of their behavior. On the mathematical representation, a relevant mathematical model is built and studied. In most cases, the mathematical model involves some parameters that require to be estimated through observation data. This stage is known as the calibration model.

At the beginning, the problem of parameter estimation appeared in statistics where the functional relationship among variables is presumed to have an explicit representation like linear, quadratic, exponential, and some other, and the set of parameters in the statistical model is assumed arising from a population with certain distribution probability [7]. In latest development, the explicit relationships were not adequate anymore to deal some problems in science and engineering. The implicit relationship among variables like differential equations subjected to initial or boundary value conditions, integral equations, or integro-differential equations had attracted the attention many mathematicians and applied scientists [13, 6, 10, 1].

The abstract formulation of parameter estimation (identification) can be described by the following operator equation.

(1.1)

where denotes the operator, constitutes the parameter vector characterizing the model and represents the observation data that might be a function of time or a discrete set of observations. Two problems in the parameter identification consist of the forward and the inverse problem. On the forward problem, the parameter is given, the output model is determined after solving the operator equation (1.1). In the real application, the parameter usually unknown and it must be estimated through the observation data . This stage is regarded as the inverse problem. The inverse problem is generally ill-posed because of noise disturbance most probably accumulated in the observation due to reading mistaken of instrument or rounding error of numerical data. This is the main obstacle in solving the inverse problem in connection with parameter estimation. This situation can be represented as

(1.2)

where , i.e. it holds (1.1) exactly whenever . The exact parameter is sometimes called the nominal, true or natural parameter and its existence is by a nature assumption.

An intuitive consideration in choosing the optimal sample that a sample containing much information for a certain parameter is believed to result a better estimator for the such parameter. The estimation quality is



Hernadi et al. 197


Surabaya Indonesia

characterized by the closeness estimator to the nominal parameter (accuracy) and the small of variance (reliability). In the real situation, the accuracy of any estimator can not be known since the true parameter is unknown. Thus, the variance is used as the measure of the estimation quality. It is hypothesized that the smaller of variance is the more accurate of estimator. This hypotheses is intuitively without any theoretical background. In the simulation, the nominal estimator is usually given prior to the design of experiment with the intention of knowing the powerful of the estimation method. In this way we can observe how the variance influences on the estimator accuracy.

The premise used on the parameter estimation is that the information content on the parameter may vary considerably from one time measurement to another [4]. Therefore, determination of the best time for measurements is required in order to obtain an optimal sample in the sense it contains the maximum information about the parameter which is being estimated.

One of the parameter model that frequently appears on the applied sciences is the system of differential equations which models some physical, sociological or biological phenomenon [4,5]:

(1.3)

where denotes the vector of state variables and constitutes the parameters vector. This model is a kind of the lumped-parameter system (LPS). For a given the admissible parameter , the solution to the initial value problem (1.2) can be obtained. Sometimes the model output emerges in the form of functional with respect the state variable , i.e. . Consequently, the model output is representable as

(1.4)

The another kind of parameter model is the distributed-parametric system (DPS) which is a model that depends on both time and spatial variables. One of the DPS model is represented by a partial differential equation subject to some initial and boundary values [12]:

(

) (1.5)

with boundary conditions (

) and initial value .

Parallel to LPS model, here we need to obtain the best for both times and locations to take measurements for data sample to estimate parameter . In practical, the DPS model is generally solved by a semi-discrete approach by discretizing one variable while keeping the another continue. The most common variable to be discretized is the spatial variable. Hence, the output of LPS model (1.4) basically can be adapted to DPS model by taking time variable fix. The problem now reduces to determine the best locations for taking measurements.

By assumption the model (1.3) is representative enough for the real problem, the existence of the nominal parameter is assured. Thus, the output system is regarded as the true output. It is also assumed that measurement at time , contains error so that two components merge in , namely the clean data and the noise. This situation is written as

[ ] (1.6)

The random process of measurement error is assumed to have zero mean, time-dependent variance , and independent.

In this paper, the inverse problem is simulated to estimate the nominal parameter through observation data. To asses the sample quality, the estimator variability based on sample is calculated from a series of experiments. For each experiment, a set of data is disturbed by artificial noises independently from one trial to another. Therefore, a series of estimators are obtained as the solution of the inverse problem in each trial. The final estimator for is taken as the mean of estimators as the approach in the Monte Carlo simulation. Two issues that will be addressed in connection with this simulation are to verify the premise used on the parameter estimation and to confirm the estimator variance with respect to the lower bound of Cramer-Rao inequality through FIM.

THE THEORETICAL BACKGROUND Considering the output model and the statistical model (observation process), the objective

function is given by a generalized functional error as:

∫

( )



Hernadi et al. 198


Surabaya Indonesia

where denotes a general measure defined on -algebra of [ ]. In particular is the probability measure. Let be the ordered set of measurement times, namely { } where and be any set in the -algebra, then for each we define the Dirac measure at by taking

{

(2.2)

Accordingly, the measure is defined as

∑

With respect to this measure, the cost functional (2.1) is reduced to the discrete version of weighted least squared functional.

∑

( )

In practical, the functional (2.3) is frequently used as the objective function instead of (2.1). It measures the weighted square discrepancies between the true output given by and the model output . In the inverse problem, parameter is determined by minimizing (2.3) in conjunction with noise imposed on according to equation (1.6).

In the application, the inverse problem is applied to an optimal sample of measurements containing the most information about parameter . The abstract problem of optimal design as introduced in [5] is how to choose ―the best‖ measure over [ ] the set of all probability measures on [ ] in the sense that

( ( )) ( )

where denotes the Fisher Information matrix (FIM) given by

∫

This abstract formulation can not proceed to realize the computation numerically since the such measure is very difficult to obtain. Instead, the set of probability measures [ ] now is characterized by mesh, or points in so that the optimization problem now is simplified to problem of choosing set of points {

} such that

( )

( )

where stands for a set of all meshes or partition on [ ] and expresses a functional acting on the set of FIM‘s. Likewise, the FIM (2.5) is approximated by its discrete version [11]:

∑

where

indicates the vector gradient of parameters. Furthermore, the functional plays as the

criteria on the optimal design. Following are three famous optimal criteria found in many literatures, e.g. [12, 5]:

1. D-optimal which maximizes the determinant of FIM, i.e.

with for determinant of .

2. E-optimal which maximizes the spectral radius of FIM, viz.

with for the largest eigen value.



Hernadi et al. 199


Surabaya Indonesia

3. A-optimal or SE-optimal which maximizes the trace of FIM, namely

with for the sum of the main diagonal entries (trace).

THE NUMERICAL SIMULATIONS According to previous explanation we have two kinds of optimization problems that must be solved in the

parameter identification. To obtain an optimal sample, we solve (2.6) based on some optimal criteria, e.g. D-optimal, E-optimal, or A-optimal. The argument of objective function consist of the matrices defined by (2.7). These are not trivial matters. The optimal sample obtained at this stage is used to estimate parameter . Considering the measurement interference of noises, the inverse problem is solved by minimizing the weighted least squared functional given in (2.3). In order that numerical simulation does work, the nominal or true parameter is prior supposed by a certain value. The numerical simulation is carried out on both system (DPS and LPS).

Numerical Simulation for DPS

Consider the model represented by partial differential equation (1.5). The solution of this model can be regarded as the true solution of the real problem. By discretizing spatial variable and keeping continue the time variable, the observation process might be represented as

( ) where [ ] and [ ]. In this case the observation vector lies in an -dimension space, i.e. , . It is assumed that the noise random process has zero mean ( ) , ( ) , and ( ) where is the Dirac delta distribution concentrated at origin, viz. if and if . In this model we want to obtain the optimal sensor location that gives the most information about parameter while observation data has been contaminated by some noises. The following algorithm is required to implement the computation in computer, in particular here we used MATLAB.

Algorithm 1

1. Discretize the time interval [ ] with sampling or by a uniform partition with step size Let be the number of times sample.

2. Take a set of potential sensor location { } by applying some strategy, or taking equi-distance points (uniformly), for a simplify. In the real optimal design, these optimal points are obtained from the solution to optimization problem (2.6).

3. For each , do the following steps: a. Generate the vector of true output with respect to assumption for all

sampling times. This vector is represented as

where .

b. Generate the artificial noises , then collect it as (

| .

c. Define the observation data ( | by taking

d. Define the objective function with

∑

e. The estimator is obtained as the minimizer of previous optimization problem, viz. .

f. Do repetition steps (a) till (e) as trials, then a set of estimators is collected. These estimators are different by artificial noises enforced on each trial. The final estimator corresponding to sensor location is taken as the mean of those estimators.

g. Calculate the standard error for the variability of estimators by formula



Hernadi et al. 200


Surabaya Indonesia

∑( )

4. The optimal sensor location is the sensor location with minimum standard error.

EXAMPLE 1 Basically the algorithm 1 introduced previously can be applied to any the distribution parameter system

(DPS). For simplicity reason, we take the model of temperature distribution on rod thin as given by the initial and boundary value problems

( )

( )

This model assumes that the rod thin is perfect insulated, no heat source, and the two end points placed in contact with ice packs at . The parameter stands for the material diffusivity forming the rod. This problem is very simple since the exact solution is of closed form given by

This problem has been considered by several authors for case study, e.g. in [12]. In the cited reference, the author in [12] did not mention the algorithm, the optimization algorithm, the way how to generate the artificial noises. In this simulation, the numerical result follows the algorithm 1 with set of noises was geneerated by function randn, the optimization algorithm used function fminsearch and fminbnd which are the build-in functions in Matlab. In the other manuscript [8, 9], the author had applied the particle swarm optimization (PSO) to find the optimum. This global optimization scheme showed a very slow convergence rate. Numerical experiment 1. For this simulation we took , and . The sampling time takes uniformly with . Suppose the candidates of optimal sensor location are given in the following set:

{ }. The experiment was carried out as times\ and the noise was generated independently among one to

other trials. The results of numerical experiment is summarized in the Figure 1 that showed the estimators for each sensor location (left panel) and corresponding their standard errors (right panel). According to simulation, the estimator obtained from location is the most accurate for that closest to the nominal parameter (left panel) and also with smallest standard error (right panel). This result indicates that the smaller variance is the more accurate estimator. In this case we say that is the most optimal location and it contains the most information about parameter being estimated. Furthermore, we need to confirm the result to the optimal criteria based on FIM as mentioned in section 2. Since this is only 1-d problem, it is easier to apply the continuous formula of FIM as given in (2.5). By elementary calculus, we can find the formulation FIM as follows:

∫ ( )

Since the integral term is always positive, the maximum of depends totally on term. It is trivial that reaches maximum at or . So, it is well-confirmed.

Figure 1. Estimators from sensor locations (left) and their corresponding standard errors (right)



Hernadi et al. 201


Surabaya Indonesia

Numerical Simulation for LPS

For the second simulation, we take the model of dynamical system with parameter as given by (1.3). Here, the objective is to choose the optimal set of measurement times. Previously, only one sensor location is involved in each trial, but now a number of points is taken account on the simulation. The following theorem is used to perform some numerical experiments.

Algorithm 2

1. Set a potential sample { } where { } is the set of measurement times. So, each sample contains the same number of points, namely .

2. Take a positive integer the number of repetitions. 3. Generate a random matrix of size where its rows consist of the noises vector . 4. For each , do the following steps:

a. Define the true output . b. For each , define the observation

where and is the random vector placed on - row of matrix . Determine the estimator

where

∑ | |

c. The final estimator for each is taken as

∑

Then, calculate the standard

error of this estimator by formula ∑ ( ) where denotes the

estimator obtained through sample at trial.

5. Repeat step 4 until all members of potential sample runs out.

EXAMPLE 2 We consider here the logistic model of Verhulst-Pearl tat describes the dynamical growth of one population

which intrinsic growth and growth with saturation due to the carrying capacity.

(

)

where denotes the carrying capacity, stands for the intrinsic growth, and signifies the initial size of population. Thus, the system of parameters is given by . For a simplification we only consider two parameters and and set as a constant. This consideration based on previous experiment that indicated parameter is less sensitive than the others [9]. This model is also simply with the analytic solution given by

(

)

.

Numerical experiment 2. In the implementation, we take in advance a set of 13 time intervals of measurement, and then put 5 points uniformly from each interval where both end points are included. In this experiment we took intervals as presented in Table 1. This strategic yields a constrained optimization which is similar as introduced in [8, 5]. Then, the parameter is estimated by data from measurement made on those intervals. The nominal parameter had been chosen , , number of trials , and the deviation standard . There is no specific deliberation with this choosing, even we can take any others number along admissible.

Table 1. Time interval for measurement

No Interval No Interval No Interval No Interval No Interval 1 [0, 4] 4 [12, 16] 7 [8, 16] 10 [10, 20] 13 [12, 20] 2 [4, 8] 5 [16, 20] 8 [16, 22] 11 [0, 25] 3 [8, 12] 6 [2, 8] 9 [0, 10] 12 [5, 10]

The optimization problem is to choose and that minimize the weighted least squared functional (2.3).

The function fminsearch had been applied in the implementation. The estimators and their standard errors obtained from each time interval is summarized in Tabel 2.



Hernadi et al. 202


Surabaya Indonesia

Tabel 2. Experiment results on each time interval

No of sample

Estimation parameter Standard errors

1 0.6954 0.1048 0.4971 0.0495 2 0.6995 0.1005 0.0119 0.0047 3 0.6996 0.1022 0.0278 0.0238 4 0.7075 1.0379 2.7343 390.79 5 1.3641 3.7687 83.300 2696.8 6 0.6985 0.1012 0.0290 0.0049 7 0.7001 0.1018 0.0290 0.0235 8 1.4344 3.6315 84.200 2474.0 9 0.6993 0.1010 0.0112 0.0054 10 0.7068 0.1344 0.3802 0.8510 11 0.6988 0.1039 0.0832 0.0320 12 0.7002 0.1000 0.0084 0.0043 13 0.7378 1.2389 18.582 18.5910

According Table 1, the worst estimation had been happened on interval [16, 20] or sample 5 where both accuracy and reliability were very poor. Both sample 4 and sample 5 give a good accuracy for parameter but poor for parameter , but the standard errors are rather big. The best result is produced by sample 12 of interval [5, 10], the accuracy and the standard errors for both parameters are highly good.

It is interesting to look the influence of measurement interval with respect to behavior of the estimation. The following the estimation from two samples of measurement are displayed on Figure 2. The left panel from the sample of interval [5, 10] shows a very reliable estimation. Otherwise, the sample 4 of interval [12, 16] on right panel exhibits a rather poor estimation, good enough for parameter but very poor for parameter . This finding indicates that any sample might contain more information about some parameters but less information for other parameters. Also, the premise that information content on the parameter may vary considerably from one time measurement to another is well-confirmed.

Figure 2. Estimators derived from the best sample (left) and the rather poor sample (right)

Finally, we make a confirmation the numerical results to the theoretical background in connection with the

optimal criteria and Fisher information matrix (FIM). The FIM for each sample is calculated by formula (2.7) with . The computation results are summarized on Table 3. Theoretically, the smaller value of optimal criteria is the better estimator as the variance is closer to the lower bound of Cramer-Rao inequality. Three optimal criteria had been applied here, namely D-optimal, E-optimal, and SE-optimal. According to this table, the smallest optimal criteria is given by sample 12 of interval [5, 10] with D-optimal criteria is -13.2531 and 0.0153 for E-optimal. It is consistent with the quality of estimation. The consistence is also shown by the worst sample of interval [16, 22] with 8.4355 for value of D-optimal and 3858.4 for E-optimal.

Table 3. The values of optimal criteria for each sample of each time interval No of

sample D-optimal E-optimal SE-optimal

1 -5.8509 0.8775 0.8808 2 -12.8318 0.0220 0.0221



Hernadi et al. 203


Surabaya Indonesia

3 -11.4845 0.0618 0.0619 4 -2.5850 10.4363 10.4435 5 8.4335 3858.4 3859.6 6 -12.3877 0.0248 0.0250 7 -10.6975 0.0865 0.0868 8 8.8811 5131.5 5132.9 9 -9.6012 0.4546 0.4548

10 -6.4705 1.4182 1.4193 11 -9.4171 0.1657 0.1662 12 -13.2531 0.0152 0.0153 13 -1.7538 18.5817 18.5910

According to numerical and theoretical analysis it could be concluded that sample 13 of interval [5, 10] is

the best since it contains the most information about both parameters and .

CONCLUSIONS The numerical simulation of inverse problems has been applied to the problem of finding the optimal

location for sensor and the optimal time for measurement. According to simulation, the optimal criteria did play important role to obtain the optimal sample. In practice, the lower bound of Cramer-Rao inequality is hard to be achieved by the value of optimal criteria, but the smaller this value is the better sampling quality. It was well-confirmed by a series of numerical experiments.

This simulation has been treated to the single sample for all parameters. According to numerical results, a sample may contain difference information among parameters. This means it might be good for a group of parameters but poor for another group. It is suggestion to take more investigation about the sample characteristic with respect to specific parameter. The conjecture is that every parameter component associated with a particular optimal sampel.

ACKNOWLEDGMENTS This work was financed by Kopertis VII East Java, Ministry of Education and Culture, Republic of

Indonesia on Fundamental Research Project under contract No. 022/SP2H/P/K7/KM/2014.

REFERENCES 1. R.C. Aster, B. Borcher and H.T. Clifford, Parameter Estimation and Inverse Problems, Academic Press,

Burlington, MA (2005). 5. H.T. Bank, A.C. Arias and F. Kappel, Parameter Selection Methods in Inverse Problem Formulation,

Technical Report, CRSC, NCS University (November 2010). 6. H.T. Bank, S. Dediu, and S.L. Ernstberger, Sensitivity Function and Their Use in Inverse Problem, J.

Inverse Ill-posed Problem 15 (2007) 683-708. 7. H.T. Bank, K. Holm, S.L. Ernstberger, and F. Kappel, Generalized Sensitivity and Optimal Experiment

Design, J. Inverse Ill-posed Problem 18 (2010) 25-88. 8. H.T. Bank, K. Holm, and F. Kappel, Comparation of Optimal Design Methods in Inverse Problems,

Inverse Problems, 27 (2011). 9. A. Doicu, Trautmann, and F. Schreier, Numerical Regulation for Atmospheric Inverse, Springer-Verlag,

Berlin (2010). 10. C. G. Goodwin and R.L. Payne, Dynamical System Identification: Experimental Design and Data

Analysis, Academic-Press (1977). 11. J. Hernadi, The Experimental Strategy for Parameter Estimation of Dynamical System Model, Preprint

(2010). 12. J. Hernadi and F. Kappel, Optimal Design of Experiment: Sensitivity and Parameter Identification,

Technical Report, KF Univ-Graz (Desember 2010). 13. W.G. Muller, Collecting Spatial Data: Optimum Design of Experiment for Random Fields, Spring-Verlag,

Berlin (2007). 14. K. Thomaseth and C. Cobelli, Generalized Sensitivity Function in Physiological System Identification,

Annal of Biomedical Engineering 27 (1999) 607-616.



Hernadi et al. 204


Surabaya Indonesia

15. D. Ucinski, Optimal Measurement Methods for Distributed Parameter System Identification, CRC Press LLC (2005).

16. C.R. Vogel, Computational Methods for Inverse Problems, SIAM, Frontier in Applied Mathematics FR34, Philadelphia (2002).



Octarina et al. 205


Surabaya Indonesia

Formulating the Linear Model of Graph Coloring Using the

Combination of Gomory Cutting Plane Method and Balas Algorithm

Sisca Octarina* , Merry Pusvita Sari and Eddy Roflin

Department of Mathematics, Faculty of Mathematics and Natural Sciences Sriwijaya University

a)Corresponding Author . Email: [email protected]

Abstract. Graph coloring is assigning colors to the vertices in graph, such that the adjacent vertices must have different colors. Graph coloring is not only assigning color to vertices, but also finding the minimum number of colors of vertices. Let be a simple graph, -coloring of is assigning colors to the vertices of in such a way that adjacent vertices are assigned different color. If has a -coloring, then is -colorable. Graph coloring can be solved by using linear equation model. This research aims to formulate a linear model of graph coloring. The combination of Gomory Cutting Plane method and Balas algorithm based on these formulations was proposed. The solution is integer 0 or 1. The value of 0 represents that the color of k is not used in i-th vertex, and otherwise the value of 1 represents that the color of k is assigned to i-th vertex. Some implementation details and initial calculations are presented.

INTRODUCTION Graph coloring is the most common problem in graph theory. Scheduling, sequencing, and planning are

some application of graph coloring problems. Graph coloring can be formulated into a linear programming problem. However, a linear programming model that formed should be structured, so that it can be resolved.

According to Díaz [3], one of the algorithms to solve graph coloring is Cutting Plane algorithm. Cutting Plane Algorithm is an important algorithm in integer linear programming problem. The main idea considered the linear relaxation and tried to strengthen it by adding inequalities. In fact, the Cutting Plane algorithm has many iterations in the cutting procedure to obtain the optimal solution. Furthermore, Branch and Cut algorithm in graph coloring has also been proposed by Diaz [4]. It divides the solution repeatedly into subsets and solves the problem of subset itself. This algorithm is more effective than Cutting Plane algorithm for branching done first before cutting procedure in obtaining the optimal solution. This algorithm depends on branching strategies, and search, the upper limit and lower limit, the linear programming relaxation and cutting techniques. Long iteration also still occurs in this algorithm.

Roflin et al, [6] combined Gomory Cutting Plane method and Balas algorithm to resolve integer programming problems. Combination of these methods is a very effective, wherein the solution obtained is an integer 0 or 1 (unlike the Gomory Cutting Plane method which only generate integer solutions) without going through the lengthy process of calculation as in the Balas algorithm. This algorithm has been tested on an integer programming 0 or 1 problem, wherein Cutting Plane Gomory techniques as a major role in determining of the optimal solution.

Based on this background, this study formulated a linear model of graph coloring. The combination of Gomory Cutting Plane method and Balas algorithm based on these formulation was proposed. The algorithm was tested on a random graph and a series of problems of the literature.

LITERATURE REVIEW

Balas Algorithm

Balas algorithm was proposed by Egon Balas [1]. This method works by successively assigning the value 1 to a certain variables. Notations used in Balas algorithm are defined as follows Set of decision variables at iteration k. It is a partial solution at iteration . Set of decision variables in objective function Set of decision variables which are not included in . All decision variables are assumed to be zero unless

it has been defined as one in advance. . Set of decision variables which can repair the final solution. Number of violated constraints when variable 1 at iteration .




Octarina et al. 206


Surabaya Indonesia

Objective value at iteration . Final objective value

Slack variable at iteration . The Balas algorithm is as follows.

STEP 1. Initialization. Set = 0, , Y(k)= B with at less one component from B< 0, , 0, { } and .

STEP 2. Determine variable in which can be moved to . Choose one set of decision variables in Pk which has objective value better than Jk. Eliminate variables which have negative value.

1 , 0k k kj j js jy y a y (1)

Variable is feasible if . The with can be eliminated and inserted to the set Nk.

Eliminate all variables causing the current objective value higher than . Let the objective function:

k

k j jj J

f c x

(2) If , is set as 1, the objective function becomes . Thus which makes the objective value greater than cannot be included in solution. Let { }, then: k k k k kk kP J RN M J N M (3)

If , then go to STEP 4. Otherwise, go to the next step. Given constraints: 1 1 2 2 1 , 0j j jn n j ja x a x a x y b y (4)

which can be written as: 1 , 0

s kx J

k k kj j js s jy y a x y

(5)

By considering only variables in , we get: , 0

s k

k kjs j j

x P

a y y

(6)

Change variable, , to 1. The variable is chosen based on Eqs. (7). argmin max( )

t k

k kt t s

x P

x I I

(7)

where,

1

min 0, ,m

k ks j js k

j

I y a x P

(8)

Defines new solution . STEP 3. Update

:

1k kj j jsy y a (9)

Update objective function: 1k k tf f c (10)

If all , then , and go to STEP 4. Otherwise, back to STEP 2.

STEP 4. Backtracking and determining optimal solution. The backtracking process is conducted towards partial solution . It is started by assuming all variables in

are one. Then, successively assigning zero to certain variables. Herein, possibilities are evaluated where is number of variables in . Then, a combination which gives the smallest value is the final objective value.

Gomory Cutting Plane

The Gomory cutting plane is an algorithm for obtaining integer solutions of the Linear Programming (LP) [5]. This algorithm works by examine a solution of the linear programming obtained by simplex method. If the solution is not in integers, a new constraint which can cut the search space so that non integer solution can be eliminated. The procedure of generating new constraints is as the following. STEP 1. Given an optimal simplex tableau. If , then



Octarina et al. 207


Surabaya Indonesia

1

,n

ji j i

j

a w

(11)

when ⌊ ⌋

and , Eqs. (11) can be transformed to:

1

nj

i j ij

a w

(12)

and

1

nj

i j i ij

a w S

(13)

where is a slack variable. Assumed is non-integer, then⌊ ⌋

and ⌊ ⌋ where

and . The additional constraint is the differences between Eqs. (11) and (13) which can be formulized as follows.

1

n

ij

ij j i ff w S

(14)

METHODOLOGY The idea of substituting Gomory cutting plane to the Balas algorithm is proposed to accelerate the procedure

of Balas algorithm. As discussed before, Balas algorithm involves a backtracking procedure which takes a long iteration. Gomory cutting plane method can replace the backtracking procedure. Herein, Gomory cutting plane condition is applied to determine which variable should be set as 1. Detail steps is given as follows. a. Formulize the LP model of the graph. b. Solve the LP model using the proposed method (combination of Gomory Cutting Plane Method and Balas

Algorithm) with this steps. STEP 1. Initialization. Transform the LP to a standard form. STEP 2. Solve the standard form LP using dual simplex method to obtain the optimal dual simplex tableau. STEP 3. If all decision variables in the optimal solution are integer, then stop. Otherwise, generate a new constraint based on Gomory procedure. The new constraint is derived from the basis variable which has the biggest non-integer solution. Solve the new LP with dual simplex method and obtain new optimal tableau. If the new solution is integer, then stop and go to STEP 4. Otherwise, back to STEP 3. STEP 4. Substitute solution the original LP form to get the final optimal solution.

RESULTS AND DISCUSSION Assume that graph G has vertices V { }, with n colors. Suppose that variable with

is a binary variable with value 0 and 1. If color k be used for coloring vertices on graph, then , otherwise . Further suppose that which states that vertices accept color , then the linear model formulation of graph coloring that obtained is as below:

Objective function Min ∑

(15)

Subject to ∑

(16)

(17) (18) { } (19) (20)

Constraint (16) assert that each vertex has most only one color. Constraint (17) shows that every pair of adjacent may not share more than colors. Constraint (18) says that for every two neighboring vertices do not have same color. Constraint (19) and (20) show that variable and are biner variables with value 0 and 1.

is a variable which use for coloring vertices, whereas is a variable where vertices accept -th color. Value 0 shows that color was not used for coloring vertices -th, otherwise value 1 shows that color k will be used for coloring vertices -th.

Supposed that given graph has 2 vertices and 1 edge as Fig 1.



Octarina et al. 208


Surabaya Indonesia

Figure 1. Graph with Two Vertices

Fig 1. has vertices { } and two colors. The vertices which establish edge { } as graph on Fig 1. is not direct graph that . Suppose that green for 1 and yellow for 2, so according to Form (1) linear model formulation that obtained is as below: Objective Function Min ∑

(21)

Subject to (22) (23) (24) (25) (26) (27) (28) (29) { } (30) (31)

The linear model function that obtained according to Eqs. (21-31) was solved by the combination of Gomory Cutting Plane method and Balas algorithm with the iteration as below: STEP 1

Eqs. (21-31) previously was changed to standard form , , , , , . Constraint with inequality ≤ must be changed to equality and added one slack variable. Constraint with barrier = must be added one artificial variable, then Eqs. (21-31) could be rewritten as: Objective Function Min ∑

∑

∑

∑

(32)

Subject to (33) (34) (35) (36) (37) (38) (39) (40) { }, (41) Eqs. (32-41) which has formed, then was inserted to simplex table, but previously substituted dan : and (42) Then the objective function is as below: ∑

∑

∑

(43)

This step was needed in order to create simplex table with the value of and are 0. STEP 2

Eqs. (32-41) was considered as common linear programming problem and to count the solution can use dual simplex method. Early iteration table has given, then by using dual simplex table as below.

Table 1. First Iteration of Dual Simplex

Basis Solution

0 0 0 0 0 0 0 0 HB

0 0 1 1 0 0 1 0 0 0 0 0 0 0 1 1

0 0 0 0 1 1 0 1 0 0 0 0 0 0 1 TM



Octarina et al. 209


Surabaya Indonesia

-1 0 1 0 0 0 0 0 1 0 0 0 0 0 0

0 -1 0 1 0 0 0 0 0 1 0 0 0 0 0 TM

-1 0 0 0 1 0 0 0 0 0 1 0 0 0 0 TM

0 -1 0 0 0 1 0 0 0 0 0 1 0 0 0 TM

0 0 1 0 1 0 0 0 0 0 0 0 1 0 1 1 0 0 0 1 0 1 0 0 0 0 0 0 0 1 1 TM

This problem need seven iterations to find the optimal solution. The last iteration for dual simplex table is as below.

Table 2. Seventh Iteration of Dual Simplex

Basis Solution

0 0 0 -1 -1 0 0 0 0 1

1 0 0 0 0 0 0 0

0

0

0

0 0 0 0 0 1 0 1

0

0

0

0 0 1 0 0 0 0 0

0

0

0

0 0 0 1 0 0 1 0

0

0

0

0 1 0 0 0 0 1 0

-1

0

0

0 0 0 0 0 0 1 -1 -1 1 1 0 0

0 0 0 0 1 0 0 0

0

0

0

0 0 0 0 0 0 -1 0 0 0 0 0 1 1 0

The value on line z doesn‘t have positive value, all variables are not positive (negative and zero). Table 2 has been optimum but the value of solution is non integer, the next step is create the new additional constraint to eliminate non integer solution become integer solution. STEP 3

Additional constraint could be formulated from variable with non integer solution as could possible avoid variable coefficient equality with value ≤ -1. From Table 2, selected as derivative of equality the new additional constraint. Before add additional constraint equality, coefficient of constraint on equality have to change previously by adding value 1. Coefficient in column and on line each has negative value then every coefficient has been changed such that get obtain positive value as below:

On coefficient ,

has changed by way

On coefficient ,

has changed by way

The next from equality ,

Coefficient show as

Coefficient , show as

Is obtained equality ,

(44)

or

(

)

) (45)

The new additional constraint equality

(46)

The next is insert additional variable into Table 3. Value on line on Table 2 doesn‘t have positive value, but the solution on Table 3 still has negative value.

According to dual simplex method rule, leaving variable is variable with the largest negative solution. The largest negative variable on Table 3 is exist on , then is selected as leaving variable, then in next iteration as non basis variable.



Octarina et al. 210


Surabaya Indonesia

Table 3. Eighth Iteration of Dual Simplex with Additional Constraint Basis Solution

0 0 0 0 0 0 -1 -1 0 0 0 0 0 1

1 0 0 0 0 0 0 0

0

0

0 0

0 0 0 0 0 1 0 1

0

0

0

0 0 1 0 0 0 0 0

0

0

0

0 0 0 1 0 0 1 0

0

0

0

0 1 0 0 0 0 1 0

-1

0

0

0 0 0 0 0 0 1 -1 -1 -1 1 1 0 0 0

0 0 0 0 1 0 0 0

0

0

0 0

0 0 0 0 0 0 -1 -1 0 0 0 0 1 1 0 0

0 0 0 0 0 0 0 0

0

0

0 1

Table 4. Eighth Iteration (Coefficient Ratio Z and

Basis

0 0 0 0 0 0 -1 -1 0 0 0 0

0 0 0 0 0 0 0 0

0

0

0 1

Ratio - - - - - - - - 2 - - - - - - From Table 4, result of the smallest ratio calculation is 2 and on column . Therefore is selected as

entering variable. Next iteration, because in previous is selected as leaving variable then will replace by .

Table 5. Nineth Iteration Dual Simplex Table with Additional Constraint

Basis Solution

0 0 0 0 0 0 0 -1 1 0 1 0 2

1 0 0 0 0 0 0 0

0

0

0 0 1

0 0 0 0 0 1 0 1

0

0

1 0

0 0 1 0 0 0 0 0

0

0

1 0

0 0 0 1 0 0 1 0

0

0

-1 1

0 1 0 0 0 0 1 0

-1

0

-1 1

0 0 0 0 0 0 1 -1 -1 -1 1 1 0 0 -2 1

0 0 0 0 1 0 0 0

0

0

0 -1 1

0 0 0 0 0 0 -1 -1 0 0 0 0 1 1 0 0

0 0 0 0 0 0 0 0

0

0

0 1 1

Solution on Table 5 has been as integer value at all, it means the solution stopped and obtained,

and STEP 4



Octarina et al. 211


Surabaya Indonesia

Optimum value from graph coloring model formulation using combination Gomory Cutting Plane method and Balas algorithm as below:

On STEP 1 has been defined

, , , , thus

and = 2

On the result, and are 1 which means that on Fig. 1 the first and second color are used. Next is 0 which means the first color is not used on first vertex, whereas on is 1 means that second color is used on second vertex. On second vertex, is 1 means that first color on second vertex is used, whereas is 0 means that the second color is not used on second vertex. The coloring vertices has been obtained using the combination of Gomory Cutting Plane methods and Balas algorithm On early solution has supposed that yellow is 2 and green is 1 so that the graph coloring will be look like Fig 2.

FIGURE 2. Graph Coloring Using the Combination of Gomory Cutting Plane Method and Balas Algorithm The LP model was also tested by software LINDO program.

Figure 3. Calculating Graph Coloring Problem Using Software LINDO

The result is with optimal solution , , , , and .

FURTHER RESEARCH The combination of Gomory Cutting Plane method and Balas algorithm can be used for coloring vertices on

graph, where the solution can be integer 0 and 1. Value 0 shows that color k is not used on vertices , and otherwise value 1 shows that color k is used for coloring vertices A full implementation is necessary to complete the proposed algorithm. Based on the initial results, there is hope that the LP bound is strong and one may not need to have a very long iterations to find optimal colorings for many structured graphs. Further exploration will explore the robustness of this framework for general graphs. It will also be interesting to see the comparison between using this scheme with a cutting plane scheme that uses modified branching scheme proposed by Diaz [4]. Finally, it will be interesting to see if this framework can be suitably exploited to solve other variations and extensions of coloring problems.

1 2



Octarina et al. 212


Surabaya Indonesia

REFERENCES 1. Aldous, J.M. & Wilson, R.J., Graphs and Applications An Introductory Approach, Great Britain:

Springer Science & Business Media, 2000. 2. Balas, E., An Additive Algorithm for Solving Linear Programs with Zero-One Variables,

Operations Research, 13(4): p. 517, 1965. 3. Diaz, I.M. & Zabala, P., A Branch and Cut Algorithm for Graph Coloring, Discrete Applied

Mathematics, 154, pp. 826-847, 2006. 4. Diaz, I.M. & Zabala, P., Cutting Plane Algorithm for Graph Coloring, Discrete Applied

Mathematic, 156, pp. 159 – 179, 2008. 5. Gomory, R.E., Outline of An Algorithm for Integer Solutions to Linear Programs, Bulletin of the

American Mathematical Society, 64(5): p. 275-278, 1958. 6. Roflin, et. al., Substituting Gomory Cutting Plane Method Towards Balas Algorithm for Solving

Binary Linear Programming, Asian Journal of Mathematics and Applications, Volume 2014, Article ID ama0156.



213


Surabaya Indonesia

POSTER PAPER OF SEMINAR (PPOS)



Nafila et al. 214


Surabaya Indonesia

Electrochemical Degradation Remazol Black B Using Nanoporous

Carbon Paste Electrodes

Nafila*, Muji Harsini, Pratiwi Pudjiastuti

Departement Chemistry FST, Airlangga University *Corresponding Author’s Email: [email protected]

Abstract. Electrochemical degradation of Remazol Black B in sasirangan waste using electrodes carbon paste nanoporous has been studied. This study aims to determine the optimum conditions include the potential degradation, pH, time degradation and characterize the degradation product of remazol black B and impairment of COD value. In this study the electrochemical degradation used carbon paste electrodes nanopore (anode) and the silver wire (cathode). The results showed that the optimum state of degradation Remazol Black B on potential 9 Volt, pH 7 and the optimum time of 35 minutes for degradation 50 mL remazol black B 50 mg/L. Degradation Remazol Black B to produce compounds that are not dangerous, and can be directly discharged into the environment with the COD after degradation for 90 minutes at 105.608 mg/L with reducing of COD is 65.37%. LC-MS results showed that remazol black B has been degraded completely and the result of degradation is composed with CO2, just little hypochlorite ion (OCl-) and chloride ions (Cl-) which is not harmful to of environmental.

INTRODUCTION The textile industry in Indonesia has been growing rapidly and resulting in increasing need of dye.

According to Al-kdasi (2004), based on the chemical structure, dyes are grouped into nitroso dye, nitro, azo, stilbene, diphenyl methane, triphenyl methane, acridine, kuinolin, indigoida, aminokinon, fetus and indophenols1. While based on the method of dyeing or coloring material to be colored, dye classified into acidic, basic, disperse, direct and others. Dyes that often used are azo group that non-biodegradable. This dye has a high toxicity in mammalian cells, acting as an agent that triggers tumor, infertility, causing damage to the liver, spleen, kidneys, and also injured skin, eyes, lungs, and bone2. According to Widodo et al., 2008 some reactive dyes are often used in textiles coloring, such as Remazol brilliant orange 3R, Remazol RNL and Remazol Black B. Remazol Black B is also called Reactive Black 5. Remazol Black B formed in black powder with a molecular formula C26H21N5Na4O19S6 and a molecular weight of 991.82 g /mol3. Remazol Black B is also known with several names such as CI Reactive Black 5, Reactive Black B, Remazol Black 5, Drimaren Black R / K-3B. Remazol Black B has a wavelength 590 nm, LD50> 2,000 mg/Kg (rats), odorless and sensitive when it comes to contact with respiratory and skin (28-33%)4.

Some people who have been doing electrodegradation (electrodecolourization) research including Widodo et al (2008), they did electrodecolourization Remazol Black B using graphite electrodes at 6.5 V (voltage) with electrodecolourization result 95.11% during 60 minutes3. Widodo, et al (2009) conducted a electrodecolourization Remazol Black B again using PbO2 electrode/graphite with results of electrodecolourization at voltage of 5.5 V for 99.64%5. Widodo, et al (2010) conducted a electrodecolourization Remazol Black B again using different electrode PbO2/Pb that can be degraded perfectly for 90 minutes6. Other researchers also have done Remazol Black B electrodecolourization using carbon as cathode and PbO2 as anode. It show results of electrodecolourization at voltage of 5.5 V as 99.69% for 120 minutes7. Indigosol and Remazol Black B Elektrodegradation using electrodes PbO2 / Pb at voltage of 8 V with electrodecolourization results 98.42% for 150 minutes and decline in the value of COD is 51.47%8.

Electrochemical degradation is a method to spliting organic compounds or converting them into other compounds that caused by the direct and indirect oxidation process using potential energy. Electrochemical degradation is a process that is carried out with very high efficiency and mainly operated under the same conditions for waste in a large quantity9. Electrode is a tool which has an important role in the process of electrochemical degradation. One material that can be used to make electrodes is carbon nanoporous. Carbon nanoporous is a unique material and has been used widely in various fields such as the separation process technology, catalysts, energy storage, gas storage and energy conversion due to the high specific surface area and porosity that are easily be regulated10.

The degraded waste analyzed its COD value (chemical oxygen demand) and liquid chromatography-mass spectrometry (LC-MS) to characterize the compounds of Remazol Black B degradation product.




Nafila et al. 215


Surabaya Indonesia


Materials

All the reagents used in our work were pro-analysis. Remazol Black B (RBB) as azo dyes were obtained from Sigma Aldrich and used without any treatment. The structures of the Remazol Black B dyes are shown in Fig. 1. The light intensity was measured by spectrophotometer at 200 to 800 nm (Shimadzu 1600). NaOH and HCl solution was used for pH control.

NaO3SOH2CH2CO2S N

N

NH2 OH

NaO3S NaSO3

N

N SO2CH2CH2OSO3Na

Figure 1. Structure of Remazol Black B

EXPERIMENTAL METHOD

Preparation Carbon Paste Nanoporous Electrodes

Carbon paste is prepared by mixing carbon powder nanoporous and paraffin with ratio 7: 3. The mixtures is heated at a temperature of 60-80°C using a hotplate. The body of electrode made of a micropipette tip in which embedded copper wire (diameter 1 mm). Carbon paste is inserted in the hole of the tip body electrode (1 cm), then smoothed on clean paper. Carbon paste Electrodes dipped in an electrolyte solution to test the current support in electrodes11.

Electrochemical Cells Degradation

The series of electrochemical degradation tools are arranged as shown in Figure 2. Carbon paste electrodes as the anode and a silver wire as the cathode is connected to a DC voltage source. At the beginning of each test run, for potential and pH optimize using 10 mg/L Remazol Black B 50,0 mL and time optimize using 50 mg/L Remazol Black B 50,0 mL which containing electrolyte supported 0,1 M NaCl was placed in a electrolysis cell. Potential and current is passed through from the electrode to solution. During the degradation process, the solution is stirred using a magnetic stirrer. The measurement Remazol Black B solution degradation products were analyzed using a spectrophotometer UV-Vis at a wavelength 597.5 nm12. The decolourization efficiency was calculated using the relation:

‐

Where, Ci and Cf are concentration values of dyes solutions before and after treatment with respect to their λmaks, respectively or Ci and Cf are initial and final values of the dyes solutions, respectively.

Figure 2. Electrochemical degradation tools

Analysis the Results of Degradation

The measurement of the electrical needs and kinetic reaction of degradation reactions used 50 mg/L Remazol Black B 50 mL. The final results of the electrochemical degradation Remazol Black B solution is



Nafila et al. 216


Surabaya Indonesia

analyzed using LC-MS-MS to determine the final result of degradation compounds, COD using method SNI 06-6989.2-2004, analysis chloride ions using Argentomentry method and hypochlorite ion using Iodometry method to detect pollutant if released into the environment.


Degradation Optimization Remazol Black B

Optimization potential degradation

Optimization potential degradation 10 mg/L Remazol Black B 50.0 mL for 15 minutes used potensial range between 0 – 15 V with interval each 2 Volt.

Figure 3. Optimization potential degradation 10 mg/L Remazol Black B during 15 minutes

Figure 3 shows of the relationship difference potential degradation with the percentage of degradation Remazol Black B after degradation. On the curve shows that the the potential increasingly for more each potential to Remazol Black B degraded. In the 9 volt potential looks Remazol Black B gives the highest peak percentage degraded at 99.65%, potentially greater than 9 volts looks stationary. Thus the 9 volt potential degradation used for pH optimization and subsequent time of electrodegradation.

Optimization pH of Remazol Black B

Optimization of the pH is did with degrading 10 mg/L Remazol Black B 50.0 mL at 9 volts for 15 minutes with a variation of pH 1-9.

Figure 4. Effect off pH on percentage of degraded 10 mg/L Remazol Black B during 15 minutes

Relationship between the pH value of the percent degradation Remazol black B can be seen in Figure 4. In Figure shows an increasingly in the percentage of degraded Remazol Black B dye in the neutral condition until pH 9 and a decline in acidic to pH 4 and re-increased on acidic condition. pH optimum for degradation Remazol Black B at pH 7 looks the percentage of degradation almost perfect of 99.61%.

Optimization Time Remazol Black B



Nafila et al. 217


Surabaya Indonesia

Optimization time of degradation to do with 50 mg/L Remazol Black B 50.0 mL at 9 volt potential and pH 7. Variation of time ranging from 5-50 minutes with intervals of 5 minutes.

Figure 5. Time Effect with the percentage degraded 50 mg/L Remazol Black B 50 mL

Figure 5 shows that the longer the time degradation, the greater the percentage of degradation. The percentage of degradation started at 10 minutes to 40 minutes, at 35 minutes the percentage of degradation realize at 99.74% and to extend the time degradation up to 120 minutes for observed the maximum wavelength in the UV region which indicates the existence of benzene ring are still appeared. The Observation of spectra Remazol Black B structure can be learned from the results of degradation profiles to change the time measured using a UV-Vis spectrophotometer.

Figure 6. UV-Vis spectra Remazol Black B to the effects of the degradation time Remazol Black B Figure 6 shows a spectrum analysis of the longer time degradation, a decline in the intensity of maximum

absorption at a wavelength 597.5 nm in the visible region are declining and lost at 20minutes. At the next extension time is to observe the maximum wavelength in the UV region which show the benzene ring are still appeared. Appearance absorption degradation time at 120 minutes still seen a peak at 200-300 nm, its indicates that the presence of benzene in compouds after degradation.

Analysis of the Electricity Fare and Chemical Kinetics

Based on the calculation of electricity consumption rates of degradation Remazol Black B using an electrochemical method with carbon paste nanoporous electrodes relatively more cheap so this method can be used as an alternative to reduction in the dye industry waste.

Figure 7. Correlation concentration 50 mg/L Remazol Black B 50 mL versus electrical energy

Based on Figure 7 can be seen energy necessity increasingly with concentrations of the to degrade also getting bigger. Rates energy necessity to degrade 1000 mg Remazol Black B required cost IDR 115.2.



Nafila et al. 218


Surabaya Indonesia

Determination of reaction kinetics is used to determine orde reaction kinetics of degradation Remazol Black B. The curve made by connecting the value of ln C versus time degradation as shown in Figure 8.

Figure 8. The kinetics first orde reaction of Remazol Black B

Figure 8 shows that the degradation Remazol Black B occur at first orde reaction, which means a decrease rate in the degradation Remazol Black B is comparable to the Remazol Black B are left with a constant rate reaction is 0.159 mg/L / min.

Analysis of The Results of Degradation

The end result of degradation Remazol Black B analyzed the compound using LC-MS-MS. This analysis aims to determine the compounds resulting from the degradation process Remazol Black B solution for the degradation is shown in Figure 9.

Figure 9. The spectrum of ESI-MS tandem MS product of result degradation Remazol Black B

Results of analysis Remazol Black B after degradation using LC-MS-MS shows the signal at m / z 225.944183 thought to be the spectrum of intermediate compounds. The presence of these compounds reinforced with spectrum at a wavelength 247 nm in the UV region remaining after 120 minutes degradation.

COD analysis shows before and after the process degradation 50 mg/L Remazol Black B for 35 minutes and 90 minutes. It shows a decline at COD values respectively for 71.90% and 72.68%, while the reference solution NaCl increased 31.94%. This suggests that the presence of Cl- ions affect the value of COD. Tests on hypochlorite ion to determine the content of hypochlorite ion (OCl-) remaining after the degradation process and the obtained results of 1.52 mg (0.00152 g) at 50 mg/L Remazol Black B and after degradation for 35 minutes. The amount of residual hypochlorite ions of the degradation process Remazol Black B is still below the maximum threshold for contamination hipochlorite that is equal to 1,633 g. Analysis of chloride ion to determine the content of residual chloride ions used during the degradation process and obtained results of 2.67 mg after 35 minutes degradation. These results indicate that there is still residual chloride ions from the degradation process that has been stopped.

Qualitative test results CO2 gas that flowed into a solution of Ba(OH)2 shows the a white precipitate BaCO3 is formed. Moreover, it can be observed visually when the degradation process lasts contained gas bubbles that stick around the electrode body that indicate the presence of CO2 gas that is formed. It can be concluded that during the process of electrochemical degradation Remazol black B will produce CO2 gas.



Nafila et al. 219


Surabaya Indonesia

CONCLUSIONS Based on research that has been done, can be observed that carbon electrodes nanopore can be used as an

anode and a silver wire as the cathode in the degradation process Remazol Black B electrochemically with optimum conditions potentially 9 volt, pH 7 and the time of degradation for 35 min to 50 mg/L Remazol Black B 50 mL. COD results before and after the degradation process of Remazol Black B decreased 71.90% at the degradation of 40 minutes and 72.68% at the degradation of 90 minutes. The characterization results of degradation product Remazol Black B does not provide spectrum in LC-MS-MS are thought to be the intermediete compounds with m / z 225.5743.

ACKNOWLEDGMENTS The Authors are grateful to BPPDN schoolarship from DIKTI for the financial support extended. Also

grateful to laboratory for their support to carry out this work.

REFERENCES 1. Al-Kdasi, A., Idris, A., Saed, K. and Guan, C.T., 2004. Treatment of textile waste water by advanced

oxidation processes. Global Nest the Int. J. 6: 222-230. 2. Singh, A., Manju, R., Suman and Bishnoi, N.R.., 2012, Azo Dye Decolorization on Immobilized Dead

Yeast Cells Employing Sequential Design of Experiments, Ecological Engineering, 47, 291– 296. 3. Widodo, D.S., Gunawan, and Kristanto, W.A., 2008, Electroremediation Contaminated Waterworks: 2.

Using Graphite on Electrodecolourization Remazol Black B Solution, J. Kim Sains & Aplikasi Vol XI No., 88-93.

4. Manurung, R., Rosdanelli H., and Irvan, 2004, Renovation Reactive Dye Colour on Anaerob-Aerob, Chemical Engineering, USU, Medan.

5. Widodo, D.S., Ismiyanto, Ikhlas, F.N., 2009, Electroremediation Contaminated Waterworks: 3. Electrodecolourization Remazol Black B Solution with PbO2/carbon and Analysis residue from decolourizaton solution, JKSA, vol.11 no. 1.

6. Widodo, D.S., Nirmasari, A.D., Haris, A., 2010, Influence of pH to Electrodecolourization Remazol Black B Dye Colour with PbO2 Electrodes, Analytical Chemistry Laboratory chemistry program FMIPA UNDIP, Semarang.

7. Kristanto, W.A., Widodo, D.S., and Gunawan, 2012, Electrodestruction Elektrodestruksi Remazol Black B Dye Colour in Artificial Waste with Lead Dioxide Electrodes, Journal science and mathematic, Vol 20 (I), 16-20.

8. Apipah, L., Widodo, D.S., and Hastuti, R., 2013, Utilization Waste Accu Eletrodes at Electrodecolourization Process Dye Colour Solution, Chem info Vol 1, No. 1, 1-10.

9. Lee, Dong, Geun, 2008, Effect of Scale During Electrochemical Degradation of Naphthalene and Salicylic Acid, Thesis, Civil Engineering, Machigan State University, USA.

10. Rohmaniyah, A., and Setiarso, P., 2012, Using Bentonite as Carbon Paste Electrodes Modifier for Pb2+ Analysis at white Shellfish Meat (Corbula faba Hinds) on Voltammetre Pulse Diferential, Unesa Journal of Chemistry Vol. 3, No 1.

11. Zayed, S. I. M., and Arida, H. A. M., 2013, Preparation of Carbon Paste Electrodes and Its Using in Voltammetric Determination of Amiloride Hydrochloride Using in the Treatment of High Blood Pressure, International Journal of Electrochemical Science, 8, 1340 – 1348.

12. Soloman, P. A., Basha, C. A., Velan, M., and Balasubramanian, N., 2010, Electro oxidation of Malachite Green and Modeling Using ANN, Chem. Biochem. Eng. Q., 24, 445–452.



Suyanto et al. 220


Surabaya Indonesia

Benzyltriethylammonium chloride As Corrosion Inhibitor on Zinc

With Soaking Method

Suyanto1*, Djony Izak R2, Ofi Rulytasari1

1Chemistry Departement, Airlangga University 2 Physic Departement, Airlangga University

Corresponding author’s Email: [email protected]

Abstract.The aim of this research is studying the influence of phase transfer catalyst benzyltriethylammonium chloride (BTEAC) as corrosion inhibitor on zinc. BTEAC was synthesized from benzyl chloride with triethylammonium. BTEAC solution 0.5 M could inhibit corrosion rate on zinc with inhibition efficiency 87.109 % . Zinc plate soaked in solution of sulphuric acid 0.5 M during 30 days, and zinc plate washed with distilate water and HNO3 5%. The weigh of sample were measured before and after treatment . Corrosion rate was determinate by Nace equation.

INTRODUCTION Phase transfer catalyst (PTC) such as crown ether can be used as corrosion inhibitor or corrosion protection

[1]. The small quantity of chemical materials or agents which able to inhibit or prevent corrosion rate on metals or non metal are known as inhibitor agents[2-5]. The organic chemical compounds which have phi (π) electrons or contain hetero atoms O, N or S, are potential as inhibitor agent, because this condition lead chemical compounds to attach on metal surface [6-23]. BTEAC is one of PTC organic compound which has phi (π) electrons and contain hetero atoms N. Therefore BTEAC is predicted have properties as corrosion protection on metal.


Materials and instruments

Benzyltriethylammonium chloride was obtained by reaction of benzyl chloride with triethylammonium, whereas zinc was obtained from commercial zinc plate. The weigh of zinc plates were measured with digital analytical balance.

Methods

Zinc plates (2 cm x 2 cm) were cleaned by abrasive paper grade 240-800 and the weigh of zinc plates was measured by digital analytical balance. Zinc plates were soaked in 25 ml solution of sulphuric acid 0.5 M and were added with 25 ml solution BTEAC 0.25 M. After 30 days, to remove the solution, zinc plates were washed with distilate water and HNO3 5%, then zinc plates were dried in oven at temperature 40oC. The weigh of zinc plate were measured. these step is known as plate treatment. Procedure for negative control is the same as plate treatment, with exception of BTEAC solution adding. The corrosion rate determine with Nace formula V = KM/AtD…..(osi1), where V is corrosion rate

(mm/day), A is initial surface area (mm), t is time treatment (day), D is density(g/ml), M is mass droping (g) and K is constant 88378,25896. The inhibition efficiency is calculated with formula

efficiency = [(Vo –Vt) : Vo] x 100%........(2) , where Vo is corrosion rate without BTEAC(negative control) and Vt is corrosion rate with BTEAC.

RESULTS AND DISCUSSION FTIR analysis resulted spectra with wave number which is shown in Table 1.



Suyanto et al. 221


Surabaya Indonesia

Table 1.The spectra IR BTEAC product synthesis and literature [24,25]

Function groups Wave numberproduct synthesis (cm-1)

Benzen, 5 adjacent H 770-730756

C - N 1155 1167

Wave numbercm-1(24,25)

The inhibition efficiency BTEAC at various concentration is shown in Table 2

Table 2. Inhibition efficiency BTEAC at various concentration

BTEAC concentration(mol) Inhibition Efficiency(%)

64.25

0.25

0.10

75.68

0.50 87.10

Table 2 show that the more inhibitor concentration increase so the higher percentage inhibition efficiency. The reason is high concentration of BTEAC provides abundant availability as inhibitor, so contact process between corrosion media and zinc plate is decreased, and the rate corrosion is slow. BTEAC is organic compound which has phi (π) electrons and contain hetero atoms N, so this compound has properties as corrosion

protection on metal, via surface layer formation of zinc plates. Due to the surface layer formation, inhibitor hindrance contact between corrosion media and zinc plate.

The inhibition mechanism is BTEAC attach on surface metal, because of hetero atom N and aromatic group with phi (π) electrons. This condition is caused by contact of metal surface and corrosion media (H+ from H2SO4) which is hindrance, so corrosion rate is decreasing.

When Zn was soaked in corrosion media without BTEAC, proton from sulphuric acid react with Zn form Zn2+ , this cation react with oxygen form Zn2O3 which known as corrosion product.

Zn + 2H+ Zn2+ + H2

Zn

H+

H+ H+

H+

H+H+

H+H+

H+H+

H+

H+

H+

BTEACBTEAC BTEAC BTEAC

BTEAC

BTEACBTEACBTEACBTEACBTEAC

BTEAC

H+

H+

H+

H+

(a) (b)

Figure 1.Zinc metal condition in corrosion media (H2SO4) without BTEAC (a) and with BTEAC (b).

CONCLUSION Phase transfer catalyst Benzyltriethylamonium Chloride (BTEAC) is able to inhibit corrosion on Zn in 0,5 M

sulphuric acid with inhibition efficiency 87.1 %

REFERENCES 1.Weber E, Vogtle F, Neutrale Organische Komplex-ligaden und ihre Alkalikomplexe II- Kroenather und

Cryptanden als Reagenzien und Ktalysatoren(TeilC), Kantakte(Merck), 1978, 2, 16 In : Weber E, Crown Compounds-Properties and Practice, Phase Transfer Catalyst, Properties and Applications, Merck-Schuhardt, p. 70.

2. J. M. Bastidas, J. L. Polo, E. Cano, C. L. Torres, Journal of Materials Science, 2000, 35, 2637-2642.



Suyanto et al. 222


Surabaya Indonesia

3.O. L. Riggs, Jr., in C. C. Nathan (Ed.), Corrosion Inhibitors, NACE, Houston, TX, 1973,11. In: S.Papavinasan, Corrosion inhibitor, Canmet Materials Technology Laboratory, Ottawa, Ontario, Canada in Uhlig's Corrosion Handbook, Second Edition, Edited by R.Winston Revie.ISBN 0-471-15777-5, John Wiley & Sons, Inc., 2000, Chapter 59, 1089.

4. J.M.Bastidas, J. De Damborena and A. J. V´ azquez, J. Appl. Electrochem , 1997, 27, 345, in J. M. Bastidas, J.L. Polo, E. Cano, C. L. Torres, Journal of Materials Science, 2000, 35, 2637 – 2642.

5. A. Singh, E. E. Ebenso, and M. A. Quraishi, Review Article International Journal of Corrosion, vol.2012, 1

6. A.Shiri, M.Etman , F.Badost, Electrochim. Acta, 1996, 41, 429. In: J.M. Batidas , P.L.Polo, E.Cano , C.L.Torres, Journal of Materials Science, 2000, 35, 2637-2642.

7.M.A.Quraishi, R.Sardar, Corrosion,2002,58(9),748-755.In:A.Sing, A.K.Sing, M.A.Quraishi, The Open Electrochemistry Journal, 2010,2, 43-51.

8. A.K.Singh, M.A.Quraishi, CorrosionSci., 2010, 52(4) 1373-1385 in A.Sing, A.K.Sing, M.A.Quraishi, The Open Electrochemistry Journal, 2010,2, 43-51.

9. M.A.Quraishi, I.Ahamad, A.K.Singh, S.K.Shukla, B.Lal, V.Singh, Mater. Chem. Phys.,2008,112(3),1035-1039. In:A.Sing, A.K.Sing, M.A.Quraishi, The open Electrochemistry Journal, 2010,2, 43-51.

10. A.K.Singh, M.A. Quraishi, J. Appl.Electrochem., 2010,40(7), 1293-1306. In: A.Sing , A.K.Sing, M.A.Quraishi, The Open Electrochemistry Journal, 2010, 2, 43-51.

11.M.Lagrenee, B.Mernari, M.Bouanis, M.Traisnel, F.Bentiss, Corrosion Sci., 2002, 44(3), 573-588. In: A.Sing, A.K.Sing, M.A.Quraishi, The open Electrochemistry Journal, 2010,2, 43-51.

12. M.A.Quraishi, F.A. Ansari, J. Appl. Electrochem., 2003, 33(3), 233-238. In: A.Sing, A. K.Sing, M.A.Quraish The open Electrochemistry Journal, 2010, 2, 43-51.

13. M. A. Quraishi, Sudheer, E. E. Ebenso, International Journal of Electrochemical Science, 2012, 7, 9920 - 9932.

14. G.N.Mu, X,Li, F.Li, Mater. Chem. Phys.,2004,86,59-68. In: E. E. Oguzie, C.B.Adindu, C.K.Enenebeaku, C.E. Ogukwe, M.A.Chidiebere, and K. L. Oguzie, Journal Physical Chemistry. C, 2012, 116, 13603−13615.

15. E.E.Oguzie, Y.Li, F.H.Wang, J. Colloid Interface Sci., 2007, 310, 90-98. In: E.E.Oguzie, C.B.Adindu, C.K.Enenebeaku, C.E.Ogukwe, M.A.Chidiebere, and K.L. Oguzie, Journal Physical ChemitryC, 2012, 116, 13603-13615.

16. M.S.Morad, A.M.K.El-Dean, Corros.Sci. 2006, 48, 3398-3412. In: E.E.Oguzie, C.B.Adindu, C.K.Enenebeaku, C.E.Ogukwe, M.A.Chidiebere, and K.L.Oguzie, Journal Physical ChemitryC, 2012, 116, 13603-13615.

17.L.Tang, G.Mu, G. Liu, Corros. Sci., 2003, 45, 2251-2262. In: E.E.Oguzie, C.B.Adindu, C.K.Enenebeaku, C.E.Ogukwe, M.A.Chidiebere, and K.L.Oguzie, Journal Physical ChemitryC, 2012, 116, 13603-13615.

18. G. Moretti, F. Guidi, G. Grion., Corros. Sc.i,2004,46,387-403. In: E. E. Oguzie, C.B.Adindu, C.K.Enenebeaku, C.E.Ogukwe, M.A.Chidiebere, and K. L. Oguzie, Journal Physical ChemitryC, 2012, 116, 13603−13615.

19. E.E.Oguzie, G.N. Onuoha, A.I.Onuchukwu, Mater. Chem.Phys., 2005,89, 305-311. In: E.E.Oguzie, C.B.Adindu, C.K.Enenebeaku, C. E. Ogukwe, M. A. Chidiebere, and K. L. Oguzie, Journal Physical Chemistry. C, 2012, 116, 13603−13615.

20. S. M. A.Shibli, V.S.Saji, Corros. Sci., 2005,47, 2213-2224. In: E. E. Oguzie, C. B. Adindu, C. K. Enenebeaku, C. E. Ogukwe, M. A. Chidiebere, and K.L. Oguzie, Journal Physical Chemistry. C, 2012,116, 13603−13615.

21. N.Bouklah, B.Hammouti, A.Aouniti, M. Benkaddour, A.Bouyanzer, Appl.Surf. Sci., 2006,252,6236-6242. In:E.E.Oguzie,C.B.Adindu,C.K.Enenebeaku,C.E.Ogukwe, M.A.Chidiebere, and K.L.Oguzie, Journal Physical Chemitry. C, 2012, 116, 13603−13615.

22. M.A.Ameer, E.Khamis, G.Al-Senani, Adsorpt. Sci. Technol.,2000,18, 177-193. In: E.E.Oguzie, C.B.Adindu, C.K.Enenebeaku, C. E. Ogukwe, M. A. Chidiebere, and K. L. Oguzie, Journal Physical Chemistry.C, 2012, 116, 13603-13615.

23.H.Ashassi-Sorkhabi, Seifzadeh, M.G.Hosseini, Corros. Sci. 2008,50, 3363-3370. In: E. E. Oguzie, C. B. Adindu, C. K. Enenebeaku, C. E. Ogukwe, M. A. Chidiebere, and K. L. Oguzie, ,Journal Physical Chemistry. C, 2012, 116, 13603-13615.

24.Sigma-Aldrich,2010, Benzylriethylammonium Chloride, Sigma-Aldrich, USA. In: M.F.Budiningsih, Skripsi, Departemen Kimia, Fakultas Sains dan Teknologi Universitas Airlangga, Su rabaya , Indonesia, 2011, p.27.

25. DH William, I Fleming, Spectroscopic methods in organic chemistry, Mc Graw Hill Book Company (UK) Limited, London, Toko, Toronto, 1980, p.62.



Husen et al. 223


Surabaya Indonesia

The Effect of Mangosteen (Garcinia mangostana L) Pericarp Extract

towards Body Weight and Fasting Blood Cholesterol Level of Diabetic

Mice

Saikhu Akhmad Husen1* and Dwi Winarni1)

1Department of Biology, Faculty of Science and Technology Airlangga University Kampus C UNAIR, Jl. Mulyorejo, Surabaya – 60115 Indonesia

*Corresponding author: [email protected]

Abstract. This study aimed to explore the activities of crude pericarp extract of mangosteen (Garcinia mangostana, L), which now cultivated by many fruit farmers in Indonesia. This purpose was based on the fact that commonly mangosteen peel was considered rubbish and not yet to be utilized. Whereas, mangosteen peel contained various active compounds which had benefits for human‘s health. In-vivo antioxidant activity test of pericarp extract was conducted using animal model of 3-4 months old male mice strain BALB C weighted 30-40 g. Mice were divided into two group; normal control (KN) which not given STZ (streptozotocin) and STZ-induced diabetic mice. STZ induction was conducted using multiple low dose method, in which 30 mg/kg bw STZ was given as much as five times (daily induction). Diabetic mice were separated into 2 subgroups; diabetic control group (KD), diabetic-metformin HCl group (KM), and crude extract treatment group (P). Before and after STZ injection, each mouse was weighted and fasting blood cholesterol level of mice at each group was noted. After mice with fasting blood glucose and cholesterol level above 140 mh/dL were obtained, diabetic mice group then treated with crude magosteen pericarp extract, with doses for each group as following: P1 = 50 mg/kg bw, P2 = 100 mg/kg bw and P3 = 200 mg/kg bw. The measurement of fasting blood glucose and cholesterol level was conducted at 1st, 7th, and 14th day of extract treatment. Mice were treated with crude extract for 14th days. Results shown that crude mangosteen pericarp extract treatment able to elevate mice body weight and decrease fasting blood cholesterol level significantly compared to diabetic mice. Negative correlation was found between the increasing body weight and decreasing fasting blood cholesterol level of diabetic mice.

INTRODUCTION Diabetes mellitus is a metabolism disorder suffered by around 6% population on whole world. Diabetes

mellitus (CM) can be differentiated into two types; DM type 1 and DM type 2 [1]. One of DM causal factor is obesity [2]. Indonesia is one of the countries with relatively high number of obesity patients, following general increase of community income and development of unhealthy lifestyle. Based on data from International Diabetes Foundation (IDF), there are around 8.5 million diabetes cases in Indonesia caused by obesity factor alone [3]. Obesity is defined as abnormal condition in where excessive fat is accumulated inside the body and can lead to health risk. Obesity is main factor on various chronic disease, including diabetes mellitus, cardiovascular disease, and cancer. Obesity is a complex condition ensued from combination of several factors, such as genetic, cultural, behavior, and environment. The main cause of obesity is excessive energy intake unequal to long-term energy discharge [4]. Obesity inclination frequently happens to individual whose lifestyle has low-level activities but digest high-calories food with low level of micro nutrient [5].

Initially, obesity concern happen only on countries with high per capita income but later, obesity number also found to be increased on developing country, especially on urban area [6]. Obesity number currently has increased two fold compared to its number on 1980. Based on WHO data, there are 42 million children below age 5 years and 1.4 billion adult which are overweight and 500 million of them suffer from obesity [6]. Hyperlipidemia condition on obesity can increase oxidative stress inside the body, leading to various complications. Obesity patients also experience rising cholesterol level (hypercholesterolemia), caused by excessive fat accumulation on the body. One negative effect of obesity is the emergence of insulin resistance, the inability of insulin to continue delivering normal biologic function (the decline of tissue sensitivity towards insulin). Obesity patients develop resistance on cellular insulin action, indicated by the lowering of insulin ability to support glucose intake on fatty and muscle tissue, thus causing extended hyperglycemia condition [7].

Hyperglycemia directly leads to upsurge of Reactive Oxygen Species (ROS) and Reactive Nitrogen Species (RNS). ROS and RNS is able to directly oxidize and damage DNA, protein, and lipid. High level of ROS and RNS can also indirectly impair macromolecules. Reactive Oxygen Species (ROS) and Reactive Nitrogen Species (RNS) are very reactive molecules able to cause damage and oxidative stress. Oxidative stress happen when imbalance occur between highly-reactive molecules (ROS and RNS) and the existed antioxidant. In many previous studies, the dysfunctionality of β cells is found to be caused by high level of free fatty acid and




Husen et al. 224


Surabaya Indonesia

glucose. β cells are highly sensitive towards ROS and RNS because these cells don‘t possess much free-radical scavenging enzyme (antioxidant), such as catalase and superoxide dismutase. Those highly-reactive molecules (ROS and RNS) oxidized sulfhydryl group of protein, amino nitrate like tyrosine, and increase lipid peroxidase [1].

Antioxidant is a substance able to hamper negative effect of free-radicals by giving electron on damaged lipid, cell wall membrane, blood vessel, DNA, and other damage caused by reactive substance, like ROS. In order to reduce negative effect of the previously mentioned free-radicals, additional antioxidant from outside (exogenous antioxidant) is required, including vitamin E, vitamin C, and also other antioxidants obtained by digesting various fruit and vegetable containing high level of antioxidant. Indonesia possessed large riches on medicine potential from natural sources. This tropical country packed with natural resources is estimated to have around 30,000 species of plant and 7,000 of them has been known to possess medicinal value [8].

One of the plant used as basic material for traditional medicine in Indonesia is mangosteen (Garcinia mangostana, L). Mangosteen is a fruit-bearing tree originated from Southeast Asia, including Indonesia, Malaysia, Thailand, and Myanmar. Lately, mangosteen is given the name ―Queen of Fruits‖, because of its benefit as medicine for various disease, such as cancer, heart disease, arthritis, diarrhea, tonsillitis, vaginal discharge, and dysentery. In addition, mangosteen pericarp extract can also employed applied as anti-hypertension, anti-inflammatory, anti-microorganism, anti-diabetic, and even anti-HIV medicine [9]. Mangosteen pericarp contains an active substance called xanthone. In addition to possess anti-hypertension and anti-inflammatory activities, xanthone is also a formidable antioxidant compared to both vitamin C and vitamin E on scavenging free-radical, preventing cell damage, and delaying cell degenerative process [10].

This study was designed to answer the problem whether lard (pig oil) administered per oral can increase body weight and fasting blood cholesterol level of mice and whether mangosteen pericarp extract administration can decrease fasting blood cholesterol level of mice with type 2 diabetes mellitus which was induced using lard and STZ. This study aimed to prove than per oral administration of pig oil able to increase body weight and fasting blood cholesterol level of mice, and mangosteen pericarp extract administration can reduce fasting blood cholesterol level of mice with STZ-induced type 2 diabetes mellitus.

Result of the study was expected to be able to improve life quality of diabetic patients by means of utilizing local natural resources, increase economic value of tropical fruits, especially mangosteen which has potential as antioxidant, and extend additional knowledge to student about the potential of local material in our surrounding environment, which could be utilized for medicinal and treatment purpose of degenerative diseases.

MATERIALS AND METHOD

This study was designed as experimental research and conducted at Reproductive Biology Laboratory on Faculty of Science and Technology, and Institute of Tropical Disease (ITD) Universitas Airlangga. Sample used was 3-4 months old adult male mice strain BALB/C weighted 25-40 gram. Materials used including mangosteen (Garcinia mangostana, L) which was obtained from Soponyono fruit market, Rungkut, Surabaya, and other materials for following purpose: freeze drying (dry ice, technical ethanol), crude extraction (ethanol), diabetic mice induction using STZ (streptozotocin), buffer citrate solution with pH 4.5, and phosphate-buffered saline (PBS), extract solvent CMC (carboxymethylcellulose), standard anti-diabetic medicine (Metformin HCl 100 mg/Kg body weight), lard (pig oil), anesthesia substance (xylazine and ketamine), glucose (10% D-glucose dissolved in distilled water) [11].

Main instrument used in this study including mice pen, which was plastic container covered with gauzed wire, drinking bottle, feeding bowl, sawdust, microscope, petri dish, analytical scales with accuracy of 4 digits after decimal point, 2-3G injection needle with lead safeguard on its tip, 1 ml insulin injection needle for blood sample collection and diabetes induction, insulin syringe for glucose tolerance test, On Call Plus glucometer, Easy Touch glucostrips, blood cholesterol strips, glass wares, rotary vacuum evaporator, and empty table.

Research procedures were initiated by mangosteen pericarp crude extraction. Mangosteen part used for extraction was the pericarp or inside part of the peel. Samples were consisted of 24 male mice, divided into normal control (KN) group and STZ-induced diabetic group. Fasting blood glucose level was taken on 7th and 14th day after STZ induction. Fasting blood glucose level measurement using glucometer was conducted to determine mice diabetic condition. Only mice with fasting blood glucose level above 170 mg/dL would be continued used as diabetic mice on the research. Animal model was separated into these following groups; non-diabetic mice grouped into normal control (KN) group, while STZ-induced diabetic mice divided into 2 control groups of diabetic control (KD) and diabetic control given 100 mg/kg bw Metformin HCl (KM) and mangosteen pericarp extract treatment group. Treatment group was divided into 3 subgroups based on dose given; P1 was administered 50 mg/Kg bw extract, P2 was administered 100 mg/Kg bw extract, and P3 was administered 200 mg/Kg bw extract.



Husen et al. 225


Surabaya Indonesia

Each group was consisted of 4 mice. Per oral administration dose of pericarp extract referred to toxicity test result previously conducted by Saikhu et al. (2013), with resulting LD50 = 630.95 mg/kg bw. Treatment was conducted for 14 days. Body weight and fasting blood cholesterol level was measured for all mice group before and after lard administration, and at 1st, 7th, and 14th day of pericarp extract treatment. Body weight was measured using digital scales, while cholesterol level was taken using Easy Touch multi-function monitoring System, which was equipped with cholesterol test strips, and conducted after mice was fasted for 8 until 10 jam.

RESULTS Reading result on mice body weight and fasting blood cholesterol level mean change before and after lard

administration was presented on Table 1 and Figure 1. Meanwhile reading result on body weight and fasting blood cholesterol level mean change after pericarp extract treatment on 1st, 7th, and 14th day was presented on Table 2 and Figure 2.

Table 1. Mean result of reading on mice body weight and fasting blood cholesterol level, before and after lard administration.

Mean Body Weight (gram) Mean Cholesterol Level (mg/dL)

Before lard administration 29.4875 ± 2.7372 141.9583 ± 9.1864 After lard administration 31.4458 ± 2.8042 154.6250 ± 15.1939

Figure 1. Graphics showing the effect of lard (pig oil) administration on mice body weight (gram) and fasting blood cholesterol level (mg/dL). Alphabet above bars represented t test result at α = 0.05. Same alphabet indicated insignificant

difference. Different alphabet indicated significant difference.

Table 2. Mean result on reading of mice body weight and fasting blood cholesterol level change, after pericarp extract administration on 1st, 7th, and 14th day

Group Body Weight Change (gram)

Cholesterol Level Change (mg/dL)

KN - 0.900 ± 2.545 4.250 ± 15.456 KD - 3.950 ± 0.597 26.250 ± 26.837 KM - 1.080 ± 0.921 18.250 ± 9.742 P1 1.200 ± 1.116 -3.750 ± 16.296 P2 1.600 ± 0.752 -24.750 ± 21.187 P3 2.000 ± 1.070 -11.750 ± 31.372



Husen et al. 226


Surabaya Indonesia

Figure 2. Graphics showing effect of mangosteen pericarp extract on body weight and fasting blood glucose level change on mice with DM. KN = normal control. KD = DM control group without Metformin HCl treatment. KM = DM control-

metformin HCl 100 mg/Kg bw group. P1= DM treated with pericarp extract 50 mg/Kg bw. P2 = DM treated with pericarp extract 100 mg/Kg bw. P3 = DM treated with pericarp extract 200 mg/Kg bw. Mice were treated mangosteen pericarp extract for 14 days. Alphabet written above bars on each group indicated LSD test result on α = 0.05. Same alphabet

indicated insignificant difference.

DISCUSSION

From the result of mean body weight and fasting blood cholesterol level of animal model reading (Table 1 and Figure 1), before and after lard administration, significant increase on mice body weight was found (t test at α = 0.05). This showed that lard administration for 21 days was able to elevate body weight from average

29.4875 ± 2.7372 gram to average 31.4458 ± 2.8042 gram, and also able to raise fasting blood cholesterol level, from 141.9583 ± 9.1864 mg/dL before mice was given lard to be 154.6250 ± 15.1939 mg/dL after lard. The increasing body weight and fasting blood cholesterol level were caused by obesity and hyperlipidemia condition, both within blood circulation and animal model tissue. Obesity occurred because of excessive fat accumulation in the body tissue, leading to various chronic diseases to emerges, such as diabetes mellitus, cardiovascular disease, and cancer. The main case for obesity was excessive energy intake unequal to long-term energy discharge.

Obesity patient also suffered from hypercholesterolemia, which was caused by excessive fat accumulation inside the body. One of obesity negative effects was insulin resistance, the inability of insulin to deliver its normal biologic function. Obesity patients developed resistance on insulin cellular action, indicated with the declining of insulin ability to support glucose intake on fatty and muscle tissue, leading to continuing hyperglycemia condition [12].

Lard administration was conducted to induce hyperlipidemia and obesity condition, which was expected to aid the development of diabetic condition on mice before injected with STZ and treated with mangosteen pericarp extract. Results on mice body weight and fasting blood glucose level change reading was presented on Table 2 and Figure 2. From Table 2 and Figure 2 above, it could be seen that hyperglicemic condition was able to cause body weight decline on all mice from diabetic control (KD) and diabetic-Metformin HCl (KM), thus on 14 days of treatment, mice body weight was found to decrease averaging 3.95 gram on diabetic group and 1.08 gram on metformin HCl group. This was caused by STZ effect on damaging pancreatic β cells and mice had not able to conduct metabolism improvement on its body. On normal group, without STZ induction, where mice only given mineral water ad libitum and treated with CMC for 14 days, body weight also found to be decreased. On this case, mice was presumed to be affected by pen environment or CMC effect on suppressing increase of mice body weight through treatment. On P1 group, several mice was found to suffer declining body weight, but overall P1 group showed increase in body weight as much as 1.2 gram through treatment period, indicating improvement progress had started to occur.

From body weight measurement on pericarp extract treatment group at 50, 100 and 200 mg/Kg bw doses, it was found that all treatment group P1, P2, and P3 was able to reduce body weight decline significantly because of STZ-induced diabetic condition. Even the average value of body weight change at all treatment group was found positive. Thus, it was proven that mangosteen pericarp extract able to reduce blood glucose level caused by diabetic condition.

Result of Pearson correlation test showed that significant negative connection was found between mean body weight and mean decline of fasting blood cholesterol level. The higher mean body weight on mice, mean fasting blood cholesterol level was found lower. This was supported by correlation number result of -0.525, on α = 0.01. The increase of body weight on P1, P2, and P3 groups indicated improvement of Langerhans β cells,



Husen et al. 227


Surabaya Indonesia

affected insulin hormone to be produced higher, thus the amount of glucose absorbed by muscle and liver cells was also elevated, leading to improvement of mice body weight.

From the study result, it could be concluded that lard administration was able to increase body weight and fasting blood cholesterol level significantly, while mangosteen pericarp extract administration was able to elevate declining mice body weight caused by STZ injection and diabetic condition, and also able to lower fasting blood cholesterol level of STZ-induced diabetic mice significantly. Considering the result, it was recommended to give counseling to the community on the benefits of mangosteen pericarp extract on treating various metabolic disorders, especially diabetes mellitus. In addition, counseling was also recommended on other benefits of mangosteen as traditional medicine, such as for sprue, anti-cholesterol, and injury, which had proven more effective and efficient in scavenging free-radical compared to vitamin C and vitamin E.

REFERENCES 1. J.L Evans., I.D. Goldfine, B.A. Maddux, and G.M. Grodsky. ―Perspective in Diabetes: Are Oxidative

Stress-Activated Signali -Cell Dysfunction?‖, Diabetes 52, pp. 1-8 (2003).

2. WHO. Diabetes Mellitus. http://www.who.int/topics/diabetes-mellitus/en/. Accessed on November 3rd 2013. (2013).

3. International Diabetes Federation. Indonesia. http://www.idf.org/membership/wp/indonesia.Accessed on November 4th 2014 (2013).

4. G. Riccardi, P. Aggett, F. Brighenti, N. Delzenne, K. Frayn, A. Nieuwenhuizen, D. Pannemans, S. Theis, S. Tuijtelaars, B. Vessby,. ―PASSCLAIM - body weight regulation. Insulin sensitivity and diabetes risk‖, European Journal of Nutrition 43 (2), II7 – II46 (2004).

5. B.A. Swinburn, I. Caterson, J.C. Seidell, W.P. Diet James, ―Nutrition and the prevention of excess weight gain and obesity‖, Public Health Nutrition 7, pp. 123-146. (2004).

6. WHO. Diabetes mellitus. http://www.who.int/topics/diabetes-mellitus/en/. Accessed on 4 November 2014 (2014).

7. J.P. Clung, C.A. Roneker, W. Mu, D.J. Lisk, P. Langlais, F. Liu, and X.G. Lei. 2004. ―Development of Insulin Resistance and Obesity in Mice Overexpressing Cellular Glutathione Peroxidase‖, PNAS. 101(24), 8852-8857 (2004).

8. Bintang, D. Keanekaragaman Spesies Tumbuhan Berguna di Kawasan Lindung PT. Bukit Batu Hutani Alam (BBHA) Kabupaten Bengkalis Provinsi Riau. (Institut Pertanian Bogor, Bogor, 2011).

9. A.E. Nugroho, Manggis (Garcinia mangostana L) : Dari Kulit Buah Yang Terbuang Hingga Menjadi Kandidat Suatu Obat. http://en.wikipedia.org/wiki/purple-mangosteen.pdf, accessed on January 25th 2013 (2012).

10. Jung Hyun Ah, Bao Ning Su, William J. Keller, Rajendra G. Mehta and A. Douglas KingHorn, ―Antioxidant Xanthones From the Pericarp of Garcinia mangostana (Mangosteen)‖, Journal of

Agricultural and Food Chemistry, 54, 2077-2082 (2006). 11. B. Sharma, SK Satapathi and P. Roy, ―Hypoglycemic and Hypolipidemic Effect of Aegle marmelos (L)

Leaf Extract on Streptozotocin Induced Diabetic Mice‖, International Journal of Pharmacology 3(6), pp.

44-45 (2007) 12. J. Park, ―Increase in Glucose-6-Phosphate Dehydrogenase in Adipocytes Stimulates Oxidative Stress and

Inflammatory Signals‖, Diabetes, 55, 2939-2949 (2006).

http://www.who.int/topics/diabetes-mellitus/en/

http://www.idf.org/membership/wp/indonesia.Accessed

http://www.who.int/topics/diabetes-mellitus/en/

http://en.wikipedia.org/wiki/purple-mangosteen.pdf



Bogoriani 228


Surabaya Indonesia

The Activity of Andong Leaf Saponin (Cordyline terminalis Kunth.)

against Cholesterol and Diphenyl Picryl Hidrazyl (DPPH) in Invitro

Ni Wayan Bogoriani*

Department of Chemistry, Faculty of Mathemathic and Natural Science, University of Udayana, Indonesia Bali *Corresponding Author’s E-mail : [email protected]

Abstract. Saponins are plant glycosides that have biological activity. One is to prevent cardiovascular disease by lowering plasma cholesterol through inhibition of cholesterol absorption in intestine. The research is aimed to determine the activity of the leaf saponin of Cordyline terminalis against cholesterol and diphenyl picryl hidrazyl (DPPH) in vitro.This research is conducted experimental with the extraction and spectrometry method. Differences in mean levels between the groups were tested by One Way ANOVA followed by Post Hoc LSD test, in which the statistical testing is significant if p <0.05.The results showed that saponin is able to bind cholesterol 85.07% higher gemfibrozil 74.83% with significant different (p <0.05). Saponin at a concentration of 10 ppm; 50 ppm; 100 ppm; and 200 ppm are also able to reduce average the levels of free radicals of DPPH respectively 5.9%; 29.15%; 44.5% and 80.2% with significant difference (p <0.05) compared to control. Based on the results of this study concluded that the Andong leaves saponins can interact with cholesterol and reduce free radicals of DPPH in vitro. Saponins potential as a predictor of cardiovascular disease and as an antioxidant.

INTRODUCTION Cardiovascular disease is the first cause of death in both developed and developing countries.The whole

world is found 50 million deaths each year from heart disease, 39 million are in developing countries (Libby and Peter ,2006 ).

Dyslipidemia is a major risk factor for atherosclerosis which is the basis of cardiovascular disease. Atherosclerosis is a condition in which artery walls thicken as a result of the buildup of fatty materials such as cholesterol and became one of the risk factors for cardiovascular disease. (Charlton-Menys and Durrington, 2006).

Prevention of dyslipidemia and atherosclerosis is done by lowering cholesterol levels in the body is by consuming natural remedy that can lower blood cholesterol levels to normal limits. Reduction in blood cholesterol can be done by lowering food intake, inhibits the absorption of cholesterol, lower the endogenous synthesis, as well as increased spending bile and excreta (Charlton-Menys and Durrington, 2006). One of the natural remedies that can be used to reduce cholesterol in the blood is a saponin.

Saponin is a natural detergent which has surface-active properties, which molecular structures consisting of a steroid or triterpene aglycone called sapogenin and glikon containing one or more sugar chains (Osbourn, 2003; Guclu-Ustundag and Mazza, 2007; Vincken et al., 2007).

Liliaceae family is known as one of the families that are very rich in saponins. Garlic also includes family Liliaceae containing steroidal saponins (Matsuura, 2001). Andong plant including the family of Liliaceae which contains saponin. Based on the results of research that leaves of andong containing steroidal saponins (Bogoriani, 2001; Bogoriani, 2008).

Some researchers report that saponins from other plants such as garlic steroid glycosides (Matsuura, 2001), and chloragin of Chlorophytum nimonii (Lakshmi, et al., 2012) and also of alfalfa triterpene glycosides of Medicago sativa L. (Khaleel et al., 2005) soyasaponins (Sun-Ok Lee et al., 2005), Quillaja saponaria (Cheeke, 2001), and ginseng (Ha and Kim, 1984) have activity as hipokolesterolimea that can inhibit the absorption of cholesterol and lowering of plasma cholesterol concentrations that have been tested on animals and humans (Kim et al., 2003; Zhao et al., 2005; Afrose et al., 2010;). However, the mechanisms responsible for these activities is not clear. Saponins could be expected to prevent miselisasi cholesterol during digestion in the small intestine, thus reducing the availability of cholesterol for absorption into enterocytes. Inhibition of cholesterol absorption in the intestine allegedly due to saponins forming complex compounds with cholesterol, but its mechanism of action is unclear.

Prevention can also be done by reducing the cholesterol from food products that are marketed so that the food consumed is food that is low in cholesterol (Lu and Jorgensen, 1987; Williams and Coleman, 1992). Some researchers report that saponins can lower cholesterol from milk, creams and butter, the binding of cholesterol to form an insoluble complex, but its activity depends on several factors such as reaction time, temperature, and amount of saponin were added (Sundfeld et al., 1993a; Sundfeld et al., 1993b; Oh et al., 1998a; Oh et al., 1998b;



Bogoriani 229


Surabaya Indonesia

Chang et al., 2001). Based on these results prove that the saponins can form complexes with cholesterol and potential as an antioxidant in vitro


Extraction of saponin

Saponins are secondary metabolites isolated from leaves of andong by extraction and precipitation. Dry powder of andong leaves 500 grams put in a 2.5 liter beaker, was added n-hexane as much as two liters and allowed to stand for 24 hours to extract the lipids in a sample. The mixture was filtered and the filtrate is collected. Extraction is repeated in the same way four times, until the extractable lipid perfect. Subsequently the residue dried at room temperature until the free n-hexane, then macerated by as much as two liters of methanol for 24 hours. The mixture was filtered and the filtrate collected. Maceration repeated five times, until all the compounds that can be extracted with methanol perfect. Filtrate is collected, combined and evaporated. The methanol extract obtained partitioned between water and n-butanol. Then the n-butanol fraction was concentrated, washed with diethylether, dissolved in methanol, and then filtered. The methanol filtrate was then added diethylether excess to form a saponin precipitate. The precipitate is filtered and allowed to dry at room temperature (Mimaki et al., 1997; Ahmed, 2012).

Saponin- cholesterol interaction in vitro

To test of complex formation of cholesterol-saponin conducted with 600 ppm and 600 ppm cholesterol saponin placed in an Erlenmeyer flask of 300 ml and 5 mm glass marbles are added to improve mixing. Flasks were placed on a water bath while shaken. After contact cholesterol and saponins, zeolite is added directly to the flask, which is shaken in a water bath for 60 minutes. Once the process is complete the mixture disentrifugase at 3500 rpm, 4 ° C, for 25 minutes. Cholesterol-containing supernatant is separated and used for the determination of cholesterol with spectophotometric IKM 30 (Chang, et al., 2001).

Interaction saponin against DPPH

Saponin concentration in methanol are made 10 ppm, 50 ppm, 100 ppm and 200 ppm, by making the sample solution 1000 ppm by weighing 50 mg and then each of saponin which is then inserted into the flask and added methanol up to the mark. Saponin solution with a concentration of 10 ppm, 50 ppm, 100 ppm and 200 ppm, made by taking a saponin solution of 100 mL, 500 mL, 1000 mL and 2000 mL. Into each flask was added 2mL DPPH solution of 1 mM and diluted with methanol to mark boundaries. The mixture is shaken until homogeneous and then put in a dark container and allowed to stand for about 30 minutes then measured with uv-vis spectrophotometer at λ = 515 nm

Statistical Analysis

Analysis of statistical was performed with system of the statistical analysis. Values are expressed as mean±SD. Results were analyzed by one way ANOVA, and differences between treatments were determined by a least-significant-difference test (LSD). Alpa < 0.05 is used to determine statistically significant differences.


Saponin Isolates of Andong Leaf

Saponins andong leaves isolated by extraction and precipitation method. Saponin extraction is done by maceration using methanol, and fractionated with a solvent of water and butanol. Concentrated butanol fraction was dissolved in methanol and excess ether was added to obtain a saponin precipitate. a stable foam formed Showed sediment saponin. Hydrolysis process and the color reaction of Burchard Libermann, thus forming a green color indicates a steroid saponin in the andong leaves . The amount of the steroid saponin compound contained in the leaves of andong is determined by analysis of HPLC (high performance liquid chromatography) and shown in Figure 1 and 2. The chromatogram in Figure 1 there are 4 compounds that prove the existence of saponin, two saponins (peak-1: 30% and peak-3: 51 %) the structure is already known steroid saponin spirostanol (Bogoriani 2001 and Bogoriani 2008) and two other compounds (peak-2: 5% and peak-4: 14%) suspected steroid saponins furostanol as evidenced by Ehrlich reagent to form a red color (Mahato , et al., 1982).



Bogoriani 230


Surabaya Indonesia

Test of in vitro activity against isolates of saponin with free radicals of DPPH (Diphenyl Picryl Hidrazyl) can be seen in Figure 2, which shows that at concentrations of 10 ppm, 50 ppm, 100 ppm and 200 ppm respectively were able to bind free radicals 5.9% ; 29.1%; 44.5% and 80.2%. Isolates of saponin at a concentration of 200 ppm have activity against free radicals of DPPH highest. Results of saponin with cholesterol activity performed in vitro as shown in Table 1 and Figure 2

Figure 1 high-performance liquid chromatogram with YMC ODS-AQ column 5 μm 120 A0 250 x 4.6 mm, mobile phase mixture of acetonitrile-water-acetic acid (50: 50: 0.05), a flow rate of 1 ml / min, the temperature column 21°C, with a UV

detector and a temperature detector 115°C

Figure 2 Structure saponin of andong leave ( Bogoriani, 2001)

Table 1. Saponin absorption of DPPH

Group DPPH Control 0.858±0.007b c d e

Saponin 10 ppm 0.807±0.007a c d e

Saponin 50 ppm 0.608±0.007a b d e

Saponin 100 ppm 0.476±0.007a b c e

Saponin 200 ppm 0.170±0.007a b c d

DPPH : Diphenyl Picryl Hidrazyl a Represents significant difference from control p< 0.05 bRrepresents significant difference from saponin 10 ppm p< 0.05 c Represents significant difference from saponin 50 ppm p< 0.05 d Represents significant difference from saponin 100 ppm p< 0.05 e Represents significant difference from saponin 200 ppm p < 0.05

D-fukopiranosida

L-ramnopiranosida

L-ramnopiranosida

sapogenin steroid spirostan

O

H

HO

HO

OH

HO

OH

H3C

HO

H3C

H

H

H

H

HH OH

HO

OH

O

27

25

2423

26

22

21

20

1917

18

161514

1312

11

109

87

65

432

1

O

OHO

HO

H

H

H

H

HCH3

H

HH

O

H2C

CH3

CH3

CH3

O

O

O(1,37 ppm, s)

0,84 ppm, s)(

1,02 ppm, d J = 6,6 Hz)(

4,01 & 4,44 ppm, dd, J = 12,0 Hz)

(

4,71 & 4,79 ppm, br s)(

5,52 ppm, br d J = 5,4 Hz)

(

4,57 ppm,d, J = 7,0 Hz)

(

5,56 ppm, br s)(

6,43 ppm, br s)(



Bogoriani 231


Surabaya Indonesia

Figure 2. Antioxidant activity of andong leave saponin to bind free radicals DPPH at λ 515 nm.

Saponin interaction with cholesterol in vitro

Saponin interaction with cholesterol also conducted in vitro with the results as shown in Table 2 and Figure 3.

Figure 3 Comparison of average cholesterol reacts in vitro

Table 2 Average consentratrion of cholesterol reacts in vitro Group cholesterol reacts

(ppm)

p cholesterol reacts

(%)

Cholesterol 598,00 ±1,08 - -

Cholesterol + Saponin 508,75 ± 6,18a 0,00 85.07

Cholesterol + Gemfibrozil 447,50 ± 1,29a 0,00 74.83 aRepresents significant difference from cholesterol p < 0,05.

Based on the results in Table 1, saponin of andong leaves can reduce free radicals of DPPH with a significant difference of 80.2% at 200 ppm compared to control. Isolate saponins provide consentration inhibitor (IC 50) of about 115 ppm which is a value that describes the ability to isolate saponins in reducing free radicals by 50%. Saponins of andong leaves have antioxidant activity with IC50% values of 115 ppm. Saponins are also able to bind cholesterol by 85.07% higher than gemfibrozil (74.83%). It showed that saponins may reduce blood cholesterol in vivo.

CONCLUSION Based on these results of the study, it can be concluded that the andong leaves saponins can reduce free

radicals of DPPH with a significant difference (p<0.05) of 80.2% at 200 ppm compared to control dan can reacts with cholesterol by 85.07% higher than gemfibrozil (74.83%).

0.0010.0020.0030.0040.0050.0060.0070.0080.0090.00

0 50 100 150 200 250

Consentration (ppm)

REDUCTION

%

508.75 ± 6.18

447.5 ± 1.29

Saponin Gemfibrozil

Cholesterol reacts (ppm)

115



Bogoriani 232


Surabaya Indonesia

ACKNOWLEDGEMENT The study was supported by my family. The author would like to thank all who have gives assistance during

the study.

REFERENCES 1. Afrose, S., Hossain, Md. S., Salma, U., Miah,A.G., and Tsujii, H. 2010. Dietary karaya Saponin and

Rhodobacter capsulatus Exert Hypocholesterolemic Effects by suppression of Hepatic Cholesterol and Promotion of Bile Acid Synthesis in Laying Hens. Cholesterol :272731 PMCID: 3065839. P. 1-9.

2. Ahmed, D. B., Chaieb, I., Salah, K. B., Boukamcha, H., Jannet, H. B., Mighri , Z. and Daami-Remadi, M. 2012. Antibacterial and antifungal activities of Cestrum parqui saponins: possible interaction with membrane sterols. International Research J. of Plant Sci. (ISSN: 2141-5447): 3:1: 001-7. Available online http://www.interesjournals.org/IRJPS.

3. Bogoriani, W. 2001. Isolasi d totalan Identifikasi Senyawa Saponin dari Daun Andong (Cordyline terminalis Kunth.) Chemical Review. 4: 3: 92-7.

4. Bogoriani, W. 2008. Isolasi dan Identifikasi Glikosida Steroid dari Daun Andong (Cordyline terminalis Kunth.) Jurnal Kimia. 2: 1: 40-4.

5. Chang, E.J., Oh, H.I and kwak, H.S. 2001. optimization of Cholesterol Removal Conditions from Homogenized Milk by Treatment with Saponin. Asian-Aust.J. Aninm.Sci., 14:6:844-9.

6. Charlton-Menys, V. and Durrington, P. N. 2007. Human cholesterol metabolism and therapeutic molecules. Experimental Physiology – Review Article:93.1:27-42.

7. Cheeke, P.R. 1996. Biological effects of feed and forage saponins and their impacts on animal production. Saponins Used in Food and Agriculture. In GR Waller and K Yamasaki (Eds.), Saponins Used in Food and Agriculture: 377-85. New York: Plenum Press.

8. Cheeke, P.R. 2001. Actual and potential applications of Yucca schidigera and Quillaja saponaria saponins in human and animal nutrition. Recent Advances in Animal Nutrition in Australia, 13:115-26.

9. Ha, C.J. and Kim, S.H. 1984. Effects of ginseng saponins on cholesterol solubility. Saengyak Hakhoechi. 15:134-38.

10. Guclu-Ustundag, O., and Mazza, G. 2007. Saponins:Preperties, Application and Processing. Critical Reviews in Food Science and Nutrition, 47:3:231-58.

11. Khaleel, A.E., Gad, M.Z., El-Maraghy, S.A., Hifnawy, M.S. and Abdel-Sattar, E. 2005. Study of Hypocholesterolemic and Antiantherosclerotic Propertis of Menticago sativa L. Cultivated in Egypt. J of Food and Drug Analy,13:3: 212-18.

12. Kim, S.-W., Park, S.-K., Rang, S.-I., Kang, H.-C., Oh, H.-J., Bae, C.-Y., and Bae, D.-H. 2003. Hypocholesterolemic property of Yucca schidigera and Quillaja saponaria extracts in human body. Arch. Pharm. Res. 26:1042-46.

13. Lakshmi, V., Mahdi, A.A., agarwal, S. K . and Khanna, A. K. 2012. Steroidal saponin from Chlorophytum nimonii (Grah) with lipid-lowering and antioxidant activity. Original article, 3:227-32.

14. Libby and Peter. 2006. Inflammation and Cardiovascular Disease Mechanisms. American Journal of Clinical Nutrition:83(suppl), 456s-60s.

15. Lu, C.D., and Jorgensen, N.A. 1987. Alfalfa saponins affect site and extent of nutrient digestion in ruminants. J of Nutr,117: 919–27.

16. Mahato, S. B., Ganguly, A. N., and sahu, N. P. (1982). Review: steroid saponins, Phytochemistry, 21 (5), 959-978.

17. Matsuura, H. 2001. Saponinsin garlic as modifiers of the risk of cardiovascular disease. J. Nutr.131:1000S-5S.

18. Mimaki, Y., Kuroda, M., Takaashi, Y., and Sashida, Y. 1997. Steroidal glucosides from leaves of Cordyline stricta, Phytochemistry, 45:6: 1229-34.

19. Oh, H.I., Chang, E.J and Kwak, H.S. 1998a. Condition of the removal of cholesterol from milk by treatment with saponin. Korean J. Dairy Sci.20:253-90.

20. Oh, H.I., Chang, E.J and Kwak, H.S. 1998b. Condition of the removal of cholesterol from cream by saponin treatment. Korean J. Food Sci. Ani. Resour,18:224-31.

21. sbourn, A.E. 2003. Saponins in cereals. Phytochemistry, 62: 1-4. 22. Sundfeld, E., Yun, S., Krochta J.M., and Richardson, T. 1993a. separation of cholesterol from

butteroil using quillaja saponins. Effects of pH, contact time and adsorbent. J.Food Process Eng. 16:191-205.



Bogoriani 233


Surabaya Indonesia

23. Sundfeld, E., Yun, S., Krochta, J.M., and Richardson, T. 1993b. separation of cholesterol from butteroil using quillaja saponins. Effects of temperature, agitation and concentration of quillaja saponin. J.Food Process Eng. 16:207-26.

24. Vincken, J.-P., Heng, L., De Groot, A., and Gruppen, H. 2007. Saponins, classification and occurrence in the plant kingdom. Phytochemistry, 68: 275-97.

25. Williams, A.G. and Coleman, G.S. 1992. The Rumen Protozoa. New York: Springer Verlag New York Inc.

26. Zhao, H.L., Sim, J-S., Shim, S.H., Ha, Y.W., Kang, S.S. and Kim, Y.S. 2005. Antiobese and hypolipidemic effects of platycodin saponins in diet-induced obese rats; evidences for lipase inhibition and calorie intake restriction. International J of obesity,29: 983-90.



Utomo et al.. 234


Surabaya Indonesia

Integrated Model of One Parameter Logistic Model and Response

Time Model

Noer Hidayah1*, Kumaidi2, Badrun Kartowagiran2

1STAIN Kediri, 2Universitas Negeri Yogyakarta

Corresponding Author’s E-mail : [email protected]

Abstract. Currently, computer based test is becoming popular to use in several areas. For instance, the implementation of computer based test in senior High School National Examination in Indonesia. Theuse of computer in a test can record not only the response of the examinee but also the time taken for doing the examination which is called as time response. This time response cannot be obtained in a paper and pencil test. Time response should be accounted in determining the examinee ability and the test item difficulty. An examineewho responses an item correctly and rapidly should be scored differently from those who take a longer time to make a correct answer. However, measurement theories on examinee‘s time response have not been developed many. Therefore, this study aims to extend a

measurement model of item response theory by integrating time response into One Parameter Logistic Model (1PLM). The goal is improving calibration process to achieve the estimation of item-related-item difficulty and examinee ability more accurately. Response time is modeled based on lognormal distribution, where the average log of time response is the function of speed and effort needed by an examineein completing a test. The model development uses joint distribution concept. Thecalibration method to obtain the parameter estimation uses Bayesian Method, Markov Chain Monte Carlo.

INTRODUCTION Item Responsse Theory (IRT), modeling of the latent trait (the ability of test takers) and the characteristics of

the items, (such as the level of difficulty, discriminant item and guessing factor) just based on the pattern of response test participants, namely by looking at the wrong answer and the correct test participants in a test , IRT concept suitable to be applied on pure power test in which the test should not be restricted by time [1].

Hambleton and Swaminathan [2] states: An implicit assumption of all commonly used item responsse models is that the test to which the model are fit

are not administered under speed conditions... when speed affects test performance, then at least two traits are impacting on test performance: speed of performance, and the trait measured by the test content.

Van der Linden and Hambleton [3] also stated: The concept of power and speed test exist as idealization..... to analyze examinee behaviour on most exixting

test realistically, models of hybrid nature are needed. The fact that these models should addres the power and speed aspect of test has important consequences. The first is that items responsse and thier latencies... both have to be collected.

Three things that need to be considered in assessing the cognitive abilities of the participants. They are the answer (response), response time and confidence (confidence) test participants. [4] Several other studies mention that the time limit in a test affect test takers [5], speededness in tests and parameters influencing the estimation of examinee ability and characteristic items [6]. Results of research conducted by Abdelfattah suggest the response time should be considered and integrated into the scoring process, because there are differences in parameters estimation if taken account response time and obey the response time.[7] The other researchers consider the response time or speed of the test participants as a component to measure the ability of test takers. These studies reinforce that the response time should be considered in the modeling of latent trait. [8],[9],[10]

In the concept of measurement, modeling the response time provides information on the speed of the test participants, while modeling the item response provides information about the accuracy in answering the questions. Speed refers to how fast the test-takers answer the question, while accuracy indicates how correct answers given by participants of the test.

At the beginning of history, speed and accuracy is considered can measure the same psychology construct. This implies the assumption that there is no difference in the ability of the test participants, if only measured by accuracy, speed alone or a combination of the two. However, further research shows that speed and accuracy are not measuring the same construct in the complex test. So when this accuracy and speed is believed to be a variable that can measure different abilities, so both need to be considered in assessing the ability of the test taker.



Hidayah et al. 235


Surabaya Indonesia

The relationship between speed and accuracy is called with trade-offs. Although the relationship between speed and accuracy is a topic that has long been discussed, but relatively ignored in psychometric. [11]

THE FOUNDATION OF THEORY Theories about the item response theory has grown rapidly and applied widely . While theories about the

response time although it has been growing, but not as much as the development and application item response theory. Although the theory of response time has long rolled out, but this theory tend to be overlooked in psychometrics.

Item Response Theory (IRT)

Item response theory is known as a new approach to the theory of measurement. The main concept is item response models. Item response model is a mathematical equation that states the probability of correct responses in the test-takers is a function of its ability. Measurement model of item response theory based on two things: (1) performance of test participants on an item can be predicted by a set of factors called latent trait (ability of examinee and characteristic of item) (2) The relationship between the performance of test participants in a question and a set of trait can be described in a increasing monotonous function called item characteristic function (ICC). ICC illustrates the relationship between the pattern of the test taker's ability level and a probability of correctly answering a question controlled by characteristic item. Function characteristics of this clause states that increasing the level of one's ability, also increase the probability of correctly answering a particular item.

Item response modeis developed model for test questions which form such dichotomous. There are three models for this, One Parameter Logistic Model (1PLM), Two Parameter Logistic Model (2PLM) and Three Parameter Logistic Model (3PLM). All three models are distinguished by a number of parameters that exist in the model. One Parameter Logistic Model (1PLM) has a item difficulty level parameters (b) and the ability level (), with the following equation.

exp1 ,

1 exp

i j

ij i j

i j

bP X b

b

(1)

Two-Parameter Logistic Model (2PLM) has a item discriminat parameters (a), item difficulty parameter (b)

and examinee ability level (), with the following mathematical equation.

exp1 ,

1 exp

j i j

ij i j

j i j

a bP X b

a b

(2)

Three Parameter Logistic Model (3PLM) has guessing factor parameter (c), item difficulty parameter (b) and examinee ability level () , with the following mathematical equation.

exp1 , 1

1 exp

j i j

ij i j i i

j i j

a bP X b c c

a b

(3)

Response Time Modeling

The response time is the time required by the test participants to read and answer a question test [12], [13]. Other literature, such as [7],[14] calls response time be the response latency. The topic of the actual response time has long been discussed. However, this topic is underdeveloped and not widely applied because of record response time is difficult. Recording the response time can only be done if the test using a computer, either Computer Adaptive Test (CAT) and Computer BasedTest(CBT).

Along with the development of science and technology, the subject of the response time experienced growth, especially in the field of psychology and education. Several studies and research related to the analysis of the response time is done with a variety of purposes, including (1) to improve estimating process in the ability of test takers (the scoring) and the parameters of item [11],[16],[17], [18],[19],[20],[21]. (2) to analyze the speed of test participants in the test and the estimated time required to solve a problem [22] (3) to test the relationship between the velocity components and component accuracy in completing the test [23],[24],[25] (4) to improve estimates



Hidayah et al. 236


Surabaya Indonesia

of the parameters, especially if the test conditions occur speededness [6], [26], [27],[28] (5) to diagnose ability of the test participants and the learning process [29],[30] (6) To know how much effort in the test participants completed the test, whether working on themselves or using suitable suitable strategy or merely guessing the answer. [31] (7) To detects unusual patterns of answers in CAT [32]

Modeling the response time for the purposes of this done in 3 different ways or strategies. The first strategy is to model the response time exclusively. This strategy can be applied to the test with a very simple matter with strict time limits and the data provide very limited information. The second strategy implies a separate analysis between response time model and item respon model. Separate analysis gives some information about the accuracy of the response and the response time, but the relationship between the two variables that can not be modeled and assumed to be independent. The third strategy is to combine the response time model to the item response model simultaneously. (Ethink). The discussion in this research led to the first goal, that estimating the ability of test takers (scoring process) and the parameters of matter. Development of the model is done by entering the response time to the item response model.

Development Simultaneous Model Between Time Response and Item Response

Simultaneous model between the response time and the item response is the item response model which includes the response time. In other words enter the response time into IRT models. The research about simultaneous model the response time and the item response was a little. Some experts who developed the model are Thissen [11], Verhelst, Verstralen & Jansen [18], Roskam [16], Wang and Hanson [19], Wang [33], Ingrisone [20] and Ingrisone II [21].

Verhelst et al. [18] and Roskam [16] included the response time model into Rasch models, and the model is suitable for testing in the form of speed test. Thissen Model [11], Wang and Hanson model [19], Ingrisone model [20] and Ingrisone II model [21] is applied to power test. Models can be classified into power test or speed test based on the item characteristic curves.

Thissen [11] stated 2PLM in as follows.

1

( 1)1 exp( )

ij

ij

P rz

(4)

Where : ij j i jz a c and

( 0) 1 ( 1)ij ijP r P r

i state the effective ability of the participant to-i, ja state item discriminant parameter to item j. and jc

state guessing factor. rij states answer to the i-th participant in answering questions j. rij = 1 if the answer is correct and rij = 0 if the participant answers incorrectly.

Roskam model as the Van Der Linden & Hambleton, (Eds) [17] is as follows :

exp1 , ,

1 exp

j ij ij ijij ij

j ij i j ij i

tP U t j i

t

(5)

where is the ability of the test participants, i is the level of difficulty and tij is the response time. ,

and are the logarithm of , and t. The equation 5 states that the effective ability to item-i is a process to mental speed of test participants. Rational of model is effectively the ability of the test participants answered questions will go up if the time used to answer questions also added. The effective mental speed of test participants is one of determinant factor to increasing it. . Roskam using the Weibull distribution –one parameters as the marginal distribution of response times.

Limitations of model proposed by Verhelst et al. improved by Wang and Hanson [19] . Wang and Hanson combines the response time to the 3PLM that can be applied to a power test. The proposed model is as follows:

1

1 , , , ,

1 exp 1.7 ( )

jj i i i j j j ij j

i jj i j

ij

cP a b c t c

da bt

(6)

a, b, c and are regular parameters as 3PLM, d is a parameter that states slowness in question (item slowness parameter), t is the response time of test participant to item j , is a parameter that states the slowness of



Hidayah et al. 237


Surabaya Indonesia

participants test (examinee slowness). Wang [33] chooses Weibull distribution -one parameters chosen as the marginal distribution of the response timebecause of the simplification.

Ingrisone [20] proposes simultaneous model as follows.

exp 1,71 , , ,

1 exp 1,7

j i ij j

ij ij i j j

j i ij j

a t bP u t a b

a t b

(7)

Equation 7 represents the probability of correct responses given response time of test participant, ijt. i

states examinee‘s ability i, where i = 1,2,3, ..., N. aj state item discrimination parameter to j (j = 1,2,3, ... J), bj expressed about the item difficulty level parameter j (j = 1,2,3, ... J) and tij is a response time by test participants of the i to the item j-th. η is a constant value which states regression coefficients of response time. If there is no effect on the response time, the probability of correct responses is zero. If there is effect of response time on the probability of correct responses, the magnitude of η is not equal to zero. Ingrisone chooses lognormal distribution as a marginal distribution of response time.[20]

Ingrisone II [21] proposes a simultaneous model as follows.

exp1,71 , , ,

1 exp1,7

i ij j

ij ij i j

i ij j

t bP x t b

t b

(8)

Equation 8 represents the probability of correct response is a function of the ability of test takers ( i ), the

level of difficulty ( jb) and response time ( ijt

). is coefficient of a response time variable which show the effects of response time to the probability of correct responses. If the amount is not equal to zero, then the response time affect the probability of correct responses and if equal to zero then the model becomes 1PLM. Ingrisone II select two parameter Weibull distribution as the distribution of response time. The distribution is the development of a single parameter Weibull distribution. [21]

RESEARCH METHODOLOGY This study is basically a development research, namely to develop 1PLM by inserting of response time

variable. Development of the model is intended to obtain a more realistic mathematical models that can be used in the practice of measurement. IRT models simulate the response of participants test based only on the answer (true-false) of test item without considering the response time. Whereas in fact a test is always limited by time. Therefore, incorporation the response time into the model, is expected to improve the concept of IRT. Estimation of examinee‘s ability and test items parameter is calibrated in accordance with the conditions that exist in a test.

Development of the model is done with the following procedures. Step 1: Formulate a simultaneous model between the item response and a response time . Formulation of a model based on the study of the theory associated with the simultaneous model , the item response and a response time.

Step 2: Estimate the parameters of the model with the Bayesian approach Markov Chain Monte Carlo. The model testing is done using simulation data. Simulation data needed to obtain broader conclusions or

generalize a variety of conditions that may be in a test. Applications in various conditions is also used to see the robust nature of the resulting mathematical model. Simulation studies conducted to generate data based on algorithms generated in the mathematical model. Simulation data in this study be set based on the number of participants and the number of test test, which is the number of participants to 500 person and the number of test items to 11.

RESULTS AND DISSCUSSION

The parameter logistic model with the integration of response time

This study intends to develop 1PLM, with a response time integrating into the model. The model is a form of simultaneous models between item response and response time. Model offered in this study will essentially improve the model developed by Ingrisone II [21] and Roskam [16] in which the both of the researchers also integrates variable response time into 1PLM Ingrisone stated model suitable for power test, but according to item



Hidayah et al. 238


Surabaya Indonesia

characteristic curve, Ingrisone II, model appropriately used for the speed test. The model that will be developed in this study is used to measure achievement test participants (achievement test). Measurements of achievement in the form of test-takers. The test item is a dichotomous, so the response of the test participants are true and false answers. About the difficulty level of the test parameters, as in models 1PLM The parameters of test participant that measured are test participants' ability and speed (measured in terms of response time)

The model was developed for the power test. Power test was used to determine the level of mastery of the material and measuring the ability of test participants. The item in the power test has a level of difficulty varies. Implementation of the power test is theoretically not limited by time, despite the fact that each test is always limited by time. Answer of participants test is not always true (probability of correct answer is not always high) though done in infinite time. The definition is used as the basis for preparing the integrated 1PLM with response time.

Mathematical model in item response theory conceived to capture the relationship between the variables with the parameters contained in the measurement process. Curve that illustrates the relationship between the variables or parameters in the simultaneousmodel between the response and the response time is called the Conditional Accuracy Function (CAF). CAF represents the probability of correct response is a function of the ability of test takers and the time required to complete the test questions.

CAF concept is basically the same as the concept of the ICC, which is a increasing monotone function. Ranger states the relationship between the test scores in the measurement response time test takers achievement (achievement test) is a increasing monotonous function, whereas in the measurement of the attitude and personality of a non tests that relationship is expressed in the function of the inverted-U. The function causes the difference in achievement measurement model and the attitude and personality measurement model. [34] CAF represents the probability of correct response test participants will be higher if the ability of test takers getting higher and available more time. Based on the above elaboration, 1PLM with the integration of response time is formulated as follows.

exp

, , , 1 , ,

1 exp

ji j

ij

ij ij i j j ij ij i j j

ji j

ij

db

tf x t b d P x t b d

db

t

(9)

Equation 9 shows the probability of correct response influenced by the ability of test takers ( j ), examinee response time (tij) and item difficulty level (bi). If tij close to infinity, then the equation 9 same as the 1PLM. The magnitude of tij to infinite can mean that the test is not limited by time. The relationship between the probability of correct responses and the parameter as well as equation 9 is described in the following CAF.

Figure 1. CAF of 1PLM With Response Time Integration

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

1

10

20

30

40

50

60

70

80

90

10

0

11

0

12

0

13

0

14

0

15

0

Probability of correct response

Response time

b = 0,5 j = 2

b = 0,5 j = 1 b = 0,5 j = 0,5

b = 0,5

j = -1



Hidayah et al. 239


Surabaya Indonesia

CAF in the figure 1 are arranged on the same difficulty level of test conditions (bi) that is equal to 0.5 and

levels of ability ( j ) is different. It is intended to compare several conditions that show the difference between the participant's ability and level of difficult. Figure 4.1 desribes the conditions which difference between the ability of test takers and the level of difficulty is positive and negative. The graph in all conditions compared indicate increasing monotonous function. The CAF means more time spent to complete the test item, the higher the probability of correct response. The greater of the difference between the examinee ability level and the level of difficulty make in higher probability of correct response. It means that test participants with a higher capacity will have a higher probability of correct response when compared with test participants with lower capabilities.

Response time, tij, as in the equation 9 is fixed variable, so as if the probability of the response time is only determined by the response time recorded by the administration of the test during the test. In fact the response time of examinees may vary according to the their condition. An examinee doing same test for several time (the test questions and setting are the same) will require the different response time to complete test. Rationally, the condition and strategies of the examinees used in completing the test questions may change from time to time and it affects the speed and response time. This shows the response time should not be fix variable, but the random variable. The response time as random variables in statistics expressed in the form of distributions. The response time is empirically has a skewed shape and is positive. Some form of distribution that meets the characteristics of the response time is a Lognormal distribution, Weibull distribution and Gamma distribution.

The distribution selected as the marginal distribution of response time in this study is the lognormal distribution that has been developed by Van der Linden [29]. The difference lies in the sense of a varians on the lognormal distribution. Van der Linden [25] defines a variance as the discriminant factor between the participants of the test. This study defines the variance in shape parameter. The resulting mathematical model as follows.

2

ln1 1; , exp

22

ij j i

i i j

ij

tf t

t

(10)

j i is the average of response time. j is the amount of effort required to complete the item test j. The

bigger of j , the amount of effort required by the examinee to complete item i also getting bigger. i is the speed of the i-th examinee in resolving item. The higher speed of the i-th test participants, cause the less time

required to complete the test item j, ,i

. is a variance. Simultaneous model between item response and response time, as in the equation 9, combined with the marginal distribution of the response time, as in the equation 9, by using the concept of joint distribution.

The Estimation of Parameter.

The method of parameter estimation used in this study are Bayesian methods, Markov Chain Monte Carlo. In the Bayesian approach, parameters in the model is a random variable. All kinds of information is expressed in the form of probability distributions. Parameter estimation is performed by determining prior distribution and likelihood distribution. Prior distribution of each parameter indicates the uncertainty in the actual value (true value) before the data was obtained. While the likelihood distribution is determined from data. Prior distribution is then combined with likelihood distribution to get the posterior distribution. Conclusions or parameter estimation is based on the posterior distribution.

Prior distribution of each parameter is selected based on the domain of parameters. It is determined according to the theory. Theoretically, the ranges of difficulty level from infinity negative to infinity positive,.

,ib , so that the distribution of prior is normal distribution, (0,1)b N

. The examinee‘s ability

level is ranged from infinity negative to infinity positive,. ,i

, so that the distribution of prior is

normal distribution (0,1)i N

. The amount of effort required to complete the test item j is lied from

infinity negative to infinity positive,. ,i

The prior distribution for j is normal distribution



Hidayah et al. 240


Surabaya Indonesia

(0,1)i N. The range of speed of the examinee in completing test

j is infinity negative to infinity

positive, ,j

. So the prior distribution for j is (0,1)j N

. The amount of jd is positive, so

that the prior distribution chosen is normal distribution with positive value, (0,1)jd N

, 0jd

. Variance of

lognormal distribution 2

is set 1, so that the selected prior distribution is uniform distribution. Likelihood distribution is determined based on the data. The data in this study is a response or answering in

the form of dichotomous, that is, if the correct answer is scored one, and if the wrong answer is scored null. Therefore the likelihood distribution that selected is a Bernoulli distribution. The concept used in this study is a joint distribution between Bernoulli distribution and lognormal distribution as a marginal distribution of response time. Full likehood distribution as follows.

1

2exp1.7ln

1 1 1, , , , , exp

221 exp1.7 1 exp1.7

x xij ijd jbi j tt ij j iij

L b d x td d tj j ij

b bi j i jt tij ij

1 1

N J

i j

(11) Joint probability distribution for the data and all parameters are as follows

1

2exp1.7ln

1 1 1, , , , , , exp

221 exp1.7 1 exp1.7

x xij ijd jbi j tt ij j iij

f x t dij ij i j i j d d tj j ijb bi j i jt tij ij

1 1

N JxAxBxCxDxExF

i j

(12) Where is :

21 1

exp22

A f b b

(12.a)

21 1

exp22

B f

(12.b)

21 1

exp22

C f

(12.c)

21 1

exp22

D f

(12.d)

21 1

exp22

E f d d

(12.e)

1

F fg f

(12.f)

Full conditional probability b is proportional to the Joint probability distribution, such as equation 12, that assumes the other parameters are constant.

1

2exp1.7ln1 1 1

, , , , , , exp22

1 exp1.7 1 exp1.7

x xij ijj

i jij j iij

j j iji j i j

ij ij

db

ttf b x t dj ij ij i j i j d d t

b bt t

1 1

1 1 2exp22

N J

i j

b

(13)

Full conditional probability is proportional to the Joint probability distribution, such as equation 12, that assumes the other parameters are constant.



Hidayah et al. 241


Surabaya Indonesia

1

2exp1.7ln1 1 1

, , , , , , exp22

1 exp1.7 1 exp1.7

x xij ijj

i jij j iij

jj j ij

i j i jij ij

db

ttf x t b dj ij ij j i j d d t

b bt t

2

1 1

1 1exp

22

N J

i j

(14)

Full conditional probability is proportional to the Joint probability distribution, such as equation 12, that assumes the other parameters are constant.

1

2exp1.7ln1 1 1

, , , , , , exp22

1 exp1.7 1 exp1.7

x xij ijj

i jij j iij

jj j ij

i j i jij ij

db

ttf x t b dj ij ij i i j d d t

b bt t

2

1 1

1 1exp

2 2

N J

i j

(15) Full conditional probability is proportional to the Joint probability distribution, such as equation 12, that

assumes the other parameters are constant.

1

2exp1.7ln1 1 1

, , , , , , exp22

1 exp1.7 1 exp1.7

x xij ijj

i jij j iij

jj j ij

i j i jij ij

db

ttf x t b dj ij ij i i j d d t

b bt t

2

1 1

1 1exp

2 2

N J

i j

(16)

Full conditional probability d is proportional to the Joint probability distribution, such as equation 12, that assumes the other parameters are constant.

1

2exp1.7ln1 1 1

, , , , , , exp22

1 exp1.7 1 exp1.7

x xij ijj

i jij j iij

jj j ij

i j i jij ij

db

ttf d x t bj ij ij i i j d d t

b bt t

2

1 1

1 1exp

2 2

N J

i j

d

(17) Full conditional probability is proportional to the Joint probability distribution, such as equation 12, that

assumes the other parameters are constant.

1

2exp1.7ln1 1 1

, , , , , , exp22

1 exp1.7 1 exp1.7

x xij ijj

i jij j iij

j jj j ij

i j i jij ij

db

ttf x t b dij ij i j d d t

b bt t

1 1

N J

i j

(18) Full conditional distribution, in the equation 13-18, can not be expressed in the form of a well-known

distribution and the distribution is not simple. Therefore, parameter estimation is done by numerical iteration with the implementation of MCMC. The algorithm used is Gibbs Sampler. The algorithm generates the value of parameters based on the full conditional distribution of the above. Implementation of the algorithm is done with the helping WinBUGS 1.4, such as the following program.

Table 9 WinBUGS program 14 for estimating model parameters

model{ for (i in 1:N){ theta[i]~dnorm(0,1) tau[i]~dnorm(0,1) for (j in 1:J){

x[i,j] ~ dbern( pi[i,j] ) pi[i,j]<-exp(theta[i]- (d[j]/t[i,j])-b[j])/(1+exp(theta[i]-(d[j]/t[i,j])-b[j])) t[i,j]~dlnorm(mu[i,j],sigma) mu[i,j]<- beta[j]- tau[i] } # priors



Hidayah et al. 242


Surabaya Indonesia

for (j in 1:11){ b[j]~dnorm( 0,1) d[j]~dnorm(0,1)I(0,) beta[j]~dnorm(0,1) }

sigma~dunif(0,1) }

The reliability of the program as table 1 in estimating of the model parameters can be seen by comparing of

the generation parameters (parameters is specified to generate the data, or so-called true parameters ) with the parameter estimation. Comparison between the generation parameters and parameter estimation are as follows.

Table 10 The comparing of Generated Parameter and Estimated Parameter

Item difficult level (bj) The effort of examinee needed

to complete item tes (j )

Generated Parameter

Estimated Parameter

The distance of both

Generated Parameter

Estimated Parameter

The distance of both

-3 -3.061 -0.061 -0.63667339 -0.6193 0.017373 -2.4 -2.478 -0.078 1.178306319 1.099 -0.07931 -1.8 -1.808 -0.008 -1.730907016 -1.738 -0.00709 -1.2 -1.189 0.011 -0.078959223 -0.1355 -0.05654 -0.6 -0.3904 0.2096 0.670851036 0.6453 -0.02555

0 0.05265 0.05265 -0.263044163 -0.26 0.003044 0.6 0.7192 0.1192 1.537099167 1.514 -0.0231 1.2 1.101 -0.099 0.507695599 0.4503 -0.0574 1.8 1.577 -0.223 -0.941789438 -0.9106 0.031189 2.4 2.025 -0.375 -0.254579273 -0.391 -0.13642 3 2.756 -0.244 0.135498108 0.1899 0.054402

Table 2 , the third column shows the distance between generated parameter and estimated parameter is very small, where is close to zero. Estimated parameter of examinee‘s ability level and speed are not shown in this

article because there are very many, as many as 500 participants test. The average destance of the generated parameter and estimated parameter for the level of examinee‘s ability is 0.107. The average distance of the

generated parameters generation and estimated parameter for the examinee‘s speed is 0.000307. The average distance of the parameters generation and parameter estimation of the four parameters mentioned above shows a small value. This shows that the parameter estimation process runs properly and produce accurate parameter values.

CONCLUSIONS ML1P containing variable response time is a more realistic model than the conventional ML1P. The model is

able to capture reality, because in fact the test is always limited by time. Parameter estimation with Bayesian approach Markov Chain Monte Carlo showed accurate results. This is indicated by the average distance of the generated parameters and estimated parameters are small. But the conclusion of the goodness of models and parameter estimation accuracy is not enough by that way. There needs to be a further test by looking at the model fit the data.

The next study should look at the fitness of the model, for example by the Posterior Predictive Model Checking (PPMC), DIC, AIC, BIC etc. In addition the model needs to be tested for some situation of tests to see how they can produces the best estimated. The situation of test can be set based on the number of participants and the number of test test, because the test always involves both of the them.

REFERENCE 1. Breukelen, G.J.P.V, Psychometric Modeling of response speed and accuracy with mixed and conditional

regression, Psychometrika, 70(2), pp. 359-376, 2005.

http://link.springer.com/journal/11336



Hidayah et al. 243


Surabaya Indonesia

2. Hambleton, R.K., & Swamnathan, H, Item Response Theory.Boston, MA : Kluwer Nijhoff Publishin, 1985. 3. Van Der Linden, W. J. & Hambleton, R. K. (Eds). Handbook of modern item response theory, New York :

Springer-Verlag, 1997. 4. Pleskac TJ, Busemeyer JR. A dynamic, stochastic theory of confidence, choice, and response time. In:

McNamara DS, Trafton JG, editors. Proceedings of the 29th Annual Cognitive Science Society. Cognitive Science Society; Austin, TX: 2007. pp. 563–568, 2007.

5. Hopkins, 1998, 6. Oshima, T.C., The Effect of Speededness on Parameter Estimation IRT, Journal of Education

Measurement, 31,(3), pp. 200-219, 1994, 7. Abdelfattah, F. A., Response Latency Effects On Classical And Item Response Theory Parameters Using

Different Scoring Procedures, PhD Dissertation, Ohio University., Ohio, 2007. 8. Schnipke, D.L.,Pashley, P.J., Assessing Subgroup Differences in Item Response Times, the Annual Meeting

of the American Educational Research Association, 1997. 9. Schnipke, D. L., & Scrams, D. J. Exploring issues of examinee behavior: Insights gained from response-time

analyses. Law School Admission Council Computerized Testing Report. LSAC Research Report Series, 1999.

10. Hornke, L. F., Item Response Times in Computerized Adaptive Testing, Psicológica, 21(1), 175 – 189, 2000

11. Thissen. D., (1983), Latent Trait Scoring of Timed Ability Tests, University of Kansas, http://www.psych.umn.edu, (25 Nopember 2012)

12. Shih, B., et.al., Response Time Model for Bottom-Out Hints as Worked Examples, downloaded from http://www.cmu.edu, (5 Juli 2014)

13. Verbić, S., Establishing the rhythm of responding to test questions, http://arxiv.org, (20 Agustus 2015). 14. Halkitis, Jones, and Pradhan, Estimating testing time: The effects of item characteristics on response latency,

the annual meeting of the American Educational Research Association in New York, New York, 1996. 15. Abdelfattah, F. A., Response Latency Effects On Classical And Item Response Theory Parameters Using

Different Scoring Procedures, Dissertation, Ohio University, 2007. 16. Roskam, E.E., Models for Speed and Time Limit Test, W. J. Van der Linden & R.K. Hambleton (eds.),

Handbook of Modern Item Response Theory, pp. 169 – 185), New York : Springer, 1997. 17. Van Der Linden, W. J. & Hambleton, R. K. (Eds), Handbook of modern item response theory, New York :

Springer-Verlag, 1997. 18. Verhelst, N.D., et.al. A Logistic Model for Time-Limit Test, W. J. Van der Linden & R.K. Hambleton

(eds.), Handbook of Modern Item Response Theory, New York : Springer, pp. 169 – 185, 1997. 19. Wang, T & Hanson, B.A, Development & Calibration of an Item Response Model that Incorporates

Response Time, Applied Psychological Measurement, 29, pp. 323-339, 2005. 20. Ingrisone, S.J., An Extended Item Response Theory Model Incorporating Item Response Time, Ph.D.

Dissertation, Florida State University, 2008. 21. Ingrisone II, J.N., Modeling The joint Distribution of Response Accuracy and Reponse Time, PhD

Dissertation, Florida State University, 2008. 22. Van Der Linden, W., J., A lognormal model for Response Times on Test Items, Journal of Educational and

Behavioral Statistics, 31, 181 – 204, 2006. 23. Van Der Linden, W. J., A Hierarchical Framework for Modeling Speed and Accuracy on Test Items,

Psychometrika, 72(3), 287 – 308, 2007. 24. Klein Entink, R.H., et.al., A Multivariate Multilevel Approach To The Modeling Of Accuracy And Speed

Of Test Takers, Psychometrika, 74(1), 21 – 48, 2009. 25. Van der Linden, W. J., Conceptual Issues in Response Time Modeling, Journal of Educational

Measurement, 46 (3), 247-272, 2009. 26. Yamamoto, K., Estimating the Effects of Test Length and Test Time on Parameter Estimation Using the

HYBRID Model, ETS Tecnical Report TR-10, Princeton, NJ : Educational Testing Service, 1995. 27. Meyer, J. P., A mixture Rasch model with item response-time components. the annual meeting of the

National Council on Measurement in Education, New York, NY, 2008. 28. Wollack, J.A. & Woo, V., Using Response time to Improve Parameter estimation for Speeded Test Item -

University of Wisconsin-Madison, the annual meeting of the National Council on Measurement in Education, San Diego, CA, 2009.

29. Meyer, J.P. & Wise, S.L., Item Response Time And Distractor Analysis Including Item Response Time in Distractor Analysis via Multivariate Kernel Smooting, James Madison University, 2005.

http://www.psych.umn.edu/psylabs/catcentral/pdf%20files/cat79session6.pdf

http://www.cmu.edu/

http://arxiv.org/



Hidayah et al. 244


Surabaya Indonesia

30. Gvozdenko, E & Chambers, D, Beyond test accuracy : Benefit of Measuring response time in computerized testing, Australian Journal of Educational Technology, 23(4), 542-558, 2007.

31. Wise, S,L., & DeMars, C.E., An Application of Item Response Time : The Effort-Moderated IRT Model, Journal of Educational Measurement, 43, 19-38, 2006.

32. Van der Linder & Van Krimpen-Stoop, Using Response Time to detect Aberrant Responses In Computer Sdaptive Testing, Psychometrika, 68(2), 251-265, 2003.

33. Wang, T. A model for the joint distribution of item response and response time unsing one-parameter Weibull distribution (CASMA Research Report 20). Iowa City, IA : Center for Advance Studies in Measurement and Assessment, 2006.

34. Ranger, J., Modeling responses and response times in personality tests with rating scales, Psychological Test and Assessment Modeling, 55(4), 361 – 382.



Utomo et al.. 245


Surabaya Indonesia

Study of Variation on the Concentration of Digestion Method to Heavy

Metal Determination in River Sediment Samples

Yudhi Utomo*, Neena Zakia

1 Department of Chemistry, Faculty of Matematic and Natural Science, State University of Malang Jl. Semarang 5 Malang-Indonesia

*Corresponding Author. Email: [email protected]

Abstract. Research on the development of analysis methods of heavy metals in sediment samples by wet digestion process had been conducted. The aim of this study was to find an appropriate digestion method for heavy metal analysis in the solid dosage particularly the sediment of acid-based solvent based and heating by hot plate and measurement by AAS instruments. In the first stage, wet digestion process with two different HNO3 concentrations, 3 M and 6 M, were studied to the heavy metals of Cu, Cd and Cr in the sediment sample from Surabaya River. It shows that the higher of HNO3 concentration used, the higher of Cu, Cd and Cr concentrations obtained. The digestion process result with different solvents is given below. The highest of Cu concentration is obtained by using the solvent mixture of HCl 6M and HNO3 6 M (1:1). Meanwhile by using aquaregia and HCl 6M solvent, the concentration of Cu, Cd and Cr are equal. However, the highest concentration of Cr and Cd metals is obtained by using HCl 6 M.

INTRODUCTION Research analysis of heavy metals in the sample or solution is almost have no obstacles, because many

instruments which can be used directly without any pretreatment complicated. The analysis of solid samples in a pretreatment, such as sediment must be done to get a sample solution which is ready to be measured and the process is a lot of methods that can be done. The researchers handling variation is often get an obstacle in deciding a good method for the analysis of certain heavy metals. Digestion to sample sediments can use concentrated acid such as HNO3, HCl, HClO4, H2SO4 or HF. That concentrated acid can be used alone or in combination such as aqua regia is a mixture of HNO3-HCl ratio of 1: 3 is widely used in the study heavy metal analysis of samples of sediment and soil, but solvents that have capabilities that are not equal to the analysis of any heavy metals ( Boyle et al., 1998; Atgin et al., 2000 and Romaquare et al., 2008). Research involving heavy metal content in the sediments is mostly done by a total analysis regardless of sediment formation process, to do this many ways of destruction with different chemicals. Chen and Ma (2001) revealed in a study of heavy metals in soil samples by comparing three destruction techniques based aqua regia, the results of these studies illustrate three destruction techniques when applied to metal analyzes of sediment samples or soil gives different results, for Cr and Pb better destruction by Microwave aqua regia + HF, Cu and Fe good enough using Microwave aqua regia destruction, inside Zn is good with conventional techniques aqua regia hotplate. In general the destruction Microwave recovery techniques aquaregia + HF is better. Romaquare et al. (2008) revealed in a comparative study of three extraction procedure for the analysis of heavy metals are gradually developed by Tessier, McGrath- Cegarra and Gimeno-Garcia from soil samples. The other result that show a variety of preliminary process in the analysis of solid samples as performed by Atgin et al. (2000) which uses HNO3 and HF for digestion by heating in an oven at 150°C for 90 minutes, while Boyle et al. (1998) revealed in an analysis of Cu and Pb heavy metals in the sediment of a lake through an extraction process using 10% HCl for 5 hours at room temperature.

MATERIALS AND METHODS Sediment samples taken directly from Surabaya River in several locations as Lengkong Baru-Canggu to

Bridge along Sidoarjo, while the sediment samples were analyzed the levels of Cr and Cu component. Sample preparation: sediment samples first separated from coarse material such as plastic or leaf litter, then dried there is a temperature of 110oC for 4 hours and repeated until dry. Sample crushed and sieved to 50-100 mesh size, then sediment of sample dried till a constant in the 105oC oven. Destruction of sediment samples using wet digestion method, dry samples finely ground and sifted 50-100 mesh further heated for 1 hour at 80°C before digestion. 5 grams of dried sample was added 20 mL concentrated HNO3, left one night and then digested with hotplate at 120oC for 2 hours. The digestion result is cooled and filtered, the filtrate diluted with distilled water (aquadest) to 50 mL, then the concentration of Cr and Cu in the sample is measured by a Atomic Absorption Spectrophotometer (AAS) instrument as total substances




Jusoh et al. 246


Surabaya Indonesia

RESULTS AND DISCUSSION Wet digestion as a first step in the analysis to destruct bonds that occur in the sediments with metals, the

strength of the solvent in the process of digestion is expected to show the actual content of the metals contained in the sediments. The results of digestion process using nitric acid solution with different concentrations on same temperature and time of heating in the determination of Cu, Cr and Cd using AAS instrument obtained data as similarly in recapitulation attachment so that the results of a similar analysis using other solvents, that is HCl mixture HCl- HNO3 (1: 1) and aqua regia. The comparison results of the digestion process on HNO3 variation concentrate is that 3 M and 6M concentration is shown in Figure 1.

Figure 1. The content of heavy metals in sediment HNO3 digestion results with different concentrations

In general, the use of higher concentration solvents capable to release heavy metals from sediment matrix in greater numbers, while the difference of the content Cu, Cr and Cd heavy metals showed that Cu heavy metals much higher contained in the Brantas River and deposited in the sediment. The analysis results of the digestion process using different solvents obtained an average yield is shown in Figure 2.

Figure 2. Concentration of heavy metal digestion results with different solvent

Sediment digestion using different solvents obtained different results for each heavy metals. Cd and Cr metals can be taken from the sediment matrix with a higher content on digestion using 6M HCl followed by a mixture of HNO3: HCl (1: 1) than aqua regia, whereas Cu in sediments are very well digested with a mixture of HNO3: HCl (1: 1) although still under the use of HNO3. Chromium (Cr) is very good digestion using 6M HCl, but Cd good enough using concentrated HNO3 .

0.00

5.00

10.00

15.00

20.00

25.00

30.00

Cu Cr Cd

27.43

0.16 0.68

29.17

0.18 0.76 Con

cent

ratio

n m

g/kg

dry

Heavy metal substance

HNO3 6M

HNO3 3M

0.00

5.00

10.00

15.00

20.00

25.00

CdCr

Cu

0.49

8.07

19.68

0.26 2.39

20.25

0.31 2.76

21.54

Con

cent

ratio

n (m

g/kg

dry

)

Heavy metal substance

HCl 6 M

Air Raja

HNO3-HCl (1:1)



Jusoh et al. 247


Surabaya Indonesia

CONCLUSION The discussion results of data research can be concluded as the more concentrated HNO3 concentration

obtained Cu, Cr and Cd heavy metals in higher sediment samples. The digestion results with different solvent obtained difference, to high levels of Cu metal by a mixture solvent 6M HCl and 6M HNO3 (1: 1), while aqua regia 6 M HCl is almost same. The higher levels of Cr and Cd metals obtained using 6 M HCl solvents.

REFERENCES 1. Anazawa, K., Kaida, Y., Shinomura, Y. Tomiyasu,T., and Sakamoto, H. 2004. Heavy-Metal Distribution in

River Waters and Sediments Around a‖Firefly Village‖, Shihoku, Japan: Application of Multivariate

Analysis. Analytical Sciences, Januari Vol. 20: 79-84 2. Atgin,R. S., Omar El-Agha, Zararsız, A., Kocatas, A., Parlak, H., and Tuncel, G. 2000. Investigation of The

Sediment Pollution in Izmir Bay: Trace Elements. Spectrochimica Acta Part B 55 : 1151-1164 3. Bentivegna, C.S., Alfano, J.E., Bugel, S.M. and Czechowicz, K. 2004. Influence of Sediment Characteristics

on Heavy Metal Toxicity in Urban Marsh. Urban Habitats, Vol.2 No.1 4. Boyle, J.F., Mackay, A.W., Rose, N.L., Flower, R.J. and Appleby, P.G. 1998. Sediment Heavy Metal

Record in Lake Baikal: Natural and Anthropogenic Sources. Journal of Paleolimnology, 20: 135-150 5. Chen, M and Ma, L.Q. 2001. Comparison of Three Agua Regia Digestion Methods for Twenty Soils. Soil

Sci.Soc.Am.J. . 65: 491-499 6. Horsfall Jr, M.and Spiff, A.I. 2002. Distribution and Partitioning of Trace Metals in Sediment of The Lower

Reaches of The New Calabar River, Port Harcourt, Nigeria. J.Environmemtal Monitoring and Assessment 78: 309-326.

7. McAlister, J.J., Smith, B.J. Neto, J.B., and Simpson, J.K. 2005. Geochemical Distribution and Bioavailability of Heavy Metals and Oxalate in Street Sediments from Rio De Janeiro, Brazil: A Preliminary Investigation. Environmental Geochemistry and Health 27: 429-441

8. Romaguera, F., Boluda,R., Fornes, F dan Abad,M. 2008. Comparison of Three Sequential Extraction Procedures for Trace Element Partitioning in Three Contaminated Mediterranean Soils. Environ Geochem Health 30:171–175.



Jusoh et al. 248


Surabaya Indonesia

NEGATIVE EFFECT OF CTAB IN SYNTHESIS OF ZnO/MSN

PHOTOCATALYST FOR DECOLORIZATION OF METHYL

ORANGE

Jusoh, N.W.C. 1, Jalil, A.A. 1,2*, Triwahyono, S. 3

1Department of Chemical Engineering, Faculty of Chemical Engineering, 2Centre of Hydrogen Energy, Institute of Future Energy,

3Department of Chemistry, Faculty of Science,Universiti Teknologi Malaysia, 81300 Johor Bahru, Johor. *Corresponding author’s Email: [email protected]

Abstract. Zinc oxide supported on mesostructured silica nanoparticles (ZnO/MSN) catalyst was prepared via a simple electrochemical method in the presence of cetyltrimethylammonium bromide (CTAB) surfactant and its photoactivity was tested on a decolorization of methyl orange (MO) dye. Previous studied confirmed that a desilication occurred in the MSN framework accompanied by isomorphous substitution of Zn2+ to form active species Zn–O–Si bond. However, the interaction between Zn2+ with the silica support was found to be impeded when the CTAB was added to the electrolysis system, which led to less formation of the Si-O-Zn bonds. This result revealed that ZnO/MSNWOC, prepared in absence of CTAB displays a higher decolorization (4.47×10-3 mM h-1) than ZnO/MSNWC (3.06×10−2 mM h-1) at the similar reaction conditions.

INTRODUCTION An advanced oxidation process (AOPs) using heterogeneous semiconductor photocatalysts such as TiO2,

FeOOH, ZnO, CuO, and ZrO2 have become a popular method for the removal of toxic pollutants from wastewater [1,2]. Among them, ZnO has been widely researched in photocatalytic applications due to its comparable performance with the TiO2. However, the efficiency of ZnO was limited due to insufficient contact between pollutant and the catalyst surface. Incorporation of ZnO onto mesoporous material support is one of the method to overcome this limitation, as well as enhanced the photocatalytic activity [3]. Herein, we reported the preparation of ZnO loaded on mesostructured silica nanoparticles (ZnO/MSN), and studied the effect of cetyltrimethylammonium bromide (CTAB) in electrolyte on the formation and photocatalytic activity of the catalyst. In general, CTAB is a typical cationic surfactant, is often used in scientific researches to manipulate the nucleation and growth of particles [4]. However, we have found that the presence of CTAB decreased the efficiency of ZnO/MSNWC catalyst toward decolorization of methyl orange (MO) dye. Herein, we proposed a mechanism to explain these phenomena.

METHODOLOGY The ZnO/MSN catalyst was prepared according to previous report protocol [5-8]. An open system electrolysis

cell was fitted with a magnetic stirrer and a platinum plate cathode (2×2 cm2) facing a zinc plate anode (2×2 cm2). A 10 mL distilled water and 4 mL of ammonium hydroxide (NH4OH) were added into 1 g of MSN. Then, 0.1 M cetyltrimethylammonium bromide (CTAB) and 0.1 M tetraethylammonium perchlorate (TEAP) were added into the mixture. The electrolysis was conducted at a constant current of 120 mA cm-2 and 0 °C under air atmosphere for 5 min 26 sec to load 5 wt% of ZnO. The sample was impregnated and dried at 383 K for 12 h before being calcined at 823 K for 3 h to yield a white powder catalyst donated as ZnO/MSNWC. For comparison, ZnO/MSNWOC was prepared using the similar method in the absence of CTAB. Next, the photoactivities of the catalysts were evaluated for the decolorization of MO dye, performed in a batch reactor fixed with UV lamp (4 X 9 Watt; 365 nm emission) and a cooling system.




Jusoh et al. 249


Surabaya Indonesia


Photocatalytic Testing

The photocatalytic activity of ZnO/MSN catalysts has been tested for decolorization of 3.06×10-2 mM MO at pH 2 using 1g L-1 catalyst, and the results were illustrated in Figure 1. A control experiment conducted in the presence of bare MSN showed that no significant decolorization was achieved after 8 h. However, the addition of 5 wt% ZnO into MSN clearly increased the initial decolorization rate of both ZnO/MSN catalysts. ZnO/MSNWOC showed higher decolorization rate (4.47×10-3 mM h-1) compared to ZnO/MSNWC (3.73×10-3 mM h-1). The different photocatalytic efficiencies were achieved for each catalyst which most likely due to the influence of the structural arrangement resulted from the addition of CTAB in the electrolyte system.

Figure 1. Initial decolorization rate of the catalyst for decolorization of MO dye

Proposed Structure of the ZnO/MSN Catalyst

A probable mechanism for the formation of both catalysts was proposed as in Figure 2. Figure 2a showed the structure of desilicated MSN in alkaline electrolyte system. In the absence of CTAB, the produced Zn2+ cations during electrolysis were isomorphously substituted into MSN framework to form active species Si-O-Zn (Figure 2b). In fact, this active species play an important role in enhancing the decolorization of MO [5]. However, when the CTAB was added into the system, the CTA+ ions were also competitively interacted with the MSN surface (Figure 2c) to lessen the formation Si-O-Zn bonds which decreased the efficiency of the decolorization.

Figure 2. Proposed mechanism of synthesis of ZnO/MSN in the presence of CTAB

ACKNOWLEDGMENT The authors are grateful for the financial support by the Fundamental Research Grant Scheme (Grant No. 4F423)

and the awards of MyPhD Scholarship (Nurfatehah Wahyuny Che Jusoh) from the Ministry of Higher Education.

REFERENCES 1. A.A. Jalil, M.A.H. Satar, S. Triwahyono, H.D. Setiabudi, N.H.N. Kamarudin, N.F. Jaafar, N. Sapawe, R.

Ahamad. 2013. J. Electroanal. Chem. 701: 50-58. 2. R. Jusoh, A.A. Jalil, S. Triwahyono, A. Idris, S. Haron, N. Sapawe, N.F. Jaafar, N.W.C. Jusoh. 2014. Appl.

Catal. A: Gen. 469: 33-44. 3. C. Bouvy, F. Piret, W. Marine, B.L. Su. 2007. Chem. Phys. Lett. 433: 350-354.

0 1 2 3 4 5

Initial decolorization rate (x10-3 mM h-1)

MSN

ZnO/MSNWC

ZnO/MSNWOC

Si

O

Si

O O

Si

O O

Si

O

H H H H H HSi

O

Si

O O

Si

O O

Si

O

H Zn HZn

Si

O

Si

O O

Si

O O

Si

O

H Zn HN N

Pt c

atho

de

Zn2+

Zn anode

N+

With addition of CTAB

Electrolysis

Without CTAB

(a)

(b) (c)



Jusoh et al. 250


Surabaya Indonesia

4. A.H. Karim, A.A. Jalil, S. Triwahyono, S.M. Sidik, N.H.N. Kamarudin, R. Jusoh, N.W.C. Jusoh, B.H. Hameed. 2012. J. Colloid. Interf. Sci. 386: 307-314.

5. N.W.C. Jusoh, A.A. Jalil, S. Triwahyono, H.D. Setiabudi, N. Sapawe, M.A.H. Satar, A.H. Karim, N.H.N. Kamarudin, R. Jusoh, N.F. Jaafar, N. Salamun, J. Efendi. 2013. Appl. Catal. A: Gen. 468: 276-287.

6. A.A. Jalil, N. Fatimah, A. Panjang, S. Akhbar, M. Sundang, N. Tajuddin, S. Triwahyono. 2007. J. Hazard. Mater. 148: 1-5.

7. A.A. Jalil, S. Triwahyono, N.A.M. Razali, N.H.H. Hairom, A. Idris, M.N.M. Muhid, 2010. J. Hazard. Mater. 174: 581-585.

8. A.A. Jalil, S. Triwahyono, N.A.M. Razali, N.H.H. Hairom, A. Idris, M.N.M. Muhid, A. Ismail, N.A.M. Yahaya, N.A.L. Ahmad, H. Dzinun. 2010. J. Hazard. Mater. 174: 581-585.



Purnama et al. 251


Surabaya Indonesia

Inovation of Collagen Based Corneal Hydrogel with the Addition of

Glycopolymer as the Solution for Irreversible Blindness by Corneal

Ulcers

Disca Sandyakala Purnama*, 2)Hendita Nur Maulida, 3)Rara Setya Angtika, 4)Astriani Hendayanti,5)Anandhika Muhammad Satria

Biomedical Technology, Faculty of Science and Technology, Airlangga University, Surabaya *Corresponding Author’s Email :[email protected]

Abstract. Indonesian Health Ministry Data showed that 5.3% of 100,000 population in Indonesia suffer from corneal ulcers. Corneal ulcer is one of the five leading causes of largest blindness morbidity and loss of vision. Corneal transplant being the only medicine accepted widely despite there are many post-surgery complications. The study was conducted to review the initials make a synthetic cornea from natural ingredients collagen type I composite with a glikopolimer from poly (1,8-octanediol-co- citrate) (POC) and poly (vynyl-alcohol) (PVA) hydrogels form to review increase the strength of mechanical and biodegradibilty as candidate synthetic cornea. The result of Fourier Transform Infra Red (FTIR) tin that strong cross Occurs on Force C = O stretching which is a cluster formation of ester IN 1735

cm-1. The result of tensile test showed that the value of elasticity modulus tends to decrease with increasing concentration of collagen with modulus of elasticity in a row for a review variation collagen 1%, 3%, and 5% is 13.26 MPa, 13.21 MPa, and 11.5 MPa. Moisture balance test results Phosphate Buffer Saline (PBS) for 7 days for a review sample of collagen Variation 1%, 3%, and 5% is 89%, 91% and 92%. The synthetic cornea from collagen type 1 and glycopolymer from poly(1,8-octanediol-co-citrate) (POC) and poly(vynyl alcohol) (PVA) have a potential to be solution for irreversible blindness caused by corneal ulcers.

INTRODUCTION WHO data showed that in 2014 there were 45 million people with blindness in the world , where one-third of

people in Southeast Asia , including Indonesia . WHO data of 2004 showed that corneal ulcers ( sores on the cornea ) is a major public health problem because it can cause prolonged morbidity , loss of vision ( either one eye or two eyes ) . According to data from 1993 , indicated that the incidence of corneal ulcers occur 5.3 % per 100,000 population of Indonesia , the cause is trauma , contact lens wear , and sometimes unknown cause ( Fandri , 2013 ) . Corneal ulcer is also the number two cause of blindness in Indonesia ( Wijaya in Fandri , 2013 ) .

The formation of ulcers on the cornea are found by the collagenase formed by new epithelial cells and inflammatory cells . There are two forms of ulcers on the cornea that are central and marginal or peripheral . Causes of corneal ulcers are a bacterium , fungus , and herpes simplex . The bacteria that often lead to ulcers , among others, alpha hemolytic Streptococcus , beta hemolytic streptococci . Stafilokokkusaureus , Moraxella likuefasiens , Pseudomonas aeruginosa , Nocardiaasteroides , Alcaligenessp , and some other bacteria ( Ilyas , 2006) . Objective symptoms of ulcers in the form of ciliaryinjection , partial loss of corneal tissue and infiltrate . In the more severe cases occur iritis with hypopyon to permanent blindness ( Ilyas, et al , 2002) .

Until now , the transplant using donor tissue to be the only treatment that can be widely accepted for irreversiblecorneal blindness. However , the treatment is a transplant donor has many shortcomings in post-operative complications such as host response ( autoimmune ) , donor limitations , discrepancies and long recovery time ( C. Deng , et al , 2010) . Therefore , synthetic corneal donor tissue as a replacement for an alternative solution that is widely used . More recently , the results of a Phase 1 clinical trials in humans showed corneal substitute material results crosslinked recombinant human collagen has been successfully help regenerate tissues including corneal cells and nerves ( C. Deng , et al , 2010)

However, for clinical conditions where there is a failure or excess collagenase endothelium, the mechanical strength of the collagen needs to be improved by combining collagen with glikopolimer material in the form of a hydrogel. In this study, used glikopolimer of poly (vinyl alcohol) (PVA) as a hydrogel-forming agent and also has properties similar to tissues, have a good bikompabilitas properties and optical properties that can be set as desired (K. Liu, et al 2008). Then added poly (1,8-octanediol) (POC) that has mechanical properties, degradation and energy is also a good surface, where these qualities are very important in controlling the biological response to the material to be implanted (Richard, et al, 2009). So as to create a composite of collagen added to glikopolimer (PVA and POC) can make the cornea hydrogels have good mechanical properties and can also support the regeneration of the corneal tissue and nerves (C. Deng, et al, 2010).

Based on that background , it is proposed the manufacture of artificial cornea of the collagen hydrogel and glikopolimer are biodegradable and permanently able to restore vision . Cornea has a very high affinity to water . Corneal tissue is placed in water or physiological fluid causes the tissue to bulge or turgid , along with it , the less

mailto::[email protected]



Purnama et al. 252


Surabaya Indonesia

transparent stroma ( Hogan , et al , 1971; 58 ) . Hogan, et al , added , needed mechanical pressure of 60 mmHg to the corneal stroma were submerged saline to maintain hydration corneal transparency and ideal . Addition glikopolimer the anterior part is hydrophilic so it will keep the cornea moist with tear fluid . The cornea will be attached to the eye cells that will be supported by the stitching on the edges .


Material

The tools used are digital scales Mettler Toledo , magnetic stirrer Yellow MAG HS 7 , freezer, freeze dryer , Fourier Transform Infra Red ( FTIR ) Perkin Elmer , autograph IMADA , UV - Vis Spectrophotometer . While the materials used are collagen type I production of red snapper BATAN , poly ( vinyl alcohol ) from Sigma Aldrich , poly ( 1,8 - octadienol ) from Sigma Aldrich and citric acid and Phosphate Body Saline ( PBS ) .

Methods

Synthesis Glikopolimer Poly ( vinyl alcohol ) 40 % w / v dissolved in distilled water . Poly ( 1,8 - octadienol ) -CO - citric - acid made of poly ( 1,8 - octadienol ) and citric acid are stirred at 150⁰ C temperature for 20 minutes and then converted on 140⁰ C temperature for 1 hour . Then add a 98% weight toluene resulting from the previous solution , and then stir until the solution becomes homogeneous. Mixedwith a solution of poly ( vinyl alcohol ) and Poly ( 1,8 - octadienol ) -CO - citric - acid and stir until homogeneous.

Synthesis Hydrogel of Collagen – Glikopolimer

Collagen 2 grams dissolved in 2 % citric acid solution . Then some variation of collagen is made which are 1 % v / v collagen , 3 % v / v collagen , and 5 % v / v collagen , then mix each solution with a solution of collagen glikopolimer that have been made previously and mix until homogenous . Then print and do the drying using a freeze drier.

Analysis Fourier Transform Infra Red( FTIR )

FTIR is used to determine the functional groups of the material used and new functional group obtained from the synthesis is done on the sample . Samples to taste then added KBr powder which is then compressed with a hydraulic clamp and placed in a specimen and irradiated with infrared wave number 4000-450 cm - 1 ( Smith , 2011) .

Water Content Test Using (Phosphate Body Saline) PBS

Test the water content is using to determine how much moisture that can be absorbed by the hydrogel is determined by calculating the weight of the hydrogel after inserted into PBS divided by the weight of the hydrogel before being put into PBS and expressed in%.

Tensile Test Using Autograph

Tensile test is used to determine the elastic properties of the material when implanted and to determine the mechanical rerspon when the material interacts with the body's tissues, the sample was formed into a dog bone shape with a length of 63.5 mm and a width of 10 mm and 5 mm on each sample variation in accordance with the American Society for Testing Materials (ASTM D 1822 L). Samples were loading 50N with a speed of 10mm / min. The results obtained are the maximum stress values (MPa) and the amount of strain and elongation (Purwanti, 2010). The formula used to describe

the relationship of stress and strain are: E = ζ / ε (1)

with Voltage: ζ = F / A, F = force, A = cross-sectional area Strain: ε = L / L, L; changes in the length, L: original length

RESULTS AND DISCUSSION Results from this study is the corneal hydrogel which hydrogel is formed through a freeze dry method in which

the sample on any variation made with a thickness of 0.5 mm with a diameter of 11- 12 mm in accordance with the original cornea ( Riordan - Eva , 2010) . The resulting hydrogel cornea then tested to conform to the parameters required to be applied as the cornea through FTIR testing , tensile

testing , moisture content and spectroscopy.



Purnama et al. 253


Surabaya Indonesia

Mod

ulus

E

last

isit

as

(MP

a)

Poly(1,8-octanediol-co-citrate)(POC) has been synthesized during this research by heating citric acid and 1,8 octanediol at 140-160°C for one hour. This treatment was purposed for creating cross-link between the two main ingredients, citric acid and 1,8 octanediol, which have boiling point at 153°C and 57-61°C respectively according to the product data released by manufacturer. In this synthesize, citric acid acted as the binder to the other chemical structure. Data of Fourier Transform Infra Red (FTIR) spectroscopy proofed that cross-linking of ester group (C=O) was formed in the synthesize process, which can be marked as C=O stretch at 1731 cm-1. According to Yang et al, 2004, the synthesize process of POC will be considered as success if the presence of ester group was identified on the FTIR spectroscopy.

Figure 1. FTIR spectrum of Poly(1,8-octanediol-co-citrate)( POC)

Tensile strength test is then performed to determine the elasticity properties of the material. The result demonstrated that the elasticity modulus of the material tends to declined as the concentration of collagen increased.

13.5

13

12.5

12

11.5

11

10.5 Kolagen 1% Kolagen 3%

Kolagen 5%

Variasi Sampel Graph 1. Result of Tensile Test

Addition of 1% collagen to the composite material has resulted to the highest increase of elasticity modulus recorded in Graph 1, where the value of modulus is 13,26 MPa. Another variations of collagen addition were 3% and 5%, which resulted in value of modulus as 13,21 MPa and 11,5 respectively. Generally, elasticity modulus declined as the concentration of collagen added increased due collagen gave the composite brittle properties.

According to Crabb et al, 2006, elasticity modulus of human cornea is 3-13 MPa, it indicates that all variations of the samples are in the value range of the functional standards of human cornea.

Three samples were soaked in Phosphate Based Saline (PBS) for 7 days to determine moisture content of the samples. Those three samples were weighted before and after soaking, resulting data that showed moisture content of each samples. The composite with addition of 1% collagen contained water as much as 89%, while two other samples, 3% and 5%, contained moisture as much as 91% and 92% respectively.



Purnama et al. 254


Surabaya Indonesia

%W

ater

C

onte

nt

93

92

91

90

89

88

87 Collagen 1% Collagen 3% Collagen

5%

Sample Variations Graph 2. Result of Water Content Test

The data showed that increasing in collagen addition to the composite affected the hydrogel‘s ability to held

water, water content increased as amount of collagen added increased. As the amount of collagen increased, the cross-linking formed between molecules were also increased. According to Righe, 1992, human cornea has water content as much as 81%, while sample with 1% collagen added contained 89% water. So the sample with addition of 1% collagen was close to the standard value of water content of human cornea.

CONCLUSION Synthesize of collagen-glycopolymer was obtained by making glycopolymere first by mixing PVA 40%

with 1,46 ml POC then added to 1% collagen. After the mixture are completely mixed, it is ready to be casted and freeze dried.FTIR spectroscopy result showed that ester group has formed and marked at wave number 1753 cm-1. It means the synthezise was considered as success. Tensile strength testing values were ranged of 11.5 to 13.26 MPa. Three variations of samples were contained water in range of 89% to 92%. Based on the result of all the test this cornea hydrogel can be a candidate of synthetic cornea.

REFERENCES 1. Arrohmah. 2007. Studi Karakteristik Klorofil Pada Daun Sebagia Material Photodetector Organic.

Surakarta : Fakultas Matematika dan Ilmu Pengetahuan Alam Universitas Sebelas Maret. 2. C. Deng, dkk. 2010. Collagen and Glycopolymer Based Hydrogel For Corneal Application. Elsevier :

Acta Biomaterialia. 3. Crabb, dkk. 2006. Biomechanical and Microstructural Characteristics of A Collagen Film-Based

Corneal Stroma Equivalent. Tissue Engineering;12(6);1565-75. 4. Fandri, MY. 2013. Penatalaksanaan Pada Pasien Ulkus Kornea Dengan Prolaps Iris Oculi Sinistra.

Medula, Volume 1, Nomor 1. 5. IlyasS., Maylangkay B. H. H., UlkusKornea. 2002. Ilmu Penyakit Mata Untuk Dokter Umum dan

Mahasiswa Kedokteran. Edisi Ke-2. Jakarta : Sagung Seto. Hal : 131-134. 6. Ilyas S., UlkusKornea. 2006.IlmuPenyakit Mata. Edisi ke-3. Jakarta: Fakultas Kedokteran

Universitas Indonesia. Hal : 159- 160. 7. Liu K, dkk. 2008. Graphite/Poly(Vinyl Alcohol) Hydrogel Composite As Porous Ringy Skirt For

Artificial Cornea. Elsevier : Materials Science and Engineering C. 8. Purwanti, Ani. 2010. Analisis Kuat Tarik dan Elongasi Plastik Kitosan Terplastisasi Sorbitol.Teknik

Kimia, Institut Sains dan Teknologi AKPRIND Yogyakarta. Jurnal Teknologi, Volume 3 Nomor 2, Desember 2010, 99-106.

9. Richard T.T, dkk. 2009. Recent Development on Citric Acid Derived Biodegradable Elastomers. Bentham Science Publicers Ltd.

10. Righe, B. 1992. Eye Contact. Chem Br ; 28:241-4. 11. Riordan, Eva P . 2010. Anatomi & Embriologi Mata In:Vaughan, Asbury.Oftalmologi Umum Edisi17.

Jakarta: EGC 12. Smith, Brian C. 2011. Fundamentals of Fourier Transform Infrared Spectroscopy, Washington: CRC

Press. 13. Yang, J, dkk. 2006. Synthesis and Evaluation of Poly(diol citrate) Biodegradable Elastomers.

Biomedical Engineering Department, Northwestern University, Evanston, IL 60208, USA.



Qulub et al. 255


Surabaya Indonesia

Properties of Nano Hydroxyapatite and Poly (1,8 Octanediol-Co- Citrate)

(POC) for Biodegradable Bone Screw

Fitriyatul Qulub*, Hendita Nur Maulida, Ewing Dian Setyadi, Systi Adi Rachmawati, Imroatus Sholikhah, Prihartini Widiyanti

1 Teknobiomedik, Fakultas Sains Dan Teknologi, Universitas Airlangga *Coresponding : [email protected]

Abstract. Innovations of biomaterial that have the main characteristics biodegradable is required to avoid reoperation and biocompatible. It has been synthesized and characterized of variation composition of nano hydroxyapatite (HA) on 1,8-octanediol Poly-Co-Citrate (POC) as Biodegradable Bone Screw. This research aims to synthesize and characterize POC-HA composition influence on mechanical properties and compatibility POC-HA composite. POC synthesis performed by condensation polymerization method, an ester bond formed group C=O stretch at 1731 cm-1 through a test of functional groups POC pre-polymer. POC pre-polymer composite with nano-HA on the variation of composition 62%,65%, 68% and 71%, followed by post-polymerization treatment. Results showed the composition of HA nanoparticles influence mechanical properties and biocompatibility. The best results on the composition of 62% HA, reviewed in mechanical hardness value of 885.57 MPa approached bone hardness (150-664 MPa). The degradation rate of 3.42% (4 weeks) corresponding period of grafting bone fracture 21 months, a pore size range of 600nm - 2193nm. Based on the results of this study characteristics, POC-HA composites as a potential candidate biodegradable bone screw.

INTRODUCTION Injuries have become a major public health problem, more than two thirds of injuries experienced by emerging

countries such as Indonesia (Riyadina, 2009). One of the medical treatment of bone fractures could be solved by orthopedic surgery. The data from a medical practitioner at Medical Faculty of Airlangga University / Soetomo Hospital in Surabaya show that the cases of orthopedic surgery is 300-400 cases per month (Sopyan, 2007). 80% of patients in the productive age and most (63%) were treated operatively using either internal fixation with a bone screw so the patient can be mobilized immediately and worked (Saleh, 1998).

Bone screws are special screws for handling bone fracture (Riyadina, 2009). The installation techniques with bone fracture fixation screw more easy and efficient. In Sri Rahayu‘s research (2012) gives innovation fixation design of

bone fractures were more simple, using the cortical bone screw that mixture from titanium and stainless steel that can be accepted by the body. Fixation technique using titanium materials provide a limited indication because the cost is very expensive. The mechanical properties of stainless steel metal seven times more greater than the bone mechanical so the bone can be relied as cantilever load at bone but its toxic. The research from Respati (2010) concluded that metal stainless steel is a material that‘s good for the bone screw because of its mechanical properties but it‘s not suitable for tissue. Stainless steel can not be used for a long period due to the body fluid occur localized corrosion. So, we need materials that are not only good for the mechanical properties, but also good for the tissue, such as the biocompatible polymer alloy or composite.

The development research of polymer based on composites as biomaterials for prosthetic so rapidly along with the advance of biomaterials technology. The main properties of biomaterials are biocompatible, biofunctional, biodegradable and resistant to corrosion (Purnamasari, 2013). One of the biopolymers that are useful as biodegradable bone screw is poly (1,8-octanediol-co-citrate) (POC). Based on research by Yang, et al., (2006) poly (1,8-octanediol-co-citrate) (POC) are biocompatible, non-toxic, readily available and economic.

Biopolymer POC have some potency as composite material to improve the mechanical properties of materials that compatible with the application of biodegradable bone screw. Efforts made by Qiu H, et al., (2006) is have been composite POC with hydroxyapatite.

Hydoxyapatite (HA) is a member of the group of ceramic material mineral apatite with a chemical formula Ca10(PO4)6(OH)2 which is biocompatible and bioactive because the mineral‘s content as good as chemical and physical bone‘s structure (Respati, 2010). These advantage makes HA material more prospective as a base material of

a prosthesis. Nano-sized hydroxyapatite is very beneficial, especially if they can function as a filler in the polymer as it can stimulate the growth and bone formation and increase strength.




Qulub et al. 256


Surabaya Indonesia

On this background, the research on alloys poly (1,8-octanediol-co-citrate) (POC) with variations in the composition of nano-hydroxyapatite as a candidate biodegradable bone screw needs to be synthesized to get the material that is biocompatible, strong and capable to activate back the growth of osteoblasts. POC have flexible degradation period. Osteoblasts can grow faster so the material that have been chosen may not be permanent and should be degradable. If the material is not biodegradable, it is needed surgery to take the polymer after the osteoblast regeneration process on bone was completed. Hydroxyapatite used as a filler to of improve the mechanical properties and good compatibility to the cell in the bone screw application and can stimulate the growth of osteoblasts.

The goal of these research was to prove the synthesis of composite biodegradable bone screw with nano-hydroxyapatite composition variation on poly (1,8-octanediol-co- citrate) (POC) and to find out the results of the characterization of biodegradable bone screw through the test of functional groups, hardness, compressive strength test, morphological test, and biodegradation test. Biodegradable Bone Screw POC-HA composite results expected to act as a bone screw for bone fracture that have corresponding period of the biodegradation process of bone grafting and biocompatible.

MATERIAL AND METHOD

Materials

The equipment used in this study include the Balance analytical (Mettler Toledo), measuring cups, glass beaker, spatula, Fourier Transform Infrared Spectrometry (FT-IR) Tensor 27, Bruker, Shimadzu Micro hardness tester of type M, Shimadzu Corporation Kyto-Japan, Scanning Electron Microscopy (SEM) (Inspect S50), EDAX analysis Ametek material division, Sputter Coater "Quorum" SX7620, Autografh, thermometer, pH-meter, Teflon dishes, magnetic stirrer, hot plate, and a Gallenkamp Vacuum Oven. The material used is 1,8-octanediol (98%) from Sigma-Aldrich (St. Louis, MO, USA), citric acid (99.5%) [Mw: 210.14] from SAP hydroxyapatite (HA) [Mw: 502.32, assay> 90%; particle size <100nm of PAIR-BATAN, Ethanol for analysis (99.99%), Simulated Body Fluid (SBF), pH 7.4, in 37 ° C, Buffer solution, pH 7, Center for Chemistry, Indonesian Institute of Sciences.

The research variables

The control variable are the pre-polymerization treatment and post-polymerization, the independent variable is the nanomaterial composition of hydroxyapatite, while the dependent variable is the result of characterization which includes the characteristics of functional groups, hardness, compressive strength, biodegradation, morphology and changes in element content.

RESULTS AND DISCUSSION Polymer synthesis on Poly (1,8-octanediol-co-citrate) is reacted from citric acid and 1.8 octanediol at the same

mole. The method used to measure the two materials by condensation method with temperatures up to 160ºC then lowered reaches 140ºC for 1 hour using a set of reflux tools with a constant flow of nitrogen. The temperature treatment intended to occur between the cross link between citric acid and 1,8 octanediol, because it is based on data product that citric acid boiling point of 153ºC and the boiling point of 1,8 octanediol 57-61ºC. The characteristic of citric acid to binds other chemical chain. Pre polymer composite was measured with nano hydroxyapatite by post- polymerization process at the temperature of 80ºC for 3 days followed by a temperature of 120°C for 1 day under vacuum conditions. Temperature and time of post- polymerization process affects the mechanical properties of these composites. The longer of time process and the optimum temperature will improve the mechanical properties of materials (Yang, et al., 2006).

Figure 1. the IR spectrum composite POC-HA



Qulub et al. 257


Surabaya Indonesia

1042,48

930,03

941,8 885,57 H

ardn

ess

Val

ue

(MP

a)

Com

pres

sive

Str

engt

h va

lue

(MP

a)

The absorption of spectrum was appeared the typical functional groups of hydroxyapatite and POC polymer. At wave number 1545 cm-1-1346 cm-1 indicates a functional group CO3 of hydroxyapatite and at wave number 1030 cm-1-563 cm-1 shows the functional groups PO43-, functional group that appear in this analysis according to research by Moradi (2013) which states that the functional group at wave number 592 cm-

1, 601 cm-1, 962 cm-1, 1030 cm-1 and 1089 cm-1 are a phosphate group of hydroxyapatite. While the peaks show the typical spectrum of the polymer poly (1,8 Octanediol co Citrate) (POC) also appear ester group at wave number 1741 cm-1 (Yang et al., 2004).

Based on the analysis of functional groups in the IR spectrum POC-HA composites showed that there are not newly formed group developed. This is because the polymer composite formed between the POC as the matrix and hydroxyapatite as a filler chemical bonding does not occur, but only bound physically (Moradi, et al., 2013).

Hardness testing using Vickers microhardness carried out to determine the resistance of material to the press deformation (indentation). This testing is done especially on the material to intersect between two components and moving each other. The testing process is done by pressing a sample of 0.5 kgf (4.903325 N) by a pyramid-shaped indenter.

Graphic of Hardness Test

POC-62 HA POC-65 HA POC-68 HA POC-71 HA Variation Composition of Nano HA

Figure 2. influence graphics composition of the composite POC-HA The data of hardness test results appears that the highest hardness values contained in the composite sample POC-

71HA with hardness value of 1042.28 MPa. It‘s different from POC-HA composite sample by varying the composition of hydroxyapatite 62% wt, 65% wt and 68% wt gain that increasing hardness value on 885.57 MPa, 930.03 MPa and 941.8 MPa. Based on the hardness value of each sample showed that increasing of hardness was followed by increasing composition of HA nanoparticles in composites. It is because POC-HA biocomposite have higher percentage composition of hydroxyapatite than POC polymer. The value of hardness on hydroxyapatite is 4.86 GPa (Mucalo, 2015).

Data from literature that known Gibson. et al., (2010) showed that the hardness of human bone (cortical and cancellous) ranging between 150-664 MPa. In the POC-62HA sample, the value is greater hardness but closer than the hardness of bone. This is supported by increasing hardness interface and internal fixation of bone as the skeletal system (Larsson, 2002).

Graphic of Compressive Strength 20

10

9.078 8.78 9.56

0 64

POC-62 HA POC-65 HA POC-68 HA POC-71 HA Variation Composition of Nano HA (%wt)

Figure 3. effect of composition graph hydroxyapatite against violence composites

Compressive strength test results showed the highest value of compressive strength reached 9.56 MPa by POC-68HA sample but the lowest value in the sample POC-71HA is 3.64 MPa. The graphic value above was fluctuative, according to the research of Qiu, H. (2006) that the addition of hydroxyapatite composition will improve the mechanical properties of compressive strength of the material. Based on the analysis of the test results of the compressive strength, compressive strength value of all the samples included in the range of compressive strength in cancellous bone based on the literature, which is 2-12 MPa (Ficai et al., 2011). However, the interaction between matrix and filler is best viewed in the SEM sample results POC-62HA (Figure 6) with a value of 9.078 MPa in compressive strength.

The comparison effect of hydroxyapatite composition of multiple samples to degradation rate showed by Figure 5 below.



Qulub et al. 258


Surabaya Indonesia

Element Wt% At%

CK 16.46 27.21

OK 37.40 46.42

NaK 02.59 02.23

PK 17.58 11.27

CaK 25.97 12.87

Matrix Correction ZAF

Figure 4. the weight lost during in vitro test biodegradability

The degradation process carried out by incubated the samples into Simulated Body Fluid at pH 7.4 and a static temperature conditions for 4 weeks. This value refers to the study of H. Qiu et al, 2006 that in the period of 0-5 weeks, the percentage of weight lost8%. The degradation rate was stable in POC-62HA sample and it‘s supported by the results of the calculation of total degradation where POC-62HA sample has a total degradation for 27 months. The accumulation of degradation period on these samples appropriate with the maximum period of bone grafting process according to Appley (Warwick DJ, et al., 2001) which needed up to 21 months. Biodegradation characterization results are supported by morphology test (Figure 6) and the results Dispresive Energy X-Ray Spectroscopy (EDX) (Table 1).

Figure 5. morphology samples POC-62HA after biodegradability process for 2 weeks

a.Before Process Biodegradation b. After Process Biodegradation

Element Wt% At%

CK 22.41 36.13

OK 32.17 38.93

NaK 01.55 01.31

PK 17.20 10.75

CaK 26.66 12.88

Matrix Correction ZAF

Analysis of pore size by morphological observation of samples in the variation of62% of hydroxyapatite which

has undergone a process of degradation for 2 weeks had a range of 600 nm - 2,193 µ m.



Qulub et al. 259


Surabaya Indonesia

EDX results showed that weight percentage (Wt%) decreasing on C element, where it‘s the one of the element that contained on POC polymer that appears on biodegradation process (Tran, RT, et al., 2009). Decreasing of 5.95% caused by the dissolution of the CX bond that react with SBF. However, the elements O and P are increasing, that proved the bioactivity of material, because the element O is possible to react in the formation of apatite in materials that soaked in SBF (Bohner and Lemaitre, 2009).

CONCLUSIONS Synthesis POC carried out by condensation polymerization method, obtained POC pre-polymer. POC pre-polymer

composite with nano-HA in the composition of 62%, 65%, 68% and 71% into a clay-like. POC pre-polymer composite with nano- HA printed on Teflon dishes for printing that appropriate on characterization and treatment of post-polymerized at 80ºC for 3 days followed 120° on vacuum condition for 24 hours. The results of the POC polymer formed by ester bonding group C = O stretch at 1735 cm-1 through a test of functional groups (FTIR). Mechanical tests show that the hardness value 885.57 MPa - 1042.48 MPa and the compressive strength characterization 3.64 MPa - 9.56 MPa. The rate of degradation of the bone healing period for 27 months.

ACKNOWLEDGEMENTS Thank you to the Directorate General of Higher Education that has funded this research through the Student

Creativity Program (PKM) 2015, Mr. Prof. Dr. Moh. Nasih, SE., M.T., Ak. As Airlangga University rectors and to everyone who helped the implementation of this research.

REFERENCES 1. Bohner, M. and Lemaitre, J. 2009. Can Bioactivity be Tested In Vitro with SBF Solution?, Biomaterials., 30

(12), 2175-2179. 2. Ficai, A., Andronescu, E., Voicu, G., Ficai, D., 2011, Advances in Collagen/Hydroxyapatite

Composite Materials. Politehnica University of Bucharest, Faculty of Applied Chemistry and Materials Science, Romania.

3. Gibson, L.J., et al. 2010. Cellular Materials in Nature and Medicine.P.131: Cambridge University Press

4. Moradi, A., et al. 2013. Fabrication and Characterization of Elastomeric Scaffolds Comprised of a Citric-Based Polyester/Hydroxyapatite Microcomposite. Material and Design 50 (2013) 446.450

5. Mucalo, Michael. 2015. Hydroxyapatite (Hap) for Biomedical Applications. Elsevier. 6. Qiu H, Yang J, Kodali P, Koh J, Ameer GA.2006.A Citric Acid-Based Hydroxyapatite Composite for

Orthopedic Implants.Biomaterials, 27(34): 5845-5854 7. Rahayu, Sri. 2012. Titanium Bone-Screw : Alternatif Fiksasi Intermaksilar pada Fraktur Mandibula

Sederhana . Majalah Kedokteran FK UKI 2012 Vol XXVIII No.2 April – Juni 8. Respati, S.M.B. 2010. Bahan Biomaterial Stainless Steel dan Keramik. Momentum, Vol. 6, No. 1, April

2010 : 5 – 8 9. Riyadina, Woro, dkk. 2009. Pola Determinan Sosiodemografi Cedera Akibat Kecelakaan Lalu Lintas

di Indonesia. Pusat Penelitian Pengenmbangan Biomedis dan Farmasi.Vol.59 No.10 10. Saleh M, Irawan E. 1998. Penanganan Penderita Fraktur di Bangsal Bedah RS Dr. Kariadi. 11. Sopyan Iis. 2007. Coral dan Gamping, Alternatif Murah Pengobatan Kanker Tulang. Makalah Jakarta:

Pusat Data dan Informasi Perhimpunan Rumah Sakit Seluruh Indonesia 12. Spielmann Horst, Andrea Seiler, Susanne Bremer, Lars Hareng, Thomas Hartung, Hans Ahr, Elaine

Faustman, Ulla Haas, Graeme J. Moffat, Heinz Nau, Philippe Vanparys, Aldert Piersma, Juan Riego Sintes and Jane Stuart 2006. The Practical Application Of Three Validated In Vitro Embryotoxicity Tests. ATLA 34, 527-538.

13. Tran,R.T, et al., 2009. Recent Developments on Citric Acid Derived Biodegradable Elastomers. Vol.2, No.3.

14. Trihapsari, Enita. 2009. Faktor-Faktor yang Berhubungan dengan Densitas Mineral Tulang Wanita ≥ 45 Tahun. Jakarta Pusat: Departemen Pendidikan Nasional

15. Warwick, D.J, et al. 2001. Apley's System of Orthopaedics and Fractures 8Ed. Francis: Taylor 16. Yang, J., et al. 2006. Synthesis and evaluation of poly(diol citrate) biodegradable elastomers.

Biomedical Engineering Department, Northwestern University, Evanston, IL 60208, USA.

Date post:	09-Feb-2022
Category:	Documents
Upload:	others
View:	1 times
Download:	0 times

ICOWOBAS 2015 - UNUD

Documents