Grl’j-I’fIASE
HOXOGJWEOTX CATALYSIS I N SHOCK W.4VES
2833
Gas-Phase Homogeneous Catalysis in Shock Waves. 111. The Oxidation of Carbon Monoxide by Oxygen in the Presence of Nickel Carbonyl by Shimpei Matsuda‘ Department of Chemistry, Harvard University, Cambridge, Massachusetts
08158
(Receiued March 3, 1972)
‘The osidrhtion of carbon monoxide by oxygen in the presence of small amounts of nickel tetracarboiiyl was irrvc?stigated in incident and reflected shock waves. it reaction mixture, typically containing 300 pprn of Ki:(’O)a: 4yoCO, and 2% 0 2 in Ar, was shock heated to temperatures between 1100 and 2400’K at li total roilcentration of 6 x 10” molecules/cc in incident shock waves. The COS formation W:LS followed by the infrared emission at 4.27 p over 2-msec particle time. The COZ concentration was found to increase linearly wit,h time, the rate of CO oxidation being much greater than in the absence of Ki(CO)+ A time-of-flight mass spectrometer coupled to a shock tube was used to investigate the formation of transient species containing Xi during the reaction behind reflected shock waves. Mass peaks due to N i 2nd NiO vere identified, the peak height, of the latter species being less than 10% of the former. A cyclic process involving ?si, NiO, and NiCO is proposed to account for the homogeneous catalytic oxidation of CO in the presence of Ni(C0)4.
Introduction The acceleration of combustion reactions by the addition of 5niali amounts of metallic compounds has bcrn a subjmt of scvrral studies. Thus induction periods of the flash-initiatd CzH2-02 rraction were shortcncd by the additlion of Si(CO)d, I?C(CO)~, and Cr(CO)G.?J Thr Cz€l,-02 rcaction in shock waves iu-hich i u a striclcij hornogcncoux system) is greatly enhanced in the prt I I C ~of only 10 ppm of Cr(C0)6.2b Thc strong catalytic rffccts of Cr(C0)S and Ice(CO), on thc oxidation of carbon monoxide in shock waves have been throughly -tudicd by Kistiakowsky and co~ v o r k e r s . 11) ~ ~ they ~ C L ~ V of the Cr(C0)6-CO-02 reaction thc formation of Cr, CrO, Cr02, and CrOs in steady-statc concmt-rationswas cox1firmc.d by means of a time-of-flight m a s iprctrometer coupled to a shock tube and a catalytic chain process CrO, i0, --- CrOn+2
CrO,,
+ CO
=
CrOn-,
+ C02
(n = 0, 1)
(1)
(m = 1, 2, 3) (2)
liar been p r o p o ~ r d . ~A similar mechanism was proposed for thr FrCCO)~-CO-02 r e a ~ t i o n . ~ The spontaneour ignition of Si(CO)., as well as the combustion of ti-butanc induced by Ni(C0) 1 have been studied by scvcral .~r.orkrrs.~-~ The mechanism, howw r r , is not 1tcll undcrstood, mainly because of the hctcrogencous nature of the reaction. Sickcl oxide as n-cll as Fez03 and CrLOaare good catalysts for the CO oxidation in the gas-solid hcterI” Thc purpose of this ogencous rractior 31 p p c r i+ to prtwnt drtails of the homogeneous catalytic cffcct of Xi, a d d d in thc form of Ni(CO)4, on tht. CO-02 rcwtion. Shock waves have proved to be tho rdral nicdia for this tvpc of study.
Experimental Section Apparatus. A 3-in. i.d. shock tube cquippcd with two InSb infrared detectors (IRST apparatus) has been described in detail previously.!’ l 2 One infrared detector monitored the rmission at 4.20 p from COz and the other tkic emission at 5.03 p from CO. The emission intensity at 4.20 p was calibratcd t o absolutc COS concentration by shock heating ltnorvn COz--hr mixturrs ovcr a tcmpcrature range uscd in this study. Shocks were gencrated in 5 or 10 Torr of reactant mixtures using Hr or Hc 1-2 a5 a drivcr gas A 1-in. 1.d. shock tube coupled to a timcwf-flight mass spcctrometcr (TOIST apparatus) has also been described prcviously. l 1 l 4 Consrcut ivc mav, spectra of a shockcd gas flowing into thc ion sourcc of the time(1) Present address: Department of Chemistry, Cornell University, Ithaca, N. Y. 14850. (2) (a) K. Erhard, Z . P h y s . Chrm. (Frank.frLrt a m M a i n ) , 36, 126 (1963); (b) S. M-latsuda arid 11. Gutman, J . Phys. Chem., 75, 2402 (1971). (3) T. P. J. h o d , G. B. Kistiakowsky, and S. Matuuda, J. Chem. Phys., 56, 1083 (1972). (4) S. Mstsuda, ibid., 57, 807 (1972). (5) E . J. Badin, 1’. C. Hunter, and It. N. Pease, J . Amw. Chem. Soc., 70, 2055 (1948). (6) A. 1’. Gwrat and 1%. W. Thompson, ,J. Chem. Soc., 1822 (1934). (7) C. E. H. Bawn, Trans. Fa’nraduy Soc., 31, 440 (1935). (8) A . Egerton and S. Rudrakanchana, Proc. 12oy. Soc., S e r . A , 225, 427 (1954). (9) NI. Katz, Aduan. Cntal., 8, 177 (1953). (IO) J. K. Dixon and J. 12. Longfieltl, “Catdysis,” Vol. 111, 1’. 1%. Emmett, Ed., Iteinhold, New York, N. Y . , 1960, 1) 281. (11) A. LM. Dean and G . B. Kistiakowsky, .I. Chrm. Phus., 53, 830 (1970). (12) J. B. Homer and 0. B. Kistiakowsky, ibid., 46, 4213 (1967). (13) .J. E. Dove and D . McL. Moulton, I’roc. IZoy. Soc., Ser. A , 283, 216 (196.5). (14) I. D. Gay, G. l3. Kistiakowsky, J. V. Michael, and R. Nilci, .I. Chem. I’hys., 43, 1720 (1965).
T h e Journal of Physical Chemistvy, Vo2. 7 6 . .Vo. SO,1972
2834
SAIMPEI MATSIJDA
of-flight m a spectrometer through a pin hole in the end wall of the shock tube were displayed on oscilloscopes every 20 psec over a total observation time of about 200 psec and recorded photographically. Mass spectra covered a range of m/e 50 to 110. The ionizing electron energy was 22.5 eV in all experiments. Material. Matheson research grade Cot, Os, and Ar were used without further purification. Matheson research grade CO was passed through a packed column cooled a t liquid nitrogen temperature before use. Matheson tank Ni(CO)4 was vacuum distilled several times before use. I n TOFST experiments Matheson research grade premixed O r N e (8/92) gas was used without further purification. Preparation of Gas Mizture. Mixtures of Ni(CO)4 and OZ ignite spontaneously after an induction period a t room temperature.'4 Therefore, mixtures of Ni(CO)4-Ar and CO-OrAr were prepared and stored separately. All gas mixtures used in IRST experiments are listed in Table I. Mixture A was prepared from a mixture containing 3% Ni(CO)r in Ar. For each series of experiments with one of mixtures B-F a new sample of mixture A was prepared in order t o avoid the change in Ni(CO)4 concentration during storage, because the dissociation of Ni(CO)4 might be appreciable even at room All gas mixtures were used within 48 hr of preparation.
Table I: Composition of Reaction Mixtures Mix-
NKCOh.
co.
t"rm
PPm
%
A B
570
C
D
E F
8.04 12.8 4.42 7.47 8.07
0%.
+
Al.
70
%
4.03 4.12 8.58 7.10 4.22
99.9 87.9 83.1 87.0 85.4 87.7
In IRST experiments the mixing of mixture A with one of mixtures B-F was carried out in the vacuum system (-2 1.) about 3 min before each shock experiment and the thus prepared Ni(CO)rCO-OrAr mixture was introduced into the shock tube approximately 1 min before shock firing. Usually a CO-OrAr mixture was introduced into the vacuum system first and then about an equal amount of the Ni(CO)4-Ar mixture. Since the duration for mixing is likely to be too short for complete mixing, the order of the introduction of two gases was reversed in some experiments. As will be shown later the rate of COZ formation was not affectedby this. In TOFST experiments 0.2 Torr of Ni(CO)4 vapor was mixed with 40 Torr of the O r N e gas in the vacuum line about 5 min before each shock experiment. The N i ( C 0 ) r O r N e mixture was then introduced into the The J r m d of Phvriml Chnniahu. Vol. 76. No. #O,lW$
(b) Figure 1. Reaction profiles of the Ni(CO),-CO-O, reaction at high and low temperatures. (a) For reaction conditions see Table I. Upper trace, CO emission at 5.03 p, 100 mV/division; lower trace, COZemission at 4.20p, SO mV/division; one division corresponds to 3.3 x 10" CO, molecules/cc. (b) mixture B A, 7 ' , = 1268'K, (Ni)o = 1.66 X lo", (CO) = 2.34 X 10'8 molecules/cc. Upper trace, CO emission, 20 mV/division; lower division, C02 emission 20 mV/division; one division corresponds to 2.6 x 10" CO, molecules/cc. Sweep time is 50 psec/division for all traces. Time increases from left to right. Arrival of shock wave8 is indicated by B sharp rise in the CO2 signals.
shock tube and the experiment completed within 1 min. It was found with the mass spectrometer that no COZwaq formed during the 5 min of mixing.
Results Figure 1 shows the typical reaction profiles in IRST experiments. The profile given in Figure la was observed in all experiments performed ahove 1600°K. The COZ emission signal is seen to rise sharply upon shock arrival; after 40 psec particle time the CO, concentration increases a t a slower constant rate which is still much greater than in the absence of Ni(C0)4." The initial sharp rise in the COzsignal w m unobservable above 2200'K. The amount of COz formed in this initial stage of the reaction was greater at lower temperatures. The linear growth region in the (Cot) us. (15) . . A. M i t a c h . 2. Phva. Chm... 40. . 1 (1902). . . (16) J. P.Day. R. G . Pearson, and F. Basolo, J . Amer. C h n . Sac.. 90, 6933 (1968).
2835
.-
~
~
-
-
Table TI : Rate of COZFormation in the Ni(C0)4-CO-Q2 Reaction (Ni)o* X 10-14,
T2I
moleoules/
OK
uo
10-18,
x io+,,,
molecules/ cc
mnIecules/
ec/mo!ecules
W 6W3
seo
2.41 2.44
6.40 6.05 3.43 3.74 2.92 2.46
2390 2309 2070 1960 1882 1749
Mixture B A (1: 2.37 1.59 1.81 2.23 2.00 2.32 2.06 2.62 1.68 2.27 1.57 2.28
2404 2137 2014
1.62 1.76 1.67
1.59 2.02 1.12 0,815
(OK-')
Figure 2. Plots of K ( = rl(CO,)/dt/(Ni),(CO)) vs. IO,OOO/T ( K i s in units of ec/molecules sec): A, mixture C A; 0 , mixture B -1- A; 0, niixture F A; 0 , mixture E A; , mixture D A.
+ +
+
+-
R/fNi)o(CO)
10-15,
+-
4000Qi f
RE" x
ICO) X
time plot was used to calculate the rates of COz forma'Gion. These rates of COz formation a t high temperalures as well as experimental conditions are summarized in Table 11 and the rates of CQz formation normalized by (Ni)o(GO), where (Ni), is the initial conccntration of Ni(CO)r, are plotted against T-1 in Figure 2 . Figure Ih shows a typical reaction profile at lower temperatures ( Ni Ji- 4CO
+ O2 +NiO + 0 AHo" = 31 kcal/mol NiO + CO -+ Ni + COz -39
Ni
602
+0
(6)
(WiU) s B / (Nilss
1
____
IC;(Oz) =
+ i) (11) ks(C0)
k6(02) /k6(CO)
(111)
Since it was found that (NiO)88/(I?i)ss$ l c 5 ( 0 2 ) , eq IT reduces t o d(CQd/dt = W W O J
(n = 1-6)
(8)
Wi(CO)n = Ni(C0).-1
+ CO
(n -= 1-4)
(9)
remains significant to higher temperatures than in the Cr(CO)6 arid P"e(C0)j cases, although NiCO stays in such low concentrations that is not reliably observable by the present. TOFST apparatus. The following mechanism is proposed R'i
-.C
60
zNiCO
+ 0%-+ NiQ -INiO + c0 --+ Ni +
NiCO
(101
@02 a 3 2
(11) (12)
The concentration of Ni(CO)z being very low, it is not included in the mechanism. Vsing a steady-state assumption for Xi, NiO, and Xi 0 (steadv-state concentrations are attained within 5 psen,, see also Figure l a ) the rate of C02 formation is obtained as
(7)
+
-.
+G
Cr(CO).-l
is measurably reversible a t these low tmnperatures. The average metal-carbon monoxide bond energy in Cr(COJe, Fe(CO)6, and Ni(CO)I i s esfimatod t o be 27.1, 27.7, and 35.2 kcal/mol, resprcLively.21 Thus, since the metal-carbon monoxide lsond is considerably stronger in Kr(CO)s, it may be suggested that the reverse reaction in
(5)
has been specifically excluded from the present scheme because the rate of COZ production in reaction 7 is less *ban 5% of the observed rate of GO2 formation even at ihe highest temperatures employed in this study.l1 Cling a steady-sta,te assumption for Xi and NiO con(centrations and a mass balance equation, i.@., (Ni)ss (NiO),, = (Xi)@,we obtain d(C0)Z dt
Cr(CO),
(4)
For the calculation of the heat of reaction the dissociation energy of KfO was taken from ref 18 and the heat of formation of other species from ref 19. It should be noted that the reaction i0 2 =
slowly, being observable even a t 150 psecZo This suggests that the reaction
(IV)
which is not consistent with the experimental finding that the rate of C 2 formation is strongly dependent on the CO concentration (see eq I). The study of the Cr(CO)s decomposition (0.4% cr(CO)6 in Xe) using the TOFST apparatus showed that the CrCO+ and Cr(CO)2+peaks disappeared completely within 40 psec of the shock reflection above PSOO"K.3 At lower temperatures (
