1 s2.0 S1364032120305670 Main

Renewable and Sustainable Energy Reviews 133 (2020) 110279
Contents lists available at ScienceDirect
Renewable and Sustainable Energy Reviews

journal homepage: http://www.elsevier.com/locate/rser
Is bioelectrochemical energy production from wastewater a reality?

Identifying and standardising the progress made in scaling up microbial
electrolysis cells
Daniel Leicester, Jaime Amezaga, Elizabeth Heidrich *
School of Engineering, Newcastle University, Newcastle-upon-Tyne, NE1 7RU, United Kingdom
A R T I C L E I N F O A B S T R A C T
Keywords: Bioelectrochemical systems (BESs) have the potential to produce energy from wastewater. However, they are far
Bioelectrochemical energy from ready to be applied into industry. The development of large and pilot-scale systems to harness energy and
Microbial electrolysis cells value-added chemicals is widely regarded as one of the greatest research challenges in this field. There are
Wastewater
several reasons for this: i) they are expensive, ii) they are difficult to engineer, iii) the data that can be derived
Hydrogen
Scale-up
from them is often limited, rarely in duplicate, and is disproportionate to the time commitment. Given these
Sustainability restrictions, systematic reviews of large and pilot-scale systems can be helpful in determining the direction of
future research. These reviews need to standardise very different reactor set-ups, operational conditions, and
methods of reporting data. Here we present an analysis of the energy production from semi-pilot and pilot-scale
BESs, and benchmark their performance against existing wastewater treatment. The parameters used include
complexity of wastewater; chemical oxygen demand (COD) loading rate; conductivity; reactor depth; volumetric
treatment rate; effluent quality; energetic treatment balance; and temperature. We find that factors which are
perceived to be problematic, such as low conductivities and temperatures, have been overcome by BESs at pilot-
scale, and that these systems have met the regulatory requirements for discharge standards. We identify reactor
depth and volumetric treatment rate as the areas for future research to focus on. The first of these issues will need
an engineering solution, while the second is likely to come from improved understanding of the complex mi
crobial digestion pathways. Material science may help both. Importantly, these pilot studies have shown that
renewable energy production from wastewater is possible, and with targeted future research, could become a
reality.
are energy intensive, accounting for as much as 3% of electricity con

sumption in developed economies [6]. Furthermore, 80% of the world’s
1. Introduction wastewater goes untreated into receiving waters [7]. This scenario is
neither sustainable nor affordable for the world’s growing population in
Increasing demands for energy and water are the biggest issues to a time of increasing energy costs. However, human waste contains en
threaten both human health and the ecosystems we depend on [1], with ergy locked up in its organic molecules [8]. Harnessing some of this
fossil fuel consumption increasing by 1300 times in the last 200 years energy whilst also treating the waste would transform this part of the
[2]. With a continuing global population rise, the world’s energy re human environment interaction.
quirements are predicted to increase by 28% by 2040 [3], and water When rediscovered in the early 2000s, the technology of Bio
demand is predicted to increase by 30% by 2050 [4]. There is a pressing electrochemical Systems (BESs) seemed to offer this much needed
need to drastically change this situation; the solutions will need to be transformative step, producing easy to use electrical energy directly
widespread and far-reaching. All aspects of the way humans interact from treating wastewater. Research in this field exploded, moving from
with the planet on which we reside will need to be considered, and re 13 papers in 2000 to 11,418 in January 2020 (based on Scopus search
ductions to their energetic costs made. One such area that has the po term on Bioelectrochemical Systems OR Microbial Fuel Cells, which was
tential for radical change is the treatment of domestic wastewater. an early and frequently used term for a type of BES). Yet despite this high
Current methods of wastewater treatment, developed in the 1900s [5],
* Corresponding author.
E-mail address: elizabeth.heidrich@ncl.ac.uk (E. Heidrich).
https://doi.org/10.1016/j.rser.2020.110279
Received 14 February 2020; Received in revised form 12 August 2020; Accepted 16 August 2020
Available online 28 August 2020
1364-0321/Crown Copyright © 2020 Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
D. Leicester et al. Renewable and Sustainable Energy Reviews 133 (2020) 110279
generated in AS enters an AD reactor, where yields are around a tenth of

Abbreviations aerobic yields [30,31]. Therefore, of the initial 100% COD, 60% is lost as
CO2 in the AS process, 1.6% ends up as AD sludge, and 38.4% is avail
AS Activated sludge able for energy production i.e. methane, which must be combusted to
AD Anaerobic digestion yield energy. So although AD is a highly efficient process, it recovers
BES Bioelectrochemical system energy after the energetically expensive AS process has taken place. The
COD Chemical oxygen demand costs of AS are two fold: firstly the input of energy for aeration, and
EAM Electrochemically active microorganisms secondly the loss of approximately 60% of the wastewater’s energy as
HRT Hydraulic retention time CO2. Thus replacement of the AS with an anaerobic process would lead
MEC Microbial electrolysis cell to substantial energy savings and potential energy gains.
MFC Microbial fuel cell Lower energy wastewater treatment options exist and are effective.
OLR Organic loading rate Trickling filters flow wastewater through a porous media, and aerobic
RAS Return activated sludge conditions are created by using a large area and allowing natural
TRL Technology readiness level ambient levels of oxygen in the air to access the bacterial biofilm. They
VTR Volumetric treatment rate are a simple low-cost wastewater treatment technology, with high
removal efficiencies of up to 90% of the influent BOD [32], and low
energy cost of 0.15–0.4 kWh/m3 [33]. However, they require a larger
land surface area than AS, typically 10 times higher due to a much
amount of research, to the authors’ best knowledge, there are yet no longer hydraulic retention time (HRT) (24 h [34]) and smaller depth
commercial or even large-scale operational BES units in existence (0.9–2.5 m deep [35]), and so are often unsuitable for areas with a high
worldwide. As a point of reference, Annamox, a biological process in the density urban population [36,37]. Rittman and McCarty [38] report the
nitrogen cycle, was discovered in 1999; the first full scale plant was in surface area treatment rates in a trickling filter are at around 0.033
operation by 2002; and by 2013 there were at least 30 full-scale plants kgBOD/m2∙day. Comparatively, using a volumetric treatment rate for
[9]. AS of 0.6 kgBOD/m3∙day, and assuming a 3 m deep tank, the equivalent
BESs are a highly complex technology, combining electrochemistry for AS is around 1.8 kgBOD/m2∙day. Clearly there will be some site and
with microbiology and wastewater engineering. The strides forward in waste stream specific issues, but there is a substantial difference in the
understanding these systems, and the microbes that facilitate them, have space requirements. Additionally, poor removal of nitrogen and phos
been significant and broad-reaching. These have been reviewed exten phorous often make AS a more desirable option [39]. Wastewater sta
sively in the literature for different reactor architectures [10–12]; anode bilization ponds are also able to effectively reduce the pollutants and
and cathode materials [13–15]; substrates used such as synthetic and toxins within wastewater. These anaerobic ponds are also low cost, with
real wastewater [16,17]; pre-treatment methods [18]; and the opera removal efficiencies of up to 85% BOD [40,41]. However, due to depth
tional conditions of BESs [11,19]. In addition scale-up, and the prospects of 2.5 m and a retention time of >1 day, land requirements are again
for use with wastewater treatment, have also been covered in several high [41,42].
reviews [17,20–24]. Comparing each of the pilot studies published in Classical AD recovers energy from wastewater by using microor
the literature to each other is important, especially as performance ganisms in the absence of oxygen to digest biodegradable matter,
criteria are often measured and reported differently. However, to un recovering energy in the form of biogas (methane) [43]. Developed in
derstand their readiness for application, we must compare their per 1895 in Exeter [44], this is a relatively simple and low cost method of
formance to the treatment technologies that already exist. This analysis wastewater treatment with energy recovery, and treatment rates can be
addresses this, simplifying and standardising the operational parameters as high as 65%, with loading rates of 1.0–3.0 kgVS/m3⋅day [45].
of each pilot-scale reactor onto a single diagram, in order to visually However, HRTs are between 5 and 40 days. AD is widely used for in
illustrate the performance gap between BESs and the status quo. dustrial wastewater treatment, sludge treatment and for wastewater
In developed countries, activated sludge (AS) is the most commonly treatment in warmer climates such as South America [46]. Unfortu
used process for wastewater treatment by volume treated [25]. It utilises nately, this microbial process is impeded at low temperatures and with
the bacteria present in the wastewater, supplying oxygen via aeration to dilute wastewaters. Research is developing and identifying bacterial
encourage the oxidation of organic compounds into carbon dioxide. The communities which are able to successfully treat raw wastewaters at
large amount of energy available to the bacteria in this aerobic digestion temperatures lower than 13 ◦ C [46–48]. However, currently AD is not
leads to rapid growth, which in turn removes the organic matter at a used for low strength domestic wastewater in the UK and other
high rate. Once the organics have been removed the bacteria die and temperate climates.
sink, producing sludge. This technology has remained largely un Bioelectrochemical systems (BESs) are an alternative anaerobic
changed in the past 100 years, and although highly effective at meeting wastewater treatment technology which may overcome some of these
discharge standards, aeration is energetically expensive, accounting for obstacles. A BES utilises electrochemically active microorganisms
around 50% of the total treatment costs [26]. (EAMs) capable of extracellular electron transfer. These microorganisms
The energetic treatment cost for AS is typically between 2.52 and 7.2 oxidise organic pollutants and can directly produce electrical current via
kJ/gCOD [27]; this energy is spent on aerating and cannot be recovered. electron transfer to a solid electrode. Electrons flow to a counter elec
The energy content of domestic wastewater is estimated to be 16.1 trode or cathode via an external circuit; protons (H+) arrive via an ion
kJ/gCOD [8]. This energy stored in the organic carbon present in the exchange membrane (IEM), and a reduction reaction can occur (see
wastewater is either released as CO2, or becomes part of the bacterial Fig. 1). As a result, wastewater treatment is achieved due to the oxi
cells that form the sludge. Recent developments at wastewater treatment disation of pollutants, and simultaneously, energy is recovered due to
plants have seen an increase in the resource recovery from this sludge the current generation from the flow of electrons [49].
through the use of anaerobic digestion (AD) [28,29]. However this only In relation to energy recovery from the treatment of wastewater
recovers the proportion of energy in the wastewater that is driven into there are two main types of BESs: a microbial fuel cell (MFC) and a
this sludge. The typical yields in AS are about 0.4 g-COD-cells/gCOD microbial electrolysis cell (MEC). In an MFC, the cathode reduction re
substrate [89], though this can be lower where systems are run with low action uses oxygen; water is produced as a waste product, and energy is
loading rates. This means that of the 100% COD entering the AS tank, a harvested in the current that flows in the circuit. An MEC operates under
maximum of 40% becomes new biomass or sludge, and 60% is ‘burnt’ completely anaerobic conditions, with the protons being reduced at the
with oxygen producing CO2 and is therefore lost. The biomass or sludge cathode to produce hydrogen gas (H2) or other chemicals such as caustic
2
Fig. 1. A simplified schematic of a BES (A), and an illustration of how this design can be modified for scale-up to produce electrode cassettes (B) where the cathode
half-cell becomes a central compartment with the anodes either side. Multiple units or electrode cassettes can then be inserted into a tank of liquid.
soda. As the conversion of organic material to hydrogen is an endo energy density of hydrogen (120 MJ/kg) makes it a highly efficient
thermic reaction, energy needs to be supplied to the system for the re energy carrier compared to methane (50 MJ/kg), gasoline (44 MJ/kg)
action to proceed. For acetate to hydrogen this is 0.14 V; however, the and ethanol (26.8 MJ/kg). It is seen as a clean, sustainable and renew
added potential also needs to overcome the system’s over-potentials, able fuel, producing zero carbon emissions [11]. If MECs can produce
which for small systems is > 0.2 V [10,50] and for larger systems can this fuel from wastewater whilst also treating it, this could be a more
be closer to 1 V [51,52]. As additional energy is added by applying this economically viable prospect than producing just electricity in an MFC
potential, achieving net energy neutral treatment in an MEC requires the [23,59]. Therefore, only MECs have been chosen to be reviewed in this
energy recovered in the hydrogen to be higher than this input. paper.
There is continuing debate over which is the best BES configuration If BESs are to achieve their “great potential to become an alternative
to use for energy neutral wastewater treatment. An MFC is technically a to conventional wastewater treatment” [23] they would need to replace
simpler bioreactor. It requires no additional energy input, and it pro the AS process, currently the most prevalent wastewater treatment
duces energy directly in the form of electrical current, which although method (by volume treated). It has been reported that the cost of
limited in amount, is readily usable for lighting, pumping, or UV aeration can range from between 50% and 90% of the total electricity
disinfection. Pilot-scale MFCs have been successfully deployed in field used by a treatment plant [60], of which all would be removed when
sites for the treatment of urine in source separated toilets [53] and in the operating a BES. Pumping costs are estimated to be around 7% [60],
laboratory for combined wastewaters [54], and faeces [55]. The mul although this is very site specific, and many sites are designed so the
tiple MFC cell stack design is of particular importance, and the feasibility wastewater flows downhill. As MECs are still a long way from devel
of this technology has been demonstrated in the UK [53] and Ghana [56] oping a commercially viable system, it is difficult to estimate the
for decentralised treatment of the urine component for wastewater. pumping costs. In a scenario of much smaller stacked units, wastewater
The development of low cost MFC air cathodes has been critical to would need to be pumped in multiple directions, and pumping costs
this success; however, it is also the barrier to larger scale application. Air could be high. If they were retrofitted into the existing AS lanes, then
cathodes require exposure to the air, meaning the space requirements pumping costs for a BES will essentially be the same, though could be
are likely to be large. The stacked MFC design used by Ieropoulos et al. offset to some degree if energy was produced. However in both cases the
[53] at the University of West England involves boxes that are 70 × 30 pumping of the return activated sludge (RAS) will be removed (esti
× 16cm, with a volumetric capacity of 25 L, each containing 36 indi mated as 2% of the total energy costs [60]), further reducing the energy
vidual MFCs. The residence time of these boxes was at least 14 days. This consumed (see Fig. 2). Additionally MECs produce less sludge than the
gives a hydraulic loading capacity of 0.008 m3/m2∙day. The hydraulic AS process [61], so this pumping cost, and in some cases the cost of
loading of trickling filters is 4–10 m3/m2∙day for medium rate systems transporting this sludge to a treatment facility could be reduced.
[57], as these are typically 0.9–2.5 m deep [35]. Though the MFCs treat This paper reviews the published work of pilot-scale MECs, and
urine only, and therefore are not directly comparable with wastewater makes an assessment of how close they are to practical implementation
treatment, this means over 500 units would need to be stacked on top of via the replacement of AS. It is acknowledged for industrial application
each other to reach the space requirements of a trickling filter, which that a thorough cost-benefit analysis and life-cycle assessment would
makes them too large for centralised urban wastewater treatment. also be needed, but this is beyond the scope of this review, and covered
Assuming a 4 cm gap to allow air circulation, this would be a 100 m tall in detail elsewhere [27,37,59,62]. The methodology chosen to review
tower. It seems unlikely that stacked MFCs would work for centralised the different pilot-scale MEC reactors involves a direct comparison with
treatment, even without considering the complexity of wiring, tubing the operational parameters of typical AS plants. The first section of the
and monitoring that this large number of small scale MFCs would entail. review identifies what these parameters are, derived mainly from text
Scaling up MFCs will require larger area cathodes [58]. An air cathode books and legal standards, and the second half compares the pilot-scale
size of 0.62 m2 has been reached, but the water pressure at a depth of MEC studies available in the literature to these parameters. Conclusions
just 0.85 m is problematic [58]. are then drawn as to which parameters we need to focus research effort
An MEC overcomes this limitation as oxygen is not required at the into in order to take this technology from a laboratory curiosity into an
cathode. Instead, a small potential is added to drive different anaerobic industrial reality.
reduction reactions. This is usually the production of hydrogen, though
the production of caustic soda is also a possibility. The gravimetric
3
Fig. 2. A flow diagram for a typical AS wastewater treatment plant, and the imagined positioning of a BES.
2. Operational parameters
Table 2
Summary of the parameters used as the Ideal System and their key references.
Wastewater treatment is a compliance based industry, and its aim is
to discharge ‘safe’ water. Safety is determined by regulatory standards. Parameter Values Units Key references
In the EU, effluent quality needs to match EU (1991) discharge standards Wastewater ≥90 no./L Huang et al. (2010) [77]
[63] (Table 1). Globally there are slight differences in these guidelines. complexity
Substrate mS/cm Henze et al. (2002) [90]
For example, in the USA COD must be < 120 mg/L [64] and in China this ≤1.25
conductivity
is < 60 mg/L [65]. Its infrastructure is typically built on 25–50 year Organic loading rate ≥1.67 kgCOD/m ⋅d3
Logan et al. (2008) [93]
cycles. New technologies would need to robustly demonstrate that they Depth ≥3 m Eckenfelder et al. (1998) [30]
meet regulatory compliance, and would ideally fit into existing infra Volumetric treatment ≥1.25 kgCOD/m3⋅d Logan et al. (2008); EU (1991)
structure to allow for less costly transition. rate [30,58]
COD removal % EU (1991) [63]
To determine the theoretical “ideal” conditions that the MECs would
≥75
Energetic treatment ≤2 kJ/gCOD Pant et al. (2011); Li et al. (2014)
be required to meet, this paper has used AS as a benchmark. As a cost [27,98]
replacement technology MECs should: be able to treat the same types Temperature ≤10 ◦
C Ali (2019) [101]
and concentrations of wastewater (complexity and conductivity); be
able to treat them at realistic temperatures; be of a size similar to
existing infrastructure (reactor size/depth); be able to cope with the
break these down would lead to poor COD removal rates and therefore
volume and strength of wastewater at current treatment plants (organic
non-compliance, as many of the other complex pollutants are not
loading rate); be able to treat the wastewater to the desired discharge
currently regulated for. Thirdly the microbiology of wastewater is highly
standards (effluent quality and volumetric treatment rate); and be able
complex, with 1015–1018 individual bacteria in an AS tank [66].
to do this using less energy (energetic treatment cost). The values for
Assuming the most abundant species represents about 10% of the pop
these parameters are found in Table 2.
ulation (NT/NMax = 10), this gives an estimated diversity of around
104–105 [67]. Some of this complexity will be helpful, providing the
2.1. Wastewater complexity food chain for complex organics; however, some may be detrimental or
competitive [68].
In order to replace or compete with AS, MECs must be able to The complex organics which act as a bacterial food source in do
function with the complex nature of real wastewater. By this we mean mestic wastewater include proteins, fats and carbohydrates [69]. These
firstly that the mixture itself is complex as there are lots of components organic compounds are typically digested in a food chain, with different
in it which will vary with space and time. Secondly, these components in groups of organisms being responsible for different stages of this chain.
the mixture will themselves be complex, for example long chain organics Velasquez-Orta et al. [69] suggests that the pathway to break down
that require multiple stages to break down. It will also contain trace these complex organic compounds within a BES is similar to AD, with
metal, chemical or pharmaceutical pollutants that may be impossible to hydrolysis followed by fermentation. Hydrolysis breaks complex organic
break down, or else they may be toxic to microorganisms. The issue of molecules into simple molecules, and then these are broken down into
long chain complex organics is of particular importance, as failure to volatile fatty acids (VFAs) via fermentation. The final step however, is
not completed by methanogens but by electrogenically active microor
Table 1 ganisms which convert acetate to CO2 and H+, and also transfer an
EU 1991 Wastewater Treatment discharge standards. electron to the electrode. In AD some of these stages are known bottle
Parameters Discharge standards Units necks. Hydrolysis (when bacteria attempt to break down complex
BOD 25 mg/L polymers into simple sugars, VFAs and amino acids) is often the rate
COD 125 mg/L limiting step [70], and subsequently research involving AD often per
TSS 35 mg/L forms pre-treatment techniques in order to boost this step [71–73].
Total Nitrogen 15 mg/L
Therefore, within an MEC, breakdown may be slow, or not even
Total Phosphate 2 mg/L
4
possible, depending on the components of the wastewater. In small scale operating in an unrealistic and advantageous scenario.
MFCs at lab temperatures Velasquez-Orta et al. [69] determined the
combined rate of hydrolysis and fermentation was 0.0024 h− 1, whereas
the rates for both fermentation alone and acetate consumption were an 2.3. Organic loading rate
order of magnitude faster, 0.018 h− 1 and 0.017 h− 1 respectively.
The use of synthetic wastewater cannot replicate or model this The organic loading rate (OLR) is the rate at which the organic
complex reality. The difference in running a system with real waste content of the wastewater is supplied to the system. It is a critical
water, compared to that of synthetic wastewater, is well observed [74]. measurement in the design of a wastewater treatment plant. It takes into
In laboratory scale MFCs, Zhang et al. [75] reported coulombic effi account both the volume and strength of the wastewater that is treated,
ciencies of 90% in acetate fed reactors, compared to 22% in identical and is calculated based on the flow rate and size of reactor (or HRT) and
domestic wastewater fed reactors. At pilot-scale, Baeza et al. [52] the COD concentration of the wastewater (Equation (1)). It enables the
observed higher removal efficiencies but lower hydrogen production size of the tank required to be determined as follows:
when using glucose (36.8% COD removal, 0.028 m3 H2/m3⋅day) and
S
glycerol (26.3% COD removal, 0.015 m3 H2/m3⋅day) as compared to Ks = (Equation 1)
HRT × 1000
domestic wastewater (6% COD removal, 0.031 m3 H2/m3⋅day) using the
same reactor. The reason for this higher production was not determined. where S is COD concentration (mg/L), HRT is hydraulic retention time
In order to make a comparison of different wastewaters and artificial (days), and Ks is OLR (kgCOD/m3⋅day).
substrates, our method requires an actual number to be assigned for Using an average COD of 500 mg/L for medium strength wastewater
complexity. For the purpose of this study we have chosen to use [91], and an average HRT of 7.2 h (0.3 days) [92], an ‘Ideal’ OLR based
complexity as the number of organic compounds present in the sub on AS has been determined at 1.67 kgCOD/m3⋅day.
strate. It is acknowledged that this does not account for complexity in This aligns with the typical loading rate for AS described by Logan
terms of variability over time and space, and may not account for the et al. [93] of 0.2–2 kgCOD/m3⋅day, or slightly above that described by
recalcitrant nature of some individual components or the microbiology. Rittman’s et al. [38] of 0.5–1 kgCOD/m3⋅day.
Detailed knowledge of the composition of wastewater is quite limited The ideal system needs to have a similar or higher OLR to the
[76]. Huang et al. [77] determined with GC/MS analysis that there were treatment process it is replacing to allow it to sit on the same land
at least 90 organic compounds within wastewater. Eriksson et al. [78] footprint. A lower OLR would possibly require the purchase of land and
found 900 compounds in grey water, ranging from trace hydrocarbon building of new infrastructure, which would add a large cost to any
pollutants such as oil and grease, to heavy metals such as manganese or proposed change and may be prohibitive. In many of the papers
zinc. Real wastewaters from an industrial process such as food and drink reviewed, the organic loading rate was not given, and sometimes the
production may contain fewer and more simple compounds. Colin et al. influent COD of the substrate was not given either. In these cases an
[79] found that ethanol, fructose and sucrose represented more than estimate of the influent COD has been made using the average effluent
90% of winery effluent. Synthetic substrates typically contain between 1 and COD removal rates. Calculation of the loading rate was possible
and 10 different compounds [52,80–82]. In this review the use of do when a HRT was reported (see Supplementary Information for detail).
mestic wastewater with an MEC is given the value of 90 using the
number found by Huang et al. [77], which is the maximum, or ideal
value, for other substrates to be set against. For simpler wastewaters the 2.4. Reactor depth
number of typical components are used as determined either in the
paper, or in other literature. For synthetic wastewaters we use the With a retention time of 6–8 h, the size of the AS tank is governed by
number of components likely to contribute as a food source. the population it serves. As these can range from small towns to large
cities, the volume will vary hugely. To limit the need for large scale
2.2. Substrate conductivity modifications to infrastructure, new systems should be compatible with
the AS tanks currently in place. One of the biggest challenges in
Conductivity is the ionic flow through a solution; it is a measure of advancing MECs is the scale–up of designs. Problems include lower
the concentration of charged ions which are free to move in a liquid. power densities, hydrogen production decline, cost increase, and
Wastewater conductivity is relatively low, ranging between 0.7 and 1.8 manufacturing [20,52,83,94]. Currently the biggest reactor operated
mS/cm. The limited number of ions which are free to flow increases was 1000 L, with several others between 20 and 200 L; yet the majority
resistance and electrolyte Ohmic losses, reducing the current density are less than 1 L. Most of the large-scale reactors have been modular in
and lowering the electricity harvested [83]. In artificial wastewaters, design, with multiple electrode pairs within the same tank. At pilot-scale
conductivity is augmented by adding buffers, such as a phosphate buffer, the number of these units has ranged from 3 to 24 [89], yet this number
typically in the range of 7.5–20 mS/cm [10,84]. An increased perfor could be increased to fill the volume of a larger tank. Size of reactor, in
mance of 0.13–0.82 m3 H2/m3⋅day was observed by Verea et al. [85] terms of volume, is therefore not such a useful comparison.
when the conductivity of the synthetic wastewater was doubled from Although AS tanks come in a wide range of sizes they all have
7.5 mS/cm to 15 mS/cm. approximately the same depth, typically between 3 and 6 m [30,95].
Systems using real wastewaters could be dosed with the same buffers The depth controls the aeration efficiency, which can range from 0.5 to
to artificially increase the conductivity and boost performance. How 1.5 kg O2/kWh when using a surface aerator [96]. They are designed to
ever, the cost for a large continuous flow reactor to be routinely dosed be deep enough to maximise oxygen contact with the bacteria, yet not so
with chemicals may be too high to consider for application to industry deep as to increase the head pressure and reduce the efficiency of the
[27]. The ideal system should therefore work with the typical conduc blower [60,97]. The depth of the tank is therefore taken in this case to
tivity of wastewater. The reported values for these in the literature are: represent the value of the ideal size the system needs to attain. Assuming
1.80 mS/cm [86]; 1.25 mS/cm [52]; and 0.8 mS/cm [87,88] for do retrofitting of existing infrastructure, and that each electrode should
mestic wastewater, and 0.7 mS/cm for winery wastewater [89]. Simi reach from the bottom of the tank to the surface, an ideal value of 3 m is
larly, Henze et al. [90] reports for high strength wastewater a value of used. Depth will be critical to many of the issues relating to scaling
1.20 mS/cm. A value of 1.25 has been selected in this study to represent MECs. Manufacturing cassettes with dimensions of 3 m or over will be
the ‘Ideal System’. If a reactor uses a substrate with a lower conductivity difficult, and factors such as resistance, water pressure, turbulence and
than 1.25 mS/cm, it shows that it is successfully coping with greater sludge accumulation will also become more significant at this scale, and
Ohmic losses, while a higher conductivity indicates the reactor is may be difficult to predict based on smaller systems.
5
2.5. Volumetric treatment rate 2.7. Temperature
The volumetric treatment rate is a measure of the ability for the The temperature at which a biological treatment is run is a critical
systems to successfully reduce the COD of the substrate to the required parameter, but will also be highly variable across the world and with
concentrations, taking into account reactor volume, flow rate, substrate seasons. Lower temperatures slow down biological activity (<20 ◦ C
strength, removal rate and effluent standards [38]. It gives a clearer [100]), and in the case of anaerobic digestion have been shown to be a
comparison between reactors that utilise different operating conditions. limiting factor [100]. In both the USA and UK winter temperatures will
A replacement system would not only need to cope with the organic reach below zero; however, there is a latent heat within wastewater
loading, as previously discussed, but also achieve national discharge caused by biological activity. Average wastewater temperatures recor
standards of <125 mg/L or >75% removal [63]. Therefore, taking the ded in the UK are in the range of 10–25 ◦ C [101]. The value of 10 ◦ C is
calculated loading rate of 1.67 kgCOD/m3⋅day, a volumetric treatment used as the ideal as this is the most challenging temperature these sys
rate of 1.25 kgCOD/m3⋅day has been determined using the 75% removal tems might face.
requirement [63]. VTR can also be calculated using Equation (2).
3. Research method
CODr
VTR = (Equation 2)
HRT × 1000
Rose diagrams are used to standardise and visualise a series of
where CODr is COD removed (mg/L), HRT is hydraulic retention time metrics about the performance of a system which may be of differing
(days), and VTR is the volumetric treatment rate (kgCOD/m3⋅day). scales and dimensions. They are commonly used in the computer game
industry, for example to plot the speed, manoeuvrability, acceleration,
and cost of a racing car. They can also be a powerful data visualisation
2.6. Energetic treatment balance method in research [102]. The multiple parameters are presented in a
circular plot where the centre value is the ‘worst’ and the outer circle the
The AS process uses energy to power the air blowers that aerate the ‘best’ (these values could be high or low). The proportion of the petal
tanks. The amount of energy can be calculated per gram of COD filled in gives an indication of how good the unit is in that parameter;
removed, giving the energetic treatment cost. This value will be highly with all petals filled in they give a quick impression of the total per
variable based on the individual equipment and the aeration regime formance across the multiple parameters. In this study, they are used to
used at different wastewater treatment plants. Pant et al. [27] deter show the performance of each of the pilot studies, by indicating how far
mined that the energy consumption for AS lies in the range of 0.7–2 they are away from the operational parameters they need to meet in
kWh/kgCOD removed, which converts to 2.52–7.2 kJ/gCOD removed, order to replace AS (Fig. 3).
and Li et al. [98] reports 1.08–2.1 kJ/gCOD. We chose the value of 2 For each pilot study used, the values for each of these parameters are
kJ/gCOD as an ambitious but realistic target. taken directly from the paper, or calculated or estimated based on the
For each pilot study reported, the energetic treatment balance is information contained within them (Fig. 4). In some cases the values
calculated by first determining the energy costs (Equation (3)), and then have been re-calculated and are different from those directly in the
the energy gains (Equation (4)) from the hydrogen produced using the paper, in order to ensure the method for calculation is comparable across
following calculations. the studies (see Supplementary Information for details). The values for
Energy Cost: each of the studies are then shown as a shaded section of the rose petal.
Where data is entirely absent from the paper and cannot be reasonably
/ (V × A) × 86.4
kJ gCOD = (Equation 3) estimated or calculated, the entire petal is removed. In a perfect situa
gCOD/day
tion all petals of the rose would be fully shaded. Each rose petal is scaled
Energy Recovery: using the parameters in Table 3. In most cases the inside of the petal is
/ LH2
L.day− 1
× Reactor Volume (L) × H2 % × Hm2 3kg × H2kgMJ
kJ gCOD =
gCOD.day− 1
(Equation 4)
Where gas has been recovered, the energy content of this has been
calculated and compared to that of grams of COD removal. Volumetric
densities used for methane and hydrogen are 0.656 kg/m3 and 0.0898
kg/m3 respectively, with energy densities (higher heating values) of
55.6 MJ/kg CH4 and 142 MJ/kg H2 [99]. The energy balance is there
fore the difference between the cost and recovery of kJ to remove 1 g of
COD.
The energetic treatment costs is an important parameter as it takes
into account the OLR, the efficiency of COD removal, power input, and
energy recovered as a usable product. In theory MECs should be energy
positive. Successful reactors will offset this cost by energy recovery from
gas production i.e. they should produce energy per gram of COD
removed. The pilot studies reviewed may not achieve this; however,
they may still have a lower energetic cost than AS.
It is noted within these costs that pumping the wastewater is not
considered. It is likely that this may be similar between AS and an MEC,
if a desired treatment capacity is similar. It is also acknowledged that the
energy recovery from the gas may be an over-estimation; for example if
the gas needed separation, purification or pressurisation, then this Fig. 3. An example rose diagram, the actual ideal values for each component
would add cost. are written within each petal.
6
zero, representing the very worst level of operation. For some of the the metabolic pathways in BES will be vital for their successful appli
parameters (conductivity, temperature and energy balance) a smaller cation. Increasing the efficiency of which waste organics are converted
number is desirable, and in this case the petals are scaled in the opposite into products will increase both treatment and resource recovery.
direction, again with the centre representing the worst level of opera
tion. In the case of the energetic balance, as described in section 2.5, two 4.2. Substrate conductivity
ideal values are plotted. The outer boundary of the petal is the point at
which the MEC is energy neutral, as would be desirable, but the inner The use of real, non-supplemented wastewaters in most of the studies
dashed line is the energetic cost of AS. A shaded petal above this line but means that the conductivity was of a level found in real wastewater,
below the outer limit is an MEC that has better energy costs than AS, but though there is inherent variability in wastewater from different places
is not yet energy neutral. Full details of the scaling are given in Sup and of different concentrations. In the case of the pilot reactor by Hei
plementary Information. drich et al. [86], wastewater conductivity was higher than the ideal
The term ‘pilot-scale’ is used differently among different studies, as is wastewater value at 1.80 mS/cm. This is possibly due to the use of raw
the word ‘scalable’. Pilot-scale, or scalable reactors should be of a design wastewater, whereas other studies used effluent from primary settle
which has the potential to be scaled; to use real wastewater; to be ment [52,87,88,94,107]. Escapa et al. [20] also used primary effluent
operated outside the laboratory; and to be continuously flowing. Wang but this had a conductivity similar to the raw wastewater, showing the
et al. [103] analysed the language used in BES research and found that of inherent differences in wastewater taken from different areas. In the
all the articles with ‘scalable or pilot-scale BES’ in the title, only 13.8% study by Heidrich et al. [86], the energy recovery was high, reaching the
meet the above criteria. San–Martin et al. [61] and Sugnaux et al. [104] energy neutral level, which may suggest this higher conductivity infer
also review work on pilot-scale BES, and discount all pilot studies which red an advantage by reducing resistance loss. However, this was also
operate solely in batch mode. For our review we have selected to use achieved by Cusick et al. [89] who had a conductivity lower than 1.25
only those studies which use continuous flow, are at a scale over 0.01 m mS/cm.
in height and run over a greater length of time than 1 month. This gives Only four of the pilot studies reviewed had a conductivity higher
us twelve pilot studies in total. than that which would be needed in our ideal system. Carmona-Martínez
et al. [81] used saline water. This conductivity was in an order of
4. Results and discussion magnitude higher than those other studies, being in the range of 9 S/m,
but this did not produce better performance in other areas such as en
4.1. Wastewater complexity ergy recovery. Luo et al. [82] used a synthetic medium with 1.70 mS/cm
conductivity, and did have high energy recovery. However, this was
Nine out of the twelve studies have a complexity value as good as lower than the real wastewater used by Heidrich et al. [86] and Escapa
that needed to replace AS. Additionally, all of these studies use real et al. [20]. In this case, the use of acetate may have been of greater
domestic wastewater, therefore the other complexities, (not simply the significance.
number of constituent parts as discussed in Section 2.1) will have been In the study by Cusick et al. [89] there was an addition of a phos
overcome. It is clear that although there may be reduced performance phate buffer, and average conductivity was 1.80 mS/cm. During boiler
with wastewater, which has been well documented in laboratory studies water dilution this was 0.70 mS/cm. Interestingly, it was reported that a
[108], this is not preventing pilot-scale MEC reactors from working. The decrease in conductivity had little to no effect on the reactor perfor
rose diagrams show that the MEC pilots which operate with real mance. Similarly, Tartakovsky et al. [110] noticed only a minor impact
wastewater do not have generally lower performance in the other sec on MEC performance when conductivity was dropped from 15 to 9
tions than those operated with synthetic substrates or simpler mS/cm. However, Verea et al. [85], observed that an increase of 7.5
wastewaters. mS/cm to 15 mS/cm resulted in hydrogen production rates of 0.13–0.82
The pilot study by Cusick et al. [89] was also run on real wastewater m3 H2/m3⋅day respectively in an acetate fed reactor.
from a winery. It is important as industrial wastewaters such as this Application of this technology cannot rely on the artificial supple
might be an easier entry point for commercialisation of MEC technology. mentation of conductivity, as this would be too costly to achieve with
Analysis of the constituent parts of this wastewater was not given, but high volumes of liquid. It is possible that this technology could be tar
Mosse et al. [109] lists an average of 26 different organic compounds in geted only at wastewaters from certain industry, which already have
10 different winery wastewaters, and therefore this value is used. This high conductivity or salinity; however this would be a very small mar
pilot study demonstrates the applicability of MECs for industrial ket. From the studies reported here it is seen that lower conductivities do
wastewaters as well as domestic or municipal wastewaters, as has been not automatically mean lower performance, and that MECs can function
reviewed elsewhere [108]. with the increased resistance caused by the conductivities experienced
Only two of the pilot studies use synthetic wastewater [81,82] and in in real wastewaters. Advances in material science and reactor design to
both cases acetate is used as the sole carbon source. In the case of Car overcome these resistances may be a better solution to the low ionic
mona-Martínez et al. [81] this coincides with good performance in terms conductivity. Supplementing conductivity with a buffer is not likely to
of energy recovery, OLRs and VTRs; however, the use of saline condi be a cost effective or practical solution for large-scale treatment, and nor
tions may have been the cause for this. The performance of the other does it seem necessary.
reactors which were fed synthetic wastewater was similar to studies
using real wastewaters. The use of synthetic substrates is sometimes 4.3. Organic loading rate
justified on the basis that it allows the reactor to function better, and
therefore enables greater understanding and optimisation, which can be The organic loading rate gives the value of the amount of organic
the case at a laboratory scale. However, this review suggests that the use substrate that the reactor receives per day, and is a balance between the
of real wastewaters is not the limiting factor in pilot systems. influent COD and the HRT. In nearly all the pilot reactors, the OLR is
Applications of BES for wastewater treatment will rely on their considerably lower than that of the ideal system. This means that reactor
ability to treat complex mixtures of organic compounds. Even relatively infrastructure, were this to be built to scale, would need to be consid
simple wastewaters which are high in sugar, such as winery waste have erably larger than that currently used for AS. This is likely to be very
different and complex components [79] that require multiple stages of costly, and may not be feasible, especially in urban settings. It therefore
digestion by different organisms. It is clear from the low conversion represents a critical area for improvement, and should be a focus for
efficiencies in the pilot studies examined, that these complex waste or future pilot studies.
ganics are not all transferred to current. Understanding and optimising The ideal OLR is based on a COD of 500 mg/L, which is higher than
7
Fig. 4. Rose diagrams to compare the operational performance of each pilot-scale MEC against the parameters for AS as indicated by the outer edge of the rose petals.
In the case of Energy balance the dotted line represents the level for AS. The outer petal is an energy neutral system, as would be the aim with an MEC.
8
Fig. 4. (continued).
many of the influent CODs used in the pilot studies. This means even in 500 mg/L, the fast HRT of 5 h resulted in a OLR of 1.67 kgCOD/m3⋅day.
those with a HRT approximating that of AS (i.e. 8 h) the OLR was the Although decreasing the residence time of the substrate has been shown
same or lower. The only pilot study that achieved the necessary OLR was to decrease COD removal [106], Cotterill et al. [87] still achieved
Cotterill et al. [87]. Here, even though the influent COD was lower than average effluent to match discharge standards at this speed. Gil-Carrera
9
et al. [106] also had a fast HRT of 4 h, but only achieved an OLR of 0.67
A table of operational performances from each pilot-scale MEC reviewed. These have either been taken directly from the papers, or calculated based on reported values. Justification and calculations of each parameter can
kgCOD/m3⋅day, due to the very low strength wastewater (COD of 112
Temperature
mg/L).
In the pilot studies with a low OLR, the low rate was often due to a
longer HRT than is typical in an AS plant. With the high strength
31.0
22.0
22.0
23.0
20.0
20.0
19.2
25.0
37.0
22.0
(◦ C)
8.5
9.9
wastewater used in the Cusick et al. [89], a reduction from 1 day to an
18-h retention time would see it fit with the ideal OLR. Baeza et al. [52],
Energy balance (kJ/
Heidrich et al. [86] and San-Martin et al. [88] used similar strength
wastewater to the average value, and so would need to reduce the HRTs
down to 8, 5.8 and 5.8 h respectively. However, the rest of the studies
− 12.695
− 0.005
gCOD)
[20,82,88,94,107] used a wastewater or synthetic equivalent with
− 1.87
− 8.95
25.20
0.69
1.24
1.13
2.08
1.80
1.80
0.87
values much lower than average wastewater. These would have to
operate at a retention time of less than 5 h, and in cases such as
Gil-Carrera et al. [106], as low as 1.5 h. This may well impact the
Energy recovered
diffusion of substrate into the biofilm layer and reduce performance, and
it could also cause biofilm detachment. Simply reducing the retention
(kJ/gCOD)
time will most likely reduce the removal rates and worsen effluent
quality (as shown by Gil-Carrera et al. [107]), although the study by
14.34
1.37
0.04
6.60
1.16
1.34
1.34
2.65
9.51 Cotterill et al. [87] shows that the HRT can be faster than in AS. Most
0
0
0
reactors in the study were only run at one single HRT. Therefore, opti
Energetic cost (kJ/
mising the HRT would be a valuable step forward in understanding how

this technology fits with existing infrastructure, and where it might best
be applied.
gCOD)
25.20
1.64
1.37
0.73
7.84
1.13
3.24
3.14
3.14
0.87
0.78
0.55
4.4. Reactor depth
Cusick et al. [89] used the biggest reactor to date [61], with 24
COD removal
modules containing 6 electrode pairs each. The electrodes in this reactor

were 0.7 m high, which is one fifth of the minimum 3 m needed. In this
(%)
study, performance was compared to small-scale laboratory reactors

70
30
64
25
84
76
64
64
15
67
75
61
(2.5 cm high), which were run in the same conditions, and it was found
that the estimated possible current density was 44% less. The authors
VTR (kgCOD/
attributed this to changes to the reactor design, specifically the use of

stainless steel mesh compared to platinum carbon cloth cathodes.
m3⋅d)
0.35
0.14
1.06
0.06
0.27
0.50
0.43
0.43
0.08
0.35
0.96
0.12
However, greater resistance and the relatively slow start up were also
mentioned. If the reactor increased to 3 m in height, it is expected that
further loss in performance would occur. If this was at the same rate seen
between the laboratory and pilot-scales, this would cause failure.
Depth
0.46
0.98
0.53
(m)
n/a
n/a
n/a
n/a
n/a
0.7
0.3
0.8
0.2
Heidrich et al. [86] adopted a different design, and reported the size
of the electrodes as 0.2 m wide by 0.3 m high, 10 times smaller than
would be needed to fit into AS tanks. The electrodes for Cotterill et al.
OLR (kgCOD/
[87] and Baeza et al. [52] were based on the Heidrich et al. [86] design,
but were 0.8 m and 0.46 m high respectively. Cotterill et al. [87]
m3⋅d)
0.50
0.45
1.67
0.25
0.32
0.66
0.67
0.67
0.54
0.52
1.28
0.20
operated 1 m2 anodes, in the dimensions of 1.2 m by 0.8 m high, which

makes them the second tallest electrodes to date. In this study a smaller
scale MEC in-situ was run alongside the larger one so that this scale-up
Conductivity (mS/
can be directly compared. Increasing the size of the electrodes by 16

times did not have a detrimental impact on current density or other
performance metrics, though hydrogen production was reduced. How
0.70
1.80
0.81
0.80
0.61
0.61
1.78
1.70
ever, microbial contamination of the cathode chamber was the likely

cm)
n/a
n/a
n/a
90
cause of this, rather than scale. The study did also highlight the unwieldy
and structurally weak design of the flat plate anodes at this size, sug
gesting that further scaling of this design may not be possible.
Complexity
be found in the Supplementary Information.
The largest electrodes used in a pilot study to date was in the study by
(no.)
San-Martín et al. [88], which used 0.98 m deep x 0.48 m wide anodes.
26
90
90
90
90
90
90
90
90
90
1
Although modular, these had a different design than previous reactors in

which the cathode chamber was equally as large as the anodic, and the
2013/14 [51,
wastewater was recirculated through both chambers to facilitate deni

2013 [105]
2013 [106]
2013 [107]
2011 [89]
2017 [87]
2017 [52]
2018 [88]
2015 [20]
2014 [94]
2015 [81]
2017 [82]
published
trification and the subsequent conversion of nitrite to molecular nitro

Date
gen. These could still be placed within an AS tank, although

86]
modification to the design might present some difficulty in order to keep

each chamber sealed. As these are the largest electrodes, they are the
Gil-Carrera (b)
Gil-Carrera (a)
Gil-Carrera (c)
closest to fitting the ideal scenario. When operated as a MFC, the author
San-Martín
Martinez
Carmona-
Heidrich
comments that power densities were much lower than similar smaller
Cotterill
Escapa
Table 3
Cusick
Brown
Baeza
Paper
lab scale designs (0.06 mW/m2 from 0.47 m2 anodes compared to

Luo
3.6W/m2 from 0.005 m2 anodes [111]).
10
The rest of the reactors were not modular, and so are not compatible a range in VTR of between 0.06 and 1.06 kgCOD/m3⋅day, the worst
with an AS tank. Carmona-Martínez et al. [81] had a 20 cm high anode, performing reactor would need to be over 17 times larger than the best
but at 1 cm thick, the scale-up to 3 m would be structurally difficult. The performing reactor to treat the same amount of wastewater. Further
remaining reactors used anodes and designs too small to be considered more, all the VTRs achieved by the pilot studies are lower than AS, so
able to scale-up to 3 m [82,94,105,106]. more space would be required. This has a significant impact on their
Depth may be a critical factor in the application of BESs. Existing applicability into existing treatment work, and would also add to the
wastewater treatment technologies, even those with a large land foot material costs of the reactors. Understanding and defining the optimal
print such as treatment ponds and trickling filters, operate at depths rate at which substrate can be transferred out of the wastewater flow and
much greater than has been tested with these pilot systems. Stacking into the BES biofilm will be vital in taking this technology forward. With
multiple using on top of each other could be a solution, however, this this value the correct size of reactor can be designed for the given flow
may be difficult to implement, and is likely to add to both building and rate, and the costs accurately estimated.
pumping costs. Pilot-scale demonstration of this technology over large
depths will be vital in understanding if this technology is suitable for 4.6. Effluent quality
applications beyond those which are very small and localised. Research
at this scale is needed, as in addition to the practical issues of building Volumetric treatment rate and effluent quality are highly connected.
structurally sound electrodes to fit this size, there may be important Running the reactors at different retention times and measuring COD
effects on the microbiology, thermodynamics, hydraulics and electro removal would indicate at what point the system ran at peak removal
chemistry caused by the pressure difference at depth. rates. Once this has been found, improving the efficiency of COD
removal at this loading rate would then help the reactor produce an
4.5. Volumetric treatment rate effluent quality closer to the ideal model of <120 mg COD/L, or a 75%
removal rate, while also fitting with the volumetric treatment rate of an
The volumetric treatment rate is a similar parameter to OLR, and is ideal system. As the organic compounds in the wastewater have to be
necessary in determining if the MEC could fit into existing infrastruc funnelled through the anode as current, the anode could be the
ture. It is arguably more important than OLR as it is the actual amount of component that needs optimisation. There have been a number of anode
wastewater that can be treated per unit area. Simply increasing the HRT pre-treatment techniques used in order to make the system more effi
could result in optimal OLR as discussed above, but if the substrate or cient, including: heat treatment, chemical treatment and high temper
wastewater is then moving too quickly through the reactor for the ature ammonia gas treatment [114,115], although these have been used
substrate to diffuse into the biofilm, it will not treat it. Due to the in the laboratory, not at pilot-scale.
modular designs of the electrodes for some reactors, more could be Escapa et al. [20], Baeza et al. [52] and Heidrich et al. [86] had the
added into the tank, reducing the anode working volume but increasing lowest COD removal, with average values of 15%, 25% and 30%
anodic surface area. This would reduce the retention time while respectively. As these reactors operated longer HRTs than desired, but
increasing the loading rate and increasing the ratio of the wastewater still did not remove enough pollutants, a greater anode to volume ratio
which comes into contact with the anodes and the biofilm. Increased to improve the loading rate may also benefit the treatment rate.
anode surface area has been shown to enable greater COD removal, and, Increased surface area of the anode has been reported to increase both
subsequently, increased reactor performance [112]. Studies to optimise current densities and subsequent organic degradation when using real
HRT must be based on the volumetric treatment rate. If a critical HRT wastewater [116].
can be found, which achieves similar removal efficiency (percentage Those reactors which achieved discharge standards [87,88,105] all
COD) but at a faster rate, this will boost the VTR. Recio-Garrido et al. used lower than average strength wastewater. Gil-Carrera et al. [106]
[113] used a combined bioelectrochemical–electrical model to investi achieved discharge standards, but was run using a wastewater already
gate HRT on removal efficiency, removal rate and power production. It below 125 mg/L COD. Cotterill et al. [87] suggested that the low
was found that a retention time of 10.5 h was the optimum for COD influent COD was detrimental to the start-up and hydrogen production
removal rates, and any higher resulted in the same effluent quality but of the reactor. Lower strength wastewater has been reported to limit
lower treatment rates. current densities and treatment rates [75]. Targeting an expensive
The highest VTR is in Cotterill et al.’s [87] reactor. This reactor was technology such as an MEC at a wastewater stream that needs minimal
run on relatively low strength wastewater, but the authors did note that treatment is unlikely to be viable. However, importantly, these studies
COD removal efficiency was higher when the strength of influent COD do show that MECs can continue to work down to low COD levels, and
was higher. It is possible therefore that if this reactor was run with that it would be possible with the correct volume of treatment space to
higher strength wastewater, the target VTR may have been reached. The reduce COD from a loading of 500 mg/l to below 125 mg/l. Reactors
treatment rate reported by Carmona-Martínez et al. [81], when the which achieved relatively high removal efficiencies, but not high
reactor was fed 0.64 g/L of acetate, is also high (0.96 kgCOD/m3⋅day). enough removal rates to achieve discharge standards, could be
This reactor was also subject to increased loading rates up 6.4 g/L of improved by optimising HRT, anode surface area and reactor size.
acetate. Although these values are not used in the rose diagrams, as they To be successful for wastewater treatment applications, BES must not
are unrealistic in comparison to the concentration of acetate found in only demonstrate that they are able to meet discharge standards, but
wastewater, it shows that a VTR of 7.04 kgCOD/m3⋅day could be ach that they can also do this reliably and consistently. Variability in the
ieved in this reactor. levels of performance in these systems to date [51,87] could be a sig
All the other reactors had very low VTR, and the worst performing nificant problem. Initial applications of these technologies may be better
reactors in terms of pollutant removal were Baeza et al. [52] and Escapa suited to places within the treatment works where full treatment in not
et al. [20] with removal rates of 0.06 and 0.08 kgCOD/m3⋅day respec required, i.e. where they would be a pre-treatment rather than a fin
tively. Baeza et al. [52] altered the HRT following the unsuccessful ishing step. They might also be better placed where the organic load is
pollutant removal. It was found that a 10-day HRT was needed. With an high, and therefore the costs of building these reactors can be mitigated
OLR of 0.05 kgCOD/m3⋅day in order to reach 72% removal, this would by the reduction in energy cost of high COD removal via aerobic
result in a VTR of 0.04 kgCOD/m3⋅day, 34.7 times lower than needed. digestion.
Volumetric treatment rate is arguably the most important parameter Currently in the UK and Europe, effluent quality is primarily focused
for understanding the applicability of this technology in the wastewater on removal of the organic load (through BOD, COD and total suspended
treatment industry. It encompasses the removal rates of the organic solids), and levels of Nitrogen and Phosphate. It is possible in the future
pollutants, flow rates and reactor size. The pilot studies examined show that these regulations will be extended to other pollutants such as metals
11
and pharmaceuticals [6]. Investigating the ability of BES to remove overpotentials on the anode would mean less energy input [12]. This
these other pollutants and harvest them as resource could offer a sig should be a goal of design modifications. Future research should also
nificant advantage of this technology over those currently available. target the efficient bioproduction of different products at the cathode
which may bring greater cost benefits. Defining a way in which to
4.7. Energetic treatment balance incorporate this into the energy balance, for example by using the en
ergy offset by typical industrial production, will also be important.
Most of the reactors had an energy treatment cost that is lower than
AS, and some reached an energy neutral and even positive state. The 4.8. Temperature
most cost efficient reactor to date in terms of energy recovery was the
reactor fed winery wastewater by Cussick et al. (2011). This had an The temperatures at which the reactors are run is critical if they are
energetic treatment cost of 1.64 kJ/gCOD removed, less than typical AS going to operate under ambient conditions in temperate climates, as
costs of ≤2.52 kJ/gCOD removed, and the energy recovered was 14.34 heating raw wastewater on an industrial scale will never be viable.
kJ/gCOD removed. This gives a net energy positive energetic treatment Unfortunately the majority of pilot studies reviewed were either set up
balance of +12 kJ/gCOD removed, although this seems very high given in a laboratory, and were artificially heated, or both, operating at over
the energy value of 16.1 kJ/gCOD for wastewater, or indeed the value 22 ◦ C. Only 2 have shown that MEC wastewater treatment can operate
for glucose of 14.6 kJ/gCOD [8]. The methane content produced was all year round outside of the laboratory [51,87]. Heidrich et al. [86]
much higher than the calculated stoichiometric conversion of current to reports minimum and maximum wastewater temperatures of 8.5–27.0
methane, and it is suggested that current was produced from COD ◦
C for the influent, highlighting the large temperature range MECs will
removal, while anaerobic digestion occurred separately. This would have to cope with if placed on a wastewater treatment site. Cotterill et al.
most likely be due to the installation of the thermostat to heat the [87] also operated a reactor at ambient temperatures, with the average
wastewater to 31 ◦ C. The energetic cost of heating was not considered in temperature during start-up being 9.9 ◦ C. It was discussed that this low
the balance. However, heating a litre of water by 10 ◦ C requires 41.9 kJ, temperature could have been the cause of the longer than expected
equivalent to the energy it would contain with a COD of 2600 mg/L. start-up. Despite laboratory work showing a correlation between per
Carmona-Martínez et al. [81] and Luo et al. [82] also achieved a positive formance and temperature [117], seasonal variations in both of these
energy balance with 1.87 and 8.954 kJ/gCOD removed respectively. studies did not map significantly onto changes in performance. This is
However, in both these cases synthetic wastewater was used, with the likely to be due to other factors masking the relationship, as all bio
substrate being acetate. logical activity is affected by temperatures. Fully understanding this
In order to compare Heidrich et al. [86] with other papers, it was relationship will be important in predicting reactor behaviour and per
necessary to recalculate the energy balance by only using one face of the formance, and it is possible they may have to run at an energetic and
anode. This recalculation shows this reactor was also very marginally financial loss in the winter and make this up in the summer.
energy positive, gaining 0.005 kJ/gCOD removed. Cotterill et al. [87] Low temperatures are described as the Achilles heel in advancing
used a similar but larger reactor design and had similar energy costs to anaerobic digestion technology [100], and so a major advantage of
power the systems. However, due to the poor hydrogen capture this MECs over AD is their ability to operate at low temperatures. Gaining a
reactor was not energy positive, and required a total net energy input of full understanding of this disparity, despite the similar food chains
0.69 kJ/gCOD removed. It should be noted that this is still less than AS. involved could be vital in advancing not only BESs, but also low tem
Gil-Carrera et al. [105], Gil-Carrera et al. [106] and Brown et al. [94] perature AD. Initially application of this technology is likely to be in
also all achieved energy balances less than AS, with 2.08, 1.8 and 0.874 places, or on waste streams, where there is a high energy load i.e. a high
kJ/gCOD respectively. In the studies by both Gil-Carrera et al. [105] and concentration of organic matter, but where anaerobic digestion is not
Gil-Carrera et al. [106] the energy cost was higher than AS, but suc possible. This may be on small-scale operations where heating is not
cessful hydrogen recovery counterbalanced this. Baeza et al. [52] had viable, or on the treating AD centrate which is too liquid to be fed into
one of the highest energy costs, 7.84 kJ/gCOD, due to the 1.5 V potential the AD process. To advance the fundamental science of these systems,
difference used. However, due to the high hydrogen production, and low and for their practical application, testing, validation and experimen
COD removed, the energy balance was still better than the AS compar tation should be done in environmentally relevant conditions. In order
ison, with a value of 1.24 kJ/gCOD removed. The pilot project by to develop a full understanding of the limits of BESs, research at the
San-Martín et al. [88] also performed less well with respect to this pilot-scale (which aims to advance the technology toward application)
parameter. The gas produced comprised of a mix of methane, CO2 and should be operated at representative temperatures for the UK, Europe
hydrogen but it was in such small quantities that virtually no energy was and North America.
recovered from the system. In the study by Escapa et al. [20], the
calculated treatment cost of 25.2 kJ/gCOD is the largest of the discussed 5. Practical and policy implications of this study
reactors (see Supplementary Information for calculations). Part of the
problem of addressing the energy balance in MECs is the consistency Currently, MEC technology sits around the technology readiness
with which terms are calculated and reported [103]. However, it has level (TRL) 5: there has been validation of the technology in a relevant
been shown that even in these first pilot studies, where build quality and environment [118]. Future research should therefore seek to optimise
conditions may not be optimal, energy neutrality can be achieved. the performance, rather than just demonstrate it. Based on the analysis
Actual application of BESs may however require a greater incentive of the different pilot studies, this paper has highlighted two key areas
than energy neutrality. Even with advancements in finding low cost required for BESs to become a competitive wastewater treatment tech
materials, these reactors will be far more costly to build than the existing nology. Firstly, achieving the optimal volumetric treatment rates will be
technologies [62]. The ability for BES to harness reducing equivalents vital in understanding the applicability of this technology. Previous
from wastewater, and then convert these into products is the technol literature shows that increasing the organic loading rate can boost this
ogy’s greatest asset. By manipulation of the cathode reaction it is volumetric treatment [119]. This can be achieved either by increasing
possible to produce different products that may have higher worth than the strength of the wastewater or by increasing its flow rate [120]. Each
the energy value alone [98]. This means that they have the potential to of these modifications will increase the rate at which the organics within
produce different things at different times, responding to industrial the wastewater are supplied into the reactor, which in turn should boost
need, market prices and incentive schemes. The energy balance remains the rate of organic removal. Only by fully understanding and optimising
an important performance parameter, and future research should seek to this mass transfer can the correct size of reactor be designed for the given
improve upon this, in particular by reducing the energy costs: lower flow rate and the costs accurately estimated.
12
Secondly, size also remains a significant problem. The largest MEC to of the metabolic pathways of waste organic digestion. Critically though,
date was 1000 L, with a hydraulic retention time of 1 day [89]. This is far in four out of the twelve pilot studies examined, the MECs were energy
from the size required at wastewater treatment operations. Many BESs neutral or even positive, demonstrating that energy recovery from
have therefore been designed with modular electrodes, multiple units wastewater treatment using this technology is possible.
that can be placed in any existing tank. However, these existing tanks are
deep, and this depth is necessary to achieve treatment on the small land
footprint in urban areas. Designing an electrode that can span to the Declaration of competing interest
depth of 3 m will need to overcome the effect that hydrostatic pressure
has on both the biofilm formation, the performance and the structural The authors declare that they have no known competing financial
integrity. It will also need to cope with the changes in the thermody interests or personal relationships that could have appeared to influence
namic and kinetic properties of the biological and electrochemical the work reported in this paper.
processes. Further research investigating these new reactor designs that
retain the same land foot print of existing assets is needed. Acknowledgements
A further issue demonstrated in this research, and which has been
highlighted previously [23,121], is the standardisation of the design, This work has been financially supported by the Engineering and
methods and the reporting of BESs. The data we present in this study Physical Sciences Research Council (EPSRC) EP/N50928/1 and the
seeks to compare the performance of each pilot-scale study and the Centre for Doctoral Training in Engineering for the Water Sector entitled
parameters used in the wastewater treatment industry. However, com Skills Technology Research and Management (STREAM) EP/L015412/
parable information was often difficult to find within the research pa 1. The authors would also like to thank Amy Green and Laura McGin
pers, and in some cases was absent or had to be calculated from other tyfor their help with the rose diagrams, Sophie Leicester for her excellent
values given. If policy makers and industry promote this technology, proof reading, and finally Northumbrian Water Ltd for their continuing
relevant data must be presented in a clear and systematic way. The support on this project.
commercial implementation of these systems will depend not only on
their ability to meet with current and future wastewater treatment Appendix A. Supplementary data
regulations, but will also need to consider policy decisions, energy tar
gets, and carbon trade schemes [122]. If BES technologies are to reach Supplementary data to this article can be found online at https://doi.
their full potential, researchers must not only provide the data to show org/10.1016/j.rser.2020.110279.
this is possible, but also present it so that it can be accessed by those
which will promote or implement these. References
The barriers to bring this technology further from TRL 5 are not only
technical but also related to policy incentives for innovation. The strict [1] Zimmerman JB, Mihelcic JR, Smith J. Global stressors on water quality and
regulatory constraints on performance and the economics of water quantity. Environ Sci Technol 2008;42:4247–54. https://doi.org/10.1021/
es0871457.
utilities do not favour investments in technologies that cannot fulfil both [2] Ritchie H, Roser M. Fossil fuels. OurworldindataOrg; 2018.
criteria straight away. However, the imperative of defining zero-carbon [3] EIA. International energy outlook 2016 whith projections to 2040. 2016. https://
pathways should benefit technologies able to reach this goal. BES doi.org/DOE/EIA-0484(2014).
[4] Boretti A, Rosa L. Reassessing the projections of the world water development
technologies fit in this category. Accordingly, there is strong argument report. Npj Clean Water 2019;2. https://doi.org/10.1038/s41545-019-0039-9.
for R&D policies able to invest in deblocking the technical constraints of [5] Wiesmann U, Choi IS, Dombrowski E-M. Historical development of wastewater
BES technologies, and for utility regulatory frameworks that encourage collection and treatment. Fundam. Biol. Wastewater treat. Berlin, Germany:
Wiley-VCH Verlag GmbH & Co. KGaA; 2006. p. 1–23. https://doi.org/10.1002/
explorative applications in real settings, in order to bring them to
9783527609604.ch1.
practice in the future. [6] Li W-W, Yu H-Q, Rittmann BE. Chemistry: reuse water pollutants. Nature 2015;
528:29–31. https://doi.org/10.1038/528029a.
[7] UN AB. Wastewater management. 2015. Geneva, Switz.
6. Conclusion [8] Dai Z, Heidrich ES, Dolfing J, Jarvis AP. Determination of the relationship
between the energy content of municipal wastewater and its chemical oxygen
The journey towards zero carbon will require far-reaching solutions demand. Environ Sci Technol Lett 2019;6:396–400. https://doi.org/10.1021/acs.
estlett.9b00253.
across all aspects of the human/environment interaction. Wastewater
[9] Ni SQ, Zhang J. Anaerobic ammonium oxidation: from laboratory to full-scale
treatment would seem like an easy win in this regard, as there is more application. BioMed Res Int 2013. https://doi.org/10.1155/2013/469360. 2013.
energy contained within the wastewater than is currently being used to [10] Call D, Logan BE. Hydrogen production in a single chamber microbial electrolysis
treat it. However, the locked-in infrastructure and strict environmental cell lacking a membrane. Environ Sci Technol 2008;42:3401–6. https://doi.org/
10.1021/es8001822.
regulation to protect receiving waters means change is difficult to [11] Kadier A, Simayi Y, Abdeshahian P. A comprehensive review of microbial
implement and taking risks are avoided. The benefits of MECs are clear, electrolysis cells ( MEC ) reactor designs and configurations for sustainable
as they are simultaneously treating wastewater while recovering the hydrogen gas production. Elsevier; 2015. p. 427–43. https://doi.org/10.1016/j.
aej.2015.10.008.
energy harnessed from the organics as valuable products. Their ability to [12] Slate AJ, Whitehead KA, Brownson DAC, Banks CE. Microbial fuel cells: an
do this with dilute wastewaters and at low temperatures sets them apart overview of current technology. Renew Sustain Energy Rev 2019;101:60–81.
from classical anaerobic digestion. However, the technology is still far https://doi.org/10.1016/j.rser.2018.09.044.
[13] Kundu A, Sahu JN, Redzwan G, Hashim MA. An overview of cathode material and
from being commercialised. By standardising the data across all of these catalysts suitable for generating hydrogen in microbial electrolysis cell. Int J
studies and benchmarking them against industry standards, we are able Hydrogen Energy 2013;38:1745–57. https://doi.org/10.1016/j.
to identify the operational parameters that MECs are repeatedly able to ijhydene.2012.11.031.
[14] Hindatu Y, Annuar MSM, Gumel AM. Mini-review: anode modification for
attain, and importantly clearly identify those where further research is improved performance of microbial fuel cell. Renew Sustain Energy Rev 2017;73:
needed. MECs have been shown to cope well with the conditions that 236–48. https://doi.org/10.1016/j.rser.2017.01.138.
might theoretically stop their performance, such as low temperatures, [15] Kadier A, Kalil MS, Abdeshahian P, Chandrasekhar K, Mohamed A, Azman NF,
et al. Recent advances and emerging challenges in microbial electrolysis cells
low conductivities and complex real wastewaters. None of these factors
(MECs) for microbial production of hydrogen and value-added chemicals. Renew
appear to be the Achilles heel of MEC operation, and therefore supple Sustain Energy Rev 2016;61:501–25. https://doi.org/10.1016/j.
menting or amending them is not necessary. There is however, a sig rser.2016.04.017.
nificant performance gap with MECs and AS in terms of the volumetric [16] Pandey P, Shinde VN, Deopurkar RL, Kale SP, Patil SA, Pant D. Recent advances
in the use of different substrates in microbial fuel cells toward wastewater
treatment rate and the reactor depth. Solving these issues will require treatment and simultaneous energy recovery. Appl Energy 2016;168:706–23.
both improved reactor design and increased fundamental understanding https://doi.org/10.1016/j.apenergy.2016.01.056.
13
[17] He L, Du P, Chen Y, Lu H, Cheng X, Chang B, et al. Advances in microbial fuel [48] Álvarez JA, Armstrong E, Gómez M, Soto M. Anaerobic treatment of low-strength
cells for wastewater treatment. Renew Sustain Energy Rev 2017;71:388–403. municipal wastewater by a two-stage pilot plant under psychrophilic conditions.
https://doi.org/10.1016/j.rser.2016.12.069. Bioresour Technol 2008;99:7051–62. https://doi.org/10.1016/j.
[18] Wei J, Liang P, Huang X. Recent progress in electrodes for microbial fuel cells. biortech.2008.01.013.
Bioresour Technol 2011;102:9335–44. https://doi.org/10.1016/j. [49] Pham TH, Aelterman P, Verstraete W. Bioanode performance in
biortech.2011.07.019. bioelectrochemical systems: recent improvements and prospects. Trends
[19] Logan BE, Hamelers B, Rozendal R, Schröder U, Keller J, Freguia S, et al. Biotechnol 2009;27:168–78. https://doi.org/10.1016/J.TIBTECH.2008.11.005.
Microbial fuel Cells: methodology and technology. Environ Sci Technol 2006;40: [50] Logan BE, Call D, Cheng S, Hamelers HVM, Sleutels THJA, Jeremiasse AW, et al.
5181–92. https://doi.org/10.1021/es0605016. Microbial electrolysis cells for high yield hydrogen gas production from organic
[20] Escapa A, San-Martín MI, Mateos R, Morán A. Scaling-up of membraneless matter. Environ Sci Technol 2008;42:8630–40. https://doi.org/10.1021/
microbial electrolysis cells (MECs) for domestic wastewater treatment: es801553z.
bottlenecks and limitations. Bioresour Technol 2015;180:72–8. https://doi.org/ [51] Heidrich ES, Edwards SR, Dolfing J, Cotterill SE, Curtis TP. Performance of a pilot
10.1016/J.BIORTECH.2014.12.096. scale microbial electrolysis cell fed on domestic wastewater at ambient
[21] Wang H, Ren ZJ. A comprehensive review of microbial electrochemical systems temperatures for a 12 month period. Bioresour Technol 2014;173:87–95. https://
as a platform technology. Biotechnol Adv 2013;31:1796–807. https://doi.org/ doi.org/10.1016/j.biortech.2014.09.083.
10.1016/j.biotechadv.2013.10.001. [52] Baeza JA, Martínez-Miró À, Guerrero J, Ruiz Y, Guisasola A. Bioelectrochemical
[22] Rousseau R, Etcheverry L, Roubaud E, Basséguy R, Délia ML, Bergel A. Microbial hydrogen production from urban wastewater on a pilot scale. J Power Sources
electrolysis cell (MEC): strengths, weaknesses and research needs from 2017;356:500–9. https://doi.org/10.1016/J.JPOWSOUR.2017.02.087.
electrochemical engineering standpoint. Appl Energy 2020;257. https://doi.org/ [53] Ieropoulos IA, Stinchcombe A, Gajda I, Forbes S, Merino-Jimenez I, Pasternak G,
10.1016/j.apenergy.2019.113938. et al. Pee power urinal-microbial fuel cell technology field trials in the context of
[23] Escapa A, Mateos R, Martínez EJ, Blanes J. Microbial electrolysis cells: an sanitation. Environ Sci Water Res Technol 2016;2:336–43. https://doi.org/
emerging technology for wastewater treatment and energy recovery. from 10.1039/c5ew00270b.
laboratory to pilot plant and beyond. Renew Sustain Energy Rev 2016;55:942–56. [54] Ghadge AN, Jadhav DA, Ghangrekar MM. Wastewater treatment in pilot-scale
https://doi.org/10.1016/j.rser.2015.11.029. microbial fuel cell using multielectrode assembly with ceramic separator suitable
[24] Zhang Y, Liu M, Zhou M, Yang H, Liang L, Gu T. Microbial fuel cell hybrid systems for field applications. Environ Prog Sustain Energy 2016;35:1809–17. https://doi.
for wastewater treatment and bioenergy production: synergistic effects, org/10.1002/ep.12403.
mechanisms and challenges. Renew Sustain Energy Rev 2019;103:13–29. https:// [55] Kretzschmar J, Riedl S, Brown RK, Schröder U, Harnisch F. eLatrine: lessons
doi.org/10.1016/j.rser.2018.12.027. learned from the development of a low-tech MFC based on cardboard electrodes
[25] Scholz M. Activated sludge processes. Wetl. Water pollut. Control. Elsevier; 2015. for the treatment of human feces. J Electrochem Soc 2017;164:H3065–72.
p. 91–105. https://doi.org/10.1016/b978-0-444-63607-2.00015-0. https://doi.org/10.1149/2.0121703jes.
[26] Shi CY. Mass flow and energy efficiency of municipal wastewater treatment [56] Castro CJ, Goodwill JE, Rogers B, Henderson M, Butler CS. Deployment of the
plants. Water Intell Online 2015;10. https://doi.org/10.2166/9781780400907. microbial fuel cell latrine in Ghana for decentralized sanitation. J Water, Sanit
9781780400907. Hyg Dev 2014;4:663–71. https://doi.org/10.2166/washdev.2014.020.
[27] Pant D, Singh A, Van Bogaert G, Gallego YA, Diels L, Vanbroekhoven K. An [57] Peavy H, Rowe D, Tchobanoglous G. Water and wastewater pumping. Environ
introduction to the life cycle assessment (LCA) of bioelectrochemical systems Eng 1986.
(BES) for sustainable energy and product generation: relevance and key aspects. [58] Rossi R, Jones D, Myung J, Zikmund E, Yang W, Gallego YA, et al. Evaluating a
Renew Sustain Energy Rev 2011;15:1305–13. https://doi.org/10.1016/J. multi-panel air cathode through electrochemical and biotic tests. Water Res 2019;
RSER.2010.10.005. 148:51–9. https://doi.org/10.1016/j.watres.2018.10.022.
[28] EPA. Urban waste water treatment in 2017. 2017. [59] Foley JM, Rozendal RA, Hertle CK, Lant PA, Rabaey K. Life cycle assessment of
[29] Bodík I, Sedláček S, Kubaská M, Hutňan M. Biogas production in municipal high-rate anaerobic treatment, microbial fuel cells, and microbial electrolysis
wastewater treatment plants - current status in EU with a focus on the Slovak cells. Environ Sci Technol 2010;44:3629–37. https://doi.org/10.1021/
Republic. Chem Biochem Eng Q 2011;25:335–40. es100125h.
[30] Eckenfelder W. Activated sludge: process design and control. Lancaster, [60] Drewnowski J, Remiszewska-Skwarek A, Duda S, Łagód G. Aeration process in
Pennsylvania, USA: Technomic Publishing Company, Inc; 1998. bioreactors as the main energy consumer in a wastewater treatment plant. Review
[31] Dinsdale R, Bryne K, Tucker D. The anaerobic digestion of textile desizing of solutions and methods of process optimization. Processes 2019;7. https://doi.
wastewater. Ecotext. W. Forw. Sustain. Dev. Text. Elsevier Ltd.; 2007. p. 163–7. org/10.3390/pr7050311.
https://doi.org/10.1533/9781845693039.4.163. [61] San-Martín MI, Leicester DD, Heidrich ES, Alonso RM, Mateos R, Escapa A.
[32] Wastewater EPA. Technology fact sheet trickling filters. 2000. Bioelectrochemical systems for energy valorization of waste streams. Energy syst.
[33] Henrich C-D, Marggraff M. Energy-efficient wastewater reuse – the renaissance of Environ. InTech; 2018. https://doi.org/10.5772/intechopen.74039.
trickling filter technology. In: Proc 9th int conf water reuse; 2013. p. 27–31. [62] Aiken DC, Curtis TP, Heidrich ES. Avenues to the financial viability of microbial
[34] USEPA. Wastewater technology fact sheet trickling filter nitrification. 2000. electrolysis cells [MEC] for domestic wastewater treatment and hydrogen
[35] Articles E. Trickling filter - classification and mechanism - engineering articles. production. Int J Hydrogen Energy 2019;44:2426–34. https://doi.org/10.1016/j.
Environ Eng 2015. http://www.engineeringarticles.org/trickling-filter-classifi ijhydene.2018.12.029.
cation-and-mechanism/. [Accessed 30 July 2020]. [63] EU. Council. Directive of 21 May 1991 concerning urban waste water treatment
[36] Deng Y. Improvements to the performance of trickling filters by inclusion of (91/271/EEC). Council of European Communities (CEC). Off J Eur Communities
alternative surface active media. 2018. 1991;135:40–52.
[37] Velasquez-Orta SB, Heidrich O, Black K, Graham D. Retrofitting options for [64] EPA. Parameters of Water Quality 2001;74:177–86.
wastewater networks to achieve climate change reduction targets. Appl Energy [65] Zhou Y, Duan N, Wu X, Fang H. COD discharge limits for urban wastewater
2018;218:430–41. https://doi.org/10.1016/j.apenergy.2018.02.168. treatment plants in China based on statistical methods, vol. 10. Switzerland:
[38] Rittmann BEMP. Environmental biotechnology: principles and applications. Water; 2018. https://doi.org/10.3390/w10060777.
2001. [66] Curtis TP, Head IM, Lunn M, Woodcock S, Schloss PD, Sloan WT. What is the
[39] Tchobanoglous G. Wastewater Engineering : treatment and resource recovery -, extent of prokaryotic diversity? Philos. Trans. R. Soc. B Biol. Sci. 2006;361:
vol. 2; 2014. 2023–37. https://doi.org/10.1098/rstb.2006.1921. Royal Society.
[40] Liu L, Hall G, Champagne P. Disinfection processes and mechanisms in [67] Curtis TP, Sloan WT, Scannell JW. Estimating prokaryotic diversity and its limits.
wastewater stabilization ponds: a review. Environ Rev 2018;26:417–29. https:// Proc Natl Acad Sci U S A 2002;99:10494–9. https://doi.org/10.1073/
doi.org/10.1139/er-2018-0006. pnas.142680199.
[41] Alexiou GE, Mara DD. Anaerobic waste stabilization ponds: a low-cost [68] Ferrera I, Sánchez O. Insights into microbial diversity in wastewater treatment
contribution to a sustainable wastewater reuse cycle. Appl Biochem Biotechnol systems: how far have we come? Biotechnol Adv 2016;34:790–802. https://doi.
Part A Enzyme Eng Biotechnol 2003;109:241–52. https://doi.org/10.1385/ org/10.1016/j.biotechadv.2016.04.003.
ABAB:109:1-3:241. [69] Velasquez-Orta SB, Yu E, Katuri KP, Head IM, Curtis TP, Scott K. Evaluation of
[42] Mara D. Low-cost urban sanitation. 1996. hydrolysis and fermentation rates in microbial fuel cells. Appl Microbiol
[43] Ward AJ, Hobbs PJ, Holliman PJ, Jones DL. Optimisation of the anaerobic Biotechnol 2011;90:789–98. https://doi.org/10.1007/s00253-011-3126-5.
digestion of agricultural resources. Bioresour Technol 2008;99:7928–40. [70] Adekunle KF, Okolie JA. A review of biochemical process of anaerobic digestion.
[44] Marsh G. Rise of the anaerobic digester. Renew Energy Focus 2008;9:28–34. Adv Biosci Biotechnol 2015;6:205–12. https://doi.org/10.4236/abb.2015.63020.
https://doi.org/10.1016/s1755-0084(08)70063-2. [71] Fernandes TV, Klaasse Bos GJ, Zeeman G, Sanders JPM, van Lier JB. Effects of
[45] Tanaka R, Kowase S, Unno M. Chiral cyclotrisiloxanes. Dalton Trans 2010;39: thermo-chemical pre-treatment on anaerobic biodegradability and hydrolysis of
9235–7. https://doi.org/10.1039/c0dt00135j. lignocellulosic biomass. Bioresour Technol 2009;100:2575–9. https://doi.org/
[46] Petropoulos E, Dolfing J, Davenport RJ, Bowen EJ, Curtis TP. Developing cold- 10.1016/j.biortech.2008.12.012.
adapted biomass for the anaerobic treatment of domestic wastewater at low [72] Valo A, Carrère H, Delgenès JP. Thermal, chemical and thermo-chemical pre-
temperatures (4, 8 and 15 ◦ C) with inocula from cold environments. Water Res treatment of waste activated sludge for anaerobic digestion. J Chem Technol
2017;112:100–9. https://doi.org/10.1016/j.watres.2016.12.009. Biotechnol 2004;79:1197–203. https://doi.org/10.1002/jctb.1106.
[47] Martin Garcia I, Mokosch M, Soares A, Pidou M, Jefferson B. Impact on reactor [73] Zhang L, Gao R, Naka A, Hendrickx TLG, Rijnaarts HHM, Zeeman G. Hydrolysis
configuration on the performance of anaerobic MBRs: treatment of settled sewage rate constants at 10–25 ◦ C can be more than doubled by a short anaerobic pre-
in temperate climates. Water Res 2013;47:4853–60. https://doi.org/10.1016/j. hydrolysis at 35 ◦ C. Water Res 2016;104:283–91. https://doi.org/10.1016/j.
watres.2013.05.008. watres.2016.07.038.
14
[74] Wagner RC, Regan JM, Oh S-E, Zuo Y, Logan BE. Hydrogen and methane [100] Bowen EJ, Dolfing J, Davenport RJ, Read FL, Curtis TP. Low-temperature
production from swine wastewater using microbial electrolysis cells. Water Res limitation of bioreactor sludge in anaerobic treatment of domestic wastewater.
2009;43:1480–8. https://doi.org/10.1016/j.watres.2008.12.037. Water Sci Technol 2014;69:1004–13. https://doi.org/10.2166/wst.2013.821.
[75] Zhang X, He W, Ren L, Stager J, Evans PJ, Logan BE. COD removal characteristics [101] Ali SF. Determining the UK’s potential for heat recovery from wastewater using
in air-cathode microbial fuel cells. Bioresour Technol 2015;176:23–31. https:// steady state and dynamic modelling - preliminary results. WEENTECH Proc
doi.org/10.1016/j.biortech.2014.11.001. Energy 2019;5:107–21. https://doi.org/10.32438/wpe.58181.
[76] Dignac M-F, Ginestel P, Bruchet A, Audic J-M, Derenne S, Largeau C. Changes in [102] Sanderson DJ, Peacock DCP. Making rose diagrams fit-for-purpose. Earth Sci Rev
the organic coposition of wastewater during bilogical treatment as studied by 2020;201:103055. https://doi.org/10.1016/j.earscirev.2019.103055.
NMR and IR spectroscopies. Water Sci Technol 2001;43:51–8. https://doi.org/ [103] Wang Z, He Z. Demystifying terms for understanding bioelectrochemical systems
10.2166/wst.2001.0072. towards sustainable wastewater treatment. Curr Opin Electrochem 2020;19:14–9.
[77] Huang M, Li Y, Gu G. Chemical composition of organic matters in domestic https://doi.org/10.1016/j.coelec.2019.09.001.
wastewater. Desalination 2010;262:36–42. https://doi.org/10.1016/J. [104] Sugnaux M, Happe M, Cachelin CP, Gasperini A, Blatter M, Fischer F. Cathode
DESAL.2010.05.037. deposits favor methane generation in microbial electrolysis cell. Chem Eng J
[78] Eriksson E, Auffarth K, Henze M, Ledin A. Characteristics of grey wastewater. 2017;324:228–36. https://doi.org/10.1016/j.cej.2017.05.028.
Urban Water 2002;4:85–104. https://doi.org/10.1016/S1462-0758(01)00064-4. [105] Gil-Carrera L, Escapa A, Mehta P, Santoyo G, Guiot SR, Moran A, et al. Microbial
[79] Colin T, Bories A, Sire Y, Perrin R. Treatment and valorisation of winery electrolysis cell scale-up for combined wastewater treatment and hydrogen
wastewater by a new biophysical process (ECCF®). Water Sci Technol 2005;51: production. Bioresour Technol 2013;130:584–91. https://doi.org/10.1016/j.
99–106. https://doi.org/10.2166/wst.2005.0012. biortech.2012.12.062.
[80] Gimkiewicz C, Harnisch F. Waste water derived electroactive microbial biofilms: [106] Gil-Carrera L, Escapa A, Moreno R, Morán A. Reduced energy consumption during
growth, maintenance, and basic characterization. J Vis Exp 2013:50800. https:// low strength domestic wastewater treatment in a semi-pilot tubular microbial
doi.org/10.3791/50800. electrolysis cell. J Environ Manag 2013;122:1–7. https://doi.org/10.1016/j.
[81] Carmona-Martínez AA, Trably E, Milferstedt K, Lacroix R, Etcheverry L, Bernet N. jenvman.2013.03.001.
Long-term continuous production of H2 in a microbial electrolysis cell (MEC) [107] Gil-Carrera L, Escapa A, Carracedo B, Morán A, Gómez X. Performance of a semi-
treating saline wastewater. Water Res 2015;81:149–56. https://doi.org/10.1016/ pilot tubular microbial electrolysis cell (MEC) under several hydraulic retention
j.watres.2015.05.041. times and applied voltages. Bioresour Technol 2013;146:63–9. https://doi.org/
[82] Luo S, Jain A, Aguilera A, He Z. Effective control of biohythane composition 10.1016/j.biortech.2013.07.020.
through operational strategies in an innovative microbial electrolysis cell. Appl [108] Kadier A, Simayi Y, Kalil MS, Abdeshahian P, Hamid AA. A review of the
Energy 2017;206:879–86. https://doi.org/10.1016/j.apenergy.2017.08.241. substrates used in microbial electrolysis cells (MECs) for producing sustainable
[83] Logan BE. Scaling up microbial fuel cells and other bioelectrochemical systems. and clean hydrogen gas. Renew Energy 2014;71:466–72. https://doi.org/
Appl Microbiol Biotechnol 2010;85. https://doi.org/10.1007/s00253-009-2378- 10.1016/j.renene.2014.05.052.
9. [109] Mosse KPM, Patti AF, Christen EW, Cavagnaro TR. Review: winery wastewater
[84] Ambler JR, Logan BE. Evaluation of stainless steel cathodes and a bicarbonate quality and treatment options in Australia. Aust J Grape Wine Res 2011;17:
buffer for hydrogen production in microbial electrolysis cells using a new method 111–22. https://doi.org/10.1111/j.1755-0238.2011.00132.x.
for measuring gas production. Int J Hydrogen Energy 2011;36:160–6. https://doi. [110] Tartakovsky B, Mehta P, Santoyo G, Guiot SR. Maximizing hydrogen production
org/10.1016/j.ijhydene.2010.09.044. in a microbial electrolysis cell by real-time optimization of applied voltage. Int J
[85] Verea L, Savadogo O, Verde A, Campos J, Ginez F, Sebastian PJ. Performance of a Hydrogen Energy 2011;36:10557–64. https://doi.org/10.1016/j.
microbial electrolysis cell (MEC) for hydrogen production with a new process for ijhydene.2011.05.162.
the biofilm formation. Int J Hydrogen Energy 2014;39:8938–46. https://doi.org/ [111] Rabaey K, Lissens G, Siciliano SD, Verstraete W. A microbial fuel cell capable of
10.1016/j.ijhydene.2014.03.203. converting glucose to electricity at high rate and efficiency. Biotechnol Lett 2003;
[86] Heidrich ES, Dolfing J, Scott K, Edwards SR, Jones C, Curtis TP. Production of 25:1531–5. https://doi.org/10.1023/A:1025484009367.
hydrogen from domestic wastewater in a pilot-scale microbial electrolysis cell. [112] Borole AP, Reguera G, Ringeisen B, Wang Z-W, Feng Y, Kim BH. Electroactive
Appl Microbiol Biotechnol 2013;97:6979–89. https://doi.org/10.1007/s00253- biofilms: current status and future research needs. Energy Environ Sci 2011;4:
012-4456-7. 4813. https://doi.org/10.1039/c1ee02511b.
[87] Cotterill SE, Dolfing J, Jones C, Curtis TP, Heidrich ES. Low temperature domestic [113] Recio-Garrido D, Perrier M, Tartakovsky B. Modeling, optimization and control of
wastewater treatment in a microbial electrolysis cell with 1 m 2 anodes: towards bioelectrochemical systems. Chem Eng J 2016;289:180–90. https://doi.org/
system scale-up. Fuel Cell 2017;17:584–92. https://doi.org/10.1002/ 10.1016/j.cej.2015.11.112.
fuce.201700034. [114] Feng Y, Yang Q, Wang X, Logan BE. Treatment of carbon fiber brush anodes for
[88] Isabel San-Martín M, Mateos R, Carracedo B, Escapa A, Morán A. Pilot-scale improving power generation in air-cathode microbial fuel cells. J Power Sources
bioelectrochemical system for simultaneous nitrogen and carbon removal in 2010;195:1841–4. https://doi.org/10.1016/j.jpowsour.2009.10.030.
urban wastewater treatment plants. J Biosci Bioeng 2018;126:758–63. https:// [115] Wang L, Liu W, Kang L, Yang C, Zhou A, Wang A. Enhanced biohydrogen
doi.org/10.1016/j.jbiosc.2018.06.008. production from waste activated sludge in combined strategy of chemical
[89] Cusick RD, Bryan B, Parker DS, Merrill MD, Mehanna M, Kiely PD, et al. pretreatment and microbial electrolysis. Int J Hydrogen Energy 2014;39:
Performance of a pilot-scale continuous flow microbial electrolysis cell fed winery 11913–9. https://doi.org/10.1016/j.ijhydene.2014.06.006.
wastewater. Appl Microbiol Biotechnol 2011;89:2053–63. [116] Di Lorenzo M, Scott K, Curtis TP, Head IM. Effect of increasing anode surface area
[90] Henze M, Harremoes P, Jansen J la C, Arvin E. Wastewater treatment: biological on the performance of a single chamber microbial fuel cell. Chem Eng J 2010;156:
and chemical processes. Springer; 2002. 40–8. https://doi.org/10.1016/j.cej.2009.09.031.
[91] Oakley M. Environmental engineering. Mil Eng 2018;110:55–7. [117] Cheng S, Xing D, Logan BE. Electricity generation of single-chamber microbial
[92] Yildiz BS. Water and wastewater treatment: biological processes. Metrop. Sustain. fuel cells at low temperatures. Biosens Bioelectron 2011;26:1913–7. https://doi.
Underst. Improv. Urban Environ. Cambridge, UK: Woodhead Publishing; 2012. org/10.1016/j.bios.2010.05.016.
p. 406–28. https://doi.org/10.1533/9780857096463.3.406. [118] European Commission. H2020 - work Programme 2018-2020 13. Europe in a
[93] Logan BE. Microbial fuel cells. State College. Pennsylvania, USA: Wiley- changing world-Inclusive, innovative and reflective societies. 2018.
Interscience; 2008. [119] Kim KY, Yang W, Evans PJ, Logan BE. Continuous treatment of high strength
[94] Brown RK, Harnisch F, Wirth S, Wahlandt H, Dockhorn T, Dichtl N, et al. wastewaters using air-cathode microbial fuel cells. Bioresour Technol 2016;221:
Evaluating the effects of scaling up on the performance of bioelectrochemical 96–101. https://doi.org/10.1016/j.biortech.2016.09.031.
systems using a technical scale microbial electrolysis cell. Bioresour Technol [120] Leicester DD, Amezaga JM, Moore A, Heidrich ES. Optimising the hydraulic
2014;163:206–13. https://doi.org/10.1016/j.biortech.2014.04.044. retention time in a pilot-scale microbial electrolysis cell to achieve high
[95] Japan Sewage Works Association. Design standard for municipal wastewater volumetric treatment rates using concentrated domestic wastewater. Molecules
treatment plants. gcus.jp/wp/wp-content/.../735735ded23d4d28db9cc4f879e8d 2020;25:2945. https://doi.org/10.3390/molecules25122945.
a24.pdf. [Accessed 29 December 2019]. [121] Ivanov I, Ren L, Siegert M, Logan BE. A quantitative method to evaluate microbial
[96] IWA. Activated sludge process. IWA Publishing. IWA Publ; 2012. https://www.iw electrolysis cell effectiveness for energy recovery and wastewater treatment. Int J
apublishing.com/news/activated-sludge-process. [Accessed 29 July 2020]. Hydrogen Energy 2013;38:13135–42. https://doi.org/10.1016/j.
[97] Snyder R, Wyant D. Activated sludge process control, training manual for ijhydene.2013.07.123.
wastewater treatment plant operators. 2009. [122] Khoshnevisan B, Tabatabaei M, Tsapekos P, Rafiee S, Aghbashlo M, Lindeneg S,
[98] Li WW, Yu HQ, He Z. Towards sustainable wastewater treatment by using et al. Environmental life cycle assessment of different biorefinery platforms
microbial fuel cells-centered technologies. Energy Environ Sci 2014;7:911–24. valorizing municipal solid waste to bioenergy, microbial protein, lactic and
https://doi.org/10.1039/c3ee43106a. succinic acid. Renew Sustain Energy Rev 2020;117:109493. https://doi.org/
[99] Xia A, Herrmann C, Murphy JD. How do we optimize third-generation algal 10.1016/j.rser.2019.109493.
biofuels? Biofuels, Bioprod Biorefining 2015;9:358–67. https://doi.org/10.1002/
bbb.1550.
15

1 s2.0 S1364032120305670 Main

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

1 s2.0 S1364032120305670 Main

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 s2.0 S1364032120305670 Main

Uploaded by

Copyright:

Available Formats

Renewable and Sustainable Energy Reviews 133 (2020) 110279

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

Is bioelectrochemical energy production from wastewater a reality?

are energy intensive, accounting for as much as 3% of electricity con­

generated in AS enters an AD reactor, where yields are around a tenth of

2.5. Volumetric treatment rate 2.7. Temperature

[20,82,88,94,107] used a wastewater or synthetic equivalent with

mising the HRT would be a valuable step forward in understanding how

4.4. Reactor depth

modules containing 6 electrode pairs each. The electrodes in this reactor

study, performance was compared to small-scale laboratory reactors

attributed this to changes to the reactor design, specifically the use of

operated 1 m2 anodes, in the dimensions of 1.2 m by 0.8 m high, which

can be directly compared. Increasing the size of the electrodes by 16

ever, microbial contamination of the cathode chamber was the likely

Although modular, these had a different design than previous reactors in

wastewater was recirculated through both chambers to facilitate deni­

trification and the subsequent conversion of nitrite to molecular nitro­

gen. These could still be placed within an AS tank, although

modification to the design might present some difficulty in order to keep

lab scale designs (0.06 mW/m2 from 0.47 m2 anodes compared to

3.6W/m2 from 0.005 m2 anodes [111]).

You might also like

are energy intensive, accounting for as much as 3% of electricity con

wastewater was recirculated through both chambers to facilitate deni

trification and the subsequent conversion of nitrite to molecular nitro