ARTICLE pubs.acs.org/Langmuir
SO2 Adsorption on Carbon-Supported Pt Electrocatalysts K. Punyawudho,*,† J. R. Monnier,‡ and J. W. Van Zee‡ † ‡
Department of Industrial Chemistry, Chiang Mai University, Thailand 50200 Department of Chemical Engineering, University of South Carolina, Columbia, South Carolina 29208 ABSTRACT: Adsorption of SO2 on a Pt/C catalyst typically used in proton exchangemembrane fuel cells (PEMFCs) has been investigated by temperature programmed desorption (TPD). SO2 concentrations in N2 were varied from 5 ppm to 1% (vol) and adsorption isotherms were determined at 25, 50, and 80 °C. Oxygen assisted (O-assisted) desorption experiments (i.e., successive TPD experiments following exposure to room temperature O2 after the first TPD event) produced an additional SO2 peak at a temperature higher than the initial SO2 peak. These two types of SO2 adsorption were identified as weakly adsorbed SO2 species desorbed between 140 and 200 °C, depending on concentration, and a strongly adsorbed, dissociated species. For the strongly adsorbed, dissociative species, 18O2 isotope introduction during O-assisted desorption yielded ratios of 50%, 36%, and 14% for SO2 masses of 64, 66, and 68, respectively. The activation energy and kinetic constant of desorption are reported for weakly adsorbed SO2 at 1% and 20 ppm SO2 using the Polanyi-Wigner equation.
’ INTRODUCTION The effect of air contaminants on fuel cell electrocatalysts has received much attention recently1-5 but fundamental data for adsorption isotherms are lacking as reviewed by St-Pierre.6 SO2 is one contaminant in air that has caused significant loss of performance and deactivation of PEMFCs,7-10 and the effect of SO 2 strongly depends on concentration and dosage. 7,8 Mohtadi et al.7 observed that partial recovery was obtained after applying contaminant-free air for 24 h. Jing et al.10 have recently observed that the cell potential decreased from 0.68 to 0.44 V at constant current density of 500 mA/cm2 when exposed SO2 concentrations as low as 1 ppm. The objective of this paper is to provide fundamental data for PEMFCs electrocatalysts that might be used to develop quantitative descriptions of these performance losses. Some background for carbon-supported Pt catalysts can be obtained from single crystal studies of Pt without the carbon support. For example, Zebisch et al.11 have used Pt(110)1 2 and concluded that the only SOx (x = 1 - 4) species capable of gas phase desorption is SO2 in agreement with the TPD results of Wilson et al.12 Further, Zebisch et al.11 showed that SO2 desorbed over a range of temperatures (100-800 K) with different desorption energies, particularly for surface coverage of SO2 > 0.5. In the same study, XPS was used to determine the S binding energies (i.e., S2p1/2 and S2p3/2) for SO2 coverages of 0.5 and 1.0 on Pt. These energy values indicated S, SO2, SO3, and SO4 were adsorbed depending on the temperature of SO2 exposure. At 140 K, only SO2 existed on the Pt surface; at 250 K, the S (S2p1/2 = 162.2 eV) binding energies indicated onset of SO2 dissociation, and at 340 K, both SO2 and SO3 were detected. For 500 K and θSO2 ≈ 0.5, SO3 became SO4, and residual S-containing species existed at 650 K. r 2011 American Chemical Society
Addition information on surface species, obtained with numerical simulation, have been reported. McKee13 used computational method and determined that SO4 is not thermodynamically stable at temperature above 150 K in the gas phase relative to its dissociation energy. Sellers and Shustorovich14 employed the bond order conservation-Morse potential (BOC-MP) method for a LangmuirHinshelwood reaction to analyze sulfur oxide chemistry for low surface coverages and found that the activation barrier of SO3 decomposition (SO3 f SO2 þ O) was lower than the barriers for desorption of SO3 for all metals on fcc(111) surface, including Pt. Thus, this decomposition prior to desorption is consistent with only SO2 being detected during TPD experiments. Sun et al.15 have studied multilayer SO2 adsorption on Pt(111) at temperatures from 100 to 300 K. Both XPS and UPS revealed that there was no dissociation at 100 K, but upon heating to 300 K, some SO2 desorbs while other species undergo dissociation to form adsorbed S, SO, and SO4. Even after flash heating the Pt(111) surface to 1000 K, some sulfur species remained on the Pt surface. Lin et al.16 have used first-principle DFT-GGA calculations to study the configuration of chemisorbed SO2 on Pt (111) surfaces. They found that η2-Sb,Oa and η3-Sa,Oa,Oa on fcc surfaces are the most energetically stable configurations, where η2 and η3 correspond to two and three surface atoms coordinated with the adsorbate, respectively, and the subscripts a and b indicate the atoms on atop and bridging Pt sites, respectively. These configurations are consistent with experimental observation discussed above. For the carbon support, it has been reported that activated carbon adsorbs SO2 during the cleaning of exhaust gases.17-20 Thus, a brief review of the adsorption of SO2 on activated carbon Received: November 22, 2010 Published: February 04, 2011 3138
dx.doi.org/10.1021/la104637w | Langmuir 2011, 27, 3138–3143
Langmuir is useful for understanding the Pt/C electrocatalyst results. Zawadzki’s observation18 of weak and strong adsorption for SO2 adsorption on carbon was confirmed by others.21-24 Zawadzki25 also used IR spectroscopy to infer interactions between O2 and H2O with adsorbed molecules of SO2, and concluded that the amount of SO2 adsorbed on carbon is greatly increased in the present of O2 and H2O.25-27 However, the coadsorption of SO2 and O2 on carbon decreased the amount of SO2 because of the competitive adsorption of SO2 and O2 on the carbon surface.24,25 Dratwa et al.28 proposed a mechanism of SO2 adsorption on carbon surfaces in the presence of oxygen and water, whereby the SO2 is converted to sulfuric acid between 380 and 440 K. They also proposed carbon regeneration at temperatures above 600 K, where sulfuric acids react with carbon releasing SO2, CO2, and water. In contrast to the above work on unsupported crystalline Pt and on carbon, the data reported here correspond to the supported Pt/C catalyst used in PEMFCs. To initiate the understanding of SO2 adsorption in PEMFCs, which produce water and consume O2, we consider SO2 adsorption on the supported catalyst from dry N2 mixtures. Further, we seek to understand any dissociative adsorption species through 18O2 assisted desorption experiments. We also seek to understand how kinetic constants and activation energies change with concentration, since contamination of PEMFCs cathode may occur with very dilute concentrations and small dosages.
ARTICLE
Figure 1. Comparison TPD for 10% Pt/C (—) and carbon Vulcan XC72 ( 3 3 3 ) for 166 ppm SO2/balance in N2 adsorbed at temperature of 25 °C.
’ EXPERIMENTAL SECTION Commercially available 10 wt % Pt/C (E-TEK) sample sizes of 0.1 g were used for all SO2 desorption experiments. Each sample was placed in a 4-mm ID quartz tubular reactor that was mounted in a custom split tube furnace (model 3210 by ATS Systems) capable of heating all samples at linear rates up to 50 °C/min. Before the temperature programmed desorption (TPD) experiments were conducted, all samples were pretreated in situ by heating at 300 °C in 10% H2/balance Ar (UHP grade) for 1 h, followed by an additional hour at 300 °C in flowing Ar. Samples were then cooled to the desired temperature (25, 50, or 80 °C) for the SO2 adsorption experiments. Mass flow controllers (Brooks 5850E) were used to reach equilibrium coverage (typically 30 min) with five concentrations of SO2 from premixed cylinders of SO2/balance N2 (UHP grade) (i.e., 5 ppm, 20 ppm, 160 ppm, 0.2%, and 1% SO2) on the catalyst surface. The SO2containing gas mixture was replaced with a gas flow of 30 SCCM Ar as a sweep gas. The Ar sweep gas was introduced into a residual gas analyzer (INFICON Transpector 2) by a variable leak valve (model ULV-075, by MDC Vacuum Products) and after baselines were attained for the peaks of interest, the TPD experiment was initiated. The ramp rate in all cases was 20 °C/min from 25 - 550 °C. The mass intensities typically monitored were m/e = 18 (H2Oþ), 28 (COþ), 32 (O2þ), 44 (CO2þ), 64 (SO2þ), and 80 (SO3þ). Temperatures were recorded using an Omega data logger (Type K, model HH306). All mass intensity and temperature data were stored and analyzed using proprietary T32 software by INFICON. To investigate other Pt-sulfur species that might exist, an additional O2 treatment was performed immediately after the initial TPD experiment was concluded. The sample was cooled in flowing Ar to 25 °C and 10% O2/balance in He (UHP) was flowed over the catalyst for 30 min to saturate the Pt surface with adsorbed oxygen (e.g., Pt-O surface species). After O2 saturation and an Ar purge to remove residual O2, the TPD experiment in 30 SCCM Ar was repeated over the same temperature range. The sequence of Pt-O formation (by O2-pretreatment of the Pt surface) followed by TPD was repeated three more times until no additional SO2 was detected. The low temperature O2 pretreatment step was used to prevent Pt-catalyzed oxidation of the carbon support at the higher temperatures of the TPD experiment, since extensive oxidation of the carbon support at T > 250 °C
Figure 2. TPD of 0.2% SO2/balance in N2 on Pt/C for an adsorption temperature of 25 °C; desorption after SO2 exposure (—), first O-assisted desorption sequence (- 3 3 -), second O-assisted desorption sequence ( 3 3 3 ), third O-assisted desorption sequence (- 3 - ), fourth O-assisted desorption sequence (- - -). was observed with the standard temperature programmed oxidation experiment. The mole fraction of SO2 desorbed during TPD was determined by calibrating the response of the RGA (Residual Gas Analyzer) at known partial pressures. Three different leak rates were used to determine a response factor (amps SO2þ/torr SO2) for each concentration of SO2. Integration of the SO2 peak [mass intensity vs time (function of temperature)] from the TPD curve and application of the SO2 response factor and sweep gas flow rate permitted the determination of the number of μmoles SO2 that desorbed from the Pt surface. Finally, the total concentration of Pt surface sites were determined to be 11.7 μmol Pt at an average particle size of 5.8 nm for the 0.102 g sample in a separate experiment by a Micromeritics 2920 Pulse Chemisorpion Analyzer using standard H2 chemisorption protocols. Consequently, this surface site is used to compute the coverage of SO2 adsorption in this work.
’ RESULTS AND DISCUSSIONS Thermal Desorption Results. An example of an SO2 TPD experiment for 10% Pt/C catalyst is illustrated in Figure 1 for 166 ppm SO2. Similar graphs for other concentrations are consistent with Figure 1 and can be found elsewhere.29 The possible species monitored during desorption were CO, SO2, SO3, and O2 but only SO2 (i.e., 64 amu) was detected during desorption. Figure 1 shows that the adsorption of SO2 on Pt/C catalyst is much greater than on carbon (Vulcan XC-72) and that the desorption temperatures are distinct (i.e., 95 °C for carbon and 140 °C for Pt/C). Thus, the adsorption of SO2 on carbon (Vulcan XC-72) can be neglected 3139
dx.doi.org/10.1021/la104637w |Langmuir 2011, 27, 3138–3143
Langmuir
ARTICLE
Figure 3. TPD of SO2 and O-assisted desorption: (a) 0.2% SO2 at Tad = 25 °C, (b) 0.2% SO2 at Tad = 80 °C, (c) 20 ppm, Tad = 25 °C, and (d) 20 ppm, Tad = 80 °C. Desorption after SO2 exposure (—), first O-assisted desorption sequence (- 3 3 -), second O-assisted desorption sequence ( 3 3 3 ), third O-assisted desorption sequence (- 3 -), and fourth O-assisted desorption sequence (- - -).
when compared to the Pt/C catalyst at the same SO2 exposure and temperature. This similar result was observed by Garsany et al.5 when Pt/C and Vulcan XC-72 were examined by XPS after exposure of S(IV) in 0.1 M HClO4 solution at Eads = 0.65 V. Further, we did not observe a distinct peak of SO2 during TPD of O-assisted desorption for Vulcan XC-72. Figure 2 shows the results of SO2 desorption and O-assisted SO2 desorption for a Pt/C catalyst exposed to 0.2% SO2. There are two distinct peaks at 140 and 300 °C. The peak at 140 °C represents SO2 desorption from Pt surface, while the four higher temperature peaks result from four consecutive applications of O2 to the SO2 exposed Pt/C sample. The O-assisted desorptions have peaks centered at ∼300 °C. Thus, we observe two types of surface bonding corresponding to weak and strong adsorption, consistent with the single crystal Pt (111) work of Astegger and Bechtold.30 The more weakly bound species desorbs at the lower temperature to yield SO2(g) and Pt sites, while the more strongly bonded species desorbs with O-assisted desorption at the higher temperature. This implies that sulfur (S) or sulfur oxide (SO) remains on the Pt surface, as discussed below. There is a shift in the temperature of the weakly bonded species from 140 to 180 °C when the adsorption temperature is changed from 25 to 80 °C at 0.2% SO2 exposure as shown in Figure 3a and b. This shift is not observed for adsorption of 20 ppm SO2, as shown in Figure 3c and d. There is some broading of this first peak observed for the 0.2% SO2 that is not shown in Figure 3c and d. On the other hand, there is no shift in temperature for the peaks from O-assisted desorption for all graphs in Figure 3. This indicates that the energy for the strong adsorption is independent of the adsorption temperature and gas phase concentration of SO2 and therefore corresponds to an adsorption site independent of the initial adsorption kinetics. Note that the intensities of
Figure 4. Isotherms for weak adsorption of SO2 on 10% Pt/C, as a function of adsorbed temperature. Legend: 1 = 80 °C, O = 50 °C, and b = 25 °C temperature of adsorption.
Figure 3a and b can not be compared with those of Figure 3c and d because of the use of different response factors and leak rates. The concentration dependence on adsorption was determined by integrating TPD data versus time for the given response factor, and the moles of adsorbed SO2 on Pt/C were calculated for each of the five concentrations (i.e., 1%, 0.2%, 166 ppm, 20 ppm, and 5 ppm). Hydrogen chemisorption of the 10% Pt/C sample gave the surface concentration of Pt sites, which was determined to be 11.7 μmol. From this value the surface coverage of SO2 was determined using the assumption of one SO2 to one Pt surface site. Thus, the coverages at 25, 50, and 80 °C are shown in Figures 4 and 5 for weak and strong adsorption, respectively. Figure 4 indicates the isotherm for weak adsorption is a function of adsorption temperature and concentration. At the same concentration, higher adsorption temperatures give lower coverage. Conversely, at similar adsorption temperatures, SO2 coverage increases 3140
dx.doi.org/10.1021/la104637w |Langmuir 2011, 27, 3138–3143
Langmuir
ARTICLE
Figure 5. Isotherms for strong adsorption of SO2 on 10% Pt/C, as a function of adsorbed temperature. Legend: 1 = 80 °C, O = 50 °C, and b = 25 °C temperature of adsorption.
Table 1. Total Coverage of SO2 on Pt/C as a Function of Adsorption Temperature and SO2 Concentrationa θSO2
θSO2
adsorption
concentration
(weak)
(strong)
total
temp (°C)
(ppm)
ML
ML
ML
5 20
0.02 0.08
0.10 0.13
0.12 0.21
166
0.18
0.28
0.46 0.57
25
50
80
2000
0.23
0.34
10000
0.24
0.35
0.60
5
0.01
0.11
0.11
20
0.06
0.14
0.19
166
0.14
0.21
0.36
2000 10000
0.21 0.20
0.28 0.30
0.49 0.50
5
0.01
0.10
0.11
20
0.05
0.12
0.17
166
0.11
0.16
0.27
2000
0.16
0.21
0.37
10000
0.16
0.23
0.38
a
The error limits are not shown in the table, but in all cases error was
