Electrochemical evaluation of molybdenum ... - Semantic Scholar
Author's personal copy i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 9 4 3 9 e9 4 4 5
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
Electrochemical evaluation of molybdenum disulfide as a catalyst for hydrogen evolution in microbial electrolysis cells Justin C. Tokash, Bruce E. Logan* Department of Civil and Environmental Engineering, The Pennsylvania State University, University Park, PA 16802, USA
article info
abstract
Article history:
There is great interest in hydrogen evolution in bioelectrochemical systems, such as
Received 25 March 2011
microbial electrolysis cells (MECs), but these systems require non-optimal near-neutral pH
Received in revised form
conditions and the use of low-cost, non-precious metal catalysts. Here we show that
3 May 2011
molybdenum disulfide (MoS2) composite cathodes have electrochemical performance
Accepted 11 May 2011
superior to stainless steel (SS) (currently the most promising low-cost, non-precious metal
Available online 16 June 2011
MEC catalyst) or Pt-based cathodes in phosphate or perchlorate electrolytes, yet they cost w4.5 times less than Pt-based composite cathodes. At current densities typical of many
Keywords:
MECs (2e5 A/m2), the optimal surface density with MoS2 particles on carbon cloth was 25 g/
Hydrogen evolution reaction
m2, achieving 31 mV less hydrogen evolution overpotential than similarly constructed Pt
Molybdenum disulfide
cathodes in galvanostatic tests with a phosphate buffer. At higher current densities
Platinum
(8e10 A/m2) the MoS2 catalyst had 82 mV less hydrogen evolution overpotential than the
Stainless steel
Pt-based catalyst. MoS2 composite cathodes performed similarly to Pt cathodes in terms of current densities, hydrogen production rates and COD removal over several batch cycles in MEC reactors. These results show that MoS2 can be used to substantially reduce the cost of cathodes used in MECs for hydrogen gas production. Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
1.
Introduction
Due to the steady increase in global demand for energy, there has been increased interest in using hydrogen as an energy carrier. One very promising technology is the use of a microbial electrolysis cell (MEC) to generate hydrogen gas [1,2]. In an MEC, bacteria attached to the anode oxidize organic matter in the solution generating protons and electrons that are combined electrocatalytically at the cathode to form hydrogen gas. Alternatives are needed to Pt and other noble metals used as catalysts for the electrochemical hydrogen evolution reaction (HER). The success of platinum for HER is primarily due to
its low free energy difference in the initial Volmer step where protons are bonded to catalytic sites of the cathode ðHþ þ e 4Hads Þ, but also the Tafel ð2Hads 4H2 Þ and/or Heyrovsky ðHads þ Hþ þ e 4H2 Þ steps that release molecular hydrogen [3]. An ideal catalyst exhibits zero change in free energy between the reactant and product states. However, many metals are not suitable HER catalysts because they either form very strong metalehydrogen bonds or, for example in the case of gold, form very weak metalehydrogen bonds, which thermodynamically hinders the HER [4]. Platinum is not only expensive and rare, but it can be poisoned by adsorption of chemicals that shift the cathodic
* Corresponding author. Tel.: þ1 814 863 7908; fax: þ1 814 863 7304. E-mail address: [email protected] (B.E. Logan). 0360-3199/$ e see front matter Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2011.05.080
Author's personal copy 9440
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 9 4 3 9 e9 4 4 5
potential to more negative values, resulting in decreased HER. This behaviour is clearly shown in the recent work of De Silva Mun˜oz et al. [5] where Pt was compared to stainless steel as a HER catalyst. While Pt out performed SS in the initial scan using linear sweep voltammetry (LSV), successive scans showed decreased performance of Pt over time due to surface adsorption of phosphate anions. In contrast, stainless steel maintained the same LSV response over repeated scans. For these reasons, SS is becoming a preferred catalyst to Pt for HER in MECs. While SS is durable, easy to fabricate into electrodes, and inexpensive, it has a high overpotential compared to Pt and therefore alternatives are needed to improve energy recovery and performance. Two promising organic compounds that can catalyze HER are hydrogenase and nitrogenase enzymes. Recent studies of these enzymes [4,6], including the use of quantum chemical simulations [4], suggested that molybdenum disulfide (MoS2) would have a small free energy difference for HER. Although MoS2 has been used as a photocatalyst for hydrogen evolution [7e9], it has not been well explored in electrochemical studies for HER. Based on the laminar structure of MoS2, it is thought that hydrogen evolution is catalyzed at edge sites of the layered MoS2 based on its graphite-like structure [4,10e13]. Very little is known about the electrochemical behaviour of MoS2 as a hydrogen evolution catalyst [9,11,14]. Therefore, we examined MoS2 particles for HER in two different electrolytes (sodium perchlorate and a phosphate buffer) to study the effects of solution chemistry, and in two different electrode configurations in order to examine how to best distribute particles to expose catalytically active edge sites. In the first electrode configuration MoS2 particles were deposited onto stainless steel, and compared for HER performance relative to bare stainless steel. For the second electrode configuration we mixed MoS2 particles with carbon black and bound them to carbon cloth using a conductive polymer binder (Nafion) (the same approach used with Pt catalysts for HER in fuel cells). We varied the mass loading ratio between MoS2 and carbon black, which affects the surface density of the catalyst, in order to determine conditions for optimal performance. Once the optimal loading ratio was found for the MoS2 composite cathodes, MEC reactors were operated in fed-batch mode to compare Pt-based composite cathodes to ones made with MoS2 catalyst.
2.
Material and methods
2.1.
Electrode construction
Flat sheets of type 304 stainless steel (Trinity Brand Industries, Inc.) were sanded smooth with silicon carbide sand paper down to a 1200 grit finish, and then polished with successively decreased sizes of alumina particle slurries (to 1 mm). Between each successive step standard metallographic procedures were followed, whereby samples were ultrasonically washed in distilled/deionized water and rinsed with acetone with a final rinse in distilled/deionized water and drying under a stream of dry nitrogen. Similarly prepared SS samples were burnished (by rubbing the powder onto the SS) with MoS2 particles (Aldrich, 99%, particle size < 2 mm) until an even
coating was obtained on the surface. The amount of MoS2 coating on SS was 7.86 104 mol/m2 (1.109 g/m2), equivalent to w1000 monolayers. MoS2 catalyst powder was mixed with 5 mg/cm2 carbon black (Cabot, VULCAN XC-72R, >99%) at mass loading ratios of 0e50% (Table 1) and 50 mL/cm2 of a 2:1 volume solution of Nafion polymer (Aldrich, 5 wt.%) and iso-propanol, vortexed for w15 s at 3200 rpm (VWR Vortex Mixer), and applied to only the solution side of the carbon cloth (E-TEK, B-1/B/30WP, 30% by weight PTFE Wet-Proofed) using a small paint brush. The electrodes were then air-dried overnight (>12 h) before testing. All MoS2 composite cathodes had the same amount of carbon black and Nafion binder so that the only variable was the amount of MoS2. Carbon cloth cathodes with a Pt catalyst were prepared using a commercially available platinum powder pre-mixed with carbon black (E-TEK, C1-10, 10 wt.% Pt on Vulcan XC-72). The platinum surface density was 5.56 g/m2, which maintained the same amount of carbon black as used in the MoS2 composite cathodes. The Pt-based composite cathodes also used the same quantity of Nafion binder as in the case of the MoS2 composite cathodes. The construction of the platinum composite cathodes has been described previously [15]. All cathodes had a geometric surface area of 12 cm2, but only 7 cm2 was exposed to the solution due to the design of the electrochemical cell.
2.2.
Electrochemical characterization
Cathodes were used as the working electrode in a threeelectrode, two-chamber electrochemical cell. Two different
Table 1 e Composition of cathodes by loading ratio (wt.% catalyst/total mass) and by catalyst surface density, and cathode overpotentials (h) based on LSV and galvanostatic polarization (GStat) data at L7.14 A/m2 in 0.1 M NaClO4 or 50 mM phosphate buffer solution (PBS). For comparison purposes the overpotential was set to zero for the 10 wt.% Pt cathode. Positive values indicate a greater (more cathodic) overpotential and negative values indicate lower overpotential. Sample ID
0 MoS2 5 MoS2 10 MoS2 15 MoS2 20 MoS2 25 MoS2 30 MoS2 35 MoS2 40 MoS2 45 MoS2 50 MoS2 10 wt% Pt Bare SS MoS2 on SS
h (V) in NaClO4
h (V) in PBS
Catalyst loading ratio (%)
Catalyst surface density (g/m2)
0.00 9.09 16.7 23.1 28.6 33.3 37.5 41.2 44.4 47.4 50.0 10.0
0.00 5.00 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 5.56
0.426 0.289 0.286 0.285 0.278 0.252 0.264 0.245 0.247 0.220 0.241 0.000
0.496 0.371 0.356 0.290 0.282 0.298 0.286 0.270 0.235 0.277 0.259 0.000
0.064 0.042 0.002 0.010 0.011 0.023 0.010 0.008 0.028 0.029 0.030 0.000
0.172 0.007 0.125 0.049 0.176 0.083 0.150 0.079 0.095 0.084 0.233 0.000
N/A N/A
N/A 1.11
0.836 0.287 0.333 0.346
1.041 0.105
1.442 0.165
LSV GStat LSV GStat
Author's personal copy i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 9 4 3 9 e9 4 4 5
aqueous electrolytes were tested: 0.1 M NaClO4 (Alfa Aesar, 98e102%, ACS Grade) at pH 5.8; and 50 mM phosphate buffer (PBS, Na2HPO4, 4.58 g/L; and NaH2PO4$H2O, 2.45 g/L) at pH 7. The perchlorate solution was used to avoid possible complicating effects of buffer adsorption (for example phosphate) onto the surfaces [5], and a phosphate buffer solution was used because it is a common electrolyte for MEC studies [1,16e20]. Solutions were degassed by nitrogen sparging for 15 min prior to being added to both cathode and anode chambers before the start of each experiment. The cathode and anode chambers were separated by a cation exchange membrane (SELEMION, CMV, Asahi Glass Co., Ltd., Japan) to limit interference by competing and/or conjugate reactions taking place at each electrode. A platinum flag was used as the counter electrode in the anode chamber with an Ag/AgCl (3 M NaCl) reference electrode (Bioanalytical Systems, Inc., RE-5B) in the cathode chamber. All recorded potentials are corrected for pH and are reported here vs. the RHE scale using E(RHE) ¼ E(Ag/AgCl) þ 0.211 V þ 0.0591 V/pH, where E(Ag/Ag/ Cl) is the reference electrode potential, 0.211 V the shift from Ag/AgCl (3 M NaCl) scale to SHE scale, and 0.0591 V/pH is the shift from SHE to RHE scale. Cathodes were characterized using two electrochemical techniques using a multichannel potentiostat (Solartron 1470E): linear sweep voltammetry (LSV); and galvanostatic polarization. For LSV tests, the cathodes were swept at 1 mV/s from their open circuit potential down to 1.189 V or until a total cathodic current density of 15 A/m2 (NaClO4) or 10 A/m2 (PBS) was reached. Current densities were normalized by the exposed geometric surface area of the electrode (7 cm2). Galvanostatic tests were also conducted to characterize the electrodes under steady state conditions. Current was held for 15 min at 1, 2, 3, 5, 8, and 10 mA for the NaClO4 electrolyte or in the range from 1 to 7 mA in 1 mA steps for the PBS electrolyte, and the steady state potential was plotted as a function of current density. Even though the applied current density is the independent variable and the potential is the dependent variable, the potential was plotted on the x-axis for ease of comparison of results with LSV tests. All electrochemical tests were performed three times to ensure consistency in the results, and all results given here were derived from the third experimental run for each test.
2.3.
9441
In order to compare the performance of MoS2 composite electrodes to standard platinum-based composite cathodes, we operated MECs first with the platinum composite cathodes, and then with one of the best performing MoS2 composite cathodes. Platinum composite cathodes were used for six batch cycles in order to obtain a performance baseline, then they were replaced with MoS2 composite cathodes for three additional fed-batch cycles. Prior to each batch cycle, the MEC reactor headspace was flushed with nitrogen gas for several minutes. During each fed-batch cycle the volume of gas produced was recorded using a respirometer (AER-200; Challenge Environmental) and the evolved gas was collected in air-tight gas bags (0.05 L capacity; Cali-5-Bond, Calibrated Instruments, Inc.) for post-cycle gas composition analysis using gas chromatography (Model 310; SRI Instruments). Hydrogen production yield and energy recovery were used to calculate the electrical, substrate, and overall energy efficiencies as previously described [16].
3.
Results and discussion
3.1.
LSV results in NaClO4
Based on LSV scans in a NaClO4 solution, the optimum MoS2 loading was 54 mg MoS2 and 60 mg carbon black, or 47 wt.% MoS2, with a surface density of 45 g/m2 (Fig. 1 and Table 1). This optimum catalyst loading had 0.616 V less HER
Microbial electrolysis cell (MEC) tests
MECs (quadruplicate reactors) were operated in fed-batch mode using single-chamber reactors [16] containing 30 mL of solution with bioanodes that had been well-established in previous MEC tests (>60 batch cycles). The solution consisted of 50 mM phosphate buffer (pH 7) as well as trace vitamins and minerals [21] and 1 g/L sodium acetate (anhydrous, SigmaeAldrich) as the organic substrate. Electrode potentials were measured with Ag/AgCl reference electrodes. The MECs were operated by potentiostatically setting the bioanode voltage to 0.276 V vs. Ag/AgCl (0.349 V vs. RHE), which was chosen based on prior results that indicated this value as the potential that yields the maximum current for the oxidation of acetate anion. The cathodes were the counter electrodes and their potentials were recorded using a PC controlled multimeter (Model 2700; Keithley Instruments, Inc.).
Fig. 1 e LSV scans of select cathodes in 0.1 M NaClO4. Solid lines: (A, red) bare stainless steel; (B, blue) MoS2 coated stainless steel; and (C, black) 10% platinum on carbon cloth. The dashed lines correspond to various loading ratios of MoS2 added to carbon black and applied to carbon cloth: 0 wt.% MoS2 (1, red); 9 wt.% MoS2 (2, orange); 23 wt.% MoS2 (3, green); 33 wt.% MoS2 (4, blue); and 47 wt.% MoS2 (5, violet). The arrow indicates increasing MoS2 loading ratio for the dashed lines. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Author's personal copy 9442
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 9 4 3 9 e9 4 4 5
overpotential than SS, but 0.220 V more overpotential than Pt on carbon cloth (Table 1). A total of 11 different MoS2 electrodes were examined, although many of these curves were omitted from the Figures for clarity as all followed the same general trend with mass loading as those included in Fig. 1. Once the mass loading ratio reached 47% there was no further improvement found by increasing the MoS2 to carbon black mass ratio. The SS electrode burnished with MoS2 had a greater catalytic activity than uncoated SS electrode (Fig. 1). The layer of MoS2 on SS enhanced hydrogen evolution as evidenced by a reduction of HER overpotential of 0.503 V compared to bare stainless steel. Based on structural information [4,10,11] that MoS2 has a laminar structure and that the electrochemically active sites occur at the edges of the crystal planes, it appears that a burnished MoS2 coating resulted in exposure of the active edge sites of MoS2, resulting in higher activity than bare SS. The overpotential for bare SS was much higher than that of Pt on carbon cloth in LSV tests, as expected from previous tests [13]. For example, the overpotential for SS determined by LSV was 0.836 V larger than Pt on carbon cloth.
3.2.
Galvanostatic results in NaClO4
Even though the sweep rate used in the LSV scans was relatively low (1 mV/s), which is generally accepted as slow enough to approximate steady state conditions, we performed additional constant current tests to better approximate the performance of these materials under steady state conditions. Under these truly steady state conditions any time-dependent response, for example capacitive surface charging/discharging, is eliminated over the long period of time used to perform the constant current tests. The cathode made with 48 mg MoS2 and 60 mg carbon black on carbon cloth exhibited the lowest overpotential compared to platinum on carbon cloth, which consisted of 44 wt.% MoS2, with a 40 g/m2 MoS2 surface density (Fig. 2 and Table 1). This electrode required 0.050e0.060 V less overpotential than bare SS for high current densities (7 to 14 A/m2), and at less negative current densities (less than 7 A/m2) the MoS2 catalyst cathodes substantially out performed bare SS. For example, at 4 A/m2 the MoS2 cathode had 0.116 V less overpotential than SS, at 3 A/m2 the HER overpotential was 0.386 V less than bare SS (which out performed 10% platinum by 0.134 V), and at 1 A/m2 the MoS2 cathode had 0.896 V less overpotential than SS for the HER (better than platinum by 0.646 V).
3.3.
LSV tests in phosphate buffer
The optimum MoS2 loading in the phosphate buffer solution based on LSV scans was 47 wt.% MoS2, in agreement with results obtained with LSV scans and a NaClO4 solution (Fig. 3 and Table 1). This cathode had 1.075 V less overpotential than bare stainless steel and 0.034 V less overpotential than a 10 wt.% Pt on carbon cloth cathode (Table 1). LSV scans (Fig. 3) suggest that a 33 wt.% loading (25 g/m2) performs better than the 47 wt.% loading at higher current densities, even though cathodes with these two loadings have nearly the same HER overpotential.
Fig. 2 e Galvanostatic polarization data for select cathodes in 0.1 M NaClO4. Solid lines: bare stainless steel (no symbol, red); MoS2 coated stainless steel ( , blue); and 10 wt.% Pt on carbon cloth (A, black). The dashed lines correspond to various loading ratios of MoS2 added to carbon black and applied to carbon cloth: 0 wt.% MoS2 ( , red); 9 wt.% MoS2 ( , orange); 23 wt.% MoS2 ( , green); 33 wt.% MoS2 ( , blue); and 44 wt.% MoS2 ( , violet). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
As the MoS2 to carbon black mass ratio increased from 0% to 50%, initially there was a steady decrease in HER overpotential up to about 17 wt.% MoS2, but beyond this loading ratio the HER overpotential did not scale directly with the amount of MoS2. For example, the HER overpotential as measured by LSV for the 50 wt.% loading was nearly the same as the cathode made with 9 wt.% MoS2. One possible reason for the decreased performance may have been too many MoS2 particles present, resulting in the MoS2 particles on the outermost surface blocking otherwise available edge sites of the underlying MoS2 particles. Another possible reason that the HER overpotential did not scale with the amount of catalyst in this case could be due to phosphate poisoning, as observed with platinum-based cathodes [5]. The SS cathode coated with MoS2 had a greater catalytic activity than the bare SS electrode (Fig. 3). The MoS2 coating improved the cathodic current density significantly at applied potentials below about 0.3 V vs. RHE. The bare SS cathode produced a nearly zero cathodic current density in the range examined here (Line “A”, Fig. 3), but with the addition of a MoS2 coating we observed a current density of nearly 2.5 A/m2 at about 1 V vs. RHE (Line “B”, Fig. 3). Cathodes constructed by coating stainless steel with MoS2 particles offer a sizeable improvement in HER performance (Table 1). The bare SS cathode had a HER overpotential using LSV of 1.041 V beyond that of the platinum-based cathode. The MoS2 coated SS cathode had a HER overpotential of only 0.105 V beyond platinum on carbon cloth. The overpotential for bare SS was much higher than that of platinum on carbon cloth in LSV tests (Table 1), consistent with previous tests [17].
Author's personal copy i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 9 4 3 9 e9 4 4 5
Fig. 3 e LSV scans of select cathodes in 50 mM PBS. Solid lines: bare stainless steel (A, red); MoS2 coated stainless steel (B, blue); and 10% platinum on carbon cloth (C, black). The dashed lines correspond to various loading ratios of MoS2 added to carbon black and applied to carbon cloth: 0 wt.% MoS2 (1, red); 9 wt.% MoS2 (2, orange); 23 wt.% MoS2 (3, green); 33 wt.% MoS2 (4, blue); and 47 wt.% MoS2 (5, violet). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
9443
Fig. 4 e Galvanostatic polarization data for select cathodes in 50 mM PBS. Solid lines: bare stainless steel (no symbol, red); MoS2 coated stainless steel ( , blue); and 10 wt.% Pt on carbon cloth (A, black). The dashed lines correspond to various loading ratios of MoS2 added to carbon black and applied to carbon cloth: 0 wt.% MoS2 ( , red); 9 wt.% MoS2 ( , orange); 23 wt.% MoS2 ( , green); 33 wt.% MoS2 ( , blue); and 47 wt.% MoS2 ( , violet). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Galvanostatic tests in phosphate buffer showed that the cathode made with 47 wt.% MoS2 in carbon black on carbon cloth (45 g/m2 MoS2 surface density) produced the lowest HER overpotential, which was only slightly better than cathodes made with 33, 41, and 23 wt.% MoS2 (Table 1). These four composite MoS2 cathodes out performed the platinum on carbon cloth cathodes (Fig. 4 and Table 1). In addition, these four electrodes had nearly 1.5 V less overpotential than bare SS. The MoS2 coated SS cathode also showed a significant improvement in performance compared to bare SS in galvanostatic tests. The HER overpotential at high current density was reduced by over 1.2 V simply by burnishing MoS2 catalyst onto SS (Table 1). Even though the performance of MoS2 coated SS was not as good as the composite MoS2 cathodes, these results confirm what we observed in the LSV tests, that MoS2 was an effective low-cost HER catalyst.
these cathodes in the MECs was 10.0 1.6 A/m2 for Pt and 10.7 1.2 A/m2 for MoS2, and the average fed-batch cycle time was 14.7 0.4 h (Fig. 5). Gas compositions were H2 ¼ 90 6%, CH4 ¼ 1 0.6%, and CO2 ¼ 8.7 5.4% for the Pt cathodes, and H2 ¼ 92 3%, CH4 ¼ 1 0.8%, and CO2 ¼ 6.7 2.3% for the MoS2 cathodes (Fig. 6). COD removals were 88 2% for MECs with Ptbased cathodes and 90 2% for the MECs with MoS2 based cathodes (Fig. 7). As a result of the similar COD removals and gas composition, the Coulombic and substrate efficiencies are very similar for the two types of cathodes. The main difference in results obtained with the two types of cathodes was that the electrical energy efficiency with the platinum-based cathodes was slightly better (158 14%) than that with the MoS2 cathodes (123 7%) (Fig. 7). The overall efficiency, which takes into account the electrical and substrate efficiencies, was very similar for the two catalyst materials, with 61 5% for the platinum-based cathodes and 57 3% for the MoS2 cathodes.
3.5.
3.6.
3.4.
Galvanostatic tests in phosphate buffer
MEC tests
Based on LSV and galvanostatic polarization tests in two electrolytes, and after consideration of the costs of electrode materials (discussed below), the 33 wt.% MoS2 composite cathode was chosen for additional tests in MECs. The performance of the MECs in terms of current densities, gas composition, COD removal and Coulombic, electrical, substrate, and combined efficiencies were not appreciably different for the MoS2 and Pt catalyst cathodes. The average current density for
Consideration of material costs and performance
A rough comparison was made of the costs of the different types of cathodes, based on purchased prices, to quantify possible savings using the different materials. Bare or MoS2 coated stainless steel cathodes have the lowest cost (approximately $57/m2 for either one). Composite MoS2 cathodes are about an order of magnitude more expensive (about $600/m2) and 10 wt.% platinum composite cathodes are about five times more expensive (about $2700/m2) than similarly constructed
Author's personal copy 9444
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 9 4 3 9 e9 4 4 5
Fig. 5 e MEC reactor current density vs. time. The first six batch cycles correspond to MECs with 10 wt.% Pt cathodes and the final three batch cycles correspond to 33 wt.% MoS2 cathodes. MEC reactors #1 through #4 correspond to black, red, green, and blue lines, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
composite MoS2 cathodes. The composite MoS2 cathodes are less expensive than the platinum-based cathodes simply due to the much lower catalysts price because they use the same carbon cloth and polymer binder. For example, the 10 wt.%
Fig. 7 e MEC efficiencies for each reactor: Coulombic efficiency (-, black); electrical efficiency ( , blue); substrate efficiency ( , red); and overall efficiency ( , green). Results for Pt were averaged over six cycles and results for MoS2 were averaged over three cycles. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
platinum in carbon black catalyst costs $2135/m2 at present prices, whereas the MoS2 in carbon black costs about $11/m2. In order to reduce the price of the composite MoS2 cathodes even further, alternatives to the carbon cloth could be used as this material contributes 80% of the cathode cost for MoS2 composite cathodes. One such material is a carbon mesh material (3K, Plain Weave Carbon Fibre Fabric, Fibre Glast Development Corp.), which only costs about $43/m2 as opposed to the carbon cloth, which costs about $470/m2 at present prices. The binder used here (Nafion) was not critical to the performance of the MoS2. We also tested two other conductive polymers (poly[bisphenol A-co-epicholorohydrin], BAEH; and poly[methyl vinyl ether-alt-maleic acid monobutyl ester], PMVEMABE) (data not shown). Both of these other polymers had a similar performance to Nafion, but offer the possibility of using a binder at a much reduced cost.
4. Fig. 6 e MEC gas composition (stacked bars, left axis) and percent COD removal (black dots, right axis). Gas composition was: H2 (blue, bottom), CH4 (black, middle), and CO2 (red, top). Separate MEC reactor results for gas composition and percent COD removal are shown as averages over six batch cycles for those using 10 wt.% Pt cathodes and over three batch cycles for MoS2 cathodes. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Conclusions
Abiotic tests using either LSV or galvanostatic polarization showed that MoS2 particles dispersed into carbon black and applied to carbon cloth performed much better than bare SS as an HER catalyst. Electrochemical tests also showed that this material could even surpass the performance of platinumbased electrodes prepared in the same manner. The overall best choice of MoS2 to carbon black loading ratio was found to be 47 wt.% based only on overpotentials. However, considering the small overpotential differences for catalysts at different loadings, and the need to reduce materials costs, the
Author's personal copy i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 9 4 3 9 e9 4 4 5
33 wt.% MoS2 cathode is the best choice. MEC tests using this optimum loading ratio showed that MoS2 cathodes performed similarly to platinum cathodes in terms of current densities, gas production, COD removal, and energy efficiency. Thus, MoS2 cathodes prepared as described here provide a new, lowcost material that has advantages of ease of fabrication and durability similar to platinum but at a much lower price.
[9]
[10]
[11]
Acknowledgements We thank Air Products and Chemicals, Inc. for the financial support of this project and appreciate the useful discussions with Dr. Matthew D. Merrill, Dr. Tomonori Saito, and Dr. Younggy Kim.
[12]
[13]
references [14] [1] Liu H, Grot S, Logan BE. Electrochemically assisted microbial production of hydrogen from acetate. Environ Sci Technol 2005;39:4317e20. [2] Cheng S, Logan BE. Sustainable and efficient biohydrogen production via electrohydrogenesis. Proc Natl Acad Sci 2007; 104:18871e3. [3] Schmickler W. Interfacial electrochemistry. New York: Oxford University Press; 1996. [4] Hinnemann B, Moses PG, Bonde J, Jørgensen KP, Nielsen JH, Horch S, et al. Biomimetic hydrogen evolution: MoS2 nanoparticles as catalyst for hydrogen evolution. J Am Chem Soc 2005;127:5308e9. [5] De Silva Mun˜oz L, Bergel A, Fe´ron D, Basse´guy R. Hydrogen production by electrolysis of a phosphate solution on a stainless steel cathode. Int J Hydrogen Energy 2010;35:8561e8. [6] Koper MTM, Bouwman E. Electrochemical hydrogen production: bridging homogeneous and heterogeneous catalysis. Angew Chem Int Ed 2010;49:3723e5. [7] Sobczynski A. Molybdenum disulfide as a hydrogen evolution catalyst for water photodecomposition on semiconductors. J Catal 1991;131:156e66. [8] Zong X, Yan H, Wu G, Ma G, Wen F, Wang L, et al. Enhancement of photocatalytic H2 evolution on CdS by
[15]
[16]
[17]
[18]
[19]
[20]
[21]
9445
loading MoS2 as cocatalyst under visible light irradiation. J Am Chem Soc 2008;130:7176e7. Zong X, Wu G, Yan H, Ma G, Shi J, Wen F, et al. Photocatalytic H2 evolution on MoS2/CdS catalysts under visible light irradiation. J Phys Chem C 1963e1968;2010:114. Elizondo-Villarreal N, Vela´zquez-Castillo R, Galva´n DH, Camacho A, Yacama´n MJ. Structure and catalytic properties of molybdenum sulfide nanoplatelets. Appl Catal A 2007;328: 88e97. Jaramillo TF, Jørgensen KP, Bonde J, Nielsen JH, Horch S, Chorkendorff I. Identification of active edge sites for electrochemical H2 evolution from MoS2 nanocatalysts. Science 6 July 2007;317:100e2. Brorson M, Carlsson A, Topsøe H. The morphology of MoS2, WS2, CoeMoeS, NieMoeS and NieWeS nanoclusters in hydrodesulfurization catalysts revealed by HAADF-STEM. Catal Today 2007;123:31e6. Kisielowski C, Ramasse QM, Hansen LP, Brorson M, Carlsson A, Molenbroek AM, et al. Imaging MoS2 nanocatalysts with single-atom sensitivity. Angew Chem Int Ed 2010;49:2708e10. Bonde J, Moses PG, Jaramillo TF, Nørskov JK, Chorkendorff I. Hydrogen evolution on nano-particulate transition metal sulfides. Faraday Discuss 2008;140:219e31. Cheng S, Liu H, Logan BE. Increased performance of singlechamber microbial fuel cells using an improved cathode structure. Electrochem Commun 2006;8:489e94. Call D, Logan BE. Hydrogen production in a single chamber microbial electrolysis cell lacking a membrane. Environ Sci Technol 2008;42:3401e6. Zhang Y, Merrill MD, Logan BE. The use and optimization of stainless steel mesh cathodes in microbial electrolysis cells. Int J Hydrogen Energy 2010;35:12020e8. Bond DR, Lovley DR. Electricity production by Geobacter sulfurreducens attached to electrodes. Appl Environ Microbiol 2003;69:1548e55. Hu H, Fan Y, Liu H. Hydrogen production using singlechamber membrane-free microbial electrolysis cells. Water Res 2008;42:4172e8. Sleutels THJA, Hamelers HVM, Rozendal RA, Buisman CJN. Ion transport resistance in microbial electrolysis cells with anion and cation exchange membranes. Int J Hydrogen Energy 2009;34:3612e20. Balch WE, Fox GE, Magrum LJ, Woese CR, Wolfe RS. Methanogens: reevaluation of a unique biological group. Microbiol Rev 1979;43:260e96.
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
Electrochemical evaluation of molybdenum disulfide as a catalyst for hydrogen evolution in microbial electrolysis cells Justin C. Tokash, Bruce E. Logan* Department of Civil and Environmental Engineering, The Pennsylvania State University, University Park, PA 16802, USA
article info
abstract
Article history:
There is great interest in hydrogen evolution in bioelectrochemical systems, such as
Received 25 March 2011
microbial electrolysis cells (MECs), but these systems require non-optimal near-neutral pH
Received in revised form
conditions and the use of low-cost, non-precious metal catalysts. Here we show that
3 May 2011
molybdenum disulfide (MoS2) composite cathodes have electrochemical performance
Accepted 11 May 2011
superior to stainless steel (SS) (currently the most promising low-cost, non-precious metal
Available online 16 June 2011
MEC catalyst) or Pt-based cathodes in phosphate or perchlorate electrolytes, yet they cost w4.5 times less than Pt-based composite cathodes. At current densities typical of many
Keywords:
MECs (2e5 A/m2), the optimal surface density with MoS2 particles on carbon cloth was 25 g/
Hydrogen evolution reaction
m2, achieving 31 mV less hydrogen evolution overpotential than similarly constructed Pt
Molybdenum disulfide
cathodes in galvanostatic tests with a phosphate buffer. At higher current densities
Platinum
(8e10 A/m2) the MoS2 catalyst had 82 mV less hydrogen evolution overpotential than the
Stainless steel
Pt-based catalyst. MoS2 composite cathodes performed similarly to Pt cathodes in terms of current densities, hydrogen production rates and COD removal over several batch cycles in MEC reactors. These results show that MoS2 can be used to substantially reduce the cost of cathodes used in MECs for hydrogen gas production. Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
1.
Introduction
Due to the steady increase in global demand for energy, there has been increased interest in using hydrogen as an energy carrier. One very promising technology is the use of a microbial electrolysis cell (MEC) to generate hydrogen gas [1,2]. In an MEC, bacteria attached to the anode oxidize organic matter in the solution generating protons and electrons that are combined electrocatalytically at the cathode to form hydrogen gas. Alternatives are needed to Pt and other noble metals used as catalysts for the electrochemical hydrogen evolution reaction (HER). The success of platinum for HER is primarily due to
its low free energy difference in the initial Volmer step where protons are bonded to catalytic sites of the cathode ðHþ þ e 4Hads Þ, but also the Tafel ð2Hads 4H2 Þ and/or Heyrovsky ðHads þ Hþ þ e 4H2 Þ steps that release molecular hydrogen [3]. An ideal catalyst exhibits zero change in free energy between the reactant and product states. However, many metals are not suitable HER catalysts because they either form very strong metalehydrogen bonds or, for example in the case of gold, form very weak metalehydrogen bonds, which thermodynamically hinders the HER [4]. Platinum is not only expensive and rare, but it can be poisoned by adsorption of chemicals that shift the cathodic
* Corresponding author. Tel.: þ1 814 863 7908; fax: þ1 814 863 7304. E-mail address: [email protected] (B.E. Logan). 0360-3199/$ e see front matter Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2011.05.080
Author's personal copy 9440
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 9 4 3 9 e9 4 4 5
potential to more negative values, resulting in decreased HER. This behaviour is clearly shown in the recent work of De Silva Mun˜oz et al. [5] where Pt was compared to stainless steel as a HER catalyst. While Pt out performed SS in the initial scan using linear sweep voltammetry (LSV), successive scans showed decreased performance of Pt over time due to surface adsorption of phosphate anions. In contrast, stainless steel maintained the same LSV response over repeated scans. For these reasons, SS is becoming a preferred catalyst to Pt for HER in MECs. While SS is durable, easy to fabricate into electrodes, and inexpensive, it has a high overpotential compared to Pt and therefore alternatives are needed to improve energy recovery and performance. Two promising organic compounds that can catalyze HER are hydrogenase and nitrogenase enzymes. Recent studies of these enzymes [4,6], including the use of quantum chemical simulations [4], suggested that molybdenum disulfide (MoS2) would have a small free energy difference for HER. Although MoS2 has been used as a photocatalyst for hydrogen evolution [7e9], it has not been well explored in electrochemical studies for HER. Based on the laminar structure of MoS2, it is thought that hydrogen evolution is catalyzed at edge sites of the layered MoS2 based on its graphite-like structure [4,10e13]. Very little is known about the electrochemical behaviour of MoS2 as a hydrogen evolution catalyst [9,11,14]. Therefore, we examined MoS2 particles for HER in two different electrolytes (sodium perchlorate and a phosphate buffer) to study the effects of solution chemistry, and in two different electrode configurations in order to examine how to best distribute particles to expose catalytically active edge sites. In the first electrode configuration MoS2 particles were deposited onto stainless steel, and compared for HER performance relative to bare stainless steel. For the second electrode configuration we mixed MoS2 particles with carbon black and bound them to carbon cloth using a conductive polymer binder (Nafion) (the same approach used with Pt catalysts for HER in fuel cells). We varied the mass loading ratio between MoS2 and carbon black, which affects the surface density of the catalyst, in order to determine conditions for optimal performance. Once the optimal loading ratio was found for the MoS2 composite cathodes, MEC reactors were operated in fed-batch mode to compare Pt-based composite cathodes to ones made with MoS2 catalyst.
2.
Material and methods
2.1.
Electrode construction
Flat sheets of type 304 stainless steel (Trinity Brand Industries, Inc.) were sanded smooth with silicon carbide sand paper down to a 1200 grit finish, and then polished with successively decreased sizes of alumina particle slurries (to 1 mm). Between each successive step standard metallographic procedures were followed, whereby samples were ultrasonically washed in distilled/deionized water and rinsed with acetone with a final rinse in distilled/deionized water and drying under a stream of dry nitrogen. Similarly prepared SS samples were burnished (by rubbing the powder onto the SS) with MoS2 particles (Aldrich, 99%, particle size < 2 mm) until an even
coating was obtained on the surface. The amount of MoS2 coating on SS was 7.86 104 mol/m2 (1.109 g/m2), equivalent to w1000 monolayers. MoS2 catalyst powder was mixed with 5 mg/cm2 carbon black (Cabot, VULCAN XC-72R, >99%) at mass loading ratios of 0e50% (Table 1) and 50 mL/cm2 of a 2:1 volume solution of Nafion polymer (Aldrich, 5 wt.%) and iso-propanol, vortexed for w15 s at 3200 rpm (VWR Vortex Mixer), and applied to only the solution side of the carbon cloth (E-TEK, B-1/B/30WP, 30% by weight PTFE Wet-Proofed) using a small paint brush. The electrodes were then air-dried overnight (>12 h) before testing. All MoS2 composite cathodes had the same amount of carbon black and Nafion binder so that the only variable was the amount of MoS2. Carbon cloth cathodes with a Pt catalyst were prepared using a commercially available platinum powder pre-mixed with carbon black (E-TEK, C1-10, 10 wt.% Pt on Vulcan XC-72). The platinum surface density was 5.56 g/m2, which maintained the same amount of carbon black as used in the MoS2 composite cathodes. The Pt-based composite cathodes also used the same quantity of Nafion binder as in the case of the MoS2 composite cathodes. The construction of the platinum composite cathodes has been described previously [15]. All cathodes had a geometric surface area of 12 cm2, but only 7 cm2 was exposed to the solution due to the design of the electrochemical cell.
2.2.
Electrochemical characterization
Cathodes were used as the working electrode in a threeelectrode, two-chamber electrochemical cell. Two different
Table 1 e Composition of cathodes by loading ratio (wt.% catalyst/total mass) and by catalyst surface density, and cathode overpotentials (h) based on LSV and galvanostatic polarization (GStat) data at L7.14 A/m2 in 0.1 M NaClO4 or 50 mM phosphate buffer solution (PBS). For comparison purposes the overpotential was set to zero for the 10 wt.% Pt cathode. Positive values indicate a greater (more cathodic) overpotential and negative values indicate lower overpotential. Sample ID
0 MoS2 5 MoS2 10 MoS2 15 MoS2 20 MoS2 25 MoS2 30 MoS2 35 MoS2 40 MoS2 45 MoS2 50 MoS2 10 wt% Pt Bare SS MoS2 on SS
h (V) in NaClO4
h (V) in PBS
Catalyst loading ratio (%)
Catalyst surface density (g/m2)
0.00 9.09 16.7 23.1 28.6 33.3 37.5 41.2 44.4 47.4 50.0 10.0
0.00 5.00 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 5.56
0.426 0.289 0.286 0.285 0.278 0.252 0.264 0.245 0.247 0.220 0.241 0.000
0.496 0.371 0.356 0.290 0.282 0.298 0.286 0.270 0.235 0.277 0.259 0.000
0.064 0.042 0.002 0.010 0.011 0.023 0.010 0.008 0.028 0.029 0.030 0.000
0.172 0.007 0.125 0.049 0.176 0.083 0.150 0.079 0.095 0.084 0.233 0.000
N/A N/A
N/A 1.11
0.836 0.287 0.333 0.346
1.041 0.105
1.442 0.165
LSV GStat LSV GStat
Author's personal copy i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 9 4 3 9 e9 4 4 5
aqueous electrolytes were tested: 0.1 M NaClO4 (Alfa Aesar, 98e102%, ACS Grade) at pH 5.8; and 50 mM phosphate buffer (PBS, Na2HPO4, 4.58 g/L; and NaH2PO4$H2O, 2.45 g/L) at pH 7. The perchlorate solution was used to avoid possible complicating effects of buffer adsorption (for example phosphate) onto the surfaces [5], and a phosphate buffer solution was used because it is a common electrolyte for MEC studies [1,16e20]. Solutions were degassed by nitrogen sparging for 15 min prior to being added to both cathode and anode chambers before the start of each experiment. The cathode and anode chambers were separated by a cation exchange membrane (SELEMION, CMV, Asahi Glass Co., Ltd., Japan) to limit interference by competing and/or conjugate reactions taking place at each electrode. A platinum flag was used as the counter electrode in the anode chamber with an Ag/AgCl (3 M NaCl) reference electrode (Bioanalytical Systems, Inc., RE-5B) in the cathode chamber. All recorded potentials are corrected for pH and are reported here vs. the RHE scale using E(RHE) ¼ E(Ag/AgCl) þ 0.211 V þ 0.0591 V/pH, where E(Ag/Ag/ Cl) is the reference electrode potential, 0.211 V the shift from Ag/AgCl (3 M NaCl) scale to SHE scale, and 0.0591 V/pH is the shift from SHE to RHE scale. Cathodes were characterized using two electrochemical techniques using a multichannel potentiostat (Solartron 1470E): linear sweep voltammetry (LSV); and galvanostatic polarization. For LSV tests, the cathodes were swept at 1 mV/s from their open circuit potential down to 1.189 V or until a total cathodic current density of 15 A/m2 (NaClO4) or 10 A/m2 (PBS) was reached. Current densities were normalized by the exposed geometric surface area of the electrode (7 cm2). Galvanostatic tests were also conducted to characterize the electrodes under steady state conditions. Current was held for 15 min at 1, 2, 3, 5, 8, and 10 mA for the NaClO4 electrolyte or in the range from 1 to 7 mA in 1 mA steps for the PBS electrolyte, and the steady state potential was plotted as a function of current density. Even though the applied current density is the independent variable and the potential is the dependent variable, the potential was plotted on the x-axis for ease of comparison of results with LSV tests. All electrochemical tests were performed three times to ensure consistency in the results, and all results given here were derived from the third experimental run for each test.
2.3.
9441
In order to compare the performance of MoS2 composite electrodes to standard platinum-based composite cathodes, we operated MECs first with the platinum composite cathodes, and then with one of the best performing MoS2 composite cathodes. Platinum composite cathodes were used for six batch cycles in order to obtain a performance baseline, then they were replaced with MoS2 composite cathodes for three additional fed-batch cycles. Prior to each batch cycle, the MEC reactor headspace was flushed with nitrogen gas for several minutes. During each fed-batch cycle the volume of gas produced was recorded using a respirometer (AER-200; Challenge Environmental) and the evolved gas was collected in air-tight gas bags (0.05 L capacity; Cali-5-Bond, Calibrated Instruments, Inc.) for post-cycle gas composition analysis using gas chromatography (Model 310; SRI Instruments). Hydrogen production yield and energy recovery were used to calculate the electrical, substrate, and overall energy efficiencies as previously described [16].
3.
Results and discussion
3.1.
LSV results in NaClO4
Based on LSV scans in a NaClO4 solution, the optimum MoS2 loading was 54 mg MoS2 and 60 mg carbon black, or 47 wt.% MoS2, with a surface density of 45 g/m2 (Fig. 1 and Table 1). This optimum catalyst loading had 0.616 V less HER
Microbial electrolysis cell (MEC) tests
MECs (quadruplicate reactors) were operated in fed-batch mode using single-chamber reactors [16] containing 30 mL of solution with bioanodes that had been well-established in previous MEC tests (>60 batch cycles). The solution consisted of 50 mM phosphate buffer (pH 7) as well as trace vitamins and minerals [21] and 1 g/L sodium acetate (anhydrous, SigmaeAldrich) as the organic substrate. Electrode potentials were measured with Ag/AgCl reference electrodes. The MECs were operated by potentiostatically setting the bioanode voltage to 0.276 V vs. Ag/AgCl (0.349 V vs. RHE), which was chosen based on prior results that indicated this value as the potential that yields the maximum current for the oxidation of acetate anion. The cathodes were the counter electrodes and their potentials were recorded using a PC controlled multimeter (Model 2700; Keithley Instruments, Inc.).
Fig. 1 e LSV scans of select cathodes in 0.1 M NaClO4. Solid lines: (A, red) bare stainless steel; (B, blue) MoS2 coated stainless steel; and (C, black) 10% platinum on carbon cloth. The dashed lines correspond to various loading ratios of MoS2 added to carbon black and applied to carbon cloth: 0 wt.% MoS2 (1, red); 9 wt.% MoS2 (2, orange); 23 wt.% MoS2 (3, green); 33 wt.% MoS2 (4, blue); and 47 wt.% MoS2 (5, violet). The arrow indicates increasing MoS2 loading ratio for the dashed lines. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Author's personal copy 9442
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 9 4 3 9 e9 4 4 5
overpotential than SS, but 0.220 V more overpotential than Pt on carbon cloth (Table 1). A total of 11 different MoS2 electrodes were examined, although many of these curves were omitted from the Figures for clarity as all followed the same general trend with mass loading as those included in Fig. 1. Once the mass loading ratio reached 47% there was no further improvement found by increasing the MoS2 to carbon black mass ratio. The SS electrode burnished with MoS2 had a greater catalytic activity than uncoated SS electrode (Fig. 1). The layer of MoS2 on SS enhanced hydrogen evolution as evidenced by a reduction of HER overpotential of 0.503 V compared to bare stainless steel. Based on structural information [4,10,11] that MoS2 has a laminar structure and that the electrochemically active sites occur at the edges of the crystal planes, it appears that a burnished MoS2 coating resulted in exposure of the active edge sites of MoS2, resulting in higher activity than bare SS. The overpotential for bare SS was much higher than that of Pt on carbon cloth in LSV tests, as expected from previous tests [13]. For example, the overpotential for SS determined by LSV was 0.836 V larger than Pt on carbon cloth.
3.2.
Galvanostatic results in NaClO4
Even though the sweep rate used in the LSV scans was relatively low (1 mV/s), which is generally accepted as slow enough to approximate steady state conditions, we performed additional constant current tests to better approximate the performance of these materials under steady state conditions. Under these truly steady state conditions any time-dependent response, for example capacitive surface charging/discharging, is eliminated over the long period of time used to perform the constant current tests. The cathode made with 48 mg MoS2 and 60 mg carbon black on carbon cloth exhibited the lowest overpotential compared to platinum on carbon cloth, which consisted of 44 wt.% MoS2, with a 40 g/m2 MoS2 surface density (Fig. 2 and Table 1). This electrode required 0.050e0.060 V less overpotential than bare SS for high current densities (7 to 14 A/m2), and at less negative current densities (less than 7 A/m2) the MoS2 catalyst cathodes substantially out performed bare SS. For example, at 4 A/m2 the MoS2 cathode had 0.116 V less overpotential than SS, at 3 A/m2 the HER overpotential was 0.386 V less than bare SS (which out performed 10% platinum by 0.134 V), and at 1 A/m2 the MoS2 cathode had 0.896 V less overpotential than SS for the HER (better than platinum by 0.646 V).
3.3.
LSV tests in phosphate buffer
The optimum MoS2 loading in the phosphate buffer solution based on LSV scans was 47 wt.% MoS2, in agreement with results obtained with LSV scans and a NaClO4 solution (Fig. 3 and Table 1). This cathode had 1.075 V less overpotential than bare stainless steel and 0.034 V less overpotential than a 10 wt.% Pt on carbon cloth cathode (Table 1). LSV scans (Fig. 3) suggest that a 33 wt.% loading (25 g/m2) performs better than the 47 wt.% loading at higher current densities, even though cathodes with these two loadings have nearly the same HER overpotential.
Fig. 2 e Galvanostatic polarization data for select cathodes in 0.1 M NaClO4. Solid lines: bare stainless steel (no symbol, red); MoS2 coated stainless steel ( , blue); and 10 wt.% Pt on carbon cloth (A, black). The dashed lines correspond to various loading ratios of MoS2 added to carbon black and applied to carbon cloth: 0 wt.% MoS2 ( , red); 9 wt.% MoS2 ( , orange); 23 wt.% MoS2 ( , green); 33 wt.% MoS2 ( , blue); and 44 wt.% MoS2 ( , violet). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
As the MoS2 to carbon black mass ratio increased from 0% to 50%, initially there was a steady decrease in HER overpotential up to about 17 wt.% MoS2, but beyond this loading ratio the HER overpotential did not scale directly with the amount of MoS2. For example, the HER overpotential as measured by LSV for the 50 wt.% loading was nearly the same as the cathode made with 9 wt.% MoS2. One possible reason for the decreased performance may have been too many MoS2 particles present, resulting in the MoS2 particles on the outermost surface blocking otherwise available edge sites of the underlying MoS2 particles. Another possible reason that the HER overpotential did not scale with the amount of catalyst in this case could be due to phosphate poisoning, as observed with platinum-based cathodes [5]. The SS cathode coated with MoS2 had a greater catalytic activity than the bare SS electrode (Fig. 3). The MoS2 coating improved the cathodic current density significantly at applied potentials below about 0.3 V vs. RHE. The bare SS cathode produced a nearly zero cathodic current density in the range examined here (Line “A”, Fig. 3), but with the addition of a MoS2 coating we observed a current density of nearly 2.5 A/m2 at about 1 V vs. RHE (Line “B”, Fig. 3). Cathodes constructed by coating stainless steel with MoS2 particles offer a sizeable improvement in HER performance (Table 1). The bare SS cathode had a HER overpotential using LSV of 1.041 V beyond that of the platinum-based cathode. The MoS2 coated SS cathode had a HER overpotential of only 0.105 V beyond platinum on carbon cloth. The overpotential for bare SS was much higher than that of platinum on carbon cloth in LSV tests (Table 1), consistent with previous tests [17].
Author's personal copy i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 9 4 3 9 e9 4 4 5
Fig. 3 e LSV scans of select cathodes in 50 mM PBS. Solid lines: bare stainless steel (A, red); MoS2 coated stainless steel (B, blue); and 10% platinum on carbon cloth (C, black). The dashed lines correspond to various loading ratios of MoS2 added to carbon black and applied to carbon cloth: 0 wt.% MoS2 (1, red); 9 wt.% MoS2 (2, orange); 23 wt.% MoS2 (3, green); 33 wt.% MoS2 (4, blue); and 47 wt.% MoS2 (5, violet). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
9443
Fig. 4 e Galvanostatic polarization data for select cathodes in 50 mM PBS. Solid lines: bare stainless steel (no symbol, red); MoS2 coated stainless steel ( , blue); and 10 wt.% Pt on carbon cloth (A, black). The dashed lines correspond to various loading ratios of MoS2 added to carbon black and applied to carbon cloth: 0 wt.% MoS2 ( , red); 9 wt.% MoS2 ( , orange); 23 wt.% MoS2 ( , green); 33 wt.% MoS2 ( , blue); and 47 wt.% MoS2 ( , violet). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Galvanostatic tests in phosphate buffer showed that the cathode made with 47 wt.% MoS2 in carbon black on carbon cloth (45 g/m2 MoS2 surface density) produced the lowest HER overpotential, which was only slightly better than cathodes made with 33, 41, and 23 wt.% MoS2 (Table 1). These four composite MoS2 cathodes out performed the platinum on carbon cloth cathodes (Fig. 4 and Table 1). In addition, these four electrodes had nearly 1.5 V less overpotential than bare SS. The MoS2 coated SS cathode also showed a significant improvement in performance compared to bare SS in galvanostatic tests. The HER overpotential at high current density was reduced by over 1.2 V simply by burnishing MoS2 catalyst onto SS (Table 1). Even though the performance of MoS2 coated SS was not as good as the composite MoS2 cathodes, these results confirm what we observed in the LSV tests, that MoS2 was an effective low-cost HER catalyst.
these cathodes in the MECs was 10.0 1.6 A/m2 for Pt and 10.7 1.2 A/m2 for MoS2, and the average fed-batch cycle time was 14.7 0.4 h (Fig. 5). Gas compositions were H2 ¼ 90 6%, CH4 ¼ 1 0.6%, and CO2 ¼ 8.7 5.4% for the Pt cathodes, and H2 ¼ 92 3%, CH4 ¼ 1 0.8%, and CO2 ¼ 6.7 2.3% for the MoS2 cathodes (Fig. 6). COD removals were 88 2% for MECs with Ptbased cathodes and 90 2% for the MECs with MoS2 based cathodes (Fig. 7). As a result of the similar COD removals and gas composition, the Coulombic and substrate efficiencies are very similar for the two types of cathodes. The main difference in results obtained with the two types of cathodes was that the electrical energy efficiency with the platinum-based cathodes was slightly better (158 14%) than that with the MoS2 cathodes (123 7%) (Fig. 7). The overall efficiency, which takes into account the electrical and substrate efficiencies, was very similar for the two catalyst materials, with 61 5% for the platinum-based cathodes and 57 3% for the MoS2 cathodes.
3.5.
3.6.
3.4.
Galvanostatic tests in phosphate buffer
MEC tests
Based on LSV and galvanostatic polarization tests in two electrolytes, and after consideration of the costs of electrode materials (discussed below), the 33 wt.% MoS2 composite cathode was chosen for additional tests in MECs. The performance of the MECs in terms of current densities, gas composition, COD removal and Coulombic, electrical, substrate, and combined efficiencies were not appreciably different for the MoS2 and Pt catalyst cathodes. The average current density for
Consideration of material costs and performance
A rough comparison was made of the costs of the different types of cathodes, based on purchased prices, to quantify possible savings using the different materials. Bare or MoS2 coated stainless steel cathodes have the lowest cost (approximately $57/m2 for either one). Composite MoS2 cathodes are about an order of magnitude more expensive (about $600/m2) and 10 wt.% platinum composite cathodes are about five times more expensive (about $2700/m2) than similarly constructed
Author's personal copy 9444
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 9 4 3 9 e9 4 4 5
Fig. 5 e MEC reactor current density vs. time. The first six batch cycles correspond to MECs with 10 wt.% Pt cathodes and the final three batch cycles correspond to 33 wt.% MoS2 cathodes. MEC reactors #1 through #4 correspond to black, red, green, and blue lines, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
composite MoS2 cathodes. The composite MoS2 cathodes are less expensive than the platinum-based cathodes simply due to the much lower catalysts price because they use the same carbon cloth and polymer binder. For example, the 10 wt.%
Fig. 7 e MEC efficiencies for each reactor: Coulombic efficiency (-, black); electrical efficiency ( , blue); substrate efficiency ( , red); and overall efficiency ( , green). Results for Pt were averaged over six cycles and results for MoS2 were averaged over three cycles. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
platinum in carbon black catalyst costs $2135/m2 at present prices, whereas the MoS2 in carbon black costs about $11/m2. In order to reduce the price of the composite MoS2 cathodes even further, alternatives to the carbon cloth could be used as this material contributes 80% of the cathode cost for MoS2 composite cathodes. One such material is a carbon mesh material (3K, Plain Weave Carbon Fibre Fabric, Fibre Glast Development Corp.), which only costs about $43/m2 as opposed to the carbon cloth, which costs about $470/m2 at present prices. The binder used here (Nafion) was not critical to the performance of the MoS2. We also tested two other conductive polymers (poly[bisphenol A-co-epicholorohydrin], BAEH; and poly[methyl vinyl ether-alt-maleic acid monobutyl ester], PMVEMABE) (data not shown). Both of these other polymers had a similar performance to Nafion, but offer the possibility of using a binder at a much reduced cost.
4. Fig. 6 e MEC gas composition (stacked bars, left axis) and percent COD removal (black dots, right axis). Gas composition was: H2 (blue, bottom), CH4 (black, middle), and CO2 (red, top). Separate MEC reactor results for gas composition and percent COD removal are shown as averages over six batch cycles for those using 10 wt.% Pt cathodes and over three batch cycles for MoS2 cathodes. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Conclusions
Abiotic tests using either LSV or galvanostatic polarization showed that MoS2 particles dispersed into carbon black and applied to carbon cloth performed much better than bare SS as an HER catalyst. Electrochemical tests also showed that this material could even surpass the performance of platinumbased electrodes prepared in the same manner. The overall best choice of MoS2 to carbon black loading ratio was found to be 47 wt.% based only on overpotentials. However, considering the small overpotential differences for catalysts at different loadings, and the need to reduce materials costs, the
Author's personal copy i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 6 ( 2 0 1 1 ) 9 4 3 9 e9 4 4 5
33 wt.% MoS2 cathode is the best choice. MEC tests using this optimum loading ratio showed that MoS2 cathodes performed similarly to platinum cathodes in terms of current densities, gas production, COD removal, and energy efficiency. Thus, MoS2 cathodes prepared as described here provide a new, lowcost material that has advantages of ease of fabrication and durability similar to platinum but at a much lower price.
[9]
[10]
[11]
Acknowledgements We thank Air Products and Chemicals, Inc. for the financial support of this project and appreciate the useful discussions with Dr. Matthew D. Merrill, Dr. Tomonori Saito, and Dr. Younggy Kim.
[12]
[13]
references [14] [1] Liu H, Grot S, Logan BE. Electrochemically assisted microbial production of hydrogen from acetate. Environ Sci Technol 2005;39:4317e20. [2] Cheng S, Logan BE. Sustainable and efficient biohydrogen production via electrohydrogenesis. Proc Natl Acad Sci 2007; 104:18871e3. [3] Schmickler W. Interfacial electrochemistry. New York: Oxford University Press; 1996. [4] Hinnemann B, Moses PG, Bonde J, Jørgensen KP, Nielsen JH, Horch S, et al. Biomimetic hydrogen evolution: MoS2 nanoparticles as catalyst for hydrogen evolution. J Am Chem Soc 2005;127:5308e9. [5] De Silva Mun˜oz L, Bergel A, Fe´ron D, Basse´guy R. Hydrogen production by electrolysis of a phosphate solution on a stainless steel cathode. Int J Hydrogen Energy 2010;35:8561e8. [6] Koper MTM, Bouwman E. Electrochemical hydrogen production: bridging homogeneous and heterogeneous catalysis. Angew Chem Int Ed 2010;49:3723e5. [7] Sobczynski A. Molybdenum disulfide as a hydrogen evolution catalyst for water photodecomposition on semiconductors. J Catal 1991;131:156e66. [8] Zong X, Yan H, Wu G, Ma G, Wen F, Wang L, et al. Enhancement of photocatalytic H2 evolution on CdS by
[15]
[16]
[17]
[18]
[19]
[20]
[21]
9445
loading MoS2 as cocatalyst under visible light irradiation. J Am Chem Soc 2008;130:7176e7. Zong X, Wu G, Yan H, Ma G, Shi J, Wen F, et al. Photocatalytic H2 evolution on MoS2/CdS catalysts under visible light irradiation. J Phys Chem C 1963e1968;2010:114. Elizondo-Villarreal N, Vela´zquez-Castillo R, Galva´n DH, Camacho A, Yacama´n MJ. Structure and catalytic properties of molybdenum sulfide nanoplatelets. Appl Catal A 2007;328: 88e97. Jaramillo TF, Jørgensen KP, Bonde J, Nielsen JH, Horch S, Chorkendorff I. Identification of active edge sites for electrochemical H2 evolution from MoS2 nanocatalysts. Science 6 July 2007;317:100e2. Brorson M, Carlsson A, Topsøe H. The morphology of MoS2, WS2, CoeMoeS, NieMoeS and NieWeS nanoclusters in hydrodesulfurization catalysts revealed by HAADF-STEM. Catal Today 2007;123:31e6. Kisielowski C, Ramasse QM, Hansen LP, Brorson M, Carlsson A, Molenbroek AM, et al. Imaging MoS2 nanocatalysts with single-atom sensitivity. Angew Chem Int Ed 2010;49:2708e10. Bonde J, Moses PG, Jaramillo TF, Nørskov JK, Chorkendorff I. Hydrogen evolution on nano-particulate transition metal sulfides. Faraday Discuss 2008;140:219e31. Cheng S, Liu H, Logan BE. Increased performance of singlechamber microbial fuel cells using an improved cathode structure. Electrochem Commun 2006;8:489e94. Call D, Logan BE. Hydrogen production in a single chamber microbial electrolysis cell lacking a membrane. Environ Sci Technol 2008;42:3401e6. Zhang Y, Merrill MD, Logan BE. The use and optimization of stainless steel mesh cathodes in microbial electrolysis cells. Int J Hydrogen Energy 2010;35:12020e8. Bond DR, Lovley DR. Electricity production by Geobacter sulfurreducens attached to electrodes. Appl Environ Microbiol 2003;69:1548e55. Hu H, Fan Y, Liu H. Hydrogen production using singlechamber membrane-free microbial electrolysis cells. Water Res 2008;42:4172e8. Sleutels THJA, Hamelers HVM, Rozendal RA, Buisman CJN. Ion transport resistance in microbial electrolysis cells with anion and cation exchange membranes. Int J Hydrogen Energy 2009;34:3612e20. Balch WE, Fox GE, Magrum LJ, Woese CR, Wolfe RS. Methanogens: reevaluation of a unique biological group. Microbiol Rev 1979;43:260e96.
