Chemical Engineering Journal 469 (2023) 144001 Contents lists available at ScienceDirect Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej Exploring key operational factors for improving hydrogen production in a pilot-scale microbial electrolysis cell treating urban wastewater Oscar Guerrero-Sodric, Juan Antonio Baeza *, Albert Guisasola GENOCOV, Department of Chemical, Biological and Environmental Engineering, School of Engineering, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain A R T I C L E I N F O A B S T R A C T Keywords: Bioelectrochemical systems (BES) are becoming popular technologies with a plethora of applications in the Bioelectrochemistry environmental field. However, research on the scale-up of these systems is scarce. To understand the limiting Hydrogen factors of hydrogen production in microbial electrolysis cell (MEC) at pilot-scale, a 135 L MEC was operated for Microbial electrolysis cell six months under a wide range of operational conditions: applied potential [0.8–1.1 V], hydraulic residence time Pilot scale Urban wastewater [1.1–3.9 d], and temperature [18–30 ◦ C], using three types of wastewater; synthetic (900 mg CODs L− 1), raw urban wastewater (200 mg CODs L− 1) and urban wastewater amended with acetate (1000 mg CODs L− 1). The synthetic wastewater yielded the maximum current density (1.23 A m− 2) and hydrogen production (0.1 m3 m− 3 d− 1) ever reported in a pilot scale MEC, with a cathodic recovery of 70% and a coulombic efficiency of 27%. In contrast, the use of low COD urban wastewater limited the plant performance. Interestingly, it was possible to improve hydrogen production by reducing the hydraulic residence time, finding the optimal applied potential or increasing the temperature. Further, the pilot plant demonstrated a robust capacity to remove the organic matter present in the wastewater under different conditions, with removal efficiencies above 70%. This study shows improved results compared to similar MEC pilot plants treating domestic wastewater in terms of hydrogen production and treatment efficiency and also compares its performance against conventional activated sludge processes. 1. Introduction Focusing on energy recovery, wastewater contains many different organic compounds (from simple carbohydrates to more complex mol­ The challenge of achieving sustainable wastewater treatment within ecules) that contain valuable energy stored within their chemical bonds the framework of addressing climate change is a critical challenge for [2]. The quantification of the energy content of domestic wastewater is a modern society. Despite advancements, municipal wastewater treat­ complex and multifaceted challenge. Numerous studies have attempted ment remains an energy-intensive process, mainly because of the large to establish correlations between this energy content and widely used aeration requirements of the activated sludge process, which constitutes parameters such as biological oxygen demand (BOD) or chemical oxy­ the primary operational cost of a standard wastewater treatment plant gen demand (COD). Combustion enthalpy of typical wastewater com­ (WWTP) [1]. However, environmental engineering is undergoing a pounds can be calculated from reference data [3], giving values (kJ g− 1 paradigm shift towards viewing wastewater as a resource rather than COD) of 13.6 for acetic acid, 14.6 for glucose and 15.6 for glutamic acid. merely waste. Then, a typical value around 15 kJ g− 1 COD seems to be a reasonable Abbreviations: AD, Anaerobic digestion; AEM, Anion exchange membrane; ARB, Anode respiring bacteria; BES, Bioelectrochemical system; BOD, Biological oxygen demand; CAS, Conventional activated sludge; CE, Coulombic Efficiency; COD, Chemical oxygen demand; CODS, Soluble chemical oxygen demand; EC, Energy consumption; EE, Energy efficiency; emf, Electromotive force; EP, Energy produced; HRAS, High-rate activated sludge; HRT, Hydraulic retention time; IEM, Ionic exchange membrane; j, Current density; MEC, Microbial electrolysis cell; MET, Microbial electrochemical technologies; NEP, Net energy production; OLR, Organic loading rate; ORR, Organic removal rate; PTFE, Polytetrafluoroethylene; PVC, Polyvinyl chloride; R2, R-squared; rCAT, Cathodic recovery; rE, Energy recovery; RSL, Return sludge liquor; SHE, Standard hydrogen electrode; SS, Stainless-steel; SWW, Synthetic wastewater; T, Temperature; TSS, Total suspended solids; UWW, Urban wastewater; UWWA, Amended urban wastewater; VFA, Volatile fatty acid; VSS, Volatile suspended solids; WAS, Waste activated sludge; WW, Wastewater; WWTP, Wastewater treatment plant; ΔGR, Gibbs free energy; ΔV, Applied potential; θ, Coefficient of temperature dependency; σ, Electrical conductivity. * Corresponding author. E-mail addresses: Oscar.Guerrero@uab.cat (O. Guerrero-Sodric), JuanAntonio.Baeza@uab.cat (J.A. Baeza), Albert.Guisasola@uab.cat (A. Guisasola). https://doi.org/10.1016/j.cej.2023.144001 Received 27 March 2023; Received in revised form 17 May 2023; Accepted 6 June 2023 Available online 7 June 2023 1385-8947/© 2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). O. Guerrero-Sodric et al. Chemical Engineering Journal 469 (2023) 144001 approach considering the inherent variability of urban wastewater such systems at pilot scale, a general reduction in reactor performance is fractionation, and it agrees with different reported values [4–6]. In a reported compared to bench-scale reactors [22,23]. Hydrogen produc­ modern society, each population equivalent (PE) produces about tion is usually poor, in the range 0.005–0.040 m3 m− 3 d− 1 [24–27], with 200–300 L of wastewater per day with an organic matter load about low CE (<50 %) and significant hydrogen leakage [28,29]. Also, the 80–120 g COD person− 1 d− 1 [7]. This means that, for example, in the reported range of organic matter removal in these studies is very broad case of Spain, with a population about 47 million people, an ideal (between 10 and 60%) due to the different influent load and composi­ maximum of 3⋅1016 J per year could be recovered from wastewater, tion, with reported hydraulic retention times (HRTs) between 4 and 48 which is equivalent to the energy of burning around 735,000 tonnes of h. crude oil in a modern power station. This value is certainly over­ Commonly organic loading rates (OLR) in systems for urban waste­ estimated, and a realistic goal may be to recover between 25 and 50% of water (UWW) treatment are in the range 0.04–1.6 g COD L− 1 d− 1 [30]. this energy [8]. In this sense, MECs must operate at a short HRT to be competitive with There are different processes for harnessing the energy contained in other technologies (e.g., activated sludge and anaerobic digestion sys­ wastewater. The most common methodology is anaerobic digestion tems) [31]. Achieving high COD removal efficiencies under short HRTs aiming at biogas production. Other promising alternatives aim at pro­ is one of the recent objectives of WWTP designers aiming only at COD ducing green hydrogen as an energy carrier. Hydrogen does not generate removal. However, poor anode-based oxidation rates are typically found greenhouse gas emissions when is transformed into electricity, and its in MECs due to low mixing [32], which is translated in higher volumes combustion enthalpy is higher than that of methane (122 kJ g− 1 vs 55 kJ or higher HRTs. Conventional stirring or recycling pumps are imple­ g− 1). Nevertheless, the world’s demand for hydrogen is nowadays mented to improve turbulence in the reactor to avoid mass transfer mainly supplied by the steam reforming process [9], and therefore, a limitations [27,29], but this may hinder the system energy balance. high environmental impact is associated with its production [10]. In this The highest hydrogen production in a double-chamber MEC pilot context, microbial electrochemical techniques (MET) have gained plant (0.031 m3 m− 3 d− 1) was reported in our previous study [27] and considerable attention as they can produce hydrogen from wastes in has been used as a reference to evaluate the technical feasibility of MECs devices that combine electrochemistry with the metabolism of electro­ [33], its environmental performance [34], and discussed in many other active microorganisms. These systems, known as microbial electrolysis works [35–37]. Despite promising results were obtained with our cells (MEC), can simultaneously treat wastewater and produce hydrogen reactor configuration, limited operational conditions were tested. [8]. For more details regarding the reactions occurring in MECs, please Hence, it was considered appropriate to further study the reactor in refer to the “Reaction Mechanisms” section in the Supplementary other scenarios to understand the limiting factors of MEC operation at Information. pilot scale. The 135 L double-chamber MEC was operated under a wide At lab-scale, MEC configurations can be divided into single and range of operational conditions and a critical assessment was performed double chamber depending on the presence of an ion exchange mem­ to determine how hydrogen yield and energy efficiency can be improved brane (IEM) that separates the anodic and cathodic compartments [18]. by modifying the operational parameters. Finally, the experimental ef­ Electroactive microorganisms grow as a biofilm on the anode whereas ficiency obtained in this study was used to evaluate the treatment of 1 abiotic hydrogen evolution takes place at the cathode. These microor­ m3 of UWW by an MEC and the results were compared with two existing ganisms, known as exoelectrogens or anode-respiring bacteria (ARB), technologies reported in the literature: conventional activated sludge can degrade organic matter using an insoluble electrode as the terminal (CAS) and high-rate activated sludge (HRAS). electron acceptor [11]. ARB possess the unique ability of transferring the electrons extracellularly, which travel from the anode to the cathode 2. Materials and methods through the electric circuit and reduce protons on the cathode, thereby forming hydrogen. Thermodynamics in an MEC show a positive Gibbs 2.1. Reactor design free energy (ΔGR) because of a negative electromotive force (emf). For instance, for the case of acetate as electron donor, the theoretical The pilot plant (Figs. S1–S3) consisted of a 220 L stainless-steel (SS) reduction potential is − 0.28 V vs standard hydrogen electrode (SHE) tank containing nine cassette-type cells, each of which working as a assuming standard conditions at a pH of 7. For the cathodic hydrogen double-chamber MEC. The cells were placed into the reactor, forcing the evolution reaction, the theoretical reduction potential is − 0.41 V vs SHE wastewater to circulate on both sides of them. A similar cell configu­ at standard conditions and pH of 7. Therefore, under these conditions, ration was used as reported in our previous study [27]. The reactor had a the emf is − 0.13 V and ΔGR is positive. Thus, the overall process is non- total anodic volume of 135 L once the nine cells were placed inside the spontaneous, and a minimum applied potential (ΔV) of 0.13 V will be reactor. The cells consisted of a central PVC frame (3 cm width × 36 cm required to drive the reactions [8]. This value is one order of magnitude length × 46 cm height) acting as a cathode chamber, containing pressed lower than the theoretical value for water electrolysis, 1.23 V [12] and is SS wool (Steel wool #2, Barlesa SL, Spain). The internal cathode section the main reason for the recent interest in this technology. However, the of each cell had a volume of 3.7 L. The cathode was separated from the inherent voltage losses increase to a large extent the ΔV requirements. anodes using an anion exchange membrane (AEM) on both sides (AMI- The common ΔV in MEC at lab-scale to drive hydrogen production is 7001S, Membranes International Inc., United States). The anodes were reported to be not lower than 0.4 V [13]), being usually between 0.6 and made of carbon felt (PX35 Carbon Felt, Zoltek™, United States) and 1.0 V [14,15]. were pressed to the membranes on both sides of the cathode through a MECs have shown promising results at laboratory scale in terms of SS mesh. The anodes were thermally pre-treated (400 ◦ C for 20 min) to hydrogen production, usually attaining values between 1–10 m3 H2 m− 3 improve biomass adhesion. Marine SS wire and SS connection strips reactor d− 1 [16–18], with high cathodic recoveries (rCAT) and high (Maranges SA, Spain) were used for the electrical connections. The coulombic efficiencies (CE) [19,20]. However, despite the large number membranes and the anodes were located between the cathode of manuscripts published about the fundamentals of bioelectrochemical compartment and two outer PVC frames (1 cm length × 36 cm width × hydrogen production, few studies are conducted at pilot scale. Jadhav 46 cm height in external dimensions with a 6 cm frame). The central et al. specified that less than 1% of the more than 20,000 studies about frame was provided with an olive fitting at the top that allowed the gas MET addressed the issue of scaling-up [21]. Even though lab-scale collection via a tube (Marprene® Tubing 9.6 mm BORE, Watson Mar­ research is essential, it may put out of sight the challenging issues low, United Kingdom) connected to a 10 L gas sample bag (FlexFoil® involved in the operation of MECs under real-scale conditions. MECs Standard, SKC, United States). The pieces were held together by tight­ need to be operated under conditions like those of WWTPs to bridge the ening 34 wing nuts onto bolts through 5 mm holes in the PVC frames. gap between lab and industrial adoption. Among the few experiences of The projected anode surface was 0.163 m2 per cell (i.e. a total of 1.47 m2 2 O. Guerrero-Sodric et al. Chemical Engineering Journal 469 (2023) 144001 in the reactor), which yielded a total area/volume ratio of 10.9 m2 m− 3. 2.4. Key performance indices (KPIs) Current density (j) in each cell was expressed as function of the 2.2. MEC operation projected anodic surface (A m− 2). Hydrogen production was expressed considering the total production in the plant (9 cells) and the anodic Each individual cell was powered by a separate power source (Pro­ volume (135 L) (m3 m− 3 d− 1) or surface (m3 m− 2 d− 1). grammable DC LAB Power Supply LABPS3005DN, Velleman Goup, The cathodic gas recovery (rCAT) compares the moles of electrons Belgium). Both voltage and current were digitally monitored using the required for hydrogen production with the moles of electrons arriving to AddControl software developed in LabWindows CVI by the research the cathode as current. The rCAT was calculated using Eq. (1): group [38]. The current intensity for each cell was recorded every 5 min for calculating several performance parameters. pH and temperature − 1 VH2 ⋅Vm,H ⋅bH2 ⋅F rCAT = ∫ tf 2 (1) were monitored in the anodic compartment using a pH probe (HACH pH t0 I dt electrode Crison5233, Spain) and a thermoresistance (Pt1000, Axiom­ atic, Spain). Conductivity was measured using a conductivity meter Being VH2: Volume of hydrogen produced (L), Vm,H2: Molar volume (COND 8, XS Instruments, Italy). The anodic temperature was controlled of hydrogen (24.06 L mol− 1 at 1 atm, 20 ◦ C), bH2: Number of electrons using the AddControl software and a silicone heating tape (HBSI 10 m, transferred per mole of hydrogen (2 mol e− mol− 1 H2), F: Faraday’s HORST GmbH, Germany) placed around the tank. constant (96485C mol− 1 e− ), t0: Initial time of the cycle (s), tf: Final time Three different types of wastewaters were used during the operation of the cycle (s), and I: Current intensity (A). of the plant (Table 1). In the first period, synthetic wastewater with The energy recovery (rE) refers to the ratio of the amount of energy acetate as sole carbon source and macronutrients was used. It was pre­ recovered as hydrogen and the electrical input. The rE was calculated pared in a refrigerated 2000 L tank with tap water, sodium acetate using Eq. (2): (2335 g), NH4Cl (305.67 g), K2HPO4 (56.19 g), KH2PO4 (43.91 g) and nH ⋅ΔHH2 50 mL of micronutrient solution [39], obtaining the concentrations re­ rE = ∫ tf 2 (2) I⋅ΔV dt ported in Table 1. In the second period, real UWW from a municipal t0 WWTP (Manresa, Spain) after primary settling was used, refilling the With nH2: Number of moles of hydrogen produced, ΔHH2: Heat 2000 L tank approximately every 10 days (see composition in Table 1). enthalpy of hydrogen (-285.8 kJ mol− 1), and ΔV: Applied potential (V). The levels of organic matter in the raw UWW after filtration (0.22 µm) The coulombic efficiency (CE) is the relation between the coulombs were in the range of 200–300 mg COD L− 1. Also, the raw influent was recovered as current and the coulombs that could be theoretically amended with acetate during the temperature experiments to increase generated from the substrate oxidation. It was calculated using Eq. (3): the organic concentration up to 1000 mg L− 1 of COD. The catholyte was ∫ tf a 5 g L− 1 NaCl solution. It was only renewed on day 30 before the CE = t0 I dt (3) continuous operation with UWW and its pH was measured weekly (it ΔC⋅VL ⋅MO− 21 ⋅bs ⋅F remained around 11–12 during the whole operation). Being ΔC: Difference of substrate concentration between the inlet The pilot plant was started-up with nine non-colonised cells, using an and the outlet of the reactor (g L− 1), VL: Anodic volume (135 L), MO2: anaerobic sludge inoculum from the same WWTP. The anodic chamber Molar mass of oxygen (32 g mol− 1), and bs: Number of moles of electrons was filled with synthetic medium and operated in batch mode with an transferred per mol of COD (4 mol e− mol− 1 COD). internal recycle of 150 L d− 1 to improve mixing conditions, using a timed peristaltic pump (520 FAM/R2, Watson Marlow, United Kingdom). After 20 days of enrichment period, the plant was moved to 2.5. Temperature effect study and response modelling continuous operation. During the whole operation, the influent was fed in the range 46–125 L d− 1 from the 2000 L supply tank and using 150 L The effect of temperature on hydrogen production and current d− 1 of recirculation flowrate. density was studied experimentally. The temperature range (18–30 ◦ C) was selected based on typical wastewater temperatures in an urban WWTP, which commonly varies between 10 and 25 ◦ C depending on the 2.3. Analytical methods season and geographic location, while temperatures above 30 ◦ C are highly unlikely [30]. To model the response, the Arrhenius equation (4) Organic matter concentration was measured using commercial COD was used and fitted to the experimental data. kits (LCK 514, Hach, United States) that covered the range of 100–2000 mg O2 L− 1 and a spectrophotometer (DR2800, Hach, United States). (4) 20) k(T) = k(20 ◦ C)⋅θ(T− Hydrogen, oxygen, nitrogen, and methane were analysed with a gas with k(T) and k(20 C) being hydrogen production rate or current ◦ chromatograph (7820-A, Agilent Technologies, United States). The total density at temperature T (◦ C) and at 20 ◦ C. hydrogen or methane production was calculated using the Gas Bag The model was fit to the experimental data by minimising the sum of Method presented by Ambler and Logan [40]. The detailed analytical the quadratic differences between the experimental and predicted data procedure can be found in the supplementary information. using the GRG nonlinear method of Excel Solver. The goodness of fit of the models was evaluated with the coefficient of determination R2 Table 1 calculated with Eq. (5). Average characteristics of the different wastewaters treated. ∑n Wastewater CODS (mg N-NH+ P-PO3− (yi − ŷi )2 4 (mg 4 (mg σ (mS pH R2 = 1 − ∑i=1 (5) L− 1) L− 1) L− 1) cm− 1) n 2 i=1 (yi − y) SWW 903 ± 6 40 ± 2 10 ± 1 2.8 ± 0.1 6.9 ± 0.1 UWW 247 ± 34 51 ± 4 5±1 1.1 ± 0.2 6.7 ± 2.6. CAS/HRAS comparison 0.2 UWWA 1069 ± 39 51 ± 4 5±1 3.2 ± 0.2 7.2 ± The MEC performance was compared to CAS and HRAS systems by 0.1 calculating the energy requirements of each technology for treating 1 m3 σ: Electrical conductivity; SWW: Synthetic wastewater; UWW: Urban waste­ of UWW with 300 mg COD L− 1. As these aerobic technologies only treat water; UWWA: Urban wastewater amended with acetate. wastewater and do not produce energy by themselves, methane 3 O. Guerrero-Sodric et al. Chemical Engineering Journal 469 (2023) 144001 production by anaerobic digestion (AD) of the purged sludge in the CAS (no primary), using a yield of 205 L CH4 kg− 1 VSSadded reported under and HRAS was considered as a revenue. CAS and HRAS performances similar conditions [44]. On the other hand, hydrogen production in the were assessed using data reported in the literature. The energy con­ MEC was calculated considering the efficiencies obtained in this study sumption was calculated considering the dominant energy requirements during the continuous operation with raw UWW at 20 ◦ C. for municipal wastewater treatment [41]: aeration energy for the bio­ The net energy production (NEP) was calculated as the difference logical processes and heating for anaerobic digester. Pumping energy between the energy consumed (EC) and the energy produced (EP), and was discarded as it was assumed that all technologies would account for the process energy efficiency (EE) was defined as the amount of energy similar energy consumption. The energy needs of the biological pro­ recovered as hydrogen or methane with respect to EC. cesses were considered as the amount of oxygen consumed in the aerobic bioreactor, and were calculated by Eq. (6) as proposed in [42]: 3. Results and discussion Mineralized COD SatDO Aeration energy = (6) 3.1. Start-up of the pilot plant with synthetic wastewater AE (SatDO − DO) With Aeration energy: Energy requirements for aeration (kWh d− 1), The cells were inoculated with anaerobic sludge and started in batch Mineralized COD: The influent COD mineralized fraction COD (kg d− 1), mode using synthetic medium for a period of 20 days. Once a stable AE: Aeration efficiency (1.5 kg O2 kWh− 1), [30], SatDO: Dissolved ox­ current density was obtained (around 1.0 A m− 2), the plant was shifted ygen concentration at saturation (10 mg L− 1 at 15 ◦ C), [42], and DO: to continuous mode with an inlet flowrate of 46 L d− 1 and an OLR of Dissolved oxygen concentration in the bioreactor (mg L− 1). 0.33 g L− 1 d− 1. The synthetic feeding was maintained for the first days The energy requirements of the anaerobic digester were considered under continuous conditions to assess the performance of the MEC pilot as the amount of energy needed to increase the temperature of the plant in an ideal scenario: i.e. readily biodegradable carbon source and a incoming waste activated sludge (WAS) to the digestor. A conservative moderate conductive medium (>2.5 mS cm− 1). value of 0.12 kWh m− 3 wastewater was assumed for the comparison The performance of the plant was studied under different fixed in­ [43], but the actual requirements were also calculated for each case by tensities (Stage 1) in each cassette cell: 0.10, 0.15 and 0.20 A cell− 1 (or Eq. (7): 0.61, 0.92 and 1.22 A m− 2, respectively) (Figs. 1. and S4). The power sources were configured to maintain the current intensity through each Q = m Cp ΔT (7) cell by automatically modifying the applied voltage within a range of − 1 Being Q: Heating energy requirements (kWh d ), m: Mass flow rate [0.00–1.00 V]. The maximum total gas production of 14.74 L d− 1 (0.01 of WAS (kg d− 1), Cp: Specific heat of WAS (considered as water: 1.17 m3 m− 2 d− 1 or 0.11 m3 m− 3 d− 1) with a 90% of hydrogen purity was 10− 3 kWh kg− 1 ◦ C− 1), and ΔT: Temperature difference between the obtained at the highest current intensity tested (0.20 A cell− 1) and at an digestor and the incoming WAS (◦ C). applied potential of 0.93 V. This is the highest hydrogen production The operational temperature in the anaerobic digestor was assumed reported at such a high MEC scale (>100L) and represents a significant to be 35 ◦ C, and the temperature of the A-stage was considered for the advancement on the way towards reaching viable hydrogen efficiencies. incoming WAS (15 ◦ C). Heat losses through the surface of the tank were Most of the studies on hydrogen production from UWW in cassette-type not considered. pilot-scale MECs report values in the range 0.005–0.031 m3 m− 3 d− 1 The gas mass flow rates were estimated using the ideal gas law under (Table 2). However, the performance and efficiency parameters ob­ normal conditions (20 ◦ C and 1 atm). Both the hydrogen produced by tained with the synthetic medium should be treated with caution, and the MEC pilot plant, and the methane produced by AD were considered not exactly considered for future scaling-up processes, since the char­ to be completely transformed into electricity. Thus, only the combustion acteristics of real UWW hinder to obtain these high-performance values, enthalpy of each gas was used. The amount of WAS generated was as will be discussed below. It was observed that rCAT varied as a function calculated performing the COD balances. Methane production was of the current intensity used and increased from 36 ± 1% with 0.10 A estimated considering the volatile suspended solids (VSS) in the WAS cell− 1 to 68 ± 2% with 0.20 A cell− 1. The major reported cause of low Fig. 1. Average key performance indices, applied potential and total gas production at different fixed intensities during the continuous operation with synthetic wastewater. Standard deviations are shown in the figure. 4 O. Guerrero-Sodric et al. Chemical Engineering Journal 469 (2023) 144001 Table 2 Summary of the data reported with cassette-type MECs at pilot-scale for urban wastewater treatment. 1 Pilot plant V WW (g COD OLR (g L− S /V (m2 HRT (h) T (◦ C) CODr rCAT CE ΔV jmax (A H2 (m3 m− 3 (L) L− 1) d− 1) m− 3) (%) (%) (%) (V) m− 2) d− 1) Heidrich et al. 120 UWW (0.45) 0.14 16.4 24 16 34 70 55 1.1 0.30 0.015 [25] Heidrich et al. 100 UWW 0.54 1/22 33 49 41 0.7 0.25 0.007 [28] (0.07–0.4) Baeza et al. 130 UWW 0.5/0.25 12.6 24/48 22 6/25 82 28 1.0 0.30 0.031/0.020 [27] (0.3–0.5) Cotterill et al. 175 UWW 1.60 34 5 11.5 63 10 21 0.9 0.29 0.005 [29] (0.3–0.5) Leicester et al. 36 RSL (2.2–4.3)a 4.4–8.6 20.2 0.36–432 10–20 52 b 1 b 6b – 1.11 b 0.011b [56] This work 135 SWW (0.9) 0.33 10.9 65 18 64 70 26 0.9 1.23 0.099 UWW 0.2–0.07 26–94 30–77 77–66 37–23 0.42–0.27 0.038–0.022 (0.2–0.3) UWWA (1.0) 0.40 65 63 88 17 0.34 0.040 V: Anodic working volume; S/V: Projected anodic surface to reactor volume ratio; CODr: COD removal; jmax: Maximum current density. a Reactor fed with return sludge liquor (concentrated filtrate of anaerobic digestion sludge). b Results reported at the optimal HRT found (0.5 d). rCAT are hydrogen leakages. It was determined that hydrogen leakages replaced by wastewater from a municipal WWTP after primary settling were constant in the range studied (6.23 ± 0.19 L H2 d− 1). Hence, the without any amendment. As expected, this real effluent showed reduced higher the hydrogen production, the lower the effect of hydrogen losses electrical conductivity (1.1 ± 0.2 mS cm− 1) and concentration of due to leakages, causing rCAT to increase with intensity.’’. organic matter (CODS = 247 ± 34 mg L− 1). The organic matter frac­ Hydrogen was the major compound in the gas collected, followed by tionation of real influents is very relevant because the substrate- methane. In all cases, the methane production remained constant utilization capabilities of most exoelectrogenic bacteria are limited to (around 1.48 ± 0.05 L CH4 d− 1). Methane could come from either the fermentation products, basically volatile fatty acids (VFA). Thus, anodic or the cathodic chamber. In the first case, methane from anaer­ fermentative bacteria are essential to provide the substrate required for obic organic matter degradation would be transferred to the cathode current generation from the potential complex molecules present in through the AEM. This hypothesis implies that most of the methane is domestic effluents. A very diverse microbial community needs to be produced in the bulk liquid and, therefore, methanogenesis depends on developed in the reactor, including exoelectrogens, hydrolytic and the anolyte conditions rather than on the anodic biofilm. Under this fermentative bacteria [18,45]. These populations establish a syntrophic hypothesis, the methane flux through the membrane should have been relation with positive interactions that favour an efficient anodic-based independent of the different applied potential, which seems to be in line oxidation of the complex organic compounds, which is the first step to with the observed results. Moreover, if methane had been produced in obtain hydrogen evolution at the cathode. However, the inherent vari­ the anode, CE would have been very low, since part of the entering ability of UWW properties challenges the ARB predominance in the substrate would be diverted to methanogenesis rather than to exoelec­ reactor and competitors for substrate can appear, thereby reducing the trogenesis. The CE values obtained were below 50% in all cases, which CE. For this reason, a comprehensive analysis of the plant performance seems to corroborate that part of the COD was not consumed by the under different scenarios is essential to understand the opportunities of exoelectrogens. these systems under real conditions. On the other hand, cathodic methane production implies the pres­ The plant was operated for>100 days under different operational ence of hydrogenotrophic methanogens in the cathode, which would conditions. In a first period (Stage 2), different applied potentials were occur if the cells were not completely sealed. In this scenario, the pH tested to find the conditions with better energy efficiency (section gradient would have been reduced by the direct contact between elec­ 3.2.1). Afterwards (Stage 3), the reactor was operated for>40 days at the trolytes and the cathode may have become biocompatible (otherwise, optimal applied potential of 0.90 V to understand the effect of temper­ the theoretical cathodic pH in a double chamber MEC should be around ature on the plant performance (section 3.2.2). Finally (Stage 4), the 12). Additionally, methanogenesis would depend on the current in­ plant was operated under different HRT values to determine its effect on tensity values as there would be different concentrations of hydrogen for the COD removal efficiency (section 3.2.3). The most important opera­ each case, thus affecting the methane formation kinetics. The pH of the tional parameters are shown in Table 3. catholyte was over 12, and hence it can be concluded that most of the methane collected came from the anodic chamber. Thus, it is necessary 3.2.1. Effect of applied potential on the plant performance to use IEM with lower permeability to prevent methane diffusion pro­ Increasing the input voltage can enhance H2 production and boost duced anaerobically in the anode. current density as shown by previous authors [46] at expenses of higher The highest rE attained (94 ± 2%) corresponded to the highest cur­ energy requirements. However, undesired electrochemical oxidation of rent density and hydrogen production, but energy neutrality was not the compounds in the wastewater and even water electrolysis may occur attained yet. The applied potential required was much higher than the when the applied potential is too high, resulting in a decrease of exoe­ theoretical value of 0.13 V (calculated under standard conditions, at a lectrogenic activity [47]. The applied potential has a significant impact pH of 7 and with acetate as sole electron donor) due to the potential losses of the system. As high current densities are required to obtain Table 3 significant hydrogen production, reducing the energy loss is essential to Key operational parameters of each experimental stage. achieve energy neutrality. Period ΔV (V) Inlet flowrate (L d− 1) T (◦ C) Stage 1 0.8–0.9 50 18 3.2. Operation in continuous mode of the pilot plant with UWW Stage 2 0.8–1.1 50 18 Stage 3 0.9 50 18–30 Stage 4 0.9 35–123 18 After the period with synthetic wastewater, the feed was entirely 5 O. Guerrero-Sodric et al. Chemical Engineering Journal 469 (2023) 144001 on energy efficiency since it is directly proportional to the power con­ focus on improving the structural design of the cells, enhancing the sumption. Thus, it is important to find its optimal value to improve MEC physical, chemical, and spatial properties of the electrodes and the performance at the lowest applied potential possible. membrane, to reduce the overall internal resistance to current flow. Fig. 2 shows the average current density and gas production values obtained at different applied potentials and Fig. S5 shows the current 3.2.2. Effect of temperature on the plant performance density profiles. The applied potential was increased in 0.10 V intervals, In this second experimental period, the effect of temperature was from 0.50 V to 1.10 V, which is the typical range for the pilot-scale studied for the first time in a pilot-scale MEC using both raw and operation with UWW. The best results for this set of experiments were amended (acetate addition) UWW to decouple the fermentation step obtained with an ΔV of 0.90 V, showing a maximum current density from exoelectrogenesis. Temperature was progressively increased from (0.21 A m− 2) and hydrogen production of 0.016 m3 m− 3 d− 1 (1.5 10− 3 18 ◦ C to 30 ◦ C in 39 days feeding raw UWW. Afterwards, the reverse m3 m− 2 d− 1), with rCAT = 64 ± 2%, rE = 91 ± 2%, and CE = 30 ± 7%. operation was carried out until room temperature was reached again, Higher ΔV were detrimental as they had no significant effect on the using amended UWW. Fig. S6 shows all the experimental profiles ob­ performance and only decreased energy efficiency to rE = 68 ± 2%. Due tained during this period. The reactor showed a robust capacity to to the inherent characteristics of real wastewater, current density and remove the organic matter present in the wastewater under different hydrogen production were down to 6 times lower than the reported in temperatures, with COD removal efficiencies above 70% (75 ± 4%) in section 3.1 using synthetic wastewater. The most likely reason is that the the case of raw UWW, and around 60% in the case of amended UWW (63 limiting step was the fermentation of complex organic matter to VFA ± 5%). Fig. 3 displays the dependence of current density and hydrogen rather than the exoelectrogenic activity. At ΔV above 0.90 V, the production on temperature for both raw and amended UWW. The cur­ exoelectrogenic ability to consume VFA was likely higher than the VFA rent density increased at 30 ◦ C up to 0.36 ± 0.02 A m− 2 with raw UWW production by fermentation. The composition of the gas was also richer which was 91 % higher than that obtained when operating at 18 ◦ C, in methane (around 30%), and although rE remained at similar values, while an increase of 40 % was observed when the influent was amended rCAT and CE were approximately 10% lower. The low COD strength of with acetate (0.50 ± 0.01 A m− 2). domestic effluents can limit MEC performance when compared to syn­ The simplified Arrhenius equations model fit are also presented in thetic wastewater. This difference in performance can lead to undersized Fig. 3, showing a good agreement (R2 up to 0.972) between experi­ designs if the productions and efficiencies of studies carried out under mental data and model prediction. The highest deviation appears for the unrealistic conditions are taken as a reference for future scale-up. Lab­ current density of UWW at the higher temperature tested (T = 30 ◦ C). oratory-scale research is essential but understanding the limitations of Typical activated sludge models [49] categorize biological processes such systems at larger scales with real effluents is crucial too. based on their temperature dependency θ: none (1.00), low (1.04), Any voltage below 0.8 V was not enough to drive hydrogen pro­ medium (1.07), and high (1.12). In this study, the operation with UWW duction. Again, the minimum ΔV to drive evolution was higher than the leads to a medium value (θ = 1.065), while operating with VFA- theoretical value (i.e. 0.13 V) due to overpotentials and ohmic losses of amended UWW resulted in a low temperature dependency (θ = the system [14]. Increasing the electrode surface area, improving the 1.031). These results suggest that the operation with raw UWW was electrode catalytic effect, increasing the temperature, or increasing the more sensitive to temperature changes than the operation with VFA- ARB-enrichment in the anodic biofilm can decrease the overpotential amended UWW. This observation agrees with fermentation being the losses [14]. Furthermore, ohmic losses can generally be minimized by limiting step, and that temperature has a strong effect on the rate of the reducing the electrode spacing, using a membrane with low resistivity, fermentation processes required to transform complex carbon sources and increasing the solution conductivity [11,15]. In real WWTPs, the use into VFA. When increasing the temperature with raw UWW, the limi­ of externally enriched cultures is unfeasible considering the large mi­ tation due to fermentation was mitigated and VFA production increased. crobial diversity that is entering in the reactor and the potential evolu­ In contrast, when the pilot plant was fed with acetate in excess, tion of the population and increasing the load or the conductivity by fermentative bacteria did not play a limiting role because there was dosing chemicals to the reactor is unfeasible [48]. Thus, research should already enough substrate for exoelectrogens and the effect of Fig. 2. Average key performance indices, current density and total gas production at different applied potential during the continuous operation with urban wastewater. Results for ΔV < 0.8 V are not shown in the figure. Standard deviations are shown in the figure. 6 O. Guerrero-Sodric et al. Chemical Engineering Journal 469 (2023) 144001 Fig. 3. Average current density and hydrogen production at different temperatures during the continuous operation with raw and amended urban wastewater. Experimental data are presented with standard deviation. Lines represent the model predictions (equation (4)). temperature on fermentation was not observed in the experimental temperature is not controlled. results. The value of rCAT (Fig. S6) increased progressively with temperature 3.2.3. Effect of HRT on the plant performance to a value of 80 ± 2% for raw UWW, and in the case of amended, up to HRT is a critical parameter for MEC performance. Low HRTs imply 85 ± 3%. Methane production was also present in this scenario for both an increase of the organic matter load at expenses of not having enough substrates, but it was not significantly affected by temperature. As a time for an efficient substrate conversion to hydrogen. In our previous result, the gas produced had a higher percentage of hydrogen. CE was work, the HRT was increased from 1 to 2 days to tackle an unsuccessful also improved by increasing the temperature with raw UWW. On one organic matter removal [27]. In fact, batch experiments showed that 10 hand, the organic removal rate (ORR) was not substantially influenced days were needed to reach 72% of organic matter removal. To the best of by temperature because it appeared to be limited by the low COD con­ our knowledge, there are no reports on the effect of HRT on a double- centration in the reactor. These low COD concentrations did not have chamber pilot-scale MEC treating real UWW, but some dynamic much room to be reduced. On the other hand, current density did in­ models are being used to study its effect [54,55]. In any case, reducing crease with temperature because COD was being fermented faster and the HRT remains a critical factor in making MEC technology feasible. VFAs availability for ARBs improved. These results suggest that ARBs In the last experimental period (from day 60 onwards), different uptake VFAs faster than methanogens. Nevertheless, a large variability volumetric flow rates were tested to study the effect of HRT on the in this parameter can be observed because of the variable influent reactor performance. Surprisingly, the plant showed 23 ± 2% increase concentration. CE decreased drastically below 15 ± 2% when the OLR in current density compared to Stage 2, although the same operational was increased from day 39 onwards (Fig. S6, VFA-amended UWW conditions were maintained (HRT = 2.9 d; ΔV = 0.9 V and T = 18 ◦ C), period), indicating an increase of the substrate consumed by other probably caused by a change in the biofilm populations after the period planktonic microorganisms present in the anode chamber, causing the with amended UWW. As a result, the pilot plant achieved energy reactor to show lower current density and gas production than that with neutrality with all volumetric flow rates tested. only synthetic wastewater, despite the OLR was increased to 0.40 g L− 1 Fig. 4 shows the average results for each HRT during the continuous d− 1. operation with raw UWW. Decreasing the HRT from 4 to 1 d caused an Some studies also showed a reduction in the MEC performance under increase of 86% in current density, reaching 0.42 A m− 2, with hydrogen low temperatures [50]. Heidrich et al. [28] operated a MEC pilot plant production following the same trend and a cathodic recovery of 77 ± fed with domestic wastewater at room temperature for 12 months of 2%. Specifically, hydrogen production increased from 0.024 to 0.038 m3 operation, but the general low levels of performance and high variance m− 3 d− 1 when the HRT decreased from 4 to 1 d (i.e. 52% increase). This in the data masked any temperature trend that may have occurred. boost in current density or hydrogen production through the inlet Moreover, the influence of temperature on the composition and pro­ flowrate can be attributed to two factors: i) increasing the OLR increases duction of VFAs in AD has been investigated by several authors. These both the COD concentration in the reactor and the exoelectrogenic ki­ systems share some of the hydrolytic and acetogenic steps with MECs netics and ii) it also decreases the residence time of the planktonic and are limited by low temperatures [51]. For instance, significant biomass in the system, favouring the washout of methanogens (i.e. less changes on the performance while increasing the temperature from anodic methane production and higher CE). On the other hand, higher 10 ◦ C to 35 ◦ C under mesophilic conditions have been reported [52,53]. flowrates increase the velocity of the liquid in the anode chamber, Thus, it can be stated that temperature plays an important role in MEC increasing the turbulence and improving external mass transfer from the operated with UWW as the metabolism of the entire microbial com­ bulk to the biofilm surface. However, the quality of the treated water munity is altered. Considering the daily and seasonal temperature was drastically reduced at low HRT. The OLR was increased from 63 ± 6 fluctuations, this system will provide variable hydrogen production if mg COD L− 1 d− 1 at HRT = 4 d to 200 ± 15 mg COD L− 1 d− 1 at HRT = 1 7 O. Guerrero-Sodric et al. Chemical Engineering Journal 469 (2023) 144001 Fig. 4. Average key performance indices, current density, and total gas production at different hydraulic residence times during the continuous operation with urban wastewater. Standard deviations are shown in the figure. d, but the maximum ORR achieved was around 70 ± 9 mg COD L− 1 d− 1 d− 1). We attribute this improvement in performance to different up­ at HRT = 1 d, leading to a significant reduction in the percentage of COD grades implemented in the pilot plant: i) the replacement of the current removed (from 75 ± 4% to 30 ± 3% (Fig. S7)). Despite the higher water and voltage monitoring system, removing the external resistance to treatment efficiencies at high HRTs, CEs were still lower due to bio­ determine the intensity in the cells, ii) the enhancement of the electrical logical organic matter degradation pathways different from exoelec­ wiring, replacing the crocodile clips with SS connection strips to avoid trogenesis. In fact, an increase in CE from 23 to 37% can be observed as rusting and iii) the replacement of PTFE tubbing with Marprene® in the the HRT decreases to 1 d. Despite the percentage of COD removed gas collection system. Furthermore, the microbial populations estab­ decreased at lower HRT, the energy recovery was higher. For the same lished in the anodic compartment could have played a critical role. reason, the amount of methane produced and transferred from the anode The organic matter removal efficiencies were similar to those ob­ was lower at low HRTs: 1.14 ± 0.22 L CH4 d− 1 at HRT = 4 d whereas tained in other studies, and HRT>3 days was required to achieve organic 0.72 ± 0.11 L CH4 d− 1 were recovered at HRT = 1 d. Thus, there is a matter removal of 70%. The only exception appears in Cotterill et al. trade-off between effluent quality and hydrogen production in the [29], who proved at pilot-scale that increasing the anode surface to choice of the optimal HRT. High HRT will result in higher percentage of volume ratio can result in a significant reduction of the required HRT, COD removal, whereas low HRT will result in higher hydrogen pro­ obtaining a good quality effluent. However, the lack of pilot-scale MEC duction. Low HRT configurations would require higher anodic surface to experiences available in the literature does not allow a reliable com­ volume ratios (m2anode m−reactor 3 ) to promote the exoelectrogenic substrate parison. More research is needed in the field. consumption over other biological processes. However, the changes in the biofilm structure due to these new hydraulic conditions could play 3.4. Comparing MECs versus conventional wastewater treatment an important role and deserve more study. technologies 3.3. Comparison with previous cassette-type MEC pilot plants MECs need to achieve high COD removal efficiencies with short HRT to compete with other existing technologies for municipal wastewater The implementation of MECs in current WWTPs has the potential to treatment, such as the A/B configuration [59,60]. A/B systems consist of transform organic pollutants into energy and valuable products as well an A-stage with short HRT and a B-stage where autotrophic nitrogen as reducing the current treatment costs with environmental benefits. removal occurs. The A-stage aims at converting the influent particulate, However, there is a wide gap between theoretical and practical perfor­ colloidal, and soluble COD into a single solid stream with minimum mance. As indicated in Table 2, the current densities and hydrogen mineralisation and low energy consumption by the stimulation of sludge production rates reported with pilot-scale cassette MECs are far from production, bacterial storage, and bioflocculation [61]. The surplus those obtained in lab-scale systems. Current densities reported in pilot sludge of the A-stage is diverted to an energy recovery process such as plants usually range between 0.25 and 0.30 A m− 2; only the plant pre­ anaerobic digestion, while the supernatant is treated in the B-stage. sented by Leicester et al. [56] reported a value of 1.11 A⋅m− 2 but The costs and possible revenues of the treatment of 1 m3 of UWW working with a smaller sized plant with anodic volume of 36L, and with a COD concentration of 300 mg L− 1 by different systems (CAS + treating return sludge liquor with higher organic matter concentration. AD, HRAS + AD and a MEC) were compared (Table 4). Wastewater On the other hand, lab-scale MECs can yield up to 2 A m− 2 [57,58]. treatment was successful in CAS with 85% of COD removal, but the EE Hydrogen production in double-chamber cassette-type MEC pilot re­ was the lowest (59 %) of all technologies investigated and the HRT actors is usually in the range 0.005–0.031 m3 m− 3 d− 1. This is much required was higher than in a HRAS. Rey-Martínez et al. [62] reported lower than 1–1.3 m3 m− 3 d− 1 reported in bench-scale [16,19]. In this the highest EC of all HRAS studied due to a higher COD mineralization, study, a maximum hydrogen production of 0.099 m3 m− 3 d− 1 was ob­ even though the maximum COD removal was achieved. On the other tained with synthetic wastewater. Furthermore, 0.038 m3 m− 3 d− 1 of hand, the best results were obtained in the HRAS published by Kinyua hydrogen were obtained feeding UWW, which is 23% higher than the et al. [63], with a NEP of − 0.140 kWh kg− 1 COD degraded and a COD maximum value reported previously under similar conditions [27]. This removal of 77%. Negative values of NEP indicated that the methane is equal to a 60% increase in the specific hydrogen production (m3 m− 2 produced contained more energy than the required by the system. 8 O. Guerrero-Sodric et al. Chemical Engineering Journal 469 (2023) 144001 Table 4 conventional wastewater treatment processes. Hence, future designs Comparison of conventional and high-rate activated sludge versus microbial should be aimed at: i) maximizing the anode surface to reactor volume electrolysis cell. ratio, making the exoelectrogenic activity in the reactor prevalent to Guven Kinyua Guven Rey- This study achieve high COD removal efficiencies at low HRT, and ii) optimizing et al. et al. et al. Martínez the cassette design to reduce the electrical resistance to current flow in 2019 2017 2019 et al. 2021 the cells. [64] [63] [64] [62] Technology CAS HRAS HRAS HRAS MEC 4. Conclusions T (◦ C) 20 20 20 18–22 18 DO (mg L− 1) 2 1 0.5 1 – • A MEC pilot plant was run under different operational conditions. HRT (h) 6 0.5 1 2.5 70 26 The best results were obtained with synthetic wastewater, yielding SRT (h) 96 13.4 9.6 28.8 the maximum current density (1.23 A m− 2) and hydrogen production – CODr (%) 86 77 59 88 74 30 H2 or CH4 0.015 0.023 0.017 0.016 0.066 0.041 (13.40 L⋅d− 1, 0.10 m3 m− 3 d− 1 and 1.0⋅10− 2 m3 m− 2 d− 1) ever re­ (m3 d− 1) ported in a pilot-scale MEC. H2 or CH4 0.057 0.098 0.098 0.059 0.325 0.455 • Low COD concentration in the UWW limited the plant performance. (m3 kg− 1 COD Specifically, an average hydrogen production of 5.1 ± 0.2 L d− 1 degraded) (0.038 m3 m− 3 d− 1 or 3.9⋅10− 3 m3 m− 2 d− 1) were obtained at HRT = EC (kWh 0.238 0.168 0.160 0.234 0.198 0.107 1 d. These are much better results than similar MEC pilot plants d− 1) treating domestic wastewater. EP (kWh 0.139 0.200 0.145 0.144 0.188 0.116 • High COD removal efficiencies were obtained with UWW under a d− 1) EC (kWh 0.924 0.727 0.904 0.886 0.924 1.189 wide range of operational conditions. kg− 1 COD • The effectiveness of double-chamber configuration in achieving degraded) viable current densities and recovering most of the hydrogen pro­ EP (kWh 0.541 0.867 0.819 0.545 0.892 1.293 duced was proved (rCAT > 80%). kg− 1 COD degraded) • The optimum applied potential for this configuration was found to be EPCOD (%) 14 22 21 14 24 33 0.9 V, far from the theoretical 0.13 V. NEP (kWh/ 0.383 − 0.140 0.085 0.341 0.031 − 0.104 • Low-moderate temperatures (<20 ◦ C) resulted in poor anode-based kg COD oxidation but when temperature was increased (>30 ◦ C), the degraded) fermentation step was improved, and hydrogen production increased EE (%) 59 119 91 62 95 109 from 2.0 to 5.7 L d− 1. DO: Dissolved oxygen; SRT: Solids retention time; EC: Energy consumed; EP: • The plant was able to produce 5.1 L H2 d− 1 at HRT = 1 d, which was Energy produced; EPCOD: Energy recovered from removed COD from influent; 52% more than that obtained operating at HRT = 4 d but followed by NEP: Net energy production; EE: Energy efficiency. a significant reduction in the percentage of COD removed (from 77 ± 1% to 30 ± 3%). Guven et al. [64] reported lower percentages of mineralized and purged • The MEC performance was compared to CAS and HRAS. MEC have COD (operating at HRT < 1 h and SRT < 14 h) than those of Kinyua et al. the potential to become an alternative to conventional wastewater [63], which resulted in a 23% reduction in methane production, treatment in terms of energy recovery, but the limited volumetric decreasing EE to 91%. treatment capacity hinders its full-scale implementation. Considering the results of our MEC pilot plant, 0.066 m3 d− 1 of hydrogen could be produced with a NEP of − 0.031 kWh kg− 1 COD Declaration of Competing Interest degraded at HRT = 70 h, while an average of 0.019 ± 0.004 m3 d− 1 of methane would be generated with the HRAS + AD (average NEP = 0.095 The authors declare that they have no known competing financial ± 0.024 kWh kg− 1 COD degraded). Additionally, a NEP of − 0.104 kWh interests or personal relationships that could have appeared to influence kg− 1 COD degraded could be achieved with an EE of 109%, if HRT in the the work reported in this paper. MEC was decreased to 26 h. Nonetheless, the MEC EC at HRT = 26 h was the highest of all technologies compared (1.19 kWh kg− 1 COD Data availability degraded). Therefore, the applied potential in the MEC must be decreased to make it less energy-intensive, since the energy consump­ Data will be made available on request. tion is directly proportional to the ΔV. MECs have the potential to become a realistic alternative to CAS or HRAS in terms of energy recovery because they recover similar or even Acknowledgements more energy from the organic compounds in wastewater compared to CAS/HRAS + AD. Additionally, the sludge post-treatment requirements This work was supported by the LIFE + NIMBUS project (LIFE19 appear to be considerably lower considering the total suspended solids ENV/ES/000191) (www.life-nimbus.eu). The authors are members of (TSS) in the outlet of the MEC pilot plant (<20 mg L− 1). Also, during the the GENOCOV research group (Grup de Recerca Consolidat de la Gen­ operation with the MEC pilot plant, an average of 15 ± 3% of N-NH+ eralitat de Catalunya, 2021 SGR 515, www.genocov.com). 4 and 8 ± 1% of P-PO3− 4 removal was observed (Fig. S8), possibly to cover the nutrient requirements of ARB and other microorganisms that might Appendix A. Supplementary data 3− have activity. By contrast, in a HRAS, N-NH+ 4 and P-PO4 removal can account up to a 30% and 12%, respectively [62]. Thus, it appears to be Supplementary data to this article can be found online at https://doi. more feasible to avoid nitrification in an MEC than in a HRAS (if it is org/10.1016/j.cej.2023.144001. intended to be used as A-stage in an A/B system). However, the HRT required in an MEC to achieve a similar COD removal to that of a HRAS References may be several times higher. The successful implementation of such [1] M. Molinos-Senante, F. Hernandez-Sancho, R. Sala-Garrido, Benchmarking in technology on a larger scale is currently hindered by the important wastewater treatment plants: a tool to save operational costs, Clean Technol. differences in volumetric treatment capacity between MECs and Environ. Policy 16 (2014) 149–161, https://doi.org/10.1007/s10098-013-0612-8. 9 O. Guerrero-Sodric et al. Chemical Engineering Journal 469 (2023) 144001 [2] Z. Dai, E.S. Heidrich, J. Dolfing, A.P. Jarvis, Determination of the relationship [28] E.S. Heidrich, S.R. Edwards, J. Dolfing, S.E. Cotterill, T.P. Curtis, Performance of a between the energy content of municipal wastewater and its chemical oxygen pilot scale microbial electrolysis cell fed on domestic wastewater at ambient demand, Environ. Sci. Technol. Lett. 6 (2019) 396–400, https://doi.org/10.1021/ temperatures for a 12 month period, Bioresour. Technol. 173 (2014) 87–95, acs.estlett.9b00253. https://doi.org/10.1016/j.biortech.2014.09.083. [3] D.R. Lide, CRC Handbook of Chemistry and Physics, 96th Edition, 2015-2016, [29] S.E. Cotterill, J. Dolfing, C. Jones, T.P. Curtis, E.S. Heidrich, Low temperature Handbook of Chemistry and Physics. (2016). doi:10.1136/oem.53.7.504. domestic wastewater treatment in a microbial electrolysis cell with 1 m 2 anodes: [4] I. Shizas, D.M. Bagley, M. Asce, Experimental determination of energy content of towards system scale-up, Fuel Cells 17 (2017) 584–592, https://doi.org/10.1002/ unknown organics in municipal wastewater streams, Am. Soc. Civil Eng. 130 fuce.201700034. (2005) 45–53, https://doi.org/10.1061/(ASCE)0733-9402(2004)130:2(45). [30] G. Tchobanoglous, H.D. Stensel, R. Tsuchihashi, F. Burton, M. Abu-Orf, G. Bowden, [5] E.S. Heidrich, T.P. Curtis, J. Dolfing, Determination of the internal chemical energy W. Pfrang, Metcalf & Eddy. Wastewater Engineering. Treatment and resource of wastewater, Environ. Sci. Tech. 45 (2011) 827–832, https://doi.org/10.1021/ recovery, 5th ed., McGray-Hill Education, 2014. es103058w. [31] J.J. Fornero, M. Rosenbaum, L.T. Angenent, Electric power generation from [6] B. Korth, T. Maskow, S. Günther, F. Harnisch, Estimating the energy content of municipal, food, and animal wastewaters using microbial fuel cells, Electroanalysis wastewater using combustion calorimetry and different drying processes, Front. 22 (2010) 832–843, https://doi.org/10.1002/elan.200980011. Energy Res. 5 (2017) 1–8, https://doi.org/10.3389/fenrg.2017.00023. [32] I.S. Michie, J.R. Kim, R.M. Dinsdale, A.J. Guwy, G.C. Premier, The influence of [7] G. Kiely, Environmental Engineering, McGraw-Hill, International (UK) Limited, anodic helical design on fluid flow and bioelectrochemical performance, Bioresour. 1997. Technol. 165 (2014) 13–20, https://doi.org/10.1016/j.biortech.2014.03.069. [8] B.E. Logan, Microbial Fuel Cells (2008). [33] D. Leicester, J. Amezaga, E. Heidrich, Is bioelectrochemical energy production [9] A. Haryanto, S. Fernando, N. Murali, S. Adhikari, Current status of hydrogen from wastewater a reality? Identifying and standardising the progress made in production techniques by steam reforming of ethanol: a review, Energy Fuel 19 scaling up microbial electrolysis cells, Renew. Sustain. Energy Rev. 133 (2020) (2005) 2098–2106, https://doi.org/10.1021/ef0500538. 110279, https://doi.org/10.1016/j.rser.2020.110279. [10] P.L. Spath, M.K. Mann, Life cycle assessment of hydrogen production via natural [34] J. Chen, W. Xu, X. Wu, N.a. Lu, T. Wang, H. Zuo, System development and gas steam reforming, National Renewable Energy Lab.(NREL), Golden, CO (United environmental performance analysis of a pilot scale microbial electrolysis cell for States), 2000. hydrogen production using urban wastewater, Energy Convers Manag. 193 (2019) [11] H. Liu, S. Grot, B.E. Logan, Electrochemically assisted microbial production of 52–63, https://doi.org/10.1016/j.enconman.2019.04.060. hydrogen from acetate, Environ. Sci. Tech. 39 (2005) 4317–4320, https://doi.org/ [35] T. Fudge, I. Bulmer, K. Bowman, S. Pathmakanthan, W. Gambier, Z. Dehouche, S. 10.1021/es050244p. M. Al-salem, A. Constantinou, Microbial Electrolysis Cells for Decentralised [12] R.A. Rozendal, H.V.M. Hamelers, R.J. Molenkamp, C.J.N. Buisman, Performance of Wastewater Treatment: The Next Steps, (2021). single chamber biocatalyzed electrolysis with different types of ion exchange [36] R. Rousseau, L. Etcheverry, E. Roubaud, R. Basséguy, M.L. Délia, A. Bergel, membranes, Water Res. 41 (2007) 1984–1994, https://doi.org/10.1016/j. Microbial electrolysis cell (MEC): Strengths, weaknesses and research needs from watres.2007.01.019. electrochemical engineering standpoint, Appl. Energy 257 (2020), 113938, [13] J. Ditzig, H. Liu, B. Logan, Production of hydrogen from domestic wastewater using https://doi.org/10.1016/j.apenergy.2019.113938. a bioelectrochemically assisted microbial reactor (BEAMR), Int. J. Hydrogen [37] P. Dange, S. Pandit, D. Jadhav, P. Shanmugam, P.K. Gupta, S. Kumar, M. Kumar, Energy 32 (2007) 2296–2304, https://doi.org/10.1016/j.ijhydene.2007.02.035. Y.-H. Yang, S.K. Bhatia, Recent developments in microbial electrolysis cell-based [14] R.A. Rozendal, H.V.M. Hamelers, G.J.W. Euverink, S.J. Metz, C.J.N. Buisman, biohydrogen production utilizing wastewater as a feedstock, Sustainability. 13 Principle and perspectives of hydrogen production through biocatalyzed (2021) 8796, https://doi.org/10.3390/su13168796. electrolysis, Int. J. Hydrogen Energy 31 (2006) 1632–1640, https://doi.org/ [38] J.A. Baeza, Advanced direct digital control (AddControl): lessons learned from 20 10.1016/j.ijhydene.2005.12.006. years of adding control to lab and pilot scale treatment systems, in: in: 13th IWA [15] S. Cheng, B.E. Logan, High hydrogen production rate of microbial electrolysis cell Conference on Instrumentation, Control and Automation. ICA 2022, Tsinghua (MEC) with reduced electrode spacing, Bioresour. Technol. 102 (2011) 3571–3574, University, Beijing (China), 2022, pp. 13–15. https://doi.org/10.1016/j.biortech.2010.10.025. [39] G.J.F. Smolders, J. van der Meij, M.C.M. van Loosdrecht, J.J. Heijnen, [16] Y. Wang, W.-Q. Guo, D.-F. Xing, J.-S. Chang, N.-Q. Ren, Hydrogen production using Stoichiometric model of the aerobic metabolism of the biological phosphorus biocathode single-chamber microbial electrolysis cells fed by molasses wastewater removal process, Biotechnol. Bioeng. 44 (7) (1994) 837–848, https://doi.org/ at low temperature, Int. J. Hydrogen Energy 39 (33) (2014) 19369–19375, https:// 10.1002/bit.260440709. doi.org/10.1016/j.ijhydene.2014.07.071. [40] J.R. Ambler, B.E. Logan, Evaluation of stainless steel cathodes and a bicarbonate [17] X.M. Guo, E. Trably, E. Latrille, H. Carrre, J.P. Steyer, Hydrogen production from buffer for hydrogen production in microbial electrolysis cells using a new method agricultural waste by dark fermentation: a review, Int. J. Hydrogen Energy 35 for measuring gas production, Int. J. Hydrogen Energy 36 (2011) 160–166, (2010) 10660–10673, https://doi.org/10.1016/j.ijhydene.2010.03.008. https://doi.org/10.1016/j.ijhydene.2010.09.044. [18] N. Montpart, L. Rago, J.A. Baeza, A. Guisasola, Hydrogen production in single [41] L.J. Pakenas, Energy efficiency in municipal wastewater treatment plants: chamber microbial electrolysis cells with different complex substrates, Water Res. technology assessment, New York State Energy Research and Development 68 (2015) 601–615, https://doi.org/10.1016/j.watres.2014.10.026. Authority, 1995. [19] R.C. Wagner, J.M. Regan, S.E. Oh, Y. Zuo, B.E. Logan, Hydrogen and methane [42] K. Svardal, H. Kroiss, Energy requirements for waste water treatment, Water Sci. production from swine wastewater using microbial electrolysis cells, Water Res. 43 Technol. 64 (2011) 1355–1361, https://doi.org/10.2166/wst.2011.221. (2009) 1480–1488, https://doi.org/10.1016/j.watres.2008.12.037. [43] R. Khiewwijit, H. Temmink, H. Rijnaarts, K.J. Keesman, Energy and nutrient [20] L. Rago, J.A. Baeza, A. Guisasola, Increased performance of hydrogen production recovery for municipal wastewater treatment: how to design a feasible plant in microbial electrolysis cells under alkaline conditions, Bioelectrochemistry 109 layout? Environ. Model. Softw. 68 (2015) 156–165, https://doi.org/10.1016/j. (2016) 57–62, https://doi.org/10.1016/j.bioelechem.2016.01.003. envsoft.2015.02.011. [21] D.A. Jadhav, S.-G. Park, S. Pandit, E. Yang, M. Ali Abdelkareem, J.-K. Jang, K.- [44] A. Trzcinski, C. Wang, D. Zhang, W. Ang, L. Lin, T. Niwa, Y. Fukuzaki, W. Ng, J. Chae, Scalability of microbial electrochemical technologies: applications and Performance of A-stage process treating combined municipal-industrial challenges, Bioresour. Technol. 345 (2022) 126498, https://doi.org/10.1016/j. wastewater, Water Sci. Technol. 75 (2017) 28–38, https://doi.org/10.2166/ biortech.2021.126498. wst.2016.511. [22] F. de Fouchécour, V. Larzillière, T. Bouchez, R. Moscoviz, Systematic and [45] M. Kokko, S. Epple, J. Gescher, S. Kerzenmacher, Effects of wastewater quantitative analysis of two decades of anodic wastewater treatment in constituents and operational conditions on the composition and dynamics of bioelectrochemical reactors, Water Res. 214 (2022) 118142, https://doi.org/ anodic microbial communities in bioelectrochemical systems, Bioresour. Technol. 10.1016/j.watres.2022.118142. 258 (2018) 376–389, https://doi.org/10.1016/j.biortech.2018.01.090. [23] A. Guisasola, J.A. Baeza, A. Marone, É. Trably, N. Bernet, Opportunities for [46] H. Hu, Y. Fan, H. Liu, Hydrogen production using single-chamber membrane-free Hydrogen Production from Urban/Industrial Wastewater in Bioelectrochemical microbial electrolysis cells, Water Res. 42 (2008) 4172–4178, https://doi.org/ Systems, in: Microbial Electrochemical Technologies, CRC Press, Boca Raton : CRC 10.1016/j.watres.2008.06.015. Press, [2020], 2020: pp. 225–243. https://doi.org/10.1201/9780429487118-15. [47] R.C. Tice, Y. Kim, Methanogenesis control by electrolytic oxygen production in [24] A. Escapa, M.I.I. San-Martín, R. Mateos, A. Morán, Scaling-up of membraneless microbial electrolysis cells, Int. J. Hydrogen Energy 39 (2014) 3079–3086, https:// microbial electrolysis cells (MECs) for domestic wastewater treatment: bottlenecks doi.org/10.1016/j.ijhydene.2013.12.103. and limitations, Bioresour. Technol. 180 (2015) 72–78, https://doi.org/10.1016/j. [48] D. Pant, A. Singh, G. van Bogaert, Y.A. Gallego, L. Diels, K. Vanbroekhoven, An biortech.2014.12.096. introduction to the life cycle assessment (LCA) of bioelectrochemical systems (BES) [25] E.S. Heidrich, J. Dolfing, K. Scott, S.R. Edwards, C. Jones, T.P. Curtis, Production of for sustainable energy and product generation: Relevance and key aspects, Renew. hydrogen from domestic wastewater in a pilot-scale microbial electrolysis cell, Sustain. Energy Rev. 15 (2011) 1305–1313, https://doi.org/10.1016/j. Appl. Microbiol. Biotechnol. 97 (2013) 6979–6989, https://doi.org/10.1007/ rser.2010.10.005. s00253-012-4456-7. [49] M. Henze, W. Gujer, T. Mino, M. van Loosedrecht, Activated sludge models ASM1, [26] L. Gil-Carrera, A. Escapa, B. Carracedo, A. Morán, X. Gómez, Performance of a ASM2, ASM2d and ASM3, Water Intell. Online 5 (0) (2006), https://doi.org/ semi-pilot tubular microbial electrolysis cell (MEC) under several hydraulic 10.2166/9781780402369. retention times and applied voltages, Bioresour. Technol. 146 (2013) 63–69, [50] A. Larrosa-Guerrero, K. Scott, I.M. Head, F. Mateo, A. Ginesta, C. Godinez, Effect of https://doi.org/10.1016/j.biortech.2013.07.020. temperature on the performance of microbial fuel cells, Fuel 89 (2010) 3985–3994, [27] J.A. Baeza, À. Martínez-Miró, J. Guerrero, Y. Ruiz, A. Guisasola, https://doi.org/10.1016/j.fuel.2010.06.025. Bioelectrochemical hydrogen production from urban wastewater on a pilot scale, [51] E.J. Bowen, J. Dolfing, R.J. Davenport, F.L. Read, T.P. Curtis, Low-temperature J. Power Sources 356 (2017) 500–509, https://doi.org/10.1016/j. limitation of bioreactor sludge in anaerobic treatment of domestic wastewater, jpowsour.2017.02.087. Water Sci. Technol. 69 (2014) 1004–1013, https://doi.org/10.2166/wst.2013.821. 10 O. Guerrero-Sodric et al. Chemical Engineering Journal 469 (2023) 144001 [52] K. Komemoto, Y.G. Lim, N. Nagao, Y. Onoue, C. Niwa, T. Toda, Effect of [59] B. Boehnke, B. Diering, S.W. Zuckut, AB process removes organics and nutrients: temperature on VFA’s and biogas production in anaerobic solubilization of food treatment combines high food-to-microorganism ratio efficiency with advanced waste, Waste Manag. 29 (2009) 2950–2955, https://doi.org/10.1016/j. activated sludge operations, Water, Environ. Technol. 9 (1997) 23–28. wasman.2009.07.011. [60] B. Wett, K. Buchauer, Energy self-sufficiency as a feasible concept for wastewater [53] L. Deng, H. Yang, G. Liu, D. Zheng, Z. Chen, Y. Liu, X. Pu, L. Song, Z. Wang, Y. Lei, treatment systems Biomethanisation View project FAFODI-Optimierte Kinetics of temperature effects and its significance to the heating strategy for Fettabscheidernutzung zur Gewinnung von Ressourcen und Reduzierung negativer anaerobic digestion of swine wastewater, Appl. Energy 134 (2014) 349–355, Umweltauswirkungen View project, 2007. https://www.researchgate.net/ https://doi.org/10.1016/j.apenergy.2014.08.027. publication/228684984. [54] J.R. Day, E.S. Heidrich, T.S. Wood, A scalable model of fluid flow, substrate [61] J. Jimenez, M. Miller, C. Bott, S. Murthy, H. de Clippeleir, B. Wett, High-rate removal and current production in microbial fuel cells, Chemosphere 291 (2022) activated sludge system for carbon management - Evaluation of crucial process 132686, https://doi.org/10.1016/j.chemosphere.2021.132686. mechanisms and design parameters, Water Res. 87 (2015) 476–482, https://doi. [55] D. Recio-Garrido, M. Perrier, B. Tartakovsky, Modeling, optimization and control org/10.1016/j.watres.2015.07.032. of bioelectrochemical systems, Chem. Eng. J. 289 (2016) 180–190, https://doi. [62] N. Rey-Martínez, A. Barreiro-López, A. Guisasola, J.A. Baeza, Comparing org/10.1016/j.cej.2015.11.112. continuous and batch operation for high-rate treatment of urban wastewater, [56] D.D. Leicester, J.M. Amezaga, A. Moore, E.S. Heidrich, Optimising the Hydraulic Biomass Bioenergy 149 (2021), 106077, https://doi.org/10.1016/j. Retention Time in a Pilot-Scale Microbial Electrolysis Cell to Achieve High biombioe.2021.106077. Volumetric Treatment Rates Using Concentrated Domestic Wastewater, Molecules [63] M.N. Kinyua, M. Elliott, B. Wett, S. Murthy, K. Chandran, C.B. Bott, The role of 25 (2020) 2945, https://doi.org/10.3390/molecules25122945. extracellular polymeric substances on carbon capture in a high rate activated [57] A. Tenca, R.D. Cusick, A. Schievano, R. Oberti, B.E. Logan, Evaluation of low cost sludge A-stage system, Chem. Eng. J. 322 (2017) 428–434, https://doi.org/ cathode materials for treatment of industrial and food processing wastewater using 10.1016/j.cej.2017.04.043. microbial electrolysis cells, Int. J. Hydrogen Energy 38 (2013) 1859–1865, https:// [64] H. Guven, R.K. Dereli, H. Ozgun, M.E. Ersahin, I. Ozturk, Towards sustainable and doi.org/10.1016/j.ijhydene.2012.11.103. energy efficient municipal wastewater treatment by up-concentration of organics, [58] L. Rago, J.A. Baeza, A. Guisasola, Bioelectrochemical hydrogen production with Prog. Energy Combust. Sci. 70 (2019) 145–168, https://doi.org/10.1016/J. cheese whey as sole substrate, J. Chem. Technol. Biotechnol. 92 (2017) 173–179, PECS.2018.10.002. https://doi.org/10.1002/jctb.4987. 11