Removal and recovery of inhibitory volatile fatty acids from mixed acid

fermentations by conventional electrodialysis

Rhys Jon Jonesa, Jaime Massanet-Nicolau, Alan Guwy, Giuliano C. Premier, Richard M. Dinsdale, Matthew

Reilly.

a Sustainable Environment Research Centre, Faculty of Computing Engineering and Science,

University of South Wales, Pontypridd CF37 1DL, United Kingdom.

E-mail: rhys.jones@southwales.ac.uk

Abstract

Hydrogen production during dark fermentation is inhibited by the co-production of

volatile fatty acids (VFAs) such as acetic and n-butyric acid. In this study, the

effectiveness of conventional electrodialysis (CED) in reducing VFA concentrations in

model solutions and hydrogen fermentation broths is evaluated. This is the first time

CED has been reported to remove VFAs from hydrogen fermentation broths. During 60

minutes of operation CED removed up to 99% of VFAs from model solutions, sucrose-

fed and grass-fed hydrogen fermentation broths, containing up to 1,200 mg·l-1 each of

acetic acid, propionic acid, i-butyric acid, n-butyric acid, i-valeric acid, and n-valeric

acid. CED's ability to remove VFAs from hydrogen fermentation broths suggests that

this technology is capable of improving hydrogen yields from dark fermentation.

Keywords

Anaerobic Digestion; Hydrogen; Grass; Electrodialysis; Volatile Fatty Acids;

Fermentation.

1. Introduction

Alternative energy sources to fossil fuels must be made readily available if the problems

1 of 27

associated with diminishing reserves, environmental damage, security of supply, and

increasing demand are to be overcome (Patterson et al., 2008; Yüksel, 2010).

Converting a biomass energy source to a hydrogen energy vector is attractive due to its

high calorific value of 122 kJ·g-1 (Azbar and Cetinkaya Dokgoz, 2010), and because it

represents a renewable and carbon neutral fuel (Meherkotay and Das, 2008). Fossil fuels

are currently more economically attractive than alternative energy vectors however and

questions regarding the long term availability of biomass in sufficient volumes to

establish a hydrogen economy need to be addressed before its large-scale production is

feasible (Hay et al., 2013).

Dark fermentation, which uses a range of substrates from industrial waste to energy

crops, is a widely researched, renewable, pathway to hydrogen. Theoretically, it is

possible to produce 4 mol H2·mol-1 hexose sugar when acetic acid is the only product

(Guwy et al., 2011), but experimental values tend to be in the range of 2.5 - 3.5 mol H2

mol-1 hexose sugar due to thermodynamic limitations (De Gioannis et al., 2014;

Rosales-Colunga et al., 2010). Hydrogen yields are also limited by the co-production of

volatile fatty acids (VFAs), in particular acetic acid and n-butyric acid, which form

during fermentation. VFA formation eventually prevents hydrogen production due to

end product inhibition (Hawkes et al., 2007; Hirata et al., 2005) and cell lysis of the

hydrogen producing bacteria (Choudhari et al., 2014; Tang et al., 2014)

Studies have demonstrated that reducing the concentrations of acetic acid and n-butyric

acid in hydrogen fermentation broths, via bipolar membrane electrodialysis, increases

hydrogen yields (Redwood et al., 2012a, 2012b; Tang et al., 2014). Bipolar membrane

electrodialysis is one of several electrodeionisation and electrodialysis processes, the

2 of 27

majority of which were described by Huang et al (Huang et al., 2007). Bipolar

membrane electrodialysis requires expensive, bipolar, water-splitting, membranes in

order to work and reported increases in hydrogen yield using glucose as a substrate are

inconsistent. For example, Tang et al. (Tang et al., 2014) report an increase from 1.7 to

2.2 mol H2·mol-1 glucose, whereas Redwood et al. (Redwood et al., 2012b) claim that

the removal of organic acids via bipolar membrane electrodialysis prolongs

fermentation time from 3 days to 3 weeks, tripling H2 production in doing so from 1 to 3

mol H2·mol-1 glucose. More recent work used activated carbon as a means to extract

inhibitory byproducts from hydrogen reactors fed by water hyacinth, and increased

hydrogen yields from 124.9 ml H2·g-1 total volatile solids to 134.9 ml H2·g-1 total

volatile solids (Cheng et al., 2015). As will be discussed below, VFAs are valuable in a

number of applications and their recovery following activated carbon treatment would

be difficult.

The work reported here compares the ability of conventional electrodialysis (CED) to

reduce VFA concentrations in model solutions, and post-fermentation broths that were

obtained following dark fermentation of sucrose and grass pellets. CED is a versatile

technology that can concentrate, dilute, and separate matter whilst producing relatively

low amounts of waste (Bailly et al., 2001; Huang et al., 2007; Lameloise and

Lewandowski, 2012). CED utilises less costly materials than bipolar membrane

electrodialysis, as was used by Redwood et al. (Redwood et al., 2012b) and Tang et al.

(Tang et al., 2014), and so in some applications it could possess an economical

advantage over bipolar membrane electrodialysis. VFAs extracted via CED would not

be contaminated as they might be using Cheng et al’s (Cheng et al., 2015) activated

carbon method, so they can easily be used as substrates in downstream bioenergy

3 of 27

systems such as photofermentation, microbial electrolysis, and microbial fuel cells,

theoretically producing up to 12 mol H2·mol-1 hexose sugar (Claassen et al., 2010;

Guwy et al., 2011; Redwood et al., 2012a, 2012b; Singhania et al., 2013, 2012).

Acetic acid, one of the VFAs that can be extracted, also has uses beyond conversion to

an energy carrier. It has been called “one of the world’s most important chemicals”

(Ijmker et al., 2014) due to widespread use in manufacturing polymers, polyethylene

terephthalate, and solvents. According to market forecasts, its global demand in 2013

was 10.4 mt, and by 2020 its market revenue is expected to reach $12.2 billion. There is

also demand for butyric acid, which is used in the cosmetic, food, and pharmaceutical

industries and along with other VFAs, it can also be used as a precursor to fuels and

chemicals (Agler et al., 2011; Blahušiak et al., 2013; Choudhari et al., 2014).

The aim of this work is to establish how well conventional electrodialysis can reduce

the VFA concentrations of post-fermentation broths compared to model solutions. It is

important to make this comparison because post-fermentation broths are high in

microbial and particulate matter, both of which are known to contribute to fouling and

thus a gradual reduction of the ion flux within a CED stack (Hongo et al., 1986; Lee et

al., 2003; Singhania et al., 2012; Strathmann, 2010). Fouling is reported to increase

electrical resistance within CED stacks, thereby reducing current efficiency, which is

widely used to quantify the performance of electrodialysis processes (Ferrer et al., 2006;

Huang et al., 2007; Lameloise and Lewandowski, 2012; Wang et al., 2011, 2010). If

CED is found to be capable of removing VFAs from post-fermentation broths in similar

quantities to model solutions, there would be scope for further studies in to the ability of

CED to remove VFAs from working reactors. The possibility exists that CED may also

4 of 27

be capable of removing VFAs from dilute aqueous solutions, which has hitherto proved

problematic for the production of chemicals and fuels from the otherwise promising

carboxylate platform (Agler et al., 2011), although this application falls outside of the

scope of this study. This is the first time the ability of conventional electrodialysis to

remove VFAs from post-fermentation broths has been reported and quantified. It is also

the first study in which a comparison has been drawn between the ability of

conventional electrodialysis to treat model solutions and post-fermentation broths.

2. Materials and Methods

2.1 Experimental design

Conventional electrodialysis was carried out on controls; model solutions that contained

VFAs and deionised water to establish baseline VFA removal data in the absence of

microbial and particulate matter. CED was then performed on hydrogen fermentation

broths collected following sucrose and grass fermentations within continuously stirred

batch reactors, which contained microbial and particulate matter, and the VFA removal

data were compared to the baseline established by CED of model solutions. CED was

carried out on two batches each of the model solution, sucrose fed reactor broth, and

grass fed reactor broth.

2.2 Apparatus

The apparatus for conventional electrodialysis (Figure 1) is typically referred to as a

CED “stack” and consists of two electrodes (one anode and one cathode) across which

an electrical potential is established by an external power supply, and which are housed

within an anode chamber and a cathode chamber. Between the electrodes are alternating

anion exchange membranes (AEMs) and cation exchange membranes (CEMs), which

5 of 27

are separated by spacers to create Diluate and Concentrate Chambers, through which

liquids flow. A wider housing encases and supports the aforementioned components. In

this study the CED stack was manufactured by PCCell GmbH (Heusweiler, Germany),

and the model was an ED 64002 with 20 cell pairs (0.128 m2 effective membrane area).

2.3 Preparation of model solutions and stock solutions

Three, distinct, solutions flow through a CED stack; concentrate, dilute, and electrolyte.

In this study, the concentrate (250 ml phosphate buffer solution: 8,000 mg·l-1 NaCl;

1,440 mg·l-1 Na2HPO4; 240 mg·l-1 KH2PO4; and 200 mg·l-1 KCl) and electrolyte (1,000

ml of Na2SO4 at 35,510 mg·l-1) remained the same throughout all experiments. The

dilute varied between experiments and consisted either of 250 ml model VFA solutions

(1,000 mg·l-1 each of acetic, propionic, i-butyric, n- butyric, i-valeric and n-valeric acid,

and 8,000 mg·l-1 NaCl; 1,440 mg·l-1 Na2HPO4; 240 mg·l-1 KH2PO4; and 200 mg·l-1 KCl),

or hydrogen fermentation broths, the compositions and preparations of which are

outlined below. VFA concentrations in model solutions where chosen to approximate

VFA levels obtained from fermentative hydrogen production using a variety of

substrates such as sewage biosolids, grass and wheat co-product (Massanet-Nicolau et

al., 2013, 2010; Yu et al., 2014).

2.4 Preparation of hydrogen fermentation broths

Two dark hydrogen fermentation reactors, one sucrose-fed, and one grass pellet fed

continually stirred batch reactors were adapted from Reilly et al’s (Reilly et al., 2014)

method as follows. The sucrose-fed continuously stirred batch reactors contained 100

ml of 15 g·l-1 sucrose solution, 120 ml of inoculum (sewage sludge from a sewage

treatment works serving a large metropolitan area), and 280 ml of water. The grass

6 of 27

pellet fed continuously stirred batch reactors were identical except 100 ml of 80 g·l-1

grass pellet suspension was added instead of sucrose solution. To inactivate

methanogenic organisms and to prevent their eventual consumption of hydrogen, the

inoculum was heated to 120°C for 40 minutes and once cooled to room temperature,

120 ml of inoculum was added to each continuously stirred batch reactor. To establish

anaerobic conditions for dark fermentation, each continuously stirred batch reactor was

sparged with carbon dioxide until its pH level reached 6.25. The headspace of each

continuously stirred batch reactor was then flushed with CO2 in order to ensure

anaerobic conditions. The four continuously stirred batch reactors were then incubated

at 35°C for 40 hours. Hydrogen yields from the sucrose and grass fermentation broths

were 240.4 ml g-1 VS and 28.2 ml g-1 VS respectively.

Following incubation, the contents of each of the four continuously stirred batch

reactors were separately centrifuged at 3,500 x g for 40 minutes, and then the clarified

liquor was aspirated off and filtered to 2.5 μm. The four resulting effluents then had

8,000 mg·l-1 NaCl; 1,440 mg·l-1 Na2HPO4; 240 mg·l-1 KH2PO4; and 200 mg·l-1 KCl

added to them to minimise the conductivity gradient between the diluate and

concentrate streams ensuring the majority of ion flux was as a result of electrodialysis.

2.5 Operation of conventional electrodialysis stack

The CED stack (Figure 2) was connected to three, separate, circuits; one fed by the

electrode rinse reservoir, another by the concentrate (phosphate buffer solution)

reservoir, and another by the diluate reservoir, which contained either model solutions

or treated hydrogen fermentation broths depending on the experiment. Each circuit was

pumped through the CED stack with peristaltic pumps at a rate of 2.0 l·hr-1. Each

7 of 27

reservoir was stirred with magnetic stirrers for the duration of the CED experiment.

Once the CED stack was filled with liquid, 18 V of electrical potential was applied

across the stack using an Array (Nanjing, China) 3644A programmable DC power

supply.

2.6 Sampling technique and analysis

Samples (1 ml) were taken from the concentrate and diluate reservoirs as soon as the

power supply was activated, and at ten minute intervals for 60 minutes thereafter. The

VFA content of the samples taken from the concentrate and diluate reservoirs were

measured using a headspace gas chromatograph, equipped with a flame ionization

detector and a free fatty acid phase column (Nukol), which operated at 190°C and at 14

psi with a nitrogen carrier gas according to the method of Cruwys et al. (2002). At the

same time as samples were taken from the concentrate and diluate circuits, the current

(A) flowing through the CED stack was recorded.

3. Results and discussion

3.1 CED of model solutions

In the first experiment, CED was used to reduce the concentration of acetic, propionic,

i- butyric, n-butyric, i-valeric, and n-valeric acids from model solutions. The initial

concentrations of all acids were in the range of 800 and 1,000 mg·l-1, and were reduced

consistently by up to two orders of magnitude following 60 minutes of CED, and in the

second run, acetic acid was reduced to below detectable levels for gas chromatography

(less than 10 mg·l-1) (Figure 3). The concurrent increases in the VFA concentrations of

the concentrate circuit were expected and demonstrated that ion flux progressed as

expected.

8 of 27

3.2 CED of hydrogen fermentation broths from sucrose-fed continuously stirred batch

reactors

Figure 4 shows that during the CED of hydrogen fermentation broths from sucrose fed

continuously stirred batch reactors, the concentrations of acetic and n-butyric acid were

successfully reduced by at least 92%, and the majority of removal occurred during the

first 30 minutes. In both instances, acetic acid was removed entirely.

3.3 CED of hydrogen fermentation broths from grass pellet fed continuously stirred

batch reactors During the CED of grass reactor hydrogen fermentation broths the acetic

and n-butyric acid concentrations were successfully reduced by up to 96%, with the

fastest rate of reduction occurring in the initial 30 minutes (Figure 5). The removal of

other VFAs is not presented as they were not present in concentrations greater than 40

mg·l-1 following dark hydrogen fermentation.

As demonstrated by the conventional electrodialysis of model solutions (Figure 3), it is

possible to reduce an initial acetic acid concentration by around 1,000 mg·l-1, and

although the n-butyric acid concentration of the diluate circuit (hydrogen fermentation

broths from grass fed continuously stirred batch reactors) is initially lower than in the

model solutions, it was still possible to reduce the concentration from 642 mg·l-1 to 18

mg·l-1. Concurrently, the concentrate circuit becomes more concentrated with acetic and

n-butyric acid, illustrating the ion flux from the diluate circuit to the concentrate circuit.

3.4 Comparing percentage reduction in VFA concentrations

Table 1 shows that in terms of percentage reduction, there is little difference between

the ability of conventional electrodialysis to treat model solutions and hydrogen

fermentation broths. Mean acetic acid concentration reduction is 98% in model

9 of 27

solutions (from 1,031 mg·l-1 to 20 mg·l-1), more than 99% in sucrose fed reactor broths

(753 mg·l-1 to less than 10 mg·l-1), and 94% in grass fed reactor broths (1,189 mg·l-1 to

71 mg·l-1). It also shows that acetic acid is typically the most readily removed of the two

VFA species, except when comparing reduction data from the grass reactor

experiments, in which n-butyric acid concentration is reduced by 96% (642 mg·l-1 to 25

mg·l-1) as opposed to 94% for acetic acid (1189 mg·l-1 to 71 mg·l-1). There is a

correlation between the molarity of the different VFA species and their rates of

transport (Figure 3) suggesting that initial concentration gradient of the acid influences

the rate at which it is removed from the concentrate stream.

After 50 minutes into the CED of grass fed reactor broths, acetic acid levels in the

diluate stream increase (Figure 5). This increase was found to be statistically significant

(P=0.0406 using a Student’s T-Test). Back diffusion, a phenomenon whereby a

concentration gradient overcomes the electrical gradient and ions move against

conventional current would explain this observation (Bailly et al., 2001).

Numerous investigations into the removal of VFAs from solutions exist, however they

employ methods other than CED and there is little uniformity between the variables

tested so comparisons between methodologies are therefore difficult. For example,

Choudari et al. (Choudhari et al., 2014) use pervaporation with polyether block amide

based composite membranes to remove butyric acid of concentration 5,950 mg·l-1 from

2 l of anaerobic digestion broths. In that study, a total membrane area of 0.01591 m2

was reported to be able to remove up to 120 g·m-2·hr-1 of butyric acid, which equates to

a 23% reduction in butyric acid concentration in 60 minutes as compared to a maximum

of over 99% in this work, which represents a fourfold increase in performance. An

10 of 27

advantage of polyether block amide based membranes is that less costly materials than

polymer-based membranes such as graphene can be used, but this appears to be at the

expense of removing larger amounts of VFAs, which, as discussed are extremely

valuable.

Tang et al. (Tang et al., 2014), used bipolar membrane electrodialysis to remove acetic

acid from ethanol type fermentation broths and report a reduction in acetic acid

concentration from 1,600 mg·l-1 to 110 mg·l-1 following 39 minutes of bipolar

membrane electrodialysis. Figures 3, 4, and 5 demonstrate that CED is capable of

similar rates of reduction, with the advantage that anion exchange membranes and

cation exchange membranes are less costly than bipolar membranes.

CED is a technology capable of reducing the acetic and n-butyric acid concentrations of

hydrogen fermentation broths provided that the pre-CED filtration steps outlined in this

work are taken. In moving towards an automated process, in which VFAs are removed

continuously with a view to improving hydrogen yields, it is important to establish how

effective CED is at treating hydrogen fermentation broths with fewer pretreatment steps

and if successful, to use CED to remove VFAs from hydrogen reactors as they evolve.

If this can be achieved, conventional electrodialysis represents a way in which to

increase hydrogen yields from a given dark fermentation reactor at potentially low cost.

3.5 pH levels

In this study, the aim of CED was to remove VFAs from model solutions and post-

fermentation broths, and in doing so pH levels were affected. Following CED and the

associated removal of VFAs, the model solution's pH rose from 3.03 to 3.89, the

sucrose fermentation broth's pH rose from 6.74 to 8.72, and the grass fermentation

11 of 27

broth's pH rose from 6.31 to 6.48. It is known that the pH level of a hydrogen reactor

has an effect on gas yields but in these instances CED was used to treat fermentation

broths that were obtained once fermentation had been completed. Further studies will

examine the effect of VFA removal together with changes in pH on ongoing

fermentative hydrogen production.

3.6 Electrical current data

The electrical current data shown in Figure 6 follow a trend consistent with their

respective acetic and n-butyric acid reduction rates shown in Figures 3, 4 and 5. For the

grass fed and sucrose fed fermentations (Figures 4 and 5) the majority of acetic and n-

butyric acid flux had occurred by 30 minutes, and this is reflected by the sharp decrease

in current from greater than 1.5 A to roughly 0.35 A (Figure 6) during the first 20

minutes of those experiments.

The diluate circuit (Figure 2) in these experiments became gradually depleted of ions,

and this is reflected by the current data (Figure 6). As ions become retained within the

concentrate circuit, overall stack resistance increases because ions were not replaced in

the diluate circuit. If CED is used to treat hydrogen fermentation broths during

continuous dark fermentation, the likelihood is that the diluate circuit will be

replenished with acetic acid and n-butyric acid and their associated ions continuously,

resulting in a consistent electrical current during CED.

4. Conclusions

CED can reduce the acetic and butyric acid concentrations of a 250 ml diluate circuit by

over 92% from an initial concentration of up to 1,200 mg·l-1 for both acids without

bipolar membranes. VFA removal rates from CED of sucrose and grass fermentation

12 of 27

broths as were similar to those from model solutions. Decreases in electrical current

within the CED stack were observed as VFAs were removed from the diluate circuit.

The ability of CED to successfully remove VFAs from fermentation broths indicates it

could be used to increase hydrogen and VFA yields during anaerobic processes and

facilitate their recovery.

Acknowledgements

The authors would like to acknowledge the support of the EPSRC Hydrogen and Fuel

Cell Consortium Funding (EP/J016454/1), University of South Wales Centenary

Scholarship Fund and the European Regional Development Fund for the

CymruH2Wales project.

References

1. Agler, M.T., Wrenn, B.A., Zinder, S.H., Angenent, L.T., 2011. Waste to

bioproduct conversion with undefined mixed cultures: the carboxylate platform.

Trends Biotechnol. 29, 70–78.

2. Azbar, N., Cetinkaya Dokgoz, F.T., 2010. The effect of dilution and l-malic acid

addition on bio-hydrogen production with Rhodopseudomonas palustris from ef-

fluent of an acidogenic anaerobic reactor. Int. J. Hydrogen Energy 35, 5028–

5033.

3. Bailly, M., Roux-de Balmann, H., Aimar, P., Lutin, F., Cheryan, M., 2001. Pro-

duction processes of fermented organic acids targeted around membrane opera-

tions: design of the concentration step by conventional electrodialysis. J. Memb.

Sci. 191, 129–142.

13 of 27

4. Blahušiak, M., Schlosser, Š., Marták, J., 2013. Extraction of butyric acid with a

solvent containing ammonium ionic liquid. Sep. Purif. Technol. 119, 102–111.

5. Cheng, J., Lin, R., Song, W., Xia, A., Zhou, J., Cen, K., 2015. Enhancement of

fermentative hydrogen production from hydrolyzed water hyacinth with activ-

ated carbon detoxification and bacteria domestication. Int. J. Hydrogen Energy

40, 2545–2551.

6. Choudhari, S.K., Cerrone, F., Woods, T., Joyce, K., Flaherty, V.O., Connor,

K.O., Babu, R., 2014. Pervaporation separation of butyric acid from aqueous

and anaerobic digestion (AD) solutions using PEBA based composite mem-

branes. J. Ind. Eng. Chem.

7. Claassen, P. a. M., de Vrije, T., Koukios, E., van Niel, E., Eroglu, I., Modigell,

M., Friedl, A., Wukovits, W., Ahrer, W., 2010. Non-thermal production of pure

hydrogen from biomass: HYVOLUTION. J. Clean. Prod. 18, S4–S8.

8. Cruwys, J.A., Dinsdale, R.M., Hawkes, F.R., Hawkes, D.L., 2002. Development

of a static headspace gas chromatographic procedure for the routine analysis of

volatile fatty acids in wastewaters. J. Chromatogr. A 945, 195–209.

9. De Gioannis, G., Friargiu, M., Massi, E., Muntoni, a., Polettini, a., Pomi, R.,

Spiga, D., 2014. Biohydrogen production from dark fermentation of cheese

whey: Influence of pH. Int. J. Hydrogen Energy 39, 20930–20941.

10. Ferrer, J.S.J., Laborie, S., Durand, G., Rakib, M., 2006. Formic acid regenera-

tion by electromembrane processes. J. Memb. Sci. 280, 509–516.

14 of 27

11. Guwy, A.J.J., Dinsdale, R.M.M., Kim, J.R.R., Massanet-Nicolau, J., Premier,

G., 2011. Fermentative biohydrogen production systems integration. Bioresour.

Technol. 102, 8534–8542.

12. Hawkes, F.R., Hussy, I., Kyazze, G., Dinsdale, R., Hawkes, D.L., 2007. Con-

tinuous dark fermentative hydrogen production by mesophilic microflora: Prin-

ciples and progress. Int. J. Hydrogen Energy 32, 172–184.

13. Hay, J.X.W., Wu, T.Y., Juan, J.C., Md. Jahim, J., 2013. Biohydrogen production

through photo fermentation or dark fermentation using waste as a substrate:

Overview, economics, and future prospects of hydrogen usage. Biofuels,

Bioprod. Biorefining 7, 334–352.

14. Hirata, M., Gao, M., Toorisaka, E., Takanashi, H., Hano, T., 2005. Production of

lactic acid by continuous electrodialysis fermentation with a glucose concentra-

tion controller. Biochem. Eng. J. 25, 159–163.

15. Hongo, M., Nomura, Y., Iwahara, M., 1986. Novel method of lactic acid pro-

duction by electrodialysis fermentation. Appl. Environ. Microbiol. 52, 314–319.

16. Huang, C., Xu, T., Zhang, Y., Xue, Y., Chen, G., 2007. Application of electro-

dialysis to the production of organic acids: State-of-the-art and recent develop-

ments. J. Memb. Sci. 288, 1–12.

17. Ijmker, H.M., Gramblička, M., Kersten, S.R.A., Ham, A.G.J. Van Der, Schuur,

B., 2014. Acetic acid extraction from aqueous solutions using fatty acids. Sep.

Purif. Technol. 125, 256–263.

15 of 27

18. Lameloise, M.-L., Lewandowski, R., 2012. Recovering l-malic acid from a

beverage industry waste water: Experimental study of the conversion stage using

bipolar membrane electrodialysis. J. Memb. Sci. 403-404, 196–202.

19. Lee, H.-J., Oh, S.-J., Moon, S.-H., 2003. Recovery of ammonium sulfate from

fermentation waste by electrodialysis. Water Res. 37, 1091–1099.

20. Massanet-Nicolau, J., Dinsdale, R., Guwy, A., Shipley, G., 2013. Use of real

time gas production data for more accurate comparison ofcontinuous single-

stage and two-stage fermentation. Bioresour. Technol. 129, 561–7.

21. Massanet-Nicolau, J., Guwy, A., Dinsdale, R., Premier, G., Esteves, S., 2010.

Production of hydrogen from sewage biosolids in a continuously fed bioreactor:

Effect of hydraulic retention time and sparging. Int. J. Hydrogen Energy 35,

469–478.

22. Meherkotay, S., Das, D., 2008. Biohydrogen as a renewable energy resource—

Prospects and potentials. Int. J. Hydrogen Energy 33, 258–263.

23. Patterson, T., Dinsdale, R., Esteves, S., 2008. Review of Energy Balances and

Emissions Associated with Biomass-Based Transport Fuels Relevant to the

United Kingdom Context. Energy & Fuels 22, 3506–3512.

24. Redwood, M.D., Orozco, R.L., Majewski, A.J., Macaskie, L.E., 2012a. An in-

tegrated biohydrogen refinery: synergy of photofermentation, extractive fer-

mentation and hydrothermal hydrolysis of food wastes. Bioresour. Technol. 119,

384–92.

16 of 27

25. Redwood, M.D., Orozco, R.L., Majewski, A.J., Macaskie, L.E., 2012b. Electro-

extractive fermentation for efficient biohydrogen production. Bioresour. Tech-

nol. 107, 166–74.

26. Reilly, M., Dinsdale, R., Guwy, A., 2014. Mesophilic biohydrogen production

from calcium hydroxide treated wheat straw. Int. J. Hydrogen Energy 39,

16891–16901.

27. Rosales-Colunga, L.M., Razo-Flores, E., De León-Rodríguez, A., 2010. Influ-

ence of pH control on hydrogen production by Escherichia coli ΔhycA ΔlacI us-

ing cheese whey as substrate. J. Biotechnol. 150, 152–153.

28. Singhania, R.R., Christophe, G., Perchet, G., Troquet, J., Larroche, C., 2012.

Immersed membrane bioreactors: an overview with special emphasis on anaer-

obic bioprocesses. Bioresour. Technol. 122, 171–80.

29. Singhania, R.R., Patel, A.K., Christophe, G., Fontanille, P., Larroche, C., 2013.

Biological upgrading of volatile fatty acids, key intermediates for the valoriza-

tion of biowaste through dark anaerobic fermentation. Bioresour. Technol. 145,

166–74.

30. Strathmann, H., 2010. Electrodialysis, a mature technology with a multitude of

new applications. Desalination 264, 268–288.

31. Tang, J., Jia, S., Qu, S., Xiao, Y., Yuan, Y., Ren, N.-Q., 2014. An integrated bio-

logical hydrogen production process based on ethanol-type fermentation and bi-

polar membrane electrodialysis. Int. J. Hydrogen Energy 39, 13375–13380.

17 of 27

32. Wang, Y., Huang, C., Xu, T., 2011. Which is more competitive for production

of organic acids, ion-exchange or electrodialysis with bipolar membranes? J.

Memb. Sci. 374, 150–156.

33. Wang, Y., Zhang, X., Xu, T., 2010. Integration of conventional electrodialysis

and electrodialysis with bipolar membranes for production of organic acids. J.

Memb. Sci. 365, 294–301.

34. Yu, L., Bule, M., Ma, J., Zhao, Q., Frear, C., Chen, S., 2014. Enhancing volatile

fatty acid (VFA) and bio-methane production from lawn grass with pretreat-

ment. Bioresour. Technol. 162, 243–249.

35. Yüksel, I., 2010. Hydropower for sustainable water and energy development.

Renew. Sustain. Energy Rev. 14, 462–469.

18 of 27

Figure 1: Configuration of CED stack used in this study. Movement of acetate and

butyrate species are shown.

Figure 2: Schematic diagram of the CED stack used in this study showing electrolyte,

concentrate and diluate circuits and reservoirs.

Figure 3: Concentration flux of individual acids in concentrate and diluate circuits

during CED of model solutions.

Figure 4: Acetic acid and n-butyric acid concentration flux in concentrate and diluate

circuits during CED of fermentation broths from sucrose fed hydrogen fermentations.

Figure 5: Acetic acid and n-butyric acid concentration flux in concentrate and diluate

circuits during the CED of fermentation broths from grass fed hydrogen fermentations.

Figure 6: The decrease in electrical current observed as the CED of model solutions,

sucrose fed continuously stirred batch reactors, and grass fed continuously stirred batch

reactors progressed.

19 of 27

Table 1: Mean concentration reductions in acetic acid and n-butyric acid in diluate

circuits composed of model solutions and hydrogen fermentation broths

20 of 27

21 of 27

22 of 27

23 of 27

24 of 27

25 of 27

26 of 27

Diluate circuitStandard deviation

Standard deviation

(mg·l-1) (%) (mg·l-1) (mg·l-1) (%) (mg·l-1)

Model solution 1,031 98% 33 919 95% 35Sucrose broth 753 >99% 97 722 97% 18Grass broth 1,118 94% 14 617 96% 7

Mean acetic acid concentration reduction (n=2)

Mean n-butyric acid concentration reduction (n=2)

27 of 27