Microsoft Word - AlChE2014 Abstract-Zeng et al.docx

CONVERSION OF FURANIC AND PHENOLIC COMPOUNDS TO HYDROGEN GAS IN A MICROBIAL ELECTROLYSIS CELL

Xiaofei Zeng1, Abhijeet P. Borole2, Spyros G. Pavlostathis1

1Georgia Institute of Technology, Atlanta GA 30332-0512

2Oak Ridge National Laboratory, Oak Ridge TN 37831

Introduction

Biofuels produced via biomass pyrolysis have been widely viewed as a promising alternative to fossil fuels. The primary product generated from biomass pyrolysis, bio-oils, have an oil phase rich in precursors of transportation fuels and an aqueous phase composed of corrosive and inhibitory compounds. Any process that removes these chemicals will be highly beneficial for improving biofuel quality and production efficiency.
The production of biofuels requires the use of hydrogen gas in the hydrogenation process where bio-oils are upgraded to drop-in fuels. The current practice for hydrogen gas production, however, is based on reforming natural gas, which increases both fossil fuel consumption and greenhouse gas emissions. Therefore, a sustainable strategy to produce hydrogen gas for use in biofuel production would be a significant improvement over the existing methods.
The goal of this study is use microbial electrolysis cell (MEC) technology to convert model organic compounds, representative of bio-oil aqueous phase, to hydrogen gas. Bacteria
growing in the MEC anode metabolize the model compounds to protons and electrons. With a small voltage input (0.3-0.8 V), protons and electrons combine into hydrogen gas in the
cathode. MEC technology achieves three times higher hydrogen yield than dark fermentation, and consumes nine times less energy than water electrolysis (Lee et al., 2010). With the aim
to develop a waste-to-resource process, the research focuses on investigating the conversion efficiency of model compounds to hydrogen gas, and gaining an in-depth understanding of the biotransformation pathways and microbial activity in the anode.

Materials and Methods

Model Compounds

Furfural (FF), 5-hydroxymethylfurfural (5-HMF), syringic acid (SA), vanillic acid (VA) and
4-hydroxybenzoic acid (HBA) were selected as model compounds for this study. These compounds are found in the bio-oil aqueous phase derived from plant biomass pyrolysis. FF
and 5-HMF represent furanic compounds, while the other three compounds represent phenolic
acids. The selected model compounds are hydrophilic with log Kow values ranging from -0.45 to 1.58, water solubility greater than 1.2 g/L, and theoretical oxygen demand ranging from
1.45 to 1.6 g O2/g. All model compounds were purchased from Sigma Aldrich (St. Louis, MO).

MFC/MEC Construction

An H-type MEC reactor was constructed with two identical 250 mL square glass bottles equipped with sampling ports and two solution displacement burettes for gas collection. The
anode electrode was made of five 0.5 x 0.5 x 2 inch carbon felt strips attached to a stainless steel rod, while the cathode electrode was a 5 x 6 cm carbon cloth with 0.5 mg Pt/cm2 on the
side facing the proton exchange membrane (Nafion 117). A potentiostat (Gamry Instruments
Inc., Warminster, PA) was used to set the voltage and to conduct electrochemical analyses.
A single-chamber MFC with an air-cathode was set up for the enrichment of a microbial consortium in the bioanode. The MFC anode was the same as the MEC, while the cathode was an electrode-membrane assembly containing 0.5 mg Pt/cm2 on the carbon cloth. A resistor was connected between the anode and the cathode with a multi-meter installed across the two electrodes to record the voltage.

MFC Operation

The MFC served for the enrichment of a bioanode microbial community capable of metabolizing the model compounds, to be further used in setting up other MFCs and MECs. The inoculum for the MFC was an anode-developed microbial community fed with a diverse mixture of organic compounds (Borole et al., 2011). A mixture of the five model compounds (FF, 5-HMF, SA, VA and HBA, each at 62.5 mg COD/L) and glucose (214 mg COD/L) were used as the substrates. The total, initial substrate chemical oxygen demand (COD) was 526 mg/L and 200 mL minimal medium was used to supply nutrients and to buffer (Borole et al.,
2011). The MFC was maintained at room temperature (22-24oC) with a bioanode N2 headspace under a low positive pressure (ca. 0.05 atm). The bioanode liquid phase was continuously mixed with a magnetic stirrer. The MFC has been operated in three, distinct
phases. Phase I lasted for 35 days, during which the MFC was fed once a week and half of the medium (100 mL) was replaced before each feeding, except for the first two cycles where no
media was replaced. Phase II (day 35-49) was designed to accelerate microbial growth by reducing resistance from 500 to 250 and then to 100 ?. In Phase III, glucose was not used and the MFC feed included only the five model compounds. Other conditions were the same as
in Phase II and the resistance was 100 ?.

Analytical Methods

Over the course of each feeding cycle, the pH and the concentration of the five model compounds, glucose, and soluble COD were measured. The model compounds are quantified
using an Agilent HPLC unit equipped with a UV-Vis detector. A BioRad HPX-87H column was used with a mobile phase of 15% acetonitrile in 5 mM sulfuric acid and a flow rate of 0.6
mL/min (Borole et al., 2011). The wavelength was set at 280 nm for FFl and 5-HMF and 210 nm for the phenolic compounds. Biotransformation metabolites were named according to the peak retention time and UV wavelength (i.e., P24-210 refers to the peak detected at 24 min at
a UV wavelength of 210 nm) and their abundance was recorded as peak area of UV spectra. COD and pH was measured according to Standard Methods (Rice et al., 2012).

Results and Discussion

Model Compounds Conversion

When the MFC anode was fed with model compounds together with glucose, effective removal of the five model compounds within 24 hours was achieved. Two biotransformation
products, P18-210 and P24-210, accumulated and were not degraded for 6 days (Figure 1). After glucose was excluded from the bioanode feed, the microbial community retained the ability to completely remove the five model compounds within 24 hours, and showed an
enhanced conversion of the metabolites. The abundance of metabolites P18-210 and P24-210 at the end of three subsequent 7-day cycles without feeding glucose decreased from 3320 mUA*s to undetectable levels and from 3600 to 460 mUA*s, respectively (Figure 2). In contrast, the abundance of the two metabolites when the feed included glucose was around
3000 mUA*s and persisted over time. The decreased accumulation of metabolites suggests that the microbial consortium was developing the capacity to convert the metabolites. Correspondingly, the soluble COD removal in the MFC anode improved (Figure 3). Thus, more electrons were extracted from the substrates generating higher current and eventually hydrogen production in the case of a MEC.

Current Density and H2 Production

The MFC bioanode developed significant electroactivity within one week from
inoculation. The current production increased with reduced external resistance, from 0.1 mA/cm2 at 500 ? to 0.23 mA/cm2 at 100 ? (Figure 4). When higher current was produced, the current density declined over time within each feeding cycle as the number of electrons stored in the substrates was the same, but the time needed to transfer these electrons became shorter when a higher current was allowed. In the cycles without glucose, the decline of current within each feeding cycle became even more substantial, which is partially due to fewer available electrons when glucose was no longer fed. It may also be that the metabolites were less susceptible to biotransformation than the model compounds, which is supported by the fact that the current dropped upon complete removal of the model compounds. The coulombic efficiency of the MFC was 30-70% in the feeding cycles without glucose.
The maximum hydrogen gas production from MEC was calculated based on the current density of the MFC (Figure 5), assuming equal cathode efficiency between MFC and MEC. The cumulative hydrogen production from 316 mg COD/L model compounds ranged from 35 to
110 mL H2 at STP/L of bioanode. Further details and data on H2 production will be presented and discussed.

Conclusions

A microbial consortium developed in the MFC anode was capable of metabolizing five model compounds typically found in the bio-oil aqueous phase resulting from the pyrolysis of plant biomass. A soluble COD conversion of 30-50% was achieved with a coulombic efficiency of 30-70%. Among the non-biodegradable COD, two biotransformation products were detected by HPLC. Long-term enrichment of the anode microbial community is promising to achieve a higher extent of metabolites biodegradation, and thus a higher yield of H2 in a MEC.

References

1. Borole, A.P., Hamilton, C.Y. and Schell, D.J. (2012) Conversion of residual organics in corn stover-derived biorefinery stream to bioenergy via a microbial fuel cell. Environmental Science
& Technology 47(1), 642-648.
2. Lee, H.S., Vermaas, W.F.J. and Rittmann, B.E. (2010) Biological hydrogen production:
prospects and challenges. Trends in Biotechnology 28(5), 262-271.
3. Rice, E.W., Baird, R.B., Eaton, A.D. and Clesceri, L.S. (2012) Standard methods for the examination of water and wastewater, 22nd ed, APHA, AWWA, WEF, Washington DC, USA.

50

5-HMF

40 FF SA

VA

30 HBA

20

10

P18-210

P24-210

7000

6000

5000

4000

3000

2000

1000

400

300

200

100

Initial

Final

0 0 0

34 36 38 40 42 44

Time (d)

Figure 1. Time course of the five model compounds and two persistent metabolites during a MFC feeding cycle.

16 17 18

Cycle

Figure 3. Removal of soluble COD in three, glucose-free MFC feeding cycles.

4000

3000

P18-210

P24-210

0.30

0.25

0.20

Resistance (???

500

250

100

100

2000

0.15

1000

0.10

0.05

0

16 17 18

Cycle

0.00

0 10 20 30 40 50 80 90 100

Time (d)

Figure 2. Accumulation of metabolites in three, glucose-free MFC feeding cycles.
Figure 4. Current density of the MFC operated with different resistance, fed with the five model compounds and glucose (glucose excluded in the last four feeding cycles).

350

300

250

200

150

100

50

0

70 75 80 85 90 95 100 105

Time (d)

Figure 5. Calculated H2 production in four, glucose-free MFC feeding cycles.