Catalysis over Supported Metals. IV. Ethane Hydrogenolysis over

Catalysis over Supported Metals. IV. Ethane Hydrogenolysis over Dilute Nickel Catalysts. W. F. Taylor, J. H. Sinfelt, and D. J. C. Yates. J. Phys. Che...
1 downloads 0 Views 726KB Size
CATALYSIS OVER SUPPORTED METALS

The resulting specific surface is 100 m.2/g., corresponding to an average particle dimension of 60 A. Because of the many assumptions needed to apply (8) and (9) to the gels made from the HTS sol, these equations can be expected to give only the order of magnitude of the specific surface, and thus the difference between the results from (8) and (9) probably is not significant. As the normalizing integral was probably underestimated, the specific surface computed from (9) may be too large, and a reasonable estimate of the order of magnitude of the specific surface would be 200 m.2/g., and the cozresponding average particle dimension would be 30 A. Since this average dimension is much nearer to the

Catalysis over Supported Metals. IV.

3857

value obtained by line broadening than to the dimension computed from nitrogen adsorption, one can conclude that in this gel most of the pores are inaccessible to nitrogen. This inaccessibility is responsible for the large difference between the average dimensions determined by adsorption and line broadening.

Acknowledgments. The author wishes to express his appreciation to K. H. McCorkle and co-workers for assistance in preparing the gels and for advice and guidance during the course of the investigation, to D. E. Andrews, James E. Thomas, and T. R. Taylor for aid in obtaining the scattering data, and to E. E. Pickett for assistance in preparing and handling the gel samples.

Ethane Hydrogenolysis

over Dilute Nickel Catalysts

by W. F. Taylor, J. H. Sinfelt, and D. J. C. Yates Process Research Division, Esso Research and Engimering Company, Linden,New Jersey (Received May 24, 1965)

The kinetics of ethane hydrogenolysis have been studied over dilute nickel catalysts containing 1 and 5% nickel and compared with previous results on catalysts containing 10% nickel. Two different supports, ailica and silica-alumina, were used for the nickel. The The nickel surface reaction rate measurements were made in the range 175 to 335'. areas of the catalysts were determined by hydrogen chemisorption. This made it possible to determine the specific catalytic activity of the nickel in terms of rate of hydrogenolysis per unit area of nickel surface. The specific catalytic activity and apparent activation energy vary significantly with the nickel concentration. The catalytic properties of the very dilute 1% nickel catalysts, in particular, are markedly d.8erent from those of higher nickel concentration and exhibit an especially large effect of the support used for the nickel.

I. Introduction The catalytic hydrogenolysis of ethane Over SUP ported nickel has previously been investigated by Taylor and co-workers,1s2using a catalyst containing 15% nickel deposited on kieselguhr. These studies

revealed some striking kinetic effects, particularly the strong inverse dependence of the reaction rate on (1)

K. Morikawa, W. S. Benedict, and H. S. Taylor, J . Am. Chem.

(2)

c. Kemball and H. s. Taylor, ibid., 70,345 (1948).

so,.., 58,

1795 (1936).

Volume 69, Number 11

November 1965

3858

hydrogen pressure. Published studies by the authors of the present paper have confirmed these results for catalysts containing 10% nickel and have demonstrated a marked effect of the support on the catalytic activity of nickel for this rea~ti0n.a~Recently, we have extended the work to dilute nickel catalysts, ie., catalysts with nickel concentrations in the range of 1 to 5% nickel. The purpose of the work was to determine whether the catalytic properties of dilute nickel catalysts are significantly different from those of the more conventional nickel catalysts of higher concentration. Evidence of a critical nickel concentration has previously been reported by Hill and Selwoods in studies of benzene hydrogenation over a series of nickel on alumina catalysts of varying nickel concentration. These investigators found that catalysts containing less than 3% nickel were virtually inactive for the benzene hydrogenation reaction, while catalysts containing more than 10% nickel were highly active. I n the present paper, the results of kinetic studies on ethane hydrogenolysis over dilute nickel catalysts containing 1 and 5% nickel are discussed and compared with previously reported data on catalysts containing 10% nickel. The studies were made using both silica and silica-alumina as supports for the nickel. Hydrogen chemisorption measurements were made on all of the catalysts to determine the extent of dispersion of the nickel on the supports. Significant differences between the very dilute nickel catalysts and those of higher nickel concentration have been observed, both from the standpoint of the magnitude of catalytic activity and the kinetic relationships and parameters found for the different catalysts.

11. Experimental Section A. Kinetic Measurements. Apparatus and Procedure. The reaction rate measurements were carried out in a flow system at atmospheric pressure. Details of the reactor have been described previously.6 The reaction gases were passed downflow through the catalyst bed, and the products were analyzed by a chromatographic unit coupled directly to the outlet of the reactor. The chromatographic column was 2 m. in length and 0.6 cm. in diameter. The column was packed with 100-mesh silica gel and was operated a t 40". Helium was used as a carrier gas, and a thermal conductivity detector was used with the column. The reactant gases, ethane and hydrogen, were passed over the catalyst in the presence of helium diluent. Gas flow rates were measured using orificetype flowmeters with manometers. A total gas flow rate of 1 l./min. was used throughout. The run procedure consisted of passing the reactant gases over The Journal of Physical Chemistry

W. F. TAYLOR, J. H. SINFELT, AND D. J. C. YATES

the catalyst for a period of 3 min., a t which time a sample of the product was taken for chromatographic analysis. The ethane was then cut out, and hydrogen flow was continued for a period of 10 min. at the reaction temperature prior to another run. I n this way, it was possible to minimize variation in catalyst activity from period to period. As an additional insurance against the complications due to varying catalyst activity, measurements at any given set of conditions were in most cases bracketed with periods at a standard set of conditions. In this way, the effect of changing a variable can be determined by comparison with the standard condition periods run immediately before and after the period in question. I n each reaction study 0.2 g. of catalyst was diluted with 0.5 g. of ground Vycor. The catalysts were reduced in the reactor in flowing HPovernight at 370". Materials. The ethane was obtained from the Matheson Co. It is estimated that any hydrocarbon impurities amounted to less than 0.01% each. High purity hydrogen was obtained from the Linde Co. and was further purified by passing it through a Deoxo unit containing palladium catalyst to remove traces of oxygen as water, prior to passage through a molecular sieve dryer. The nickel catalysts were prepared by impregnation of silica or silica-alumina with an aqueous solution of Xi(XO3)2.6HzO. The silica employed was Cabosil HS-5 (surface area 340 m.Z/g., particle size less than 1 p ) , obtained from the Cabot Corp., Boston, Mass. The silica-alumina used was DA-1 (nominally 13% A1203,87% SiOt), obtained from the Davison Chemical Co., Baltimore, hld. The surface area of the silicaalumina was 450 m.2/g., and the particle size was less than 50 p. I n the case of the nickel on silica catalysts, the concentration of the impregnating solution ranged from 0.7 to 7.0 g. of Ni(N03)z.6Hz0/100ml. of water, depending on the final nickel concentration desired. For the nickel on silica-alumina catalysts, the concentration ranged from 7.0 to 70 g./lOO ml. of HzO. The lower concentrations in the case of the nickel on silica were necessary because much more solution was required to obtain satisfactory wetting of the Cabosil. After impregnation, the catalysts were dried overnight a t 105". The dried catalysts were then pressed into wafers at 8000 p.s.i., after which the wafers were (3) D. J. C. Yates, W. F. Taylor, and J. H. Sinfelt, J.Am. Chem. Soc., 86,2996 (1964). (4)W.F. Taylor, D. J. C. Yates, and J. H. Sinfelt, J . Phys. Chem., 68, 2962 (1964). (5) F. N. Hill and P. W. Selwood, J . Am. Chem. SOC.,71, 2522 (1949). (6) J. H.Sinfelt, J. Phys. Chem., 68, 344 (1964).

CATALYSIS OVER SUPPORTED METALS

crushed and screened. The 45-60-mesh fractions were used in the experiments. B. Chemisorption Measurements. Apparatus and Materials. A conventional glass vacuum system was used, with a mechanical backing pump, an SO-l./sec. oil diffusion pump, and a trap cooled with either solid carbon dioxide or liquid nitrogen. The ultimate dynamic vacuum reached by the system was about lo-' torr and was measured by an Alpert-type ionization gauge. Details of the apparatus have been described previ~usly.~ Hydrogen of 99.984% purity was purchased from the Linde Co., Linden, 3. J. Traces of oxygen were removed from it by a Deoxo purifier, obtained from Englehard Industries, Inc., Newark, N. J. The water formed in the Deoxo unit was removed by a trap cooled with solid carbon dioxide. The helium, used in calibrating the cells, was obtained from the Bureau of Mines, Amarillo, Texas, and had a purity of 99,98% or better. It was dried by passage through a trap cooled by liquid nitrogen before use. Procedure. The sample was glass-blown into the cell, no greased joint being used. The cell was weighed before and after putting in the sample so that corrections could be made for water driven off during the reduction and subsequent evacuation. The sample was then heated and evacuated until 100" was reached. After evacuating at this temperature for 1 hr., the hydrogen was passed through the sample at a flow rate of 500 ~ m . ~ / m i n .The temperature was then slowly raised to 370", while maintaining the above flow. The sample was then reduced overnight at 370" and 500 cm.a/rnin. The hydrogen was then cut off, and the sample was evacuated at 370" for 1 hr. ,4t the end of this time, pressures in the region of low6 torr were obtained. The sample cell was then isolated from the pumps and cooled to 20". The hydrogen isotherm was then run, usually three or four points being obtained with the highest pressure used being about 30 cm.

111. Results The hydrogen chemisorption isotherms obtained on the nickel catalysts are plotted in Figure 1. The amount of hydrogen adsorbed at a pressure of 10 cm. was taken as a measure of the monolayer point. Nickel surface areas were calculated on the basis that each nickel atom in the surface adsorbs one hydrogen atom and that each nickel atom occupies 6.5 A.z of surface.? Nickel surface areas of the various catalysts are given in Table I. As the nickel content of the catalyst decreased, it can be seen that the amount of hydrogen adsorbed, and correspondingly the nickel surface area,

3859

O L

-- - ,

0

10 20 Equilibrium pressure, om.

30

Figure 1. Adsorption isotherms for Hz on supported nickel catalysts at room temperature: 0, 10% Ni on SiOz; 0, 5% Ni on SiOz; A, 1% Ni on SiOzO; 0, 10% Ni on SiOz-Al~Oa; . , 5% Ni on SiOz-AWa; A, 1% Ni on Si02-Al~O~.

decreased. Since the amounts of hydrogen adsorbed on the 1% nickel content catalysts are relatively low, it is probable that the nickel area measurements on these catalysts are less reliable than those on the higher nickel content catalysts. Adsorption of hydrogen on the supports was a negligible part of the total adsorption, except possibly for the 1% nickel on silica-alumina catalyst. I n this case, the total volume of hydrogen adsorbed at 10 cm. approaches the limit of detectability of the adsorption apparatus, and, hence, it is difficult to make a reliable judgment about the contribution of the support to the total adsorption. However, it will become evident from subsequent results that a precise knowledge of the nickel surface area of this catalyst is not necessary for a comparison of its catalytic properties with those of the other catalysts. Table I : Summary of Nickel Surface Areas Determined by HZChemisorption Catalyst

Ni surface area, m.z/g. of catalyst

Ni on SiOz 1% Ni 5% Ni 10% Ni

0.7 5.9

13.6

Ni on SiO~-Al208 1% Ni 5% Ni 10% Ni

0.1

3.1

6.8 ~

(7) D. F. Klemperer and F. 9. Stone, Proc. Roy. SOC. (London), A243, 375 (1958).

Volume 69, Number 11 November 1966

3860

W. E'. TAYLOR, J. H. SINFELT, AND D. J. C. YATES

10-1

The hydrogenolysis of ethane to form methane was studied over the various catalysts at temperatures above about 180". In studying the kinetics, the approach taken was to measure the rates of reaction at low conversion levels. The degree of conversion in the present work ranged from 0.02 to about 7%, most of the data having been obtained at conversion levels below 1.0%. Consequently, the partial pressures of the reactants (ethane and hydrogen) do not vary much through the reaction zone, and the system approaches that of a differential reactor. The reaction rates per gram of catalyst were determined from the relation F r = --z (1) W where F represents the feed rate of ethane to the reactor in moles per hour, W represents the weight in grams of the catalyst charged to the reactor, and z represents the fraction of ethane converted to methane. I n an actual run to determine reaction rates, the catalyst was first prereduced with flowing hydrogen, after which the reactor was cooled in flowing hydrogen to a convenient reaction temperature. At a standard set of conditions of hydrogen and ethane partial pressures, p H and p E , respectively, the activity of the freshly reduced catalyst was determined. Following this, reaction rates were measured at a series of temperatures in a rising-temperature sequence. The data are shown in the Arrhenius plots in Figure 2. Data are not shown for the 1% nickel on silica-alumina catalyst since the rates were too low to measure even at 540°,which is close to the limit of the apparatus. After determining the effect of temperature on rates over the freshly reduced catalyst, the temperature was lowered to an intermediate value in the range studied, and a series of measurements was made to determine the effects of the partial pressures of hydrogen and ethane on the rates. Since it had been observed from preliminary experiments that a series of such measurements over an extended period of time resulted in some loss of activity, it was decided to bracket all of the rate measurements with measurements at a standard set of conditions. In this way it was possible to detect variations in catalyst activity during the series of measurements. This procedure has been discussed in detail el~ewhere.~,GThe effect of a kinetic variable such as hydrogen or ethane partial pressure was then determined by comparing the rate at a given set of conditions with the average of the rates a t the standard conditions immediately before and after the period in question. For each set of conditions the rate r relative to the rate ro at the standard conditions can be expressed by the ratio r/ro, which should be reasonably The Journal of Physical Chemistry

I

10-4 1.6

I

1.8 2.0 1000/T(°K.).

I

1

2.2

2.4

Figure 2. Effect of temperature on rate of ethane hydrogenolysis over supported nickel catalysts of varying nickel concentration; PE = 0.030 atm.; PH = 0.20 atm.: 0, 10% Ni on Si024; 0, 5% Ni on SiOz; A, 1% Ni on SiOz; 0, 10% Ni on SiOz-AlzOs; U, 5% Ni on SiOz-AlZOa.

independent of moderate variations in catalyst activity. The value for r/ro is unity by definition at the standard conditions ( p =~ 0.20 atm., p~ = 0.030 atm.). Values of the relative rates r / r o as a function of ethane and hydrogen pressures are given in Table 11.

Table 11: Fklative Rates of Ethane Hydrogenolysis as a Function of Ethane and Hydrogen Partial Pressures r/m5 W

p ~ , & t m . PE, r t m .

0.10 0.20 0.40 0.20 0.20 0.20

0.030 0.030 0.030 0.010 0.030 0.100

N

i on SiO-

1% Ni6

5 % Nic

10% Nid

-Ni

5 % Ni8

on SiOFAlzO10% Nif

1.51 1.00 0.34 0.39 1.00 2.33

3.44 1.00 0.28 0.60 1.00 2.28

4.03 1.00 0.14 0.37 1.00 3.91

2.44 1.00 0.21 0.46 1.00 1.94

' Rate relative to the rate a t the standard conditions

2.66 1.00 0.27 0.33 1.00 2.40 (pH =

0.20 atm., p~ = 0.030 atm.) for the particular catalyst and tem-

perature in question; the r/ro values cannot be used by themselves to compare the activities of the catalysts. * Determined Determined a t 218". Determined a t 177-191°.4 a t 287". e Determined a t 304". Determined a t 246°.4

'

The data show that the rate of ethane hydrogenolysis increases with increasing ethane pressure but decreases markedly with increasing hydrogen partial pressure. The dependence of the rate on the partial pressures of ethane and hydrogen can be expressed in the form of a

CATALYSIS OVER SUPPORTED METALS

3861

simple power law, r = kpEnpnm. Approximate values of the exponents n and m as derived from experimental data are summarized in Table 111. It can be seen that the effect of ethane and hydrogen pressure is directionally the same for all the catalysts.

Table 111: Summary of Kinetic Parameters for Ethane Hydrogenolysis -----.--tsylatC a on SiO-

-Ni

Ni concn., wt. yo Apparent activation energy, kcal./mole Reaction orders" n, CzHs m, HZ Temp., oC.b

1

5

28.7

38.2

0.8 0.6 -1.1 -1.8 287 218

Ni on Si02-AlnOa

10

40.6

5 32.1

39.2

1.0 0.6 1.0 -2.2 -1.6 -1.8 177 304 246

2

0

10

4

6

8

10

Nickel content, wt. % of catalyst.

Figure 3. Effect of nickel concentration on apparent activation energy of ethane hydrogenolysis over supported nickel catalysts: 0, Ni on SiOz; 0, Ni on SiOrAlzOa.

Orders with respect to ethane and hydrogen in the power rate law, T = k p ~ " p ~ " . Temperatures a t which the reaction orders were determined.

Apparent activation energies derived from the slopes of the Arrhenius plots in Figure 2 are included in Table 111. In Figure 3 the apparent activation energy is plotted as a function of the nickel concentration of the catalysts. The apparent activation energy for ethane hydrogenolysis is seen to increase with increasing nickel concentration. A comparison of the specific catalytic activities of the various nickel catalysts at a given temperature can be made by dividing the observed reaction rate per gram of catalyst by the corresponding nickel area per gram of catalyst. A temperature of 205" has arbitrarily been chosen for such a comparison. For those catalysts on which reaction rates could not be measured at 205", the rates were obtained by extrapolation of the Arrhenius plots in Figure 2. The specific catalytic activity, rs, thus obtained is plotted as a function of the nickel content of the catalyst in Figure 4. It can be seen that the specific activity increases with increasing nickel concentration. Also, for a given nickel content, the specific activity of nickel on silica is higher than that of nickel on silica-alumina. No quantitative data on specific catalytic activity and apparent activation energy could be obtained on the 1% nickel on silica-alumina catalyst because it was inactive at temperatures up to 537". It is estimated that the specific activity of this catalyst at 537" is less than 4 X mole/hr. m.2of nickel. Assuming an apparent activation energy of 20 kcal./mole, which is an extrapolation of the curve for the silica-alumina-

1

, I

1

0 - V L - L ~ L I ' ' 0 2 4 6 8 Nickel content, wt. % of catalyst.

'

10

Figure 4. Specific catalytic activity of supported nickel as a function of nickel concentration: 0, Ni on SiOz; 0 , Ni on SiOZ-AlzOa. Conditions: 205"; p~ = 0.030 atm.; p a = 0.20 atm.

supported catalysts in Figure 3, the specific activity of the 1% nickel on silica-alumina catalyst at 205" would be less than 8 X lom8mole/hr. m.2 of nickel. This indicates that the specific activity for the silicaalumina-supported catalyst also continues to decrease with decreasing nickel content in the range from 5 to 1% nickel.

IV. Discussion The kinetics of ethane hydrogenolysis over the dilute nickel catalysts used in this work show some striking Volume 69, Number 11 November 1966

3862

differences from the results observed with catalysts of higher nickel concentration. The apparent activation energies observed with the dilute nickel catalysts are significantly lower than the values observed for the 10% nickel catalysts. I n addition, the specific catalytic activities are much lower. I n the case of the very dilute 1% nickel on silica catalyst, the dependence of the reaction rate on hydrogen pressure is also quite different. These results demonstrate that dilute nickel catalysts are very different from more concentrated nickel systems with regard to their catalytic properties. The data show an interesting parallel with the earlier results of Hill and Selwood5 on benzene hydrogenation over nickel on alumina catalysts. These investigators reported a critical nickel concentration (about 5%), below which the catalysts were inactive. They also found that the magnetic properties of the nickel in these catalysts varied markedly, the specific ferromagnetism increasing substantially with nickel concentration. They suggested that at low nickel concentrations few nickel particles are large enough to produce ferromagnetism and that the low catalytic activity results because the aggregates of nickel atoms are too small. In other words, they proposed that a geometric factor is involved such that a minimumsize aggregate of nickel atoms is required to catalyze the hydrogenation of benzene. I n the work reported in the present paper, the hydrogen chemisorption data do not indicate that the average crystallite size of the nickel decreases with decreasing nickel concentration. This is inferred from the values of the nickel surface areas, which do not show an increase in the surface area (per unit weight of nickel) at the low concentrations, as would have been expected if the nickel aggregates were smaller. To account for the large differences in the catalytic properties of dilute nickel catalysts from those of higher nickel concentration, it is suggested that the results are a consequence of an interaction of the metal with the support. Such an interaction could well be most pronounced at low nickel concentrations. The first increments of nickel impregnated on the surface of the support could selectively interact with the most energetic sites on the surface. This would lead to heterogeneity of the nickeI sites. There is a close analogy between this suggestion and the extensive evidence for the heterogeneous nature of surfaces as derived from studies on chemisorption of gases, The effect of the support on the catalytic properties of nickel is strongly demonstrated in comparisons of nickel supported on silica and on silica-alumina. The catalytic activity of nickel for ethane hydrogenolysis is markedly higher when the nickel is supported The Journal of Phyakal Chemistry

w. F. TAYLOR, J. H. SINFELT, A N D D. J. c. YATES

on silica. The effect is evident over the whole range of nickel concentrations investigated but is particularly pronounced a t the lowest nickel concentrations. The detailed nature of the interaction between the nickel and support is not known. However, it seems reasonable to expect that the metal would interact electronically with the support since electron transfer would, in general, occur a t the junction between a metal and a semiconductor. The general features of the kinetics of ethane hydrogenolysis have been reported previously by Taylor and co-workersBfor a commercial nickel on kieselguhr catalyst. They proposed a mechanism involving a preliminary dehydrogenation of the ethane to an unsaturated radical on the surface, followed by attack of the surface radical by hydrogen

C2HB $ C2Hz

Hz + C2HZ3 CH,

+ aHz

+ CH, E+CHI

where a is equal to (6 - 2)/2. Assuming that the dehydrogenation step is an equilibrated one and that the rupture of the carbon-carbon bond is the rate-limiting step, the following rate law results8

where b is the equilibrium constant of the initial dehydrogenation step and k is the rate constant of the subsequent step leading to carbon-carbon bond rupture. It is conceivable that the differences in hydrogen pressure effects observed with the 1 and 10% nickel catalysts may, in part, be due to differences in the extent of dissociation of hydrogen from the ethane molecule in the initial dehydrogenation step. Thus, while a appears to have a value of 3 for catalysts containing 10-15% n i ~ k e l , ~it? ~isJ possible that the value of a for a 1% nickel catalyst is lower, corresponding to a more saturated surface radical. This would account for the smaller inverse dependence of the rate on hydrogen pressure. This would imply that the properties of the nickel a t low concentrations are quite different from the properties at high concentrations, probably because the first increments of nickel impregnated on the support interact strongly and selectively with certain sites on the support. Furthermore, if the initial dehydrogenation step is endothermic,g the lower (8)A. Cimino, M. Boudart, and H. S. Taylor, J . Phys. Chem., 58, 796 (1954). (9) J. H. Sinfelt, W. F. (1965).

Taylor, and D. J. C. Yates, ibid., 69, 95

HEATO F FORMATION O F NITRICOXIDE@

3863

consequence of a lower endothermicity of the initial dehydrogenation step, corresponding to less extensive dissociation of hydrogen from the molecule.

apparent activation energy observed at the lower nickel concentrations can also be understood. The lower apparent activation energy could, in part, be a

The Heat of Formation of Nitric Oxide(g)

by Margaret A. Frisch and John L. Margrave1 Department of Chemistry, University of Wisconsin, &fadison, Wisconsin (Received M a y 88, 1966)

+

+

The heat of the reaction NO(g) CO(g) = l/tNz(g) COz(g)has been measured calorimetrically, and from these data the standard heat of formation of NO(g), AHtozg8, has been established as 21.556 f 0.060 kcal./mole.

Introduction Thermochemical studies of NO had their beginning late in the nineteenth century when Berthelot2 obtained 21.6 kcal./mole for the heat of formation of NO by treating cyanogen and ethylene with NO. Koerner and Daniels,s in 1952, studied the reaction of P and NO and their measurement gave 21.8 f 0.3 kcal./mole for a H f o i g s Of NO@. The derivation of this quantity from spectroscopic and electron impact studies was not so straightforward. For nearly two decades, the preferred dissociation energies of Nz and NO varied because of uncertainties in the interpretation of the spectroscopic data. An early study of Schmid and Cero‘ listed 4.29 e.v. for Do(NO). Glocker,6,*in 1948, chose Do(NO) = 6.49 e.v. and Do(Nz) = 9.765 e.v. At the same time, Hagstrum’ concluded that the only values consistent with both elect,ron impact and band spectroscopic results were Do(NO) = 5.30 e.v. and Do(Nz) = 7.384 e.v. Rosen,* upon examination of the problem, agreed with HerzbergQthat the value for Do(NO)must be 5.29 e.v., while Gaydonlofavored the higher value on the basis of temperature measurements on the cyanogen-oxygen flame.” In 1954, Brook and Kaplan12 presented very strong arguments that the higher value was correct since they had observed the vibrational levels of NO up to v“ = 23 which indicated that a sudden drop in

AG(2r”) values would have to occur in order to agree with the 5.29-e.v. value. Their data led to Do(N0) = 6.48 e.v. and Do(Nz) = 9.76 e.v. Herzberg, et ~ l . , ’ ~ made a second measurement and obtained 6.50 e.v. for Do(N0). Brewer and Searcy14reviewed the existing data and recommended the following as the most probable (1) Rice University, Houston, Texas, to whom correspondence should be addressed. (2) (a) M. Berthelot, Ann. chim. phys., (5)6, 178 (1875); (b) ibid., (5)20,255 (1880). (3) W.E.Koerner and E’. Daniels, J . Chem. Phys., 20, 113 (1952). (4) R. Schmid and L. Gero, Math. naturw. Anz. ungar. Akad. Wiss., 62, 408 (1943). (5) G. Glocker, J . Chem. Phys., 16, 604 (1948). (6) G. Glocker, J . chim. phys., 46, 103 (1949). (7) H. D. Hagstrum, J . Chem. Phys., 16, 848 (1948). (8) B. Rosen, Mem. 8oc. roy. SCi. Liege, 12, 317 (1952). (9) G. Hersberg, “Molecular Spectra and Molecular Structure. I. Spectra of Diatomic Molecules,” D. Van Nostrand Co., Inc., New York, N. Y., 1950,p. 448. (10) A. G. Gaydon, “Dissociation Energies and Spectra of Diatomic Molecules,” 2nd Ed., Chapman and Hall, London, 1952. (11)N. Thomas, A. G. Gaydon, and L. Brewer, J . Chem. Phys., 20, 369 (1952). (12)M. Brook and J. Kaplan, Phys. Rev., 96, 1540 (1956). (13) G.Herzberg, A. Lagerquist, and E. Miescher, Can. J . Phys., 34, 622 (1956). (14) L. Brewer and A. W. Searcy, Ann. Rev. Phys. Chem., 1, 259 (1 956).

Volume 69, Number 11 November 1966