Analysis of sulfur poisoning on a PEM fuel cell electrode

Abstract

The extent of irreversible deactivation of Pt towards hydrogen oxidation reaction (HOR) due to sulfur adsorption and subsequent electrochemical oxidation is quantified in a functional polymer electrolyte membrane (PEM) fuel cell. At 70 °C, sequential hydrogen sulfide (H₂S) exposure and electrochemical oxidation experiments indicate that as much as 6% of total Pt sites are deactivated per monolayer sulfur adsorption at open-circuit potential of a PEM fuel cell followed by its removal. The extent of such deactivation is much higher when the electrode is exposed to H₂S while the fuel cell is operating at a finite load, and is dependent on the local overpotential as well as the duration of exposure. Regardless of this deactivation, the H₂/O₂ polarization curves obtained on post-recovery electrodes do not show performance losses suggesting that such performance curves alone cannot be used to assess the extent of recovery due to sulfur poisoning. A concise mechanism for the adsorption and electro-oxidation of H₂S on Pt anode is presented. H₂S dissociatively adsorbs onto Pt as two different sulfur species and at intermediate oxidation potentials, undergoes electro-oxidation to sulfur and then to sulfur dioxide. This mechanism is validated by charge balances between hydrogen desorption and sulfur electro-oxidation on Pt. The ignition potential for sulfur oxidation decreases with increase in temperature, which coupled with faster electro-oxidation kinetics result in the easier removal of adsorbed sulfur at higher temperatures. Furthermore, the adsorption potential is found to influence sulfur coverage of an electrode exposed to H₂S. As an implication, the local potential of a PEM fuel cell anode exposed to H₂S contaminated fuel should be kept below the equilibrium potential for sulfur oxidation to prevent irreversible loss of Pt sites.

Introduction

Though extensive research had been done on the issue of CO poisoning in polymer electrolyte membrane (PEM) fuel cells, there is much less in the literature on sulfur and H₂S poisoning [1], [2], [3]. If the anode fuel is obtained by reforming hydrocarbon fuels that have sulfur (e.g., coal mined from United States [4]), they may contain up to 5 ppm H₂S even after desulfurization [5]. Uribe et al. [6], [7] showed that the poisoning effect due to H₂S on a PEM fuel cell Pt anode is cumulative and irreversible. According to them, after H₂S poisoning, total recovery with neat hydrogen was not possible, and a partial recovery was possible by a potential scan between 0 and 1.4 V vs. dynamic hydrogen electrode (DHE). Mohtadi et al. [8], [9] found that the degree of recovery of a PEM fuel cell anode (Pt) poisoned by H₂S depended on the degree of oxidation of two surface species observed in the cyclic voltammogram (CV) as distinct peaks. Further they reported that the increase in Pt loading increased the partial recovery with neat hydrogen and by a potential scan between 0 and 1.4 V vs. DHE. Loučka [10], the first to study the kinetics of H₂S adsorption and oxidation on single-crystal platinum electrodes in aqueous phase at 25 °C, found that H₂S became completely dehydrogenated on adsorption, and that the hydrogen thus formed became anodically oxidized at positive electrode potentials. Also, the charges for oxidation of adsorbed sulfur were too high to account for the oxidation of a mere monolayer of adsorbed sulfur. This was later explained by the formation of a poorly reducible oxide on the electrode and not due to the presence of multiple layers of adsorbed sulfur atoms. Further, according to Loučka, complete removal by oxidation of adsorbed sulfur could not be attained by holding the poisoned Pt electrode at 1.6 V vs. DHE unless the degree of S coverage on the electrode was very low. Complete oxidation was reached only by periodic polarization to such positive electrode potentials. Loučka proposed the following reactions:

$$\text{Pt}\text{S}+3{\text{H}}_2\text{O}\rightleftharpoons \text{S}{\text{O}}_3+6{\text{H}}^{+}+6{\text{e}}^{-}+\text{Pt}$$

$$\text{Pt}\text{S}+4{\text{H}}_2\text{O}\rightleftharpoons \text{S}{\text{O}}_4^{2-}+8{\text{H}}^{+}+6{\text{e}}^{-}+\text{Pt}$$

Najdekar and Bishop [11] attributed the formation of the poorly reducible oxide to the sulfidation of Pt electrode. They attributed the large oxidation peak in the 1.25–1.42 V vs. standard hydrogen electrode (SHE) range to oxidation of platinum. They compared the oxidation and the reduction charges of each cycle of the cyclic voltammogram, and postulated that platinum oxide reacted with sulfur released at the electrode surface with the regeneration of sulfide. Using potentiodynamic oxidation at elevated temperatures (i.e., >60 °C), Contractor and Lal [12] demonstrated the presence of two forms of chemisorbed sulfur distinguished by the number of platinum sites occupied per sulfur atom. Based on electrons per site (eps) calculations, they attributed the first peak to the oxidation of linear-bonded sulfur, and the second peak to the oxidation of bridge-bonded sulfur. Lamy-Pitara et al. [13] also confirmed this presence of two forms of chemisorbed sulfur. They reported that the adsorption of sulfur was sensitive to the nature of the platinum surface. While, one sulfur atom covered one Pt atom when H₂S was adsorbed on a smooth Pt atom in zero valence state, the charge of the adsorbed sulfur depended on the degree of its coverage on a rough platinum surface, ranging between 1.5 and 2 at low coverage to 1 at higher coverage. In another study [14], they showed that the sulfur species adsorbed on the surface of the platinum were likely to be composed of sulfur and sulfides.

Farooque and Fahidy studied the oxidation of H₂S on a rotating tripolar wiper-blade electrode in the low potential region [15] (0–0.45 V vs. SHE), and in the Tafel region [16] (0.45–1.4 V vs. SHE), and reported that at lower oxidation potentials, the anodic oxidation of H₂S followed a two-electron process to yield elemental sulfur, protons and electrons. Using likelihood approach, a statistical tool to validate the most likely model from a set of contending models, they were able to conclude that the low potential oxidation of H₂S most likely followed the mechanism given below:

$$\text{Pt}+{\text{H}}_2\text{S}\rightleftharpoons \text{Pt}–{\text{H}}_2{\text{S}}_{\text{ads}}$$

$$\text{Pt}+\text{S}{\text{H}}^{-}\rightleftharpoons \text{Pt}-\text{S}{\text{H}}_{\text{ads}}$$

$$\text{Pt}-{\text{H}}_2{\text{S}}_{\text{ads}}\to \text{Pt}-\text{S}{\text{H}}_{\text{ads}}+{\text{H}}^{+}+{\text{e}}^{-}$$

$$\text{Pt}-\text{S}{\text{H}}_{\text{ads}}\to \text{Pt}-{\text{S}}_{\text{ads}}+{\text{H}}^{+}+{\text{e}}^{-}$$

The chemical reactions (3), (4) were faster than the electrochemical reactions (5), (6). This confirmed the two-electron oxidation mechanism put forward by Loučka [10]. Also, they reported that the oxidation of H₂S at higher potentials yielded colloidal sulfur. In their experiment, the wiper-blade electrode system continuously cleaned the surface by piperidine (a selective solvent for sulfur) to remove the colloidal sulfur formed. Since the electrode was always clean for further H₂S adsorption and oxidation, sulfur was the main product both in the lower and at the higher oxidation potentials.

H₂S poisoning studies have also been done for a variety of fuel cell systems. Uribe and Zawodzinski [6] and Mohtadi et al. [8], [17] studied the H₂S poisoning effects in a PEM fuel cell system. Chin and Howard [18] investigated the poisoning effect of H₂S on the anodic oxidation of hydrogen on Pt in a 94 wt% phosphoric acid electrolyte fuel cell (PAFC) over a temperature range of 25–170 °C. They reported that the extent of H₂S poisoning decreased with increasing temperature, and increased with increasing electrode potential. Further, at sufficiently high anodic potentials, a layer of adsorbed elemental sulfur was found to form on the electrode surface, which suppressed the formation of platinum oxide at the oxygen adsorption potentials. According to Kawase et al. [19], who studied the effect of H₂S on molten carbonate fuel cells, large potential losses occurred after the cell was exposed to 5 ppm H₂S. They attributed this to the formation of SO₄²⁻ and S²⁻ on the nickel electrode.

Paál et al. [20] investigated gas phase H₂S adsorption on platinum in the presence of H₂. They were able to identify the presence of sulfide and sulfate species on the poisoned surface using XPS. Also, studies by Mathieu and Primet [21] on gas phase chemisorption of H₂S on Pt showed that H₂S adsorbed dissociatively on Pt, and that dissociation lead to adsorbed sulfur and gaseous hydrogen. While investigating the effects of sulfur poisoning on platinum supported on alumina, Chang et al. [22] found that the adsorbed sulfur induced Pt agglomeration by weakening the metal-support interaction and caused migration of Pt clusters in the process. Their observation was based on the size of Pt clusters measured before and after H₂S exposure [23].

Donini et al. [24] described an electrochemical process for decomposing H₂S to produce hydrogen and sulfur. They used a divided electrolytic cell and a mixture of H₂S and volatile basic solution as the electrolyte to produce a polysulfide solution at the anode compartment. The polysulfide solution was then distilled to produce elemental sulfur. They later extended this invention to produce sulfur directly in a gas phase electrolysis process, where H₂S is oxidized at high potentials in a composite solid polymer electrolyte (CSPE)–Pt anode at elevated temperatures (>120 °C) [25]. This is in concert with the previous studies on the dissociative nature of H₂S adsorption leading to sulfur adsorption on Pt at low electrode potentials.

More recently, Wang et al. [26] developed an amperometric H₂S sensor based on its electrochemical oxidation route on a composite Pt electrode. They found that the electro-oxidation products of H₂S depended on the local electrode potential at the time of adsorption. Using XPS, they found that the main oxidation product was elemental sulfur at lower potentials and SO₄²⁻ at higher oxidation potentials. They reported that even at higher potentials, the elemental sulfur was difficult to remove from the surface of the electrode. This finding agrees with that of Loučka [10], Najdekar and Bishop [11] and Contractor and Lal [12]. Further, they tested the durability of their H₂S sensor [27], and reported that the deposition of elemental sulfur on the composite Pt electrode was the main factor affecting the life of the sensor. However, they reported that the tolerance levels of the composite Pt electrode were significantly better than that of planar Pt electrodes and they attributed this to the highly porous nature of the former.

In summary, the following could be deduced from the literature reports on the kinetics of H₂S adsorption and oxidation, both in liquid phase (planar [10], [11], [12], [13], [14], [15], [16] and composite [24], [28], [29], [30] electrodes) and in gas phase (planar [18], [19], [20], [21], [22], [23], [24] and composite electrodes [6], [8], [9], [25], [26], [27], [28]):

• H₂S adsorption on Pt is dissociative, and results in surface adsorbed sulfur species and hydrogen. The hydrogen thus formed undergoes oxidation at positive electrode potentials.

• Sulfur adsorption might result in linear-bonded sulfur species, Pt–S, and bridge-bonded sulfur species, (Pt)₂–S. The nature of this adsorption is a strong function of temperature.

• The Pt–S or (Pt)₂–S thus formed undergoes further oxidation at high electrode potentials to yield SO₂ or SO₄²⁻.

• After several potential scans, sulfur and sulfur oxidation products are largely removed.

• In the process of H₂S and S adsorption and oxidation, a small percentage of the catalyst sites appear to become inactive.

• Some of the explanations for the irreversible loss due to H₂S poisoning are deposition of organosulfur [31], sulfur induced Pt agglomeration, sub-surface sulfur and its interaction with Pt, formation of platinum sulfides and oxides that are difficult to reduce and the migration of Pt clusters due to loss in the metal-support interaction. This deactivation of Pt sites is reported to be same regardless of the type of sulfur contamination (H₂S, SO₂ or COS) [32].

Based on the above observations, a comprehensive list of reactions H₂S, as linear and bridge-bonded species on a CSPE–Pt electrode is presented:

$$\text{Pt}+{\text{H}}_2\text{S}\rightleftharpoons \text{Pt}-{\text{H}}_2{\text{S}}_{\text{ads}}$$

$$\text{Pt}-{\text{H}}_2{\text{S}}_{\text{ads}}\to \text{Pt}-\text{S}{\text{H}}_{\text{ads}}+{\text{H}}^{+}+{\text{e}}^{-}$$

$$\text{Pt}-\text{S}{\text{H}}_{\text{ads}}\to \text{Pt}-{\text{S}}_{\text{ads}}+{\text{H}}^{+}+{\text{e}}^{-}$$

$$\text{Pt}-{\text{S}}_{\text{ads}}+2{\text{H}}_2\text{O}\to \text{S}{\text{O}}_2+4{\text{H}}^{+}+4{\text{e}}^{-}+\text{Pt}$$

where reaction (7) represents adsorption and desorption of H₂S on Pt surface, reactions (8), (9) represent the oxidation of the adsorbed H₂S resulting in sulfur adsorption, and reaction 10 represents the oxidation of adsorbed sulfur to SO₂. Note that reactions (7), (8), (9) are similar to the ones proposed by Farooque and Fahidy [15], [16] yielding elemental sulfur at low oxidation potentials where as reaction 10 results in the intermediate potentials. At higher potentials, reactions 1 and 2 could occur resulting in SO₃ and SO₄²⁻, respectively. SO₂ has been reported as a poison in a number of air-contaminant studies, where it is exposed to the cathode sides of a working fuel cell [33], [34], [35]. It has been shown that SO₂ adsorbs strongly to Pt. However, its coverage on Pt depends on the adsorption potential—for example, Punyawadho has shown that the surface coverage due to adsorption is much less when fed into the anode-gas stream compared to that of a cathode-gas stream [36].

Though all these studies discuss certain aspects of H₂S induced sulfur poisoning on a Pt electrode, there is not a thorough understanding of the mechanism and a quantitative analysis of the extent of irreversible deactivation of catalytic sites on a composite PEM fuel cell electrode reported in the literature. Therefore, the objectives of this article are: (a) to quantify the deactivation of Pt catalyst after each monolayer H₂S adsorption and subsequent electrochemical oxidation at open-circuit conditions, (b) to quantify the extent of Pt catalyst deactivation after each poisoning and recovery cycle in a PEM fuel cell operating under load, (c) to state and validate a consistent mechanism for H₂S adsorption and oxidation on a PEM fuel cell electrode (at relevant anode overpotentials) which explains the deactivation, (d) to study the effect of temperature on H₂S surface coverage, equilibrium potential for H₂S electro-oxidation and the individual oxidation rates of two types of adsorbed sulfur species and (e) to study the effect of adsorption potential on the surface coverage of sulfur. These objectives are accomplished by sequential H₂S adsorption and electro-oxidation experiments on a PEM fuel cell anode at open circuit as well as load conditions. Further, charge balances between H desorption and sulfur electro-oxidation reactions are made to quantify the nature of adsorbates, and the number of electrons transferred per Pt site.