1

Utilising biohydrogen to increase methane production, energy

yields and process efficiency via two stage anaerobic digestion of

grass.

Jaime Massanet- Nicolaua, Richard Dinsdale, Alan Guwy, Gary Shipley.

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

University of South Wales, Pontypridd CF37 1DL, United Kingdom. E-mail:

jmassane@glam.ac.uk

Abstract

Real time measurement of gas production and composition were used to examine the benefits of

two stage anaerobic digestion (AD) over a single stage AD, using pelletized grass as a

feedstock. Controlled, parallel digestion experiments were performed in order to directly

compare a two stage digestion system producing hydrogen and methane, with a single stage

system producing just methane. The results indicated that as well as producing additional

energy in the form of hydrogen, two stage digestion also resulted in significant increases to

methane production, overall energy yields, and digester stability (as indicated by bicarbonate

alkalinity and volatile fatty acid removal). Two stage AD resulted in an increase in energy

yields from 10.36 MJ kg-1 VS to 11.74 MJ kg-1 VS, an increase of 13.4%. Using a two stage

system also permitted a much shorter hydraulic retention time of 12 days whilst maintaining

process stability.

1. Introduction

Increasing global energy consumption combined with concerns about the environmental

damage caused by continued dependence on fossil fuels has prompted a great deal of research

into alternative sources of energy which are both sustainable and carbon neutral. One such

energy source is biomass which can be converted into energy rich gases such as hydrogen and

1

2

methane via anaerobic digestion (AD). A large range of biomass types are compatible with

anaerobic digestion, including municipal and industrial waste streams as well as crops grown

specifically for energy production. One such energy crop is grass, which can be grown on

marginal land not used for the production of food (Tilman et al., 2006). Such grassland accounts

for 30% of the UKs land use and includes agricultural land which has been set aside under EU

agricultural policies. Grasses are typically high in sugars such as fructose and lower in

lignocellulose than many other bioenergy crops and so are ideally suited to fermentative energy

production (Adler et al., 2006; Allison et al., 2009). Maintaining grass lands confers other

environmental benefits such as sequestering carbon in the soil (Murphy and Power, 2009).

Additionally, if the reduction in sheep and cattle farming in the UK over the last decade

continues, a surplus of grass will be available for bioenergy production (CROPGEN, 2007).

Mathematical models developed at Aberystwyth University indicate that if 25% of the UK’s

permanent grassland were used, 12,945 million tonnes of grass of grass per annum could be

produced (Toop, 2013). Even at currently reported hydrogen and methane yields this would

represent a significant source of sustainable, carbon neutral bioenergy.

Anaerobic digestion as a means of producing bioenergy is a well-established process

worldwide; however innovations are constantly being made, enabling the process to cope with a

wider range of feedstocks, to produce higher yields of energy, and to operate at lower costs and

greater efficiency. One such innovation is two stage digestion in which the feedstock is digested

in two separate stages, a high rate acidogenic stage and a slower methanogenic stage. Many

advantages have been claimed for two stage anaerobic digestion, including greater process

stability and higher yields (De Gioannis et al., 2008; Lee and Chung, 2010). Two stage AD can

also be used to produce hydrogen during the acidogenic stage in addition to methane (Guwy et

al., 2011).

Two stage AD has been studied at various scales using a range of feedstocks (Dareioti and

Kornaros, 2014; Zuo et al., 2014), however there are comparatively few robustly controlled

2

3

studies where it is compared with single stage AD using complex feedstocks. In many cases

researchers conduct two stage AD trials and compare their results with those obtained by other

groups using single stage AD (Chu et al., 2008). This approach is problematic since there is

little consistency with regards to feedstock, methodology or process performance measurement.

The operation of two stage AD experiments in the laboratory are also limited by the capability

of lab scale digestion apparatus. In particular due to the problematic rheology of feedstocks with

a high solids content, digesters must be fed manually, usually only once or twice per day, the

same is true of gas production and composition measurements. Methane and hydrogen

production can vary greatly in response to feeding events, consequently depending on when

digesters are fed and when gas production is measured, yields of methane or hydrogen can be

severely under or over estimated.

The lack of detailed, controlled and robust evaluations of two stage AD using complex

feedstocks has meant that its advantages are not well understood within the industrial sector.

This often leads to more costly and energy intensive methods being used to improve methane

yields such as thermal hydrolysis. Despite this, some well designed and controlled studies have

clearly demonstrated the benefits arising from two stage AD. Nielsen et al (2004) demonstrated

improvements in performance using two stage AD to treat cow manure, and recently the

University of South Wales demonstrated a 38% increase in energy yields using two stage AD

with flour milling co-product (Massanet-Nicolau et al., 2013).

The research reported here is a comparison of two stage and single stage digestion systems

using pelletized grass as a feedstock. The study is designed to address the shortcomings of

previous evaluations of two stage digestion discussed above; two stage and single stage digester

systems are evaluated simultaneously using exactly the same batches of feedstock and gas

production rates and composition are measured in real time to avoid bias when calculating

methane and hydrogen yields. The study quantifies the advantages of two stage

3

4

hydrogen/methane digestion in terms of overall energy yield, process stability and process

efficiency.

2. Materials and Methods

2.1 Digestion experiments

Three different digester systems were evaluated in parallel, each using the same feedstock. The

first of these was a single stage digester with a relatively long hydraulic retention time (HRT) of

20 days, producing just methane. This configuration is similar to conventional AD methodology

employed at sewage treatment works. Secondly, a two stage system was evaluated, comprising

a hydrogen producing digester with an 18h HRT and a methane producing digester with a HRT

of 11.25 days for an overall HRT of 12 days. Finally another two stage system was evaluated,

again with a hydrogen stage of 18h HRT, but with a methane stage of 19.25 days so that the

overall HRT was 20 days, equal to the single stage system being evaluated. Figure 1 is a

schematic showing how these digestion systems were evaluated in parallel.

2.2 Hydrogen digester

A continuously stirred hydrogen digester with a working volume of 10 L and a headspace

volume of 2L was used in these experiments. The hydrogen digester was equipped with

instrumentation allowing pH, redox potential, and temperature to be monitored in real time

during digestion experiments. The hydrogen digester was equipped with sensors for continuous

measurement of both gas production and composition (H2, CO2 and CH4). Data from these

sensors were recorded using a PC equipped with a data acquisition card and a custom

monitoring program written using the LabViewTM programming application. The contents of the

hydrogen digester were maintained at 35oC using a thermostatically controlled electric heating

jacket. The pH of the digester was maintained at 5.5 via the automated addition of 2M NaOH.

The digester was fed automatically, using computer controlled valves once per hour with

sufficient feedstock to maintain a HRT of 18 hours.

4

5

The hydrogen producing digester was started by filling it with 5% heat treated inoculum and

95% feedstock by volume. In order to build up levels of hydrogen producing microorganisms,

the digester was initially operated in batch mode (with no additional feeding) until production of

hydrogen occurred (approximately 18 hours). Continuous feeding then commenced and the

digester was operated for a period of 30 days prior to the commencement of this study in order

to allow hydrogen production to stabilize as indicated by the stability of key parameters

including biogas production and composition, pH and VFA production.

2.3 Methane digesters

Three identical methane digesters with working volumes of 25 L were used in these

experiments. As with the hydrogen digester, these were equipped with sensors enabling

continuous measurement and recording of gas production as well as CO2 and CH4 content. The

contents of the digester were maintained at 35oC using a thermostatically controlled water bath.

The pH of the digesters was not actively controlled but was monitored daily to ensure it

remained at pH 7.0 +/- 0.5. The digesters were fed manually once per day; the single-stage

digester was fed with feedstock in sufficient quantity to maintain a HRT of 20 days. The two

two stage digesters were fed with effluent from the hydrogen digester no more than 24 hours

old. They were maintained at HRTs of 11.25 days, and 19.25 days, so that together with the

hydrogen digester, they formed digestion systems with overall HRTs of 12 days and 20 days

respectively (See Figure 1).The digesters were started by filling them with inoculum then

feeding them as described above. As with the hydrogen digester, this operation continued for a

period of 30 days prior to the commencement of this study, to enable a steady state to be

reached with regards to methane production as indicated by the stability of key parameters

including biogas production and composition, bicarbonate alkalinity and VFA production.

2.4 Feedstock and inoculum

5

6

The substrate used during these experiments was pelletized grass which is sold as animal feed

(Baileys Horse Feeds UK). The pellets are made from grass harvested from commercial grass

leys and are representative of grass varieties typically used in UK grass lands. In addition to

soluble sugars, the grass is known to contain significant levels of structurally complex

carbohydrates (Kyazze et al., 2008) so to improve digestion they were hydrolysed using alkali,

before being fed into the digesters. The pellets were soaked in water overnight in at 2-5oC

causing them to break apart. They were then diluted with water and sufficient NaOH to raise the

pH to 12 and to obtain a nominal volatile solids (VS) content of 50g L-1. This feed was then

transferred to a feed storage tank where it was pumped into the digester as required. The storage

tank was maintained at a temperature of 1.5-2.5oC to limit microbial growth. The high pH

resulting from the alkali pre-treatment limits microbial activity during storage as well as

hydrolysing complex carbohydrates.

The inoculum used in these experiments was anaerobic digester effluent taken from a local

sewage treatment works. Prior to use in the hydrogen digester, the inoculum was heated to

110oC for 20 minutes to selectively inactivate methanogenic microorganisms. In the

methanogenic digesters the effluent was used without modification.

2.5 Compositional analysis

The composition of the feedstock used in these experiments, along with samples taken from the

hydrogen and methane digesters during operation, was determined. Volatile solids (VS) were

measured in triplicate using standard methods (Clesceri et al., 1999). Volatile fatty acids

(VFAs) were measured according to the method of Cruwys et al. (2002) using a Perkin Elmer

head space gas chromatograph in conjunction with a flame ionisation detector and a Nukol free

fatty acid phase column running at 190oC and 14 psi with nitrogen as a carrier gas. Total

carbohydrate was measured in using a phenol sulphuric assay (Dubois et al., 1956), and

chemical oxygen demand was determined colourimetrically using a Hach COD kit

6

7

(concentration range = 10 to 1500 mg L-1). Bicarbonate alkalinity (BA) in the methane

digesters was measured via titration to an endpoint of pH 5.75 using hydrochloric acid (0.1M).

3. Results and Discussion

3.1 Hydrogen production

Hydrogen was measured in the first stage of the two stage systems for a period of 17 days,

beginning when the digester had reached a steady operating state 30 days after start up. Over

this 17 day period hydrogen production rates remained between 2 cm3 min-1 and 4 cm3 min-1.

The mean hydrogen production rate was 3.611 cm3 H2 min-1 which equates to a yield of 6.7 L H2

kg-1 VS (Figure 2). This is comparable with hydrogen production rates obtained from other low

grade biomass sources in particular those obtained in this lab from enzymatically pretreated

sewage biosolids (18.4 L H2 kg-1 DS) and wheat co-product (7.0 L H2 kg-1 VS) (Massanet-

Nicolau et al., 2013, 2008). Other researchers studying hydrogen production from crop residues

report similar hydrogen yields (Guo et al., 2010). It should be noted however that the molar

yield achieved here (0.55 mol H2 mol-1 glucose) is only a small fraction of the 4 mol H2 mol-1 which

is theoretically possible (Fenchel and Finlay, 1995). This suggests that further optimisation of

the hydrogen producing digestion stage is possible, which based on findings in this and previous

studies, may further increase the synergistic effects of a two stage digestion system.

3.2 Methane and overall energy yields

Over the 17 day period of the experiment the two stage, 20 day HRT system consistently

outperformed the single stage 20 day HRT system (Figure 2). The mean methane production

rate from the single stage system was 12.6 cm3 min-1, 17.23 % lower than the 14.8 cm3 min-1

obtained from the two stage 20 day system. The two stage 12 day HRT system also produced

more methane than the single stage 23.6 cm3 min-1, however since the HRT of this system was

much shorter, its organic loading rate was higher (Figure 1) and so consideration of methane

production rate alone is misleading. For a fully normalised comparison of the performance of

7

8

the three digestion systems the yields of hydrogen and methane obtained from each system

along with their overall energy value must be compared. Methane yields from the single stage

20 day HRT system were 310.0 L kg-1 VS 12.7% lower than the 349.4 L kg-1 VS obtained from

the two stage 20 day HRT system. Yields obtained from the two stage 12 day HRT system were

slightly lower than the single stage 20 day HRT system (305.3 L kg-1 VS), however large cost

and energy savings result from operating a digester at a shorter HRT and these should not be

overlooked when evaluating these digestion systems.

Figure 2 shows the overall energy value of the biogas obtained for each digestion system based

on the lower heating value of the gases. In the case of the single stage system this energy was

derived solely from methane and amounted to 10.36 MJ kg-1 VS. In the case of the two stage

system energy was derived not only from the methane but from the hydrogen produced in the

first stage also. In the two stage 20 day HRT system 11.74 MJ kg-1 VS was obtained, 13.4%

greater than a single stage system. The two stage 12 day system produced 10.27 MJ kg-1 VS,

marginally less than the single stage system although such a system would require less energy

for heating and mixing.

3.3 Bicarbonate Alkalinity and Volatile Fatty Acid Production

As well as biogas and energy yields, the performance of the three digestion systems were

evaluated with respect to volatile fatty acid (VFA) production and bicarbonate alkalinity which

are key indicators of digester health and are among the first parameters to change in response to

perturbations in environmental conditions or feedstock composition which could disrupt

methanogenesis (Hawkes et al., 1994). Figure 3 shows that bicarbonate alkalinity levels in both

two stage digestion systems were consistently higher than in the single stage system. It is

notable that simply switching from a single stage to a two stage digestion system will allow

digestion at almost twice the organic loading rate without jeopardising the stability of the

methanogenic process as indicated by bicarbonate alkalinity.

8

9

The level of VFAs in the methanogenic digester is also a valuable indicator of process stability.

The effluent from the hydrogen digester will contain high levels of VFA since these are by-

products of fermentative hydrogen production. However similarly high levels in the effluent of

the methanogenic digester would indicate that the acetoclastic methanogenesis is not keeping

pace with VFA production which will eventually lead to a pH drop in the digester and inhibition

of methanogenesis. Figure 4 shows that in the methanogenic stage of each digestion system,

single or two stage, VFA levels were 1-2 orders of magnitude lower that those present in the

hydrogen producing stage. No large differences were observed in the total VFA concentration

of the two stage systems when compared to the single stage system; mean total VFA in the

methane digesters was between 89 mg L-1 and 107 mg L-1 indicating that although the second

stage of the two stage systems were fed with substrate containing high VFA concentrations, this

did not compromise their ability to reduce VFA. Again it should be noted that these data

indicate that two stage digestion permits higher OLRs to be used while still maintaining low

VFA concentrations in bioreactor effluent.

3.4 Chemical Characterisation

Total solids, volatile solids, chemical oxygen demand (COD) and carbohydrate were measured

regularly throughout the experiment. Both the feedstock and the effluent of each digestion

system were analysed (Table 1). These parameters can be used as indicators of substrate

available for conversion to hydrogen or methane and therefore give an indication of substrate

utilization by the microbial populations of each digestion system. For each parameter

investigated, comparison of feedstock and effluent show that differences in substrate utilization,

particularly between the single stage 20 day HRT system and the two stage 20 day HRT system

were small and not proportional to the increases in methane and hydrogen yields obtained

during digestion. This indicates that the increase in bioenergy obtained via two stage digestion

is not due solely to increased substrate turnover, but by a greater molar conversion of substrate

9

10

to methane. This result is similar to that obtained when experiments of this type were conducted

using wheat co-product (Massanet-Nicolau et al., 2013).

Conventionally digester performance has been strongly linked to solids destruction, however,

the fact that there is a disparity between substrate turnover and increases in methane production

in this and other studies (Grimberg et al., 2015) indicates that the situation may be more

complex. Increases in methane could occur because conditions in two stage systems may be less

conducive to microbes utilising catabolic pathways (such as lactate, solvent or propionate

production) which consume substrate but which do not produce methane or its direct precursors

(Acetate, hydrogen and carbon dioxide) (Kleerebezem and van Loosdrecht, 2007) . In the

acidogenic stage, heat treatment of the inoculum, low pH and short HRTs may inhibit such

pathways, whereas in the methanogenic stage, decreased ratios of hydrogen to acetate and less

easily available carbohydrate may select for methanogens as opposed to sulphur reducing

bacteria which compete with them (Chen et al., 2008). This hypothesis, though speculative,

would explain the increase in methane in the absence of increased substrate turnover and could

be examined by determining the abundance of microbial groups which compete with

methanogens for substrate. Molecular techniques are in the process of being adapted to monitor

population shifts in two stage digestion systems using complex non sterile feedstock and mixed

inoculum (Kinnunen et al., 2014). These could be used to provide additional evidence

supporting a more complex ecological impact of two stage digestion on methane production.

4. Conclusion

Two stage digestion was used to produce hydrogen and methane from pelletized grass. Two

stage digestion resulted in a greater than 13% increase in overall energy yields. Based on the

models produced by Toop (2013) if 25% of available UK grassland were used to produce

feedstock, switching to two stage digestion would yield an additional 17.9 billion MJ of energy

per annum when compared with a single stage system. Two stage digestion could also be

10

11

performed at a shorter HRT (12 days instead of 20 days) without impacting on yields or process

stability.

Acknowledgments

The authors would like to acknowledge financial support from the ERDF H2Wales project and

the EPRSC Hydrogen and Fuel Cell Supergen Hub, Grant Ref. EP/J016454/1.

References

1. Adler, P.R., Sanderson, M. a., Boateng, A. a., Weimer, P.J., Jung, H.J.G., 2006. Biomass yield and biofuel quality of switchgrass harvested in fall or spring. Agron. J. 98, 1518–1525.

2. Allison, G.G., Thain, S.C., Morris, P., Morris, C., Hawkins, S., Hauck, B., Barraclough, T., Yates, N., Shield, I., Bridgwater, A. V., Donnison, I.S., 2009. Quantification of hydroxycinnamic acids and lignin in perennial forage and energy grasses by Fourier-transform infrared spectroscopy and partial least squares regression. Bioresour. Technol. 100, 1252–1261.

3. Chen, Y., Cheng, J.J., Creamer, K.S., 2008. Inhibition of anaerobic digestion process: A review. Bioresour. Technol. 99, 4044–4064.

4. Chu, C.-F., Li, Y.-Y., Xu, K.-Q., Ebie, Y., Inamori, Y., Kong, H.-N., 2008. A pH- and temperature-phased two-stage process for hydrogen and methane production from food waste. Int. J. Hydrogen Energy 33, 4739–4746.

5. Clesceri, L.S., Greenberg, A.E., Eaton, A.D., 1999. Standard Methods for Examination of Water & Wastewater, 20th ed. American Public Health Association.

6. CROPGEN, 2007. Assessment of the potential for crop-derived biogas as an energy source in the EU , taking into account technical and environmental issues and socio-economic impact, Renewable energy from crops and agrowastes.

7. 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.

8. Dareioti, M.A., Kornaros, M., 2014. Effect of hydraulic retention time (HRT) on the anaerobic co-digestion of agro-industrial wastes in a two-stage CSTR system. Bioresour. Technol. 167, 407–415.

9. De Gioannis, G., Diaz, L.F., Muntoni, A., Pisanu, A., 2008. Two-phase anaerobic digestion within a solid waste/wastewater integrated management system. Waste Manag. 28, 1801–1808.

11

12

10. Dubois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A., Smith, F., 1956. Colorimetric method for determination of sugars and related substances. Anal. Chem. 28, 350–356.

11. Fenchel, T., Finlay, B.J., 1995. Ecology and evolution in anoxic worlds, Oxford series in ecology and evolution. Oxford University Press, Oxford ; New York.

12. Grimberg, S.J., Hilderbrandt, D., Kinnunen, M., Rogers, S., 2015. Bioresource Technology Anaerobic digestion of food waste through the operation of a mesophilic two-phase pilot scale digester – Assessment of variable loadings on system performance. Bioresour. Technol. 178, 226–229.

13. Guo, X.M., Trably, E., Latrille, E., Carrère, H., Steyer, J.-P., 2010. Hydrogen production from agricultural waste by dark fermentation: A review. Int. J. Hydrogen Energy 35, 10660–10673.

14. 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.

15. Hawkes, F.R., Guwy, A.J., Hawkes, D.L., Rozzi, A.G., 1994. On-line monitoring of anaerobic digestion: application of a device for continuous measurement of bicarbonate alkalinity. Water Sci. Technol. 30, 1–10.

16. Kinnunen, M., Hilderbrandt, D., Grimberg, S., Rogers, S., Mondal, S., 2014. Comparative study of methanogens in one- and two-stage anaerobic digester treating food waste. Renew. Agric. Food Syst. 1–9.

17. Kleerebezem, R., van Loosdrecht, M.C., 2007. Mixed culture biotechnology for bioenergy production. Curr. Opin. Biotechnol. 18, 207–212.

18. Kyazze, G., Dinsdale, R., Hawkes, F.R.R., Guwy, A.J.J., Premier, G.C.C., Donnison, I.S.S., 2008. Direct fermentation of fodder maize, chicory fructans and perennial ryegrass to hydrogen using mixed microflora. Bioresour. Technol. 99, 8833–8839.

19. Lee, Y.-W., Chung, J., 2010. Bioproduction of hydrogen from food waste by pilot-scale combined hydrogen/methane fermentation. Int. J. Hydrogen Energy 35, 11746–11755.

20. Massanet-Nicolau, J., Dinsdale, R., Guwy, A., 2008. Hydrogen production from sewage sludge using mixed microflora inoculum: Effect of pH and enzymatic pretreatment. Bioresour. Technol. 99, 6325–6331.

21. 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.

22. Murphy, J.D., Power, N.M., 2009. An argument for using biomethane generated from grass as a biofuel in Ireland. Biomass and Bioenergy 33, 504–512.

23. Nielsen, H.B., Mladenovska, Z., Westermann, P., Ahring, B.K., 2004. Comparison of two‐stage thermophilic (68°C/55°C) anaerobic digestion with one‐stage thermophilic (55°C) digestion of cattle manure. Biotechnol. Bioeng. 86, 291–300.

12

13

24. Tilman, D., Hill, J., Lehman, C., 2006. Carbon-negative biofuels from low-input high-diversity grassland biomass. Science 314, 1598–1600.

25. Toop, T.A., 2013. Sustainable grasslands for biological energy. Aberystwyth University.

26. Zuo, Z., Wu, S., Zhang, W., Dong, R., 2014. Performance of two-stage vegetable waste anaerobic digestion depending on varying recirculation rates. Bioresour. Technol. 162, 266–272.

13

14

Figure 1. Experimental design of parallel single and two stage digestion experiments

Figure 2. Overall hydrogen and methane yields obtained from each digestion system evaluated. Column labels denote total energy yields from the digestion system based on LHV of methane and or hydrogen.

14

15

Figure 3. Bicarbonate alkalinity of each digestion system.

Figure 4. VFA production at different stages in the digestion process (error bars denote standard deviation).

15

16

Table 1. Chemical characterisation of methanogenic stage bioreactor in each digestion system.

16