J. Phys. Chem. 1984, 88.4444-4446
4444
an acceleration of the isothermal decomposition rate with decreasing coverage; they also were unable to account for this result. Many previous studies of formic acid on nickel have focused attention on dehydrogenation vs. dehydration as measured by the CO/CO2 product r a t i ~ . ~We ~ ~have ~ ' purposefully not emphasized that issue since the CO/CO2 product ratio can be skewed by secondary reactions. Nickel is known to catalyze C 0 2 decomposition into CO and adsorbed oxygen, C O and CO, hydrogenation, and the water gas shift reaction. Since the reaction kinetics more closely reflect the behavior of the primary reactions, it is our contention that reaction kinetics provide a more meaningful comparison between the UHV and moderate-pressure regimes.
VACANT SITE COVERAGE
co
FORMATE
Figure 6. Phase diagram for formate and CO adsorbed on Ni at 548 K. Tie lines correspond to the following gas-phase compositions: (a) PCO = 0.067 torr, PHcmH = 1.6 torr; (b) Pco = 0.10 torr, P H C ~=H 2.0 torr; (c) Pco = 0.29 torr, P H C m H = 4.0 torr; (d) PCO= 0.62 torr, P H C ~=H 7.1 torr. Plait point at PCo = 0.061 torr, PHCOOH= 1.5 torr. whereas Figures 3b and 5b are in the single-phase region and little effect on the reaction rate is noted. The formate-formate interactions are not sufficiently attractive to induce surface-phase condensation in the two-component formate-empty site system; however, addition of a third species such as CO can induce phase separation. The enhancement of the reaction rate by the addition of hydrogen is due to a reduction in stability of the formate intermediates. Addition of hydrogen causes a slight reduction in the number of formate-formate nearest-neighbor pairs, which accelerates the rate. The hydrogen-formate interactions appear to be negligible, so that no phase separation occurs and no dramatic increase in the reaction rate is observed. Using the heat of adsorption of hydrogen noted above and no hydrogen-formate interactions, the model is able to reproduce the experimental results shown in Figure 4. There are several anomalous results in previous studies of formic acid decomposition on nickel at moderate pressures that can now be explained in terms of adsorbate interactions. Walton and Verhoek observed that as C O was added to formic acid vapor, the isothermal reaction rate initially increased and then decreased.40 They were unable to account for this result, though it is now apparent that it results from destabilization of the formate intermediates by adsorbed CO. It seems that adsorbed CO probably played an important role in the supercatalytic activity of nickel wires reported by Duell and Robertson4* and also by W i l l h ~ f t .Evidence ~~ for attractive interactions between formate intermediates is presented by Giner and R i s ~ m a nwho , ~ ~observed (48) M. J. Duell and A. J. B. Robertson, Trans. Faraday SOC.,57, 1416 (1961). (49) E. M. A. Willhoft, J. Chem. SOC.,Chem. Commun., 146 (1968).
Conclusions The results presented in this paper have elucidated the mechanism of formic acid decomposition observed in many TPD experiments. Formic acid dimers formed at low temperatures dehydrate on nickel surfaces, leaving adsorbed CO and formate on the surface. Formic acid monomers dehydrogenate to form formate intermediates. The TPD experiments indicated that adsorbate interactions caused the unusual kinetics. It was found that formate-formate interactions were attractive, while COformate interactions were strongly repulsive. Kinetic parameters determined from the TPD experiments were used to predict reaction kinetics at moderate pressurs successfully. Several important features that could be represented by the model include (1) a maximum in the isothermal reaction rate as a function of formic acid pressure, (2) a dramatic maximum in the isothermal reaction rate with the addition of carbon monoxide, and (3) enhancement of the isothermal rate of reaction with the addition of hydrogen. None of these results can be accounted for by simple Hougen-Watson type rate expressions, and they can be predicted only by considering adsorbate interactions. The enhancement of the reaction rate by the addition of a nonreactive component is an important finding. In this work CO had a dramatic effect on the rate of formate decomposition by destabilizing the formate. It was found the surface phases could be formed by adding CO, a nonreactive component. The formation of surface phases greatly affected the reaction rate. In this way CO acted as a cocatalyst, similar to a coenzyme in a biological system. These results suggest the possibility of adding nonreactive components to facilitate other heterogeneous chemical reactions. In cases of bimolecular reactions it may be possible to add a third component that would push the reactive components closer together to facilitate reaction. This concept clearly needs further investigation. Acknowledgment. We thank the Air Force Office of Scientific Research for their financial support of this work. Registry No. Ni, 7440-02-0; CO, 630-08-0; formic acid, 64-18-6; H,, 1333-74-0. (50) J. Giner and E. Rissman, J . Catal., 9, 115 (1967). (51) P. Mars, J. J. F. Scholten, and P. Zwietering, Adv. Caral., 14, 35 ( 1963).
COMMENTS Concernlng the Alleged Efficiency of Photoaquatlon in the Ultraviolet Photolysls of Bromo(pentammlne)cobalt( I I I ) S i c In a recent paper in this journal, Kirk et al.' have reported that ultraviolet irradiations of C O ( N H ~ ) ~ Bproduce ~ ~ + Co0022-3654/84/2088-4444$01.50/0
(NH,),OH?+ and Co2+in a wavelength-independent0.46: 1 ratio. This stands in contradiction to our earlier report^^,^ that we were unable to detect photoaquation products, and the contradiction has led us to reinvestigate the 254-nm photolysis of C O ( N H ~ ) ~ B ~ ~ + in acidic aqueous solution^.^ Owing to the uncertainties intrinsic to spectrophotometric determinations of small amounts of photolysis products in the 0 1984 American Chemical Society
The Journal of Physical Chemistry, Vol. 88, No. 19, 1984 4445
Comments
TABLE I: Summary of Experimental Observations in the 254-nm Photolysis of Co(NH3)5Br2t quantum yields no. of % exptl conditionsa methods of analysis for Co(I1I) species dtmns photolysis CoZt C O ( N H ~ ) ~ O H ? +A [ C O ( N H , ) ~ B ~ ~ + ] [ C O ( N H ~ ) ~ B ~ ~=~ ] , , , ,spectrophotometric, ~ with tar = 1.75 X 6 I20 0.31 f 0.02 0.03 f 0.08 0.44 f 0.09 lo4 and €OH2 = 3.1 X 10' M-' cm-' (8.6-5.7) X lo4 M; at 256 nm, €'ar = 7.1 X lo2 and €'OH2 = 0.1 M HCIO4 38 M-' cm-' at 325 nm [ C O ( N H ~ ) ~ B ~ ~=+ ] , , , ~ chromatography on Dowex 50WX100 3