Oscillatory Chemical Reactions in Heterogeneous Systems - American

---

J. Phys. Chem. 1996, 100, 19141-19147

19141

Oscillatory Chemical Reactions in Heterogeneous Systems: Oxidation of Hydrogen on Platinum Surface by Strong Oxidants in Aqueous Solutions Krisztina Kurin-Cso1 rgei and Miklo´ s Orba´ n* Department of Inorganic and Analytical Chemistry, L. Eo¨ tVo¨ s UniVersity, H-1518 Budapest 112., P.O. Box 32., Hungary ReceiVed: June 11, 1996; In Final Form: September 20, 1996X

The oxidation of hydrogen gas in aqueous solution by oxidants such as H2O2, Na2O2, K2Cr2O7, CrO3 + H2O, ClO2-, and ClO3- + Cl- has been found to take place in an oscillatory manner on the surface of platinum. Oscillations in the resting potential of a Pt electrode were observed without applying external voltage or current. The primary cause of the oscillations is not related to the cations or anions added to the solution. The dependence of the oscillatory behavior on the concentration of the oxidants and on the type of acid in the supporting electrolyte, the effect of added halide ions on the oscillations, and the behavior of other noble metal surfaces in the hydrogen-oxidant systems have been studied. The heterogeneous (gas-solid-liquid) H2-(Pt)-oxidant oscillators seem to represent a link between the oscillatory heterogeneous catalytic surface (solid-gas) reactions and the oscillatory electrochemical (solid-liquid) reactions.

Introduction Oscillatory chemical reactions and related temporal and spatial phenomena have been observed in many homogeneous and heterogeneous systems.1 In homogeneous gas or liquid phases the concentration of some species in the bulk oscillates, while in the heterogeneous solid-gas or solid-liquid systems the oscillations can occur at the interface, and the state of the surface changes periodically. Two types of heterogeneous oscillatory reactions may be distinguished. One is the surface-catalyzed oxidation of gases (e.g., H2, CO, NH3, small organic molecules) at high temperatures under a wide pressure regime. The other involves electrochemical oscillations during which electrodissolution/ electrodeposition of metals or oxidation/reduction of substances in an appropriate supporting electrolyte occurs in oscillatory fashion on the surface of the externally polarized metal electrodes. Excellent reviews have summarized the earlier and recent developments in the field of both the surface-catalyzed2 and the electrochemical3,4 oscillatory systems. In this paper we report on oscillations in heterogeneous systems observed under conditions which are very different from those applied for surface-catalyzed reactions and, in a sense, from the electrochemical reactions as well. In an undivided batch reactor filled with aqueous solutions of strong oxidants and equipped with a Pt electrode and a Hg/Hg2SO4/K2SO4 reference electrode, oscillations in the potential of the Pt electrode were recorded when H2 gas was introduced into the cell at room temperature without applying external electric force on the electrode pair. In each case the coverage of the Pt surface changes periodically, and no bulk oscillation occurs; therefore, the H2-(Pt)-oxidant oscillators seem to represent a link between the surface-catalyzed chemical and the potentiostatic or galvanostatic electrochemical oscillatory reactions. Experimental Section Materials. All reagents used in the experiments were of the highest purity commercially available and were applied without * To whom correspondence should be addressed. X Abstract published in AdVance ACS Abstracts, November 1, 1996.

S0022-3654(96)01714-5 CCC: $12.00

further purification. The H2O2 (Carlo Elba, 30%) contained stabilizer; therefore, parallel runs with Na2O2 (Merck, pro analysi) were also carried out. No noticeable difference between the potential vs time curves taken with H2O2 and Na2O2 was observed; however, in the Na2O2 systems the Pt electrode showed oscillatory response for a much longer time (more than 10 h compared to 2-4 h observed in H2O2 solution), indicating that the Na2O2 surely was not contaminated with any surfaceblocking impurities. Special attention was paid to the quality of the water: to remove trace amounts of anion and cation impurities which could adsorb on the Pt surface, the water was distilled twice over alkaline KMnO4 in a glass apparatus. The resistance of the water used was 18.2 Mohm‚cm. Apparatus and Methods. The reaction vessel was an undivided thermostated glass beaker of volume 150 cm3, equipped with a Pt electrode and Hg/Hg2SO4/K2SO4 reference electrode. The potential differences developed across the electrode pair were collected with a DTK 286-D-type personal computer installed with an eight-channel ADC converter (Labtech PCL-7US). This technique allowed us to analyze the oscillations even though they appeared with high frequency (e.g., about 1 oscillation/s, obtained in many compositions of the H2-(Pt)H2O2 system). The signal was also recorded on a multichannel recorder (Kipp & Zonen BD 41). The supporting electrolyte was dilute (0.1-0.001 M) sulfuric acid, and the temperature was 25 °C in most experiments. The experiments were carried out in batch and flow configurations. In the batch experiments the supporting electrolyte was placed in the beaker and saturated with H2 gas. A glass capillary tube served to introduce H2 in fine bubbles at a constant rate. The bubbles were distributed by magnetic stirring. Monitoring the potential of the Pt electrode in time indicated the extent of saturation of the solution with H2. When the potential reached its lowest value, the solution of the oxidant was added while maintaining the bubbling of H2 and recording the potential. In some cases CSTR (continuously fed stirred tank reactor) experiments were carried out. The supporting electrolyte was first saturated with H2 in a separate container and then fed into a flow-through reactor using a peristaltic pump. A second inlet served to introduce the solution of oxidant. An interesting way © 1996 American Chemical Society

19142 J. Phys. Chem., Vol. 100, No. 49, 1996 was to supply the reagent H2 internally: adding Zn powder to the acidic supporting electrolyte causes H2 gas to evolve. Oscillations in the potential then appear on adding the oxidant. (The products, Zn2+ ions, seem not to disturb the oscillatory behavior; i.e., their adsorption on the Pt is negligible.) No mechanical stirring was required in this kind of experiment, and a practically noise-free oscillatory response was obtained. The concentration of the H2 in the electrolyte solutions can be approximated by the solubility of H2 in water (8.2 × 10-4 M/dm3). Analytical determinations indicate a very small degree of consumption of the applied oxidants during the experiments; e.g., 89% of the initial concentration of H2O2 measured by KMnO4 titration was still present after a 3.5 h long run during which about 5000 oscillations had been recorded. Due to the slight consumption of the reagents during one oscillatory cycle, many peaks could be recorded even with the H2 bubbling turned off. Due to the nearly constant reagent concentrations during even long experiments, the reactor can be regarded as a semibatch arrangement, and practically sustained oscillations are expected if they appear, unless the Pt surface becomes blocked by an adsorbed inhibiting species. To obtain reproducible oscillatory traces in the above systems, it is important to apply identical physical parameters during the different runs. Care should be taken to keep relative positions of the electrodes and H2 inlet tube unchanged and to maintain the rate of H2 bubbling, the rate of the magnetic stirrer, and the temperature constant. The most reproducible results and the most noise-free oscillatory curves are obtained when the mechanical mixing in the system is minimized, i.e., when the H2 bubbling and stirring are stopped. At fixed physical parameters the dependence of the system’s behavior on its chemical composition was studied by varying the concentration of the oxidants, the pH, and the type of acid. The effect of different anions capable of being adsorbed strongly on the Pt surface was also tested. For studying the dynamical behavior of the system, the rotation rate of a Pt disk-ring electrode was chosen as a bifurcation parameter. Varying the rotation speed allows to change gradually the composition at the solution-Pt surface interface; i.e., the mass transport becomes adjustable. Standardization of the Pt Electrode. In our experiments several kinds of Pt electrodes were used, including the needle type, bright Pt plate electrodes of 1-8 cm2 geometric surface area and a rotating disk-ring electrode assembly (Beckman U.S.A.). Preliminary experiments showed that the reproducibility of the oscillatory behavior greatly depends on the state and standardization of the Pt surface. The oxidants used seem to influence the state of the Pt surface, probably forming a Pt oxide film or other kind of adsorption layer which can prevent starting the oscillations or can stop the oscillations after a number of cycles. The number of these cycles is determined by the nature of the oxidant; for example, H2O2 (and Na2O2) produced sustained oscillations almost indefinitely, while K2Cr2O7 produced only several dozen oscillations and S2O82- produced none. At the termination of the oscillations (or upon applying S2O82-, MnO4-, etc.) the potential of the Pt electrode reached a steady high value which could not be decreased by increasing the stirring rate to ensure a higher H2 input concentration. In such cases renewal of the Pt surface was necessary. Three methodssone chemical (i), one mechanical (ii), and one electrochemical (iii)swere found to be suitable for the standarization of the Pt surface. (i) Soaking the electrode in a freshly prepared, saturated FeSO4 solution renewed the Pt

Kurin-Cso¨rgei and Orba´n surface when halogen (Br2, Cl2) or oxyhalogen (BrO3-, ClO2-, IO3-) was the applied oxidant. (ii) Polishing the electrode surface with a suspension of Al2O3 (Gamma Micropolish II., deagglomerated alumina, particle size 0.05 µm, Buehler, Union Carbide) ensured good and reproducible starting conditions in all cases when the surface of the rotating disk-ring electrode was to be cleaned. (iii) Cyclic voltammetry (based on the application of a periodic triangular potential-time sweep to the electrode and recording the current flowing as a function of potential) proved to be most useful for regeneration of the Pt electrode. When a blocked or exhausted Pt electrode is applied (this electrode does not show the potential of the hydrogen electrode when it is immersed into a supporting electrolyte saturated with H2), its CV curve is free from the peaks characteristic for the oxidation and reduction of the hydrogen on the Pt surface. As the sweeps are repeated more and more times, the CV curve characteristic of the clean Pt surface in 1.0 M H2SO4, (shown for example on p 10 of ref 6) is recorded clearly, indicating that the Pt electrode is in an acceptable and reproducible state. About 40-50 cycles (in 1.0 M H2SO4) were needed to produce a clean and fresh Pt surface. Results When a Pt electrode is immersed into a sulfuric acid solution saturated with H2 gas, its potential approaches a value close to zero (vs SHE). Introducing an aqueous solution of oxidants, the potential rises and reaches either a steady high value or shows oscillations. The following oxidants have been tested: H2O2 and Na2O2, K2Cr2O7 and CrO3, Ce4+, Fe3+, MnO4-, S2O82-, and ClO2-. S2O82- immediately poisoned the electrode, preventing oscillations from being observed, and the electrode in this case needed the largest number of cycles (among the tested oxidants) to recover by the CV technique. The MnO4-, Ce4+, and Fe3+ solutions yielded very noisy potential vs time traces. The noise was irregular, its maximum and minimum values were 100150 mV, and it could not be smoothed or regulated by varying the composition of the reaction mixture or other experimental parameters. In spite of their irregular appearance, these noisy potential vs time curves may be regarded as exhibiting very high-frequency oscillations. However, very regular and reproducible oscillations were obtained under certain chemical and physical conditions when the applied oxidants were H2O2 or Na2O2, K2Cr2O7 or CrO3, and ClO2- or ClO3- + Cl-. We have focused on studying these systems in detail. H2O2/Na2O2-(Pt)-H2. Three peroxides (H2O2, Na2O2, and S2O82-) were applied to oxidize the H2 on the Pt surface. As already mentioned, in the presence of S2O82- on the Pt electrode did not yield oscillations. However, the H2O2 proved to be one of the best oxidants for oscillatory oxidation of the adsorbed hydrogen. We observed long lasting (3-4 h), high-frequency (1-20 s) oscillations with amplitude of 100-150 mV over a wide range of reagent concentrations ([H2O2] ) 0.01-0.08 M and [H2SO4] ) 0.02-2.0 M), employing either a Pt plate, Pt sheet, or rotating Pt disk-ring electrode. The oscillations were maintained for an even longer time (10 h) when the peroxide was applied in a purer chemical form as Na2O2. An interesting and unexpected observation was the relative insensitivity of the oscillations to the geometric surface of the electrode: both a Pt needle and Pt sheet electrode (2 cm2 surface area) showed potential oscillations when they were immersed in solutions of identical chemical composition. On the other hand, applying two Pt electrode-reference electrode pairs to the same solution, we found that the oscillatory curves monitored on a two-channel recorder differed in frequency (and slightly

Oscillatory Chemical Reactions in Heterogeneous Systems

J. Phys. Chem., Vol. 100, No. 49, 1996 19143

Figure 1. Oscillations in the potential of a Pt disk-ring electrode vs SHE obtained in the unstirred Na2O2-H2SO4-H2 system with increasing rotation rate. [Na2O2] ) 2 × 10-2 M, [H2SO4] ) 3 × 10-1 M, [H2] ∼ 8 × 10-4 M (saturated, H2 bubbling is stopped), t ) 25 °C.

Figure 2. Complex oscillations in the potential of a Pt disk-ring electrode at different rotation rates in the Na2O2-H2SO4-H2 unstirred system. [Na2O2] ) 2 × 10-2 M, [H2SO4] ) 5 × 10-1 M, [H2] ∼ 8 × 10-4 M (saturated, H2 bubbling is stopped), t ) 25 °C. Rotation rates: (a) 120, (b) 180, (c) 240, and (d) 360 rpm.

in amplitude) even though the surfaces of the Pt electrodes were similar in size and roughness. These findings suggest that (i) the oscillations occur on the surface of the electrode and not in the bulk of the solution and (ii) only a small fraction of the total Pt surface is involved in bringing about the oscillations. Most of the experiments were carried out using the rotating Pt disk-ring electrode because this setup provided the most regular and almost noise-free oscillations. A typical set of potential vs time curves at different rotation rates is shown in Figure 1.

At the chemical composition noted in Figure 1, oscillations appear at rotation rates between 120 and 2400 rpm. The amplitude generally increases and the frequency decreases with increasing rotation rates and also with increasing reagent concentrations. Under certain conditions complex oscillations were observed (Figure 2). The irregular oscillations at low rotation rate become regular complex ones with a sequence of one big and two small peaks, then turn again irregular, and finally settle in regular, high-amplitude single oscillations as the rotation rate is gradually increased. Fetner and Hudson5

19144 J. Phys. Chem., Vol. 100, No. 49, 1996

Kurin-Cso¨rgei and Orba´n

Figure 3. Potential vs time traces recorded in different acids under identical external parameters. [Na2O2] ) 2 × 10-2 M, [acid] ) 2 × 10-1 M, H2 (constant bubbling). Electrode: Pt sheet (2 cm2 geometric surface area). Curves: (a) HClO4, (b) H2SO4, (c) H3PO4, and (d) HCl.

have already demonstrated such a transition of a regular singlepeak to multipeak oscillations when the rotation rate of a disk electrode was varied during the reduction of H2O2 on a Pt electrode under galvanostatic conditions. Dependence of the Oscillatory Behavior on the Type of Acid in the Supporting Electrolyte. It is known that the position and intensity of the peaks in the hydrogen region of the CV curves taken in dilute HClO4, H2SO4, H3PO4, and HCl are rather different6,7 due to the anion adsorption effect on the Pt surface. The anions of the acids compete with hydrogen for the adsorption sites and may occupy part or all of the Pt surface. The following sequence has been established for the adsorbability of the anions on Pt: ClO4- < HSO4- < H2PO4- < Cl-.12 Our experiments carried out in different acids indicate that anion adsorption influences the oscillatory oxidation of H2 on Pt: The stronger the anion’s adsorptivity, the stronger the damping effect. Figure 3 presents the potential vs time responses in dilute HClO4, H2SO4, H3PO4, and HCl taken under the same physical and chemical conditions. From the curves it can be seen that using HClO4 solution is the best for avoiding complications due to anion adsorption, while Cl- from HCl completely suppresses the oscillations. When halide ions were added in portions to the oscillatory H2-Pt-Na2O2/H2SO4 system shown in Figure 3b, the following effects were observed: 10-4 M Cl- extended significantly the period time (it became 5 times longer), and 10-4 M Br- or Isuppressed the periodic behavior. Application of Other Noble Metal Electrodes. The noble metals such as Pd, Ir, Rh, and Au bear some resemblance to Pt in their hydrogen adsorption properties; therefore, we have also tested the behavior of these metals in the H2-H2O2-H2SO4 system. Among the noble metal electrodes only the response

of the Ir showed such character, high-amplitude (about 200 mV), high-frequency (1 oscillation/s) signals which can be regarded as irregular oscillations. The responses of the other electrodes in an acidic H2-H2O2 system were smooth after reaching an equilibrium potential. From the analysis of the CV curves taken in dilute H2SO4 by many authors6 the behavior of the noble metal electrodes in the H2-H2O2 systems may be understood. There are important differences in the characteristics of their voltammograms; e.g., the hydrogen adsorption on Rh commences at a significantly more cathodic potential than on Pt, and no hydrogen adsorption peaks at all were observed on Pd and Au, although the Pd absorbs a remarkable amount of hydrogen in its metal lattice. On the other hand, in the CV curve taken for Ir the hydrogen adsorption commences at the same potential and gives rise to two peaks at approximately the same places as on Pt. One may conclude that a cyclic voltammogram, which indicates the capacity of the metal surface with respect to hydrogen adsorption, may be used to predict whether the metal electrode allows hydrogen oxidation on its surface to proceed in an oscillatory fashion. Effect of Temperature. The effect of temperature on the oscillatory behavior was tested for several compositions with steady H2 bubbling. At low temperatures (it is 4 °C for the composition of [Na2O2] ) 0.02 M, [H2SO4] ) 0.2 M, and rotation rate 480 rpm) the potential approaches a steady low value around the minimum value of the oscillating potential. With increasing temperature the oscillations start with high amplitude (180 mV) and low frequency (5 s/cycle). The amplitude gradually decreases and the frequency increases as the temperature is further risen; finally, the potential becomes steady again, reaching the maximum value of the oscillatory

Oscillatory Chemical Reactions in Heterogeneous Systems

Figure 4. Oscillations in the potential of Pt sheet electrode (2 cm2 geometric surface area) in the K2Cr2O7-(Pt)-H2 unstirred system. [K2Cr2O7] ) 1 × 10-3 M, [H2SO4] ) 3 × 10-1 M, H2 (slow bubbling).

potential at about 50 °C. No hysteresis was found when the variation of the temperature was reversed. In some compositions the oscillations turned complex before they terminated again at low temperature. The effect of the temperature on the oscillatory curves is probably associated with the temperature-dependent concentration of the hydrogen gas in the bulk and on the adsorption layer, which also decreases with increasing temperature. K2Cr2O7/(CrO3 + H2O + H+)-(Pt)-H2. Potential oscillations with an amplitude of about 400 mV between 240 and 790 mV (compared to standard hydrogen electrode) and with a period of 1-10 min were observed on both Pt sheet and rotating Pt disk-ring electrodes in the concentration ranges [K2Cr2O7] ) 3 × 10-4-10-2 M and [H2SO4] ) 0.1-0.5 M. No oscillations were noted outside these ranges. The oscillations appeared after an induction period of 1-40 min; the higher the K2Cr2O7 and H2SO4 concentrations, the shorter were the induction period and the period of the oscillations. The oscillations were much fewer in number compared to those observed in the peroxide-hydrogen-Pt system: Only a few oscillations appeared in dilute solutions, and the number of peaks did not exceed 40-50 even at the highest applied concentrations. The number of oscillations in this system depends on the surface area of the Pt as well. In a solution of 0.2 M H2SO4 and 10-3 M K2Cr2O7 with gentle H2 bubbling (but without stirring) a Pt sheet (2 cm2 surface area) and a Pt needle (∼0.15 cm2 surface area) electrode exhibited 45 and 4 oscillations, respectively. When the oscillations terminated the chemical composition of the system was about the same as at the start (the concentration of K2Cr2O7 was followed by iodometric titration), and new oscillations could be initiated when the poisoned Pt electrode was treated with the CV technique or replaced. This again is evidence that the surface oscillations are hindered if the surface becomes covered with a Pt oxide film due to contact with certain oxidants. We may exclude the modification of the surface by the adsorption of the potassium ions: Using the acidic solution of CrO3 (CrO3 + H2O h H2CrO4, 2H2CrO4 h H2Cr2O7 + H2O) instead of K2Cr2O7, no difference was found in the behavior of the system. Figure 4 shows a typical oscillatory response vs time curve taken in a K2Cr2O7-H2-H2SO4 system. NaClO2-(Pt)-H2. Bromate ion was reported to be capable of oxidizing hydrogen in an oscillatory fashion on a platinum surface.8 Our preliminary experiments showed similar behavior for chlorite ions as well. Now we have studied the chlorite system in detail. When choosing the experimental conditions, we had to bear in mind that chlorite ions decompose in acid solutions; however, at pH > 4 the chlorite is stable for a reasonable time. With variation of the pH between 4 and 9 in the supporting electrolyte the ClO2--H2-Pt system readily showed long-lasting oscillations over a wide concentration range

J. Phys. Chem., Vol. 100, No. 49, 1996 19145 of chlorite (10-4-2.5 × 10-2 M), with an amplitude of 400500 mV and period between 10 s and 7 min using any kind of Pt electrode. Complex oscillations were observed at certain combinations of [ClO2-] and [H+]. The transition from simple oscillations to complex ones can also be induced by varying the rotation rate of a Pt disk-ring electrode. Figure 5 depicts simple and complex oscillations at different rotation rates. NaClO3-Cl- - (Pt) - H2. Chlorate ion is known to be kinetically inert in its reactions; therefore, oscillatory oxidation of the hydrogen by chlorate on Pt is not expected. However, oscillations (∼300 mV amplitude, ∼2 min period time) in the chlorate system were induced when a high concentration of chloride (0.2-0.3 M) was also present and high acid concentration (>1 M H2SO4) was applied. Under such conditions chlorine can be formed in the reaction between chlorate and chloride and the system must be regarded as Cl2-(Pt)-H2 rather than a ClO3--(Pt)-H2 oscillator. In the X2-(Pt)-H2 flow systems (X2 ) Cl2 or Br2), oscillations in the potential of Pt were observed when an acidic solution of halogen was pumped and hydrogen gas was simultaneously introduced into a CSTR.9 Contrary to the X2-(Pt)-H2 oscillators, the ClO3--Cl--H+(Pt)-H2 oscillates in batch as well, indicating that the continuous supply of chlorine necessary for the oscillations is ensured by the reaction between chlorate and chloride. It is worthwhile to emphasize that the ClO3--(Pt)-H2-acid oscillator is highly insensitive to chloride inhibition, suggesting a less important role of chloride adsorption on platinum in the mechanism responsible for the oscillations in this system. Discussion In our earlier studies, large-amplitude potential oscillations on Pt electrode (vs Hg/Hg2SO4/K2SO4 reference electrode) were observed in heterogeneous systems consisting of a Pt surface, aqueous solution of bromate, bromine, or chlorine, and hydrogen gas. These oscillations were supposed to originate in a surface reaction between hydrogen and elementary halogens.8,9 For describing the oscillatory behavior the surface coveragedependent activation energy model proposed first by Belyaev et al.10 was chosen from those suggested for explaining the oscillations in heterogeneous catalytic reactions. (A compilation of the applicable models are discussed in ref 2.) Here we have shown that besides Br2 and Cl2 many other oxidants, first of all hydrogen peroxide, dichromate, and chlorite ions, can also induce potential oscillations on Pt in the presence of H2 gas. However, oscillatory oxidation of H2 gas can also be carried out on the surface of Pt electrode in electrochemical cells under either galvanostatic or potentiostatic conditions. (For detailed listing of works see refs 3 and 4.) For example, in an early paper of Sawyer and Seo11 potential oscillations during the electrochemical H2 oxidation were reported, and cyclic formation and reduction of Pt oxides was suggested to include into the oscillatory mechanism. Later, Hora´nyi and Visy12 observed and explained potential oscillations under galvanostatic conditions in the presence of electrosorbing cations. The role of the external galvanostatic circuit is to ensure a constant flux of electrons needed for the oxidation/reduction processes. The role of the cations is to inhibit the oxidation by blocking the sites on the surface for H2 adsorption. Hora´nyi and Rizmayer13 have pointed out and proved experimentally that galvanostatic potential oscillations during some electrooxidation or electroreduction processes can be similar to open circuit potential oscillations. The open circuit situation means that no externally imposed voltage or current is applied to induce potential oscillations on the Pt. They showed that both electrocatalytic reduction of HNO3 on Pt and the chemical reduction of HNO3

19146 J. Phys. Chem., Vol. 100, No. 49, 1996

Kurin-Cso¨rgei and Orba´n

Figure 5. Simple and complex oscillations in the potential observed in the chlorite-(Pt)-H2 system at different rotation rates of a disk-ring electrode. [NaClO2] ) 2 × 10-4 M, pH ) 4 (adjusted with H2SO4), t ) 25 °C. Rotation rates: (a) 100, (b) 300, (c) 500, and (d) 600 rpm.

by H2 is an open circuit cell can induce potential oscillations on a platinized Pt working electrode. The presence of an appropriate amount of strongly adsorbable Cl- ions is absolutely necessary for the oscillations in H2-(Pt)-HNO3 reaction. In our experiments only H+, Na+, or K+ as cations and HSO4- and ClO4- as anions were applied (unless the effect of other ions on the oscillatory behavior was studied), and it is unprobable that the adsorption of these ions on Pt plays any definite role in bringing about the oscillatory process. In the H2-(Pt)-oxidant oscillators the electron flux needed for the redox reaction between H2 and the oxidant is ensured exclusively from the reactants, but the reaction cannot occur unless the reaction takes place in a galvanic cell arrangement. The primary driving force of the potential oscillations here is the chemical process (oxidation of H2 by the oxidant) occurring on the Pt surface. In contrary, when potentiostatic or galvanostatic conditions are applied, the external force produces the chemical changes on the electrode, giving rise to a change in its potential. Comparing our present results with those observed on galvanostatic or potentiostatic oscillations has made us consider the H2-(Pt-vs reference)-oxidant systems as a twoelectrode galvanic cell in which the Pt surface transmits the electrons between the reactants under galvanostatic-like conditions where the current passing through the circuit is practically zero. In the oscillatory catalytic surface reactions and in the electrocatalytic oscillators, similarly to the H2-(Pt)-oxidant systems, the coverage of the surface of the catalyst changes periodically; therefore, these oscillators seem to be interrelated, serving the H2-(Pt)-oxidant system as a link between the other two groups. The origin of the oscillations in the H2-(Pt)-oxidant systems is not clear at this moment. It can be a nonlinearity in the overall

reaction which arises from physicochemical processes attributed to oscillatory behavior in heterogeneous catalysis or in pure electrochemical systems, e.g., coverage-dependent adsorption and reaction rate, the autocatalytic free site generation on the surface, the oxide formation and reduction, the inhibitory effect of the product adsorption, etc., which may include further complications; namely, the extent of these phenomena can be potential dependent. In the H2-(Pt)-oxidant oscillators the net chemical reaction is very simple:

H2 + ox f H+ + red The H2 is known to diffuse fast to and adsorb strongly on the Pt surface where it reacts with the oxidant. Because no bulk oscillations are observed (a slow reaction between H2 and the oxidant in the bulk definitely occurs), only the adsorbed H2 is considered to react in an oscillatory fashion. The product H+ appears as H2O if the oxidant is H2O2 or Cr2O72-, but formation of HX (X ) halogenide ion) cannot be excluded when the oxidant is halogen (Br2, Cl2) or oxyhalogenate ions (BrO3-, ClO2-). The oxidation of H2 and the reduction of the oxidant take place simultaneously on the surface of the single Pt electrode, and the ratio of their rates generates the potential of the Pt. This mixed potential may change periodically if the kinetics of the reaction and the mass transport match properly. The main aim of this paper is to present our experimental results, and now we do not intend to test known models to simulate the H2-(Pt)-oxidant oscillators. We have already shown that the surface coverage-dependent activation energy model works for the H2-(Pt)-Br2/Cl2 systems,9 but we feel

Oscillatory Chemical Reactions in Heterogeneous Systems that a model which implies electrochemical considerations as well may serve better for modeling the H2-(Pt)-oxidant oscillators. Since the potential oscillations occur usually between +0.3 and +0.8 V, the Pt(II/IV) oxide/hydroxide formation and reduction may not be neglected. Although no electrosorbing cations or anions were purposely added to the present systems for generating oscillations, the adsorption of the product OH-, Cr3+, Cl-, and Br- may also be a factor to deal with. Acknowledgment. This work was supported by the Hungarian Scientific Research Fund (OTKA) through Grants T 016680 and F 017073. References and Notes (1) For one of the most recent review on the field, see e.g.: Gray, P.; Scott, S. K. Chemical Oscillations and Instabilities; Claredon Press: Oxford, 1994.

J. Phys. Chem., Vol. 100, No. 49, 1996 19147 (2) Schu¨tz, F.; Henry, B. E.; Schmidt, L. D. AdV. Catal. 1993, 39, 51-127. (3) Wojtowicz, J. In Modern Aspects of Electrochemistry; Bockris, J., Conway, B. E., Eds.; Plenum Press: New York, 1972; Vol. 8, pp 47-120. (4) Hudson, J. L.; Tsotsis, T. T. Chem. Eng. Sci. 1994, 49, 14931572. (5) Fetner, N.; Hudson, J. L. J. Phys. Chem. 1990, 94, 6506. (6) Woods, R. Chemisorption at Electrodes. In Elelctroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1976; Vol. 9, pp 2-162. (7) Bakos, I.; Hora´nyi, G. Magy. Ke´m. Foly. (in Hungarian) 1992, 98, 146. (8) Orba´n, M.; Epstein, I. R. J. Am. Chem. Soc. 1981, 103, 3723. (9) Orba´n, M.; Epstein, I. R. In Synergetics; Vidal, C., Pacault, A., Eds.; Springer-Verlag: Berlin, 1981; Vol. 12, pp 197-200. (10) Belyaev, V. D.; Slinko, M. M.; Slinko, M. G. Proc. Congr. Catal. 6th 1977, 2, 758. (11) Sawyer, D. T.; Seo, E. T. J. Electroanal. Chem. 1963, 5, 23. (12) Hora´nyi, G.; Visy, C. J. Electroanal. Chem. 1979, 103, 353. (13) Hora´nyi, G.; Rizmayer, E. M. J. Electroanal. Chem. 1982, 140, 347.

JP961714S