ww.sciencedirect.com

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 2 2 9e3 2 3 9

Available online at w

journal homepage: www.elsevier .com/locate/watres

Impact of salinity and pH on the UVC/H2O2 treatment ofreverse osmosis concentrate produced from municipalwastewater reclamation

Kai Liu, Felicity A. Roddick*, Linhua Fan

School of Civil, Environmental and Chemical Engineering, RMIT University, GPO Box 2476, Melbourne, Victoria 3001, Australia

a r t i c l e i n f o

Article history:

Received 27 September 2011

Received in revised form

19 February 2012

Accepted 10 March 2012

Available online 2 April 2012

Keywords:

Reverse osmosis concentrate

Organic pollutants

UVC/H2O2

pH

Salinity

Biodegradability

* Corresponding author. Tel.: þ61 3 9925 369E-mail address: felicity.roddick@rmit.edu

0043-1354/$ e see front matter ª 2012 Elsevdoi:10.1016/j.watres.2012.03.024

a b s t r a c t

While reverse osmosis (RO) technology is playing an increasingly important role in the

reclamation of municipal wastewater, safe disposal of the resulting RO concentrate (ROC),

which can have high levels of effluent organic pollutants, remains a challenge to the water

industry. The potential of UVC/H2O2 treatment for degrading the organic pollutants and

increasing their biodegradability has been demonstrated in several studies, and in this

work the impact of the water quality variables pH, salinity and initial organic concentra-

tion on the UVC/H2O2 (3 mM) treatment of a municipal ROC was investigated. The reduc-

tion in chemical oxygen demand and dissolved organic carbon was markedly faster and

greater under acidic conditions, and the treatment performance was apparently not

affected by salinity as increasing the ROC salinity 4-fold had only minimal impact on

organics reduction. The biodegradability of the ROC (as indicated by biodegradable dis-

solved organic carbon (BDOC) level) was at least doubled after 2 h UVC/H2O2 treatment

under various reaction conditions. However, the production of biodegradable intermedi-

ates was limited after 30 min treatment, which was associated with the depletion of the

conjugated compounds. Overall, more than 80% of the DOC was removed after 2 h UVC/

3 mM H2O2 treatment followed by biological treatment (BDOC test) for the ROC at pH 4e8.5

and electrical conductivity up to 11.16 mS/cm. However, shorter UV irradiation time gave

markedly higher energy efficiency (e.g., EE/O 50 kWh/m3 at 30 min (63% DOC removal) cf.

112 kWh/m3 at 2 h). No toxicity was detected for the treated ROC using Microtox� tests.

Although the trihalomethane formation potential increased after the UVC/H2O2 treatment,

it was reduced to below that of the raw ROC after the biological treatment.

ª 2012 Elsevier Ltd. All rights reserved.

1. Introduction secondary effluent, RO-based tertiary/advanced wastewater

Reverse osmosis (RO) technology has been used increasingly in

municipal wastewater reclamation over the past decade to

address freshwater shortages in many regions. Due to the

excellent performance of RO membranes in rejecting organic

and inorganic pollutants present in biologically treated

2; fax: þ61 3 9925 3746..au (F.A. Roddick).ier Ltd. All rights reserved

treatment processes can produce very high quality water

which is suitable for a wide range of reuse purposes. However,

the brine streams (also referred to as reverse osmosis concen-

trate (ROC) streams) generated from the RO systems may pose

major health and environmental risks if they are discharged to

the receiving environmentwithout appropriate treatment. The

.

Table 1 e Characteristics of the ROC.

Parameter Value Ions mg/L

�

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 2 2 9e3 2 3 93230

risks are due to the ROC containing almost all of the pollutants

in the original secondary effluent at elevated levels (i.e.,

commonly 4e6 times higher concentration). Depending on the

wastewater source, the organic pollutants in the ROC may be

toxic and/or bioaccumulative (Shon et al., 2006).

Advanced oxidation processes (AOPs) are regarded as an

effective means for degrading the organic matter in ROC. In

general, AOPs utilise the highly oxidising hydroxyl radical

(�OH) to break down organic matter into smaller (often more

biodegradable) molecules, and eventually to CO2. Several

AOPs, including UV/TiO2 (Westerhoff et al., 2009; Zhou et al.,

2010), UV/H2O2 (Liu et al., 2011; Bagastyo et al., 2011), ozone

(Zhou et al., 2010; Lee et al., 2009) and electrochemical oxida-

tion (Perez et al., 2010), have been demonstrated to be effective

for treating the ROC produced from the reclamation of

municipal secondary effluent. As H2O2 is very soluble inwater,

it can be used as an effective source for �OH production in the

presence of UV irradiation. Moreover, UV/H2O2 processes are

simple to design and operate (Buchanan et al., 2004). In most

cases, the partially oxidised intermediates have less inherent

environmental and human health risk (Westerhoff et al.,

2009). Although Zhou et al. (2010) reported that UV/H2O2

treatment was not very effective, they used UVA (365 nm) at

which wavelength H2O2 has a very low molecular extinction

coefficient (approximately 0.01 L mol�1 cm�1) compared with

at the more commonly used UVC (254 nm, 19.6 L mol�1 cm�1)

(H2O2.com, 2009). In our recent study using a UVC/H2O2

system to treat an ROC produced from a municipal secondary

effluent, the potential for decreasing the concentration of

organic contaminants and increasing the biodegradability of

the ROC was demonstrated (Liu et al., 2011). Furthermore,

Bagastyo et al. (2011) reported that compared with other

treatments including alum and ferric coagulation, and ion

exchange, UVC/H2O2 was the most efficient treatment for the

organic content of a municipal ROC.

As the pH, salinity and organic pollutant level of municipal

ROC can vary with source, season or treatment method, the

aim of this work was to investigate the effects of these water

quality variables on the efficiency of UVC/H2O2 treatment. The

treatment performance was characterised using COD, DOC,

A254, colour and fluorescence excitationeemission matrix

(EEM) spectroscopy. Size exclusion chromatography using

liquid chromatography with organic carbon detection

(LC-OCD) was employed to determine the molecular size

changes during the treatment. The biodegradability of the

ROC before and after the treatment was determined as

biodegradable dissolved organic carbon (BDOC). An indication

of the potential toxicity of the treated ROC was obtained

through Microtox� assay and trihalomethane formation

potential (THMFP) measurement.

DOC (mg/L) 21 Cl 780

COD (mg/L) 65 PO3�4 38

A254 (/cm) 0.41 SO2�4 233

Colour (PteCo mg/L) 88 NO�3 35

pH 8.5 Naþ 529

TDS (mg/L) 1685 Mg2þ 56

Electrical Conductivity

(mS/cm)

2.82 Kþ 66

BDOC (mg/L) 3.2 Ca2þ 68

Alkalinity (as CaCO3, mg/L) 295 Fe3þ 8

Zn2þ 20

2. Materials and methods

2.1. Source of wastewater and preparation of ROC

A biologically treated municipal wastewater from a local

wastewater treatment plant was used for the preparation of

the ROC. To prepare the ROC, the wastewater was subjected to

microfiltration (Microza�, Part No. UMP-153, PALL) followed by

reverse osmosis using a Sepa cell crossflow RO module

(GE-Osmonics, Minnetonka, MN) with a commercial poly-

amide membrane (AG; GE-Osmonics, Minnetonka, MN). The

characteristics of the resultant ROC are presented in Table 1.

H2SO4 (1 M) and NaOH (1 M) were used for adjusting the pH of

the ROC; NaCl (Analytical Reagent Grade) and MgSO4 (BDH

Chemicals, General Purpose Reagent) were used for the

adjustment of its salinity.

2.2. UV irradiation experiments

Irradiation was conducted using an annular reactor with

a centrally mounted lamp. The ROC was dosed with 3 mM

H2O2 (Australian Chemical Reagents, 50% w/w) which was

found to be the optimum dosage in our previous study

(Liu et al., 2011), aerated by humidified air during irradiation

and sampled periodically. The average irradiated area was

464 cm2 with a pathlength of 1.94 cm, other UV reactor

conditions are reported elsewhere (Thomson et al., 2004). The

UVC lamp emitted at 254 nm, and was manufactured by

Australian Ultra Violet Services (G36T15NU, energy input

39 W). H2O2 actinometry (Beltran et al., 1995) was used for

measuring the intensity of the UVC lamp, and the average

fluence rate of the lamp was determined as 12.89 mJ/s/cm2.

Duplicate experiments were undertaken and average results

reported. The enzyme catalase (from Aspergillus niger,

Calbiochem�) was used to decompose the residual H2O2 and

so remove its interference in water quality measurement. To

every 20 mL sample, 8 mL (activity of 16 units) of the catalase

was added and the sample was shaken at 100 rpm until H2O2

was less than 0.5 mg/L, which is considered negligible (Kang

et al., 1999). The resultant increase in COD and DOC due to

the added catalase was determined as <1 mg/L for COD and

w0.05 mg/L for DOC.

2.3. Analytical methods

Samples were filtered (0.45 mmcellulose acetate, ADVANTEC�)

prior to the following analyses. A Sievers 5310 TOC analyser

with an auto-sampler and an inorganic carbon removal

module (Sievers 900 ICR; GE, Boulder, Co) was used for DOC

measurement. Inorganic carbon was measured by the same

TOC analyser without utilising the inorganic carbon removal

module. The COD was determined with Hach Method 8000

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 2 2 9e3 2 3 9 3231

using a Hach spectrophotometer (DR/4000U). The absorbance

at 254 nm (A254) was measured using a Unicam UV/vis spec-

trophotometer. The true colour of the samples was measured

in PteCo units at 455 nmusing a Hach spectrophotometer (DR/

4000U). The pH and conductivity were measured using a Hach

Sension 156 pH/conductivity meter. The concentration of

hydrogen peroxide was measured using Merckoquant�

peroxide test sticks. An ion chromatograph (Dionex, 2010i

with Ionpac A54A-SC column, 4 � 250 mm) was used for

analysing the anion content of the ROC.

The EEM spectra of the ROC samples were obtained with

a PerkinElmer LS55 fluorescence spectrometer after adjust-

ment of the DOC to 7 mg/L. The results were processed by FL

Winlab software (PerkinElmer Applications). An add-on soft-

ware (3D Exporter) was used for exporting the 3D EEM data,

which were used for the calculation of the EEM volumes by

mathematical integration using MS Excel. Molecular size

distribution was determined using LC-OCD at the Water

Research Centre of the University of New South Wales,

Australia. The LC-OCD system (LC-OCD Model 8, DOC-Labor

Dr. Huber, Germany) utilised a size exclusion chromatog-

raphy column (Toyopearl TSK HW-50S, diameter 2 cm, length

25 cm) and the chromatograms were processed using the

Labview based program Fiffikus (DOC-Labor Dr. Huber,

Germany). BDOC (Joret and Levi, 1986) was used to determine

the biodegradability of the organic content in the variously

treated samples. BDOC measurement involved the exposure

of the samples to thoroughly washed biologically active sand

for five days under aerobic conditions. The DOC was

measured daily and the BDOCwas calculated as the difference

of the initial DOC and the lowest DOC recorded over a five-day

period. Variation in BDOC determinations was �1.2%.

Trihalomethane formation potential (THMFP) and Micro-

tox� analyses were conducted by Analytic Chemistry and

0

0.2

0.4

0.6

0.8

1

Time (min)

No

rm

alis

ed

C

OD

pH 10Original pH (8.5)pH 6pH 4

a b

0

0.2

0.4

0.6

0.8

1

Time (min)

No

rm

alis

ed

A

25

4

c

d

0 30 60 90 120

No

rm

alis

ed

D

OC

0 30 60 90 120

pH

Fig. 1 e (a) COD, (b) DOC, (c) A254 reduction at different pH condi

Testing Services (Melbourne). For THMFP analysis, the sample

was chlorinated according to standard methods 5710B (Eaton

and Franson, 2005). The samplewas then added directly to the

purge and trap, where the volatiles were purged and concen-

trated before being analysed by GC/MS. Ecotoxicity assess-

ment was performed using Microtox� tests. The tests, which

employ the luminescent marine bacterium Vibrio fischeri, were

conducted according to the protocol provided with the

Microtox 500 Analyser.

3. Results and discussion

3.1. Effect of pH

The degradation of the organic pollutants in the ROC was

enhanced with decreasing pH (Fig. 1a and b). A similar trend

was observed by Zhou et al. (2010) where onlyw2% of the DOC

(initially 18 mg/L) was removed by UVA/H2O2 treatment of

a municipal ROC at pH 6.9 whereas the removal was greatly

improved to 17% at pH 5. As indicated by its alkalinity, the ROC

had a high HCO�3 =CO

2�3 content (Table 1). Bicarbonate/

carbonate species are strong �OH scavengers as shown in

Equations (1) and (2), and the resultant CO��

3 has a much lower

oxidation potential and a higher selectivity in its reactionwith

organic compounds compared with �OH (Liao et al., 2001). At

pH 10, HCO�3 =CO

2�3 species exist in approximately the same

proportions, whereas at the original pH (8.5) HCO�3 exists

almost exclusively (Oppenlander, 2003). Buxton and Elliot

(1986) reported that CO2�3 reacts with �OH approximately two

magnitudes faster than does HCO�3 , as indicated by their

reaction constants Equations (1) and (2). Therefore, the higher

concentration of CO2�3 led to less organic degradation at pH 10

compared with pH 8.5. In addition, another radical scavenger

0

0.2

0.4

0.6

0.8

1

0 30 60 90 120Time (min)

4

5

6

7

8

9

10

0 30 60 90 120

Time (min)

tions, (d) pH change during the UVC/3 mM H2O2 treatment.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 2 2 9e3 2 3 93232

HCO�2 , which would be generated in the alkaline conditions

(Equation (3)) (AlHamedi et al., 2009), may have contributed to

further reduction in process performance. However, the COD

degradation rates for the first 30 min of irradiation were

similar at pH 6e10. This was attributed to the initially high

concentration of H2O2, and thus �OH, which was sufficiently

high over this period that scavenging did not greatly affect the

degradation rate, combined with the presence of some fast-

reacting organics.

CO2�3 þ�

OH/CO��

3 þH2O k ¼ 4:2� 108dm3mol�1s�1 (1)

HCO�3 þ�

OH/CO��

3 þOH� k ¼ 5:8� 106dm3mol�1s�1 (2)

H2O2 þOH�/HO�2 þ�

OH/H2OþO�2 (3)

�OHþ Cl�/HOCl� k ¼ 4:3� 109dm3mol�1s�1 (4)

According to Oppenlander (2003), further decreasing the pH

to 6 would lead to the transformation of approximately 50% of

the HCO�3 to H2CO3, which has a very low reactivity with �OH

(Liao et al., 2001). Hence the scavenging effect by the

HCO�3 =CO

2�3 species was reduced at pH 6, resulting in

enhanced organic degradation. The main inorganic carbon

component at pH 4 is dissolved CO2, only 0.1% of which reacts

with water to form H2CO3 (Oppenlander, 2003). Thus, the

scavenging effect caused by the HCO�3 =CO

2�3 species was

almost completely eliminated at pH 4, leading to better

oxidation performance than at pH 6 for up to 90 min.

Nevertheless, at pH 4 there was no further decrease in COD

and DOC after 90 min; this was attributed to the presence of

organics recalcitrant to the conditions used. At all pH levels,

reduction in A254 was significantly faster than of COD and

DOC. This suggested that �OH preferentially attacked the

organic molecules contributing to A254, i.e., conjugated bonds,

which was consistent with the observation by Westerhoff

et al. (2009).

Solution pH was monitored during the treatment for

a better understanding of the process. As shown in Fig. 1d,

when the ROC at original pH was treated, the pH initially

dropped from 8.5 to 7.7 and then gradually increased to 8.4.

The decrease in pH can be explained by the oxidation of the

organics to mineral acids, carbon dioxide and their acidic

intermediates (Chin et al., 2009), and the subsequent pH

increasemay be attributed to eventualmineralisation of these

acidic intermediates. For the processes at pH 4 and pH 6 the

pH increased with irradiation time; this may have been due to

the combination of Hþ with CO2 resulting from the minerali-

sation of the organics to formH2CO3 or HCO�3 (Zouboulis et al.,

2007). The small increase for pH 4 was attributed to the two-

magnitude higher concentration of Hþ at pH 4 than at pH 6.

The process at pH 10 showed a slower and smaller decrease in

pH than at the original pH. This was considered to be mainly

due to the 1.5-magnitude higher concentration of OH� at pH

10 than at pH 8.5.

Fluorescence regional integration (FRI) (Chen et al., 2003)

was used to quantify the proportion of each class of the fluo-

rescent organic species, namely, aromatic proteins (AR I & II),

fulvic acid-like substances (FA), soluble microbial products

(SMP) and humic acid-like substances (HA), during the UVC/

H2O2 treatment. The classification of the fluorescent organics

is operationally defined as described by Chen et al. (2003).

Fig. 2a shows the changes in fluorescence during 2 h UVC/

3mMH2O2 treatment at the original pHafter adjustment of the

DOC of each sample to 7 mg/L (to avoid the quenching which

can occurwhenDOCexceeds 10mg/L). Thefluorescent species

in the ROC consisted of twomajor groups, 57% humic acid-like

and 29% fulvic acid-like substances. The organic molecules in

both groups were broken down rapidly so that 72% of fulvic

acid-like and 64% of humic acid-like substances were broken

down after only 10 min treatment. After 30 min, 93% of the

total fluorescence was removed and then it gradually pla-

teaued. Taking into account the effect of pH on the fluores-

cence (Fig. 1, Supplementary Information), higher pH led to

slightly lower removal of fluorescence at 20 min and this was

mainly due to the lower reductionof fluorescence in FA species

(Fig. 2b).TheEEMvolume reduction for theHA fractionwas less

than for the FA fraction or total ROC organics. This may indi-

cate that the FA molecules were more susceptible to the AOP

treatment compared with the HA molecules under the exper-

imental conditions. Therewas a small amount of precipitation

during the pH adjustment process (pH 8.5e10) which may be

due to the removal of some higher MW HA molecules,

consistent with the EEM volume for HA being lower at pH 10

than at pH 6 and 8.5. Collectively, at least 85% of the fluores-

cence was removed after 20 min, indicating that fluorescent

species can be quickly destroyed over the pH range tested.

TheuntreatedROChada lowbiodegradability of 13% (Fig. 3).

DOC removal of 66% was achieved after 2 h UVC/H2O2 treat-

ment at pH 4 and the BDOC increased to 24%, accounting for

68% of the remaining DOC. At pH 10, mineralisation by the

UVC/H2O2 process was the lowest (31%) but DOC removal by

BDOC was the highest (38%). This was attributed to less

participation of �OH in thedegradation of the organics at higher

pH due to scavenging by the HCO�3 =CO

2�3 species resulting in

more biodegradable intermediates remaining after 2 h irradi-

ation. The total DOC removal (i.e., after both treatments) was

increased with decreasing pH due to the alleviation of scav-

enging by the HCO�3 =CO

2�3 species. However, there was only

a little improvement in theoverall reduction ofDOCat pH6and

4 due to the remaining organic pollutants being recalcitrant to

the reaction conditions used. Westerhoff et al. (2009) made

a similar observation in the oxidation of amunicipal ROCusing

the UVC/TiO2 process and suggested that there was no addi-

tional benefit in using pH � 5. The results imply that the ROCs

from different sources would need to be tested to ensure that

the concentration of the radical scavengers is sufficiently low

for the AOP process to be feasible.

3.2. Effect of salinity

For investigating the effect of salinity, two of the major

constituent salts were added to adjust the salinity of the ROC

(i.e., NaCl and MgSO4 at a molar ratio of 6:1, as found in the

original ROC). Experiments were conducted at four different

salinity levels: 1.5, 2, 3 and 4 times the original salinity cor-

responding to the EC values of 4.45, 5.90, 8.16 and 11.16 mS/

cm, respectively. The addition of the salts led to only a minor

decrease in pH (maximum of 0.2 units), and little impact on

Fig. 2 e (a) EEM volumes during 2 h UVC/3mMH2O2 treatment at original pH (8.5). (b) EEM volumes after 20min of UVC/3mM

H2O2 treatment at different pH. (API & II: aromatic protein I&II; FA: fulvic acid-like; SMP: soluble microbial products; HA:

humic acid-like.)

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 2 2 9e3 2 3 9 3233

the EEM volumes (Fig. 2, Supplementary Information). The

reduction in COD, DOC and A254 for the ROC samples at

elevated salinity was fairly similar to that for the ROC at the

original salinity (Fig. 4). Furthermore, the FRI results showed

Fig. 3 e DOC reduction after 2 h UVC/3 mM H2O2 treatment

followed by BDOC at different pH.

that the EEM volumes after 20 min treatment at the original

salinity were comparable to those at elevated salinity (Fig. 5).

Therefore, it appeared that salinity was not amajor influential

factor in the UVC/H2O2 treatment of the organic pollutants in

the ROC. The processes at elevated salinity led to slightly

lower DOC removal by biodegradation than at the original

salinity (Fig. 6); this may be attributed to the dehydration of

the bacteria at the higher salinity (Garcıa and Hernandez,

1996).

Overall, the final DOC removal at elevated salinity was

similar to that at the original salinity, providing further

evidence that the UVC/H2O2 process was not susceptible to

salt concentration over the tested range. This may also

suggest that the presence of other ions (Naþ, Mg2þ, Cl�, SO2�4 )

did not greatly negatively impact the oxidation process.

Therefore, the UVC/H2O2 process is not greatly influenced by

the salinity, indicating its applicability for treating various

types of municipal ROCs or even ROCs from brackish sources.

3.3. Effect of initial DOC concentration

The loading of organic pollutants in the ROC may fluctuate

depending on many factors such as catchment, water

recovery and season. Consequently the effect of initial organic

pollutant concentration, in terms of DOC, on the efficiency of

0

0.2

0.4

0.6

0.8

1

0 30 60 90 120Time (min)

No

rm

alis

ed

C

OD

4S3S2S1.5S1S

ba

0

0.2

0.4

0.6

0.8

1

0 30 60 90 120

Time (min)

No

rm

alis

ed

A

25

4c

0

0.2

0.4

0.6

0.8

1

0 30 60 90 120Time (min)

No

rm

alis

ed

D

OC

Fig. 4 e (a) COD, (b) DOC and (c) A254 reduction after UVC/3 mM H2O2 treatment at different salinity levels (denoted as 1d4 S,

where 1 S indicates original salinity, 4 S four times original salinity).

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 2 2 9e3 2 3 93234

the UVC/H2O2 process was investigated. ROCswith initial DOC

concentrations of 21 (original), 26 and 30 mg/L, denoted as

D21, D26 and D30, respectively, were used. The ROCs with

higher DOC (i.e., D26 and D30) were prepared using the same

method as described in Section 2.1. The DOC concentrations

formostmunicipal ROCs reported in the literature (Zhou et al.,

2010; Bagastyo et al., 2011; Lee et al., 2009; Perez et al., 2010) fall

within this range. Similar reduction trends were obtained for

normalised COD, DOC and A254. At first glance, it would seem

that initial organic concentration did not affect the efficiency

of the process. This, of course, indicates that there was

Fig. 5 e EEM volumes after 20 min of UVC/3 mM H2O2 treatmen

indicates original salinity, 4 S four times original salinity).

a higher net degradation in absolute terms with increasing

initial DOC level (Table 2).

The BDOC increased with increasing initial DOC level

(5.9 mg/L for D21, 6.8 mg/L for D26 and 9 mg/L for D30). The

percentage of the DOC removed by the UVC/H2O2-BDOC

treatment was comparable for the three ROCs (Fig. 7),

meaning higher net reduction in DOC with increasing initial

DOC concentration. It seems that the UVC/H2O2-BDOC treat-

ment performance was not affected over the range of initial

DOC concentrations tested in this work. On the other hand,

increasing initial DOC concentration also increased the

t at different salinity levels (denoted as 1e4 S, where 1 S

5661 60 58 59

13

2821 22 23 21

0%

25%

50%

75%

100%

RO

C

1S

1.5S

2S

3S

4S

DO

C R

ed

uctio

n (%

)

After UV/H2O2 After BDOC Remaining

Fig. 6 e DOC reduction after 2 h UVC/3 mM H2O2 treatment

followed by BDOC at different salinity levels (denoted as

1e4 S, where 1 S indicates original salinity, 4 S four times

original salinity).

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 2 2 9e3 2 3 9 3235

residual DOC of the treated ROC, which may be a problem for

its disposal or reuse. Nevertheless, over the DOC range tested

(i.e., 20e30 mg/L), the residual DOC after the UVC/H2O2-BDOC

treatment was still lower than in the secondary effluent

(9.6 mg/L), i.e., 3.4 mg/L for D21, 4.6 mg/L for D26 and 5.4 mg/L

for D30.

3.4. Impact of irradiation time on biodegradability

To investigate the development of biodegradability during the

UVC/H2O2 process, the BDOC of the treated sample at original

pH was measured periodically over 120 min. The UVC/H2O2

process mineralised only 10% of the DOC but almost doubled

the BDOC after 10 min (Fig. 8a). Increasing irradiation time (to

120min) led to greater DOC removal by the UVC/H2O2 process;

however, the increase for the second hour (10%) wasmarkedly

lower than for the first hour (42%). The biodegradable fraction

reached a maximum value of 35% at 30 min and then slowly

decreased to 30% at 120 min Westerhoff et al. (2009) observed

a similar trend for BDOC when they treated a municipal ROC

using aUVC/TiO2 process. In theirwork, the BDOC increased to

amaximumvalue of 30% and then decreased slowly to the end

of the treatment. This was because the biodegradable organic

fraction resulting from the UV-mediated process mainly

Table 2 e Net and percentage reduction in various water paramconcentrations.

Parameters D21

Initial value Net reduction (%) Initial value

COD (mg/L) 65 36 55 76

DOC (mg/L) 21 11.8 56 26

A254 (/cm) 0.42 0.38 91 0.53

consisted of simple organic acids which were mineralised

slowly by �OH (Westerhoff et al., 2009). Overall, the total DOC

removal increased at a markedly lower rate after 30 min UVC/

H2O2 treatment, with a further increase of 19% by 120 min.

It was observed that A254 was inversely correlated

(R2 ¼ 0.99) with the biodegradability of the remaining organics

(i.e., percentage of BDOC in the remaining DOC after the UVC/

H2O2 treatment, denoted as BDOCR) (Fig. 8b). A similar obser-

vation was made by Buchanan et al. (2004) when they treated

natural organic matter in drinking water by UVC irradiation.

The total EEM volume was also correlated to BDOCR (R2 ¼ 0.87)

(Fig. 8c). These relationships suggest that the conjugated and

fluorescent compounds were the principal source for the

production of the biodegradable products. This explains why

the production of BDOC was limited after 30 min because by

then most of these compounds had been degraded as indi-

cated by the reduction in A254 (Fig. 1c).

The molecular size distribution of the organics during the

UVC/H2O2 process was determined by LC-OCD to provide

additional information about the biodegradability. Table 3

summarises the molecular size distribution of the organics

in the ROC after the UVC/H2O2 and UVC/H2O2-BDOC processes

at different irradiation time.

The UVC/H2O2 process led to some decrease in the large

compounds after 15 min, 13% for biopolymers and only 3% for

humics. The removal of small compounds after 15 min was

greater, 21% for building blocks and 33% for LMW neutrals.

This was initially surprising because �OH preferentially

attacks large molecules (Atkinson et al., 1979). However, the

specific UV absorbance (SUVA) for humics was significantly

reduced from 2.79 to 1.23 L/mgm after 15 min. SUVA is a good

surrogate measure for the aromaticity of humic substances

(Edzwald and Benschoten, 1990). Therefore, given the short

irradiation time of 15 min, it is suggested that �OH destroyed

the conjugated bonds of the humics via electrophilic addition

because this is usually the first step for oxidation of conju-

gated organics (Legrini et al., 1993). The further breakdown of

the humics to smaller products required longer irradiation

time. This was confirmed by the markedly faster reduction in

humics between 30 and 120 min where they were reduced

from 8.13 to 1.46mg/L, corresponding to a removal of 85%, this

was also accompanied by decrease in aromaticity. The

biopolymers followed a similar trend as for humics and 70%

was removed after 120 min UVC/H2O2 treatment.

The reduction in building blocks, usually the fragments

from humics, was slower over 0e30 min compared with

30e120 min due to the continuous production of some

building blocks from fragmentation of the large compounds.

After 120 min, a removal of 80% was obtained for building

eters after 2 h UVC/3 mM H2O2 treatment at different DOC

D26 D30

Net reduction (%) Initial value Net reduction (%)

41 54 91 50 55

14.6 56 30 15.5 52

0.48 91 0.62 0.56 91

5652

28 2630

56

0%

25%

50%

75%

100%

D21

D26

D30

DO

C red

uctio

n (%

)

After UV/H2O2 After BDOC Remaining

Fig. 7 e DOC reduction after 2 h UVC/3 mM H2O2 followed

by BDOC at three different initial DOC levels.

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 2 2 9e3 2 3 93236

blocks. The LMW neutrals were greatly reduced from 7.14 to

4.17 mg/L in the first 30 min of the UVC/H2O2 process, after

which the concentration increased to 6.1 mg/L after 120 min,

due to the production of LMWneutrals from the breakdown of

the building blocks.

After 30 min UVC/H2O2 and subsequent BDOC treatment,

the biopolymers and humics were further reduced by 44% and

49%, respectively, which contributed to 56% of the BDOC

concentration; this was even higher than the contribution

10 1218

27

13

23

29

30

35

0%

20%

40%

60%

80%

100%

0 10 15 20 30Irradiation

DO

C R

ed

uc

tio

n (%

)

After UV/H2O2 Af

R2

= 0.99

0

0.1

0.2

0.3

BDOCR (%)

A2

54

1 5 3 0 4 5 6 0 75

a

b

Fig. 8 e (a) DOC reduction after UVC/3 mM H2O2 followed by BDO

and BDOCR. (c) Relationship between normalised total EEM volu

from the LMW fractions (building blocks þ LMW neutrals). A

similar trend was observed for 75 min UVC/H2O2-BDOCwhere

the removals were 60% and 57% for biopolymers and humics,

respectively, corresponding to 48% of the BDOC. This was

unexpected because these large humic compounds are

usually considered to be non-biodegradable. The aromaticity

of the humics after BDOC increased from 0.86 to 1.32 L/mg$m

for 30 min, and from 0.93 to 1.52 L/mg$m for 75 min UVC/H2O2

treatment. This helps to explain this phenomenon. Humics

with a high aliphatic content seem to be more accessible for

bacteria than those with a high degree of conjugation

(Tranvik, 1998). As the conjugated bonds of the humics in the

ROC were degraded by �OH, the structures of some were

transformed to be mainly aliphatic. Consequently, the resul-

tant aliphatic humics were degraded by the micro-organisms

in the BDOC test; however, the conjugated humics were still

resistant to biodegradation and therefore remained after the

BDOC test, leading to increase in aromaticity as shown by

the SUVA value. This indicated that the loss of conjugation in

themolecular structure greatly enhanced the biodegradability

of the organics, and provides additional evidence that the few

conjugated compounds remaining after 30 min limited the

further production of BDOC. Furthermore, the DOC mineral-

ised after 30 min would be mainly biodegradable intermedi-

ates. Hence, the final DOC removal after UVC/H2O2-BDOC

treatment was not greatly improved with longer irradiation

time. This implies that the 2 h irradiation time can be greatly

shortened, since the conjugated bonds of the compounds in

the ROC were quickly destroyed by the UVC/H2O2 treatment.

3540 42 45

50 52

34

3333

3331

30

40 50 60 75 90 120 Time (min)

ter BDOC Remaining

R2

= 0.87

0

0.1

0.2

0.3

0.4

15 30 45 60 75

BDOCR (%)

No

rm

alis

ed

V

olu

me

c

C at different treatment time. (b) Relationship between A254

me and BDOCR (number of experiments [ 2).

Table 3 e Molecular size distribution of the ROC after UVC/3 mM H2O2 and BDOC treatment.

Biopolymers Humics (w1000 Da) Building blocks(300e500 Da) mg/L

LMW neutrals

(>>20,000 Da)mg/L

mg/L Aromaticity (SUVA-HS)L/mg$m

(<350 Da)mg/L

DOCmg/L

ROC 3.24 9.51 2.79 5.83 7.14 25.72

15 min UV/H2O2 2.82 9.23 1.23 4.58 4.75 21.38

30 min UV/H2O2 1.84 8.13 0.86 4.07 4.17 18.21

UV/H2O2-BDOC 1.03 4.15 1.32 2.37 2.14 9.69

75 min UV/H2O2 1.21 5.74 0.93 2.68 4.21 13.84

UV/H2O2-BDOC 0.49 2.49 1.52 1.17 1.35 5.5

120 min UV/H2O2 0.98 1.46 0.79 1.95 6.1 10.49

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 2 2 9e3 2 3 9 3237

3.5. Ecotoxicity and trihalomethane formation potential

As the UVC/3 mM H2O2 process involved radical-mediated

reactions and led to major changes in the chemical proper-

ties of the original organic compounds in the ROC, it was

necessary to determine the possible formation of toxic prod-

ucts. TheMicrotox� test indicated that the untreated ROCwas

non-toxic to the microorganism used in the test, Vibrio fisheri.

Similarly, no toxicity was apparent for the ROC after 30 and

75 min treatment by UVC/3 mM H2O2, nor for these samples

after BDOC treatment. This demonstrated that no significant

concentration of toxic by-products occurred after either the

UVC/H2O2 or the UVC/H2O2-BDOC process.

Disinfection by chlorination is usually required before

recycled water is discharged or reused; this may lead

to formation of disinfection by-products (DBPs). These

compounds have potential adverse health effects, e.g., cause

cancer in humans (Hrudey, 2002). Thus, the trihalomethane

formation potential (THMFP) of selected samples was deter-

mined. The THMFP of the raw ROC was 1.22 mg/L, and

increased to 1.51 mg/L after 30 min of the UVC/H2O2 treat-

ment. The increase was attributed to the production of low

molecular weight THM precursor species from the fragmen-

tation of the complex compounds. Extending the irradiation

time to 75 min reduced the THMFP to its original level, which

was mainly due to the breakdown of the precursors and less

DOC being available to form more THM precursors (Kleiser

and Frimmel, 2000). Similarly, the THMFP was further

reduced after BDOC due to the decrease in DOC. Exposure to

R2 = 0.99

20

40

60

80

100

120

Time (min)

EE

/O

(k

Wh

/m

3

)

00 15 30 45 60 75 90 10 1205

a

Fig. 9 e EE/O values for the UVC/3 mM H2O2 treatment followed

concentration (number of experiments [ 2).

UVC/H2O2 for 30 min and 75 min followed by BDOC treatment

reduced the THMFP to 1.20 and 0.86mg/L, respectively. Longer

treatment would be required to reduce the THMFP to below

the permitted limit in drinking water (0.25 mg/L) (ADWG,

2004). However, this may be not necessary because the recy-

cled water may not be destined for potable use. Furthermore,

the values reported here are formation potentials and so the

actual THM levels would be much lower as the chlorine dose

in real chlorination processes is lower than those used in

THMFP analysis (Buchanan et al., 2006).

3.6. Preliminary energy consumption assessment

A figure-of-merit, electrical energy per order (EE/O), was used

for the preliminary assessment of the energy consumption of

the UVC/3 mM H2O2 process for initial DOC of 25 mg/L and pH

8.5. The definition of EE/O is described in Equation (5) (Bolton

et al., 2001). The energy consumption for the biological treat-

ment was not taken into account because it was considered to

be undertaken by a low energy process such as a constructed

wetland and so negligible compared with that for the UV

irradiation phase.

EE=O ¼ P� t� 1000

V� 60� log�Ci=Cf

� (5)

where P is the lamp power (0.039 kW), t is time (min), V is the

volume of irradiated sample (0.9 L), and Ci and Cf are the initial

and final DOC concentrations (mg/L).

0

20

40

60

80

100

120

05101520residual DOC (mg/L)

EE

/O

(k

Wh

/m

3

)

b

by BDOC with (a) irradiation time and (b) residual DOC

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 2 2 9e3 2 3 93238

The EE/O for the UVC/3 mM H2O2 followed by biological

treatment was approximately w50 kWh/m3 until 30 min and

then increased linearly (R2 ¼ 0.99) with reaction time to

120min (Fig. 9a). This showed that the energy efficiency of the

process started decreasing significantly after 30 min. Irradia-

tion time of 120min gave an EE/O value of 112 kWh/m3, which

was more than twice that for 30 min; however, the final DOC

removal for 120 min was only 20% higher than for 30 min

(63%). The EE/O for the UVC/H2O2 process alone was 160 kWh/

m3 at 10 min which then increased to 262 kWh/m3 after

120 min. This clearly demonstrated the great economic and

environmental benefits of coupling a biological treatment

with the UVC/H2O2 process.

For a target residual DOC of 10 mg/L, the corresponding

EE/Owas approximately 50 kWh/m3 (Fig. 9b). Further decrease

in residual DOC can be achieved at the expense of the energy

efficiency of the UVC/H2O2-BDOC process, leading to

increasing unit cost for the product water. Therefore, an

economic study is strongly recommended to establish a trade-

off between operational cost and customer requirements.

4. Conclusions

UVC/H2O2 advanced oxidation for ROC from municipal

secondary effluent was evaluated under different pH, salinity

and initial DOC conditions. The oxidation process performed

better in acidic than alkaline conditions due to reduced

scavenging of �OH by HCO�3 =CO

2�3 species. The UVC/H2O2

process was effective over a wide range of salinity and initial

DOC concentrations, demonstrating the wide applicability of

this technique for treating municipal ROC or potentially ROCs

from brackish sources.

The production of biodegradable organics by the UVC/H2O2

treatment was limited after 30 min, mainly due to the deple-

tion of the conjugated compounds. Both the EEM and LC-OCD

results showed that loss of conjugation in the molecular

structure of the humic-like substances greatly enhanced the

biodegradability of the organics. A DOC removal of 27% was

achieved after 30 min of the UVC/H2O2 treatment and 35% of

the DOC could be further degraded by biological treatment

resulting in a residual DOC of 9.3 mg/L.

No ecotoxicity of the ROC before and after UVC/H2O2 and

UVC/H2O2-BDOC treatment was apparent, as measured by the

Microtox� test. Although the THMFP was increased by UVC/

H2O2 treatment, it was reduced to below that of the raw ROC

after the biological treatment. However, the actual THM level

for the treated ROC after chlorination was expected to be

lower since the chlorine dose in real chlorination processes is

lower than those used in THMFP analysis. Overall, this

demonstrated the potential of recycling the UVC/H2O2-bio-

logically treated ROC for non-potable purposes.

Energy assessment using EE/O indicated that the energy

efficiency of the UVC/H2O2 treatment was greatly enhanced

(at least twice) by coupling with biological treatment (as per

BDOC test). The energy efficiency decreased rapidly after

30 min, this was accompanied by little improvement in final

DOC removal. Therefore, an irradiation time of 30 min or less

was suggested for the UVC/H2O2-BDOC treatment used in this

study. A residual DOC of 10 mg/L (initial 25 mg/L) was

achievable by UVC/H2O2-BDOC treatment with an irradiation

time of 30 min. With further development and research, e.g.,

improved design of the lamp and reactor for large-scale

processes, UV/H2O2 followed by biological treatment may

eventually become a viable option for treatingmunicipal ROC.

Appendix A. Supplementary information

Supplementary data related to this article can be found online

at doi:10.1016/j.watres.2012.03.024.

r e f e r e n c e s

ADWG, 2004. Australia Drinking Water Guidelines. NationalHealth and Medical Research Council, Australia.

AlHamedi, F.H., Rauf, M.A., Ashraf, S.S., 2009. Degradation studiesof Rhodamine B in the presence of UV/H2O2. Desalination 239(1e3), 159e166.

Atkinson, R., Darnall, K.R., Lloyd, A.C., Winer, A.M., Pitts, J.N.,1979. Kinetics and Mechanisms of the Reactions of theHydroxyl Radical with Organic Compounds in the Gas Phase.John Wiley & Sons, Inc.

Bagastyo, A.Y., Keller, J., Poussade, Y., Batstone, D.J., 2011.Characterisation and removal of recalcitrants in reverseosmosis concentrates from water reclamation plants. WaterResearch 45 (7), 2415e2427.

Beltran, F.J., Ovejero, G., Garcia-Araya, J.F., Rivas, J., 1995.Oxidation of polynuclear aromatic hydrocarbons in water. 2.UV radiation and ozonation in the presence of UV radiation.Industrial and Engineering Chemistry Research 34, 1607e1615.

Bolton, J.R., Bircher, K.G., Tumas, W., Tolman, C.A., 2001. Figures-of-merit for the technical development and application ofadvanced oxidation technologies for both electric- and solar-driven systems. Pure and Applied Chemistry 73 (4), 627e637.

Buchanan, W., Roddick, F., Porter, N., Drikas, M., 2004. Enhancedbiodegradability of UV and VUV pre-treated natural organicmatter. Water Science and Technology 4 (4), 103e111.

Buchanan, W., Roddick, F., Porter, N., 2006. Formation ofhazardous by-products resulting from the irradiation ofnatural organic matter: comparison between UV and VUVirradiation. Chemosphere 63 (7), 1130e1141.

Buxton, G.V., Elliot, A.J., 1986. Rate constant for reaction ofhydroxyl radicals with bicarbonate ions. International Journalof Radiation Applications and Instrumentation. Part C.Radiation Physics and Chemistry 27 (3), 241e243.

Chen, W., Westerhoff, P., Leenheer, J.A., Booksh, K., 2003.Fluorescence excitation-emission matrix regional integrationto quantify spectra for dissolved organic matter. EnvironentalScience and Technology 37, 5701e5710.

Chin, W.H., Roddick, F.A., Harris, J.L., 2009. Grey water treatmentby UVC/H2O2. Water Research 43 (16), 3940e3947.

Eaton, A.D., Franson, M.A.H., 2005. Standard Methods for theExamination of Water and Wastewater. American PublicHealth Association.

Edzwald, Benschoten, V., 1990. In: Hahn, H.H., Klute, R. (Eds.),Chemical Water and Wastewater Treatment. Springer, Berlin,pp. 341e359.

Garcıa, C., Hernandez, T., 1996. Influence of salinity on thebiological and biochemical activity of a calciorthird soil. PlantSoil 178 (2), 255e263.

H2O2.com, 2009. Ultraviolet Absorption Spectrum.Hrudey, S.E., 2002. Drinking water disinfection by-products:

when, what and why?. In: Disinfection By-Products and

wat e r r e s e a r c h 4 6 ( 2 0 1 2 ) 3 2 2 9e3 2 3 9 3239

Health Effects. Occasional Paper 5: Disfection By-Products.Co-operative Research Centre for Water quality andtreatment, pp. 7e14.

Joret, J.C., Levi, Y., 1986. Rapid method for estimation ofbiodegradable organic carbon in waters. Tribune duCEBEDEAU 39 (510), 3e9.

Kang, Y.W., Cho, M.-J., Hwang, K.-Y., 1999. Correction of hydrogenperoxide interference on standard chemical oxygen demandtest. Water Research 33 (5), 1247e1251.

Kleiser, G., Frimmel, F.H., 2000. Removal of precursors fordisinfection by-products (DBPs) e differences between ozone-and OH-radical-induced oxidation. Science of the TotalEnvironment 256 (1), 1e9.

Lee, L.Y.,Ng,H.Y.,Ong,S.L.,Hu, J.Y., Tao,G.,Kekre,K.,Viswanath,B.,Lay, W., Seah, H., 2009. Ozone-biological activated carbon asa pretreatment process for reverse osmosis brine treatment andrecovery. Water Research 43 (16), 3948e3955.

Legrini, O., Oliveros, E., Braun, A.M., 1993. Photochemicalprocesses for water treatment. Chemical Reviews(Washington, DC, United States) 93, 671e698.

Liao, C.H., Kang, S.F., Wu, F.A., 2001. Hydroxyl radical scavengingrole of chloride and bicarbonate ions in the H2O2/UV process.Chemosphere 44 (5), 1193e1200.

Liu, K., Roddick, F., Fan, L., 2011. Potential of UV/H2O2 oxidationfor enhancing the biodegradability of municipal reverseosmosis concentrates. Water Science and Technology 63 (11),2605e2611.

Oppenlander, T., 2003. Photochemical Purification of Water andAir. WILEY-VCH, Weinheim, Germany.

Perez, G., Fernandez-Alba, A.R., Urtiaga, A.M., Ortiz, I., 2010.Electro-oxidation of reverse osmosis concentrates generatedin tertiary water treatment. Water Research 44 (9), 2763e2772.

Shon, H., Vigneswaran, S., Snyder, S., 2006. Effluent organicmatter (EfOM) in wastewater: constituents, effects, andtreatment. Critical Reviews in Environmental Science andTechnology 36 (4), 327e374.

Thomson, J., Roddick, F.A., Drikas, M., 2004. Vacuum ultravioletirradiation for natural organic matter removal. Journal ofWater Supply Research and Technology-Aqua 53 (4), 193e205.

Tranvik, L.J., 1998. In: Hessen, D.O., Tranvik, L.J. (Eds.), AquaticHumic Substances e Ecology and Biogeochemistry. Springer-Verlag, Berlin, pp. 268e283.

Westerhoff, P., Moon, H., Minakata, D., Crittenden, J., 2009.Oxidation of organics in retentates from reverse osmosiswastewater reuse facilities. Water Research 43 (16),3992e3998.

Zhou, T., Lim, T.-T., Chin, S.-S., Fane, A.G., 2010. Treatment oforganics in reverse osmosis concentrate from a municipalwastewater reclamation plant: feasibility test of advancedoxidation processes with/without pretreatment. ChemicalEngineering Journal 166 (3), 932e939.

Zouboulis, A., Samaras, P., Ntampou, X., Petala, M., 2007. Potentialozone applications for water/wastewater treatment.Separation Science and Technology 42 (7), 1433e1446.