Oxidation and Adsorption of Hydrocarbons on Noble Metal Electrodes

17 Oxidation and Adsorption of Hydrocarbons on Noble Metal Electrodes VI. A Discussion of the Mechanism of Saturated Hydro

Downloaded by UNIV OF LEEDS on June 18, 2016 | http://pubs.acs.org Publication Date: June 1, 1969 | doi: 10.1021/ba-1969-0090.ch017

carbon Oxidation on Platinum S. B. B R U M M E R Tyco Laboratories, Inc., Waltham, Mass. 02154

The anodic oxidation of saturated hydrocarbons discussed.

Coulombic

efficiency for

CO

2

on Pt is

production

is

100% even for complex fuels. The need to examine adsorp tion processes in mechanistic

studies is emphasized.

Such

studies reveal three materials in the steady state:

CH-β

(polymeric), CH-α (alkyl radicals), and O-type (oxygenated, C -species). 1

O-type is predominant

and oxidizes the most

readily at high potentials. Its coverage is high below about 0.4 volt (R.H.E.)

and insensitive to hydrocarbon

pressure.

This observation, and reports that the overall reaction order is unity, suggest that O-type is a poison for the overall reac tion.

However,

recent results indicate that in the

region

of high coverage, the overall reaction order is less than unity and sometimes negative. It is suggested that the slow step of the overall reaction involves oxidation of an adsorbed species—e.g.,

O-type—present

at high coverage.

*Tphe anodic oxidation of saturated hydrocarbons is required for the economic utilization of the fuel cell principle.

Consequently, there

has been much interest i n this area i n the past five years. The earliest reports of saturated hydrocarbon oxidation at low temperatures and interesting rates were by Grubb and Niedrach i n 1963 ( 20).

They used

extensive amounts of Pt as the anode catalyst and concentrated H3PO4 at 150°C. as the electrolyte and achieved moderate current densities at reasonable potentials. Since then there has been a large number of studies and it is clear that while many saturated hydrocarbons can be oxidized 223

Baker; Fuel Cell Systems-II Advances in Chemistry; American Chemical Society: Washington, DC, 1969.


224

F U E L

C E L L

SYSTEMS

II

the kinetics are rather unfavorable. The ultimate utilization of the hydro carbon anode requires an enhancement of these kinetics, or requires the development of a sufficiently cheap—i.e., non-noble metal catalyst that the electrode area can be appropriately expanded. It appears that maximum activity is found for the C 2 - C 4 (J, 9) range, that straight chain hydrocarbons are better than bridged chains (18) and, that Pt is one of the better catalysts that we have (17). In our own work, we have undertaken to investigate the mechanism of anodic hydrocarbon oxidation, with a view to elucidating those features, structural, electronic and ionic, which determine the path and the rate of the overall reaction. In the first instance, we have studied the adsorp tion and oxidation of C H (3, 5, 6) and n - C H i (7) at elevated tem peratures using hot concentrated H P 0 and smooth Pt electrodes. W h i l e the chemical mechanism has not been fully unravelled at this point, a number of significant features of the reaction path have been revealed. The present paper summarizes the result of this work and discusses mechanistic conclusions, which can be drawn, both from our own work and from that of others. 3

8

6

3

Overall

Course of

4

4

Reaction

It is well known that saturated hydrocarbons are very stable materi als and it would be expected that their oxidation would be very difficult. Indeed, as mentioned, stringent conditions are required for the reaction to occur at any reasonable rate. By analogy with gas phase reactions and, i n particular, the heterogeneous catalysis reactions of these com pounds, one might expect that the reaction path would be very compli cated and as a consequence that there would be a number of products. However, it has been shown that the general reaction C„H

(2n +2 )

+ 2 n H 0 - » n C 0 + (6n + 2 ) H + (6n + 2)e~ 2

+

2

(1)

occurs to completion with no side products accumulating i n the solution or i n the gas phase (1, 10,19). This 100% faradaic efficiency for the production of C 0 is remark able and unexpected and has important mechanistic implications. Thus, following the initial adsorption, all the intermediate products between the reactant and the final product, C 0 , are adsorbed on the electrode. There is no other way to account for the observations. This suggests that the fruitful way to investigate the reaction mechanism is to study the adsorption processes. This has been the prime approach of our experi ments and, i n conjunction with the work of others, notably Gilman (13, 14) and Niedrach (21, 22, 23, 24), it has led to a considerable under standing of the nature of the products formed on the electrode. 2

2


17.

BRUMMER

Nature

225

Saturated Hydrocarbon Oxidation

of Adsorbed

Products

The methods we have used to examine the adsorbed products involve their determination with anodic stripping and with Η-atom deposition. Using these techniques in conjunction with controlled potential adsorp tion regimes onto a clean electrode, we can define the charge to oxidize the adsorbed species, O C 0 .

(2) (3)

2

In this mechanism, the main reaction (Reaction 3) occurs on that part of the electrode which is not occupied by adsorbed species, or at least on that part not occupied by the poisoning adsorbed species. It was suggested (6, 7) that reduced C0 —i.e., O-type, was the worst kind of poisoning species since, being a C i species, it very likely consumed just those C i reactive radicals which maintain the rate of the overall reaction. This view is i n sharp contrast to that of Niedrach et al. and suggests that those reactions which lead to the production of O-type are undesirable. Since some adsorbed species must be controlling the rate of the over all reaction, we suggested the possibility that the CH-α species comprises several parts (6,7). Some of these parts, following Niedrach et al. (24), are undesirable but some of them, the CH-a tive, were required to pro mote the overall reaction. 2

ac

Reaction

Order

As indicated, a central observation leading to the divergence of our view from that proposed by Niedrach et al. is the reaction order. I n an early study, we reported that the reaction order on smooth Pt is positive (27). Recently it has been reported (10, 26) that on platinized Pt the reaction order for C H and n - C H i is accurately unity. Those experi ments, on platinized Pt, were the reasons for our adopting the views expressed above. However, we have reexamined the reaction, using platinized Pt electrodes, and two important points have emerged: Firstly; we have found that coverage with O-type is much more dependent on fuel con centration than it is on smooth Pt. F o r example, at 0.35 volt on smooth Pt at 130 °C. the O-type concentration does not change when the C H pressure increases from 2.2 mm. to 223 mm. (6). Similar insensitivity is 3

8

6

4

3


8

17.

BRUMMER

227


found with n - C H i (7). O n platinized Pt, however, the concentration of O-type increases by from 0 to ~ 250 jucoul./rcm. for C H under the same conditions (28). W e cannot speculate, at this stage about this difference i n behavior between smooth and platinized Pt. The observa tions prompted us to reexamine the overall reaction order on platinized Pt. The second, and crucial, finding was that at low potentials—e.g., below 0.35 volt, the reaction order is not unity. In fact, we find a small negative order for the overall reaction rate (28). This is precisely the kind of complication which would be expected if the reaction were coming under extensive adsorption control at relative high coverage. There is little experimental conflict between our results and those re ported by Gileadi et al. (10) since most of their studies refer to the potential region above 0.35 volt, where the coverage with adsorbed materials becomes low. G

4


2

Relation

of Adsorbed Products

to Overall

3

8

Reaction

These observations show that there is still a sharp disagreement between the orders of the overall reaction and of O-type coverage on platinized-Pt. However, the results are more ambiguous than was previ ously thought (6). The unity order for the overall reaction reported by Gileadi et al. for C H oxidation (10) implies a relatively simple mecha nism. Specifically, the reaction could not proceed via O-type since the coverage with O-type was thought to be high and virtually independent of C H -pressure (6). For this reason, we postulated that the reaction proceeded on the part of the surface not occupied by O-type, a similar poisoning mechanism to H C O O H oxidation (2, 4). The negative reaction order we have observed, in conjunction with the increase in O-type coverage with pressure on platinized Pt (28), could certainly be inter preted in terms of such a poisoning mechanism (Reactions 2 and 3). However, these results on the overall reaction order show that at low potentials the reaction could proceed via an adsorbed species present at high coverage—e.g., O-type. Consideration of such a mechanism allows an alternative interpreta tion of the data. Thus, we note that coverage with C H - α has an even stronger dependence on C H pressure than has O-type for smooth (6) and for platinized Pt (28). If the oxidation of O-type involved a "reactant pair" mechanism, involving the adsorbate itself and free sites, as postu lated by Gilman for C O (11, 12), C H - α could act as a poison for the reaction because of its occupation of surface. Since CH-α is so pressure dependent, we can entertain this mechanism as an explanation of the above conflict in reaction orders while still asserting that the main re action proceeds via O-type. W e have found an increase i n the rate of 3

3

8

8

3

8

a d s


228

F U E L

C E L L

S Y S T E M S

II

O-type oxidation i n circumstances where we have deliberately removed C H - a (28) and this supports such a mechanism. These observations allow possibly the maintenance of the views expressed by Niedrach et al. concerning the role of O-type. A third observation with platinized Pt electrodes leads to further elaboration of the mechanism. W e have investigated the oxidation kinetics of the adsorbed species on platinized Pt in presence of C H with anodic chronopotentiograms. These were taken at sufficiently low current densities that the adsorbate oxidation occurs well before the electrode's oxidation but at sufficiently high current densities that insignificant s o l u t i o n - C H is oxidized. Two predominant features are observed, a prewave and a more extensive potential plateau (28). The behavior of the plateau leaves no doubt that it corresponds to the oxidation of O-type material—i.e., reduced C 0 . The oxidation current corresponding to the plateau is about 40-50% of the overall reaction current for oxidation of the hydro carbon. This is a reasonable value if O-type lies between C H and C 0 and indeed strongly suggests this conclusion. 3

8


3

8

2

3

8

2

The prewave is associated with the reactions of an adsorbed material on the electrode other than O-type, presumably C H - α . A t low currents, the prewave is found at lower potentials than O-type but at high oxidation rates it merges into the O-type wave. (It is for this reason that Giner (16), who used less rapid scanning techniques to explore hydrocarbon adsorbates, d i d not report any C H - α . ) Eventually, both processes are displaced to the region of electrode oxidation. The crucial point is that whereas at high potentials—e.g., in our "anodic desorption" experiments (β, 7) or in fast linear sweeps (13, 14, 24) C H - α w i l l oxidize less easily than O-type, it oxidizes more readily at low potentials. The relative oxidation rates at low potentials are what are significant for the operation of the hydrocarbon anode. Hence, we can no longer assert that formation of C H - α is necessarily undesirable. Similarly, our previous conclusion (6, 7) that C H - α production is parallel to O-type production can no longer be maintained. There is no reason to suppose that at low potentials—e.g., ^ 0.35 volt the reaction sequence (4) (5) (6) does not occur. If this is so, the oxidation of O-type must be rate-limiting at least as far up the homologous series as n - C H i . This is because O-type's cover age is so high. A n additional complication also arises from the above6

4


17.

BRUMMER

229


mentioned poisoning of O-type oxidation by the C H - α occupying the surface.

A s the homologous series is ascended, C H - α coverage increases

and at some stage it is clear that its oxidation reaction, Reaction 5, becomes rate limiting. F r o m the coulometry of the adsorbed species, we have the following numbers of electrons i n these various reactions C H 3

8

6-7 ->

suggested (28).

3-6 8-10 C H - α -> O-type - » 3 C 0 .

(7)

2

Confirmation of this reaction mechanism is required.

This, a n d


elucidation of the chemical structures of the intermediates are the aims of our further work. Acknowledgments It is a pleasure to thank M . J. Turner who carried out the experiments whose results are described above.

This work is supported b y the U . S.

A r m y Engineer Research and Development Laboratories, Fort Belvoir, Virginia, on Contract D A 44-009-AMC-1408(T). Literature

Cited

(1) Binder, H . , Köhling, Α., Krupp, H . , Richter, K., Sandstede, J. Electro chem. Soc. 111, 1015 (1964). (2) Brummer, S. B., J. Electrochem. Soc. 113, 1041 (1966). (3) Brummer, S. B., Ford, J. I., Turner, M . J., J. Phys. Chem. 69, 3434 (1965). (4) Brummer, S. B., Makrides, A. C., J. Phys. Chem. 68, 1448 (1964). (5) Brummer, S. B., Turner, M . J., "Hydrocarbon Fuel Cell Technology," p. 409, B. S. Baker, E d . , Academic Press, New York, Ν. Y., 1965. (6) Brummer, S. B., Turner, M . J., J. Phys. Chem. 71, 2825 (1967). (7) Ibid., 71, 3494 (1967). (8) Ibid., 71, 3902 (1967). (9) General Electric Co. Rept. U.S.A.E.R.D.L., Contract D A 44-009-ENG4909 (Dec. 31, 1963). (10) Gileadi, E . , Stoner, G . , Bockris, J. O ' M . , Univ. of Pennsylvania, Rept. U.S.A.E.R.D.L., Contract D A 44-009-AMC-469(T) (April 1966). (11) Gilman, S., J. Phys. Chem. 67, 1878 (1963). (12) Ibid., 68, 70 (1964). (13) Gilman, S., Trans. Faraday Soc. 61, 2546 (1965). (14) Ibid., 61, 2561 (1965). (15) Giner, J., Electrochim. Acta 8, 857 (1963). (16) Giner, J., Proc. 15th C.I.T.C.E. Meeting, London (1964). (17) Grubb, W . T., J. Electrochem. Soc. 113, 191 (1966). (18) Grubb, W . T . , Michalske, C . J., Proc. 18th Power Sources Conf., p. 7 (1964). (19) Grubb, W . T., Michalske, C. J., J. Electrochem. Soc. 111, 1015 (1964). (20) Grubb, W . T., Niedrach, L . W . , J. Electrochem. Soc. 110, 1086 (1963). (21) Niedrach, L . W . , J. Electrochem. Soc. 111, 1309 (1964). (22) Niedrach, L . W . , "Hydrocarbon Fuel Cell Technology," p. 377, B. S. Baker, E d . , Academic Press, New York, Ν. Y., 1965.


230

F U E L C E L L SYSTEMS

II

(23) Niedrach, L. W., J. Electrochem. Soc. 113, 645 (1966). (24) Niedrach, L. W., Gilman, S., Weinstock, I., J. Electrochem. Soc. 112, 1161 (1965). (25) Niedrach, L. W., Tochner, M., J. Electrochem. Soc. 114, 17 (1967). (26) Stoner, G., Bockris, J. O'M., (Univ. of Pennsylvania), Rept. U.S.A. E.R.D.L., Contract DA 44-009-AMC-469(T) (Oct. 1966). (27) Tyco Laboratories, Rept. U.S.A.E.R.D.L., Contract DA 44-009-AMC410(T) (Nov. 1964). (28) Tyco Laboratories, Rept. U.S.A.E.R.D.L., Contract DA 44-009-AMC1408(T) (June 1967).


RECEIVED November 20, 1967.


Oxidation and Adsorption of Hydrocarbons on Noble Metal Electrodes

Recommend Documents