Electrochemical treatment for greywater reuse: effects of cell configuration on COD reduction and disinfection byproduct formation and removal

ANSI

American National Standards Institute

BDCM

Bromodichloromethane

BDD

Boron-doped diamond

COD

Chemical oxidation demand

Cu

Copper

DBCM

Dibromochloromethane

DBP

Disinfection byproducts

DC

Direct current

DPD

N,N Diethyl-1,4 Phenylenediamine Sulfate

EC

Electrochemical

EPA

Environmental Protection Agency

ESTCP

Environmental Security Technology Certification Program

H₂O₂

Hydrogen peroxide

HAA

Halo-acetic acid

IrO₂

Iridium(IV) oxide

MBRs

Membrane bioreactors

NSF

National Science Foundation

O_2⁻

Superoxide radicals

O₃

Ozone

OH

Hydroxyl radicals

ROS

Reactive oxygen species

RuO₂

Ruthenium oxide

TBM

Tribromomethane

TCM

Trichloromethane

THM

Trihalomethane

TiO₂

Titanium dioxide

UV

Ultraviolet

Zn

Zinc

Greywater, defined as wastewater from kitchen, bath, and laundry facilities, represents an important resource for reuse in irrigation, toilet flushing, and similar applications (Ghaitidak & Yadav 2013). Decentralizing water treatment and greywater reuse would alleviate domestic water demands and reduce the volumetric burden to wastewater treatment plants leading to considerable environmental and economic benefit (Imhof & Muhlemann 2005). In the interest of capitalizing on greywater reuse, increased attention has been paid to various decentralization methods such as anaerobic filtering followed by UV treatment (Couto et al. 2015), aerobic digestion with hydrogen peroxide treatment for disinfection (Teh et al. 2015), TiO₂ photocatalytic treatment for chemical oxygen demand (COD) removal (Chong et al. 2015a), and implementation of decentralized membrane bioreactors (MBRs) (Santasmasas et al. 2013). Electrochemical (EC) treatment of greywater, while not widely studied, offers advantages over other treatment methods such as operation at ambient temperature and pressures, no waste generation, and its ability to be readily combined with other treatment strategies (Anglada et al. 2009; Radjenovic & Sedlak 2015).

EC cells have been studied extensively for their ability to remove organic and inorganic contaminants in various industrial wastewater and water recycling streams (Bagastyo et al. 2013a; Radjenovic & Sedlak 2015), as well as for disinfection (Schaefer et al. 2017). The treatment mechanisms include direct and indirect oxidation/reduction as well as electrocoagulation and precipitation of suspended solids. Electrolysis of water can produce several reactive oxygen species (ROS) by oxidation of water at the anode, such as hydroxyl radicals (•OH), ozone (O₃), superoxide radicals (•O₂⁻), active chlorine species, and hydrogen peroxide (H₂O₂) all of which have the potential to remove COD in wastewater (Gu et al. 2006; Chong et al. 2015b). Here this concept was similarly applied within the context of greywater treatment.

EC oxidation of chloride-containing water produces chlorine and hypochlorous acid/hypochlorite which reacts with electron-rich moieties to form halogenated disinfection byproducts (DBPs) (Deborde & von Gunten 2008; Prasse et al. 2015). Several studies have shown that EC treatment results in formation of trihalomethanes (THMs), and haloacetic acids (HAA), and perchlorate that exceed regulatory levels (Bagastyo et al. 2011a, 2013a; Schaefer et al. 2017). Thus, mitigation of these DBPs is needed for the effective application of EC treatment for potable water reuse applications.

Several approaches have been studied to mitigate DBPs in electrochemically treated water, including photocatalysis (Uyak et al. 2007; Mayer et al. 2014), enhanced coagulation, and sorption to powdered activated carbon (Gerrity et al. 2009; Mayer et al. 2014). Cathodic reduction of THMs has been observed, but with the exception of experiments performed using a Cu/Zn cathode, treatment of trichloromethane has been poor (Li et al. 2004; Radjenovic & Sedlak 2015; Wang et al. 2016). Other studies have demonstrated the reductive cathodic dehalogenation of HAAs, although complete dehalogenation beyond monochloracetic acid may be difficult (Korshin & Jensen 2001; Esclapez et al. 2012). These findings, coupled with several recent studies that suggest divided EC cell configurations with sequential anodic-cathodic treatment can enhance removal of COD and organic contaminants (Montanaro et al. 2008; Yu & Kupferle 2008; Wang et al. 2009; El-Ghenymy et al. 2014; Schaefer et al. 2015).

In this current study, EC treatment for greywater reuse is evaluated. Specifically, we focus on COD removal using both undivided and divided EC cells where the former takes advantage of simultaneous oxidation and reduction whereas the latter separates the anolyte and catholyte with a membrane preventing parasitic consumption of reactive oxidants. In addition, the formation and reduction of DBPs is evaluated along with overall energy demands. Several cathode materials are evaluated. To our knowledge this is the first greywater reuse study to observe the potential for implementing divided EC treatment for COD removal and coupled with cathodic treatment of DBPs.

Characteristics of the raw and filtered greywater are presented in Table 1. Collection details and amendments to greywater are summarized in the Supplementary Material (available with the online version of this paper).

Parallel plate EC cells (ElectroCell North America, Inc.) were used for all bench-scale testing. The direct current (DC) power source was a Keithley 2220G-30-1 Dual-Output Programmable DC Power Supply. A summary of the experiments performed is provided in Table 2 and a schematic of the divided cell configuration can be found in Figure 1. Other EC cell configuration information can be found in the Supplementary Material.

Analytical methods are detailed in the Supplementary Material.

Results from the undivided experiments can be found in the Supplementary Material (available with the online version of this paper).

COD removal rates for the sequential anodic-cathodic divided EC cell experiments performed at 12.5 mA/cm² are shown for each cathode material in Figure 2. Overall, in the absence of sparging, COD removal was lowest using the stainless steel cathode and greatest using the boron-doped diamond (BDD) and graphite cathodes. Consistent with previous studies comparing anodic and cathodic COD removal from wastewater (Wang et al. 2009), COD removal was greater in the anolyte than in the catholyte (except for the stainless steel cathode), as oxidation of COD in the anolyte likely was facilitated by both hydroxyl radical generation, oxidation via oxidized chlorine species, and/or direct electron transfer (Bagastyo et al. 2011b, 2013b). Removal of COD in the catholyte can occur via generation of hydrogen peroxide and/or reduction at the cathode (Rao & Venkatarangaiah 2014). Measured hydrogen peroxide levels during cathodic treatment were typically below 1 mg/L, so it is unlikely that cathodic hydrogen peroxide generation played a dominant role in COD removal. The relatively poor anodic removal of COD when using the stainless steel cathode without sparging (panel (c) in Figure 2) is not readily explained. Air sparging in the anolyte solution resulted in an increase in the anodic COD removal rate of nearly a factor of 3 (panels (c) and (d) in Figure 2). This increase in COD removal in the anolyte may have been due to enhanced removal of competing electron donors such as Cl₂. Alternately, sparging could have removed some volatile intermediate oxidation products from the greywater COD.

The net EC energy required for a 50% COD removal in the divided EC cell configuration (sequential anodic-cathodic) for both the BDD and graphite cathodes was approximately 13 W-h L⁻¹, which is 24% less than that observed in the undivided cell configuration. For the stainless steel (not sparged), the energy demand for 50% COD removal was approximately 20 W-h L⁻¹, as opposed to 17 W-h L⁻¹ observed in the undivided configuration. The EC energy demand using a stainless steel cathode with anodic sparging was 11 W-h L⁻¹. These removal efficiencies are comparable to electrocoagulation combined with subsequent electrooxidation with BDD cathodes where a current density of 30 mA/cm² resulted in >90% COD removal wastewater at 30 mA/cm² (Dermentzis et al. 2016).

On the cathodic side, the improved energy efficiency for the BDD and graphite cathodes likely is due to direct cathodic reduction of the organics present in the greywater. The differences in COD removal rate in the anodic compartment among the three cathodes tested (particularly for the stainless steel) is not readily explained, although differences in anodic treatment have been previously observed to be dependent upon the cathode materials in divided EC cells (Rao & Venkatarangaiah 2014). The measured cell voltage and pH were not measurably impacted by the anode material, thereby eliminating these parameters as an explanation. One possible explanation is that the anode material impacted migration of cations across the membrane (via transformation at the cathode surface), thereby indirectly affecting anodic treatment.

Generation of free chlorine and chlorinated byproducts generally differed from observations made in the undivided EC cell experiments. Figure S1 (available online) shows, for all three anodes tested, free chlorine concentrations increased in the unsparged anolyte to between 100 and 200 mg/L, then free chlorine decreased in the catholyte by 45% in BDD, 53% in graphite, and 53% in stainless steel. Free chlorine generation was accompanied by chloride removal during anodic treatment, while molar increases in chloride were observed during cathodic treatment in divided cells as free chlorine was reduced. Free chlorine levels were similar to those observed under undivided conditions, but the free chlorine in the anolyte (pH = 2.2) was likely present as predominantly Cl₂, and as hypochlorite in the catholyte (pH = 12). Free chlorine levels were slightly greater when using the BDD and graphite cathodes compared to the stainless steel. This increased rate of chloride oxidation using the BDD was consistent with the higher rate of removal of COD, indicating that a greater rate of oxidation was observed using BDD cathodes (>50% loss in chloride) as opposed to 47% with the stainless steel cathode. In the presence of anodic sparging, free chlorine levels were substantially diminished from approximately 100 mg/L (non-sparged stainless steel cathode experiment) to less than 2 mg/L free chlorine by the end of cathodic treatment due to removal of Cl₂ in the anolyte.

THM results for each of the three cathode materials are shown in Figure S2 and HAA5 results for the BDD cathode are shown in Figure S3 (available online). For each cathode material, THMs did not exhibit any appreciable accumulation in the anolyte. Previous studies have shown HAA5 formation is favored at low pH (Singer 1994; Liang & Singer 2003; Bagastyo et al. 2011b), which likely explains why HAA5 generation was observed, but not THMs, in the anolyte. The rate of HAA5 formation in the anolyte (Figure S3) is similar (within a factor of 2) to that observed for the undivided configuration. Replication of HAA5 data was difficult to achieve especially on the anode side.

Once the anolyte solution was transferred to the catholyte and cathodic treatment was initiated, THM concentrations greatly increased to levels that were 4 times greater than those observed in the 12.5 mA/cm² condition for the undivided configuration. This result suggests that pH had a great impact on the extent of THM generation, and increasing the pH resulted in rapid formation of THMs (Figure S2). The initial increase in THMs upon transfer to the catholyte was similar for all three cathodes, suggesting THM formation was driven by bulk solution chemistry and the transformation of dissolved chlorine and hypochlorous acid to hypochlorite. For the sparged condition (panel (d) in Figure 3), final THM generation was near the analytical detection limit of 20 μg/L, likely because the free chlorine levels were greatly diminished.

The molar ratios of the individual THMs in the anolyte were generally consistent with those observed for the undivided electrochemical experiments (Figure S8, available online). These ratios remained consistent in the catholyte, even as THMs were being removed. For both the BDD and stainless steel cathodes, THM levels began to decline in the anolyte after the initial THM spike. This decrease was likely due to cathodic reduction at the anode (Esclapez et al. 2013; Garcia-Segura et al. 2015). Removal of HAA5 in the catholyte was also observed when using a BDD cathode (Figure S3). Thus, sequential anodic-cathodic treatment provides a potential approach for mitigating the net generation of DBPs.

Perchlorate generation was not observed under divided EC conditions, as perchlorate remained below the analytical detection limit of 1 μg/L in both the anolyte and catholyte. Thus, operating under divided EC conditions serves as a potential means to mitigate perchlorate formation.

To further evaluate the cathodic rate of DBP reduction, HAA5 and THM-spiked greywater (not subjected to anodic pre-treatment) was cathodically treated to verify and quantify HAA5 and THM removal. THM results are shown in Figure 3 and HAA5 results are shown in Figure 4. THM losses with zero applied current are likely due to volatilization and/or sorption; no HAA5 losses were observed at zero applied current. The difference in THMs and HAA5 between the control (no applied current) and applied current tests are assumed to be due to cathodic reductive treatment.

Regressed first-order rate constants for the individual THMs are shown in Table 3. Results show THM removal is reasonably described by first-order kinetics. THM removal is generally greater for the stainless steel cathode than for the BDD cathode. For the stainless steel cathode, individual THM removal rates were such that TCM > BDCM > DBCM ∼ TBM; this trend was not observed for the BDD cathode where all the THMs were removed at the same rate.

HAA5 removal was not reasonably described by first-order kinetics, thus no regression analyses are presented. However, comparison of the zero current controls to the applied current data show that measurable cathodic treatment occurred. Poly-halogenated acetic acids were removed at a greater rate than the mono-halogenated acetic acids, and cathodic reduction of the poly-halogenated acetic acids did not result in any apparent increase in the mono-halogenated acetic acids.

• Of the configurations evaluated, divided EC cells equipped with BDD cathodes present the most viable option for low energy COD removal in a greywater reuse scenario where COD is removed directly on the anode side and indirectly on the cathode side. Divided cells required less energy to achieve the same COD removal as undivided cells.

• Cathodic treatment using either BDD or stainless steel cathodes resulted in reduction of both THMs and HAA5. Poly-halogenated DBPs were removed more quickly without any accumulation of monohalogenated DBPs.

• Elevated perchlorate concentrations were observed using mixed metal oxide anodes and were not observed using the divided EC cell configuration. To our knowledge this has not been previously observed and should be considered when using MMO anodes for this type of treatment.

• Sparging of an acidic anolyte under divided conditions resulted in decreased DBP formation in the catholyte.

• Collectively, these findings suggest that divided electrochemical cells employing a boron doped diamond cathode and sparging of the anolyte show promise in terms of COD removal and DBP formation management for greywater reuse.

This project was funded by Environmental Security Technology Certification Program (ESTCP) Project #ER-201637. The authors would like to thank Paul Ho and Dung Nguyen for analytical support. The results and conclusions presented herein are those of the authors, and do not necessarily represent those of the United States Government, and no endorsement of the described technology is implied.