Bioresource Technology xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Review

A critical review on characterization strategies of organic matter for wastewater and water treatment processes Ngoc Han Tran a, Huu Hao Ngo b, Taro Urase c, Karina Yew-Hoong Gin d,⇑ a

NUS Environmental Research Institute, National University of Singapore, 5A Engineering Drive 1, T-Lab Building, Singapore 117411, Singapore School of Civil and Environmental Engineering, University of Technology Sydney, PO Box 123, Broadway, NSW 2007, Australia c School of Bioscience and Biotechnology, Tokyo University of Technology, Katakura 1404-1, Hachioji, Tokyo 1920982, Japan d Department of Civil and Environmental Engineering, Faculty of Engineering, National University of Singapore, 1 Engineering Drive 2, Singapore 117576, Singapore b

h i g h l i g h t s  Hydrophilic and neutral fractions of organic matter are the most significant foulants.  Hydrophobic acids showed higher THMs formation reactivity than hydrophilic fractions.  Humic substances were efficiently broken-down by ozone or Fenton reactions.  Low molecular weight acids are not effectively removed by ozone.

a r t i c l e

i n f o

Article history: Received 6 May 2015 Received in revised form 19 June 2015 Accepted 20 June 2015 Available online xxxx Keywords: Activated sludge models (ASMs) Chemical oxygen demand (COD) COD fractionation Dissolved organic matter (DOM) Molecular weight size distribution

a b s t r a c t The presence of organic matter (OM) in raw wastewater, treated wastewater effluents, and natural water samples has been known to cause many problems in wastewater treatment and water reclamation processes, such as treatability, membrane fouling, and the formation of potentially toxic by-products during wastewater treatment. This paper summarizes the current knowledge on the methods for characterization and quantification of OM in water samples in relation to wastewater and water treatment processes including: (i) characterization based on the biodegradability; (ii) characterization based on particle size distribution; (iii) fractionation based on the hydrophilic/hydrophobic properties; (iv) characterization based on the molecular weight (MW) size distribution; and (v) characterization based on fluorescence excitation emission matrix. In addition, the advantages, disadvantages and applications of these methods are discussed in detail. The establishment of correlations among biodegradability, hydrophobic/hydrophilic fractions, MW size distribution of OM, membrane fouling and formation of toxic by-products potential is highly recommended for further studies. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction Many countries all over the world are currently facing severe freshwater resource shortages due to increasing population, rapid economic growth, and extreme weather events brought on by climate change. In particular, the scarcity of fresh water is a critical problem that causes great concern in countries where natural water resources are limited. For this reason, wastewater reclamation plays an important role in developing strategies for water resource management worldwide. To date, the reuse of treated wastewater effluents from wastewater reclamation plants (WRPs) is considered to be an alternative water resource to ⇑ Corresponding author. Tel.: +65 6546 8104. E-mail address: [email protected] (K.Y.-H. Gin).

overcome the shortage of fresh water resources. However, the use of reclaimed water from treated wastewater effluents can pose potential health effects associated with microbial pathogens, heavy metals, pharmaceuticals and personal care products (PPCPs), endocrine disrupting chemicals (EDCs), and other recalcitrant organic compounds, as secondary effluents are considered a major source of these contaminants (Luo et al., 2014; Tran et al., 2014a,b). The acceptability of reclaimed water for a given water reuse is dependent on the physical, chemical, and microbiological quality of the water. Up to now, there has been widespread agreement that a desirable treated wastewater effluent for the water reuse purpose is always required to be not only low in concentrations of organic and inorganic pollutants, but also free from biological entities, such as, faecal bacteria, enteric viruses and other emerging pathogens. It is therefore necessary to design, control, and operate a wastewater

http://dx.doi.org/10.1016/j.biortech.2015.06.091 0960-8524/Ó 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Tran, N.H., et al. A critical review on characterization strategies of organic matter for wastewater and water treatment processes. Bioresour. Technol. (2015), http://dx.doi.org/10.1016/j.biortech.2015.06.091

2

N.H. Tran et al. / Bioresource Technology xxx (2015) xxx–xxx

Nomenclatures ASMs AOPs BOD bCOD COD CAS DBPs DOM ECPs GAC YH HPLC HiA HiB HiN HoA HoB HPI HPO HoN MBRs MW OCD

activated sludge models advanced oxidation processes biochemical oxygen demand biodegradable COD chemical oxygen demand conventional activated sludge disinfection byproducts dissolved organic matter extracellular polymers granular activated carbon heterotrophic biomass yield coefficient high performance liquid chromatography hydrophilic acids hydrophilic bases hydrophilic neutrals hydrophobic acids hydrophobic bases hydrophilic fraction hydrophobic fraction hydrophobic neutrals membrane bioreactors molecular weight organic carbon detection

reclamation system capable of efficiently removing undesirable pollutants from wastewater cost-effectively that plays a key role in facilitating further water reuse applications. In the past, conventional activated sludge processes (CAS) which involve the natural biodegradation of pollutants by heterotrophic and autotrophic bacteria, such as nitrifying bacteria in the activated sludge in aerated bioreactors, are regularly employed to treat municipal/industrial wastewaters. However, the use of CAS for wastewater reclamation usually shows some drawbacks due to the difficulties in separating the suspended solids (SS) and low removal efficiency for many inorganic/organic pollutants. Subsequently, the treated wastewater effluents from CAS are limited for further water reuse applications. In recent decades, the combination of biological treatment with microfiltration (MF) or ultra-filtration (UF) has been considered as an attractive technology for industrial and municipal wastewater reclamation due to its high removal efficiency of pollutants and cost effectiveness. For instance, it is well acknowledged that the submerged membrane bioreactor (sMBR) shows better technical feasibility for the reclamation and reuse of municipal wastewater compared to CAS. In reality, membrane technology, such as MF, UF, reverse osmosis (RO), and nano-filtration (NF), always provides a high degree of treatment in terms of SS, turbidity, faecal coliform, nitrogen, phosphorus and organic removal (Shon et al., 2004, 2006). However, it is evident that a major drawback of membrane-based treatment technologies is the membrane fouling that is caused by DOM fractions in raw wastewater/treated wastewater effluents. The membrane fouling affects the membrane performance, such as permeability and the OM rejection (Shon et al., 2006). Earlier studies reported that the humic substances (HS) fraction of OM is a major foulant, which controls the rate and extent of fouling, while recent studies have claimed that hydrophilic and neutral fractions of OM might be the most significant foulants (Tang et al., 2010; Lamsal et al., 2012). Thus, a good characterization of OM in wastewater samples can help optimize the performance of membrane filtration of biologically treated wastewater effluents through using appropriate pre-treatment processes, such as

OND Xp SH RO SEC Sp SMPs KS

lH,max XI SS XS SI CS CT CI XT ST TOC TPI UF UVD XH

organic nitrogen detection particulate inert microbial products rapidly hydrolysable fraction reverse osmosis size exclusion chromatography soluble microbial products soluble microbial products the half-velocity constant the maximum specific growth rate the particulate inert fraction the readily biodegradable fraction the slowly biodegradable fraction the soluble inert fraction the total biodegradable COD the total COD the total inert COD the total particulate COD the total soluble COD total organic carbon transphilic fraction ultra-filtration ultraviolet detection at 254 nm viable heterotrophic biomass

flocculation with FeCl3 and adsorption with powder activated carbon (Shon et al., 2004). In addition to membrane technologies, the applications of advanced oxidation processes (AOPs), such as catalytic ozonation (He et al., 2013), Fenton/photo-Fenton oxidation (Nousheen et al., 2014), and photocatalytic oxidation involving UV/H2O2 (Shon et al., 2007; González et al., 2013) have been recognized as high potential technologies for wastewater reclamation and water reuse. However, AOPs are normally used to partially oxidize non-biodegradable OM with high molecular weight or recalcitrant OM to more biodegradable compounds due to the large consumption of energy and chemicals for complete oxidation and mineralization. The efficiency of AOPs in wastewater reclamation is influenced not only by the concentrations of OM but also by its physicochemical properties and molecular weight (Selcuk et al., 2006; González et al., 2013; Molnar et al., 2013). More recently, it has been demonstrated that the formation processes of disinfection by-products (DBPs) during chlorination and chloramination of secondary effluent for the production of high quality recycled water is affected by both the concentrations of DOM and DOM characteristics (Zhuo et al., 2001; Zhang et al., 2009; Doederer et al., 2014; Ma et al., 2014). Apparently, the design, control, and operation of a wastewater reclamation system is largely dependent on how much is known about the characteristics of influent streams, the complexity of wastewater, and the degree of treatment required to meet the applicable discharge limits or reuse requirements. The design and operation of a wastewater reclamation system are reliable and effective only when the main components of wastewater are well characterized, particularly with regards to the characteristics of OM. However, the fact is that wastewater and reclaimed waters are complex and may contain a broad spectrum of both inorganic and organic pollutants that could affect the treatability. It is therefore necessary to choose suitable methods for characterizing OM in wastewater or reclaimed water samples for purposes of wastewater reclamation and further water reuse applications. This review summarizes the current knowledge on the characterization of OM in wastewater and reclaimed water. In addition, the advantages and disadvantages of characterization methods of OM are also discussed.

Please cite this article in press as: Tran, N.H., et al. A critical review on characterization strategies of organic matter for wastewater and water treatment processes. Bioresour. Technol. (2015), http://dx.doi.org/10.1016/j.biortech.2015.06.091

N.H. Tran et al. / Bioresource Technology xxx (2015) xxx–xxx

2. Characterization of OM in wastewater and water samples Generally, wastewater contains a variety of organic and inorganic components with different physicochemical characteristics. The pollutants or kinds of pollutants present in wastewater are dependent on pollution sources. For example, it has been shown that high concentrations of COD, oil, grease, sulphide, ammonia, organic solvents, and polycyclic aromatic hydrocarbons (PAHs) are often detected in the wastewater samples derived from petrochemical refineries and chemical processing industries (Botalova et al., 2009), whereas a high concentration of heavy metals and a low COD value is observed in wastewater samples from electroplating/electrical industries (Kurniawan et al., 2006). In addition, the design of sewer line systems (i.e. separate sewer or combined sewer systems) also directly affects the wastewater composition. It is impossible and uneconomical to monitor all unknown inorganic and organic components in wastewater. The selection of the main monitoring parameters/target pollutants is dependent on the specific purposes of the study. This review only focuses on summarizing and elucidating advantages and limitations of the methods for characterization of OM in wastewater and water samples for wastewater reclamation and water reuse purposes.

2.1. Characterization based on the biodegradability Generally speaking, the pollution strength of wastewaters is assessed through the chemical oxygen demand (COD). It is hard, however, to give a clearly defined range to distinguish among the low, medium, and high strength of a wastewater if it is only based on the COD values. The high strength of wastewater is usually related to large quantities of components that are found in wastewater, such as high amount of COD, ammonia, suspended solids, or heavy metals. There is widespread agreement that the low strength of municipal wastewater as COD is typically less than 1000 mg/L (Gopala Krishna et al., 2008). However, for the case of petrochemical/chemical processing wastewaters, the COD value (i.e. 1000 mg/L) is still considered as high strength because petrochemical/chemical processing industries often contain hard COD with a high content of non-biodegradable compounds such as heavy metals, whereas a high content of biodegradable compounds, such as nitrogen or phosphorous elements, are frequently observed in municipal wastewaters. It has been demonstrated that one of major the drawbacks in using COD to evaluate the composition of OM in wastewater samples is that it covers both biodegradable and non-biodegradable compounds. As a result, it is very challenging to evaluate the biodegradability of wastewater by using COD as an index. The evaluation of pollution strength of wastewater based on the BOD or BOD5 is also limited since the BOD test merely indicates the biodegradable fraction of the wastewater. However, it is possible to use the ratio of BOD5/COD as an index to assess the biodegradability of OM contained in wastewater. For example, a high BOD5/COD ratio indicates the readily biodegradability of OM in wastewater samples, whereas a poor biodegradability of OM exists when the BOD5/COD ratio is low. Based on the BOD5/COD ratio of the wastewater sample, it is possible to select suitable treatment technologies for the wastewater. For instance, biological wastewater treatment technologies is feasible for treating wastewater samples with BOD5/COD ratio greater than 0.5 (Mutamim et al., 2012). On the contrary, for wastewater with BOD5/COD ratios less than 0.5, pre-treatments using physical or chemical processes are needed prior to biological treatment processes. In general, the use of lumped parameters, such as total and filtered COD, is quite sufficient for the design and operation of a

3

wastewater reclamation system. However, this is based on the premise that there are no major changes in the wastewater characteristics. In particular, it has been demonstrated that the characteristics of OM in wastewater have a major impact on the performance of biological wastewater treatments. Therefore, more detailed classification of OMs based on solubility and biodegradability allows better understanding of the composition of wastewater. This can improve process design, operation, and modelling for biological treatment processes. However, a detailed evaluation of chemical composition for the purpose of wastewater characterization is time consuming and thus, receives limited attention. Normally, the OM of wastewater can be fractionated into five different COD fractions as shown in Fig. 1, including (i) the readily biodegradable component (SS) which is in general soluble; (ii) the slowly biodegradable COD (XS) that is often related to particulate form of organic matter; (iii) the rapidly hydrolysable fraction (SH); (iv) the soluble inert fraction (SI); and (v) the particulate inert fraction (XI) (Henze et al., 2000). The COD fractionation allows the prediction of the biological treatment efficiency, the oxygen demand and sludge production. For example, SS and XI fractions are of greatest significance in process design. For BOD removal and nitrification processes, the SS concentration is important for evaluating the oxygen demand profiles for plug-flow, staged and batch-fed processes. The effect of XI concentration in the influent is significant in sludge production and aeration volume requirements. It is noted that the level of characterization required depends on the specific objectives for modelling and process design. For instance, to evaluate the biodegradability of OM in a wastewater sample, total COD (CT) is evaluated in terms of total biodegradable COD (CS) and total inert COD (CI) as described in Eq. (1):

CT ¼ CS þ CI

ð1Þ

The total inert COD (CI) can be further divided into two groups, i.e. the soluble inert COD (SI) and particulate inert COD (XI).

C I ¼ SI þ X I

ð2Þ

Similarly, the total biodegradable COD can be further categorized into three major fractions, including the readily biodegradable component (SS), rapidly hydrolysable fraction (SH) and slowly biodegradable fraction (XS), as shown in Eq. (3):

C S ¼ SS þ SH þ X S

ð3Þ

In additional, it is critically needed to differentiate the soluble and particulate fractions of the COD in the biological reactor in the evaluation of a model. For this purpose, filtration is conveniently used as a practical index of solubility. This differentiation also defines the major process components in currently used mechanistic models as shown in Fig. 2. Thus, the total soluble COD includes the following fractions: readily biodegradable component (SS); rapidly hydrolysable fraction (SH); soluble inert COD (SI) and soluble inert microbial products (Sp), while the total particulate COD consists of the following fractions: slowly biodegradable COD fraction (XS); viable heterotrophic biomass fraction (XH); particulate inert COD (XI); and particulate inert microbial products (Xp).

C T ¼ ST þ X T

ð4Þ

ST ¼ SS þ S H þ S I þ S P

ð5Þ

XT ¼ XS þ XH þ XI þ XP

ð6Þ

To date, numerous efforts have been made to establish suitable experimental procedures for fractionating COD through either physical–chemical or biological (respirometry) approaches

Please cite this article in press as: Tran, N.H., et al. A critical review on characterization strategies of organic matter for wastewater and water treatment processes. Bioresour. Technol. (2015), http://dx.doi.org/10.1016/j.biortech.2015.06.091

4

N.H. Tran et al. / Bioresource Technology xxx (2015) xxx–xxx

Fig. 1. Scheme of distribution of major COD fractions in wastewater.

Fig. 2. Scheme of fractionation of major COD components in wastewater/treated wastewater effluent samples.

(Mathieu and Etienne, 2000; Roeleveld and van Loosdrecht, 2002; Orhon and Okutman, 2003; Gatti et al., 2010). However, one of the main differences between the various fractionation methods is the way by which the readily biodegradable fraction (SS) or the soluble inert fraction (SI) is determined. Many previous studies have reported that respirometry is the most suitable method to measure the readily biodegradable fraction (SS) (Mathieu and Etienne, 2000; Karahan et al., 2008; Fall et al., 2012), while other studies have claimed that simplified physical–chemical methods, with the advantages of being faster and easier than respirometry methods, can be used to determine COD fractions. Therefore, this review aims to summarize the current knowledge on using both physical–chemical and respirometry methods for COD fractionation in order to elucidate the advantages and drawbacks of these methods in the characterization of OM in wastewater. 2.1.1. COD fractionation using physical–chemical methods Till now, several physical–chemical methods have been employed to fractionate the COD fractions via measurement of the soluble COD and biodegradable COD values of the influent and effluent of the WWTP (Roeleveld and van Loosdrecht, 2002; Gatti et al., 2010; Fall et al., 2012). The suitability of these methods is related to reproducibility and practical applicability. It is reported that the fraction of readily biodegradable COD (SS) strictly depends on the choice of filter pore size. In the past, it was assumed that the use of 0.45 lm membrane filters would allow obtaining filtrate without the content of colloidal matter. However, it was found that there was a difference in SS values between the pre-precipitated samples and the filtrate using a pore size membrane of 0.45 lm (Fall et al., 2011). Nowadays, it is widely accepted that a membrane filter of 0.1 lm is considered as the

actual cut-off size of the soluble fraction for physical–chemical characterization. For physical–chemical methods, it is accepted that the inert COD fraction (SI) can be calculated by measuring the inert soluble COD in the effluent of WWTP (i.e. COD-effluent after membrane filtration through 0.1 lm) and the readily biodegradable COD fraction (SI) is computed by subtracting the fraction SI from the soluble COD in the influent of WWTP (Gatti et al., 2010). Other studies have proposed that raw wastewater is flocculated then passed through a 0.45 lm membrane filter to obtain the total soluble COD. The inert soluble COD (SI) can be measured by using the same method on the treated effluent sample from the WWTP. The difference between the total soluble COD and SI is considered to be the readily biodegradable COD (SS). If WWTP data is not available, it is necessary to set up a lab-scale continuous or batch reactor operated under sludge ages ranging from 10 to 20 days for the determination of soluble inert COD. It should be noted that the physical–chemical method is only applicable for determining SS and cannot reflect the slowly biodegradable COD fraction (XS). To calculate XS, it is possible to subtract the fraction SS from the total biodegradable COD (bCOD). Nonetheless, it is challenging to estimate accurately the bCOD. Previously, it was assumed that the total biodegradable COD fraction (bCOD) may be considered as the BOD value at 20 days, since 95–99% of the biodegradable COD is oxidized after 20 days of incubation (Namour and Müller, 1998; Gatti et al., 2010). However, the measurement of the BOD20 is not reliable and therefore BOD20 is not recommended. On the other hand, from theory, the total biodegradable COD is always greater than the ultimate BOD, Roeleveld and van Loosdrecht (2002) recommended that bCOD should be evaluated based on the ultimate BOD (BODultimate), in

Please cite this article in press as: Tran, N.H., et al. A critical review on characterization strategies of organic matter for wastewater and water treatment processes. Bioresour. Technol. (2015), http://dx.doi.org/10.1016/j.biortech.2015.06.091

5

N.H. Tran et al. / Bioresource Technology xxx (2015) xxx–xxx

which BOD is measured as a function of time. This can be easily computed from Eq. (7).

BODultimate ¼

BODt 1  ekBOD t

ð7Þ

where kBOD is the first order rate constant of the BODt versus time measurement. It can be determined by fitting the BOD curve through the measured data with the ultimate BOD, as described by Eq. (7). It is noted that BOD measurements must be conducted with non-filtered samples and with addition of allylthiourea (ATU) at a final concentration of about 20 mg/L to inhibit nitrification. Studies have shown that kBOD is dependent on the division between SS and XS that is impacted by the type and length of the sewer line system, the fraction of industrial wastewater, and the application of pre-treatment (Roeleved and van Loosdrecht, 2002). When for a certain wastewater the kBOD is determined, the bCOD can be estimated from the BOD5. It is noted that during the measurement of BOD there is an interaction of growth and decay of biomass. As a consequence, there is a conversion of a part of biodegradable COD into an inert COD fraction. This leads to the initial biodegradable COD being higher than the determined ultimate BOD; subsequently the use of a correction factor fBOD is required to estimate the bCOD from the ultimate BOD. The total biodegradable COD fraction is calculated using the following equation.

C S ¼ bCOD ¼

BODultimate 1  f BOD

ð8Þ

In addition, Roeleveld and van Loosdrecht (2002) suggested a simple model to estimate the bCOD based on the BOD5 measurement, as described in Eq. (9):

CS ¼

BOD5 0:7  ð1  f BOD Þ

ð9Þ

where the factor 0.7 denotes the BOD5/BOD ultimate ratio; fBOD = 0.15 is a correction factor that accounts for the inert COD generated by biomass hydrolysis (Roeleveld and van Loosdrecht, 2002). Similarly, the particulate inert COD (XI) can be determined as the difference between total COD in the raw wastewater minus other COD fractions, such as SS, SI, and XS. The COD fractionation procedure can be mathematically described in Eqs. (10–13):

SI ¼ CODeff;sol

ð10Þ

SS ¼ CODinf;sol  SI

ð11Þ

X S ¼ bCODinf;tot  SS

ð12Þ

(Mathieu and Etienne, 2000; Orhon and Okutman, 2003; Gatti et al., 2010). Recently, the respirometry method has been used broadly not only for determination of the COD fractions, but also for identification of kinetic parameters in calibration processes. The OUR profiles generated in an aerated batch bioreactor initially fed with a pre-selected ratio of wastewater and biomass is a useful tool in reflecting the equivalence between the biodegradable COD and the corresponding dissolved oxygen utilization for the tested wastewater sample (Gatti et al., 2010). The biodegradable OM is estimated from the fundamental mass balance between substrate utilization, biomass growth and electron acceptor consumption (Orhon and Okutman, 2003; Gatti et al., 2010). In practice, to determine the COD fractions, a series of OUR tests are usually carried out on three types of wastewater samples: (i) a non-filtered (raw) wastewater sample (NFS); (ii) a filtered wastewater sample (FS) on membrane filter (0.45 lm); and (iii) a filtered wastewater sample (0.45 lm) after coagulation-flocculation treatment (CFS) (Munz et al., 2008). It is assumed that the COD in NFS (CODNFS) includes all COD fractions; while the COD in FS (CODFS) consists of the soluble and the rapidly hydrolysable COD fractions and the COD in the CFS (CODCFS) represents only the soluble COD fractions (Munz et al., 2008). For the respirometric experiments on heterotrophic microbes to determine COD fractions, it is required to inhibit nitrification by addition of allylthiourea (ATU) at a defined concentration (Tran et al., 2014c). In addition, it is emphasized that the quality of the kinetic information contained in the respirometry method is dependent on the ratio of the initial substrate concentration to the initial biomass concentration (Fo/Xo). Previous studies suggested that the accuracy of respirometric tests was improved when the Fo/Xo ratio was less than 0.2 mg COD/mg MLVSS (Mathieu and Etienne, 2000; Gatti et al., 2010). The readily biodegradable COD fraction (SS) is determined based on the CFS. It can be determined by the area under the exogenous oxygen uptake rate curve, obtained from the OUR curve after deduction of the endogenous respiration rate (Gatti et al., 2010).

SS ¼

SS ¼

2.1.2. COD fractionation based on respirometry methods As discussed above, there are several drawbacks in assessing biodegradability via measurements of BOD, such as the difficulty in selecting an appropriate dilution factor and the long experimental course in analyses of BOD. Due to these limitations, a comparatively efficient method known as the respirometry method that measures the oxygen uptake rate (OUR) during biodegradation processes, has been used as an alternative to the BOD measurement for determining the wastewater biodegradability fractions

t

OURCFS dt

ð14Þ

0

DO2ðCFSÞ ð1  Y H Þ

ð15Þ

Similarly, the rapidly hydrolysable COD fraction (SH) can be estimated from the FS which can be described in the following equation:

ð13Þ

For a particular wastewater in a WWTP, the OM characterization based on physical–chemical methods can be a simple and rapid tool for the determination of the biodegradable fractions. However, the choice of membrane filter pore size and/or type of sample treatment will play an important role in the determination of truly soluble COD fraction.

Z

It can also be rearranged as in Eq. (15).

SH ¼ X I ¼ CODinf;tot  X S  SS  SI

1 ð1  Y H Þ

DO2ðFSÞ DO2ðFSÞ  DO2ðCFSÞ  SS ¼ ð1  Y H Þ ð1  Y H Þ

ð16Þ

Likewise, the total biodegradable COD fraction (CS) in the sample can be determined from the NFS as presented in Eq. (17). It is therefore easy to compute the slowly biodegradable fraction (XS).

C S ¼ SS þ SH þ X S ¼

XS ¼

DO2ðNFSÞ ð1  Y H Þ

ð17Þ

DO2ðNFSÞ DO2ðNFSÞ  DO2ðFSÞ  ðSS þ SH Þ ¼ ð1  Y H Þ ð1  Y H Þ

ð18Þ

The soluble and particulate inert COD fractions can be computed from Eqs. (19) and (20):

SI ¼ CODFS  ðSH þ X S Þ ¼ CODFS  X I ¼ CODNFS  CODFS  X S

DO2ðFSÞ ð1  Y H Þ

ð19Þ ð20Þ

Please cite this article in press as: Tran, N.H., et al. A critical review on characterization strategies of organic matter for wastewater and water treatment processes. Bioresour. Technol. (2015), http://dx.doi.org/10.1016/j.biortech.2015.06.091

6

N.H. Tran et al. / Bioresource Technology xxx (2015) xxx–xxx

where YH is the heterotrophic biomass yield coefficient (mg cell/mg COD removed). This parameter is very important for estimating the biodegradable fraction. It has been illustrated that the YH value is relatively dependent on the ratio of the initial substrate concentration to the initial biomass concentration (Fo/Xo) (Gatti et al., 2010). Therefore, in order to evaluate the errors of the respirometry method, several replicates (at least four times for each experiment) should be carried out at low, medium, and high Fo/Xo ratio by adding a solution with known concentration of acetic acid. Gatti et al. (2010) reported that YH values were in the range of 0.58–0.61 (mg cell/mg COD removed), while others found the YH was about 0.67 (mg cell/mg COD removed). The characterization of heterotrophic biomass should be taken into account in design and modelling WWTPs. Normally, the estimation of the decay coefficient of the heterotrophic biomass is obtained using a single batch experiment. The kinetic parameters of growth and decay of the heterotrophic biomass are calculated using model calibration of the OUR profiles resulting from the three kinds of wastewater mentioned above (NFS, FS, and CFS), in which the maximum specific growth rate (lH,max) and the half-velocity constant (KS) are estimated via the model calibration of the OUR profile rising from the addition of CFS. However, one of the drawbacks of the respirometry method compared to the physical–chemical analysis is that it is time consuming and usually relies on many default parameter values for calculations of COD biodegradability fractions. For instance, the heterotrophic biomass yield coefficient YH is usually assumed to be 0.67 mg cell COD/mg COD for calculation of SS from the OUR profiles. The use of default parameters in the determination of COD fractions can be controversial because of differences in wastewater characteristics. The parameters associated with the determination of fractions, such as SS, XS and SI, must be validated carefully by using complicated models. It has been found that there is a notable difference in the readily biodegradable fraction obtained from physical–chemical and respirometry methods (Fall et al., 2011). In addition, another main issue associated with the evaluation of soluble and particulate inert COD fractions in wastewater by the respirometry method is the interference of residual microbial products generated during substrate utilization. Therefore, the use of a simple physical–chemical method can still be used as a potential alternative to the respirometry method for the determination of wastewater COD fractions (i.e. SS, XS, SI, and XI). 2.2. Characterization based on the particle size distribution (PSD) It has been demonstrated that the particle size distribution (PSD) of OM is an important index of wastewater treatability and can be used to evaluate the suitability of physical, chemical, or biological processes for eliminating OM in wastewater. The particle size of OM in domestic wastewater regularly varies from nano-scale to a few millimetres. Usually, the variation of PSD in a WWTP can be dependent on the nature of the suspended solids in the influent wastewater. Previous studies have reported that the PSD of OM can directly affect its biodegradability (Dulekgurgen et al., 2006; Garcia-Mesa et al., 2010). The small size OM can be consumed by microorganisms easily since smaller particle size increases the surface area available to the microorganisms, while larger particles need to be hydrolysed prior to consumption (Dimock and Morgenroth, 2006; Karahan et al., 2008). The PSD of particulate OM can be considered a significant factor affecting the biodegradation process (Dimock and Morgenroth, 2006; Dulekgurgen et al., 2006; Karahan et al., 2008). The use of PSD is particularly useful in the evaluation of separation efficiency of the membrane-based treatment technologies, such as membrane bioreactors (MBRs), ultra-filtration (UF), reverse osmosis (RO), and nano-filtration (NF) (Sophonsiri and

Morgenroth, 2004; Dulekgurgen et al., 2006; Karahan et al., 2008). In addition, the fouling of membranes is known to correlate with the particle size of OM in the solution. For this reason, a better understanding of the PSD of OM in wastewater and reclaimed water samples is needed to evaluate the treatability of OM. Generally, the PSD of OM in wastewater can be assessed by sequential filtrations (Dulekgurgen et al., 2006; Karahan et al., 2008) or laser scattering techniques (Wu and He, 2012) to evaluate the biodegradability of fractions. Based on the PSD of OM fractions, it is possible to categorize OM fractions in wastewater or reclaimed water into four main groups, including (i) settleable (e.g. >100 lm); (ii) supra colloidal (e.g. 1–100 lm); (iii) colloidal (e.g. 0.08–1 lm); and (iv) soluble (10 lm for the agricultural wastewater) leaving a gap in the size range of large macromolecules and colloids. Interestingly, Sophonsiri and Morgenroth (2004) also found that the relative protein and carbohydrate concentrations varied for the different size fractions compared to the measured chemical oxygen demand (COD) in the corresponding size fraction. In another study, Dulekgurgen et al. (2006) also demonstrated that there was a correlation between particle size distribution and chemical oxygen demand fractionations. For example, the bulk (around 65%) of COD in domestic sewage consists of particulate organic matter with particle size greater than 1600 nm and only 14% is soluble with particle size less than 2 nm (Dulekgurgen et al., 2006). It is interesting to note that there were differences between COD values in the influent and effluent samples at different particle size intervals. The COD fractions of the influent with particle size ( hydrophobic acids > transphilic acids > hydrophilic charged. Nevertheless, other studies have claimed that HPO compound character, such as humic substances, are the main foulant associated with adsorptive fouling or hydrophobic interaction, particularly during filtration with hydrophobic membranes (Jarusutthirak et al., 2002; Shon et al., 2006; Zularisam et al., 2006). In wastewater/water reclamation, it has been recognized that the treatability of DOM is dependent on both its concentration

7

and physicochemical properties. For example, previous studies reported that the HoB and HoN fractions could be removed effectively by coagulation (Shon et al., 2006; Haberkamp et al., 2007). In recent years, it has been revealed that the concentrations and characteristics of DOM in treated wastewater effluents also directly affect the formation of potentially harmful disinfection by products (DBPs) such as trihalomethanes (THMs) and haloacetic acids (HAAs) (Zhuo et al., 2001; Ye et al., 2009; Zhang et al., 2009; Lamsal et al., 2012; Doederer et al., 2014; Ma et al., 2014). For example, Zhang et al. (2009) and Lamsal et al. (2012) found that hydrophobic organic compounds especially HoA showed much higher THMs formation reactivity than HPI compounds. However, one of the main drawbacks of the use of XAD-4 and XAD-8 resins to fractionate DOM is that this method may produce significant overlap between HPO and HPI fractions. For instance, HPI fractions, such as low molecular weight acids and neutrals which are neutral acids at a pH of 2.0, are trapped within the pores of the XAD-8 resin and eluted as HPO. Similarly, a huge amount of humic substances can adsorb on XAD-4, as the contact time is too short to allow humic substances to adsorb onto the XAD-8 resin. Consequently, remaining humic substances adsorbed onto XAD-4 are eluted as TPI. Another limitation of the resin fractionation method is that the recovery of DOM from the resins using a NaOH or NaOH/NaCl eluent varies significantly (27.2–280%) depending on each type of resin and experiment. Kim and Dempsey (2012) found that more than 90% of organic acids were recovered from the DEAE resin, while only 79% of DOM was recovered from XAD-8 resin. The incomplete recovery of adsorbed DOM from XAD-8 was suggested to be likely due to simultaneous removal of HPO neutrals and acids. Theoretically, the HPO neutrals are not extracted with an alkaline eluent. Interestingly, Kim and Dempsey (2012) observed that the IRA-958 resin that was used to fractionate was degraded by the eluent and therefore, the extract was not an accurate representation of DOM in the sample. Other disadvantages of resin fractionation are connected to the discrepancies, including possible chemical and physical alterations of DOM caused by extreme pH levels and changes in pH during fractionations, irreversible adsorption of DOM onto the resin, contamination from resin bleeding and size-exclusion effects. Therefore, the development of alternative methods for characterization of DOM to diagnose the cause of membrane fouling, the formation of toxic by-products, and treatability in wastewater/water reclamation is essential. 2.4. Characterization based on the molecular weight size distribution DOM in raw wastewater, biologically treated wastewater effluents and natural waters is a complex mixture of various organic materials. It consists of different molecular weight (MW) fractions ranging from low MW substances (e.g. amino acids, carboxylic acids, alcohols, aldehydes, etc.) to high MW compounds such as humic substances, polysaccharides, and proteins (Leenheer and Croue, 2003). For example, Shon et al. (2004) reported that the MW distribution of DOM in biologically treated wastewater effluent varied from 300 to 40,000 Da, of which the highest fraction was in the range of 300–5,000 Da. It has been reported that soluble microbial products (SMPs) are the most significant components of DOM in effluents of membrane bioreactors (MBRs) (Jarusutthirak et al., 2002; Romera-Castillo et al., 2014). To date, it has been revealed that molecular properties of DOM, particularly its molecular weight, strongly affect its reactivity in both natural and engineered aquatic systems (Her et al., 2003; Leenheer and Croue, 2003; Shon et al., 2004, 2007; Yan et al., 2007, 2012). For this reason, it is possible to use the MW profiles of DOM as a helpful tool to evaluate its treatability in coagulation (Conte and Piccolo, 1999; Yan et al., 2007). For example, in a study to assess the effect of

Please cite this article in press as: Tran, N.H., et al. A critical review on characterization strategies of organic matter for wastewater and water treatment processes. Bioresour. Technol. (2015), http://dx.doi.org/10.1016/j.biortech.2015.06.091

8

N.H. Tran et al. / Bioresource Technology xxx (2015) xxx–xxx

MW size on coagulation, Yan et al. (2007) revealed that the DOM segments with MW ranging from 3000 to 10,000 Da could be removed efficiently by coagulation. In another study to investigate the effects of solar radiation on the MW size distribution from fulvic acids in pulp mill effluents, Carvalho et al. (2008) also observed that high MW fractions were readily photo-degraded to low MW constituents. Similarly, González et al. (2013) recently found that a large reduction of biopolymers (MW > 20,000 Da) and humic substances (MW 800–1000 Da) in treated wastewater to low MW fractions was observed by pre-treatment with ozone or UV/H2O2. Moreover, the MW size distribution characteristics of DOM have also been proven to have an effect on the toxicity and mobility of its associated pollutants (Al-Reasi et al., 2013). For this reason, an in-depth understanding about the MW size distribution of DOM is very critical for the design of water and wastewater treatment systems. However, so far there has been no standardized procedure to identify the MW distribution of DOM fractions in water samples. Mostly, the MW size of DOM fractions has been determined based on the apparent molecular weight distribution. In the past, several methods were employed to estimate the MW of DOM, such as ultra-filtration, small-angle X-ray scattering, and gel permeation chromatography. However, it was found that there were some drawbacks in these techniques, such as membrane pore size distribution, sample temperature, stirred cell pressure, pH, and membrane materials, which could affect the transport of organics through membranes. In recent years, high performance liquid chromatography coupled with size exclusion chromatography (HPLC–SEC) has become one of the most widely used techniques for characterizing DOM due to its advantages, such as minimal amount of pre-treatment, small injection volume, ease and speed of analysis, and the ability to determine both number average and weight-average MW, e.g. Mn and Mw, respectively (Her et al., 2003; Yan et al., 2012). In principle, the separation of molecules in a SEC column is based on size, and compounds with lower apparent MW are eluted later than those with higher apparent MW. The molecule shape and some interaction characteristics may influence the results. However, the interpretation of HPLC-SEC data is greatly affected by the type of post-column detection used (Her et al., 2003; Yan et al., 2012). More recently, HPLC-SEC is often conducted with sequential online detectors, such as organic carbon detector (OCD), organic nitrogen detector (OND) or UV detector at 254 nm (UVD), called LC–OCD–OND–UVD. This method has been known as a powerful tool to separate the pool of DOM into five major fractions with different MW sizes, including (i) biopolymers (MW >10,000 Da); (ii) humic substances (MW 800–1000 Da); (iii) building blocks (MW 350–500 Da); (iv) low MW acids (MW