Post on 11-Mar-2018
transcript
Extraction of key components from cellular material :aspects of product and process designZderic, A.
Published: 01/01/2015
Document VersionPublisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:
• A submitted manuscript is the author's version of the article upon submission and before peer-review. There can be important differencesbetween the submitted version and the official published version of record. People interested in the research are advised to contact theauthor for the final version of the publication, or visit the DOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
Citation for published version (APA):Zderic, A. (2015). Extraction of key components from cellular material : aspects of product and process designEindhoven: Technische Universiteit Eindhoven
General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ?
Take down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.
Download date: 14. May. 2018
Extraction of key components from cellular material:
aspects of product and process design
PROEFSCHRIFT
ter verkrijging van de graad van doctor aan de Technische Universiteit Eindhoven,
op gezag van de rector magnificus prof.dr.ir. F.P.T. Baaijens,
voor een commissie aangewezen door het College voor Promoties, in het openbaar
te verdedigen op dinsdag 26 mei 2015 om 16:00 uur
door
Aleksandra Žderić
geboren te Šabac, Servië
Dit proefschrift is goedgekeurd door de promotoren en de samenstelling van de
promotiecommissie is als volgt:
voorzitter: prof.dr.ir. J.C. Schouten
1e promotor: prof.dr. J. Meuldijk
copromotor(en): dr.ir. E. Zondervan
leden: prof.dr.ir. M.C. Kroon
Prof.Dr.-Ing. G. Schembecker (TU Dortmund)
prof.dr.ing. M.H.M. Eppink (Wageningen UR)
dr. O. Trifunovic (Unilever)
adviseur(s): ir. G.D. Mooiweer (DE Master Blenders)
Extraction of key components from cellular material: aspects of product and
process design
A.Žderić
Eindhoven University of Technology
The research described in this thesis was sponsored by the Institute for Sustainable
Process Technology (ISPT), project number: FO-10-06
A catalogue record is available from the Eindhoven University of Technology
Library ISBN: 978-90-386-3854-6
To my family
Table of Contents
Summary ........................................................................................................ i
1 Introduction ................................................................................................ 1
1.1. Fresh tea leaves and polyphenols ......................................................................... 2
1.2. Soybeans and oil bodies ....................................................................................... 3
1.2.1. Seed structure and chemical composition .................................................... 4
1.2.2. Soybean oil bodies ....................................................................................... 4
1.3. Mild separation pre-treatment techniques in food industry .................................. 6
1.4. Product Driven Process Synthesis (PDPS) methodology ..................................... 8
1.5. Objectives ............................................................................................................. 9
1.6. Outline ................................................................................................................ 10
1.7. References .......................................................................................................... 11
2 A study of mechanism involved during polyphenol extraction from fresh
tea leaves by pulsed electric field ................................................................ 15
2.1. Introduction ........................................................................................................ 16
2.2. Experimental procedure ..................................................................................... 17
2.2.1. Composition of tea leaf ............................................................................. 17
2.2.2. Moisture content analysis .......................................................................... 17
2.2.5. Specific energy input ................................................................................. 19
2.2.6. Determination of extraction yield - measurement of total polyphenols
content 20
2.2.7. Statistical analysis ..................................................................................... 20
2.3. Results and discussion ........................................................................................ 21
2.3.1. Characterization of pulsed electric field process and electric measurements
21
2.3.2. Effect of electric field strength on extraction yield of polyphenols ........... 23
2.3.3. Effect of the total treatment time on the extraction yield of polyphenols .. 25
2.4. Conclusion ......................................................................................................... 27
2.5. References .......................................................................................................... 27
3 Product-driven process synthesis for the extraction of polyphenols from
fresh tea leaves ............................................................................................ 31
3.1. Introduction ........................................................................................................ 32
3.2. Framing level and product ideas ........................................................................ 33
3.3. Input/output level ............................................................................................... 34
3.4. Task network ...................................................................................................... 37
3.5. Mechanism and operating window ..................................................................... 43
3.5.1. Design of experiments ............................................................................... 44
3.5.2. Variables (factors) ..................................................................................... 45
3.5.3. Statistical analysis ..................................................................................... 48
3.5.4. Response contour plots .............................................................................. 52
3.6. Conclusions ........................................................................................................ 53
3.7. References .......................................................................................................... 54
4 Isolation of oil bodies from soybeans in a mild way: definition of
operating window for process design ......................................................... 59
4.1. Introduction ........................................................................................................ 60
4.2. Materials and methods ....................................................................................... 61
4.2.1. Preparation of soybean flour ...................................................................... 61
4.2.2. Aqueous extraction of soybean oil bodies ................................................. 62
4.2.3. Enzyme/ultrasound-assisted aqueous extraction ....................................... 63
4.2.4. Recoveries ................................................................................................. 63
4.3. Results and discussion ........................................................................................ 64
4.3.1. Effect of particle size on the aqueous extraction of soybean oil bodies .... 64
4.3.2. Enzymatic hydrolysis and ultrasonication of the coarse soy flour on the
aqueous extraction of OBs .......................................................................................... 69
4.4. Conclusion ......................................................................................................... 71
4.5. References .......................................................................................................... 72
5 Product-driven process synthesis for the extraction of oil bodies from
soybeans ....................................................................................................... 75
5.1. Introduction ........................................................................................................ 76
5.2. Framing level ..................................................................................................... 77
5.3. Consumer wants and product ideas .................................................................... 78
5.4. Input/output level ............................................................................................... 81
5.5. Task network ...................................................................................................... 82
5.6. Mechanism and operating window ..................................................................... 86
5.7. Equipment integration ........................................................................................ 88
5.8. Conclusions ........................................................................................................ 92
5.9. References .......................................................................................................... 92
6 Conclusions and Outlook ......................................................................... 97
6.1. Conclusions ........................................................................................................ 97
6.1.1. Polyphenols from tea leaves ...................................................................... 98
6.1.2. Oil bodies from soybeans .......................................................................... 99
6.2. Outlook ............................................................................................................. 100
6.2.1. Wax removal from the surface of tea leaves ............................................ 100
6.2.2. Pulsed electric field method .................................................................... 100
6.2.3. Product and equipment integration .......................................................... 101
Acknowledgements................................................................................... 129
List of publications ................................................................................... 131
Curriculum vitae ...................................................................................... 133
i
Summary
Summary
Extraction of key components from cellular material:
aspects of product and process design
Novel engineering techniques promote the development of new technologies that improve
product quality. The increasing consumption volume of food additives/ingredients demands
for stringent requirements to comply with current food industry principles. Extracts from
plant material are a rich source of important compounds for nutraceutical or pharmaceutical
applications.
Tea leaves are rich in polyphenols which can be used in e.g. food, cosmetic and
pharmaceutical industry. Polyphenols have been considered to have a health-promoting
potential as antioxidant agents, which predominantly destroy free radicals.
Soybean grains contain oil bodies that can be used among others as emulsifying agent for a
wide variety of products, ranging from e.g. vaccines, food, cosmetics and personal care
products.
The objective of the work reported in this thesis is to provide basic knowledge and to
subsequently elaborate a conceptual process design for isolation of key components from
fresh tea leaves and soybeans in a mild way. Process costs should be acceptable and
product quality/purity should be suitable. Mild conditions are defined as follows: only use
food grade solvents, avoid strong acidic or alkaline conditions, and applies only mild
temperatures (< 40 oC). The Product Driven Process Synthesis (PDPS) methodology is
applied as a structural approach to deliver the appropriate processing routes for polyphenols
extraction from tea leaves and isolation of oil bodies from soy. The PDPS methodology
includes nine hierarchical levels that connect product design and process synthesis taking
into account the laws of thermodynamics, food, organic, and physical chemistry as well as
chemical engineering principles e.g. transport phenomena and separation technology.
Among different technologies, pulsed electric field (PEF) has been selected due to the fact
that it is a non-invasive method for opening the cell structure in tea leaves. The PEF
application to the tea leaves may result in an increase of temperature (by ohmic heating)
and in electrically stimulated damage of cell membranes (by electroporation). The total
specific energy input and the observed temperature increment point to a non-thermally
based increase of the permeability of the cell membranes for polyphenols. The temperature
of the aqueous suspension of the tea leaves after PEF treatment increased upon
intensification of the applied pulsed electrical energy. The highest treatment temperature
ii
Summary
increase achieved in this study was 7 °C after PEF treatment, using a pulsed electrical
energy of 29 kJ/kg at the highest electric field strength of 1.1kV/cm. The obtained results
indicate that under used conditions the increase of temperature did not exceed 10 oC. This
limited temperature increase provided a valid evidence that pulsed electric field (PEF)
processing is a non-thermal and mild method.
In the input-output level of PDPS it was found that polyphenols (PPs) in fresh tea leaves are
located in plant cell organelles called vacuoles. Prior to the actual extraction step, the cell
structure needs to be open in order to reach the PPs. This finding led into the task network
level where screening of different technologies (the “mechanism” its PDPS nomenclature)
for opening the cell structure was performed.
The extraction process has been described with a polynomial model generated from
statistical analysis of the results obtained by the Design of Experiments (DoE) method
using Box-Behnken design. Three factors that are expected to have a significant impact on
the extraction yield were evaluated: the electric field strength (E), the pulse duration or
pulse width (PD), and the number of pulses (N). The results of statistical analysis
demonstrated that the electric field strength was the most significant factor for the
extraction yield. The model developed was used to optimize the process conditions to
maximize the extraction yield of PPs. The optimal combination of E, PD and N allows 32
% extraction yield of PPs.
The pulsed electric field method is a promising technique for opening the cell structure and
for extraction of cellular material in molecular form.
In the case of soybeans, the aqueous extraction process (AEP) was experimentally tested as
an alternative for the solvent based oil extraction process with the objective to isolate intact
oil bodies (OBs) in a mild way. A simple AEP or benchmark was applied on two soy flours
obtained with different grinding methods, coarse (d90=300 µm) and fine (d90=40 µm). The
extractability of the protein coarse flour was significantly better as compared to that the fine
flour (respectively, 48 % and 40 % based on the total protein content). The oil recovery
from the coarse and fine flour was not significantly different (respectively, 23 % and 24.5
% based on the total soybean oil). In addition, to enhance the extraction yield of protein and
oil, three different pretreatments were applied to the aqueous extraction process. The
pretreatments included enzymatic hydrolysis, application of ultrasound and a combination
of these two. We reported that pretreatment with ultrasound reduced the remaining
insoluble fraction and increased the amount of solids extracted into the aqueous phase. The
combination of ultrasound and enzymes resulted in the cream with the highest lipid-to-
protein weight ratio, i.e. 10:1.
The PDPS approach was applied for the design of a process for isolation of OBs. In the
product design part of PDPS, i.e. in the consumer wants and product ideas level, the
iii
Summary
consumer wants were translated into product formulations. The House of Quality method
was used to provide a link between consumer’s wants and the design of a process bringing
together product attributes and measurable product properties. Depending on the
application, whether OBs will be used as a food additive (e.g. as a natural antioxidant in
mayonnaise) or in cosmetics (e.g. in face creams), the product i.e. OBs themselves requires
different physical, chemical, and microbiological properties. In our particular case, OBs
were produced as an intermediate product and they could be used as food natural
antioxidant in, for instance, a mayonnaise.
In the task network level of PDPS methodology, two alternatives were generated based on
experimental results. The major differences between the two proposed task network
alternatives rely on the mechanism associated to the cell wall disruption task. In the first
alternative, cell wall disruption was performed only with enzymes In the second alternative,
enzymes were combined with ultrasonic treatment. These two different treatments were
applied to the soy flour to improve the extraction of the OBs. Cell wall degrading enzymes
were applied to break the cotyledon cell and to make the structure more permeable.
Ultrasound was applied to increase the transport of elements through cellular membranes
and to extract cellular structures from damaged cells by cavitation. These alternatives were
experimentally verified. At the end a combination of enzymes and ultrasound was chosen,
because this combination improved the purity of the cream (final product), enhanced the
protein-lipid separation, and reduced the processing time. For the selected processing route
an overall process design was made. In addition, the estimation of the economic potential
showed that the proposed process for extraction of OBs had a higher economic potential
than the conventional process for the production of soybean oil due to the use of hexane.
Owing to the complexity of food matrices, the product driven process synthesis
methodology has been successfully applied as a useful and powerful tool for conceptual
process design for the isolation of key components from raw materials e.g. polyphenols
from fresh tea leaves and oil bodies from soybeans.
1
Chapter 1
Chapter 1
Introduction
During the last decade, the food industry has been facing on technical and economic
changes in society and in the manufacturing as well as food processing. These changes
however, have a significant impact on the entire food supply chain. Food products have to
meet the consumers' demand for a healthy lifestyle. Consumer requirements in the field of
food production have changed considerably: in fact, consumers increasingly believe that
food contribute directly to their health. For example, animal fat contribute to e.g. increases
the risk for a heart disease, cancer, and raises the cholesterol level in blood. On the other
hand, seed oils such as soybean oil and safflower oil provide cancer-preventive antioxidants
by lowering “bad cholesterol”. Vegetable seed oils have not been consumed until the 20th
century, simply because we did not have technologies to extract them.
The consumption of processed seed and vegetable oils has increased dramatically in the
past century. These oils are produced and extracted from seeds like soybean, cottonseed,
sunflower and a few others. Figure 1.1 shows how consumption of polyunsaturated fats
present in seed material has increased in the past century (right) and at the same time
consumption of animal origin fats has decreased (left). Both graphs contain data obtained in
United States.
Figure 1.1. Consumption of animal fat (left) and vegetable and seed oils (right) in the United States in the past
century. Source: Authority Nutrition (2011)
However, conventional processing of different seed materials involves harsh process
conditions which destroy natural ordering, protection and preservation of components in the
plant matrix. First, seeds are crushed and free oil is released. Exposure to air leads to
oxidation that causes off-flavors. A second step is extraction step in oil production. For
extraction an organic solvent such as n-hexane is used. A reason for the use of n-hexane is
the high oil solubility but on the other hand there is a possibility of hexane residue in the
2
Introduction
final product. Later in the process, anti-oxidants need to be added to prevent the unsaturated
oils from going rancid. Rancidity results in a number of undesired effects, such as loss in
quality and/or a more difficult separation of components of interest afterwards. Knowing all
these drawbacks of the current production, the necessity for milder release/disclosure of
plant components is obvious.
This PhD project focuses on conceptual process design of novel and mild process routes for
isolation of valuable components from two different types of raw materials: fresh tea leaves
and soybeans. These two raw materials are different in composition: fresh tea leaves, being
naturally high in water content and soybeans having a low water content. Both raw
materials need their own approach, as the properties differ a lot in terms of components, as
well as in terms of interactions between the plant matrix and organelles. In the case of fresh
tea leaves, polyphenols are compounds of interest. Oil bodies and native (not denatured)
proteins are target compounds in soybeans.
1.1. Fresh tea leaves and polyphenols
Green tea processed from Camellia sinensis leaves is a common, globally consumed
beverage. There are two major kinds of tea, black tea and green tea. Tea contains large
amounts of tannins or phenolic substances (5–27 % w/w) consisting of catechin (flavanol)
and gallic acid units, with those in green tea being higher than those in black tea (Leung
and Foster, 1980). Both green and black teas contain caffeine (1–5 % w/w) with small
amounts of other alkaloids also present. Tea composition varies with climate, season, tea
variety, and age of the leaf. In general, fresh green tea leaves contain 36 % w/w
polyphenols, among which catechins prevail. Pharmacological properties of tea are
primarily due to its alkaloids (caffeine) and catechins, which are divided into four primary
compounds (see Figure 1.2), epicatechin (EC), epicatechin-3-gallate (ECG),
epigallocatechin (EGC), epigallocatechin-3-gallate (EGCG), and four secondary
compounds, catechin (C), catechin gallate (CG), gallocatechin (GC), and gallocatechin
gallate (GCG). EGCG is the predominant catechin present in green tea leaves (48–55 % of
total polyphenols) (Perva-Uzunalić et al., 2006).
3
Chapter 1
Figure 1.2. Chemical structure of major catechins in green tea leaves
Green tea is considered to be one of the world's healthiest drinks and contains the highest
amounts of antioxidants of any tea. Polyphenols in green tea are thought to provide anti-
inflammatory and anti-carcinogenic effects. Epigallocatechin-3-gallate (EGCG) is the most
studied (Perva-Uzunalić et al., 2006) and bioactive polyphenol in tea and has been shown
to be the most effective in eliminating free radicals.
1.2. Soybeans and oil bodies
Soybean (Glycine max L.) is a seed belonging to the family Leguminosae. It is presently the
world’s most important oilseed in terms of total production volume and international trade
(Kapchie, et al., 2011). Soybeans account for about 35 % of total harvested area devoted to
annual and perennial oil crops. Worldwide, soybean meal (~98 %) is the primary source of
protein for poultry and livestock industries. Regardless of its nutritional value, only 2 % is
consumed directly by humans, in the form of soy food products (Jung, 2009; Karki, et al.,
2012).
4
Introduction
1.2.1. Seed structure and chemical composition
Soybean seeds have three major parts, the seed shell or hull (8 %), the germ (2 %) and the
cotyledon (90 %) (Salunkhe, et al., 1992). The structure of soybean seed is presented in
Figure 1.3. Typical cotyledon cells have a cylindrical shape; diameter and length of
cotyledon cells about 30 µm in diameter and 70-80 µm respectively (Campbell, 2010). The
wall consists of polysaccharides, proteins and phenolic compounds (lignins) (Ouhida, et al.,
2002). Oil bodies are lipid storage organelles of about 0.2–2.0 μm in diameter for most
oilseeds and 0.2–0.5 μm in the case of soybean (Bair and Snyder, 1980; Murphy, 1993 and
Wu et al., 2012). Oil bodies consist of a lipid core and shell of oleosins and phospholipids.
Figure 1.3. Structure of the soybean seed. Larger organelles in cells represent protein bodies and lipid bodies are
oil bodies (Rosenthal, et al., 1998)
1.2.2. Soybean oil bodies
In nature, soybean oil is encapsulated within micron-sized oil bodies that are present within
mature or germinating soybean seeds, see Figures 1.2 and 1.3. In Figure 1.4 transmission
electronmicrograph of soybean is shown.
5
Chapter 1
Figure 1.4. Transmission electronmicrograph of a soybean cell (Rosenthal, et al., 1998)
Plant cells store lipids as a food reserve for germinative and postgerminative growth. These
lipids are located in the cytoplasm and they are stored in subcellular particles called oil
bodies (OBs) or oleosomes. These oil bodies consist of a central neutral lipid core (94–
98 % w/w) that is surrounded by a phospholipid monolayer (0.5–2 % w/w) and a shell of
strongly amphiphilic oleosin (0.5–3.5 % w/w) (Li et al., 2002; Chen, et al., 2012). To
stabilize the storage lipids, the OBs are coated with a layer of phospholipids. Such particles
alone, however, would not be stable towards coalescence nor their chemical stability is very
limited due to the presence of phospholipases (Kapchie, et al., 2009). A complete layer of
protein (called oleosin) improves the stability in two ways. Firstly, it protects the
phospholipids monolayer from attack by the phospholipases present in the cell. Secondly, it
gives the oil bodies a negatively charged surface, thereby providing colloidal stability i.e.
preventing aggregation (Tzen, 1992; Hsieh and Huang, 2004). Oleosin proteins provide the
oil bodies with physical and chemical protection against environmental stresses, such as
moisture variations, temperature fluctuations, and the presence of oxidative reagents (Chen,
et al., 1998). Soybean OBs have similar or a better physicochemical stability as compared
to the oil droplets in soybean oil-in-water emulsions produced from isolated ingredients
(Chen et al., 1998). Based on both experimental evidence and theoretical calculations, a
structural model of OBs has been proposed (see Figure 1.5) (Rosenthal et al., 1998). Oil
bodies can be depicted as an oil core (94–98 % w/w), surrounded by a monolayer of
phospholipids (0.6–2 % w/w) containing embedded oleosins (0.6–4 % w/w) (Hsieh
and Huang, 2004).
6
Introduction
Figure 1.5. Oil bodies in seed. Left image is a transmission electron micrograph showing large and conspicuous
storage protein bodies and small but numerous oil bodies. Model of an oil body, which includes a matrix of oil (in
blue) enclosed by a layer of phospholipids (red) and the structural protein oleosin (yellow) is presented. The three
types of molecules are drawn to similar scales, whereas the diameter of the oil body has been reduced 24 times to
magnify the surface structure. (Rosenthal, et al., 1998)
Intact OBs are considered as a natural form of an emulsion that, in situ, protects the oil
from oxidation during storage (Campbell, 2010). Moreover, OBs have the advantage over
solvent extracted oil that they required neither emulsifiers nor high pressure
homogenization during processing (Kapchie, et al., 2011). The high stability of the OBs
makes them suitable application in e.g. for many food products, cosmetic, pharmaceutical.
For example, soybean OBs can be used in food products like dressings, sauces, dips,
beverages, and desserts. OBs can also be used as biocapsules for the encapsulation and
controlled delivery of functional components (Iwanaga, et al., 2007).
1.3. Mild separation pre-treatment techniques in food industry
Unfortunately for the food processing, nature does not provide materials of uniform
chemical and/or physical properties. Raw “materials” for foods have important physical
properties, which influence the separation technique to be selected. For example if there is a
difference in density, centrifugation could be applied as a separation technique. In the case
of proteins, usually solubility as physico-chemical property is used to separate them. This
could be done either by adding a solvent (one type of proteins is soluble in certain solvents,
the other not) or by heat supply (thermal denaturation).
7
Chapter 1
In general, physical separation techniques are defined as those operations which isolate
specific ingredients from a mixture without using a chemical reaction. The separations
usually aim to achieve removal of specific components, in order to increase the added value
of the products, which may be the residue, the extracted components or both. All
separations are based on exploiting the differences in physical and chemical properties of
specific components. The main bottleneck is to separate different fractions without
destabilization of one or more of the product fractions. To be able to exploit the complete
raw material without destroying the cellular content, it is necessary to use mild conditions.
Novel techniques to overcome these bottlenecks should be applicable for a variety of end
products aiming to achieve sufficient quality at large quantities.
An example of a mild, non-invasive technique is pulsed electric field (PEF). PEF is
technique in which high electric field strengths are used for simultaneous cell wall and cell
membrane perforation. One of the constraints for using PEF as a cell-opening method is
that processing material has to contain at least 70 % of moisture. Since fresh tea leaves
contain 75 % of moisture, this was reason why PEF was proposed as cell-opening method
in the case of fresh tea leaves. Tea leaf cells contain very rigid cell wall which makes cell
disruption more energy demanding. Therefore, it is needed to first disrupt the cell
walls/membranes to release cell content after which the polyphenols could be extracted.
Another mild technique is ultrasound. Ultrasound utilizes the process of cavitation to
disrupt the cell wall. Collapsing bubbles driven by bulk pressure variation due to ultrasonic
waves can cause cell and even molecule breakage (Alliger and Ciervo, 1975). In the
extraction of oil bodies from soybeans, an ultrasonic device for cell wall degradation was
used. Ultrasound is very low energy intensive technique compared to conventional cell
disruption techniques.
Degradation by enzymes is also a mild cell wall disruption method. Enzymes are not widely
used in industry because of the high costs of cell lysing enzymes. However, enzymes are
very selective in catalyst for cell wall degradation process. The performance depends on the
composition of the cell wall. Therefore, using enzymes is a mild and effective technique for
cell wall disruption. In a soy project we tested enzymes to degrade the strong cell wall of
soybean flour. To degrade such strong cell wall mechanically, a lot of energy is required.
Specific enzymes degrade specific cell wall compounds and weaken the cell wall so further
mechanical degradation can be carried out at a low energy demand. Moreover, the exact
composition of the cell wall is important to be able to select the proper (mixture) of
enzymes.
8
Introduction
1.4. Product Driven Process Synthesis (PDPS) methodology
In this work, the Product Driven Process Synthesis (PDPS) methodology proposed by
Bongers and Almeida-Rivera (2012) is used for conceptual process synthesis. The PDPS
methodology represents basically the extension of conventional process design taking also
product design into account. Design of processes for structured products (e.g. cosmetic
creams and lotions, margarine, and ice cream) is more difficult when using only
conventional process synthesis tools. Structured products have high added value and they
are often complex multi-phase materials where performance of the product is determined
by their microstructure (Bongers and Almeida-Rivera, 2009).
There are nine levels in the PDPS methodology, see Figure 1.6. It is relevant to mention
that the as the scope of the approach expands it becomes wider for multi-product
integration, scheduling and control (Bongers and Almeida-Rivera, 2012). PDPS has the
advantage that by decomposing the problem into a hierarchy of design levels of increasing
refinement, where complex and emerging decisions are made to proceed from one level to
another (Bongers and Almeida-Rivera, 2009).
9
Chapter 1
Figure 1.6. Product Driven Process Synthesis methodology (Bongers and Almeida-Rivera, 2009)
1.5. Objectives
The objective of the work presented in the thesis is to design processes for isolation of key
components from fresh tea leaves and soybeans in a mild way with acceptable process costs
and a suitable product quality/purity. Processes to be designed need to satisfy specific
10
Introduction
demands in the food industry. Only food grade solvents and materials can be used, due to
the fact that end products will be used as food additives. Processes have to be hygienic and
environmentally friendly.
In the thesis, the PDPS methodology has been used as structured approach in order to
deliver the appropriate processing routes for both raw materials: fresh tea leaves and
soybeans. The PDPS methodology is fed by consumer’s preference studies using House of
Quality tool to weigh and rank them. In the scoping stage of PDPS more effort is put in the
process synthesis by selecting appropriate fundamental tasks that can convert raw material
into desired product. Using “heuristics” for decision making, the number of feasible and
possible process alternatives has been reduced. At the end the outcome of the PDPS is how
the new optimal flowsheet is able to realize the production of the desired product cost
effectively under mild process conditions.
1.6. Outline
In the thesis two raw materials: fresh tea leaves and soybeans are considered. Chapters 2
and 3 cover the fundamental aspects and design of the process for extraction of polyphenols
from tea leaves, respectively.
In Chapter 2, the pulsed electric field (PEF) technique is used as a non-invasive method for
opening the cell structure and extraction of polyphenols (PPs). Upon the PEF treatment,
subsequent extraction of PPs occurred. The disintegration of the cellular membrane is
detected indirectly by a total PPs content measurement. The exact mechanism of electric
field applied on leaf tissue to increase permeability is explained based on collected
experimental results. Parameters such as total energy input and electric resistance are
obtained in order to (better) understand and to evaluate mechanism that takes place in the
cellular membrane under exposure of external electric field (Soliva-Fortuny, et al. 2009;
Leong, et al., 2014)
Chapter 3 describes the Product Driven Process Synthesis (PDPS) methodology is applied
for the conceptual design of an extraction process for polyphenols from fresh tea leaves. To
define operating window for PEF mechanism, a design of experiments (DoE) has been
setup and executed. Three factors that influence extraction yield have been studied: electric
field strength, pulse duration and number of pulses. DoE is used to determine which factors
or interactions between factors significantly influence the extraction yield. Optimization is
performed to maximize extraction yield of PPs.
Chapters 4 and 5 report physical background for extraction of oil bodies from soybeans and
conceptual process design, respectively.
11
Chapter 1
In Chapter 4 the result of several lab-scale aqueous extraction processes (AEP) have been
reported for the isolation of intact oil bodies (OBs). First, simple AEP has been performed
on two different flours (coarse and fine) to determine effect of the particle size on OBs
extraction. Second, the soy flour with larger particle size is pretreated with a commercial
enzyme (E), and with ultrasonication (U) during a specific period. These pretreatments are
applied to investigate whether the mass transfer of the cellular components increased,
comparing to a benchmark (no pretreatment) AEP.
Design of the soy process by applying the PDPS has been explained and described in
Chapter 5. In the first part of PDPS, product design was defined taking into consideration
consumer wants. A House of Quality diagram is used to weigh and rank consumer wants.
This led to different product ideas being either food additive (intact OBs can be used as
natural antioxidants) or cosmetics (face/hand creams). Based on the product application,
two process alternatives were generated which were not obvious. Feasibility of these
alternatives was confirmed experimentally. Economic evaluation of conventional process
and proposed alternatives are calculated.
Finally, Chapter 6 summarizes the major conclusions and findings of this PhD project.
Furthermore, an outlook for the future work is presented.
1.7. References
Asavasanti, S., Ersus, S., Ristenpart, W., Stroeve, P., & Barrett, D. M. (2010). Critical
electric field strengths of onion tissues treated by pulsed electric fields. Journal of
Food Science, 75(7), E433–43.
Bhatla, S. C., Kaushik, V., & Yadav, M. K. (2010). Use of oil bodies and oleosins in
recombinant protein production and other biotechnological applications.
Biotechnology Advances, 28(3), 293–300.
Bongers, P. M. M., & Almeida-Rivera, C. (2009). 19th European Symposium on Computer
Aided Process Engineering. Computer Aided Chemical Engineering (Vol. 26, pp.
231–236). Elsevier.
Bongers, P. M. M., & Almeida-Rivera, C. (2012). 11th International Symposium on
Process Systems Engineering. Computer Aided Chemical Engineering (Vol. 31, pp.
195–199). Elsevier.
Campbell, K. a., & Glatz, C. E. (2009). Mechanisms of aqueous extraction of soybean oil.
Journal of Agricultural and Food Chemistry, 57(22), 10904–10912.
doi:10.1021/jf902298a
12
Introduction
Canatella, P. J., Karr, J. F., Petros, J. A., & Prausnitz, M. R. (2001). Quantitative study of
electroporation-mediated molecular uptake and cell viability. Biophysical Journal,
80(2), 755–64.
Cussler, E.L. and Moggridge, G. D. (2011). Chemical Product Design (Second edi.). New
York: Cambridge
Edwards, M. F. (2006). Product Engineering. Chemical Engineering Research and Design,
84(4), 255–260.
Haberl, S., Miklavcic, D., Sersa, G., Frey, W., & Rubinsky, B. (2013). Cell membrane
electroporation-Part 2: the applications. IEEE Electrical Insulation Magazine, 29(1),
29–37.
Huisman, M. M. H., Schols, H. A., & Voragen, A. G. J. (1999). Enzymatic degradation of
cell wall polysaccharides from soybean meal. Carbohydrate Polymers, 38(4), 299–
307.
Iwanaga, D., Gray, D., & Fisk, I. (2007). Extraction and characterization of oil bodies from
soy beans: a natural source of pre-emulsified soybean oil. Journal of Agricultural …,
8711–8716.
Jung, S. (2009). Aqueous Extraction of Oil and Protein From Soybean and Lupin: a
Comparative Study. Journal of Food Processing and Preservation, 33(4), 547–559.
Kapchie, V. N., Towa, L. T., Hauck, C. C., & Murphy, P. a. (2011). Recovery and
Functional Properties of Soy Storage Proteins from Lab- and Pilot-Plant Scale
Oleosome Production. Journal of the American Oil Chemists’ Society, 89(5), 947–
956.
Lebovka, N. I., Bazhal, M. I., & Vorobiev, E. (2000). Simulation and experimental
investigation of food material breakage using pulsed electric field treatment. Journal
of Food Engineering, 44(4), 213–223.
Lebovka, N. I., Shynkaryk, M. V., El-Belghiti, K., Benjelloun, H., & Vorobiev, E. (2007).
Plasmolysis of sugarbeet: Pulsed electric fields and thermal treatment. Journal of
Food Engineering, 80(2), 639–644.
Li, Z., Huang, D., Tang, Z., & Deng, C. (2010). Microwave-assisted extraction followed by
CE for determination of catechin and epicatechin in green tea. Journal of Separation
Science, 33(8), 1079–84.
Liu, F., & Tang, C.-H. (2013). Soy protein nanoparticle aggregates as pickering stabilizers
for oil-in-water emulsions. Journal of Agricultural and Food Chemistry, 61(37),
8888–98.
13
Chapter 1
Moggridge, G. D., & Cussler, E. L. (2000). An Introduction to Chemical Product Design.
Chemical Engineering Research and Design, 78(1), 5–11.
Monsanto, M., Trifunovic, O., Bongers, P., Meuldijk, J., & Zondervan, E. (2014). Black tea
cream effect on polyphenols optimization using statistical analysis. Computers &
Chemical Engineering, 66, 12–21.
Montgomery, D. (2013). Design and Analysis of Experiments (8th ed.). New York: wiley.
Murphy, D. J. (1993). Structure, function and biogenesis of storage lipid bodies and
oleosins in plants. Progress in Lipid Research, 32(3), 247–280.
Nikiforidis, C. V, & Kiosseoglou, V. (2009). Aqueous extraction of oil bodies from maize
germ (Zea mays) and characterization of the resulting natural oil-in-water emulsion.
Journal of Agricultural and Food Chemistry, 57(12), 5591–6.
Ouhida, I., Pérez, J., & Gasa, J. (2002). Soybean (Glycine max) cell wall composition and
availability to feed enzymes. Journal of Agricultural and Food
Perva-Uzunalić, A., Škerget, M., Knez, Ž., Weinreich, B., Otto, F., & Grüner, S. (2006).
Extraction of active ingredients from green tea (Camellia sinensis): Extraction
efficiency of major catechins and caffeine. Food Chemistry, 96(4), 597–605. 15
Rosenthal, a., Pyle, D. L., & Niranjan, K. (1996). Aqueous and enzymatic processes for
edible oil extraction. Enzyme and Microbial Technology, 19(6), 402–420.
Rosenthal, A., Pyle, D. ., Niranjan, K., Gilmour, S., & Trinca, L. (2001). Combined effect
of operational variables and enzyme activity on aqueous enzymatic extraction of oil
and protein from soybean. Enzyme and Microbial Technology, 28(6), 499–509.
Ruan, J., Haerdter, R., & Gerendás, J. (2010). Impact of nitrogen supply on carbon/nitrogen
allocation: a case study on amino acids and catechins in green tea [Camellia sinensis
(L.) O. Kuntze] plants. Plant Biology (Stuttgart, Germany), 12(5), 724–34.
Salunkhe, D.K., Chavan, J.K., Adsule, R.N. & Kadam, S. S. (n.d.). World Oilseeds.
Chemistry, Technology and Utilization. New York: Van Nostrum Reinhold.
Shirsath, S. R., Sonawane, S. H., & Gogate, P. R. (2012). Intensification of extraction of
natural products using ultrasonic irradiations—A review of current status. Chemical
Engineering and Processing: Process Intensification, 53, 10–23.
Soliva-Fortuny, R., Balasa, A., Knorr, D., & Martín-Belloso, O. (2009). Effects of pulsed
electric fields on bioactive compounds in foods: a review. Trends in Food Science &
Technology, 20(11-12), 544–556.
14
Introduction
Tzen, J. T. (1992). Surface structure and properties of plant seed oil bodies. The Journal of
Cell Biology, 117(2), 327–335. 7
Weaver, J. C. (2000). Electroporation of cells and tissues. IEEE Transactions on Plasma
Science, 28(1), 24–33. doi:10.1109/27.842820
Wilczek, M., Bertling, J., & Hintemann, D. (2004). Optimised technologies for cryogenic
grinding. International Journal of Mineral Processing, 74, S425–S434.
Yarmush, M. L., Golberg, A., Serša, G., Kotnik, T., & Miklavčič, D. (2014).
Electroporation-based technologies for medicine: principles, applications, and
challenges. Annual Review of Biomedical Engineering, 16, 295–320
15
Chapter 2
Chapter 2
A study of mechanism involved during polyphenol
extraction from fresh tea leaves by pulsed electric field
ABSTRACT
The major interest in pulsed electric field (PEF) treatment of biological tissues is derived
from its non-thermal application: increasing cell permeability. This application has an
important implication in extraction of complex organic molecules. In this work, pulsed
electric field treatment is investigated as a mild (non-thermal) processing method for
opening the cell structure in fresh tea leaves. Pulsed electric field utilizes short-duration
high-voltage pulses for opening the cell structure by the process called electroporation.
Upon the treatment subsequent extraction of complex organic molecules, particularly
polyphenols occurs. The amount of extracted polyphenols (in this case the extraction yield)
has been determined as a function of electric field strength, duration and number of applied
pulses, as well as energy input per unit of mass of the sample. The results indicate that the
used conditions during the treatment increase in temperature didn’t exceed 10 oC. This
limited temperature rise provides a valid evidence that pulsed electric field (PEF)
processing is a non-thermal method applied under used conditions.
16
A study of mechanism involved during polyphenol extraction from fresh tea
leaves by pulsed electric field
2.1. Introduction
Recently, commercial interest in extraction of intracellular organic compounds and liquids
from cellular plant tissue using various “solid”-liquid extraction methods has been growing.
One of the factors influencing the extraction process is the degree of cell membrane
disintegration (Bazhal, et al., 2003). Various physical, chemical or biological treatments
damage of the cellular membrane. Pulsed electric field (PEF) technique is a non-thermal
processing method for heterogeneous food materials. Moreover, PEF is based on cell
transformation or rupture under exposure of an external electric field resulting in an
increase of the electrical conductivity and permeability of the cell membrane (Lebovka,et
al., 2000; Soliva-Fortuny, et al., 2009, (Haberl, et al., 2013). The application of electric
fields for a short duration of a few to several hundred microseconds is capable of inducing
cell membrane permeabilization through a phenomenon called “electroporation”
(Asavasanti, et al., 2011). The term electroporation is used to describe the phenomena that
accompany the exposure of cells to transmembrane electrical pulses (Weaver, 2000).
Applying an external electric field to the cells results in pore formation in the membrane.
Because pore formation is a dynamic process depending on the intensity of the PEF
treatment, electroporation can be reversible or irreversible. When induced pores are small
in comparison to the membrane area and if they are generated with PEF treatment of low
intensity the electric breakdown is reversible (Angersbach et al., 2000; Soliva-Fortuny, et
al., 2009). Increasing the intensity of the treatment by increasing the electric field strength
(E) and/or treatment time (t) (which considers the number of pulses and the pulse width
applied in the system) will result in the formation of large pores and reversible
permeabilization will turn into irreversible disruption of the cell membrane. The
irreversible permeabilization of the cell membrane in the plant tissue provides a wide range
of process applications where disruption of the cell membrane is required. Both reversible
and irreversible electroporation have found their application in different disciplines such as
biomedicine, biotechnology and environmental science see e.g. (Haberl, et al., 2013;
Yarmush, et al., 2014). In food processing currently intensive research has been done in
non-thermal preservation and sterilization by microbial inactivation (Wan, et al., 2009;
Monfort et al., 2010; Walkling-Ribeiro et al., 2011; Evrendilek, et al., 2013).
Electric field strength is an important factor that controls the efficiency of electroporation
of cellular tissue. Bazhal and authors (2003) presented a classification of the PEF modes as
low (E≤100-200 V/cm), moderate (E=300-1500 V/cm) and high (E>1500 V/cm). With a
low electric field strength, the treatment time should be longer for electroporation of the
cellular membranes. It has been found experimentally that the time needed for
electroporation of cellular membranes of the different biological tissues is inversely
proportional to the electric field strength (Bouzrara and Vorobiev, 2003).
17
Chapter 2
PEF processing of foods involves the application of short pulses (duration of micro- to
milliseconds) of high electric field intensity. The pulsed electric field application to the tea
leaves can result in increase of temperature (through ohmic heating) and electrically
stimulated damage of cell walls/membranes (through electroporation mechanism).
Therefore, in practice, it is difficult to separate two mechanisms (thermal and/or
electroporation mechanism) and to make conclusions about the real contribution of the
electric field to electroporation followed by transfer of polyphenols from the cell interior to
the surrounding liquid. By measuring the total amount of polyphenols (PPs) transferred it
was possible to monitor the process of pulse-induced membrane permeabilization.
Therefore, the objective of this work is to study the effect of operational parameters
(electric field strength, pulse duration and number of pulses) on extraction yield of
polyphenols. Total specific energy input (kJ/kg leaves) and temperature increment (oC) are
chosen as parameters to describe treatment intensity as non-thermal.
2.2. Experimental procedure
2.2.1. Composition of tea leaf
Fresh tea leaves from Kenya (Camellia Sinensis variety) were used. Tea leaf contains 30 %
of catechins, 2 % simple polyphenols (gallic acid), 3 % caffeine and the rest (proteins,
minerals, organic acids and carbohydrates). The composition is given as weight fraction (%
wt/wt) on dry mater. Catechins, gallic acid and caffeine concentrations have been measured
by HPLC.
2.2.2. Moisture content analysis
Fresh tea leaves were plucked one day before shipment to Europe (The Netherlands) and
they were stored at +5 oC until required. Before treatment, leaves were allowed to reach the
ambient temperature of approximately 20 o
C. To determine the moisture content, fresh
leaves were freeze dried and subsequently subjected to a hot-air oven at 105 o
C. Moisture
content was 75-80 wt %. Samples that were used for the moisture content analysis were
separated from the samples used for PEF treatment. For all experiments, thin slices of fresh
tea leaves were used (approximately 1 cm width).
18
A study of mechanism involved during polyphenol extraction from fresh tea
leaves by pulsed electric field
2.2.3. Natural extraction without electrical treatment: “control”
experiments
For control experiments without electrical treatment, slices of fresh tea leaves
(untreated, m = 20 g) were placed into a cylindrical glass beaker. Distilled water (200 g)
was added at 20 °C, and then diffusion was studied. A careful agitation at 250 rpm was
provided. The concentrations of total polyphenols in water were measured after 20 min.
Using the Fourier number it is possible to determine the length of experiments in order to
reach the steady state. In that moment amount of extracted polyphenols is in equilibrium
with the rest of polyphenols in the spent leaves. This dimensionless number represents the
current time to reach the steady-state.
20
R
tDF d
(eq. 2.1)
– Diffusion coefficient (m2/s)
– Time (s)
– Particle radius (m)
When Fourier number reaches 1, steady-state is achieved. Time necessary for that is the
time needed for our experiments. Equilibrium can be presented with the following mass
balance. Letter “d” in subscript presents dispersed PPs (in spent leaves) and “c”, continuous
phase (water) in t=0 and t=∞.
)()( 0,,,0, cccddd ccVCcV (eq. 2.2)
– Volume (m3)
– Concentration (kg/m3)
2.2.4. Pulsed electric field treatments
Tea leaves samples were treated using pulsed electric field (PEF) equipment with batch
treatment configuration of the Nutri-Pulse NP110-60 System (IXL Netherlands B.V.)
which consists of a PEF treatment chamber and a high voltage generator. The batch
treatment chamber (100 mm length × 70 mm width × 50 mm height, 350 mL capacity)
consisted of two parallel stainless steel electrodes of 5 mm thickness separated by a
distance of 20 mm (see Figure 2.1a). A high voltage generator provided rectangular pulses
in the range of 0.1*10-3
-0.1 s with maximum voltage of 2.2 kV and maximum number of
19
Chapter 2
pulses 50. Samples were placed in the treatment chamber between the two stainless steel
electrodes filled with an sterile salt solution (Figure 2.1a). The conductivity of the sterile
salt solution was adjusted to correspond with the conductivity of the sample (σ=3.5
mS/cm). NaCl as a salt was used for preparation of the salt solution. After cutting, the
leaves were subjected to various PEF treatments. All experiments were carried out using
electric field strength (E) ranging from 0.1 to 1.1 kV/cm, pulse duration (PD) from 0.1*10-3
to 0.1 s, and number of pulses (N) from 10 to 50. The total treatment time was defined as
the product of the number of pulses and the pulse width applied in the system (Figure 2.1b).
Apart from the treatment time pause between two pulses (PBP) is the interval between two
pulses and represents the relaxation time.
(a) (b)
Figure 2.1. Experimental set up (lab scale). Scheme of PEF treatment chamber (a) and PEF pulse protocol (b).
The temperature was measured both at the inlet and outlet of the treatment chamber. In all
experiments, the increase of the temperature due to the treatment never exceeded 10 o
C. All
experiments were duplicated.
2.2.5. Specific energy input
The pulse shape (square wave bipolar) was monitored on-line with an oscilloscope (Model
RTM 2000) during the treatment. For a square-wave pulse, the specific energy input for
each PEF-treated sample was calculated based on Eq. (2.3). The total specific energy input
was calculated by multiplying the energy delivered per pulse with the number of pulses
applied into the system:
20
A study of mechanism involved during polyphenol extraction from fresh tea
leaves by pulsed electric field
n t
spec
teratment
dttItUm
nW
0 0
)()( (eq.2.3)
Wspec is specific energy input (in kJ/kg). The energy per pulse (in kJ) is calculated from the
power of the pulse (Ppulse) multiplied by pulse duration (s). Ppulse is the result of output
voltage U(t) and the total electric current I(t) supplied to the sample on the basis of Ohm’s
Law. n is pulse number applied to the system (dimensionless) and m is the total weight of
sample (kg) charged to the treatment chamber for the PEF treatment applied.
2.2.6. Determination of extraction yield - measurement of total
polyphenols content
The disintegration of cellular membrane was detected indirectly by a total polyphenols
(PPs) content measurement. Based on the fact that the cellular membrane is ruptured,
transport of the cell material together with polyphenols occurred and the extraction yield is
defined as fractiont of extracted PPs. After treatment, samples were stored at ambient
temperature and left in the aqueous solution for 20 min (extraction time). Upon extraction,
leaves were separated from the aqueous solution. The amount of extracted PPs was
measured in the solution. The total amount of phenols was determined by direct reading of
the absorbance at 725 nm (SpectraMax 190 Absorbance Microplate Reader, USA) of
diluted samples 1/10 (v/v). The total amount of PPs was expressed as gallic acid
equivalents (GAE) by means of a corresponding calibration curve with standard
gallic acid. The extraction yield of PPs was calculated according to equation (eq.2.4):
𝑒𝑥𝑡𝑟𝑎𝑐𝑡𝑖𝑜𝑛 𝑦𝑖𝑒𝑙𝑑 =𝑝𝑜𝑙𝑦𝑝ℎ𝑒𝑛𝑜𝑙𝑠 (𝑎𝑞𝑢𝑜𝑒𝑢𝑠 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛)
𝑝𝑜𝑙𝑦𝑝ℎ𝑒𝑛𝑜𝑙𝑠 (𝑓𝑟𝑒𝑠ℎ 𝑡𝑒𝑎 𝑙𝑒𝑎𝑣𝑒𝑠). 100% (eq.2.4)
2.2.7. Statistical analysis
All experiments and measurements of characteristics were repeated at least twice. One-way
analysis of variance was used for statistical analysis of the data using the Statgraphics
Centurion XVI. For each analysis, significance level of 5 % was assumed. The error bars
presented in the figures correspond to the standard deviations.
21
Chapter 2
2.3. Results and discussion
Fresh tea leaves were subjected to pulsed electric field (PEF) treatments. Extraction of
polyphenols occurred due to pore formation and disintegration of cellular membrane.
Therefore, when PEF treated leaves were in contact with an aqueous solution after 20 min
of extraction time, color had changed in yellowish brown. This phenomenon occurs due to
the transport of cellular material into the surrounding aqueous solution and the reaction of
PPs catalyzed by the enzyme polyphenol oxidase (PPO). PPO is located in the cytoplasm of
the plant cells and PPs are in organelles called vacuoles. When PPO and PPs are brought
into contact, oxidized PPs are produced as result of enzymatic oxidation. The formation of
oxidized PPs was identified by visible observation: a change of color of the solution during
the oxidation reaction occurs. Further confirmation was conducted using UV
spectrophotometer analysis, where total PPs were determined by reading the absorbance at
725 nm.
2.3.1. Characterization of pulsed electric field process and electric
measurements
In the experiments performed, electrical measurements: voltage and current were reordered
for each PEF experiment at regular time intervals. Figure 2.2 shows typical recording of
voltage (V) and current (mA) using an oscilloscope.
Figure 2.2. Typical recording of voltage (blue dotted line) and current (red dotted line) during a PEF experiment
-100
0
100
200
300
400
500
600
0
500
1000
1500
2000
2500
-2.00E-04-1.00E-040.00E+001.00E-042.00E-043.00E-044.00E-045.00E-046.00E-04
Cu
rren
t, m
A
Volt
ag
e, V
Time, s
22
A study of mechanism involved during polyphenol extraction from fresh tea
leaves by pulsed electric field
In addition, in the experiments carried out in this study next to the electrical measurements
(voltage and current), temperature and resistance were also monitored and recorded at
regular intervals during each experiment. Figure 2.3 shows a typical resistance and
temperature time history during a PEF experiment.
Figure 2.3. Typical recording of resistance (blue dotted line) and temperature (red dotted line) during a PEF
experiment
Consequently, it was necessary to record these parameters (temperature and resistance)
during each PEF experiment in order to obtain a better insight for proper evaluation of the
results. Although PEF was intended to be a non-thermal technique, a temperature rise was
present due to the electric current through in the liquid food (ohmic heating). The average
temperature rise (ΔT) in the sample can roughly be estimated. Table 2.1, collects the PEF
treatment conditions for tea samples as well as the treatment impact on the changes in
voltage, current and temperature.
Table 2.1. Summary of PEF treatment conditions for each PEF experiment and the treatment impact on the
changes in electrical parameters and temperature.
Pulse duration,
[10-3,s]
Number of
pulses, [-]
Specific energy
input, [kJ/kg]
Change in
temperature, oC
Low Electric field strength=0.1 kV/cm
0.1 30 0.9 0.1
50 10 1.4 0.3
50 30 4.0 0.9
21.10
21.30
21.50
21.70
21.90
0.00
2.00
4.00
6.00
8.00
10.00
0 5 10 15
Tem
per
atu
re,
C
Res
ista
nce
, Ω
Time, s
23
Chapter 2
100 30 4.0 1.0
Moderate Electric field strength=0.6 kV/cm
0.1 10 1.7 0.4
0.1 50 7.8 1.9
50 30 7.7 1.8
100 10 2.6 0.6
100 50 12.5 3.0
High Electric field strength=1.1 kV/cm
0.1 30 12.0 2.9
50 10 5.9 1.4
50 50 29.7 7.1
100 30 17.5 4.2
When tea leaves were exposed to different intensities of the pulsed electrical energy,
temperature was measured and calculated to evaluate the degree of cell permeabilization
and the effect of pulsed electric field (PEF) treatment. On average, the temperature of
untreated leaves before subjected to PEF treatment was 20oC. As summarized in Table 2.1,
the temperature increased with the increase in applied pulsed electrical energy at the same
level of the applied electric field strength. This increase of temperature is particularly
significant at 0.6 and 1.1 kV/cm for a number of pulses of 30 and 50. The temperature of
the sample with tea leaves after PEF treatment increased when the applied pulsed electrical
energy was intensified. The highest treatment temperature increase observed in this study
was 7.1 °C after PEF treatment, using pulsed electrical energy of 29.7 kg
kJat electric field
strength level of 1.1 kV/cm.
2.3.2. Effect of electric field strength on extraction yield of
polyphenols
Sale and Hamilton (1967) identified the electric field strength E and the total treatment time
(which considers the number of pulses and the pulse duration applied in the system) as the
main variables determining the efficiency of the PEF damage of the plant tissue. Higher
24
A study of mechanism involved during polyphenol extraction from fresh tea
leaves by pulsed electric field
electric field strengths lead to a better damage efficiency (Canatella, et al., 2001; Toepfl, et
al., 2007). However, it was noticed that an optimal value of the electric field strength for
many vegetables and fruit tissue is within 300 to 500 V/cm.
Figure 2.4 presents experimental results for the extraction yield of PPs from fresh tea leaves
as a function of the electric field strengths for various pulse protocols: N=30, PD=0.05 s
and PBP(EXA)=0.5 s (protocol I); N=30, PD=0.05 s and PBP(EXB)=3 s (protocol II). The
only difference between protocol I and II is interval time between two pulse series, see
Figure 2.1b.
The extraction yield (EY) values of the samples subjected to both protocols increase with
increasing electric field strength. The PEF treatment accelerates the rate of the extraction of
PPs from fresh tea leaves to the surrounding liquid. This is in agreement with the behavior
observed for different fruit and vegetable tissues (Lebovka, 2009). Protocol I resulted in a
maximum value for the extraction yield of 27 % when the electric field strength is 0.9
kV/cm. However, when the interval between pulses is equal to 3 s an electric field of 1.1
kV/cm is needed to obtain the same extraction yield (protocol II).
Figure 2.4. Experimental results for extraction yield of polyphenols from fresh tea leaves vs electric field strength
at two pulse protocols: N=30, PD=0.05 s and PBP=0.5 s (protocol I) and N=30, PD=0.05 s and PBP=3 s
(protocol II). Protocol I is presented with dotted bars, and protocol II with grey squared lines. Displayed error
bars represent extraction yield values +/- standard experimental error. In all cases, a sample was taken 20 min
after the end of the PEF treatment and stored at -30 oC until the extracted amount of total PPs was analyzed. All
experiments were duplicated
For both protocols the same extraction yield was obtained, but at different electric field
strengths. When the interval between two pulse series (pause between pulses) is short
0.0
5.0
10.0
15.0
20.0
25.0
30.0
0.1 0.4 0.6 0.9 1.1
Ex
tract
ion
yie
ld,
%
Electric field strength, kV/cm
25
Chapter 2
(protocol I), a moderate electric field seems to be sufficient to cause cell membrane rupture.
However, protocol II allows a longer relaxation time and a higher electric field is needed to
achieve an extraction yield of 27 %. So these results point to the rate of extraction that is
significantly dependent on the time interval between to pulse series. Also, a longer relation
time asks for higher electric fields to achieve the same extraction yield.
The variation in the electric field strength to obtain the same extraction yield between
protocol I and II can also be related to the ionic transport between the electrodes and
electrolysis and the tissue surface. Depending on the details of contact (geometry and size
of the samples, orientation of the leaf slices, etc.) as well as the composition of the samples
(bud leaf, 1st, 2
nd and 3
rd open leaf), electrolysis may give rise to different amounts of stable
ionic compounds which would result in an increase of conductivity and tissue
disintegration.
2.3.3. Effect of the total treatment time on the extraction yield of
polyphenols
The degree of disintegration strongly depends on the treatment time and the electric field
strength ( Lebovka, 2009). For long times of PEF treatment a smaller electric field is
required. Figure 2.5 presents experimental results for the extraction yield at different
treatment times for two electric field strengths 0.4 and 0.9 kV/cm, respectively.
Lebovka, et al. (2007) reported the effect of pulse duration on the efficiency of PEF-
treatment on sugar beet. Experiments showed that a longer pulse duration see Figure 2.1b
was more effective. This influence of pulse duration was particularly pronounced at
moderate electric field strength (E=0.3 kV/cm). This is partially in agreement with the
observation because the highest extraction yield of 27 % is obtained for moderate electric
field strength (E=0.4 kV/cm) and a treatment time of 2.5 s (Figure 2.5). However, almost
the same extraction yield (EY=26.6 %) is obtained for a higher electric field (E=0.9 kV/cm)
and a shorter treatment time of 1.5 s. When unfortunately the total treatment time was 5 s
(for both electric field strengths) the experiment could not be performed due to problems
with PEF unit itself (low conductivity was detected in the PEF chamber). A possible
explanation lies in Ohm’s law. Electrical resistance is the ratio of voltage over current by
Ohm’s law and conductivity is inversely proportional to resistance. Therefore, the
resistance of the treatment chamber is an important parameter since the maximum allowed
pulse current by the power switch is 600 mA. This means that at 4kV the minimum
resistance is 6 Ohm. If the resistance is lower than this the maximum pulse current of 600
26
A study of mechanism involved during polyphenol extraction from fresh tea
leaves by pulsed electric field
mA will be exceeded and if this situation continues for more than 5 pulses the system will
automatically shut down to avoid damage to the high voltage switch. In this particular case,
treatment time was 5 s the total number of pulses (N=50) and the pulse duration (PD=0.1 s).
For the applied electric field strengths of 0.4 and 0.9 kV/cm, respectively, the critical
situation mentioned above is exceed and PEF equipment will shut down.
Figure 2.5. Experimental results for extraction yield of polyphenols from fresh tea leaves vs different treatment
times at different values of electric field strength 0.4 (blue dotted bars) and 0.9 dark dotted bars) kV/cm,
respectively for the fixed value of PBP=0.5 s. Displayed error bars represent extraction yield values +/- standard
experimental error (95 % confidence interval). In all cases, sample was taken 20 min after the end of PEF
treatment and stored at -30 oC until extracted amount of total PPs was analyzed. All experiments were duplicated.
Existing work mainly discusses the effects of pulse duration in the PEF induced
inactivation of different microorganisms. Some authors have demonstrated that inactivation
was more efficient at longer pulse duration (Belloso et al. 1997; Abram et al., 2003).
However, other reported little effect of the pulse duration on inactivation (Raso et al. 2000;
Manas et al. 2000; Sampedro et al. 2007). The effect of pulse duration seems to vary
depending on the electric field strength and a general relationship between PEF-treatment
protocols, type and quality of the soft tissue, contact parameters (geometry and size of the
samples and orientation of the leaf slices), and the resulting degree of material
disintegration requires more thorough study. A quantitative description of the performance
of PEF for extraction of key components from cellular material (e.g. PPs from fresh tea
leaves and proteins from sugar beet leaves) from first principle of chemistry and physics
0
5
10
15
20
25
30
35
40
0.005 0.75 1.25 1.5 2.5 5
Ex
tract
ion
yie
ld,
%
Treatment time, s
27
Chapter 2
cannot be given at the moment. Therefore, a model developed from statistical analysis of
the experimental results will be given in Chapter 3.
2.4. Conclusion
This study provides valid evidence that pulsed electric field (PEF) processing is non-
thermal method (depending on the used conditions) for extraction of polyphenols from
fresh tea leaves. Electric field strength as well as treatment time play an important role in
polyphenols extraction. Moreover, the efficiency of the PEF treatment was indirectly
connected to electric field effect i.e. the relaxation time after a series of pulses. Different
modes of PEF treatment (electric field strengths and total treatment times) were applied to
investigate their effect on the extraction of polyphenols from fresh tea leaves. The amount
of extracted polyphenols from the leaves into the aqueous media strongly depends on the
setting of PEF treatment. Protocol I (EXA) resulted in a maximum extraction yield of 27 %
when the electric field strength is 0.9 kV/cm. However, when the interval between pulses is
longer and equal to 3 s an electric field of 1.1 kV/cm is needed to obtain the same
extraction yield (protocol II-EXB).
The total treatment time was presented as the product of the number of pulses and pulse
width (duration). Experimental results demonstrated that longer pulses which means a
longer treatment time of 2.5 s were more effective for moderate electric field (E=0.4
kV/cm). Moreover, to achieve the same extraction yield of 27 % but with a shorter total
treatment time=1.5 s, a higher electric field (E=0.9 kV/cm) is required. Also a total
treatment time of 5 s (for both electric field strengths) the experiment was unfortunately not
possible due to limitation of the PEF unit.
2.5. References
Asavasanti, S., Ersus, S., Ristenpart, W., Stroeve, P., & Barrett, D. M. (2010). Critical
electric field strengths of onion tissues treated by pulsed electric fields. Journal of
Food Science, 75(7), E433–43.
Bair, C. W., Snyder, H. E., & Technology, F. (1980). , Electron Microscopy of Soybean-
Lipid Bodies Lipid Protein, (September), 279–282.
Bazhal, M. I., Ngadi, M. O., & Raghavan, V. G. S. (2003). Influence of Pulsed
Electroplasmolysis on the Porous Structure of Apple Tissue. Biosystems Engineering,
86(1), 51–57.
28
A study of mechanism involved during polyphenol extraction from fresh tea
leaves by pulsed electric field
Ben Ammar, J., Lanoisellé, J.-L., Lebovka, N. I., Van Hecke, E., & Vorobiev, E. (2011).
Impact of a pulsed electric field on damage of plant tissues: effects of cell size and
tissue electrical conductivity. Journal of Food Science, 76(1), E90–7.
Bouzrara, H., & Vorobiev, E. (2003). Solid–liquid expression of cellular materials
enhanced by pulsed electric field. Chemical Engineering and Processing: Process
Intensification, 42(4), 249–257.
Canatella, P. J., Karr, J. F., Petros, J. A., & Prausnitz, M. R. (2001). Quantitative study of
electroporation-mediated molecular uptake and cell viability. Biophysical Journal,
80(2), 755–64.
Evrendilek, G. A., Altuntas, J., Sangun, M. K., & Zhang, H. Q. (2013). Apricot Nectar
Processing by Pulsed Electric Fields. International Journal of Food Properties,
16(1), 216–227.
Gonzalez, M. E., & Barrett, D. M. (2010). Thermal, high pressure, and electric field
processing effects on plant cell membrane integrity and relevance to fruit and
vegetable quality. Journal of Food Science, 75(7), R121–30.
Haberl, S., Miklavcic, D., Sersa, G., Frey, W., & Rubinsky, B. (2013). Cell membrane
electroporation-Part 2: the applications. IEEE Electrical Insulation Magazine, 29(1),
29–37.
Lebovka, N. (2009). Electrotechnologies for Extraction from Food Plants and
Biomaterials. New York, NY: Springer New York.
Lebovka, N. ., Bazhal, M. ., & Vorobiev, E. (2001). Pulsed electric field breakage of
cellular tissues: visualisation of percolative properties. Innovative Food Science &
Emerging Technologies, 2(2), 113–125.
Lebovka, N. ., Bazhal, M. ., & Vorobiev, E. (2002). Estimation of characteristic damage
time of food materials in pulsed-electric fields. Journal of Food Engineering, 54(4),
337–346.
Lebovka, N. I., Bazhal, M. I., & Vorobiev, E. (2000). Simulation and experimental
investigation of food material breakage using pulsed electric field treatment. Journal
of Food Engineering, 44(4), 213–223.
Lebovka, N. I., Shynkaryk, M. V., El-Belghiti, K., Benjelloun, H., & Vorobiev, E. (2007).
Plasmolysis of sugarbeet: Pulsed electric fields and thermal treatment. Journal of
Food Engineering, 80(2), 639–644.
Lee, K. J., & Lee, S. H. (2008). Extraction behavior of caffeine and EGCG from green and
black tea. Biotechnology and Bioprocess Engineering, 13(5), 646–649.
29
Chapter 2
Leong, S. Y., Richter, L.-K., Knorr, D., & Oey, I. (2014). Feasibility of using pulsed
electric field processing to inactivate enzymes and reduce the cutting force of carrot
(Daucus carota var. Nantes). Innovative Food Science & Emerging Technologies.
Leung, A., & Foster, S. (1980). Encyclopedia of common natural ingredients. (A. Y.
Leung, Ed.)Encyclopedia of Common Natural Ingredients used in Food Drugs and
Cosmetics (pp. 142–143). John Wiley & Sons,.
Monfort, S., Gayán, E., Saldaña, G., Puértolas, E., Condón, S., Raso, J., & Álvarez, I.
(2010). Inactivation of Salmonella Typhimurium and Staphylococcus aureus by
pulsed electric fields in liquid whole egg. Innovative Food Science & Emerging
Technologies, 11(2), 306–313. doi:10.1016/j.ifset.2009.11.007
Monsanto, M., Trifunovic, O., Bongers, P., Meuldijk, J., & Zondervan, E. (2014). Black tea
cream effect on polyphenols optimization using statistical analysis. Computers &
Chemical Engineering, 66, 12–21.
Perva-Uzunalić, A., Škerget, M., Knez, Ž., Weinreich, B., Otto, F., & Grüner, S. (2006).
Extraction of active ingredients from green tea (Camellia sinensis): Extraction
efficiency of major catechins and caffeine. Food Chemistry, 96(4), 597–605.
Ruan, J., Haerdter, R., & Gerendás, J. (2010). Impact of nitrogen supply on carbon/nitrogen
allocation: a case study on amino acids and catechins in green tea [Camellia sinensis
(L.) O. Kuntze] plants. Plant Biology (Stuttgart, Germany), 12(5), 724–34.
Sale, A., & Hamilton, W. (1967). Effects of high electric fields on microorganismsI. Killing
of bacteria and yeasts. Biochimica et Biophysica Acta (BBA) - General Subjects,
148(3), 781–788.
Soliva-Fortuny, R., Balasa, A., Knorr, D., & Martín-Belloso, O. (2009). Effects of pulsed
electric fields on bioactive compounds in foods: a review. Trends in Food Science &
Technology, 20(11-12), 544–556.
Stadnik, J., & Dolatowski, Z. J. (2011). Influence of sonication on Warner-Bratzler shear
force, colour and myoglobin of beef (m. semimembranosus). European Food
Research and Technology, 233(4), 553–559.
Toepfl, S., Heinz, V., & Knorr, D. (2007). High intensity pulsed electric fields applied for
food preservation. Chemical Engineering and Processing: Process Intensification,
46(6), 537–546.
Vilkhu, K., Mawson, R., Simons, L., & Bates, D. (2008). Applications and opportunities
for ultrasound assisted extraction in the food industry — A review. Innovative Food
Science & Emerging Technologies, 9(2), 161–169.
30
A study of mechanism involved during polyphenol extraction from fresh tea
leaves by pulsed electric field
Walkling-Ribeiro, M., Rodríguez-González, O., Jayaram, S., & Griffiths, M. W. (2011).
Microbial inactivation and shelf life comparison of “cold” hurdle processing with
pulsed electric fields and microfiltration, and conventional thermal pasteurisation in
skim milk. International Journal of Food Microbiology, 144(3), 379–86.
Wan, J., Coventry, J., Swiergon, P., Sanguansri, P., & Versteeg, C. (2009). Advances in
innovative processing technologies for microbial inactivation and enhancement of
food safety – pulsed electric field and low-temperature plasma. Trends in Food
Science & Technology, 20(9), 414–424.
Wang, H., Provan, G. J., & Helliwell, K. (2003). HPLC determination of catechins in tea
leaves and tea extracts using relative response factors. Food Chemistry, 81(2), 307–
312.
Weaver, J. C. (2000). Electroporation of cells and tissues. IEEE Transactions on Plasma
Science, 28(1), 24–33.
Yarmush, M. L., Golberg, A., Serša, G., Kotnik, T., & Miklavčič, D. (2014).
Electroporation-based technologies for medicine: principles, applications, and
challenges. Annual Review of Biomedical Engineering, 16, 295–320.
Zhu, W.-L., Shi, H.-S., Wei, Y.-M., Wang, S.-J., Sun, C.-Y., Ding, Z.-B., & Lu, L. (2012).
Green tea polyphenols produce antidepressant-like effects in adult mice.
Pharmacological Research : The Official Journal of the Italian Pharmacological
Society, 65(1), 74–80.
31
Chapter 3
Chapter 3
Product-driven process synthesis for the extraction of
polyphenols from fresh tea leaves
ABSTRACT
Polyphenolics present in fresh tea leaves may reduce the risk of a variety of illnesses,
including cancer and coronary heart diseases. The usefulness of these polyphenols may be
extended by combining them with other consumer products such as food items and vitamin
supplements. However, during conventional tea processing that includes cutting, rolling,
and drying, the extraction step is missing. Therefore, there is a need for an alternative
process to isolate polyphenols under mild conditions. The Product-Driven Process
Synthesis methodology is used as a well-defined structured approach for the conceptual
design of an extraction process for polyphenols from fresh tea leaves. A detailed
specification of the starting material (fresh tea leaves) and product (polyphenols) leads to a
subsequent definition of fundamental tasks to convert our raw material into the desired
product. Among the different mechanisms and techniques that could be used to perform the
tasks under mild conditions, pulsed electric field has been selected as non-invasive and
non-thermal method for cell wall disruption. To define the operating window for pulsed
electric field method an experimental design has been defined and executed (varying
several settings of the pulsed electric field). From the collected data, the analysis of
variance has been used to determine which variables are significant i.e. electric field
strength, pulse duration and number of pulses. Box-Behnken design has been used as part
of statistical analysis to find optimal pulsed electric field settings to maximize the
extraction yield of polyphenols (extraction yield). With obtained optimum settings maximum
value of 32 % of extraction yield was achieved.
32
Product-driven process synthesis for the extraction of polyphenols from fresh
tea leaves
3.1. Introduction
Technological and scientific progress in food industry has been pushed towards extraction
of intracellular compounds and liquids from cellular plant tissue. The shift from
conventional processing of food plant material towards design of smart operation
techniques is made to explore a novel product range. Therefore, in the last ten years the link
between process design and the development of novel consumer products became
increasingly important, especially regarding to the design of processes for structured
products. These structured products have high added value and they are often complex
multi-phase materials (e.g. cosmetics creams and lotions, margarine, ice cream, etc.) see
e.g. Edwards (2006). The Product-Driven Process Synthesis (PDPS) methodology proposed
by Bongers and Almeida-Rivera (2012) connects product design with process synthesis.
The PDPS method comprises a multi-level decision hierarchy with increasing level of
complexity that helps the user in the development of new products and processes.
Table 3.1. General description of each level of PDPS approach (Bongers and Almeida-Rivera, 2009)
Level General description of each level
(0) Framing level
Description of the project background
(BOSCARD) and business context as well as
overall supply chain
(1) Consumer wants
Consumer wants are obtained in qualitative
descriptions consumer likings, focus groups,
interviews and translate them into quantifiable
product attributes
(2) Product ideas
Potential products that meet the consumer
wants are identified and mapped the
quantifiable product attributes onto the product
properties, which are measurable
(3) Input output level
Complete specification of the output is done.
Inputs (ingredients or raw materials) and the
descriptors of the output (e.g. microstructure)
are specified.
(4) Task network
Definition of the fundamental tasks needed to
convert input into output, taken from a cluster
of tasks and its subgroup is performed. Then, a
network is made from the selected tasks and
clusters.
33
Chapter 3
(5) Mechanism and operating
window
Mechanisms* that can perform tasks are
defined. And for every selected task, operating
window needs to be defined.
(6) Multiproduct integration
The outcomes of steps 3 –5 for the different
products are compared to look for overlap and
possibilities to combine the production.
(7) Equipment selection and
design
The operating units are selected taking into
consideration integration possibilities and
controllability.
(8) Multi product-equipment
integration
Multi-stage scheduling of the multiple
products is applied, including plant-wide
control.
* “Mechanisms” is the nomenclature used in the paper of Almeida and Bongers (2010)
In Table 3.1 hierarchy that starts at the framing level and ultimately leads to a complete
conceptual process design (including equipment design, process control and multi-product
equipment integration) is presented. Bongers and Almeida-Rivera (2009) explained the
complete hierarchy in detail.
This contribution will focus on opportunities for product design and process synthesis and
in the strategies towards novel molecular and functional products. In particular, we will use
the isolation of polyphenols from fresh tea leaves to illustrate the applicability and scope of
the PDPS methodology.
3.2. Framing level and product ideas
At the framing level, the background of the project, the business context and the potential
of polyphenols (PPs) as food additives are identified. Polyphenols present in fresh tea
leaves may reduce the risk of a variety of illnesses, including cancer and coronary heart
diseases. The usefulness of these polyphenols may be extended by combining them with
other consumer products such as food items and vitamin supplements. The presence of PPs
in fresh tea leaves is responsible for health benefits associated with green tea consumption.
PPs are now gaining significant attention from both technical and consumer point of view
due to potential health benefit. It has been reported that PPs may reduce risk of cancers,
cardiovascular diseases, dental decay, obesity, diabetes, and improvement in the immune
system (Wang et al., 2003).
34
Product-driven process synthesis for the extraction of polyphenols from fresh
tea leaves
Industrial tea extraction is mainly based on the maceration method (rolling) combined with
stirring, circulation, ultrasonic, microwave or enzymatic treatment (Perva-Uzunalić et al.,
2006; Lee and Lee, 2008; Zhu et al., 2012). These methods require large volumes of
solvent, high temperatures and lot energy. Extraction efficiencies are generally low.
However, during conventional tea processing that includes cutting, rolling, and drying, the
extraction step is missing. Moreover, process conditions are harsh for example, to destroy
the remaining enzymes and to reduce the moisture content in a leaf, high temperatures have
to be applied (above 80 oC). Therefore, there is a need for an alternative process for the
extraction of polyphenols under mild conditions. A novel alternative process has to be an
effective and green method that will allow extraction of polyphenols from fresh tea leaves.
3.3. Input/output level
At the input output level a complete specification of all exchange streams to the process
inputs (raw material(s)) and target output (product(s)) are identified. Fresh tea leaves
(input) contain approximately 75 wt % moisture and 25 wt % of total solids. PPs account
for approximately 30 % of dry weight of fresh tea leaves (Zhu et al., 2012). Fresh tea leaves
contain caffeine, polyphenols, polysaccharides, and necessary nutrients, such as proteins,
amino acids, lipids, and vitamins. Generally, some chemical components – free amino
acids, total tea polyphenols, and soluble sugars – are considered important indicators of tea
quality (Ruan et al., 2010).
The output of the process is a fraction of polyphenols. Green tea leaves contain low-
molecular-weight polyphenols consisting mainly of flavanol (flavan-3-ol) monomers,
which are referred to as catechins (see Figure 3.1, left). There are several isomers of this
compound: catechin, catechin gallate (CG), gallocatechin, gallocatechin gallate (GCG),
epicatechin, epicatechin gallate (ECG), epigallocatechin, and epigallocatechin gallate
(EGCG). Normally, 10–20 % of the catechins in green tea leaves are epigallocatechin and
EGCG. Black tea polyphenols are formed during the enzymatic oxidation of green tea
leaves. In green tea (non-oxidized leaves) mostly catechins are present, while in black tea
several types of polyphenols formed by enzymatic polymerization of catechins, including
theaflavins can be found (Figure 3.1, right).
35
Chapter 3
Figure 3.1: Catechins structure (left) and theaflavin structure (right) (Monsanto et al., 2014)
The performance of input-output level is assessed by simple economic analysis, which
basically estimates economic potential of the process. The difference between product
revenue and raw material costs is computed on a year basis for 1 ton/hr. To estimate the
economic potential (EP) (Cussler and Moggridge 2011) of the proposed process the
following equation was applied:
EP=Product revenues-Raw material cost
𝐸𝑃 = ∑ 𝐶𝑖 ∙ 𝐹𝑖 − ∑ 𝐶𝑗 ∙ 𝐹𝑗𝑗−𝑟𝑎𝑤 𝑚𝑎𝑡.𝑖−𝑝𝑟𝑜𝑑𝑢𝑐𝑡 eq. (3.1)
where Ci and Cj are the sales prices of the products i and the costs of the raw materials j,
respectively. F denotes the annual flow of products and raw materials. The input materials
for this estimation include the fresh tea leaves, water, and NaCl. NaCl as a salt was used for
preparation of the aqueous medium. As output, three different products are considered: pure
(unconverted) catechins, theaflavins (TF) and thearubigins (TR).TR and TF are oxidized
form of catechins. In Table 3.2 the raw materials and final products costs are collected in
detail. For the calculation of the EP, several assumptions are made. The capacity of the
plant is set at 1 ton/h of fresh tea leaves and 20 h/day, giving a final annual flow of 7300 t
of processed leaves. The mass balance obtained is considered as well as the composition
(dry-weight basis) of the output, since the price will depend on the purity of the product
(amount of catechins, and TF in the output).
36
Product-driven process synthesis for the extraction of polyphenols from fresh
tea leaves
Table 3.2. Raw materials and final product prices for economic potential
Materials Costs of the raw materials
Input
Fresh tea leavesa €263/ton
NaClb €50/ton
Waterc €1.25/m3
Sales prices of the products
Output
Catechins* €80/ton catechins
Theaflavins (TF)** €238/ton (TF)
Thearubgins (TR)f €119/ton (TR)
a Indexmundi; cEUROSTAT; *estimated price with price of commercial catechins standard (Sunphenon 90LB); **
estimated price with price of commercial theaflavins standard (TF60); f rough estimation
The economic potential for the process for extraction of PPs from fresh tea leaves is around
3 M€/year, which can be used to invest in equipment, facilities, cover operational costs,
labor costs, especially for the downstream processing (Table 3.3).
Table 3.3. Economic potential at input-output level for extraction of polyphenols from fresh tea leaves
M€/year
Cost of raw
materials
Sales prices of
products Economic Potential
2.3 5.23 2.93
37
Chapter 3
3.4. Task network
The next step in the methodology is to identify the fundamental tasks that are needed to
convert the input into the desired output. i.e. determination of the task network . The aim is
to isolate PPs from fresh tea leaves originally present inside organelle (vacuole) in the cells.
To make the polyphenols accessible the following tasks need to be executed (reference
code is included for the sake of simplicity), see also Table 3.4:
1. Size reduction of leaves (C2)
2. Cell wall disruption (C3)
3. Separation of a system into two systems with different composition (G1)
4. Physical/biological stabilization (J1)
These fundamental tasks are the critical and essential tasks based on the requirement to be
obeyed to obtain the desired final properties of the product. These four tasks (size
reduction, cell wall disruption, separation task and biological stabilization) can be
combined leading to 24 (4!=24) different processing routes. Two feasible task networks can
be formulated. Figures 3.2.1 and 3.2.2 depict our proposed task networks alternatives.
C2Size
reduction
C3Cell wall
disruption
G1Extraction of
PPs
J1Physical/chemical
stability
Figure 3.2.1: Feasible task network route 1
C2Size
reduction
C3Cell wall
disruption
G1Extraction of
PPs
J1Physical/chemical
stability
Figure 3.2.2: Feasible task network route 2
The major difference between the two proposed task networks is order of tasks. Task
network route 1 starts with C2: size reduction. However, task network 2 starts directly with
cell wall disruption step (C3). Both routes are possible and feasible due to the fact that tea
leaves tissue is soft tissue containing approximately 75 % wt/wt moisture.
For every task, there are different “mechanisms” that could be used to perform task. In
Table 3.4 all possible mechanisms for execution of these tasks are presented. Heuristics,
38
Product-driven process synthesis for the extraction of polyphenols from fresh
tea leaves
domain knowledge and project constraints to use mild conditions (i.e. exclusion of organic
solvents, avoiding high temperatures, extreme pH values: strong acidic or alkaline
conditions) are used to eliminate several “mechanisms” and hence reducing the number of
potential alternatives for process synthesis.
Table 3.4: Fundamental tasks and mechanisms
Steps Fundamental task* Mechanisms*
Size reduction of leaves C2
C21: attrition
C22: impact
C23: ultrasound
C24: cutting
C25: enzymes
Cell wall disruption C3
C31: internal cell phase change
C32: electro-magnetic fields
(PEF, ultrasound)
C33: shear
C34: enzymes
C35: chemical
Extraction of PPs G1
G11: molecular size
G12: particle size
G13: electrical charge
G14: solubility
G15: chemical affinity
G16: chemical reaction
G17: (vapour) pressure
G18: gravity
G19: molecular size and
electrical charge
G20: shear
Physical/chemical
stabilization J1 J11: freezing
39
Chapter 3
J12: cooling
* “Mechanisms” is the nomenclature used in the paper of Almeida and Bongers (2010)
All possible combinations result in more than 100 routes that could be followed for
isolation of polyphenols from tea leaves. As this number of alternatives is far from
manageable a further simplification has been proposed for the network based on the
following engineering-driven heuristics (H) and project constraints:
H1: Mechanisms C21, C22, C23 and C25 are not considered because fresh tea leaves and
components inside during processing would be damaged. Only mechanism C24 remains.
H2: C32 ultrasound mechanism for cell wall disruption has not been considered due to
effects of cavitation.
H3: C33, C34, and G20 have been excluded due to the project constraints (i.e. no harsh
conditions).
H4: C35, G15 and G16 have been excluded because of environmental reasons. This project
aims at isolating PPs from fresh tea leaves under mild conditions, and without the use of
chemicals. Furthermore, PPs will be used in the food industry and these compounds have to
be approved by obeying strict law regulations.
H5: G12 has not been considered because PPs are molecules, not particles.
H6: J12 is rejected because PPs (as a final product) have to be stored at -18 oC to avoid
contamination of the product. Therefore, only freezing is applicable.
Applying these heuristics reduce the number of possible and feasible mechanisms. The
selected mechanisms are identified and presented in Table 3.5.
40
Product-driven process synthesis for the extraction of polyphenols from fresh
tea leaves
Table 3.5: Selected tasks and mechanism
Fundamental task Mechanisms
C2
C24: cutting
C3 C32: electro-magnetic
fields- pulsed electric field
G1
G11: molecular size
G13:electrical charge
G14: solubility
G19: molecular size and
electrical charge
J1 J11: freezing
All mechanisms listed in Table 3.5 are experimentally tested. In the cell wall disruption
step, pulsed electric field (PEF) has been selected as a promising technique, because PEF is
a non-thermal and non-invasive method. Two feasible task networks can be formulated on
the basis of the appropriate mechanisms. Figures 3.3.1 and 3.3.2 depict our proposed task
networks alternatives.
41
Chapter 3
Figure 3.3.1: Proposed task network alternative 1. Codes in the scheme are given in Table 3.5
Figure 3.3.2: Proposed task network alternative 2. Codes in the scheme are given in Table 3.5
After first set of preliminary experiments and testing both alternatives, conclusion was that
alternative 1 was favorable. As a consequence of PEF exposure (mechanism C32), tea
tissue get perforated and transport of cellular material occurred. If however, after PEF
cutting is performed, there is a leakage cellular material. Therefore, alternative 1 has been
selected for further study.
PEF processing of foods involves the application of short pulses (duration of micro- to
milliseconds) with a high electric field intensity inducing cell membrane permeabilization
42
Product-driven process synthesis for the extraction of polyphenols from fresh
tea leaves
through a phenomenon called “electroporation” (Asavasanti et al., 2011)(Asavasanti, Ersus,
Ristenpart, Stroeve, & Barrett, 2010)(Asavasanti, Ersus, Ristenpart, Stroeve, & Barrett,
2010)(Asavasanti, Ersus, Ristenpart, Stroeve, & Barrett, 2010)(Asavasanti, Ersus,
Ristenpart, Stroeve, & Barrett, 2010). Term electroporation is used to describe the
phenomena that accompany the exposure of cells to transmembrane electrical pulses (Esser,
Smith, Gowrishankar, Vasilkoski, & Weaver, 2010). In Figure 3.4 electroporation process
is presented. Biological membranes are bilayers composed of phospholipids that contain
proteins inserted within the lipid matrix (Gonzalez and Barrett, 2010). Applying an external
electric field to the plant cells results in pore formation on the membrane. Pore formation is
a dynamic process depending on the intensity of the PEF treatment; electroporation can be
reversible or irreversible.
Figure 3.4: Electroporation process of the cell membrane under external electric field exposure (Yang et al.,
2008)
In this work, tea leaves samples were treated using pulsed electric field (PEF) equipment
with batch treatment configuration of the Nutri-Pulse NP110-60 System (IXL Netherlands
B.V.) which consists of a PEF treatment chamber and a high voltage generator. High
voltage generator provided rectangular pulses (see Figure 3.5b) in the range of 0.0001-0.1 s
with maximum voltage of 2.2 kV and maximum number of pulses 50. Samples were placed
in the treatment chamber between two stainless steel electrodes filled with a sterile salt
solution (Figure 3.5a).
43
Chapter 3
(a) (b)
Figure 3.5: Experimental set up (lab scale). Scheme of pulsed electric field treatment chamber (a) and PEF
pulsing protocol (b)
PEF applied on fresh tea leaves (depending on the settings) causes opening the cell
structure and transport of cellular material from the interior to the surrounding liquid
occurs. By measuring the amount of PPs in the surrounding aqueous solution after PEF
treatment, it was possible to monitor the effect of pulsed electric field on opening the cell
structure. Note that here we assume cell opening under PEF exposure not thermal Tg effect.
3.5. Mechanism and operating window
For each mechanism the operating window has to be defined. Since all experiments were
performed on lab scale, for the first step, C24: “cutting” the leaves were cut manually (1 cm
width). The most critical step in the task network is cell wall disruption (C3: pulsed electric
field) and for this step an operating window has to be defined. Three factors that are
expected to have significant impact on the extraction yield (EY) are studied: electric field
strength (E), pulse duration or pulse width (PD), and number of pulses (N). EY is defined
as following eq. (3.2):
𝐸𝑌 (%) =Amount of extracted PPs in aqueous media
Total amount of PPs 𝑥 100% eq. (3.2)
Design of experiments technique has been used to elaborate the influence of E, PD and N
on the extraction yield.
44
Product-driven process synthesis for the extraction of polyphenols from fresh
tea leaves
3.5.1. Design of experiments
Design of experiments (DoE) as a systematic tool has been applied for statistical analysis
and optimization of applied electric field strength, pulse duration and number of pulses. For
these three variables a maximum (1), minimum (-1) and center point (middle, 0) has been
defined, see Table 3.6.
Table 3.6: Minimum, center point and maximum for selected influence factors for DoE
Factors A: Electric field
strength, kV/cm B: Pulse duration, s C: Number of pulses, #
Minimum (low) -1 -1 -1
Middle 0 0 0
Maximum
(high) 1 1 1
DoE has been employed to fully characterize the selected design space with a polynomial
model, with the objective to maximize the amount of extracted polyphenols, i.e. the
extraction yield (|EY). The resulting model can be used to optimize the response and can
also be extrapolated outside boundaries of the design space see Figure 3.6. Design space
has been defined between minimum (-1), and maximum (+1) for each variable A, B and C.
Black dots in between represent middle (0). Empty dot is center point (0, 0, 0).
45
Chapter 3
Figure 3.6: Design space for three selected independent variables A, B and C minimum (-1) and
maximum (1). Black dots in between represent middle (0). Empty dot is center point (0, 0, 0)
The ranges for the three selected independent variables (factors) i.e. the independent
variables: electric field strength (E), pulse duration or pulse width (PD), and number of
pulses (N) have been selected based on literature.
3.5.2. Variables (factors)
As mentioned before three independent variables (factors) are chosen for the DoE; A:
electric field strength (E), B: pulse duration (PD), and C: number of pulses (N). Besides
selecting these three independent variables it is also necessary to define the DoE design
space, by choosing the range for each variable. For each independent variable the range is
selected based on literature:
(A) electric field strength: for food plant material such as apple, potato and carrot
applied electric field strength was from 400 to 1000 V/cm to enhance extraction of
cellular material (Ben Ammar et al., 2011). Lebovka et al. (2000) used electric
field strength in the range from 200 to 700 V/cm for PEF treatment of potato, pear
and courgette to increase degree of cell damage. In the present paper PEF
treatment has been used to open the cell structure in fresh tea leaves. The range for
electric field strength has been chosen 100 and 1100 V/cm, based on the above
mentioned literature.
(B) pulse duration: De Vito and others (2007) investigated the effect of pulse durations
(i.e.10, 100, and 1000 μs) as well as different numbers of pulses applied on the
A
B
C
46
Product-driven process synthesis for the extraction of polyphenols from fresh
tea leaves
efficiency of the PEF treatment of sugar beet and apple tissues for electric field
strengths ranging from 100 to 400 V/cm. The same authors reported that samples
exposed to the same PEF treatment time showed noticeably higher disintegration
efficiency for the larger pulse durations. Pulse duration available for this study
from 0.1*10-3
till 0.1 s is selected. Reason for this was the limitation of PEF
equipment. Note that it was not possible to run the equipment at pulse duration
below 100 μs.
(C) number of pulses: Asavasanti et al. (2010) reported that with increasing the
number of pulses from 10 to 100 damage to onion tissue increased at constant
electric field strength 333 V/cm. Above 200 pulses, PEF treatment at E = 333
V/cm caused no significant change in degree of tissue damage, so damage reaches
its maximum limit. This result is in the line with conclusions from other studies
(Lebovka et al., 2001; Lebovka et al., 2002) which demonstrated that PEF has no
effect on membrane breakdown above certain maximum limit. Depending on the
applied electric field strength, the threshold of pulse number can be different. A
slight increase in electric field strength results in a dramatic decrease in number of
pulses required to get the same degree of tissue damage. For example, increasing
E from 200 to 267 V/cm can remarkably reduce the number of pulses from 100 to
10 pulses (Asavasanti et al., 2010). This observation suggests that higher field
strengths result in lower thresholds of pulse number suggesting that there is a
saturation threshold.
Range intervals for selected independent variables are summarized and presented in Table
3.7.
Table 3.7: Range interval for selected independent variables for DoE
Analysis and model fitting have been performed using the coded design variables (A, B and
C) and not design factors with their original units. When the original units are used, the
obtained numerical results in comparison to the coded unit analysis are different and not
Factors Electric field
strength, kV/cm A
Pulse
duration, s B
Number of
pulses, # C
Minimum
(low) 0.1 -1 0.1*10-3 -1 10 -1
Middle 0.6 0 0.05 0 30 0
Maximum
(high) 1.1 1 0.1 1 50 1
47
Chapter 3
easy to interpret (Montgomery, 2013). After defining ranges for selected independent
variables, in order to evaluate DoE objective maximizing the amount of extracted PPs in
aqueous phase, extraction yield (EY) of PPs is chosen response.
The experiments are generated using the StatGraphics ® Centurion XVI software. Note that
center point experiment has been carried out in triplicate, a total of 15 experiments are
collected in Table 3.8. The center points are providing information about the reproducibility
of the process. The experiments are generated in duplicates in a randomized way.
Table 3.8: Experimental design table generated using StatGraphics software
Experimental run Independent variables (factors)
A B C
1 1 -1 0
2 0 -1 -1
3 1 0 1
4 0 -1 1
5 1 0 -1
6 (center point) 0 0 0
7 (center point) 0 0 0
8 -1 0 1
9 (center point) 0 0 0
10 1 1 0
11 0 1 -1
12 -1 0 -1
13 0 1 1
14 -1 1 0
15 -1 -1 0
Since in this point it is not known whether the independent variables have linear or
nonlinear behavior, a response surface methodology (RSM) and a Box–Behnken design are
used for modeling and optimization of the operational values of E, PD and N. The RSM can
describe linear effects as well as nonlinear (quadratic effects). Statistical tests for
48
Product-driven process synthesis for the extraction of polyphenols from fresh
tea leaves
significance and the development of quadratic relationships that link the independent
variables to the responses, can be used to optimize the settings that maximize the extracted
amount of polyphenols. The objective of the RSM is to find the best expression for the
function f, while minimizing the number of experiments. The function EY is a polynomial
series that can be represented by eq. (3.3) (Montgomery, 2013):
𝑌 = 𝑏0 + 𝑏1𝐴 + 𝑏2𝐵 + 𝑏3𝐶 + 𝑏12𝐴𝐵 + 𝑏13𝐴𝐶 + 𝑏23𝐵𝐶 + 𝑏11𝐴2 + 𝑏22𝐵2 + 𝑏33𝐶2 + 𝜀
eq .(3.3)
where Y is response (predicted EY of PPs). A, B and C stand for the electric field strength
(E), pulse duration (PD) and number of pulses (N), respectively. b0 is a constant, b1, b2, and
b3 are linear coefficients, b12, b13, and b23 are cross product coefficients, b11, b22, and b33 are
quadratic coefficients and ε represents residuals. When two different letters appear
combined (e.g. AB, BC, . . .), they represent the interaction between two independent
variables. However, when the same letter appears twice (e.g. AA, BB, . . .) it represents a
second order effect for that variable. The observed responses EY are fitted to equation (3.3)
that represents the correlation with the independent variables. The selected RSM design is
the Box–Behnken design. The design consists of triplicated center points and set of points
lying at the midpoints of each edge of the multidimensional cube that defines operational
region (minimum and maximum value of each independent variable), (see Figure 3.5). The
experimental data were processed using the StatGraphics Software.
3.5.3. Statistical analysis
Statistical analysis is performed by using StatGraphics software and Box-Behnken design is
selected to determine optimal settings for three experimental independent variables. In
DoE, analysis of variance (ANOVA) is used to test the statistical significance by comparing
variation within replicated runs with the residual (model error) variation. To test the
significance of the variability in the responses, for each of the selected effects, the ANOVA
procedure calculates the variation ratio (F-ratio). The F-ratio of F-test statistics is the ratio
of the explained-variance-per-degree-of-freedom-used to the unexplained-variance-per-
degree-of-freedom-unused. This ratio determines the significance of the effects under
investigation with respect to the variance of all the terms included in the error term, for a
chosen significance level (Montgomery, 2013; Monsanto et al., 2014). The development of
quadratic relationships that link the influence factors to the responses can be used to
optimize the operational variables, in this case to maximize the extraction yield of
polyphenols. The results are presented in Table 3.9 for a 5 % significance level.
49
Chapter 3
Table 3.9: Variance analysis results (F-ratio)
Source of variance Sum of squares F-ratio
A:electric field strength 380.33 116.7
B:pulse duration 50.552 15.52
C:number of pulses 13.599 4.180
AA 189.93 58.31
AB 0.6889 0.210
AC 40.322 12.38
BB 0.1545 0.05
BC 4.4310 1.36
CC 3.8761 1.19
Figure 3.7 presents the standardized Pareto chart. A Pareto chart allows visualization of the
statistically significant effects. In StatGraphics software statistically significant effects are
calculated using ANOVA procedure.
Figure 3.7: Standardized Pareto chart for selected response (extraction yield of PPs). Black bars represent
positive effect on the response and light color bars indicates an antagonistic effect on the response
The chart presented in Figure 3.7 displays the effects in decreasing order of significance. A
vertical line represents the border of statistical significance. In the Pareto chart the length of
each bar is proportional to the value of a t-statistic calculated for the corresponding effect.
Any bar crossing the vertical line is statistically significant at the selected significance level
50
Product-driven process synthesis for the extraction of polyphenols from fresh
tea leaves
(5 %). The black (plus) bars means that the effect has a positive effect on the response and
the light color (minus) bars represent a negative (antagonistic) effect. According to the
results collected in Figure 3.7, there are three statistically significant effects for the
extraction yield of PPs; electric field strength (A), second order effect of electric field
strength (AA) and pulse duration (B). The results also clearly show that electric field
strength has the strongest impact on response, i.e. the extraction yield.
Figure 3.8: Main effects plot for extraction yield of PPs (response) variation E for PD=0.005 s and N=30;
variation of PD for E=0.6 kV/cm and N=30; and variation N for E=0.6 kV/cm and PD=0.005 s
Figure 3.8 presents the main effects providing a clear representation of the linear and the
quadratic effect of each factor on the response. The lines indicate the estimated response
change with changing each variable from its low level to its high level, while keeping the
values of variables constant half way their lowest and highest value. Figure 3.8 clearly
shows that none of the factors has a completely linear influence on the extraction yield. All
effects have some curvature in the selected range, meaning that there are second order
effects.
For the selected response (extraction yield of PPs) a regression model has been developed
based on the influence of the independent variables. All regression coefficients have been
calculated, see eq. (3.4):
2001.005.1205.35818.062.16293.25075.059.4427.2346.18 CCBBCABAACBAEY
eq. (3.4)
51
Chapter 3
Moreover, a Parity plot for the response is used to show the fitting of the model with
experimental data. The results in Figure 3.9 show that regression model and experimental
data are in good agreement.
Figure 3.9: Parity plot for the response=extraction yield representing the results of regression model (equation
(3.4)) and the experimental data. In green rectangle, three center points are highlighted.
Figure 3.10: Residual plot representing the difference between results of experimental (measured)data and the
predicted (from regression model) data. In green rectangle, three center points are highlighted
Once a reasonable model has been fit, the residuals from the fit should be examined. In
general, a residual may be thought of as the difference between the observed value of EY
and the value predicted by the model: residual = experimental (measured) EY – predicted
(from regression model) EY. Residual plot is presented in Figure 3.10.
After generating the polynomial equation eq. (3.4) relating dependent and independent
variables, an optimization step was performed in order to maximize extraction yield (EY).
52
Product-driven process synthesis for the extraction of polyphenols from fresh
tea leaves
Optimization has been performed using concept of desirability function (Monsanto et al.,
2014). Desirability function is an established technique for the determination of the
optimum settings of input variables that can determine the optimum performance levels for
one or more responses. The desirability procedure involves two steps: (1) finding the levels
of the independent variables that simultaneously produce the most desirable predicted
responses for the dependent variables and (2) maximize the overall desirability with respect
to the controllable variables. In this particular case, there is one response (extraction yield
of PPs) to be maximized. The lower and upper limit values of the extraction yield are taken
from the Box–Behnken design levels. The optimization procedure was conducted under
these boundaries. Obtained result from numerical optimization provides optimal value of
extraction yield of 32.5 % that could be achieved for Eoptimal=1.1 kV/cm. PDoptimal=0.1*10-3
s and Noptimal=50 pulses. In Appendix I, detailed explanation how to solve optimization
problem is presented.
3.5.4. Response contour plots
The contour plots are used as the graphical representation to show interactions among three
variables (E, PD and N). Figure 3.11 presents the contour plots for each extraction yield
value obtained from the regression model. Because interactions between the variables are
taken into account in the model, the contour lines of constant extraction yield are
curved. It is desirable to operate in the region where extraction yield is between 30 and 35
%. The contour plots show that several combinations of E, PD and N could obey this
objective. In the Figure 3.11a (E versus PD and for fixed N=50) the extraction yield
increases with an increasing E and decreases with increasing of N reaching maximum of
32.5 % when E is 1.1 kV/cm and is 50 pulses. N versus E for fixed PD=0.1*10-3
s the
extraction yield increases with E and increases with N, reaching a maximum at N=50 and
E=1.1 kV/cm, see Figure 3.11b. On the other hand, at lower pulse duration, extraction yield
is higher which means that increasing pulse duration, extraction yield decreases (Figure
3.11c).
53
Chapter 3
(a) (b) (c)
Figure 3.11: Contour plot representing extraction yield of PPs versus: electric field strength (E) and pulse
duration (PD) for fixed number of pulses N=50 (a); electric field strength (E) and number of pulses for fixed
PD=0.0001s (b) and pulse duration (PD) and number of pulses (N) for fixed electric field strength 1.1 kV/cm (c).
Optimal value for EY is in the circle.
3.6. Conclusions
This chapter presents a sketch of a design methodology which extends the scope of
traditional process design to the molecular level. The extended scope introduces the use of
the product-driven process synthesis methodology for the conceptual design of polyphenols
extraction from fresh tea leaves. In the framing level and product ideas we discussed the
usefulness of polyphenols present in fresh tea leaves. A detailed specification of the input
(fresh tea leaves) and output (polyphenols) that leads to subsequent investigation of
fundamental tasks to convert raw material into the desired product is done. Two feasible
and possible tasks network routes have been proposed. Among the different mechanisms
that could be used to perform the tasks, pulsed electric field has been selected as a non-
invasive and non-thermal “mechanism” for cell wall disruption. To define an operating
window for pulsed electric field technique an experimental design has been setup and
executed (varying several settings of the pulsed electric field). From the collected
experimental data, the analysis of variance has been used to determine which variables i.e.
electric field strength, pulse duration and number of pulses as well as combinations are
significant. Box-Behnken design is used as part of the statistical analysis to find optimal
pulsed electric field settings to maximize the amount of extracted polyphenols. Within the
chosen design settings it has been found that the optimal pulsed electric field settings are:
1.1 kV/cm field strength, 0.1*10-3
s pulse duration and number of pulses 50. With obtained
optimum settings maximum value of 32.5 % of extraction yield was achieved.
54
Product-driven process synthesis for the extraction of polyphenols from fresh
tea leaves
3.7. References
Asavasanti, S., Ristenpart, W., Stroeve, P., & Barrett, D. M. (2011). Permeabilization of
plant tissues by monopolar pulsed electric fields: effect of frequency. Journal of Food
Science, 76(1), E98–111.
Bazhal, M. I., Ngadi, M. O., & Raghavan, V. G. S. (2003). Influence of Pulsed
Electroplasmolysis on the Porous Structure of Apple Tissue. Biosystems Engineering,
86(1), 51–57.
Ben Ammar, J., Lanoisellé, J.-L., Lebovka, N. I., Van Hecke, E., & Vorobiev, E. (2011).
Impact of a pulsed electric field on damage of plant tissues: effects of cell size and
tissue electrical conductivity. Journal of Food Science, 76(1), E90–7.
Bhatla, S. C., Kaushik, V., & Yadav, M. K. (2010). Use of oil bodies and oleosins in
recombinant protein production and other biotechnological applications.
Biotechnology Advances, 28(3), 293–300.
Bongers, P. M. M., & Almeida-Rivera, C. (2009). 19th European Symposium on Computer
Aided Process Engineering. Computer Aided Chemical Engineering (Vol. 26, pp.
231–236).
Bongers, P. M. M., & Almeida-Rivera, C. (2012). 11th International Symposium on
Process Systems Engineering. Computer Aided Chemical Engineering (Vol. 31, pp.
195–199).
Bouzrara, H., & Vorobiev, E. (2003). Solid–liquid expression of cellular materials
enhanced by pulsed electric field. Chemical Engineering and Processing: Process
Intensification, 42(4), 249–257.
Canatella, P. J., Karr, J. F., Petros, J. A., & Prausnitz, M. R. (2001). Quantitative study of
electroporation-mediated molecular uptake and cell viability. Biophysical Journal,
80(2), 755–64.
De Vito, F., Ferrari, G., I. Lebovka, N., V. Shynkaryk, N., & Vorobiev, E. (2007). Pulse
Duration and Efficiency of Soft Cellular Tissue Disintegration by Pulsed Electric
Fields. Food and Bioprocess Technology, 1(4), 307–313. doi:10.1007/s11947-007-
0017-y
Evrendilek, G. A., Altuntas, J., Sangun, M. K., & Zhang, H. Q. (2013). Apricot Nectar
Processing by Pulsed Electric Fields. International Journal of Food Properties,
16(1), 216–227.
55
Chapter 3
Gonzalez, M. E., & Barrett, D. M. (2010). Thermal, high pressure, and electric field
processing effects on plant cell membrane integrity and relevance to fruit and
vegetable quality. Journal of Food Science, 75(7), R121–30.
Haberl, S., Miklavcic, D., Sersa, G., Frey, W., & Rubinsky, B. (2013). Cell membrane
electroporation-Part 2: the applications. IEEE Electrical Insulation Magazine, 29(1),
29–37.
Lamsal, B. P., & Johnson, L. A. (2007). Separating Oil from Aqueous Extraction Fractions
of Soybean. Journal of the American Oil Chemists’ Society, 84(8), 785–792.
Lebovka, N. (2009). Electrotechnologies for Extraction from Food Plants and
Biomaterials. New York, NY: Springer New York. doi:10.1007/978-0-387-79374-0
Lebovka, N. ., Bazhal, M. ., & Vorobiev, E. (2001). Pulsed electric field breakage of
cellular tissues: visualisation of percolative properties. Innovative Food Science &
Emerging Technologies, 2(2), 113–125.
Lebovka, N. ., Bazhal, M. ., & Vorobiev, E. (2002). Estimation of characteristic damage
time of food materials in pulsed-electric fields. Journal of Food Engineering, 54(4),
337–346.
Lebovka, N. I., Bazhal, M. I., & Vorobiev, E. (2000). Simulation and experimental
investigation of food material breakage using pulsed electric field treatment. Journal
of Food Engineering, 44(4), 213–223.
Lebovka, N. I., Shynkaryk, M. V., El-Belghiti, K., Benjelloun, H., & Vorobiev, E. (2007).
Plasmolysis of sugarbeet: Pulsed electric fields and thermal treatment. Journal of
Food Engineering, 80(2), 639–644.
Lee, K. J., & Lee, S. H. (2008). Extraction behavior of caffeine and EGCG from green and
black tea. Biotechnology and Bioprocess Engineering, 13(5), 646–649.
Leong, S. Y., Richter, L.-K., Knorr, D., & Oey, I. (2014). Feasibility of using pulsed
electric field processing to inactivate enzymes and reduce the cutting force of carrot
(Daucus carota var. Nantes). Innovative Food Science & Emerging Technologies.
Leung, A., & Foster, S. (1980). Encyclopedia of common natural ingredients. (A. Y.
Leung, Ed.)Encyclopedia of Common Natural Ingredients used in Food Drugs and
Cosmetics (pp. 142–143). John Wiley & Sons,.
Li, Z., Huang, D., Tang, Z., & Deng, C. (2010). Microwave-assisted extraction followed by
CE for determination of catechin and epicatechin in green tea. Journal of Separation
Science, 33(8), 1079–84.
56
Product-driven process synthesis for the extraction of polyphenols from fresh
tea leaves
Moggridge, G. D., & Cussler, E. L. (2000). An Introduction to Chemical Product Design.
Chemical Engineering Research and Design, 78(1), 5–11.
Monfort, S., Gayán, E., Saldaña, G., Puértolas, E., Condón, S., Raso, J., & Álvarez, I.
(2010). Inactivation of Salmonella Typhimurium and Staphylococcus aureus by
pulsed electric fields in liquid whole egg. Innovative Food Science & Emerging
Technologies, 11(2), 306–313.
Monsanto, M., Trifunovic, O., Bongers, P., Meuldijk, J., & Zondervan, E. (2014). Black tea
cream effect on polyphenols optimization using statistical analysis. Computers &
Chemical Engineering, 66, 12–21.
Montgomery, D. (2013). Design and Analysis of Experiments (8th ed.). New York: wiley.
Perva-Uzunalić, A., Škerget, M., Knez, Ž., Weinreich, B., Otto, F., & Grüner, S. (2006).
Extraction of active ingredients from green tea (Camellia sinensis): Extraction
efficiency of major catechins and caffeine. Food Chemistry, 96(4), 597–605.
Ruan, J., Haerdter, R., & Gerendás, J. (2010). Impact of nitrogen supply on carbon/nitrogen
allocation: a case study on amino acids and catechins in green tea [Camellia sinensis
(L.) O. Kuntze] plants. Plant Biology (Stuttgart, Germany), 12(5), 724–34.
Sale, A., & Hamilton, W. (1967). Effects of high electric fields on microorganismsI. Killing
of bacteria and yeasts. Biochimica et Biophysica Acta (BBA) - General Subjects,
148(3), 781–788.
Salunkhe, D.K., Chavan, J.K., Adsule, R.N. & Kadam, S. S. (n.d.). World Oilseeds.
Chemistry, Technology and Utilization. New York: Van Nostrum Reinhold.
Shirsath, S. R., Sonawane, S. H., & Gogate, P. R. (2012). Intensification of extraction of
natural products using ultrasonic irradiations—A review of current status. Chemical
Engineering and Processing: Process Intensification, 53, 10–23.
Soliva-Fortuny, R., Balasa, A., Knorr, D., & Martín-Belloso, O. (2009). Effects of pulsed
electric fields on bioactive compounds in foods: a review. Trends in Food Science &
Technology, 20(11-12), 544–556.
Stadnik, J., & Dolatowski, Z. J. (2011). Influence of sonication on Warner-Bratzler shear
force, colour and myoglobin of beef (m. semimembranosus). European Food
Research and Technology, 233(4), 553–559.
Toepfl, S., Heinz, V., & Knorr, D. (2007). High intensity pulsed electric fields applied for
food preservation. Chemical Engineering and Processing: Process Intensification,
46(6), 537–546.
57
Chapter 3
Wan, J., Coventry, J., Swiergon, P., Sanguansri, P., & Versteeg, C. (2009). Advances in
innovative processing technologies for microbial inactivation and enhancement of
food safety – pulsed electric field and low-temperature plasma. Trends in Food
Science & Technology, 20(9), 414–424.
Wang, H., Provan, G. J., & Helliwell, K. (2003). HPLC determination of catechins in tea
leaves and tea extracts using relative response factors. Food Chemistry, 81(2), 307–
312.
59
Chapter 4
Chapter 4
Isolation of oil bodies from soybeans in a mild way:
definition of operating window for process design
ABSTRACT
In this work, experiments were performed in order to define operating windows for process
parameters for isolation of oil bodies from soybeans in a mild way. Aqueous extraction
process for simultaneous separation of oil bodies and proteins from soybean was tested. In
the first part, the effect of the particle size on the extraction of oil bodies with two different
grinding methods, to obtain one coarse flour (d90 300 µm) and one fine flour (d90 40 µm)
was studied. The extractability of the coarse flour was better compare to fine flour: oil
recoveries from the cream were very similar (23 % and 24.5 % of the total soybean oil),
and the protein extraction yield was higher for the coarse flour (48 % against 40 % of the
total protein). Second, to enhance extraction yield of protein and oil, three different
pretreatments were applied to aqueous extraction process. The pretreatments included
enzymatic hydrolysis, ultrasound and the combination of the two. We found that
pretreatment with ultrasound reduced the remaining insoluble fraction and increased the
amount of solids extracted into the aqueous phase. The combination of ultrasound and
enzymes resulted in the cream with the highest lipid-to-protein ratio of 10:1. Different
aqueous extraction process alternatives were compared with benchmark process (neither
enzymes nor ultrasound).
60
Isolation of oil bodies from soybeans in a mild way: definition of operating
window for process design
4.1. Introduction
Over the last 15 years, a growing interest in developing mild extraction processes for plant
materials has developed. The extraction of intact cellular components is a promising
method for obtaining high-added value products, with a reduced environmental impact.
Soybeans are an important crop worldwide, with a high nutritional value. However, this
value is not efficiently maintained by the current processing conditions. The goal is to
develop an extraction process able to efficiently retrieve intact oil bodies (OBs), native soy
proteins and fibers. Therefore, the need for effective extraction of biologically active
components from plants, without any loss of functionality and high purity, has resulted in
development of novel extraction processes (Shirsath, Sonawane, and Gogate, 2012). The
aqueous extraction process (AEP) is originally suggested as an alternative for the solvent
oil extraction process (Rosenthal et al. 1996). In AEP, water is used as an extracting
medium to remove oil as an emulsion or free oil, unlike organic solvents, which dissolve
the oil (Campbell and Glatz 2009). The AEP of oil can be improved by any treatment that
enhances the dissolution of these other water soluble components (mainly proteins), for
instance, by using enzymes or increasing the temperature (Rosenthal et al. 1996).
Intact OBs can be considered as a natural emulsion that, in situ, protects the lipids from
oxidation during storage (Kerry Alan Campbell, 2010). Moreover, OBs have the advantage
over solvent extracted oil that they required neither emulsifiers nor homogenization during
processing (Kapchie et al., 2011). The high stability of the OBs makes them suitable for
e.g. food, cosmetic, and pharmaceutical applications. OBs may also be interesting for
application in biobased micro capsules and delivery of functional components (Iwanaga et
al., 2007). Harsh process conditions in the current soy process destroy the OBs native
structure. In the present work, the first objective is to study the effect of the particle size of
the soybean flour on the aqueous extraction of OBs by applying mild conditions. Mild
conditions are defined as follows: only use food grade solvents, no extreme pH values (no
strong acidic or alkaline conditions), and mild temperatures (< 40oC). A simple AEP is
performed with two different particle size soy flours (coarse and fine). The yield of oil and
protein extracted from the flour is calculated and the stability of the cream is measured to
determine the integrity of the OBs (diameter is 0.2-0.5 µm). The second objective is to
study the effect of simultaneous enzymatic hydrolysis and ultrasonication of the soybean
flour on the performance of aqueous extraction of OBs. The soy flour is pretreated with a
commercial enzyme mixture, containing different types of cell-wall degrading enzymes,
and with ultrasonication during a specific period. These pretreatments are applied to
investigate whether the mass transfer of the cellular components increased, comparing to a
benchmark AEP. Therefore, based on the experimental results, operating window for
61
Chapter 4
process parameters (i.e. pH, temperature, enzyme, particle size etc.) will be defined and
later used for conceptual design of the process.
4.2. Materials and methods
4.2.1. Preparation of soybean flour
Flour A and B were made from the same soybeans. To prepare soy flour A soybeans were
milled on a Polymix© mill (Kinematica) at 2500 rpm with a screen of 2.0 mm. The full-fat
soybean flour (coarse) was classified using a Vibratory Sieve Shaker AS 200 digit (Retsch)
at a frequency of 70 kHz, with 1.0 mm, 500 µm, 250 µm and 125 µm sieves. The fraction
between 125-250 µm was used for the extraction process. The flour was stored in sealed
aluminum bags at 4 °C until it was used. Soy flour B was obtained by cryogenic milling on
a pilot-plant scale mill (fine flour). Cryogenically milled soy flour was made by placing the
beans in liquid nitrogen and grinding them using a Contraplex CW mill. The particle size
distribution of the different flours (A and B), measured with a Malvern Mastersizer
analyzer 2000, is shown in Figure 4.1. A double distribution was obtained for soy flour A
(fraction between 125 to 250 µm); 90 % of the particles had a particle size lower than 300
µm and around 65 % had a particle size lower than 100 µm. in Appendix D particle size
measurements are presented. Because some oil being extracted from the disrupted cells
(especially since the oil is in liquid state). Free oil caused particles to stick to each other,
and made the separation of the smallest particles (< 125 µm) very difficult. Flour B
presented a normal (Gaussian) distribution; 90 % of the particles had a particle size lower
than 40 µm. Taking into account that cotyledon cells are about 15-20 µm in diameter and
70-80 µm in length (Rosenthal et al. 1996), this results a high proportion of rupturing the
cells. It is important to note that the particle size distribution was measured on the wet
flour; therefore, the result reflects the size of the hydrated particles. This is the reason for
obtaining a higher particle size in flour A than the expected according to the used sieve
(250 µm).
62
Isolation of oil bodies from soybeans in a mild way: definition of operating
window for process design
Figure 4.1. Particle size distribution of the soybean flours used for the extraction process of oil bodies (A coarse
flour; B flour cryogenically milled fine flour)
4.2.2. Aqueous extraction of soybean oil bodies
OBs were physically isolated from soybean full-fat flour using an water-based flotation
centrifugation method. Soy flour was hydrated for 30 min at 4 °C in a sodium phosphate
buffer (0.1 M, pH=7.2), in a ratio of 1:6 (w/v). The slurry was mixed vigorously using a
high-speed vortex mixer for 1 min, and centrifuged (Sigma 6-16K) at 4700 rpm at 4 °C for
30 min. By the end of the centrifugation a cream fraction (OBs) on the top, a supernatant or
skim, and a residue on the bottom were obtained. The newly formed cream fraction
obtained after the aqueous-extraction was re-suspended with TRIS buffer (0.1 M, pH=8.6)
in a 1:2 solid-liquid weight ratio, and mixed vigorously to remove the storage (non-
oleosins) proteins from the cream. The mixture was centrifuged (4700 rpm at 4 °C for 15
min) to obtain a cream pad on top, a supernatant and a creamy residue at the bottom. The
cream was collected with a spatula; the supernatant was separated and stored at 4 °C for
63
Chapter 4
analysis, and the creamy residue was combined back again with the cream pad. This
“purified” cream was then re-suspended in TRIS buffer (0.1M, pH=7.2) in a 1:2 solid to
liquid weight ratio, before storage at 4 °C. All experiments were performed in duplicate.
4.2.3. Enzyme/ultrasound-assisted aqueous extraction
Detailed description of enzyme assisted and ultrasound extraction is presented in Appendix
A. Soybean flour was hydrated for 30 min at ambient temperature with a sodium phosphate
buffer (0.1 M, pH= 7.2, NaCl concentration 0.25 M), in a 1:6 solid to liquid weight ratio.
Different enzyme mixtures are tested and results are presented in Appendix B. The slurry
was mixed with enzyme solution (5 v/w%) (Ultrazyme AFP L, Novozymes) and placed in
an ultrasound bath (Elma TI-H-20) at 25 kHz (100 % power) at 40 °C for 3h (samples with
no ultrasound pretreatment were incubated in a Medline BS-21 water bath at 40 °C and 150
rpm). After the incubation period, the slurry was mixed vigorously with high speed mixer
for 1 min and centrifuge at 4700 rpm at 4 °C for 1 h. The cream phase was collected with a
spatula. The supernatant and residue fractions were mixed and centrifuged two times more
at 4700 rpm and 4 °C for 30 min. After the centrifugation cycles, all the cream phase was
washed with TRIS buffer (0.1 M, pH=8.6, NaCl concentration 0.25 M) in a 1:2 solid-liquid
weight ratio. This cream slurry was centrifuged again (4700 rpm at 4 °C for 15 min). The
supernatant was separated and re-centrifuged one more time. In this study, the top cream
and the bottom residue were kept separately, and stored at -18°C for further analysis. All
experiments were performed in duplicate.
4.2.4. Recoveries
The mass balances of oil and protein were determined for all procedures. Recoveries were
calculated as follows: protein and oil recovered from cream, supernatants and residues were
calculated as the percentage of total protein or oil present in the unprocessed soybean flour
(starting material). The protein content of the soy flour, residue, supernatant and cream
fraction was calculated by converting the nitrogen (N) content in the samples, using a
multiplication factor of 6.25. The N was determined by the Dumas method using the
Elementar Vario Max CNS Analyser. Glutamic acid was used as standard (N 9.52 %, C
40.72 %) and butter milk as a control (blank) sample. In Appendix C protein and lipid
recovery results from different aqueous extraction procedure are reported.
64
Isolation of oil bodies from soybeans in a mild way: definition of operating
window for process design
The oil content in the soy flour was determined by extraction with hexane using the Soxtec
System HT6 and 1043 Extraction Unit, according to the manufacture’s manual. Samples (5
g) were extracted with 50 ml of hexane (Program: 20 min under boiling, 40 min rinsing).
Solvent was dried under vacuum (~ 130 mbar, 40 °C), and the remaining fat weighted. The
oil content in the different fractions (cream, supernatant and residue) was determined using
a CEM SMART Trac System. Samples (1-2 g) were dried by microwave and the oil content
determined by Nuclear Magnetic Resonance, according to the manufacture’s manual. The
used method involved drying the sample at 110 °C until constant weight, before measuring
the oil content with NMR.
4.3.Results and discussion
4.3.1. Effect of particle size on the aqueous extraction of soybean oil
bodies
Two different full-fat soybean flours coarse (A-coarse and B-fine) with different levels of
grinding are used as starting material for the OBs extraction process. According to
Rosenthal et al. (1996) the critical step in aqueous extraction process is the grinding
operation, which determined the seed particle size. Efficient grinding which breaks down
the walls of cotyledon cells is essential to extract the cellular content. These two flours (A
and B) are subjected to the aqueous extraction process explained in the Materials &
Methods section. Table 4.1 shows the mass balance of the extraction process is presented.
Table 4.1. Mass balance of the aqueous extraction process of oil bodies from two different soy flours
Soy flour Amount (% w/w dry-weight basis)
Cream
Skim
Residue
A (coarse) 9.6 38.4 47.9
B (fine) 14.7 30.8 52.3
Assuming an ideal separation of the cellular components after extraction, the cream is
expected to consist of only OBs (20 % of the soybean), the skim contains the proteins and
65
Chapter 4
water soluble carbohydrates (~ 65 %), and the residue consists of the soybean fibers (~ 15
%). A bigger amount of cream is obtained from the soy flour B, while less solids are
present in the skim fraction. The higher amount of cream obtained from soy flour B could
be explained by the smaller particle size of the flour and a better extraction of cellular
contents. Although, a high amount of cream does not necessarily mean a higher amount of
OBs. Larger molecules, e.g. storage proteins are part of the emulsion and at the same time
they increase the cream yield. Demonstrated by Table 1 the simple extraction is not
efficient: there are still many soybean solids (~ 50 %) in the residue fraction. These solids
are still inside the cells, but according to the particle size of the flours (especially that of
soy flour B) most of the cells should be broken. On the other hand, extraction conditions
(mixing, temperature, osmotic pressure) are not favoring the transport of cell components to
the extraction medium. Temperature enhances the dissolution of water-soluble components
(e.g. proteins), thus, improving the extraction performance of the OBs from the cell matrix.
Therefore, maximum oil recoveries have occurred at temperatures where soy proteins
remain soluble (not denatured), normally between 40-60 °C (Lamsal & Johnson, 2007). To
keep mild conditions, no heating involved during the extraction process. The protein and oil
composition of the different fractions, as well as that of the starting material, is shown in
Table 4.2.
Table 4.2. Mass balance of the aqueous extraction process of oil bodies from two different soy flours
Treatment
Composition (% w/w dry-weight basis)
Soy flour Cream Skim Residue
Protein Oil Protein Oil Protein Oil Protein Oil
A (coarse) 40.4 21.6 25.4 68.9 47.2 N.D 28.3 18.3
B (fine) 42.4 21.1 33.0 42.2 48.4 N.D 29.8 15.9
The composition of the flours was within the range expected from literature (Salunkhe, et
al. 1992); differences were result of the different used soybean variety. Even though less
cream was obtained from the flour A compare to flour B, cream from flour A contained
almost 70 % of total oil content and 25 % of total proteins content on a dry-weight basis.
On the other hand, the cream from flour B had less lipids (42 %) but more proteins (33 %),
and sum of these two components together comprise 75 % of the total solids in this cream.
The ratio lipid-to-protein in the cream fraction A and B is 2.8 and 1.3, respectively.
According to Campbell and authors Campbell et al. (2010), oil-to-protein ratio in purified
66
Isolation of oil bodies from soybeans in a mild way: definition of operating
window for process design
OBs of 0.5 µm in diameter, containing an oleosin layer of 3.2 mg/m2,would be 20:1.
Therefore, more “impurities” were present in the cream from flour B. When transferred into
water, the OBs are accompanied by either organelles and/or water-soluble compounds,
including soluble carbohydrates and protein bodies that reduce their purity. The extracted
proteins interact with the surfaces of the OBs and form a secondary layer that impacts the
stability of the OBs (Nikiforidis and Kiosseoglou, 2009). Both creams obtained after the
AEP still contained other storage soy proteins as part of the emulsion.
Both flours (A and B) in the skim contained around 48 % of proteins and no oil were
detected. Separation from the cream phase was difficult and some residual cream was
observed in the skim phase. The detection limit of the used analysis method was not
significant to demonstrate the presence of highly diluted oil in the water phase. No
significant differences were found between the remaining residues of the two soy flours.
Both residues contained high amounts of “lost” lipids and proteins.
Figure 4.2. Protein and oil recovery from the soybean flours in each fraction after aqueous extraction process (A-
course flour, B-fine flour). Blue bars represent protein content, red bars represent oil content.
From the total amount of lipids present in the soybean, 24.5 % were retrieved in the
“purified” cream fraction from soy flour A, and 23 % from soy flour B. In both cases, most
67
Chapter 4
of the oil remained trapped in the residue fraction. The cream fraction from flour B
contained twice as much protein as the cream obtained from flour A. This confirms that the
higher amount of cream obtained for soy flour B was caused by more proteins present in
the emulsion, rather than more oil or OBs as compared with flour A. the extraction of
proteins to the skim fraction however was better for the soy flour A than the flour B.
according to the results collected in the Figure 4.2, 48 % and 40 % of the total proteins
present in the starting flour were retrieved in the skim fraction for soy flour A and soy flour
B, respectively. The residue form soy flour A remained with 44 % of oil and 36 % of the
soy proteins. The residue remaining from the soy flour B contained 43 % of the total oil and
42 % of the total proteins. Overall, from Figure 4.2 it can be observed that the extraction
yield of oil and specially proteins from the soy flour A was slightly higher than extraction
yield achieved with soy flour B. (Rosenthal and co-authors (1998) studied the effect of
particle size of flour on oil and protein aqueous extraction yield. They extracted flour with
mean particle sizes from 800 µm down to 150 µm and demonstrated that oil and protein
extraction yields were directly proportional to the inverse of flour particle size. This result
was attributed to cellular disruption enabling oil and protein release. However, according to
Campbell and Glatz (2009) the mechanism of oil mobility and release is also determined
by other factors, such as the matrix structure. The matrix structure is determined by the
native cellular geometry. The mode of cellular disruption used (e.g. kind of milling), and
the water solubility of the materials in the intercellular space (Campbell and Glatz 2009).
The fact that soy flour B was stored for a long period of time (2 years) had a negative effect
the cellular structure due to the aging. This could be responsible for the lower extractability
of the soy flour B, despite its smaller particle size. Looking at the total values for the AEP,
the lipid recovery was quite low. Between 66 and 68 % of the total lipids were recovered in
the different AEP fractions. The rest most probably remained in the skim fraction, because
the applied centrifugation speed and time were not sufficient to transfer those lipids to the
cream pad layer.
The particle size and micro structure of the cream fractions, obtained from the AEP, was
determined to demonstrate the presence of intact OBs. To obtain more information about
the microstructure of the creams, Confocal Scanning Laser Microscopy (CSLM) was used
to analyze the creams (Figure 4.3).
68
Isolation of oil bodies from soybeans in a mild way: definition of operating
window for process design
Figure 4.3. Confocal sections of the “purified” cream fraction extracted from soy flour B and stained with Nile
blue (bars=10 µm). Small individual green particles are intact oil bodies. Larger red particles are protein entities.
CSLM allows materials to be observed from different depths. The pictures in the Figure 4.3
show that particles are in the range of 10 µm, as expected from the particle size distribution
(PSD) measurements (data are not shown). However, it is possible to differentiate small
individual green particles (smaller than 1 µm). Explanation of aggregation of particles is the
presence of storage proteins (stained in red). Larger proteins have affinity (e.g. electrostatic
interactions) with the oleosin present in the OBs surface, making those aggregates.
Microscopic studies demonstrated that OBs have an apparent affinity to cell wall, bigger
protein and endoplasmic reticulum Campbell (2010). Also, it is important to consider that,
although the pH of the storage buffer was set at 7.2, where oleosins are expected to have a
negative charge, the pH was not controlled during the process. Note that OBs tend to
aggregate already at pH 6.8 Tzen (1992). Overall, the presence of small individual green
particles so close to each other provides an indication that the OBs were still intact. Free oil
droplets in a aqueous solution would tend to coalesce and form larger droplets. Light
microscopy image of one sample of soy flour A is presented in Appendix E.
OBs fill up the space between protein bodies in the cells, and are enclosed in a matrix of
cytoplasmic proteins (Campbell and Glatz, 2009). Therefore, OB-protein interactions play
an important role in the mechanism of OB release during the AEP. Conditions that favor
protein extraction i.e. temperature below the temperature of denaturation, pH different from
the isoelectric point, use of several extraction steps) generally favor OB mobility and
transfer. When it comes to the effect of the particle size of the flour on the extraction of
OBs and proteins, the obtained result was unexpected. The extraction of oil and proteins
from the soybean was a bit higher for the flour with the larger particle size. However, there
is a second factor that influenced obtained results. As already mentioned, the long storage
time of the cryogenically milled flour (~2 years old) could be responsible for its lower
extractability. Many physical and chemical changes may have occurred (i.e. denaturation)
during the storage period that could affect the solubility of the proteins and subsequently,
the extraction yield of the OBs. Above all, cryogenic milling demands high investment and
operational costs for the AEP, especially the infrastructure necessary to use and recycle
69
Chapter 4
liquid nitrogen (Wilczek, Bertling, & Hintemann, 2004). The utilization of less fine soy
flour and some process aids (i.e. ultrasonication of an aqueous suspension) to improve the
cell wall disruption and transfer of OBs might be a better alternative to improve the
extraction yields.
4.3.2. Enzymatic hydrolysis and ultrasonication of the coarse soy flour
on the aqueous extraction of OBs
It has been reported that the low extraction yields of aqueous processes can be improved by
e.g. using enzymes that hydrolyze the structural polysaccharides forming the cell wall of
oilseeds, or by application of ultrasonication that can increase the mass transfer rates and as
a consequence decrease the extraction times (Rosenthal et al. 2001; Shirsath et al. 2012). A
new AEP process, which included simultaneous enzymatic hydrolysis and ultrasound, was
applied to the coarse flour (earlier flour B). The particle size fraction between 125 and 250
µm was used for the extraction of OBs.
The new AEP involved the appropriate process temperature, the extraction buffers and the
soy flour B pretreatment. First, the process temperature was increased from room
temperature to 40 °C in this new AEP. Generally, higher temperatures are associated with
enhanced extraction. The mass transfer rate is favored by increased solute solubility and
diffusion into the bulk solvent. However, degradation of thermolabile components (i.e.
proteins) should be considered when working at high temperatures (Karki et al., 2012). For
AEP, (Domínguez, Núñez, & Lema, 1994) obtained a maximum oil yield in the extraction
of soybean at temperatures between 40-60 °C, while (Rosenthal et al. (2001) reported a
slight decrease in oil yield for temperatures above 50°C, which they attributed to protein
denaturation. Since the goal was to design a mild aqueous process to extract OBs from
soybean, and it is known that the OBs extraction yield is directly related to the protein
extractability, and extraction temperature of 40 °C was chosen.
Secondly, to improve the separation of exogenous proteins from the OBs cream fraction,
the ionic strength of the extraction buffers was increased to 250 mM NaCl. Interactions
between OBs surface proteins (oleosins) and exogenous proteins may include e.g.
electrostatic repulsions and van der Waals attraction forces, hydration effects, and hydrogen
bonding. When salt is used at low concentration, salt ions provide charge shielding or ion
binding on the charged proteins (Tsumoto, Ejima, Senczuk, Kita, & Arakawa, 2007). Salt
stabilizes proteins against dissociation and heat denaturation. Liu and Tang (2013) found
that increasing salt concentrations from 0.05 to 2.0 M affected the denaturation
70
Isolation of oil bodies from soybeans in a mild way: definition of operating
window for process design
temperature of the soy protein β-conglycinin to increase from 77 to 100 °C, and for the
protein glycinin to increase from 92 to 113 °C at pH=7.0.
Finally, two different pretreatments were done to the soy flour to improve the extraction of
OBs. The application of cell-wall degrading enzymes was done to break the cotyledon cell
and to make the structure more permeable (Rosenthal et al. 1996). The use of
cellulases/pectinases can potentially increase extraction yield and also allows the
simultaneous extraction of undenatured proteins (Kapchie et al., 2011). Moreover,
ultrasound was applied to increase the transport of elements through cellular membranes,
and extract cellular structures from cells damaged by cavitation (Vilkhu, Mawson, Simons,
& Bates, 2008). In terms of frequency, low frequencies (20-100 kHz) are recommended for
dominant physical effects of cavitation, so as to intensify the mass transfer rates (Shirsath et
al., 2012). The combination of ultrasound and enzyme-assisted extraction is a green
alternative that has shown a synergistic effect by increasing enzyme activity, decreasing the
processing time, thus improving the extraction performance in yield and time (Stadnik and
Dolatowski 2011; Easson et al. 2011). The inaccessibility of the enzymes to their substrate
has been a problem when trying to degrade soybean cell walls (Huisman et al. 1999;
Ouhida et al. 2002). Ultrasound cavitation can enhance enzyme efficiency by improving the
dispersion of the enzymes and opening-up the structure, thus facilitating transport of the
enzyme molecules to the substrate surface (Stadnik and Dolatowski 2011). For this specific
study, the commercial enzyme Ultrazyme AFL P was used simultaneously with
ultrasonication of 25kHz. This enzyme ingredient was chosen after considering its
cellulose, pectinase and xyloglucanase activities (data presented in Appendix E).
Table 4.3 shows demonstrate the mass balances obtained for the different experiments.
Table 4.3. Mass balance of the three main fractions obtained after the new AEP of OBs from soy flour
Treatment Mass balance** (% dry-weight basis)
Cream Skim Residue
Benchmark 17.0 53.0 29.9
E 15.1 57.2 27.4
U 14.3 58.7 27.0
EU 13.5 60.5 26.1
*Benchmark no pretreatment before extraction; E enzymatic hydrolysis (5 % v/w enzymes); U ultrasonication (25
kHz); EU simultaneous enzymatic hydrolysis with ultrasonication.
The results in Table 4.3 demonstrate that the application of ultrasound (with and without
enzyme addition) favored the transport of solids to the aqueous skim fraction. The highest
71
Chapter 4
amount was obtained for the EU experiment with 60.5 % of the solids in the skim, being
not significantly different from that the U experiment. The application of a pretreatment,
irrespective of which one, significantly decreased the amount of residual solids after the
extraction. Cellulolytic enzymes are breaking down large polysaccharides present in the
cell wall and releasing smaller saccharides which may become soluble; on the other hand,
ultrasound can cause the disruption of the cells and decrease the particle size (Stadnik and
Dolatowski, 2011). The combination of enzymatic hydrolysis and ultrasound, therefore,
resulted in the highest decrease of insoluble residue (26.1 % residue).
From the compositional balances (Table 4.4) results show that the lipid-to-protein ratio
increased from 5:1 for the experiment with enzyme pretreatment and to 10:1 for the
enzyme-ultrasound experiment, meaning that the pretreatment (especially ultrasound)
improved the “purity” of the cream. According to OBs structure lipid-to-protein ratio for
intact and “pure” OBs is 20:1, indicating that pretreatment with enzymes combined with
ultrasound gives better results for the lipid-protein separation than pretreatment only with
enzymes.
Table 4.4. Protein and lipid composition of the main aqueous extraction fractions obtained from soy flour
Treatment
Composition (% dry-weight basis)
Cream Skim Residue
Protein Lipids Protein Lipids* Protein Lipids
Benchmark 12.8 75.8 49.4
<10%
23.4 9.5
E 13.2 74.9 45.0 23.5 10.5
U 10.4 81.4 44.9 23.7 7.3
EU 8.6 84.4 43.7 25.0 6.8
* In the skim fraction, no lipids were detected. During the extraction process in the skim fraction, two phases were
observed: the residual (solid) and the supernatant (aqueous) phase. In the supernatant phase of the skim fraction,
the concentration of lipids was less than 1 % on wet basis.
4.4. Conclusion
Aqueous extraction of intact OBs was possible under mild conditions. The results of the
investigation of the particle size effect demonstrated that fine milling, like cryogenic
milling, favored the cream yield, without improving the lipid extraction yield or the purity
of the obtained OBs cream. On the other hand, the soy flour with the larger particle size
resulted in a higher protein extraction yield in the skim fraction. The application of
72
Isolation of oil bodies from soybeans in a mild way: definition of operating
window for process design
ultrasound did enhance the purity of the final recovered cream fraction, with an oil-to-
protein ratio up to 10:1. The highest amount obtained for the simultaneous enzymatic
hydrolysis with ultrasonication experiment with 60.5 % of the solids in the skim. The
combination of enzymatic hydrolysis and ultrasound, therefore, resulted in the highest
decrease of insoluble residue, improved the purity of the cream, and reduced the processing
time. Finally, operating windows for different extraction parameters (extraction
temperature, pH, particle size, enzyme concentration and ultrasound frequency) were
defined and they formed basis for further conceptual process design.
4.5. References
Bair, C. W., Snyder, H. E., & Technology, F. (1980). , Electron Microscopy of Soybean-
Lipid Bodies lOO [] Lipid ] Protein, (September), 279–282.
Bhatla, S. C., Kaushik, V., & Yadav, M. K. (2010). Use of oil bodies and oleosins in
recombinant protein production and other biotechnological applications.
Biotechnology Advances, 28(3), 293–300.
Campbell, K. a, & Glatz, C. E. (2010). Protein recovery from enzyme-assisted aqueous
extraction of soybean. Biotechnology Progress, 26(2), 488–95.
Campbell, K. a., & Glatz, C. E. (2009). Mechanisms of aqueous extraction of soybean oil.
Journal of Agricultural and Food Chemistry, 57(22), 10904–10912.
Cells, T., Hsieh, K., & Huang, A. H. C. (2004). Endoplasmic Reticulum , Oleosins , and
Oils in Seeds, 136(November), 3427–3434.
Deckers, H.M., van Rooijen, G., Boothe, J., Goll, J. and Moloney, M. M. (2003). Product
for Topical Applications Comprising Oil Bodies. United States.
Domínguez, H., Núñez, M. J., & Lema, J. M. (1994). Enzymatic pretreatment to enhance
oil extraction from fruits and oilseeds: a review. Food Chemistry, 49(3), 271–286.
Easson, M. W., Condon, B., Dien, B. S., Iten, L., Slopek, R., Yoshioka-Tarver, M., …
Smith, J. (2011). The application of ultrasound in the enzymatic hydrolysis of
switchgrass. Applied Biochemistry and Biotechnology, 165(5-6), 1322–31.
Huisman, M. M. H., Schols, H. A., & Voragen, A. G. J. (1999). Enzymatic degradation of
cell wall polysaccharides from soybean meal. Carbohydrate Polymers, 38(4), 299–
307.
73
Chapter 4
Iwanaga, D., Gray, D., & Fisk, I. (2007). Extraction and characterization of oil bodies from
soy beans: a natural source of pre-emulsified soybean oil. Journal of Agricultural ,
8711–8716.
Jung, S. (2009). Aqueous Extraction of Oil and Protein From Soybean and Lupin: a
Comparative Study. Journal of Food Processing and Preservation, 33(4), 547–559.
Kapchie, V. N., Towa, L. T., Hauck, C. C., & Murphy, P. a. (2011). Recovery and
Functional Properties of Soy Storage Proteins from Lab- and Pilot-Plant Scale
Oleosome Production. Journal of the American Oil Chemists’ Society, 89(5), 947–
956.
Kapchie, V. N., Towa, L. T., Hauck, C., & Murphy, P. a. (2009). Recycling of Aqueous
Supernatants in Soybean Oleosome Isolation. Journal of the American Oil Chemists’
Society, 87(2), 223–231.
Kapchie, V. N., Towa, L. T., Hauck, C., & Murphy, P. a. (2010). Evaluation of enzyme
efficiency for soy oleosome isolation and ultrastructural aspects. Food Research
International, 43(1), 241–247.
Karki, B., Maurer, D., Box, S., Kim, T. H., & Jung, S. (2012). Ethanol Production from
Soybean Fiber, a Co-product of Aqueous Oil Extraction, Using a Soaking in Aqueous
Ammonia Pretreatment. Journal of the American Oil Chemists’ Society, 1345–1353.
Lamsal, B. P., & Johnson, L. A. (2007). Separating Oil from Aqueous Extraction Fractions
of Soybean. Journal of the American Oil Chemists’ Society, 84(8), 785–792.
Liu, F., & Tang, C.-H. (2013). Soy protein nanoparticle aggregates as pickering stabilizers
for oil-in-water emulsions. Journal of Agricultural and Food Chemistry, 61(37),
8888–98.
Murphy, D. J. (1993). Structure, function and biogenesis of storage lipid bodies and
oleosins in plants. Progress in Lipid Research, 32(3), 247–280.
Nikiforidis, C. V, & Kiosseoglou, V. (2009). Aqueous extraction of oil bodies from maize
germ (Zea mays) and characterization of the resulting natural oil-in-water emulsion.
Journal of Agricultural and Food Chemistry, 57(12), 5591–6.
Ouhida, I., Pérez, J., & Gasa, J. (2002). Soybean (Glycine max) cell wall composition and
availability to feed enzymes. Journal of Agricultural and Food
Rosenthal, a., Pyle, D. L., & Niranjan, K. (1996). Aqueous and enzymatic processes for
edible oil extraction. Enzyme and Microbial Technology, 19(6), 402–420.
74
Isolation of oil bodies from soybeans in a mild way: definition of operating
window for process design
Rosenthal, A., Pyle, D. ., Niranjan, K., Gilmour, S., & Trinca, L. (2001). Combined effect
of operational variables and enzyme activity on aqueous enzymatic extraction of oil
and protein from soybean. Enzyme and Microbial Technology, 28(6), 499–509.
Salunkhe, D.K., Chavan, J.K., Adsule, R.N. & Kadam, S. S. (n.d.). World Oilseeds.
Chemistry, Technology and Utilization. New York: Van Nostrum Reinhold.
Shirsath, S. R., Sonawane, S. H., & Gogate, P. R. (2012). Intensification of extraction of
natural products using ultrasonic irradiations—A review of current status. Chemical
Engineering and Processing: Process Intensification, 53, 10–23.
Tzen, J. T. (1992). Surface structure and properties of plant seed oil bodies. The Journal of
Cell Biology, 117(2), 327–335.
Vilkhu, K., Mawson, R., Simons, L., & Bates, D. (2008). Applications and opportunities
for ultrasound assisted extraction in the food industry — A review. Innovative Food
Science & Emerging Technologies, 9(2), 161–169.
Wilczek, M., Bertling, J., & Hintemann, D. (2004). Optimised technologies for cryogenic
grinding. International Journal of Mineral Processing, 74, S425–S434.
Wu, N.-N., Huang, X., Yang, X.-Q., Guo, J., Zheng, E.-L., Yin, S.-W., Zhang, J.-B. (2012).
Stabilization of soybean oil body emulsions using ι-carrageenan: Effects of salt,
thermal treatment and freeze-thaw cycling. Food Hydrocolloids, 28(1), 110–120.
75
Chapter 5
Chapter 5
Product-driven process synthesis for the extraction of
oil bodies from soybeans
ABSTRACT
In this chapter the Product-driven Process Synthesis methodology was used as a well-
defined structured approach for the conceptual design of an extraction process for the
isolation of intact oil bodies from soybeans. In the first part of this chapter, the product-
related design stages of the Product-driven Process Synthesis methodology were addressed.
This was done by defining and framing the product problem and subsequently mapping
consumer’s wants onto product attributes. The concept of House of Quality was used to link
both design spaces and additionally to map the product attributes onto product properties.
From this exercise a problem formulation was generated that basically established the
input (soybeans) and the output (intact oil bodies). Further, we identified the fundamental
tasks to convert raw materials into a final product which led ultimately to two alternative
process flow-sheets. These process flow-sheets were not obvious and hence were
experimentally verified. In this chapter, we demonstrated that the process alternative based
on combined enzymes and ultrasound improved the purity of the final product, provided a
better separation for the protein-lipid, and reduced processing time. For the selected
alternative, we defined the operating window and developed an overall process design.
76
Product-driven process synthesis for the extraction of oil bodies from soybeans
5.1. Introduction
In classical petrochemical industry the product design is separate from process design;
product properties are not so much dependent on the way they are produced as long as
sufficient purity is achieved. Chemical product design by Moggridge and Cussler (2000) is
explained in detail. On the other hand, Douglas (1985) introduced process synthesis as a
structured route for developing a process flow-sheet. Therefore, in the last ten years the link
between process design and the development of novel consumer products became
increasingly important. Especially in regard to the design of processes for structured
products, which are more difficult when using only process synthesis tools. These
structured products have high added value and they are often complex multi-phase
materials (e.g. cosmetics creams and lotions, margarine, ice cream, etc.) Edwards (2006).
The Product-Driven Process Synthesis (PDPS) methodology proposed by Bongers and
Almeida-Rivera (2012) connects product design with process synthesis. The product-driven
process synthesis method comprises a multi-level decision hierarchy with increasing level
of complexity that aids the user in the development of new products and processes.
Figure 1.5 in Introduction Section shows the hierarchy that starts at the framing level and
ultimately leads to a complete conceptual process design (including equipment design and
multi-product equipment integration). Bongers and Almeida-Rivera (2009) explained the
complete hierarchy in detail. There are nine levels in the PDPS methodology, see Figure
3.1in the Chapter 3. It is relevant to mention that the scope of the approach expands over
design spaces of multi product integration, scheduling and control (Bongers and Almeida-
Rivera, 2012).
In this chapter we will use the isolation of intact oil bodies (OBs) from soy beans to
illustrate the applicability and scope of the methodology. In Figure 5.1, images of soybean
and oil bodies are presented. Oil bodies are located in cotyledon cell surrounded by protein
bodies.
77
Chapter 5
Figure 5.1: Structure of a soybean grain (left) and transmission electron micrograph of a soybean cell (A.
Rosenthal et al., 1998) (right)
5.2. Framing level
At the framing level, the background of the project, the business context and the potential
of OBs as food additives are identified. Oil bodies (OBs) obtained from oilseeds have been
exploited for a variety of applications in biotechnology. These applications are based on
their non-coalescing nature, ease of extraction and the presence of unique membrane
proteins (oleosins). In a suspension, OBs exist as separate entities and hence they can be
used as emulsifying agent for a wide variety of products, ranging from vaccines, food,
cosmetics and personal care products (Bhatla, Kaushik, & Yadav, 2010; Kapchie, Yao,
Hauck, Wang, & Murphy, 2013). Currently, OBs are mainly used in the personal care and
cosmetic industry. OBs have a significant affinity to the skin, and can be used as delivery
systems for exogenous oils or vitamins. Some of the products include sunscreens, make-up
removers and hair products. Most of those applications involve synthetic OBs. To date,
SemBioSys Ltd (Canada) is the only company with patents on natural OBs. This company
obtains OBs from (genetically modified) safflower and using them as protein carriers in
personal care and pharmaceutical applications (Deckers et al. 2003). The process used by
SemBioSys company is based on the aqueous extraction process (AEP) of seeds. From an
experimental study we concluded that AEP is a suitable mild extraction technique for
extraction of OBs. In addition, the experimental results helped us to define the operational
windows for all process conditions e.g. process temperature, pH, desired particle size of soy
flour, salt concentration. Soybean OBs can be used in food products like dressings, sauces,
78
Product-driven process synthesis for the extraction of oil bodies from soybeans
dips, beverages, and desserts. In food, OBs are used either to prevent lipids against
oxidation as “natural antioxidants”. Lipids are protected by a layer of proteins and therefore
protected from oxidation. In fat-containing food products (e.g. mayonnaise or margarine)
OBs could be used as “natural antioxidants”.
5.3. Consumer wants and product ideas
At the level of consumer wants and product ideas we translate what a consumer actually
wants into product characteristics such as smoothness, whiteness and creaminess. In this
section we propose a flow diagram that assists in translating the consumer wishes into
product formulations (see Figure 5.2). Three different shapes can be distinguished in the
diagram. The rectangles represent “processing” steps such as detailed interviews with
consumers, design of product recipe, measurements of functional requirements, tasting
sessions by panels and optimizing the product formulation. Blocks are reserved for
collecting the data after interviews with consumers, scoring of product characteristics and at
the end when we finally have a product that satisfies the consumer wants. The diamond
shape represents the decision making moment. Round shaped blocks represent start and end
of product design algorithm.
79
Chapter 5
Rough interview with
consumers
(first round)
Detailed interview
(second round)
Ranking of
consumer
needs
Scoring of the
product
characteristics
Design product
recipe
Measurement of
functional
requirements
Testing session
Scoring of the
product
characteristics
Consumers
satisfied
Modify recipe of
the product
New designed
product and
correlation between
attributes and
characteristics
Process Synthesis part
(continuation of PDPS
methodology)
YES
NO
Figure 5.2: Product design algorithm with collecting data steps (blocks), processing data steps (rectangles) and
decision making (diamond)
The most efficient and direct method for collecting product ideas is to ask consumers what
type of product they would like to get and to record those ideas. The ideas collected in this
way form the core of our further research. Our starting point in the product design
algorithm is, therefore, called a “rough interview with consumers - first round”. At this
stage, for instance, we could introduce to the consumer a new type of mayonnaise that
contains OBs. Firstly, we need to evaluate what the consumer values the most. For
example, we could find out whether “the product should be healthy”, or, “the product
should be easy to use”. After collecting all possible ideas that are coming directly from
consumers, we rank and sort them. For instance, a “healthy product” could be considered to
be more important than an “easy-to-open package”; or offering a cheap product could be
equally important as the product appearance. Organizing the ideas which we have generated
from the consumer interviews in this way can closely approach the final product idea.
After ranking and sorting the product ideas, a second, more detailed interview with
consumers takes place. At this stage, consumers can be asked to define in more detail what
the expected characteristics of the product should be. Some examples of the questions
used for the second (detailed) interview include the following:
1. What do you consider as a healthy product?
2. Do you prefer a solid or a liquid product?
80
Product-driven process synthesis for the extraction of oil bodies from soybeans
This is a qualitative way to describe product characteristics. The characteristics are scored
and those with the highest scores are those that will have the most outspoken effect on the
overall design of the product recipe. To analyze the collected data in a more structured and
organized manner, i.e. the House of Quality (HoQ) or sometimes referred to as the Quality
Function Deployment (QFD) concept can be used (see for example Dawson and Askin
(1999) and Karsak et al. (2003)).
The HoQ is a diagram resembling a house that is composed of several blocks (the consumer
requirements, a planning matrix, technical requirements, interrelationships, and the “roof”
of the house). Figure 5.4 presents the HoQ diagram. At the right hand side of the house, the
consumer requirements are listed. In this particular case, we are producing OBs, which are
an intermediate product and will be used as food natural antioxidant in, for instance, a
mayonnaise. Upon adding OBs to mayonnaise we have to ensure that the newly designed
mayonnaise with OBs has the same characteristics as the existing one. This means that the
designed product has to satisfy the following requirements identified by the consumer as
relevant: to have good taste, to have good texture, to be a healthy product and to have an
attractive price for value. All these requirements are collected on the left hand side in the
HoQ as demanded quality/consumer requirements. Detailed descrption of HoQ method is
presented in Appendix H.
These consumer wants (what’s) are mapped onto a set of product attributes (how’s).
Examples of attributes include creaminess, thickness, and ease to be removed from the
spoon, smooth appearance and microbiologically stability. The relationships between the
consumer- and product attributes are shown in Figure 5.3 with four grades of correlation
(strong, moderate, weak and none). For example, taste and creaminess have a strong
relationship, which is provided by the level of fat in the recipe. In fact, high fat level
products are expected to be deliver high creaminess scores and taste profiles.
A strong advantage of this matrix is the “roof” block where we correlate the product
attributes (how’s) that characterize the product. This is done by using the symbols ++
(strong positive), + (positive) and – (negative) correlation. In this particular case the couple
creaminess-smoothness is strongly correlated. The reason for this is obvious as a smooth
mayonnaise has to be creamy as well. Any increase in creaminess is directly improving the
product smoothness. Moreover, to make clear how the designed product attributes should
look like, an improvement block is included below the “roof” of the house. At this block it
we have the possibility to state which technical product attribute should be maximized (e.g.
microbiologically safety) and which needs hit a specific target (e.g. appearance, creaminess
and smoothness).
81
Chapter 5
Figure 5.3: House of Quality diagram including all blocks (consumer requirements, planning matrix, technical
requirements, interrelationships, and “roof” of the house). Legend: Θ-strong relationship, Ο-moderate
relationship, ∆-weak relationship; +: positive correlation, +-: moderate correlation.
5.4. Input/output level
At this level a complete specification of all exchange streams to the process (inputs/raw
material(s) and target output/product(s) are identified. The consumer wants and ideas level
leads to the formulation of a product formulation, for instance a mayonnaise that contains
OBs as “natural antioxidants”. In this case we define as input (feed) the soybeans with their
composition and as output (product) the amounts and the condition in which we would like
to isolate the OBs from the soy. In soy beans the OBs are located in the cells and
surrounded by protein bodies. The product should contain approximately 80 % (w/w) of the
intact OBs present in the soy bean. Visual inspection of the microstructure can be used as
an indicator for intact OBs. A desired microstructure (intact OBs) is identified by Confocal
Scanning Laser Microscopy (CSLM) which can be used to analyze the obtained OBs
fraction. Figure 5.4 shows a CSLM picture of intact OBs and protein bodies as well as
shape and size of these entities and their position in the microstructure. Most of the soybean
82
Product-driven process synthesis for the extraction of oil bodies from soybeans
protein and oil are stored in the cotyledon tissue in the organelles (also called protein
bodies) and oil bodies (oleosomes), respectively (images presented earlier in Introduction
section, page 12). The typical soybean composition is 1 % (w/w) free oil, 21 % (w/w) OBs,
38 % (w/w) of protein bodies, 35 % (w/w) carbohydrates and 5 % (w/w) ash.
Figure 5.4: Confocal Scanning micrographs of product (intact OBs) staining with Nile blue dye (bar 10 µm);
(small individual green particles (intact OBs) and bigger particles stained in red (protein bodies)
5.5. Task network
The next step in the methodology is to identify the fundamental tasks that are needed to
convert the input into the desired output. i.e. determination of the task network . The aim is
to isolate intact OBs from soybeans originally present inside cotyledon cells. To make the
oil bodies accessible the following tasks need to be executed (reference code is included for
the sake of simplicity), see also Table 5.1:
1. Size reduction of particulates (C2)
2. Cell wall disruption (C3)
3. Separation of a system into two systems with different composition (G1)
4. Physical/biological stabilization (J1)
These fundamental tasks are the critical and essential tasks based on the required to obey
desired final structure of the product. These four tasks (size reduction, cell wall disruption,
separation task and biological stabilization) can be combined leading to 24 (4!=24)
different processing routes. For example, task C3 could be performed first then task C2,
83
Chapter 5
after that J1 and at the final task G1. Among these possible combinations, only one is
feasible: C2, C3, G1 and J1. Moreover, for every task, there are different “mechanisms”
that could be used to perform task. In Table 5.1 all possible mechanisms for execution of
these tasks are presented. Heuristics, domain knowledge and project constraints to use mild
conditions (i.e. exclusion of organic solvents, high temperatures, extreme pH values: strong
acidic or alkaline conditions) are used to eliminate several of mechanisms and hence
reducing the number of potential alternatives.
Table 5.1: Fundamental tasks and mechanisms
Steps Fundamental task* Mechanisms*
Particle size reduction C2
C21: attrition
C22: impact
C23: ultrasound
C24: cutting
C25: enzymes
Cell wall disruption C3
C31: internal cell phase change
C32: electro-magnetic fields
(PEF, ultrasound)
C33: shear
C34: enzymes
C35: chemical
Extraction of OBs G1
G11: molecular size
G12: particle size
G13:electrical charge
G14: solubility
G15: chemical affinity
G16: chemical reaction
G17: (vapour) pressure
G18: gravity
G19: molecular size and electrical
charge
G20: shear
Physical/biological
stabilization J1
J11: freezing
J12: cooling * “Mechanisms” is the nomenclature used in the paper of Almeida and Bongers (2010)
All possible combinations result in more than 100 routes that could be followed. As this
number of alternatives is far from manageable a further simplification is proposed for the
network based on the following engineering-driven heuristics (H) and project constraints:
H1: Mechanisms C22 and C24 are not considered because soybeans are solid.
H2: C23 and C25 are excluded because these mechanisms will be used for the cell
wall disruption step.
84
Product-driven process synthesis for the extraction of oil bodies from soybeans
H3: C32 has been excluded due to the low water content in soybeans. PEF was
considered after soaking, but an external electric field could increase temperature
at micro level. This fact is crucial as OBs are heat sensitive materials.
H4: G17 and G20 are also not considered due to project constraints (i.e. no harsh
conditions).
H5: Experiments showed that proteins present in the cream fractions were
identified based on their molecular size and electrical charge by sodium duodecyl
sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). But this method was
not used to separate protein and oil bodies. This is the reason why task G19 has
not been considered.
H6: C35, G15 and G16 have been excluded because of environmental reasons.
This project aims at isolating OBs from soybeans under mild conditions, hence
without the use of chemicals. Furthermore, OBs will be used in the food industry
and these compounds have to be approved by obeying strict law regulations.
H7: J12 is rejected because OBs (as a final product) have to be stored at -18 oC to
avoid contamination of the product. Therefore, cooling is not sufficient.
Table 5.2: Selected tasks and mechanism
Steps Fundamental task Mechanisms
Particle size reduction C2 C21: attrition
Cell wall disruption C3
C32: electro-magnetic fields-
ultrasound
C34: enzymes
Extraction of OBs G1
G11: molecular size
G12: particle size
G13: electrical charge
G14: solubility
Physical/biological
stabilization J1 J11: freezing
After applying these heuristics, the remaining possible and feasible mechanisms are
identified (Table 5.2). All listed mechanisms in Table 5.2 have been experimentally tested.
Two feasible task networks can be formulated on the basis of the appropriate mechanisms.
Figures 5.5.1 and 5.5.2 depict our proposed task network alternatives.
85
Chapter 5
Figure 5.5.1: Proposed task network for alternative 1. Codes in the scheme are given in Table 5.1
Figure 5.5.2: Proposed task network for alternative 2. Codes in the scheme are given in Table 5.1.
86
Product-driven process synthesis for the extraction of oil bodies from soybeans
The major differences between the two proposed task network alternatives rely on the
mechanism associated to the cell wall disruption task. In alternative 1 a cell wall disruption
is performed only with enzymes, and in alterative 2 enzymes are combined with ultrasonic
treatment. These two different treatments are applied to the soy flour to improve the
extraction of the OBs. Cell wall degrading enzymes are applied to break the cotyledon cell
and to make the structure more permeable. Moreover, ultrasound is applied to increase the
transport of elements through cellular membranes, and to extract cellular structures from
damaged cells by cavitation (Vilkhu et al., 2008). In Table 5.3 the compositional mass
balance collected from experiments for two task network alternatives is presented. The total
protein content in the different fractions is around 80 % of the total solids. This is due to the
fact that some of the proteins remained trapped in the residue. From the compositional
balances it can be seen that the lipid-to-protein ratio increased from 5:1 for the experiment
with enzyme pretreatment and to 10:1 for the enzyme-ultrasound experiment, meaning that
the pretreatment (especially ultrasound) improved the “purity” of the cream. The theoretical
lipid-to-protein ratio for intact and “pure” OBs is 20:1, indicating that pretreatment with
enzymes combined with ultrasound gives better results for the lipid-protein separation than
pretreatment only with enzymes. Moreover, ultrasound increased the enzyme activity which
reduced the processing time. Finally, among the two task networks, task network 2 has
been selected because of its better results for extraction of OBs in cream fraction.
Table 5.3. Protein and lipid composition obtained from two different alternatives
Treatment
Composition (% dry-weight basis)
Cream Skim Residue
Protein Lipids Protein Lipids* Protein Lipids
Enzyme
(alternative 1) 13.2 74.9 45.0
< 10%
23.5 10.5
Enzyme combined
with ultrasound
(alternative 2)
8.6 84.4 43.7 25.0 6.8
* In the skim fraction, no lipids were detected. During the extraction process in the skim fraction, two phases were
observed: the residual (solid) and the supernatant (aqueous) phase. In the supernatant phase of the skim fraction,
the concentration of lipids was less than 1 % on wet basis.
5.6. Mechanism and operating window
For each mechanism the operating window has to be defined. The first step in the process is
the milling/grinding operation which determines the particle size of the soy flour. For
soybean extraction using ultrasound and enzymes a frequency of 25 kHz is selected,
because low frequencies (20-100 kHz) are recommended for dominant physical effects of
87
Chapter 5
cavitation. In addition, this frequency range intensifies the mass transfer rates (Shirsath et
al., 2012). The application of a cell wall degrading enzyme solution (cellulases/pectinases)
of around 2-5 % v/w during an incubation time of 1-3 hours is proposed. The process
temperature is increased from room temperature to 40°C while incubating. The mass
transfer rate is favored by increased solute solubility and diffusion into the bulk solvent;
however, degradation of thermo-labile components (i.e. proteins) should be considered
when working at high temperatures. For aqueous extraction Jung (2009) obtained a
maximum oil yield in the extraction of soybean at temperatures between 40-60 °C, while
Rosenthal et al. (2001) noted a slight decrease in oil yield for temperatures above 50 °C,
which they attributed to protein denaturation. This was one of the reasons that in our case a
temperature of 40 °C was chosen.
The bulk of proteins is stored in protein bodies, which may vary in size from 2-20 μm in
diameter. Oil is located in smaller oil bodies which are 0.2-0.5 μm in diameter and are
surrounded by the protein bodies. Separation of oil and protein bodies can be done based on
their molecular size. With filtration the oil bodies can be further separated from larger
molecules (protein and other large molecules). OBs are coated with a layer of proteins
(oleosins) that protects the phospholipids monolayer from contact with phospholipases that
are present in the cell. Moreover, this layer gives the OBs a negatively charged surface,
while at the same time preventing OBs from aggregation. To improve the separation of
exogenous proteins from the OBs cream, the ionic strength of the extraction buffers plays
an important role. Interactions between OBs surface proteins and exogenous proteins
include electrostatic and van der Waals forces, hydration effects, hydrogen bonding, salt
bridging, and ion binding. When salt is used at low concentration, salt ions provide charge
shielding or ion binding on the charged proteins (Tsumoto et al., 2007). This can result in
repulsive interactions between the protein bodies and oleosins proteins, and a subsequent
increase in OBs purity (electrical charge effect). Finally, we can summarize the
fundamental tasks, mechanisms and operating window in Table 5.4.
Table 5.4. Selected mechanisms and operating windows
Fundamental task Mechanisms Operating window
C2
C21: attrition
C21: Milling / size ring: 0.2μm
C3
C32: electro-magnetic fields-
ultrasound
C34: enzymes
C32: Frequency: 25 kHz
C34:Concentration:5%(v/w);
incubation: 3 h at 40oC
G1
G11: molecular size
G12: particle size
G13:electrical charge
G11: Membrane size: ≥0.6μm
G12: Centrifuge: 4700rpm at 4oC
G13: Buffer: 250mM NaCl
88
Product-driven process synthesis for the extraction of oil bodies from soybeans
G14: solubility
G14: Hydration for 30min at 4oC
in sodium phosphate buffer 0.1M
pH7.2
J1 J11: freezing J11: -18oC
5.7. Equipment integration
Finally, the fundamental tasks and operating window are translated into suitable processing
equipment. In Figure 5.6 the proposed process flow sheet is presented. It is noted that the
purity of the product could be chosen differently because it depends on the application of
the product (food or cosmetics). One of the main constraints is that only food grade type of
solvents may be used. For this reason water is selected as a solvent. Firstly, the soy is
grinded in a grinding mill to create soy flour of the preferred particle size. Secondly, the
soy flour is hydrated with a sodium phosphate buffer in a mixing tank. After mixing the
slurry is blended with an (2-5 % (w/v)) enzyme solution and the mixture is exposed to an
ultrasonic treatment while it is reacting in a stirred tank reactor. After reaction a three
phase centrifuge is used to separate the cream fraction with OBs from the supernatant
(skim) and residue of the bottom. A washing steps with sodium chloride follows. A
secondary treatment with a three phase centrifuge is required to separate the purified cream
from the supernatant and residue left after washing. Ultimately an ultra-filtration membrane
unit separates the oil bodies from the proteins on the basis of their molecular size.
89
Chapter 5
Figure 5.6. Proposed flow sheet for the process of extraction of OBs from soybeans
90
Product-driven process synthesis for the extraction of oil bodies from soybeans
The performance of the proposed process is assessed by a simple economic analysis and
compared with the conventional process. The difference between product revenue and raw
material costs is computed on a year basis for 1 ton/hr. To estimate the economic potential
(EP) (Cussler and Moggridge 2011) of the proposed process the following equation was
applied:
EP=Product revenues-Raw material cost
𝐸𝑃 = ∑ 𝐶𝑖 ∙ 𝐹𝑖 − ∑ 𝐶𝑗 ∙ 𝐹𝑗𝑗−𝑟𝑎𝑤 𝑚𝑎𝑡.𝑖−𝑝𝑟𝑜𝑑𝑢𝑐𝑡 eq. (5.1)
where Ci and Cj are the sales prices of the products i and the costs of the raw materials j,
respectively. F denotes the annual flow of products and raw materials. The input materials
for this estimation include the soybeans, water, NaCl and the enzyme. As output, the three
fractions obtained from the aqueous extraction process (cream, skim and residue) are
considered. The cream fraction can be sold as a high-value soybean OBs isolate (OBs). The
skim requires some concentration (by membrane filtration or evaporation) to be sold as a
high quality protein concentrate (SPC) for the food industry, but more natural and with
better functionality than the current products in the market. Finally, the residue can be sold
as soybean meal (SBM) for the feed industry or as a source for fermentable sugars for the
production of bioethanol or enzymes. In Table 5.5 the raw materials and final products
costs are detailed.
91
Chapter 5
Table 5.5. Raw materials and final product prices assumed for economic potential for the AEP of OB from
soybeans
Materials Costs of the raw materials
Input
Soybeansa €550/ton
Enzymee €8.7/kg
NaClc €50/ton
Waterd €1.25/m3
Sales prices of the products
Output
OB isolate (OBs)e** €5000/ton oil
Soy protein concentrate (SPC)f €1200/ton protein
Soybean meal (SBM)a €450/ton
* aChicago board of trade; c Indexmundi; d EUROSTAT; e Rough estimation; f Campbell and Glatz (2010)
**Price estimated with the price of high quality vegetable oil
For calculation of the EP for proposed process, several assumptions are made. The capacity
of the plant is set at 1 ton/h of soybeans and 20 h/day, giving a final annual flow of 7300
tons of processed soybeans. The composition mass balance (dry-weight basis) of each of
the streams is considered, since the price depends on the purity of the product (e.g. amount
of oil in the OBs isolate, and amount of protein in the SPC). Finally EP for proposed is
compared with conventional process for production of soybean oil.
Table 5.6. Economic potential for proposed process and conventional soybean oil process
Process
M/year
Cost of raw
materials
Sales prices of
products Economic Potential
Conventional process (with
hexane) € 7.3* € 7.7 € 0.4
Proposed process € 4.1 € 7.6 € 3.5
*In the conventional process use of hexane increases costs of raw materials
92
Product-driven process synthesis for the extraction of oil bodies from soybeans
The estimation of the economic potential showed that the conventional process involved
high costs: the use of hexane and it represented around 40 % of the total input costs. High
economic potential is obtained for proposed process with a value of around €3.5
million/year, available for investment in the process.
5.8. Conclusions
We have demonstrated the use of the product-driven process synthesis methodology for the
conceptual design of OBs extraction from soy beans. In the first part of the methodology
the House of Quality method was used to provide a link between consumer’s wants and the
design of the process bringing together product attributes and measurable product
properties. Depending on the application, whether OBs will be used as food additive (e.g.
as a natural antioxidant in mayonnaise) or in cosmetics (e.g. in face creams), the product
(OBs) itself requires different physical, chemical, and microbiological properties. Next, we
identified the fundamental tasks to convert raw materials into a final product which led to
two alternative process flow sheets which were not obvious. These alternatives were
experimentally verified. In this work, we have shown that the alternative with a
combination of enzymes and ultrasound improved the purity of the cream (final product),
enhanced the protein-lipid separation, and reduced the processing time. For the selected
processing route, we defined an operating window and constructed an overall process
design. In addition, the estimation of economic potential showed that proposed process for
extraction of OBs has higher EP compare to the conventional process for production of
soybean oil. The use of hexane in the conventional process ensures purity of the final
product (soybean oil purity > 95 %), but at the same time increases costs of raw materials.
5.9. References
Bair, C. W., Snyder, H. E., & Technology, F. (1980). , Electron Microscopy of Soybean-
Lipid Bodies and Lipid- Protein, (September), 279–282.
Bhatla, S. C., Kaushik, V., & Yadav, M. K. (2010). Use of oil bodies and oleosins in
recombinant protein production and other biotechnological applications.
Biotechnology Advances, 28(3), 293–300.
Bongers, P. M. M., & Almeida-Rivera, C. (2009). 19th European Symposium on Computer
Aided Process Engineering. Computer Aided Chemical Engineering (Vol. 26, pp.
231–236). Elsevier.
93
Chapter 5
Bongers, P. M. M., & Almeida-Rivera, C. (2012). 11th International Symposium on
Process Systems Engineering. Computer Aided Chemical Engineering (Vol. 31, pp.
195–199). Elsevier.
Bouzrara, H., & Vorobiev, E. (2003). Solid–liquid expression of cellular materials
enhanced by pulsed electric field. Chemical Engineering and Processing: Process
Intensification, 42(4), 249–257.
Campbell, K. A, & Glatz, C. E. (2010). Protein recovery from enzyme-assisted aqueous
extraction of soybean. Biotechnology Progress, 26(2), 488–95.
Campbell, K. A., & Glatz, C. E. (2009). Mechanisms of aqueous extraction of soybean oil.
Journal of Agricultural and Food Chemistry, 57(22), 10904–10912.
Campbell, K. A., Glatz, C. E., Johnson, L. A., Jung, S., de Moura, J. M. N., Kapchie, V., &
Murphy, P. (2010). Advances in Aqueous Extraction Processing of Soybeans.
Journal of the American Oil Chemists’ Society, 88(4), 449–465.
Canatella, P. J., Karr, J. F., Petros, J. A., & Prausnitz, M. R. (2001). Quantitative study of
electroporation-mediated molecular uptake and cell viability. Biophysical Journal,
80(2), 755–64.
Cells, T., Hsieh, K., & Huang, A. H. C. (2004). Endoplasmic Reticulum , Oleosins , and
Oils in Seeds, 136(November), 3427–3434. d
Cussler, E.L. and Moggridge, G. D. (2011). Chemical Product Design (Second edi.). New
York: Cambridge University Press.
Deckers, H.M., van Rooijen, G., Boothe, J., Goll, J. and Moloney, M. M. (2003). Product
for Topical Applications Comprising Oil Bodies. United States.
Domínguez, H., Núñez, M. J., & Lema, J. M. (1994). Enzymatic pretreatment to enhance
oil extraction from fruits and oilseeds: a review. Food Chemistry, 49(3), 271–286.
Esser, A. T., Smith, K. C., Gowrishankar, T. R., Vasilkoski, Z., & Weaver, J. C. (2010).
Mechanisms for the intracellular manipulation of organelles by conventional
electroporation. Biophysical Journal, 98(11),
Huisman, M. M. H., Schols, H. A., & Voragen, A. G. J. (1999). Enzymatic degradation of
cell wall polysaccharides from soybean meal. Carbohydrate Polymers, 38(4), 299–
307.
Iwanaga, D., Gray, D., & Fisk, I. (2007). Extraction and characterization of oil bodies from
soy beans: a natural source of pre-emulsified soybean oil. Journal of Agricultural …,
8711–8716.
94
Product-driven process synthesis for the extraction of oil bodies from soybeans
Jung, S. (2009). Aqueous Extraction of Oil and Protein From Soybean and Lupin: a
Comparative Study. Journal of Food Processing and Preservation, 33(4), 547–559.
Kapchie, V. N., Towa, L. T., Hauck, C. C., & Murphy, P. a. (2011). Recovery and
Functional Properties of Soy Storage Proteins from Lab- and Pilot-Plant Scale
Oleosome Production. Journal of the American Oil Chemists’ Society, 89(5), 947–
956.
Kapchie, V. N., Towa, L. T., Hauck, C., & Murphy, P. a. (2009). Recycling of Aqueous
Supernatants in Soybean Oleosome Isolation. Journal of the American Oil Chemists’
Society, 87(2), 223–231.
Kapchie, V. N., Towa, L. T., Hauck, C., & Murphy, P. a. (2010). Evaluation of enzyme
efficiency for soy oleosome isolation and ultrastructural aspects. Food Research
International, 43(1), 241–247.
Karki, B., Maurer, D., Box, S., Kim, T. H., & Jung, S. (2012). Ethanol Production from
Soybean Fiber, a Co-product of Aqueous Oil Extraction, Using a Soaking in Aqueous
Ammonia Pretreatment. Journal of the American Oil Chemists’ Society, 1345–1353.
Lamsal, B. P., & Johnson, L. A. (2007). Separating Oil from Aqueous Extraction Fractions
of Soybean. Journal of the American Oil Chemists’ Society, 84(8), 785–792.
Murphy, D. J. (1993). Structure, function and biogenesis of storage lipid bodies and
oleosins in plants. Progress in Lipid Research, 32(3), 247–280.
Nikiforidis, C. V, & Kiosseoglou, V. (2009). Aqueous extraction of oil bodies from maize
germ (Zea mays) and characterization of the resulting natural oil-in-water emulsion.
Journal of Agricultural and Food Chemistry, 57(12), 5591–6.
Ouhida, I., Pérez, J., & Gasa, J. (2002). Soybean (Glycine max) cell wall composition and
availability to feed enzymes. Journal of Agricultural and Food
Perva-Uzunalić, A., Škerget, M., Knez, Ž., Weinreich, B., Otto, F., & Grüner, S. (2006).
Extraction of active ingredients from green tea (Camellia sinensis): Extraction
efficiency of major catechins and caffeine. Food Chemistry, 96(4), 597–605.
Rosenthal, a., Pyle, D. L., & Niranjan, K. (1996). Aqueous and enzymatic processes for
edible oil extraction. Enzyme and Microbial Technology, 19(6), 402–420.
Rosenthal, A., Pyle, D. ., Niranjan, K., Gilmour, S., & Trinca, L. (2001). Combined effect
of operational variables and enzyme activity on aqueous enzymatic extraction of oil
and protein from soybean. Enzyme and Microbial Technology, 28(6), 499–509.
95
Chapter 5
Rosenthal, A., Pyle, D. L., & Niranjan, K. (1996). Aqueous and enzymatic processes for
edible oil extraction. Enzyme and Microbial Technology, 19(6), 402–420.
Ruan, J., Haerdter, R., & Gerendás, J. (2010). Impact of nitrogen supply on carbon/nitrogen
allocation: a case study on amino acids and catechins in green tea [Camellia sinensis
(L.) O. Kuntze] plants. Plant Biology (Stuttgart, Germany), 12(5), 724–34.
SALE, A., & HAMILTON, W. (1967). Effects of high electric fields on microorganismsI.
Killing of bacteria and yeasts. Biochimica et Biophysica Acta (BBA) - General
Subjects, 148(3), 781–788.
Salunkhe, D.K., Chavan, J.K., Adsule, R.N. & Kadam, S. S. (n.d.). World Oilseeds.
Chemistry, Technology and Utilization. New York: Van Nostrum Reinhold.
Shirsath, S. R., Sonawane, S. H., & Gogate, P. R. (2012). Intensification of extraction of
natural products using ultrasonic irradiations—A review of current status. Chemical
Engineering and Processing: Process Intensification, 53, 10–23.
Tzen, J. T. (1992). Surface structure and properties of plant seed oil bodies. The Journal of
Cell Biology, 117(2), 327–335.
Vilkhu, K., Mawson, R., Simons, L., & Bates, D. (2008). Applications and opportunities
for ultrasound assisted extraction in the food industry — A review. Innovative Food
Science & Emerging Technologies, 9(2), 161–169.
Wilczek, M., Bertling, J., & Hintemann, D. (2004). Optimised technologies for cryogenic
grinding. International Journal of Mineral Processing, 74, S425–S434.
97
Chapter 6
Chapter 6
Conclusions and Outlook
6.1. Conclusions
The food industry is facing the challenge of developing new food products with
excellent health benefits and meeting consumers’ appreciation. Plant extracts have
as compared to products with an animal source or to synthetic products benefits,
especially towards environmental burden. There is also a need to develop novel
dietary strategies, especially with reference to the potential health properties such
as the antioxidant property of polyphenols from fresh tea leaves.
The structured foods that use natural ingredients with a vegetable origin as food
additives are also a trend in the food producing industry. Oil bodies from soybeans
represent a good example as a natural food additive that can be used in e.g. sauces,
margarine, and mayonnaise.
This thesis deals with development of food grade and sustainable processes for
isolation of key components from two different raw materials: fresh tea leaves and
soybeans. The development of such processes that use only food grade solvents
provide not only additional health benefits beyond basic nutrients, but products
with label “natural” based on a natural source.
To meet economic and environmental requirements for processes to process two
completely different plant materials i.e. fresh tea leaves and soybeans, the Product
Driven Process Synthesis (PDPS) methodology introduced by Bongers and
Almeida-Rivera (2009) has been applied. PDPS combines product design and
process synthesis in a structured and systematic approach.
Separation of polyphenols from fresh tea leaves has been difficult due to leaves and
polyphenols instability i.e. undesired degradation reactions. Isolation of oil bodies
from soybeans has been proven to be difficult due to their fragile structure and
degradation. Furthermore, the presence of other cell constituents such as e.g.
protein bodies that interact with oil bodies increases the level of complexity for the
isolation of the oil bodies.
98
Conclusions and Outlook
Increasing the knowledge about how components are present and anchored in the
original vegetable matrix and how their individual properties depend on the
conditions (moisture, pH, ionic strength, temperature, etc.) is prerequisite to
provide the possibility to allow a successful contribution of these components to
the food market of the future.
6.1.1. Polyphenols from tea leaves
One interesting process alternative to extract polyphenols is to open the cell
structure in the fresh tea leaves by applying intensive electrical pulses. The
disadvantage would be the increase in temperature on micro (cellular) level due to
the high electric field strength and the related currents. The work reported in this
thesis provides a valid evidence that pulsed electric field (PEF) processing is a non-
thermal method applied under the used conditions. With a correct combination of
operational factors (electric field strength, pulse duration and number of pulses)
from the pulsed electric field (PEF) method we observed a very limited
temperature increase. Opening of the cell membranes seems to be the key factor in
allowing polyphenols to be transported from the interior of the cell to the
surrounding liquid. Both electric field strength and the total treatment time (product
of pulse duration and number of pulses) play an important role in opening the cell
structure and subsequent polyphenols extraction. An extraction yield of about 30 %
has been obtained in the work reported in this thesis. This extraction yield has been
obtained for the total treatment time of 2.5 s and field strength of 0.4 kV/cm. To
achieve the same extraction yield but with shorter total treatment time (1.5 s)
higher electric field strength (0.9 kV/cm) is required. When total treatment time
was 5 s (for both electric field strengths 0.4 and 0.9 kV/cm) the experiment was not
performed due to limitations of with PEF equipment (low conductivity was
detected in the PEF chamber).
As a part of the PDPS approach the statistical analysis applied by using the Design
of the Experiments method demonstrates that electric field strength is a key factor
to maximize the amount of extracted polyphenols from fresh tea leaves while
minimizing pulse duration. Polynomial models developed from a Box-Behnken
design of experimental approach were used to determine the optimal conditions for
PEF. Operating factors that have a significant influence on the extraction yield of
99
Chapter 6
polyphenols are electric filed strength, pulse duration and number of pulses. The
optimal results allow an extraction yield of polyphenols of 32 %.
The pulsed electric field method is a promising technique for opening the cell
structure and for extraction of cellular material in molecular form.
6.1.2. Oil bodies from soybeans
To prevent degradation of oil bodies (OBs) by applying high temperatures and to
avoid organic solvents (e.g. n-hexane) residue in the final product, an aqueous
extraction process is a promising alternative for the separation of oil bodies and
proteins from soybeans. Depending on the application, whether OBs will be used
as a food additive (e.g. as a natural antioxidant in mayonnaise) or in cosmetics (e.g.
in face creams), the product (OBs) itself requires different physical, chemical, and
microbiological properties.
The outcome of the task network level of PDPS methodology leads to two
alternative process flow sheets which were not obvious. In a systematic screening
study three pretreatments have been tested: enzymatic degradation of the cell wall,
ultrasound induced opening of the cell wall and a combination of both.
Experimental results show that pretreatment with ultrasound reduced the remaining
insoluble fraction in the residue and increased the amount of oil bodies extracted
into the surrounding aqueous phase. In addition, an estimation of the economic
potential showed that the proposed extraction process of oil bodies based on the
combination of enzymes and ultrasound has a higher economic potential as
compared to the conventional process for production of soybean oil. For the
proposed process the input materials for economic estimation include the soybeans,
water, NaCl and the enzyme. As output, the three fractions obtained from the
aqueous extraction process (cream, skim and residue) are considered. On the other
hand, for the conventional process input are soybeans and n-hexane necessary for
extraction of the oil. The outcome of the conventional process is pure soybean oil.
The conventional process involved high costs: the use of hexane and it represented
around 40 % of the total input costs. For the capacity of 1000 kg soybeans per hour
high economic potential is obtained for proposed novel process with added value of
around 3.5 M€/year. The use of n-hexane in the conventional process ensures
100
Conclusions and Outlook
purity of the final product (soybean oil purity > 95 %), but at the same time
increases costs of raw materials.
6.2. Outlook
6.2.1. Wax removal from the surface of tea leaves
Besides extraction of polyphenols from fresh tea leaves, removal of the waxy layer
from the leaf surface in an early stage of processing would be beneficial. Waxes
are complicated mixtures of long aliphatic alcohols, fatty acid esters, etc. and
introduce an additional mass transfer resistance during drying of tea leaves. In tea
production, waxes participate in complex formation reactions with polyphenols and
cause sediment formations that later in the process have to be additionally
separated. This leads to product losses i.e. polyphenols and addition of separation
tasks and high process investment which can be probably avoided if the waxes are
removed prior.
6.2.2. Pulsed electric field method
As described throughout this thesis, pulsed electric field (PEF) method is a
promising technique for the opening the cell membrane. According to the electrical
measurements presented in section 2.3.1. pulsed electric field exerted opening the
cell membrane. This means that under the operating settings the temperature
increment was not pronounced (< 10 oC). Phospholipid is the main component of
the cell membrane. A phospholipid based vesicle could serve as a good model
species to examine and to elucidate in detail electroporation (opening the
membrane structure) and thermal effects (Tg - glass transition temperature effect)
caused by external electric field. Further, the study of the physicochemical process
involved in phospholipid vesicle membrane can provide useful information about
the complex phenomenon of electroporation. Also fundamental insight can be
gained for the multi stage extraction process.
101
Chapter 6
Next to this project, PEF method was used as a pretreatment prior blending or
pressing during processing of sugar beet leaves in the same project cluster (A.
Kiskini, Wageningen University and ISPT). PEF was introduced to enhance the
protein extraction yield from sugar beet leaves. Experiments were carried out on
semi-industrial continuous equipment with different operating factors as compared
to the batch laboratory scale PEF equipment used in this project. The continuous
PEF process is more difficult to control as compared to batch PEF unit due to the
fact that vegetable material is placed on the belt passing between two high voltage
electrodes. Due to the different size and orientation of sugar beet leaves, exposure
of the material under high voltage is not uniform.
6.2.3. Product and equipment integration
In the last level of PDPS, idea is to integrate different products and/or operational
units. In the case of soybeans depending on the application of oil bodies whether
they will be used as food additives or in cosmetics, purity demands of the final
products are different. The proposed process for extraction of oil bodies from
soybeans could be used for production of various products. However, the order of
the tasks execution would be different. For example, purity of oil body’s stream
that will be used in personal care products is not necessary to be high, e.g. around
60 % of oil bodies in the stream. On the other hand, oil bodies that will be used as a
food additive require a high purity above 90 %. Therefore an additional separation
steps should be introduced in order to reach desired product purity.
As described in the thesis, equipment for production of oil bodies from soybeans
has been selected. However, there is also room to optimize selected unit operations.
In the proposed process for isolation of oil bodies from soybeans to disrupt the cell
wall a combination of enzymes and ultrasound was reported. The time constant for
enzymes degradation of the cell wall in a bioreactor will take around 3 hours. On
the other hand, ultrasound process is several seconds. Therefore, control of such
system is complicated and requires optimization of each unit operation.
Owing to the complexity of food matrices, product driven process synthesis
methodology has been applied as a useful tool for conceptual process design for
isolation of key components from raw materials e.g. fresh tea leaves and soybeans.
103
Appendices
Appendix A
A.1. Enzyme-Assisted Extraction
The cell wall is the primary mechanical barrier for the extraction of the OBs from the cells,
therefore it must be broken for any significant extraction to occur (Rosenthal et al., 1998).
In plant cells, the primary cell wall is constructed of pectins (homogalacturonan,
rhamnogalacturonan, xylogalacturonan), hemicelluloses (xyloglucan, arabinoxylan), and
crystalline micro fibrils of cellulose, crosslinked with proteins. Within the primary cell wall
is a secondary cell wall of cellulose and hemicellulose. The cells are held together by a
middle lamella composed of mostly pectins ( Campbell, 2010).
Figure A.1. Structure of the primary cell wall (Cosgrove, 2005)
The utilization of hydrolytic enzymes, such as cellulases, hemicellulases and pectinases, to
break the cell walls and assist the extraction process can potentially increase the extraction
yield. Moreover, cell wall hydrolysis allows the extraction of undenatured proteins, which
is an important advantage over traditional extraction processes.
Soybean cell wall is considered to have very complex fibers. It contains pectins with a
considerable amount of arabinan, galactan or arabinogalactan side chains (Huisman et al.,
104
Appendices
1999). Several attempts have been made to completely degrade the soybean cell wall with
many enzymes. Huisman and authors (Huisman et al., 1999) applied a powerful
commercial enzyme mixture into the intact soybean cell wall polysaccharides. However,
the network of the cell wall polysaccharides present in soybean was too complex or too
dense to be penetrated by the applied enzymes. Ouhida et al. (2002) needed an extensive
sequential fractionation of the cell wall polysaccharides (using chelating and alkali
solutions) to increase the enzyme accessibility to the substrate, especially to the xylans and
cellulose. Kasai and others (Kasai et al., 2004) needed prolonged reaction times,
mechanical breaking and strong heat treatments (autoclaving) to obtain a solubilisation of
83-85 % of okara (residue from soymilk production containing the fibers). Although, the
complete degradation of the soybean cell wall is not the goal during OBs extraction. The
role of hydrolytic enzymes such as cellulases, hemicellulases and pectinases in these
processes is to make the structure more permeable (Rosenthal et al., 1996) and to increase
the diffusion of cell components (protein and oil bodies) to the extraction medium without
damage (Jung, 2012).
Kapchie et al. (Kapchie, 2008) could recover twice the amount of oil from OBs just by
using a mixture of hydrolytic enzymes during the AEP. After four consecutive extraction
steps, with 3 % of enzyme cocktail (Multifect Pectinase FE, Cellulase A, and Multifect CX
13L), 85 % of the total soybean oil was recovered as oleosomes. According to these
researchers, the enzyme assisted procedure mainly depended on the enzyme concentration
combined to the mechanical disruption of the cells to obtain maximum yields. Figure A.2
shows the effect of enzyme treatment on a previously grinded seed. The enzyme action
makes the structure more permeable; the extend will depend on the particle size (Rosenthal,
1996).
105
Appendices
Figure A.2. Effect of milling and enzymatic treatment on oilseed cell structure (Rosenthal et al., 1996)
The EA-AEP has successfully recovered up to 93 % of oil (pilot plant scale) from the
starting material and 40-50 % of protein in the skim fraction (Kapchie et al., 2011).
However, most of the applied extraction procedures involved long treatment times (15 to 20
h) (Kapchie, et al., 2010). Reducing the incubation and/or extraction times is essential for
reducing microbial risks and costs in EA-AEP (Karki et al., 2012).
A.2. Ultrasound-Assisted Extraction
The application of ultrasound-assisted extraction (UAE) in food processing is of interest for
enhancing existing extraction processes, but also for enabling novel extraction
opportunities. Some of the applications of UAE include the extraction of herbal, oil, protein
and bioactives compounds from plant materials (e.g. flavones, polyphenols) (Vilkhu et al.,
2008).
Intensification of extraction efficacy using ultrasound has been attributed to the propagation
of ultrasound pressure waves through the solvent, resulting in cavitation phenomena.
Ultrasonic cavitation creates significant shear forces that can disrupt the natural liquid
layers system close the phase boundaries, and consequently stimulate the process of mass
transfer. Cavitation provides another advantage in the form of destruction of cellular
structure, and consequently the release of cell contents into the surrounding solution (see
106
Appendices
Figure A.3) (Stadnik and Dolatowski, 2011). If the substrate is dry, then ultrasound may be
used to facilitate swelling and hydration and cause enlargement of plant cell wall.
Figure A.3. The mechanism of cell wall disruption (a) breaking the cell wall due to cavitation, (b) diffusion of
solvent into the cell structure (Shirsath et al., 2012)
Diffusion through the plant cell walls, disruption and washing out of the cell contents are
attributed to improved extraction performance. The corresponding reduction in the size of
the vegetable material particles by ultrasound disintegration will increase the number of
cells directly exposed to the extraction solvent and ultrasonic cavitation (Vilkhu et al.,
2008).
It has been reported that the major impact of ultrasound on the efficacy of extraction was in
terms of reduction in required time and slight increase in the overall extraction yield
(Vilkhu et al., 2008). Li and authors (Li, e al., 2010) studied the UAE of oil from soybean
flour, using a 20 kHz ultrasonic generator for a period up to 3 h. Compared with a non-
sonicated control, the oil yield increased 11.2 % (intensity of 47.6 W/cm2). Results of the
GC fatty acid profile showed that ultrasonication did not noticeably influence the
composition of the extracted oil (0.52 % decreased in linoleic acid). UEA can also provide
the opportunity for enhanced extraction of heat sensitive bioactive and food components at
lower processing temperatures (Vilkhu et al., 2008).
Furthermore, UAE can be conveniently coupled with other extraction techniques such as
supercritical fluid extraction, microwave assisted extraction, vacuum distillation and
enzymatic treatment (Shirsath et al., 2012). The combination of ultrasound with enzyme
treatment have shown synergistic effects (Stadnik and Dolatowski, 2011). Ultrasound
improves the transport of enzyme without generating an excessive amount of highly
107
Appendices
reactive intermediates which may cause deactivation of enzymes. Moreover, ultrasound can
also activate the catalytic performance of the enzyme adsorbed onto the surface of substrate
and enhance removal of the products from the reaction zone. Thus ultrasound increases the
efficiency of enzymatic treatment with higher extraction yield and lower times (Li et al.,
2010). Kapchie et al. (2009) investigated the effect of ultrasound on the EA-AEP of OBs
from soybean. Only 3 min ultrasonication prior enzymatic hydrolysis resulted in high oil
recoveries in the oleosome fraction (~80 % with 3 % enzymes). Four consecutive extraction
cycles were needed to obtain the same oil yield without ultrasound pretreatment. No
confirmation of OB integrity was reported in this study.
In the following table A.1, a summary of some of the most recent work in OBs AEP is
presented. Typically, the processes involved the use of alkaline TRIS buffers with moderate
salt concentrations, to buffers with 0.5 M NaCl and 0.4 M sucrose, under acidic conditions.
It is believed that high sucrose and salt concentrations are necessary to preserve the OBs
organelle integrity. Moreover, using high salt concentration increases protein solubility at
low pH, allowing cellulase-assisted extraction to occur near the protein pI (4.3-4.5).
However, the separation and purification of the proteins from those osmotic solutions (e.g.
by ultrafiltration or isoelectric precipitation) still need to be assessed to obtain a valuable
fraction at the end (Campbell et al., 2010).
108
Appendices
Table A.1. Summary of some of the latest works intended to extract intact oil bodies
NR not reported; AEP Aqueous extraction process; EA-AEP Enzyme-assisted aqueous extraction process a Taking the initial amount of oil or protein present in the original material as 100 %
Reference Starting
Material Pretreatment Extraction process
Extraction
medium
Recoverya Oil:Protein
ratio in
cream Oil from
cream
Protein from
supernatant
Iwanaga, et
al., 2007 Soybeans
Soaking in
buffer solution
(pH 8.6), 4-
6°C, overnight
Alkaline AEP
3mM MgCl2,
100mM Tris-HCl
buffer, pH 8.6
36% NR 4.6:1
Kapchie, et
al., 2008
Full-fat
soybean
flour
Ultrasonication
(3min, 70W)
EA- AEP
(incubation 3%
enzyme 57°C/20h)
2 to 4 cycles of
residue extraction
Flour extraction:
Potassium acetate,
pH 4.6 with 0.5M
NaCl and 0.4M
sucrose
Residue extraction:
Tris-HCl buffer,
pH 7.2, with 0.5M
NaCl and 0.4M
sucrose
No
pretreatment
(4 cycles):
84.65±1.46%
Pretreatment
(2-cycles):
78.87±7.19%
No
pretreatment:56.76±0.56% NR
Nikiforidis,
(2009)
Maize
germ
flour
(<0.8mm)
Soaking in
water, RT, 24h
Alkaline AEP
3 extraction cycles
NaOH solution (pH
9.0) 75.50% NR 4.3:1
Kapchie, et
al., 2010
Full-fat
dehulled
soybean
flour
Soaking at
57°C/16h in
osmotic
solution (0.4M
sucrose, 0.5M
NaCl)
EA-AEP
(incubation 3%
enzyme 57°C/20h)
4 extraction cycles
Potassium acetate,
pH 4.6 with 0.5M
NaCl and 0.4M
sucrose
77.42% 43.19% NR
109
Appendices
Kapchie, et
al., 2011
Full-fat
dehulled
soybean
flour
Pilot-Plant Scale
Continuous EA-EP
(incubation 3%
enzymes, 57°C/15h);
Recirculation of
supernatant and
residue 8h
Potassium acetate,
pH 4.6 with 0.5M
NaCl and 0.4M
sucrose
91.44±1.31% 79% 4.2:1
Chen et al.,
2012 Soybeans
Soaking
overnight 4-
6°C in buffer
(pH 8.6)
AEP 3mM MgCl2 in
Tris-HCl, pH 8.6 65% NR 14:1
110
Appendices
Appendix B
B.1. Measurement of enzyme activity of different commercial enzymes
In order to select an enzyme mixture to treat the soy flour, the enzymatic activity of
different commercials enzymes was measured by the 3,5-dinitrosalicilyc acid (DNSA)
assay. The cellulase, pectinase and xyloglucanase activity of each commercial preparation
was measured by using sodium carboxymethyl cellulose (9M31XF, Hercules), low
methylated pectic (32 % DE, Danisco) and tamarind xyloglucan (Megazyme) solutions (1
% w/v) as substrates, respectively. The activity was measured at 40 °C and pH 7.2 (see
Table B.1).
Table B.1. Enzyme activity measured by the DNSA assay and price of the different commercial enzymes
Enzyme activity (Ua/ml)
Substrate*
Costs (€/kg)
CMC LMP XG SF
Rapidase TF (DSM) 4.34 6.22 22.13 1.32 18
Ultrazyme AFP L (Novozymes) 1.93 13.15 14.49 1.85 8.7
CMC Carboxymethyl cellulose; LMP Low methylated pectic; XG Xyloglucan; SF soy flour. aU=µmol of glucose release per min, at 40°C and pH 7.2.
Enzyme preparations were purified using a PD-10 Sephadex column (GE Healthcare) to get
rid of free sugars, and diluted 100 times before activity measurement. A volume of 0.3 ml
of the substrate solution was mixed with 0.3 ml of the enzyme solution, and incubated at 40
°C/ 30min. After this period, 0.6 ml of the DNSA reagent was added, and the reaction
mixture was incubated at 95 °C/5 min to develop color. Samples were cooled; centrifuged
(14000 rpm/10 min) and diluted with distilled water, to finally measure the absorbance at
540 nm (Shimadzu UV-1601). Substrate and enzyme blanks, as well as calibration with
different glucose solutions (0-6mg/ml), were taken into consideration. Glucose calibration
curve is presented in the Figure B.1. An enzyme unit (U) was defined as the µmols of
glucose released per minute, under the conditions used.
111
Appendices
Figure B.2. Glucose calibration curve for the enzyme activity DNSA assay
112
Appendices
Appendix C
C.1. Protein and lipid recoveries from different extraction procedures
In the following table, the dry-weight composition of the different AEP fractions is shown.
Table C.1. Protein and lipid composition of the main aqueous extraction fractions obtained from soy flour
Treatment
Composition (% dry-weight basis)
Cream Skim Residue
Protein Lipids Protein Lipids* Protein Lipids
Control 12.8 75.8 49.4
<10%
23.4 9.5
E 13.2 74.9 45.0 23.5 10.5
U 10.4 81.4 44.9 23.7 7.3
EU 8.6 84.4 43.7 25.0 6.8
113
Appendices
Figure C.3. Protein and lipid recoveries in the different fractions from the aqueous extraction process of soy flour (the Total yield was obtained by addition of the protein
or lipids found in the skim1, residue, skim2, top and bottom cream)
114
Appendices
Appendix D
D.1. Particle size measurement: effect of the particle size on extraction of
OBs
Figure D.1. Particle size distribution of the soybean flours used for the AEP of OBs (A flour produce by a lab
scale Polymix mill; B flour cryogenically milled in a pilot plant scale Contraplex mill)
Right after milling the beans, flour A was immediately sieved using a vibratory sieve
shaker, however, efficient size classification was not achieved, as it can be seen in Figure
115
Appendices
D.1A. A double distribution was obtained with smaller particles than expected (125 to 250
µm); 90 % of the particles had a particle size lower than 300 µm and around 65 % had a
particle size lower than 100 µm. A possible reason is that, no matter the level of milling,
there is always some oil being extracted from the disrupted cells (especially since the oil is
in liquid state). Therefore, it is possible that free oil caused particles to stick to each other,
and made the separation of the smallest particles (<12 5µm) very difficult.
Flour B presented a normal distribution; 90 % of the particles had a particle size lower than
40 µm. Considering that cotyledon cells are about 15-20 µm in diameter and 70-80 µm in
length , such a particle size probably resulted in the rupture of a high proportion of the cells.
It is important to note that the particle size distribution was measured on the wet flour;
therefore, the result reflects the size of the hydrated particles. This might be the reason for
obtaining a higher particle size in flour A than the expected according to the sieve used
(250 µm).
Figure D.2. Particle size distribution of the fresh and aged (6 days at 4°C) “purified” cream fractions, obtained
from the AEP of two different soy flours (A produce by a lab scale Polymix mill; B cryogenically milled in a pilot
plant scale Contraplex mill)
The particle size and micro structure of the cream fractions, obtained from the AEP, was
analyzed as an attempt to determine the presence of intact OBs. In the following Figure
D.2, the average particle size of the fresh and aged (6 days at 4°C) cream fractions is
presented.
Intact soybean OBs have a diameter between 0.2-0.5 µm. However, as it can be seen in
Figure D.2, the average particle size of the majority of the particles was in the range
between 1 to 10 µm. The presence of non-oleosins proteins in the cream (as confirmed by
the lipid-to-protein ratio of the cream fractions) most probably caused the higher particle
116
Appendices
size in the emulsions. Moreover, the cream from flour B, which contained a higher amount
of proteins, had a wider size distribution, confirming that non-oleosins proteins are part of
the emulsion and increasing the particle size.
The particle size distribution was also measured after 6 days of storage to check if the
extracted OBs were intact and stable against coalescence. As seen in the Figure D.2, the
particle size of both creams did not change significantly over time. Moreover, no free oil
was observed in the creams after storage. Although no direct conclusion can be made, this
is an indication that the emulsion was stable and that the oil was protected against
coalescence.
D.2. Particle size measurement: effect of the pretreatments on extraction of
OBs
The particle size distribution of the top “purified” cream was measured by light scattering
to check if the size of the particles were similar to the expected OBs size (Figure D.3).
Figure D.3. Particle size distribution of the cream fraction obtained from the different AEP of soy flour
All fractions had similar particle size distribution: 90 % of the particles had a size of 1.5-1.9
µm or less; showing smaller particle size those creams from the ultrasonicated experiments.
In this case, particles as small as 0.4 µm were obtained, which are in the range of OBs size
117
Appendices
(0.2-0.5 µm). Nikiforidis and Kiosseoglou (Nikiforidis, 2009) have reported that maize
OBs contained a secondary external layer of extraneous germ seed proteins, which
interacted between each other and formed an OBs network. As already discussed before,
the “purified” top cream still contained some extra proteins, which could be forming a
complex network and slightly increasing the OB particle size.
The cream fraction was also stored for 7 days at 4°C, and the particle size was measured
again to check whether the emulsion was stable or coalescence of OBs occurred (Figure
C.4):
Figure D.4. Particle size distribution of the cream fraction after 7 days storage at 4°C
As seen in the Figure D.4, no big differences were obtained from the different cream
fractions. In the case of the cream obtained by enzymatic hydrolysis, the particle size
slightly increased, which could mean that the enzymes had some protease side activity and
oleosins were degraded, making the OBs unstable. However, no free oil was detected on the
top of the cream. Differences can be also a result of the normal variation of the instrument
measurement.
Overall, no big differences were found between the different AEP procedures. However,
some trends were observed: the application of a pretreatment favored the extraction of
lipids and proteins from the soybean solids to the aqueous media. Irrespectively of the
pretreatment, the remaining residue always decreased. However, the recovery of those
extracted components was not as expected. The Control treatment showed the highest oil
118
Appendices
recovery in the cream fraction. Meanwhile the ultrasonicated experiments showed the
highest protein recovery in the skim fraction. The lost of OBs into the skim fraction and the
inefficient detection method, were critical factors that might have caused the apparent high
oil losses. The combination of enzymatic hydrolysis with ultrasound resulted in the highest
lipid-to-protein ratio in the cream (10:1), while the washing step permitted a further
increment (as high as 14:1). However, some non-oleosins proteins are strongly bound to the
OBs, making the purification difficult.
119
Appendices
Appendix E
E.1. Microscopy analysis
One sample of the cream from soy flour A was analyzed by light microscopy (Figure E.1).
A few big droplets are seen in the range on 20-30 µm. Disruption of the natural OB
structure is difficult to avoid during the grinding of the soybeans; therefore, oil coalescence
may occurred during the AEP. However, many very small droplets are also appreciated in
the picture, which could be intact OBs.
Figure E.1. Image of cream fraction form soy flour A with a 20x (left) objective
In order to obtain more information on the microstructure of the creams, Confocal Scanning
Laser Microscopy (CSLM) was used to analyze the creams. When CSLM is used, materials
can be observed from different depths. The microscope makes an image of a certain section
of specific thickness; how deep in the sample the image is taken can be controlled by the
instrument operator. Therefore particles can be seen from the top, the bottom or the inside.
In this particular case, the sample was stained with Nile blue, which stains based on
polarity. Highly non-polar components are colored green; less non-polar components are
colored red, and polar components (i.e. water) are in black.
120
Appendices
Appendix F
F.1. Differential scanning calorimetry (DSC) results for the cryogenically
milled soy flour
Methodology:
DSC thermographs were recorded on a DSC 8500 (Perkin Elmer) using 0.1 ml vessels. Soy
flour and the extraction fraction were heated from 15 to 140 °C, at a heating rate of 10 °C/
min and subsequently cooled to 15 °C at the same rate. After this first heating cycle, a
second heating cycle was done to investigate the reversibility of the denaturation. The
instrument was calibrated with indium.
Results:
Protein denaturation is the unfolding of the protein from a structured native state into an
(partially) unstructured state with no or little fixed residual structure, which is not far from
a random coil (De Graaf, 2000).
Figure F.1. DSC thermogram of the soy flour B (red line first heating cycle, green line second heating cycle)
The thermogram of the soy flour B (cryogenically grinded) is given below, and shows two
small peaks at temperatures of 54 °C and 90 °C, which disappear after the first heating. The
main proteins in soy are β-conglycinin (7S) and glycinin (11S). The temperature of
denaturation ranges from 68-82 °C for 7S and 83-95 °C for 11S (Cramp, 2007). However,
121
Appendices
most of the thermal denaturation studies in soy proteins have been done in model systems
with solutions of purified proteins.
One of the possible reasons for these low peak areas is that soy flour was analyzed dry;
neither hydration nor a solution was made before analysis. Water is essential to allow
thermal motion of molecules (Sochava, 1997). The denaturation temperature of the protein,
Td, strongly depends on the water content up to a water content of 10–20 % (De Graaf,
2000), as detailed in the next figure.
Figure F.2. Effect of moisture on thermal denaturation of pure soy proteins (triangles glycinin, black circles β-
conglycinin) (Sessa, 1992)
The flour sample contained around 7 % of moisture; at such water contents, Td was
expected to be higher than the maximum temperature used during analysis (140 °C).
Rosenthal et al (Rosenthal, 1998), encountered a wide absorption peak between 150-180 °C
for soy flour, while the absorption peaks below 100 °C were negligible. Moreover, soy
flour is a complex system and interaction between proteins and other components (fat) may
have occurred, especially at low moisture contents (Sessa, 1992), which may have affected
Td as well.
It can be concluded that the small peaks at 54 and 90 °C cannot be attributed to soy protein
denaturation. Other thermodynamic changes may have occurred in the flour giving the
above result. It is necessary to increase the temperature of analysis of the water content of
the sample to be able to see the endotherms of the soy proteins.
122
Appendices
Appendix H
H.1. House of Quality (HoQ) method
HoQ represents matrix form or diagram resembling the house and it is the most recognized
for of Quality Function Deployment (QFD). It is utilised by multi-disciplinary team to
translate set of customer requirements (wishes and needs), market research, and technical
engineering data to meet a new product design. HoQ matrix is built of several blocks (see
Figure H.1):
1. Customer requirements (red block)
2. Planning matrix (green block)
3. Technical requirements (purple block)
4. Inter-relationships (yellow block)
5. Roof of the house
Figure H. 1. House of Quality scheme including all blocks (customer requirements, planning matrix, technical
requirements, inter-relationships, and roof)
1. Customer requirements
This is generally the first part of HoQ matrix that has to be completed and also very
important. It documents a structured list of consumer requirements described in their own
words. This information is usually gathered through conversation with customer in which
123
Appendices
they are encouraged to describe their needs and problems. The list of their requirements
gathered in such exercise has to be structured before its entry HoQ matrix.
2. Planning matrix
The planning matrix is attached to right side of HoQ matrix and serves several purposes.
Firstly, it quantifies customers’ requirement priorities and their perception of existing
product. Secondly it allows these properties to be adjusted based on the issues that concerns
design team.
The measures used in this part of HoQ matrix are gathered form questionnaire that
customers filled in before. The first and most important measure is important weighting.
This figure quantifies the relative importance of each of the customer requirements from
their own perspective (described in the left hand side of HoQ). This measure is often shown
alongside the customer requirements description in the block: customer requirements. The
questionnaire is used to collect these important weightings. To demonstrate this in the
Table H.1, example of mayonnaise is presented.
Table H. 1. Planning matrix example of mayonnaise
Rank the following requirements on their relative importance (5 – very important, 1 –
unimportant)
Easy to spread 1 2 3 4 5
Light in colour 1 2 3 4 5
Full mouth-feel 1 2 3 4 5
124
Appendices
Table H. 2. Planning matrix
When combining questionnaire data gathered from a certain and predefined number of
customers, one more thing should be taken into consideration: single market segment. This
means, that HoQ matrix is valid and perform for one market (e.g. Western Europe).
Customers from different regions have different habits in food, beverages, preparation of
meals. If sample includes different market segment, mean figure will not useful nor be of
any value to the product design team.
Moreover, planning matrix provides a measure of the satisfaction of customers with
available/existing products. Customers are asked to consider the performance of each of the
existing products in fulfilling their specified requirements.
3. Technical requirements
This section of HoQ is referred to as the engineering characteristics. It basically describes
the product in terms of the company. This information is generated by company team who
can identify all the measurable characteristics of the product. These measurable
characteristics that are perceived by company team are related with specified customer
requirements. In the same way that customer requirements are analyzed and structured, the
same approach has to be applied to interpret product characteristics.
4. Interrelationships
This section forms the main body of the HoQ matrix and can be very time consuming to
complete. Purpose is to translate the requirements as expressed by the customer into the
product technical characteristics of the product. Structure is simple-standard two
dimensional matrix with cells that relate combinations of individual customer and technical
requirements. The level of interrelationship is weighted usually on four scale point scale
(high, medium, low, none) and a symbol representing this level of interrelationship is
entered into matrix cell.
Imp
ort
ance
Easy to spread 5
Light in colour 2
Full mouth-feel 3
125
Appendices
Figure H. 2. Example of interrelationship matrix (Θ - strong relationship; Ο – medium; ▲- weak relationship)
5. Roof of the House
The triangular “roof” matrix of the HoQ is used to identify where the technical
requirements that characterize the product, support or impede each other. For example, does
improving one requirement cause deterioration or improvement in the other technical
requirement? Where the answer is a deterioration we product design team is going for
another option: engineering trade-off. To keep it structured in the “roof” matrix, a symbol is
entered that represents this (usually ‘-‘). Where improving one requirement automatically
leads to an improvement in the other requirement, an alternative symbol is entered (‘+’).
126
Appendices
Figure H. 3. Example of “roof” matrix (++ strong positive correlation; + positive correlation; - negative
correlation)
This information recorded in the “roof” matrix is useful to the design team in several ways.
It highlights where a focused design improvement could lead to a range of benefits to the
product. Also, it focuses on the negative relationships in the design. These can represent
opportunities for innovative solutions to be developed.
127
Appendices
Appendix I
Solving optimization models with the Lagrange multiplier method
In Chapter 2, optimization was used to maximize the extraction yield of polyphenols as
function of three variables (electric field strength, pulse duration and number of pulses). In
this appendix we will explain how the Lagrange multiplier method can be used to optimize
a given objective function. The general form of an optimization model is as follows:
max 𝑓(𝑥)
s.t.
𝑔𝑖(𝑥) = 𝑏 (P1)
ℎ𝑗(𝑥) > 𝑐
𝑥 > 0
where f(x) is the objective, gi(x) the equality constraints and hj(x) the inequality constraints.
All x’s are positive. To find a local optimum of P1 the Langrange multiplier method could
be used.
The Langrange multiplier method is based on solving the necessary and sufficient
conditions of optimality. We first start with the formulation the so called Lagrangian
function-
𝐿 = 𝑓(𝑥) + ∑ 𝜆𝑖(𝑔𝑖(𝑥) − 𝑏)𝑖 + ∑ 𝜈𝑖(ℎ𝑗(𝑥) − 𝑐)𝑗 (E2)
Where and are the Lagrange multipliers. From the necessary and sufficient conditions
of optimality follows that an optimum x* can be found if the partial derivatives of the
Lagrangian function with respect to the decision variables x and the Lagrange multipliers
are set to zero and solved:
∇𝐿|𝑥∗ = 0 (E3)
The Langrange multiplier method can be demonstrated with a small numerical example.
128
Appendices
Suppose we have following optimization model:
max 𝑓(𝑥) = 𝑥12 + 𝑥2
2
s.t.
𝑔(𝑥) = 2𝑥1 − 5𝑥2 = 0 (P2)
ℎ(𝑥) = 𝑥1 + 𝑥2 = 3
𝑥1, 𝑥2 > 0
We first define the Lagrangian function according to E2:
𝐿 = (𝑥12 + 𝑥2
2) + 𝜆1(2𝑥1 − 5𝑥2) + 𝜆2(𝑥1 + 𝑥2 − 3) (E4)
Now we set the partial derivatives with respect to x and the multipliers to zero:
𝜕𝐿
𝜕𝑥1= 0 = 2𝑥1 + 2𝜆1 + 𝜆2 (E5)
𝜕𝐿
𝜕𝑥2= 0 = 2𝑥2 − 5𝜆1 + 𝜆2 (E6)
𝜕𝐿
𝜕𝜆1= 0 = 2𝑥1 − 5𝑥2 (E7)
𝜕𝐿
𝜕𝜆2= 0 = 𝑥1 + 𝑥2 − 3 (E8)
In this case equations 5 to 8 form a linear system of four equations with four unknowns that
can be solved with for example Gaussian elimination. The optimum is located at x1=2.1429,
x2=0.8571, 1=-0.3673, 2=-3.5510. The objective value at optimum is f=5.3265.
129
Acknowledgements
Acknowledgements
After four years, my PhD journey is approaching the end. For me this was a great adventure
where I met lot of people, I travelled and I learned something about engineering and food.
But all of this would not be possible without amazing people who helped me through all
these years.
First of all I would like to thank Prof. Peter Bongers†. I met him when I was a post master
(PPD) student. You selected me among other students to do the second year project in
Unilever and after six months you offered me to continue with a PhD. Unfortunately, today
you are not among us to see the final result of the project. It was a great pleasure to work
with you.
I would like to thank professor Jan Meuldijk for all the support and understanding I needed
when I moved to his group in 2012. His extensive knowledge as well as supervision helped
me to bring this project till the end.
Especially I would like to thank my daily supervisor Edwin Zondervan. At the beginning
of my PhD, it was not clear who would be my daily supervisor. The decision was made.
Peter Bongers decided that you will be my daily supervisor. It was really a great pleasure
working with you all these years. From all people within the project, you are the only one
that stayed with me till the end. I truly thank you for all meetings, advices, discussions…
Dear Edwin, thank you for pushing me forward and supporting me in difficult moments in
my life.
This project was supported by the Institute for Sustainable Process Technology (ISPT). It
was nice to be part of this consortium. In particular, I would like to thank Frans van den
Akker and Daniella Vrijling.
In addition, I would like to thank Unilever for hosting me in the first three years of the
project. During my PhD I had two nice ladies as my project leaders, Olivera Trifunovic and
Nasim Hooshyar. Also, a word of appreciation to Ardjan Krijgsman, Cristhian Almeida-
Rivera, Hilde Wijngaard, and Hans Hoogland for the valuable input as well as pleasant time
that we had together.
I greatly acknowledge the members of my doctoral committee: prof.dr.ir. M.C. Kroon,
prof.dr.-ing. G. Schembecker, prof.dr.ing. M.H.M. Eppink, dr. O. Trifunovic and ir. G.D.
Mooiweer. Not only for taking part in this committee, but also for the feedback on this
thesis.
130
Acknowledgements
I would also like to thank the other industrial project members Reinoud Noordman from
Heineken, Edwin Poiesz and Tjerk van Mil from Cosun. In addition, prof. Harry Gruppen,
Jean-Paul Vincken, Atze Jan van den Goot, Peter Wierenga from Wageningen University
for the nice discussions and valuable feedback during our project meetings. In addition a
warm word of thanks to my project ladies Alexandra, Annewieke and Laura. Dear ladies, it
was really a pleasure to work and to travel with you.
My special and huge thanks to all my students: Carla, Tugba and Feipeng for your hard
work and contribution to this thesis. I enjoyed working with you guys.
I would like to thank all members from the Chemical Reactor Engineering (SCR) group:
Emila, Dulce, Paola, Shohreh, Lara, Violeta, Carlos, Michiel, Slavisa, Lana, Vladan. I also
thank Denise for help throughout my PhD. A big thank to my officemates: Miguelito,
Tom, Martijn and Arend. Thank you guys for the pleasant trips, coffee breaks, biertjes,
sports and inburgering …
To all the students of the PPD program for all the parties, coffee breaks, drinks in the
FORT and the Zwarte Doos. A big thank to Leontien for all help and support in the last four
years. Big thank to my Serbian crew Dragana and Jovana (mojim Cucama). Ladies you
make my life even more crazy
I would like to thank all my friends in Serbia: Nevena, Smiljka, Irena, Vojkan, Marija,
tetka Ljilja i cika Momir, kumovi Jokici. Hvala na podršci i razumevanju Not to forget
Vucicevic in USA
At the end I would like to thank my family for all support and understanding in the last four
years. I thank my sister Jelena, my nephew Bogdan and brother in law Dusan. Without you
at my side, all of this would not have been possible. Mati, još jedna stepenica, i još jedna
diploma. Kao što ti meni uvek kažeš “najkraćim putem do pobede”.
131
List of publications
List of publications
Journal Publications
Zderic, A. Mastwijk, H., Zondervan, E., Meuldijk, J. A study of mechanism involved
during polyphenol extraction from fresh tea leaves by pulsed electric field (2015),
submitted to Biosystems Engineering
Zderic, A., Meuldijk, J., Zondervan, E. Product-driven process synthesis for the extraction
of polyphenols from fresh tea leaves (2015), submitted to Industrial and Engineering
Chemistry Research
Zderic, A., Araya-Cluotier, C., Zondervan, E., Meuldijk, J. Isolation of oil bodies from
soybeans in a mild way: definition of operating window for process design (2015),
submitted to Applied Biochemistry and Biotechnology Journal
Zderic, A., Tarakci, T., Almeida-Rivera, C., Meuldijk, J., Zondervan, E. Product-driven
process synthesis for the extraction of oil bodies from soybeans (2015), submitted to American Institute of Chemical Engineers Journal
Peer Reviewed Conference Proceedings
Zderic, A., Zondervan, E., Meuldijk, J. Breakage of Cellular Tissue by Pulsed Electric
Field: Extraction of Polyphenols from Fresh Tea Leaves, Proceeding of the 11th
International Conference on Chemical and Process Engineering (ICheaP 11), 2-5 June
2013, Milan, Italy, pp. 1795-1800
Zderic, A., Tarakci, T., Hooshyar, N., Zondervan, E., Meuldijk, J. Process Design for
Extraction of Soybean Oil Bodies by Applying the Product Driven Process Synthesis
Methodology, Proceeding of the 24th
European Symposium on Computer Aided Process
Engineering (ESCAPE 24), 15-18 June 2014, Budapest, Hungary, pp. 193–198
Zderic, A., Meuldijk, J., Zondervan, E. Product-driven process synthesis for the extraction
of polyphenols from fresh tea leaves, Proceeding of the 12th
International Conference on
Chemical and Process Engineering (ICheaP 12), 19-22 May, 2015, Milan, Italy
132
List of publications
Oral Presentations
Zderic, A., Zondervan, E., Meuldijk, J. Breakage of Cellular Tissue by Pulsed Electric
Field: Extraction of Polyphenols from Fresh Tea Leaves, 11th
International Conference on
Chemical and Process Engineering (ICheaP 11), 2-5 June 2013, Milan, Italy
Zderic, A., Zondervan, E., Meuldijk, J. Extraction of protein in a mild way, Computer
Aided Process Engineering (CAPE Forum), 12-14 May 2014, Milan, Italy
Poster Presentations
Zderic, A., Zondervan, E., Trifunovic, O., Bongers, P. Selective opening and fractionation
of the natural raw material, Netherlands Process Technology Symposium (NPS), 24-26
October 2011, Papendal, The Netherlands
Zderic, A., Zondervan, E., Meuldijk, J. Statistical analysis of data from pulsed electric field
tests to extract polyphenols, 9th
European Congress of Chemical Engineering (ECCE9),
21-25 April 2013, The Hague, The Netherlands
Zderic, A., Tarakci, T., Hooshyar, N., Zondervan, E., Meuldijk, J. Process Design for
Extraction of Soybean Oil Bodies by Applying the Product Driven Process Synthesis
Methodology, 24th
European Symposium on Computer Aided Process Engineering
(ESCAPE 24), 15-18 June 2014, Budapest, Hungary
Zderic, A., Zondervan, E., Meuldijk, J. Pulsed electric field as cell opening method:
extraction of polyphenols from fresh tea leaves, Netherlands Process Technology
Symposium (NPS), 3-5 November 2014, Maarsen, The Netherlands
Zderic, A., Meuldijk, J., Zondervan, E. Product-driven process synthesis for the extraction
of polyphenols from fresh tea leaves, Proceeding of the 12th
International Conference on
Chemical and Process Engineering (ICheaP 12), 19-22 May 2015, Milan, Italy
133
Curriculum vitae
Curriculum vitae
Aleksandra Žderić was born on 10 July 1981 in Šabac, Serbia. After finishing secondary
school in 2001, she started Chemical Engineering studies in the Faculty of Technology and
Metallurgy, University of Belgrade. In 2007, she did her graduation project (6 months) at
the Eindhoven University, The Netherlands in the Process Systems Engineering group
under the supervision of prof.dr.ir. Andre de Haan. In 2008 she obtained her Master’s
degree in Chemical Engineering from Belgrade University. Her Master thesis was in the
field of separation technology on the project “Strategies for host-guest extraction of
Immunoglobulin G”. After her Master studies in 2009, she enrolled in the two years Post-
Master program “Process and Product Design”. In 2010 she started the second year project
with Unilever R&D Vlaardingen under the supervision of prof.dr.ir. Peter Bongers and
prof.dr. Jan Meuldijk. She wrote her thesis on “New routes for liquid tea extraction”. In
March 2011 she received her PDEng diploma from Eindhoven University of Technology.
In April 2011, she started her PhD in the field of process and product design on the project
“Selective opening and fractionation of natural raw materials” at the Eindhoven University
of Technology in the group of prof.dr.ir. Peter Bongers and dr.ir. Edwin Zondervan. Since
2012 she moved to the Polymer Reaction Engineering group under the supervision of
prof.dr. Jan Meuldijk and dr.ir. Edwin Zondervan. Her project was sponsored by the
Institute for Sustainable Process Technology (ISPT) in cooperation with several industrial
partners such as Unilever, Heineken, Cosun, DSM and Synthon and her work during this
PhD project led to this thesis.