Energy Fuels 2010, 24, 1176–1183 Published on Web 01/20/2010
: DOI:10.1021/ef9009636
Performance and Byproduct Analysis of Coal Gas Solid Oxide Fuel Cell Rahul Singh,† Felipe Guzman, Rajesh Khatri,‡ and Steven S. C. Chuang* Department of Chemical and Biomolecular Engineering, The University of Akron, 200 East Buchtel Commons, Akron, Ohio 44325-3906. †Present address: Topsoe Fuel Cell A/S, Nymollevej 60, Kgs, Lyngby DK-2800, Denmark. ‡Present address: Center for Applied Energy Research, Lexington, KY 40511. Received August 31, 2009. Revised Manuscript Received November 2, 2009
The direct electrochemical oxidation of coal gas was studied by pyrolyzing a sample of Ohio #5 coal in flowing Ar at 700, 800, 900, and 950 °C and transporting the resulting gaseous products to a Cu (Copper) anode solid oxide fuel cell (SOFC) operated at 950 °C. Pyrolysis of coal at 700 °C produced a H2-rich coal gas containing 89% H2, 4% CO, 6% CH4, 1% CO2, and sulfur compounds (i.e., 1% COS and 1% SO2), which yielded a maximum current density of 320 mA/cm2 at 0.5 V. Raising the pyrolysis temperature from 700 to 950 °C increased the CO concentration in the coal gas (i.e., 42 vol % CO), which in turn reduced the fuel cell maximum current density. No sulfur compounds were present in the coal gas produced at temperatures higher than 700 °C. Coal gas fuel operation did not degrade the fuel cell performance. A fuel composition of 25 vol % CH4 in He generated a current density of 340 mA/cm2 at 0.5 V. These results demonstrated that the Cu anode is effective for the electrochemical oxidation of sulfur-containing coal gas at 950 °C. The addition of CO2 and D2O to the pyrolysis reactor led to the formation of CO and HD, indicating the occurrence of reforming reactions. Diffuse reflectance infrared Fourier transformation (DRIFT) spectra showed that coal pyrolysis proceeded by dehydrogenation of hydroaromatics, dealkylation of aromatics, and oxidation reactions, leading to the formation of coke with a surface containing C-O, C-S, and S-O bonds. The results of the fuel cell performance strongly support the feasibility of direct power generation from coal gas in a Cu-anode SOFC.
deactivation of the anode by the coal gas. The majority of SOFCs use Ni (Nickel) anodes that have been optimized for the electrochemical oxidation of H2 and sulfur-free syngas.5 These anodes are susceptible to deactivation by sulfur poisoning and coking.6-9 Therefore, the development of highly active and durable coal gas fuel cells requires the use of alternate anode catalysts. Cu may serve as an excellent anode catalyst because of its oxidation ability as well as carbon and sulfur resistance.10-12 Previous studies have demonstrated the use of Cu-based anodes for the electrochemical oxidation of CH4 and n-decane fuel with 5000 ppm of thiophene at 700-800 °C without deactivation.11 The resistance to deactivation of the Cu-based anode has been attributed to its low activities for C-C and Cu-S bond formation, which are reaction steps leading to carbon and sulfur poisoning.12 Although Cu anodes have shown carbon and sulfur tolerance, they can be susceptible to sintering because of the low melting point of Cu metal. The objective of this study is to evaluate the feasibility of generating power by electrochemically oxidizing the sulfur- and
Introduction Coal-fired power plants supply about 50% of the electricity consumed in the United States.1 Two main drawbacks of producing electricity in a coal-fired power plant are low efficiency and emission of CO2, NO, SO2, and Hg.1-3 The low efficiency is a result of combining thermal/mechanical processes, such as those presented in Figure 1a: (i) combustion of coal to produce thermal energy (i.e., heat), (ii) generation of steam to drive a turbine, and (iii) usage of kinetic energy of the turbine to generate power. One potential approach to improve the efficiency of coal power generation, shown in Figure 1b, would be to pyrolyze coal to produce coal gas and coke byproduct. The coal gas, which contains high concentrations of hydrocarbons and sulfur-containing species, including COS, SO2, and H2S, can be electrochemically oxidized in a solid oxide fuel cell (SOFC) to produce electricity. The coke byproduct can be converted to high value activated carbon.4 Hydrocarbons and sulfur-containing species present in the coal gas can poison the anode catalyst of the fuel cell. Therefore, the key issue of developing a coal gas fuel cell (a SOFC with a direct feed of coal gas) lies in enhancing the electrochemical oxidation activity and improving the resistance to
(5) Yu, Z.; Chuang, S. S. C. Appl. Catal., A 2007, 327 (2), 147–156. (6) Chen, X. J.; Khor, K. A.; Chan, S. H. Electrochem. Solid-State Lett. 2005, 8 (2), A79–A82. (7) Dong, J.; Cheng, Z.; Zha, S.; Liu, M. J. Power Sources 2006, 156 (2), 461–465. (8) Matsuzaki, Y.; Yasuda, I. Solid State Ionics 2000, 132 (3,4), 261– 269. (9) Weber, A.; Sauer, B.; Muller, A. C.; Herbstritt, D.; Ivers-Tiffee, E. Solid State Ionics 2002, 152-153, 543–550. (10) Gorte, R. J.; Vohs, J. M. J. Catal. 2003, 216 (1-2), 477–486. (11) Kim, H.; Vohs, J. M.; Gorte, R. J. Chem. Commun. 2001, 22, 2334–2335. (12) Lu, Z.; Pei, L.; He, T.-m.; Huang, X.-q.; Liu, Z.-g.; Ji, Y.; Zhao, X.-h.; Su, W.-h. J. Alloys Compd. 2002, 334 (1-2), 299–303.
*To whom correspondence should be addressed. Fax: 1-330-9725856. E-mail:
[email protected]. (1) http://www.energy.gov/energysources/electricpower.htm (accessed on Aug 20, 2009). (2) http://www.fossil.energy.gov/programs/powersystems/gasification/ index.html (accessed on Aug 25, 2009). (3) Cheng, C.-M.; Hack, P.; Chu, P.; Chang, Y.-N.; Lin, T.-Y.; Ko, C.-S.; Chiang, P.-H.; He, C.-C.; Lai, Y.-M.; Pan, W.-P. Energy Fuels 2009, 23 (10), 4805–4816. (4) Alonso, A.; Ruiz, V.; Blanco, C.; Santamaria, R.; Granda, M.; Menendez, R.; de Jager, S. G. E. Carbon 2006, 44 (3), 441–446. r 2010 American Chemical Society
1176
pubs.acs.org/EF
Energy Fuels 2010, 24, 1176–1183
: DOI:10.1021/ef9009636
Singh et al. Table 1. Proximate Analysis of Ohio #5 Coal proximate analysis moisture (%) dry ash (%) dry volatile matter (%) dry fixed carbon (%)
ultimate analysis
4.15 4.80 37.98
carbon (%) hydrogen (%) nitrogen (%)
57.22
oxygen (%)
83.99 5.50 1.88 8.6
sulfur forms pyritic (%) organic (%) sulfate (%)
0.70 1.21 0.01
total (%)
1.92
It is important to note that the anode impregnation step was carried out after calcination of the cathode because of the low melting point of Cu. The cathode was prepared by (i) mixing LSM (La0.8Sr0.2 MnO3-x) powders and glycerine in a 1/1.3 (LSM/glycerine) weight ratio, (ii) applying this mixture on the electrolyte disk, and (iii) calcining at 1200 °C for 2 h. LSM powders were prepared by the soft chemical synthesis technique,13 which involved (i) preparing a NH4-ethylenediaminetetraacetic acid (EDTA) solution by mixing EDTA (99.9%, Avocado) in 25% (v/v) NH4OH solution, (ii) dissolving stoichiometric amounts of La(NO3)3 3 6H2O, Sr(NO3)2, and Mn(CH3COO)2 3 4H2O in NH4-EDTA at the weight ratio of 1/1.5 (La þ Sr þ Mn/ EDTA), (iii) heating to 250 °C to initiate a self-combustion reaction forming an ash material, and (iv) calcining the resultant ash material at 800 °C for 1 h, producing a black-colored LSM powder. The chemical precursors used in the fuel cell preparation were purchased from Alfa-Aesar, unless specified otherwise. The anode and cathode materials were characterized by the X-ray diffraction technique (Philips Analytical, λ = 1.54 A˚). The fuel cell was sealed to an alumina tube (McDanel, Inc.) with a high-temperature ceramic paste (Product 503, Aremco) and was cured in two steps: (i) 85 °C for 2 h and (ii) 235 °C for 2 h. Pt grids and wires (99.9%, Alfa-Aesar) were attached to the cathode and anode with Pt paste (Engelhard, Lot 44338A) and were fired at 950 °C for 30 min. The inset of Figure 2 shows an image of the SOFC assembly. The current-voltage performance of the Cu SOFC was recorded by an in-house LabVIEW system (National Instruments) under different types of flow: (a) pure coal gas, (b) coal gas saturated with D2O, and (c) coal gas mixed with CO2. Experimental Apparatus. Figure 2 shows the experimental apparatus used to study the electrochemical oxidation of the coal gas produced from coal pyrolysis. A 3.5 g sample of Ohio #5 coal was placed in a quartz pyrolysis reactor and heated from 25 to 700 °C at a heating rate of 20 °C/min under flowing 40 cm3/min Ar stream. Coal gas was generated at 700, 800, 900, and 950 °C by maintaining the pyrolysis reactor at each temperature for 70 min, as shown in Figure 2b. The coal gas produced was transported by the Ar stream to the fuel cell system, which was placed inside a furnace operating at 950 °C. The fuel cell current-voltage performance was recorded for (a) pure coal gas, (b) coal gas generated with the addition of D2O, and (c) coal gas generated with the addition of CO2, as shown in the sequence of experimental runs in Figure 2b. D2O and CO2 were introduced into the pyrolysis reactor by a step method to study their effect on the composition of coal gas produced. The step method employs a four-port valve, which instantaneously changes the inlet flow to the reactor from Ar (40 cm3/min) to Ar/CO2 (37/3 cm3/min, Ar mixed with CO2) or Ar/D2O (3 vol %, Ar mixed with D2O). The gas flows were controlled by mass flow controllers (Brooks Instrument, 1580E). The composition of the coal gas was monitored by a Fourier transform infrared (FTIR) spectrometer (560 series, Nicolet) and a mass spectrometer (MS, QMS200 Balzers-Pfeiffer). The following m/e ratios were monitored on MS: 2 (H2), 3 (HD), 15 (CH4), 18 (H2O), 28 (CO), 32 (O2), 40 (Ar), 44 (CO2), and 64 (SO2).
Figure 1. Schematic representation of the (a) coal-fired power plant, (b) integrated gasification fuel cell (IGFC), and (c) coal gas SOFC.
hydrocarbon-containing coal gas in a Cu-anode SOFC. Coal gas was generated by pyrolyzing Ohio #5 coal at 700, 800, 900, and 950 °C and transporting the gaseous products to the fuel cell at 950 °C. The power generation characteristics of the SOFC were determined by comparing voltage-current polarization plots for various compositions of coal gas. CO2 and D2O were introduced into the pyrolysis reactor to probe their effect on the composition of coal gas produced and, ultimately, the power generation characteristics of the fuel cell. The initial Ohio #5 coal and the resultant coke were further characterized by diffuse reflectance infrared Fourier transformation (DRIFT) and Raman spectroscopy. Experimental Section Ohio #5 Coal and Coke Characterization. Ohio #5 coal and coke were characterized by DRIFT and Raman spectroscopy. The samples for DRIFT analysis were prepared by sieving the coal and coke through a screen opening size of 94 mesh, diluting the screened material in KBr (1:20 weight ratio), and loading the samples in a DRIFT cell (Spectra-tech) placed in an IR bench (Nicolet 560). DRIFT spectra were recorded at room temperature, co-adding 32 scan at a resolution of 4 cm-1. The samples for Raman analysis were recorded with a Raman spectrophotometer equipped with a triple monochromator (Jobin Yvon T6400) at a frequency of 514.5 nm. No dilution with KBr was required for Raman analysis. Proximate analysis of Ohio #5 coal is shown in Table 1. Fuel Cell Preparation and Characterization. The SOFC used for this study was composed of an yttria-stabilized zirconia (YSZ) electrolyte disk (8 mol % Y2O3, 15 mm in diameter, 1 mm in thickness, Tosoh), a Cu/YSZ powder-YSZ fibers anode (20 μm in thickness), and a lanthanum strontium manganite (LSM) YSZ/LSM cathode (40 μm in thickness). The anode was prepared by (i) applying a thin layer of a mixture of dense YSZ powder (Tosoh), porous zirconia fibers (Zircar), and glycerine (1:3:4 weight ratio) on the electrolyte disk, (ii) calcining at 1400 °C to obtain a 20 μm thickness of the anode, (iii) impregnating 15 wt % CuCl3.6H2O (Sigma-Aldrich) onto the anode layer, and (iv) calcining at 950 °C. The weight percent has been calculated on the basis of the amount of YSZ powder and fibers on the anode.
(13) Huang, Y.-H.; Xu, Z.-G.; Yan, C.-H.; Wang, Z.-M.; Zhu, T.; Liao, C.-S.; Gao, S.; Xu, G.-X. Solid State Commun. 2000, 114 (1), 43–47.
1177
Energy Fuels 2010, 24, 1176–1183
: DOI:10.1021/ef9009636
Singh et al.
Figure 2. (a) Experimental setup and (b) sequence of experimental runs.
of a higher melting point metal, which can raise the melting point of a Cu alloy anode.25 IR and MS Analysis of Coal Pyrolysis. Figure 4 shows the IR spectra of the coal gas produced during the pyrolysis of Ohio #5 coal. The IR spectrum of the coal gas produced at 320 °C shows the presence of CO2 at 2349 cm-1, CO at 2183 cm-1, water at 3406 cm-1, COS at 2941 cm-1, SO2 at 1341 cm-1, and oxygenated species (CdO) at 1653 cm-1. The IR spectra of the coal gas produced from 465 to 950 °C showed the following trend: (i) appearance of CH4 from 640 °C (i.e., IR band at 3016 cm-1), (ii) disappearance of COS and SO2 at 800 °C, and (iii) increase in CO and water formation at 800 °C. The IR results were complemented with MS profiles of gaseous species produced from coal pyrolysis, shown in Figure 5a. The coal pyrolytic reaction initiated with the formation of gaseous species (i.e., CO2 at 200 °C, CO at 240 °C, SO2 at 260 °C, CH4 at 465 °C, and H2 at 480 °C). The trend of gas-phase species observed on MS is consistent with that observed in the IR. It should be noted that variations of MS and IR profiles of the coal gas components, which were continuously fed to the fuel cell, reflect the transient nature of the coal pyrolysis experiment. The decrease in the SO2 profile at temperatures above 550 °C can be attributed to the semi-batch operation of the pyrolysis reactor, where coal was not constantly fed and sulfur can be depleted. Figure 5b shows the open circuit voltage (OCV) produced by the fuel cell during exposure to this coal gas. Table 2 summarizes the coal gas composition and the observed OCV. The rapid rise of the OCV began when the pyrolysis reactor reached 550 °C, producing a significantly high concentration of CH4.
Results and Discussion XRD Analysis of LSM and Cu-YSZ. Figure 3a shows the XRD pattern of the LSM cathode, which reveals the formation of the cubic-phase perovskite structure with a lattice parameter of 3.9 A˚.13-16 The small hump observed in the lower diffraction angles indicates the presence of a residual amount of amorphous LSM. Figure 3b shows the pattern of the fresh Cu-YSZ, evidencing the CuO and cubic zirconia phase of YSZ.17-21 Analysis of the Cu-YSZ anode after the coal gas experiment, shown in Figure 3c, revealed the formation of CuS.22-24 The increase in the intensity of the CuO peaks of the used anode indicates the growth of the CuO crystallites, suggesting the occurrence of sintering of Cu metal particles. The sintering could be abated by the addition
(14) Grossin, D.; Noudem, J. G. Solid State Sci. 2004, 6 (9), 939–944. (15) Im, H. S.; Chon, G. B.; Lee, S. M.; Koo, B. H.; Lee, C. G.; Jung, M. H. J. Magn. Magn. Mater. 2007, 310 (2, Part 3), 2668–2670. (16) Ma, Y. R.; Chueh, C. H.; Kuang, W. L.; Liou, Y.; Yao, Y. D. Surf. Sci. 2002, 507-510, 573–576. (17) Avgouropoulos, G.; Ioannides, T. Appl. Catal., B 2006, 67 (1-2), 1–11. (18) Callon, G. J.; Goldie, D. M.; Dibb, M. F.; Cairns, J. A.; Paton, J. J. Mater. Sci. Lett. 2000, 19 (19), 1689–1691. (19) Kikuchi, N.; Tonooka, K.; Kusano, E. Vacuum 2006, 80 (7), 756– 760. (20) Lanke, U. D.; Vedawyas, M. Nucl. Instrum. Methods Phys. Res., Sect. B 1999, 155 (1,2), 97–101. (21) Papadimitropoulos, G.; Vourdas, N.; Vamvakas, V. E.; Davazoglou, D. J. Phys.: Conf. Ser. 2005, 10, 182–185. (22) Chen, L.; Chen, J.; Zhou, H.; Liu, L.; Wan, H. Mater. Lett. 2007, 61 (10), 1974–1977. (23) Isac, L.; Duta, A.; Kriza, A.; Manolache, S.; Nanu, M. Thin Solid Films 2007, 515 (15), 5755–5758. (24) Qin, A.-M.; Fang, Y.-P.; Ou, H.-D.; Liu, H.-Q.; Su, C.-Y. Cryst. Growth Des. 2005, 5 (3), 855–860.
(25) Gross, M. D.; Vohs, J. M.; Gorte, R. J. Electrochim. Acta 2006, 52 (5), 1951–1957.
1178
Energy Fuels 2010, 24, 1176–1183
: DOI:10.1021/ef9009636
Singh et al.
Figure 3. X-ray diffraction patterns of (a) LSM cathode, (b) fresh Cu-YSZ anode, and (c) Cu-YSZ anode after exposure to coal gas at 950 °C.
Figure 5. (a) MS results of coal pyrolysis and (b) OCV of the coal gas fuel cell during the temperature ramp from 25 to 700 °C. The variation of the IR intensity of the COS band at 2064 cm-1 has been included for completeness.
pyrolysis, forming carbonaceous species that may further enhance the anode electronic conductivity. The formation of gaseous products can be understood by considering the semi-batch operation of the coal pyrolysis reactor, which can be described by the mole balance on a species A as dCA ν0 ¼ rA - CA dt V which can be simplified to ln
ν0 dNA =dt EA n NA ¼ lnðCA AWÞ - ln 1 þ ðν0 =VÞNA V RT
where n is the order of the reaction for the formation of species A, ν0 is the volumetric flow rate (cm3/min), V is the volume, W is the weight of coal (g), A is the pre-exponential factor, EA is the activation energy, T is the temperature, and NA is the number of moles. The activation energies for the evolution of gaseous species, calculated by plotting ln(NA) versus 1/T, were determined to be 133 kJ/mol for H2, 83 kJ/mol for CH4, 67.6 kJ/mol for CO, and 37.6 kJ/mol for CO2. These activation energies are in agreement with those previously reported
Figure 4. IR spectra of coal gas produced between 25 and 950 °C.
Exposure of the Cu anode to this high concentration CH4 at 950 °C is expected to cause the reduction of CuO to Cu metal, generating the observed OCV. Achieving an OCV of 1.02 V, close to the theoretical Nernst potential of CH4 electrochemical oxidation, indicates that CuO is completely reduced to Cu metal, which provides oxidation activity and electronic conductivity. CH4 could also decompose at 950 °C through 1179
Energy Fuels 2010, 24, 1176–1183
: DOI:10.1021/ef9009636
Singh et al.
Table 2. Effluent Concentration of Coal Gas and Respective OCV of the Coal Gas Fuel Cell position
T (°C)
H2 (%)
CO (%)
CH4 (%)
CO2 (%)
COS (%)
SO2 (%)
OCV (V)
coal gas coal gas þ D2O coal gas þ CO2 coal gas coal gas þ D2O coal gas þ CO2 coal gas coal gas þ D2O coal gas þ CO2 coal gas coal gas þ D2O coal gas þ CO2
700 700 700 800 800 800 900 900 900 950 950 950
88.29 86.16 22.77 89.48 86.26 32.5 79.77 78.82 33.0 48.8 67.24 25.2
1.06 5.38 26.9 7.52 8.09 43.5 19.6 18.6 61.1 41.9 28.44 70.0
4.76 4.24 2.82 2.27 4.07 0.69 0.56 1.07 0.5 1.42 1.72 0.54
0.88 1.11 48.54 0.711 1.56 23.24 0 1.42 4.4 8.6 2.58 3.05
1.06 1.06
1.00 1.00
0.99 0.96 0.81 0.90 0.88 0.85 0.75 0.81 0.73 0.76 0.77 0.64
Figure 6. MS profiles of the effluents of the coal pyrolysis reactor at 700, 800, 900, and 950 °C temperature segments.
from pyrolysis of coals at 700 °C (146 kJ/mol for H2, 132 kJ/ mol for CH4, 88-96 kJ/mol for CO, and 59-63 kJ/mol for CO2).26,27 The activation energies of H2 and CH4 were higher than those of CO and CO2, indicating that dehydrogenation and dealkylation reactions during coal pyrolysis occur through a higher activation energy pathway than oxidation reactions in coal. This could be due to the immediate availability of oxygen ions on the surface of the coal to oxidize the carbon. An active dehydrogenation and dealkylation catalyst would be necessary to enhance the formation of H2 and CH4 from coal pyrolysis. Figure 6 shows the MS profiles of the coal gas produced at 700, 800, 900, and 950 °C. Increasing the temperature from 700 to 950 °C led to (i) a decrease in the formation of H2, CH4, and CO2 and (ii) an increase in the formation of CO, which are consistent with the observations on IR. The observed trend in the composition of the coal gas can be explained by the thermochemical conversion reactions of coal, where the evolution of products is controlled by the degree of bond breaking and bond formation, cross-linking reactions, and chemical moiety rearrangements among the species present. The occurrence of dehydrogenation and molecular rearrangement reactions during coal pyrolysis resulted in the release of H2, CH4, CO, CO2, and sulfur compounds and the formation of a liquid-phase pitch product at 450 °C. The pitch product consists of dimers, trimers, and higher oligomers (Mw ∼ 1000) formed by polymerization and condensation reactions of the chemical constituents present in coal. Further heating of the pitch product to 950 °C resulted in the formation of coke. This process has been
suggested to occur by the following pathway: (i) collision and coalescence of the liquid droplets in the pitch, forming a complete anisotropic liquid phase, and (ii) continued polymerization and solidification of the anisotropic liquid to produce coke with a polymeric structure.28 The MS results for the addition of D2O and CO2 into the coal gas stream shown in Figure 6 will be further discussed. Performance of Coal Gas SOFC. Figure 7a shows the voltage-current performance (V-I curves) recorded at 950 °C during exposure of the Cu-anode SOFC to the coal gas generated at 700, 800, and 950 °C. Figure 7b presents the SOFC performance in H2 and CH4 for comparison purposes. The fuel cell OCV and maximum current density, which reflect the ability of the cell to convert the chemical energy of the coal gas into electricity, decreased gradually as the temperature of the pyrolysis reactor was increased from 700 to 950 °C, and the concentration of CH4 and H2 in the coal gas declined, while that of CO increased. This result suggests that the Cu anode exhibits a higher activity for the electrochemical oxidation of CH4 and H2 than that for COrich coal gas. A separate study on a Ni-anode fuel cell with a similar Pt current collector/electrolyte/cathode as the Cuanode fuel cell showed significantly low electrochemical oxidation activity of CH4, ruling out the contribution of the Pt current collector to the catalytic activity of the Cu anode, in agreement with previous studies.29 The addition of D2O/He and CO2/He to the pyrolysis reactor at 700 °C decreased the fuel cell OCV from 0.99 to 0.96 and 0.82 V, respectively. The maximum current density of the fuel cell also decreased from 330 to 320 mA/cm2 after
(26) Porada, S. Fuel Sci. Technol. 2004, 83 (9), 1191–1196. (27) Schuchardt, U.; Schusterman de Cencig, E. G. Proc. Int. Conf. Coal Sci. 1985, 925–928.
(28) Singer, L. S.; Lewis, I. C. Appl. Spectrosc. 1982, 36 (1), 52–57. (29) Park, S.; Vohs, J. M.; Gorte, R. J. Nature 2000, 404 (6775), 265– 267.
1180
Energy Fuels 2010, 24, 1176–1183
: DOI:10.1021/ef9009636
Singh et al.
Figure 8. MS profile of HD (m/e = 3) release during heating of the coal pyrolysis reactor from 700 to 800 °C, from 800 to 900 °C, and from 900 to 950 °C.
Figure 6 indicate that the formation of HD (m/e = 3) was not observed after D2O exposure at 700 °C. However, heating the reactor from 700 to 800 °C resulted in the appearance of HD species, shown by the positive profile on MS in Figure 8. This result suggests that HD species were indeed produced during exposure to D2O and were adsorbed in the porous structure of coal. The evolution of HD at 800 °C confirm the occurrence of reforming type reactions at 700 °C, although the rate of desorption of isotopic product (HD) was negligible and required heating of the pyrolysis reactor to detect it. Coal gas mixed with D2O at 700 °C generated a maximum current density of 310 mA/cm2 at 0.48, which decreased to 52 mA/cm2 at 0.1 V when the temperature was raised to 950 °C. This decrease in maximum current density is accompanied by an increase in the concentration of CO and decrease in the concentration of H2 and CH4 in the coal gas. Exposure of CO2 to Coal Gas. The addition of Ar/CO2 (37/3 cm3/min) into the inlet stream at 700, 800, 900, and 950 °C, shown in Figure 6, was performed to probe the effect of CO2 on the coal pyrolysis reaction and the performance of the fuel cell in a concentrated CO2 environment. The OCV and maximum current density for CO2 concentrated coal gas feed decreased with temperature from 700 to 950 °C. The introduction of CO2 flow resulted in the formation of CO (m/e = 28) by the reforming reaction of carbon dioxide with carbon and/or hydrocarbons, i.e., CO2 þ C f 2CO.33 Performance of the Cu Anode after Exposure to Coal Gas. The activity of the Cu anode after exposure to coal gas with a high CO2 concentration and 1.0% sulfur (shown in Figure 7b) was evaluated by flowing a 25% CH4/Ar stream and producing an OCV of 1.03 V and a maximum current density of 340 mA/cm2 at 0.6 V. This result further confirms that the decrease in the current density produced by the fuel cell in coal gas generated at 800-950 °C was due to the reduction in the H2 and CH4 concentrations. The absence of a strong sulfur-inhibiting effect can be attributed to the dominance of CuO over CuS crystallites, as shown in Figure 3c. The formation of CuS is very limited, even though the Cu anode has been exposed to coal gas in the entire period (i.e., 350 min) of this study. Copper sulfides can be formed even during exposure to low sulfur concentrations
Figure 7. Voltage-current polarization plot for (a) coal gas fuel cell and (b) 25% CH4 after coal gas exposure and 25% H2 at 950 °C.
the addition of D2O and to 270 mA/cm2 after addition of CO2. The decrease in the OCV and maximum current density resulted from the lower concentration of H2 and CH4 during the addition of D2O and CO2. The adition of CO2 produced CO, as shown by the MS profile of Figure 6, generating an OCV close to the theoretical Nernst potential of CO electrochemical oxidation30 and a maximum current density significantly lower than those of pure coal gas or coal gas with D2O. The maximum achievable current density (i.e., current density at 0 V) at 950 °C was estimated to be 2900 mA/cm2 by neglecting the electrode polarization and plotting a hypothetical I-V curve from an OCV of 1 V with a slope of 0.34 Ω cm2. This value of area-specific resistance (ASR) was calculated from a 1 mm YSZ electrolyte, with a YSZ conductivity of 3 10-1 S/cm at 950 °C.31,32 The lower current densities observed in the coal gas experiment (i.e., 350-400 mA/cm2 at a voltage of ∼0.5 V) with respect to the maximum achievable current density indicates that the current fuel cell has not reached its full potential and further improvements on the Cu-anode structure and composition are needed. Exposure of D2O to Coal Gas Stream. The addition of Ar/ D2O (3 vol % in Ar) to the coal pyrolysis reactor at 700, 800, 900, and 950 °C in Figure 6 was performed to examine the reforming reaction of D2O with coal and coal gas and its impact on the performance of the fuel cell. The MS profiles in (30) Chuang, S. S. C. Catalysis 2005, 18, 186–198. (31) Mauvy, F.; Lenormand, P.; Lalanne, C.; Ansart, F.; Bassat, J. M.; Grenier, J. C. J. Power Sources 2007, 171 (2), 783–788. (32) Sammes, N. M.; Cai, Z. Solid State Ionics 1997, 100 (1,2), 39–44.
(33) Froment, G. F.; Bischoff, K. B. Chemical Reactor Analysis and Design; John Wiley and Sons: New York, 1979; p 800.
1181
Energy Fuels 2010, 24, 1176–1183
: DOI:10.1021/ef9009636
Singh et al.
Figure 9. DRIFT spectra of Ohio #5 coal and coke. The IR intensity of coke has been enhanced 7 times from the original.
below the parts per million order.34 Owing to the semi-batch operation of the pyrolysis reactor, exposure to small concentrations of sulfur would not cause complete anode deactivation, as reported in previous studies.35 XRD analysis carried out before and after the coal gas fuel cell experiment revealed that the Cu anode was composed of CuO crystallites with grain sizes in the range of 30 nm. These CuO grain sizes are in agreement with those reported in a previous study of a Cu-YSZ anode prepared for the electrochemical oxidation of high-purity H2 fuel, which produced crystallites in the range of 34-44 nm.36 Although the electrochemical oxidation of CH4 and toluene has also been studied on these Cu anodes, no results have been reported on the effect of Cu grain size on the fuel cell performance.10 A comparison of the grain size of the Cu anode to those Cu catalysts that have been reported to be highly active for CO oxidation37 suggests that small particle sizes (i.e., ∼5 nm) may be required to increase the oxidation activity of the fuel cell anode. DRIFT Analysis of Ohio #5 Coal and Coke. Figure 9 shows the single-beam DRIFT spectra of Ohio #5 coal and coke produced after the pyrolytic reaction of coal at 950 °C. The assignment of the characteristic bands present in the IR spectra of coal is well-established in the literature, and similar bands were observed for Ohio #5 coal. The characteristic IR absorptions show the presence of saturated C-H, aromatic C-H, aromatic and polyaromatic CdC, oxygenates, m- and p-substituted benzenes, and sulfur
compounds in the coal.38-41 The substituted benzenes are a class of organic compounds derived from benzene, where one or more of the hydrogen atoms are replaced by other organic functional groups, for example, phenol, toluene, and styrene. The IR spectrum of coal is expected to vary with the volatile, ash, organic, and inorganic matter contents. The examination of the IR spectrum of Ohio #5 coal indicates the presence of a large amount of volatile organic matter and the absence of alumino-silicate species at ∼ 700 cm-1. The alumino-silicate species are characteristics of a high ash content coal. Similar IR spectra have been reported for coal with a significant amount of organic matter in comparison to inorganic ash content.42 The coal pyrolysis process resulted in the removal of (i) hydroxyl bands, (ii) C-H bands of paraffins/olefins/ aromatics, and (iii) CdC bands of aromatics in the IR spectrum of coke. These changes suggest that coal pyrolysis resulted in breaking of C-C, CdC, C-H, and O-H bonds, leading to the formation of H2, CH4, CO, CO2, and H2O. The IR bands at 750-1000 cm-1 (C-O, C-S, and S-O) in coke are more intensified than saturated and unsaturated hydrocarbon regions.43 This observation indicates that the pyrolysis of coal produced coke-containing oxygenates (thioethers) and sulfonates on its surface. Furthermore, the absence of aliphatic and aromatic CH vibrations at 28003100 cm-1 in coke implies that C-H bands present in the aromatic and aliphatic compounds in coal are completely removed by dealkylation, dehydrogenation, oxidation, and cross-linking reactions during pyrolysis.
(34) Kim, H.-T.; Kim, S.-M.; Jun, K.-W.; Yoon, Y.-S.; Kim, J.-H. Int. J. Hydrogen Energy 2007, 32 (15), 3603–3608. (35) McIntosh, S.; Vohs, J. M.; Gorte, R. J. J. Electrochem. Soc. 2003, 150 (4), A470–A476. (36) Grgicak, C. M.; Green, R. G.; Giorgi, J. B. J. Mater. Chem. 2006, 16 (9), 885–897. (37) Tu, C.-H.; Wang, A.-Q.; Zheng, M.-Y.; Wang, X.-D.; Zhang, T. Appl. Catal., A 2006, 297 (1), 40–47. (38) Bandopadhyay, A. K.; Sen, R. Fuel Sci. Technol. 1998, 17 (1), 31– 33.
(39) Painter, P. C.; Snyder, R. W.; Starsinic, M.; Coleman, M. M.; Kuehn, D. W.; Davis, A. Appl. Spectrosc. 1981, 35 (5), 475–85. (40) Sobkowiak, M.; Painter, P. Energy Fuels 1995, 9 (2), 359–363. (41) Solomon, P. R.; Carangelo, R. M. Fuel Sci. Technol. 1988, 67 (7), 949–959. (42) Ritz, M. Vib. Spectrosc. 2007, 43 (2), 319–323. (43) Prauchner, M. J.; Pasa, V. M. D.; Otani, C.; Otani, S. Energy Fuels 2001, 15 (2), 449–454.
1182
Energy Fuels 2010, 24, 1176–1183
: DOI:10.1021/ef9009636
Singh et al. Table 3. Capital Cost of Electricity Generation Technologies technology
size (MW)
capital costs in 2008 ($/kW)
coal-fired power plant coal gas fuel cell approach integrated gasification fuel cell (IGFC) integrated coal-gasification combined cycle (IGCC) conventional combustion turbine
600 100 100
2058 1996 1800
380
2378
160
670
dehydrogenation reactions of hydroaromatics, dealkylation of aromatics, and cross-linking polymerization reactions during coal pyrolysis. Coal Gas Fuel Cell Capital Costs. The economic feasibility of a full size coal gas fuel cell can be evaluated by considering (i) the electrochemical conversion of coal gas in the SOFC and (ii) the thermochemical conversion of the residual coke (i.e., conversion of the coke residue from coal pyrolysis in a conventional coal-fired power plant). Assuming the similar performance of the coal gas fuel cell and other conventional fuel cells, as suggested by the H2 and coal gas I-V curves in Figure 7, and comparable characteristics between the pyrolysis reactor and fuel cell gasification/reforming units, it can be inferred that the capital cost of the electrochemical conversion of coal gas would closely resemble that of the IGFC systems shown in Table 3.51,52 Following an analogous argument, the capital costs of the coke thermochemical conversion can be assumed as those reported for conventional coal-fired power plants in Table 3. Thus, the overall capital cost of the coal gas fuel cell would lie between that of IGFC and conventional power plants.
Figure 10. Raman spectra of Ohio #5 coal and coke.
Raman Characterization of Ohio #5 Coal and Coke. Figure 10 shows the Raman spectra of Ohio #5 coal and coke in the region consisting of bands at (i) 1590 cm-1 (G band) assigned to ring breathing vibrations of aromatics, (ii) 1354 cm-1 (D band) assigned to C-C vibrations between aromatic rings and aromatics composed of six-fused benzene rings or more, and (iii) 2200-3400 cm-1 assigned to C-C vibrations characteristic of poorly ordered graphitic structures.44-46 The variations in Raman wavenumber shifts, relative band intensities, and band intensity ratios for several carbonaceous materials, including graphite, pyrocarbons, single-crystal graphite, carbon fibers, and glassy carbons, have been related to structure changes during the pyrolysis process.47-49 In particular, the ratio of the intensity of the D and G bands (ID/IG), which increased from 0.66 for Ohio #5 coal to 0.92 for the coke byproduct, indicates a relative increase in the concentration of aromatics with at least six benzene rings, in agreement with previous gasification studies of the Victorian coal.50 An increase in the concentration of fused benzene rings in coke is the outcome of
Conclusions Coal pyrolysis produced H2- and CH4-rich coal gas that yielded a maximum current density comparable to that produced by a 25% CH4 in He feed. The presence of SO2 and COS did not cause the degradation of the Cu anode. The addition of D2O to the pyrolysis reactor led to the formation of HD; CO2 addition led to the formation of CO, revealing the occurrence of reforming reactions and yielding a CO-rich coal gas. Because of the presence of a large CuO particle on the anode, the performance of the fuel cell decreased when the CO-rich coal gas was fed. IR and Raman spectra revealed that the coke resulting from pyrolysis contains C-C, C-O, S-O, and C-S bond structures. Coal pyrolysis coupled to a Cu-anode SOFC constitutes a technically feasible approach for direct electric power generation with potentially valuable coke byproducts.
(44) Beyssac, O.; Goffe, B.; Petitet, J.-P.; Froigneux, E.; Moreau, M.; Rouzaud, J.-N. Spectrochim. Acta, Part A 2003, 59A (10), 2267–2276. (45) Johnson, C. A.; Thomas, K. M. Fuel Sci. Technol. 1987, 66 (1), 17–21. (46) Li, X.; Li, C.-Z. Fuel Sci. Technol. 2006, 85 (10-11), 1518–1525. (47) Dong, S.; Alvarez, P.; Paterson, N.; Dugwell, D. R.; Kandiyoti, R. Energy Fuels 2009, 23 (3), 1651–1661. (48) Ferrari, A. C.; Robertson, J. Phys. Rev. B: Condens. Matter Mater. Phys. 2000, 61 (20), 14095–14107. (49) Matthews, M. J.; Pimenta, M. A.; Dresselhaus, G.; Dresselhaus, M. S.; Endo, M. Phys. Rev. B: Condens. Matter Mater. Phys. 1999, 59 (10), R6585–R6588. (50) Li, X.; Hayashi, J.-i.; Li, C.-Z. Fuel Sci. Technol. 2006, 85 (12-13), 1700–1707.
Acknowledgment. The authors acknowledge the financial support received for this research from the Department of Energy (DE-FC36-06GO86055), Ohio Coal Development Office, and FirstEnergy Corporation. (51) http://www.eia.doe.gov/oiaf/aeo/assumption/pdf/electricity.pdf (accessed on Oct 28, 2009). (52) http://www.netl.doe.gov/technologies/coalpower/fuelcells/publications/IGFCPerformance.pdf (accessed on Oct 30, 2009).
1183