Enhanced digestion of waste activated sludge using microbial electrolysis cells at ambient temperature

Water Research xxx (2015) 1e10

Contents lists available at ScienceDirect

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

Enhanced digestion of waste activated sludge using microbial electrolysis cells at ambient temperature Joseph R. Asztalos, Younggy Kim* Department of Civil Engineering, McMaster University, 1280 Main Street West, Hamilton, ON, L8S 4L8, Canada

a r t i c l e i n f o

a b s t r a c t

Article history: Received 30 January 2015 Received in revised form 14 May 2015 Accepted 23 May 2015 Available online xxx

This study examined the effects of the microbial electrolysis cell (MEC) reactions on anaerobic digestion of waste activated sludge from municipal wastewater treatment under ambient temperature conditions (22e23
C). Two lab-scale digesters, a control anaerobic digester and an electrically-assisted digester (EAD e equipped with a MEC bioanode and cathode) were operated under three solids retention times (SRT ¼ 7, 10 and 14 days) at 22.5 ± 0.5
C. A numerical model was also built by including the MEC electrode reactions in Anaerobic Digestion Model No.1. In experiments, the EAD showed reduced concentration of acetic acid, propionic acid, n-butyric acid and iso-butyric acid. This improved performance of the EAD is thought to be achieved by direct oxidation of the short-chain fatty acids at the bioanode as well as indirect contribution of low acetic acid concentration to enhancing beta-oxidation. The VSS and COD removal was consistently higher in the EAD by 5e10% compared to the control digester for all SRT conditions at 22.5 ± 0.5
C. When compared to mathematical model results, this additional COD removal in the EAD was equivalent to that which would be achieved with conventional digesters at mesophilic temperatures. The magnitude of electric current in the EAD was governed by the organic loading rate while conductivity and acetic acid concentration showed negligible effects on current generation. Very high methane content (~95%) in the biogas from both the EAD and control digester implies that the waste activated sludge contained large amounts of lipids and other complex polymeric substances compared to primary sludge. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Psychrophilic anaerobic digestion Waste activated sludge Bioelectrochemical system Exoelectrogenic bacteria Anaerobic Digestion Model No. 1 Short-chain fatty acids Electrically-assisted digestion

1. Introduction In municipal wastewater sludge treatment, anaerobic digestion is typically operated under mesophilic conditions at temperatures ranging from 35 to 40
C (Metcalf & Eddy et al., 2004). To maintain this temperature requirement for large sludge volumes, a substantial amount of energy is therefore required. However, low temperature conditions below 35
C slow down the biosolids destruction with reduced rates of microbial growth and biogas production (Connaughton et al., 2006). As a result, lower temperature digesters require a substantially long solids retention time (SRT) for adequate performance. For example, a digester operated at 24
C or lower would require an SRT of longer than 20 days to digest municipal wastewater sludge under well-mixed conditions (Reynolds and Richards, 1995; Metcalf & Eddy et al., 2004;

* Corresponding author. E-mail address: [email protected] (Y. Kim).

Bolzonella et al., 2005). However, there are a number of benefits of operating digestion systems at a lower temperature, such as reduced energy input and significantly reduced digester construction cost without heating systems and insulation walls, allowing for small wastewater treatment facilities to operate sludge digesters. Since low temperature conditions substantially decrease the rate of sludge digestion, the primary objective of this study was to enhance the rate of volatile suspended solids (VSS) and chemical oxygen demand (COD) removal using an electrically-assisted digester (EAD) at ambient temperature conditions (22e23
C). In anaerobic digestion, the destruction of biosolids is achieved through a series of biological reactions (Fig. 1). Under mesophilic conditions, the hydrolysis of carbohydrates and proteins is relatively quick, requiring 1e3 days; while lipids require 6e8 days for hydrolytic decomposition (Grady et al., 2011). Many studies have reported that if digester influent contains a large amount of complex lipids, the hydrolysis step starts to govern the overall rate of biosolids destruction (Ariunbaater et al., 2014; Izumi et al., 2010; Ma et al., 2011; Valo et al., 2004). Hydrolyzed soluble organics

http://dx.doi.org/10.1016/j.watres.2015.05.045 0043-1354/© 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Asztalos, J.R., Kim, Y., Enhanced digestion of waste activated sludge using microbial electrolysis cells at ambient temperature, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.05.045

2

J.R. Asztalos, Y. Kim / Water Research xxx (2015) 1e10

Nomenclature1 ADM1 CE COD DHCH4 EAD Eap F FID GC I LCFA

Anaerobic Digestion Model No.1 Coulombic efficiency () chemical oxygen demand (mg/L) heat of combustion of CH4 (890.8 kJ/mol) electrically-assisted digester applied electric voltage (V) Faraday constant (96,485 C/mol) flame ionization detector gas chromatography electric current (A) long chain fatty acid

(monosaccharides, amino acids and long-chain fatty acids) are decomposed to short-chain organic acids and hydrogen gas in acidogenesis reactions, such as fermentation and beta-oxidation. Acetoclastic methanogenesis and hydrogenotrophic methanogenesis are the final steps converting acetate and hydrogen gas to methane gas, respectively. At 35
C, hydrogenotrophic methanogens are known to grow rapidly and convert hydrogen gas to methane in less than 1 day (Grady et al., 2011). Acetoclastic methanogenesis, however, requires a substantially long time as acetoclastic methanogens need 3e5 days (Methanosarcina spp.) and at least 12 days (Methanosaeta spp.) to sustain growth (Grady et al., 2011), indicating that the time requirement for acetoclastic methanogenesis is even longer at ambient temperature conditions (20e25
C). The majority of acetoclastic methanogenesis is driven by Methanosaeta species which result in the process being another rate-limiting step in anaerobic digestion. When the digester influent contains easily degradable substrates with a relatively small amount of lipids, acetoclastic methanogenesis has been reported to be the dominant rate-limiting step (Ariunbaater et al., 2014; Grady et al., 2011; Rittmann and McCarty, 2001). In domestic wastewater sludge digestion, the influent does not typically contain high levels of complex substrates which results in acetoclastic methanogenesis being the key rate-limiting step. In this study, microbial electrolysis cell (MEC) technology was integrated into a lab-scale anaerobic digester in order to expedite the rate of biosolids destruction as previously described (Asztalos and Kim, 2015, accepted) but under even lower temperature conditions (22
C). An MEC consists of a bioanode and cathode that are electrically connected with an external power supplier (Liu et al., 2005b; Rozendal et al., 2006; Logan et al., 2008). Acetate is oxidized by exoelectrogenic bacteria at the bioanode and hydrogen gas is produced at the cathode via electrolytic water reduction. Hydrogenotrophic methanogens quickly convert the produced hydrogen gas into methane gas. The MEC bioanode oxidizes a certain portion of the available acetate in the digester and creates an additional pathway for acetate removal as previous demonstrated in the EAD (electrically-assisted digester) under mesophilic condition (40
C) (Asztalos and Kim, 2015, accepted). While expedited volatile suspended solids (VSS) and chemical oxygen demand (COD) removals were demonstrated, decomposition of organic acids (acetic acid, propionic acid, butyric acid and valeric acid) in the EAD was not clearly investigated. Experimental examination of such organic acids is necessary to ensure proper destruction of long-chain fatty acids (LCFAs) via beta-oxidation. We also focused on how the electric current is governed in wastewater sludge

1

The symbols in ADM1 are defined in Tables 1 and 2.

MEC nCH4 R rE SRT TCD TSS V VSS WAS WE WCH4

microbial electrolysis cell amount of CH4 produced (mol) Gas constant (8.314 J/mol/K) energy recovery from the EAD () solids retention time (d) thermal conductivity detector total suspended solids (mg/L) volume of sludge (L) volatile suspended solids (mg/L) waste activated sludge electric energy consumed to drive the MEC reactions (J) energy recovered as CH4 from the EAD (J)

digestion by finding correlations with potential limiting factors, such as organic loading rate, conductivity and acetic acid concentration. Another aspect of this study is to explain whether the MEC reactions can increase the methane content in biogas production because the MEC cathode reaction creates an extra amount of hydrogen gas and hydrogenotrophic methanogenesis consumes carbon dioxide (4H2 þ CO2 / CH4 þ 2H2O). We also built a numerical model similar to Anaerobic Digestion Model No.1 (ADM1) developed by the IWA Task Group (Batstone et al., 2002) with the addition of the MEC component (Asztalos and Kim, 2015, accepted). The numerical model allowed us to examine a variety of components and microbes under various SRT, temperature and electrical current conditions. For example, experimental results on improved digestion performance with the MEC reactions under ambient temperature conditions were compared with conventional digester performance under various temperature conditions using the model. In addition, model simulation results can be used to explain how the MEC electrode reactions affect other biological reactions and contribute to enhancing anaerobic digestion performance. We also used the model to find equivalent temperature increases that would result in the same degree of improvement achieved with the MEC reactions. This application of the model will provide a meaningful conclusion on which option is more beneficial between increasing operation temperature and employing the MEC reactions in anaerobic digestion. Based on this comparison, we will be able to further discuss the energy requirement for the MEC reactions and that for heating wastewater sludge to attain mesophilic conditions. 2. Materials and methods 2.1. Reactor construction Two lab-scale anaerobic digestion reactors, a control digester and an electrically-assisted digester (EAD), were constructed with MEC components. The reactor bodies were made out of a thick polypropylene block in which a cylindrical hole (6.5 cm diameter and 6.5 cm depth with the effective liquid volume of 180 mL) was drilled. Two end-plates were fastened to the top and bottom of the bodies using metal tie rods and nuts placed along the perimeter of the reactor bodies (Fig. 2). Three carbon fiber brushes (2 cm diameter and 2.5 cm in length; Mill-Rose, OH) were pretreated in a muffle furnace at 450
C for 30 min (Wang et al., 2009) before they were placed in each digester as bioanodes. A single layer of stainless steel mesh was used as the MEC cathode without the use of any precious metal catalysts (total projected area of 135 cm2, AISI 304, 100-mesh, McMaster-Carr, OH). The stainless steel mesh was

Please cite this article in press as: Asztalos, J.R., Kim, Y., Enhanced digestion of waste activated sludge using microbial electrolysis cells at ambient temperature, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.05.045

J.R. Asztalos, Y. Kim / Water Research xxx (2015) 1e10

3

Fig. 1. (A) Reaction pathways in conventional anaerobic digestion where the final methanogenesis step is driven mainly by slow acetoclastic methangens. (B) Reaction pathways in electrically-assisted digestion where the final step is driven mainly by rapid hydrogenotrophic methanogens.

Fig. 2. (A) Schematic of electrically assisted digester. (B) Top view of the lab-scale electrically assisted digester.

wrapped around the interior wall of the reactor. The three bioanodes were fit through the top end-plate with an average distance of ~2 cm from the stainless steel mesh cathode. A small hole was drilled through the top plate to allow for feeding and withdrawing solution from the digesters. A plastic tube was glued to the top plate for biogas collection. 2.2. Reactor operation The constructed digesters were started with an influent containing 50% digested sludge from other lab-scale reactors and 50% waste activated sludge (WAS) from secondary clarifiers at a local municipal wastewater treatment facility. Note that the lab-scale inoculum reactors, which were originally started with digested sludge from a full-scale mesophilic high-rate anaerobic digester at a local wastewater treatment facility, were operated for at least 3 months with WAS at 22e23
C under a 14-day SRT condition. After the one-time start-up cycle, the digesters of this study were directly fed with WAS. The collected sludge was stored for up to 2 weeks at 4
C and was unaltered without any pretreatments. To avoid potential enrichment of psychro-tolerant microorganisms, the feed WAS was not stored longer than 2 weeks in a 4
C fridge. The composition of collected WAS from a local wastewater

treatment facility was not consistent with substantially varying chemical oxygen demand (7.89 ± 1.88 g COD/L) and volatile suspended solids (5.20 ± 1.20 g VSS/L) throughout the 5-month experiment. The digesters were operated under a stable temperature condition (22.5 ± 0.5
C) and were continuously mixed using magnetic stirrers. The MEC reactions in the EAD were induced using an external power supplier (GPS-1850D; GW Instek, CA) while the control digester was operated as a typical anaerobic digester by disconnecting the electrodes. The electric potential application (Eap) was 0.8 V during the EAD start-up for ~2 weeks and then held constant at 1.2 V throughout the experiment. Three different SRT conditions (7, 10 and 14 days) were examined in the experiment. For a given SRT condition, the digesters were operated under a continuous fed-batch mode where 90 mL (one half of the effective reactor volume) was replaced with untreated secondary sludge (i.e., WAS from the local facility). For instance, a 14-day SRT condition was achieved by feeding the
 mL ¼ 14 d . While the real SRT condigester every 7 days 90180 mL=7 d dition gradually increases with time during a continuous fed-batch cycle (e.g., 7e14 days for the 14-day SRT condition), the conventional definition for the mean retention time (reactor volume

Please cite this article in press as: Asztalos, J.R., Kim, Y., Enhanced digestion of waste activated sludge using microbial electrolysis cells at ambient temperature, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.05.045

4

J.R. Asztalos, Y. Kim / Water Research xxx (2015) 1e10

divided by the rate of sludge replacement) was used in this study to indicate the experimental condition. The gas head-space in the reactors was purged using nitrogen gas at the beginning of each fed-batch cycle. The initial SRT was 7 days and was lengthened to 10 days and then finally 14 days. For each SRT condition, at least 5 fed-batch cycles were repeated and results from all cycles were taken for analysis and discussion.

et al., 2008):

Z WE ¼

IEap dt

(2)

The energy recovered as methane gas (WCH4) was also determined by (Logan et al., 2008):

WCH4 ¼ nCH4 DHCH4 2.3. Experimental measurements For each cycle, influent and effluent sludge samples were measured for total suspended solids (TSS), volatile suspended solids (VSS) and total chemical oxygen demand (COD) in accordance with the standard method (Eaton et al., 2005). The percent VSS and COD removal for a given SRT condition was determined by averaging VSS and COD analysis results over 5 consecutive continuous fed-batch cycles. The sludge was also analyzed for conductivity and pH (SevenMulti, Mettler Toledo, Switzerland). The raw sludge pH was stable and neutral throughout the experiment at 6.8 ± 0.3. The conductivity of the sludge was also stable and relatively low at 1.26 ± 0.2 mS/cm. While high conductivity sludge is favorable in bioelectrochemical wastewater treatment, no additional electrolytes were added in the feed sludge in order to demonstrate that this EAD method is applicable for typical domestic wastewater sludge with low conductivity. Short-chain fatty acids were analyzed using a flame ionization detector-gas chromatography (FID-GC) instrument (Varian CP3800) equipped with a Stabilwax-DA column (Restek Corporation, PA). Prior to the FID-GC analysis, the sludge samples were centrifuged at 7000 RPM for 10 min and the supernatant was then acidified using a 3% vol. phosphoric acid solution as previously described (Eaton et al., 2005). Biogas produced in each digester was collected using a gas bag (250 mL capacity, Cali-5-Bond, Calibrated Instruments Inc., NY). Collected biogas was analyzed for CH4, CO2, N2, O2 and H2 using two thermal conductivity detector-gas chromatography (TCD-GC) instruments (Varian Star 3400 CX, Agilent Technologies, CA). One TCD-GC was equipped with a Porapak-Q packed column (Chromatographic Specialties Inc., Canada) for the separation of CH4, CO2, and N2 using helium as a carrier gas. The other TCD-GC was used to analyze H2 and O2 using a Molecular Sieve 5a column with nitrogen as a carrier gas. Electric current in the EAD (I) was determined by measuring the electric potential drop across a 10-U resistor every 20 min using a multimeter and data acquisition system (Model 2700, Keithley Instruments, OH). The electric current was normalized by the effective sludge volume in the digester (180 mL) to calculate the volume-based current density (or specific current). 2.4. Efficiency and recovery calculations The VSS and COD removal (in percent) was calculated by comparing experimental analysis results of the influent and effluent sludge. As defined by Logan et al. (2008), Coulombic efficiency (CE) is the ratio of the COD degraded by exoelectrogenic bacteria to the total COD removal (DCOD) on an electron basis.

Z 8 CE ¼

Idt

FVDCOD

(1)

I is the electric current in the EAD; F is the Faraday constant (96,485 C/mol); and V is the sludge volume (180 mL). The electric energy consumption in the EAD (WE) was calculated using (Logan

(3)

DHCH4 is the heat of combustion of methane (890.8 kJ/mol) (Haynes, 2013) and nCH4 is the amount of produced methane in moles. The methane production in moles (nCH4) was approximated from DCOD as demonstrated in Metcalf and Eddy et al. (2004): 
1 mol  CH4 nCH4 ¼ VDCOD 64 g  COD

(4)

The conversion factor between mol-CH4 and g-COD was found from oxidation of methane (CH4 þ 2O2 / CO2 þ 2H2O). The energy recovery (rE) is the ratio between WCH4 and WE as:

rE ¼

WCH4 WE

(5)

2.5. Numerical model development A steady-state version of Anaerobic Digestion Model No. 1 (ADM1) (Batstone et al., 2002) was modified and used to simulate the rate of biosolids destruction, microbial growth and change in organic concentration in both the EAD and control digester (Asztalos and Kim, 2015, accepted). For each of the 21 included components (Table 1), a steady-state mass balance equation was developed using the kinetic rate expressions provided by ADM1 (Table A in Supplementary Information). The system of nonlinear equations was solved using fixed-point iteration (Appendix B in Supplementary Information). Verification of the model was completed by using an example simulation provided by Rosen and Jeppsson (2006) (Table C in Supplementary Information). Note that the short-chain fatty acids (valerate, butyrate, propionate and acetate) in the influent sludge (Table 1) were obtained by averaging the experimental analysis results. The other components in Table 1

Table 1 Influent composition of sludge used for the mathematical model. Influent parameters were selected to match the total COD and fatty acid composition of the influent used in experimentation as well as the typical breakdown found in waste activated sludge. Model component

Symbol

Influent (mg-COD/L)

Composites Particulate inerts Carbohydrates Proteins Lipids Monosaccharide degraders Amino acid degraders LCFA degraders Valerate and butyrate degraders Propionate degraders Acetoclastic methanogens Hydrogenotrophic methanogens Monosaccharides Amino acids Long chain fatty acids Valerate Butyrate Propionate Acetate Hydrogen gas Methane gas

Xc Xin Xch Xpr Xli Xsu Xaa Xfa Xc4 Xpro Xac Xh2 Ssu Saa Sfa Sva Sbu Spro Sac Sh2 Sch4

4600 1300 320 320 500 10 10 10 30 30 30 30 300 300 10 13.58 15.23 4.00 43.51 0 0

Please cite this article in press as: Asztalos, J.R., Kim, Y., Enhanced digestion of waste activated sludge using microbial electrolysis cells at ambient temperature, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.05.045

J.R. Asztalos, Y. Kim / Water Research xxx (2015) 1e10

5

Table 2 Kinetic constants used for mathematical model. Parameters selected were the suggested values for ADM1 (Batstone et al., 2002) and were adjusted for 22
C. The pH for both digesters was fixed at 7. Model parameter

Symbol

Value

Unit

q

Max. specific disintegration rate Microbial decay rate (all) Max. specific hydrolysis rate (all) Half-saturation value for sugar utilization Max. specific sugar utilization rate Half-saturation value for amino acid utilization Max. specific amino acid utilization rate Half-saturation coefficient for LCFA utilization Max. specific LCAFA utilization rate Half-saturation value for butyrate/valerate utilization Max. specific butyrate/valerate utilization Half-saturation value for propionate utilization Max. specific propionate utilization Half-saturation value for acetoclastic methanogensis Max. specific acetoclastic methanogenesis rate Half-saturation value for hydrogenotrophic methanogenesis Max. specific hydrogenotrophic methanogenesis rate Yield of sugar degraders Yield of amino acid degraders Yield of LCFA degraders Yield of butyrate/valerate degraders Yield of propionate degraders Yield of acetoclastic methanogens Yield of hydrogenotrophic methanogens Fraction of inert particulate from composite decomposition Fraction of carbohydrate from composite decomposition Fraction of protein from composite decomposition Fraction of lipid from composite decomposition Fraction of LCFA from lipid decomposition Fraction of valerate from amino acid decomposition Fraction of butyrate from sugar decomposition Fraction of butyrate from amino acid decomposition Fraction of propionate from sugar decomposition Fraction of propionate from amino acid decomposition Fraction of acetate from sugar decomposition Faction of acetate from amino acid decomposition Fraction of H2 gas from sugar decomposition Fraction of H2 gas from sugar decomposition

kdis kdec khyd Ks,su ksu Ks,aa kaa Ks,fa kfa Ks,c4 kc4 Ks,pro kpro Ks,ac kac Ks,h2 kh2 Ysu Yaa Yfa Yc4 Ypro Yac Yh2 fi,xc fch,xc fpr,xc fli,xc ffa,li fva,aa fbu,su fbu,aa fpro,su fpro,aa fac,su fac,aa fh2,su fh2,aa

0.319 0.0127 7.018 318.64 40.178 300 40.178 400 4.305 127.46 15.366 48.964 9.825 95.592 5.098 0.002 35 0.1 0.08 0.06 0.06 0.04 0.05 0.06 0.3 0.2 0.2 0.3 0.95 0.23 0.13 0.26 0.27 0.05 0.41 0.4 0.19 0.06

d1 d1 d1 mg-COD/L d1 mg-COD/L d1 mg-COD/L d1 mg-COD/L d1 mg-COD/L d1 mg-COD/L d1 mg-COD/L d1 e e e e e e e e e e e e e e e e e e e e e

1.035 1.035 1 1.035 1.017 1 1.017 1 1.026 1.035 1.020 1.056 1.022 1.035 1.053 1.103 1

were fractionated from the average of the measured total COD in the experiment. To account for the MEC reactions in the EAD, the following equations were implemented in the model for acetate removal at the bioanode (Eq. (6)) and hydrogen gas production at the cathode (Eq. (7)) (Logan et al., 2008)

þ  CH3 COO þ 4H2 O/2HCO 3 þ 9H þ 8e

(6)

2Hþ þ 2e /H2

(7)

A fixed electric current density governed the rate of the electrode reactions in the model simulation. All simulations were performed at 22
C unless otherwise stated. The ADM1 kinetic parameters for 22
C are summarized in Table 2 (Batstone et al., 2002). For simulations at another temperature condition, the Arrhenius equation was used to adjust the kinetic parameters (Grady et al., 2011).

kðTÞ ¼ kð22
CÞqðT22Þ

(8)

Note that k is the kinetic parameter, T is the temperature in
C and q is the Arrhenius constant.

3. Results 3.1. VSS and COD removal The MEC reactions in the EAD expedited the removal of VSS by 5e10% under the 7- and 14-day SRT conditions (Fig. 3A). For the 7day SRT condition, a statistically significant difference was noted (p-value ¼ 0.002). For the 14-day SRT condition, although there was an observed difference in VSS removal between the control digester and EAD it was not found to be statistically significant (pvalue ¼ 0.297). A significant difference could not be established between the control digester and EAD for the 10-day SRT condition. It should be noted that the EAD performance in VSS removal was less dependent on the SRT condition with a small increase in the VSS removal from 26 to 28% while the control digester showed a greater variation from 16 to 22% VSS removal when the SRT was increased. The COD removal trend was consistent with that of the VSS removal result (Fig. 3B). The EAD removed 5e10% more total COD compared to the control digester for the 7- and 14-day SRT conditions. The 7-day COD removal results between the two digesters were found to be statistically significantly different (pvalue ¼ 0.038). Once again for the 14-day SRT condition, the observed difference between the EAD and control was not found to be statistically significant (p-value ¼ 0.442). The EAD showed no improvement in COD removal over the control digester for the 10day SRT condition. It should be noted that there is a more gradual increase in the percent COD removal with the increasing SRT for the

Please cite this article in press as: Asztalos, J.R., Kim, Y., Enhanced digestion of waste activated sludge using microbial electrolysis cells at ambient temperature, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.05.045

6

J.R. Asztalos, Y. Kim / Water Research xxx (2015) 1e10

Fig. 4. Effect of SRT on electric current generation in the EAD. The x-axis (number of continuous fed-batch cycles) was prepared by normalizing time by the length of a fedbatch cycle; thus one cycle unit is 7 days (14-day SRT), 5 days (10-day SRT) and 3.5 days (7-day SRT). The electric current density (specific current) was obtained by normalizing electric current by the sludge volume in the EAD (180 mL).

EAD (30%e34%) compared to the control digester (20%e28%). These results emphasize that the total COD removal was less dependent on SRT when MEC reactions were present.

the resistive potential loss was less than 2% of Eap of 1.2 V. Thus, the reactor was effectively designed for the treatment of low conductivity WAS. The Coulombic efficiency (CE) in the EAD was 15%, 16% and 13% for 7-, 10- and 14-d SRT, respectively, indicating that the MEC reactions contributed to the total COD removal by 13e16%. Also, these CE results are relatively independent on the examined SRT conditions, and this independency can be explained by the linear correlation between the organic loading rate and electric current (Fig. 5A).

3.2. Electric current and organic loading rate

3.3. Short-chain fatty acids

The volume-based electric current density (specific current) in the EAD varied substantially as the sludge composition was not consistent throughout the experiment. However, the current density was generally high between 15 and 25 A/m3 for the 7-day SRT while it was relatively low, usually below 10 A/m3 for the 10- and 14-day SRT (Fig. 4). Even though the sludge composition was not consistent, a clear linear correlation (R2 ¼ 0.82) was found between the average current density over each fed-batch cycle and the organic loading rate (Fig. 5A). Exoelectrogenic bacteria in bioelectrochemical systems are known to prefer acetate as an organic substrate to complicated organic compounds, such as sugars and other organic acids (Cheng and Logan, 2007; Catal et al., 2008). However, the average current density was not dependent on the influent or effluent acetic acid concentration (Fig. 5B and C), indicating that the MEC reactions are not rate-limiting the overall biosolids digestion. The conductivity of influent sludge was consistently low at 1.31 ± 0.25 mS/cm without any sludge pretreatment. The effluent conductivity was almost doubled in the EAD to 2.36 ± 0.23 mS/cm without any clear dependency on the SRT condition. The average current density was also not clearly correlated with the average conductivity (Fig. 5D), indicating that the resistive loss between the bioanode and cathode did not limit the electrode reaction. With the effluent conductivity of 2.36 mS/cm and average distance of 2.0 cm between the bioanodes and cathode, the resistive potential loss was 0.01e0.02 V for current density between 10 and 20 A/m3, indicating

The MEC reactions in the EAD not only accelerated acetate consumption, but also improved the removal of other shortchain fatty acids (Fig. 6). The acetic acid concentration in the EAD was consistently lower by 30e40% than that in the control digester effluent (Fig. 6A). The MEC reactions also noticeably enhanced the removal of propionic acid, iso-butyric acid and nbutyric acid (Fig. 6B, C and 6D). A discernable trend could not be established between exoelectrogenic activities and the removal of iso-valeric and n-valeric acid. For example, iso-valeric acid removal was not significantly enhanced by MEC reactions (Fig. 6E), and n-valeric acid concentration was significantly lower in the EAD (2.3 mg-COD/L) compared to the control (6.9 mgCOD/L) only when the digesters were operated under a 10-day SRT (Fig. 6F).

Fig. 3. (A) VSS removal and (B) COD removal in the control and EAD reactors. Results shown are the average concentration across 5 continuous fed-batch cycles. The error bars indicate the magnitude of the standard deviation (n ¼ 5).

3.4. Biogas and energy recovery Hydrogen gas was not detected in the GC analysis throughout the experiment and the biogas consisted primarily of CH4 (~95%) with a small fraction of CO2 (~5%). In addition, this biogas composition was consistent between the EAD and control digester. These results are similar to a recent study by Bo et al. (2014); however, they reported that only their MEC digester showed such a high CH4 content in biogas composition. The energy recovery (rE) increased gradually with the increasing SRT condition (326%, 336% and 371% for the 7-, 10- and

Please cite this article in press as: Asztalos, J.R., Kim, Y., Enhanced digestion of waste activated sludge using microbial electrolysis cells at ambient temperature, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.05.045

J.R. Asztalos, Y. Kim / Water Research xxx (2015) 1e10

7

Fig. 5. Average current density vs. (A) organic loading rate, (B) influent acetic acid concentration, (C) effluent acetic acid concentration, and (D) average conductivity. Acetic acid concentration data are from 10-day and 14-day SRTs only.

14-d SRT, respectively). The relatively high energy recovery indicates that the EAD can be operated as a net energy producer while treating wastewater sludge under low temperature conditions. The rate of CH4 production was estimated using Eq. (4). For the control digester, the CH4 production rate decreased gradually from 17.4 to 11.0 mL/d with the increasing SRT while the CH4 production rate was greater from 25.6 to 14.0 mL/d in the EAD (22.5
C). The enhanced CH4 production can be explained by the MEC reactions in the EAD. Note that the contribution of the MEC reactions to the CH4 production is equivalent to that to the COD removal (i.e., Coulombic efficiency), indicating 13e16% of CH4 produced in the EAD was contributed by the MEC reactions.

3.5. Simulation results The mathematical model predicted that the EAD (I ¼ 10 and 20 A/m3) achieves improved total COD removal only for SRT conditions between 2 and 15 days compared to the zero current condition (Fig. 7A). Also, the increasing electric current density from 10 to 20 A/m3 resulted with an improved total COD removal. The acetoclastic methanogen population (Xac) is expected to be lower in the EAD than that found in the control digester (Fig. 7B), indicating the bioanode replaces the role of acetoclastic methanogens. The hydrogenotrophic methanogen population (Xh2) was consistently higher in the EAD (Fig. 7C). This difference in the hydrogenotrophic methanogen population (Xh2) is a direct result of the enhanced hydrogen gas production at the MEC cathode. Acetate concentration (Sac) was consistently lower in the EAD due to the bioanode reaction accelerating acetate destruction (Fig. 7D). Hydrogen gas was rapidly consumed and converted to methane gas by hydrogenotrophic methanogens in both the EAD and control digester. As

a result, hydrogen gas concentration (Sh2) was always very low below 0.001 mg-COD/L.

4. Discussion 4.1. MEC reactions enhancing organic acid removal The bioanode in the EAD is known to directly oxidize acetate (Logan et al., 2008). Both the experimental results (Fig. 6A) and mathematical model results (Fig. 7D) show that the bioanode was successful at reducing the acetate (or acetic acid) concentration. An acetate accumulation was observed in the control reactor for the 10-day SRT, indicating the rate-limiting role of acetoclastic methanogens for the short SRT condition. Both digesters showed a trend with decreasing acetic acid concentration with increasing SRT (Fig. 6A). This trend was also found in the simulation results as the acetate concentration (Sac) starts to decrease for SRTs longer than 7 days (Fig. 7D). The enhanced removal of propionic acid, iso-butyric acid and nbutyric acid in the EAD (Fig. 6B, C and 6D) can be partially explained by their direct oxidation at the bioanode as exoelectrogenic bacteria are known to directly utilize the short-chain fatty acids (Liu et al., 2005a; Cheng and Logan, 2007). However, this trend becomes less noticeable for n- and iso-valeric acid (Fig. 6E and F), indicating that exoelectrogens do not preferably utilize valeric acids with presence of shorter chain fatty acids (such as acetic acid, propionic acid and butyric acid). The enhanced removal of propionic acid, n-butyric acid and isobutyric acid can also be induced indirectly via beta-oxidation as previously discussed (Asztalos and Kim, 2015, accepted). Since acetate is a product of the beta-oxidation reaction (Fig. 1), a lower concentration of acetate makes the reaction more

Please cite this article in press as: Asztalos, J.R., Kim, Y., Enhanced digestion of waste activated sludge using microbial electrolysis cells at ambient temperature, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.05.045

8

J.R. Asztalos, Y. Kim / Water Research xxx (2015) 1e10

Fig. 6. Organic acid concentrations. (A) Acetic Acid; (B) Propionic Acid; (C) isoButyric Acid; (D) nButyric Acid; (E) isoValeric Acid; (F) nValeric Acid. Data set for 7-day SRT condition was not included as there were not enough data points.

thermodynamically spontaneous, leading to a more favorable condition for the growth of beta-oxidizing microorganisms. However, the lower acetate concentration did not always result in enhanced removal of valeric acid in the EAD. This inconsistent observation with valeric acid can be explained by the small fraction (31%) of acetate produced from beta-oxidation whereas the majority of beta-oxidized valeric acid is converted to propionic acid and hydrogen gas (Batstone et al., 2002). It was reported that 57% of propionate and 80% of butyrate (Batstone et al., 2002) are converted to acetate, explaining why the MEC reactions had a more significant impact on beta-oxidation of propionic and butyric acids compared to valeric acids. In the model development, we assumed that short-chain fatty acids except for acetate are not directly oxidized by the bioanode because exoelectrogenic bacteria predominantly utilize acetic acid when it is present with other short-chain fatty acids. Thus, propionic acid, butyric acid and valeric acid remained relatively unchanged with the electric current in the mathematical model simulation (not shown). In addition, the yield coefficient for the growth of beta-oxidizing bacteria was also constant (Table 2) without considering the positive effect of low acetate concentration on their growth. In future work, the direct and indirect contributions of exoelectrogenic bacteria to oxidation of various organic

acids will be investigated to develop more sophisticated mathematical models. Since the total COD of the influent sludge varied substantially in the experiment (Fig. 3B), direct comparisons with the simulation results were not available. However, the percent COD removal in the control was relatively small (20e29%) while the EAD showed the greater COD removal (30e36%) in the experiment (Fig. 3B). This trend of the improved COD removal in the EAD was consistent with the model simulation results. The percent COD removal was 23e43% in the EAD (10 A/m3) for the SRT range from 7 to 14 days while it was smaller at 17e40% without the electric current (Fig. 7A). Note that the experimental results were not completely fit to the simulation results mainly due to the varying sludge composition and limited information on the influent sludge composition. However, the substantially reduced organic acid concentration and resulting improved COD removal in the EAD were consistently found both in the experimental and simulation results. 4.2. High purity methane biogas The biogas composition found from both digesters contained a high fraction of CH4 gas (~95%). In many recent studies, it has been

Please cite this article in press as: Asztalos, J.R., Kim, Y., Enhanced digestion of waste activated sludge using microbial electrolysis cells at ambient temperature, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.05.045

J.R. Asztalos, Y. Kim / Water Research xxx (2015) 1e10

9

Fig. 7. Model simulation results with varying SRT. (A) Effluent total COD; (B) Population of acetoclastic methanogens (Xac); (C) Population of hydrogenotrophic methanogens (Xh2); and (D) Acetate concentration (Sac).

reported that the MEC reactions do not invoke a change in the overall gas composition (Asztalos and Kim, 2015, accepted; Sun et al., 2015; Feng et al., 2015). In these previous studies, the biogas composition was similar to that found from conventional anaerobic digestion systems (i.e. ~65% CH4 and ~35% CO2). However, in this study, the collected biogas contained a consistently high CH4 fraction (93% in the EAD and 95% in the control). The relatively low temperature condition (22e23
C) is not expected to have a significant impact on the biogas composition as a previous study reported biogas compositions similar to those collected from mesophilic anaerobic digesters (Connaughton et al., 2006). Therefore, the observed high CH4 content in our experiments can hardly be attributed to the low temperature. Another potential reason for the different CH4 content from our previous work is the sludge composition, as WAS was used in this study while a mixture of primary sludge and WAS was used in our previous work (Asztalos and Kim, 2015, accepted). Digestion of agricultural and food wastes typically result in a high yield of methane gas (Zhang et al., 2014; Rincon et al., 2010) because they contain large amounts of lipids and other polymeric substances which stoichiometrically lead to higher CH4 content. Although no evident explanation was provided, the substantially high CH4 content of 95% in the biogas implies that the secondary sludge used in this study contained a large amount of complex lipids and polymeric substances compared to primary sludge. Based on this discussion, we suggest that secondary sludge be digested separately from primary sludge to produce high purity CH4 gas from digestion.

removal by up to 10% compared to that in the control digester at an SRT of 7 days (Fig. 3). In order to achieve the same degree of improvement (i.e., additional 10% COD removal) by increasing operation temperature, our simulation results indicate that the control digester would need to be operated at a temperature of ~35
C (Fig. 8). The total electrical energy input for the EAD per 7 day SRT cycle was approximately 1.0 kJ which is greater than energy required to heat a liquid volume of ~180 mL from 20 to 35
C

4.3. MEC reactions supplementing low temperature digestion With the induced MEC reactions, the EAD showed improved COD

Fig. 8. Model simulation for temperature effect on total COD removal in conventional anaerobic digestion (no electric current).

Please cite this article in press as: Asztalos, J.R., Kim, Y., Enhanced digestion of waste activated sludge using microbial electrolysis cells at ambient temperature, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.05.045

10

J.R. Asztalos, Y. Kim / Water Research xxx (2015) 1e10

(0.75 kJ). However, considering an additional energy requirement for maintaining the temperature at 35
C for 7 days, an EAD operating at ambient temperatures can substantially save the cost for digester operation with minimized heating energy input while it achieves a similar degree of sludge digestion to a conventional digester operated at mesophilic temperatures. Note that this discussion was derived based on the numerical simulation results; thus, additional future studies are necessary for experimental verification. 5. Conclusions The electrically assisted digester (EAD) expedited the VSS and COD removal by 5e10% compared to the control digester for all examined SRT conditions. The trends in the VSS and COD removal indicate that when the MEC reactions are present, the SRT has relatively minor effects on the performance of the digester. The MEC reactions in the EAD reduced the steady-state concentration of short-chain fatty acids, such as acetic acid, propionic acid, n-butyric acid and iso-butyric acid. Exoelectrogenic bacteria are considered to directly contribute to the oxidation of these short-chain fatty acids. In addition, the reduced acetate concentration in the EAD also made beta-oxidation reactions more thermodynamically favorable, indirectly enhancing the degradation of the organic acids. The biogas composition was not affected by the MEC reactions and both digesters produced high purity methane gas (~95%) from waste activated sludge. By comparing the experimental and mathematical simulation results, an EAD under ambient temperature conditions was found to perform similar to a conventional digester operating at mesophilic conditions. These results suggest that an EAD can substantially reduce construction and operation costs for wastewater sludge digestion. Acknowledgments This study was supported by New Faculty Start-up Fund (Department of Civil Engineering, McMaster University), Discovery Grants (RGPIN/435547-2013, Natural Sciences and Engineering Research Council of Canada) and Canada Research Chairs Program (950-2320518). The authors thank Ms. Anna Robertson and Mr. Peter Koudys for their help on equipment operation and reactor construction. The authors also thank the City of Hamilton for providing wastewater treatment sludge. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.watres.2015.05.045. References Ariunbaater, J., Panico, A., Esposito, G., Pirozzi, F., Lens, P.N.L., 2014. Pretreatment methods to enhance anaerobic digestion of organic solid waste. Appl. Energy 123, 143e156. Asztalos, J.R., Kim, Y., 2015. Lab-scale experiment and model study on enhanced digestion of wastewater sludge using bioelectrochemical systems. J. Environ.

Informatics. http://dx.doi.org/10.3808/jei.201500308 (accepted). Batstone, D.J., Keller, J., Angelidaki, I., Kalyuzhnyi, S.V., Pavlostathis, S.G., Rozzi, A., Sanders, W.T.M., Siegrist, H., Vavilin, V.A., 2002. Anaerobic Digestion Model No. 1 (ADM1). IWA Publishing, London, 77p. Bo, T., Zhu, X., Zhang, L., Tao, Y., He, X., Li, D., Yan, Z., 2014. A new upgraded biogas production process: coupling electrolysis cell and anaerobic digestion in singlechamber, barrel-shape stainless steel reactor. Electrochem. Commun. 45, 67e70. Bolzonella, B., Pavan, P., Battistoni, P., Cecchi, F., 2005. Meshophilic anaerobic digestion of waste activated sludge: influence of the solid retention time in the wastewater treatment process. Process Biochem. 40, 1453e1460. Catal, T., Li, K., Bermek, H., Hong, L., 2008. Electricity production from twelve monosaccharides using microbial fuel cells. J. Power Sources 175 (1), 196e200. Cheng, S., Logan, B.E., 2007. Sustainable and efficient biohydrogen production via electrohydrogenesis. Proc. Natl. Acad. Sci. U. S. A. 104, 18871e18873. Connaughton, S., Collins, F., O'Flaherty, V., 2006. Psychrophilic and mesophilic anaerobic digestion of brewery effluent: a comparative study. Water Res. 40 (13), 2506e2510. Eaton, A.D., Clesceri, L.S., Rice, E.W., Greenberg, A.E., 2005. Standard Methods for Examination of Water and Wastewater, 21st ed. APHA (American Public Health Association). Feng, Y., Zhang, Y., Chen, S., Quan, X., 2015. Enhanced production of methane from waste activated sludge by the combination of high-solid anaerobic digestion and microbial electrolysis cell with iron-graphite electrode. Chem. Eng. J. 259, 787e794. Grady, C.P.L., Daigger, G.T., Love, N.G., Filipe, C.D.M., 2011. Biological Wastewater Treatment, third ed. Taylor and Francis Group, Boca Raton. 1022 p. Haynes, W.M., 2013. CRC Handbook of Chemistry and Physics, 94th ed. CRC Press. Izumi, K., Okishio, Y.K., Niwa, C., Yamamoto, S., Toda, T., 2010. Effects of particle size on anaerobic digestion of food waste. Int. Bio-deterior. Biodegr. 64, 601e608. Liu, H., Cheng, S., Logan, B.E., 2005a. Production of electricity from acetate or butyrate using a single-chamber microbial fuel cell. Environ. Sci. Technol. 39, 658e662. Liu, H., Grot, S., Logan, B.E., 2005b. Electrochemically assisted microbial production of hydrogen gas from acetate. Environ. Sci. Technol. 39, 4317e4320. Logan, B.E., Call, D., Cheng, S., Hamelers, H.V.M., Sletuels, T.H.J.A., Jeremiasse, A., Rozendal, R.A., 2008. Microbial electrolysis cells for high yield hydrogen gas production from organic matter. Environ. Sci. Technol. 42, 8630e8640. Ma, J., Duong, T.H., Smits, M., Vestraete, W., Carballa, M., 2011. Enhanced biomethanation of kitchen waste by different pretreatments. Bioresour. Technol. 102, 592e599. Metcalf & Eddy, Inc, Tchobangolous, G., Burton, F.L., Stensel, H.D., 2004. Wastewater Engineering: Treatment and Reuse, fourth ed. McGraw Hill, New York. 1819 p. Reynolds, T.D., Richards, P., 1995. Unit Operations and Processes in Environmental Engineering, second ed. PWS Publishing Company, Boston. Rincon, B., Banks, C.J., Heaven, S., 2010. Biochemical methane potential of winter wheat (Triticum aestivum L.): Influence of growth stage and storage practice. Bioresour. Technol. 101 (21), 8179e8184. Rittmann, B.E., McCarty, P.L., 2001. Environmental Biotechnology: Principles and Applications. McGraw Hill, New York. Rosen, C., Jeppsson, U., 2006. Aspects on ADM1 Implementation within the BSM2 Framework. Technical Report. Dept of Industrial Electrical Engineering and Automation Lund University, Sweden. Rozendal, R.A., Hamelers, H.V.M., Euverink, G.J.W., Metz, S.J., Buisman, C.J.N., 2006. Principle and perspectives of hydrogen and production through biocatalyzed electrolysis. Int. J. Hydrog. Energy 31, 1632e1640. Sun, R., Zhou, A., Jia, J., Liang, Q., Liu, Q., Xing, D., Ren, N., 2015. Characterization of methane production and microbial community shifts during waste activated sludge degradation in microbial electrolysis cells. Bioresour. Technol. 175, 68e74. Valo, A., Carrere, H., Delgenes, J.P., 2004. Thermal, chemical and thermos-chemical pretreatment of waste activated sludge for anaerobic digestion. J. Chem. Technol. Biotechnol. 79, 1197e1203. Wang, X., Cheng, S., Feng, Y., Merril, M.D., Saito, T., Logan, B.E., 2009. The use of carbon mesh anodes and the effect of different pretreatment methods on power production in microbial fuel cells. Environ. Sci. Technol. 43, 6870e6874. Zhang, P., Lin, C.-J., Liu, J., Pongprueksa, P., Evers, S.A., Hart, P., 2014. Biogas production from brown grease using a pilot-scale high-rate anaerobic digester. Renew. Energy 68, 304e313.

Please cite this article in press as: Asztalos, J.R., Kim, Y., Enhanced digestion of waste activated sludge using microbial electrolysis cells at ambient temperature, Water Research (2015), http://dx.doi.org/10.1016/j.watres.2015.05.045