The Pyrolysis of Formaldehyde1,2 - American Chemical

By Ralph Klein,361 Milton D. Scheer,3b and. Louis J. Schoen361. Received June 15, 1955. The pyrolysis of formaldehyde is shown to occur predominantly ...
15 downloads 0 Views 359KB Size
50

RALPHKLEIN,MILTOND. SCHEER AND LOUISJ. SCHOEN

VOl. 78

[CONTRIBUTION FROM BUREAUOF MINES]

The Pyrolysis of Formaldehyde's2 BY

RALPHK L E I N , 3 a

MILTON

D.

S C H E E R , 3 b AND

LOUISJ.

SCHOEN3a

RECEIVED JUNE 15, 1955 The pyrolysis of formaldehyde is shown to occur predominantly by a chain reaction. Carbon monoxide and hydrogen are formed in accordance with second-order rate laws. Methyl alcohol results from a homogeneous bimolecular reaction. An unspecified condensation product, amounting to a small fraction of the reacted formaldehyde, is also indicated. The thermal decomposition of mixtures of CHz0 and CDzO gives HP,DZand H D in proportions such that PgD/Pn2PD, is a constant equal to about 3.6. The pyrolysis of CH20 and Dz mixtures shows t h a t the hydrogen atom concentration, although small, exceeds that which would arise from the thermal decomposition of hydrogen alone.

Introduction

any, effect on the rate.4c6 We have confirmed these observations, using fused silica reaction vessels On the basis of pressure measurements, Fletcher4 concluded that the decomposition of formaldehyde differing in their S / V ratios by a factor of 9. The between 500 and 600" proceeded by a bimolecular, decomposition rates are, however, sensitive to the homogeneous process. Patat and Sachsse,s using previous history and treatment of the reaction an ortho-parahydrogen technique, estimated the bulb. The rates of pressure rise increase, although not regularly, as the number of runs increase. steady-state H atom concentration to be The most consistent data were obtained after the of the value computed on the basis of a RiceHerzfeld scheme and inferred that a chain mech- reaction vessel had been vacuum baked a t 800" anism was not applicable. However, the work of prior to a run. In Table 11, experiments 1 through Longfield6on the sensitized decomposition of form- 4 were made with baked out bulbs while in l a aldehyde suggests the existence of a reaction chain. and 4a about 20 runs had been made previously The present investigation was undertaken to without a bake out. Table I1 also shows the course-of-reaction data. determine unambiguously whether or not formaldehyde decomposes thermally by a chain mech- PCO- P H ,is taken as methanol. We have identified it as a product of the reaction mass spectroanism. metrically. Longfield had previously demonstrated Results conclusively that it occurs in the pyrolysis of If in the hydrogen isotope species formed in the formaldehyde. If the pyrolysis gave only CO, should equal A P . decomposition of rnixtures of CHZO and C D 2 0 Hz and CH30H as products, PH? little or no H D is present, the mechanism for the PH,is always somewhat larger than AP,indicating formaldehyde pyrolysis would certainly not involve a condensation reaction. The characterization of a chain process. A series of experiments was this condensation is not clear; it is, however, small. An upper limit for the H-atom concentration carried out in which 1 : l mixtures of CHzO and during the pyrolysis of formaldehyde was deterCDzO were pyrolyzed (the method of formaldehyde preparation and gas handling has been described mined. This was done by pyrolyzing a 10 t o 1 elsewhere7), and the hydrogen isotope distribution mixture of CHzO to D2g a t 165 mm. total pressure in the reaction products was determined. The a t 547". Mass spectrometric analysis of the experimental results are given in Table I. SubTABLE I stantial quantities of H D are formed. Samples of ISOTOPEDISTRIBUTION I N SON-COYDENSABLES the formaldehyde residue after a 30-second run HYDROGEN DECOMPOSITIOS OF CH?O-CD?O MIXTURES AT 537" gave the following analysis8: CH20, - 1 . ? 1 . 2 ~ ~FROM ; A N D 140 MM. TOTAL PRESSURE^ CD20, 50.6%; CHDO, 4.2%. The CHDO found is due completely to the CHDO present in the original CDzO sample. There is, therefore, no isotope 0.5 44.6 11.6 1 1 43.8 3.7 mixing between formaldehydes. 43.9 12.7 13.4 3.4 2 1 1.0 The kinetics of the pyrolysis of formaldehyde is 40.3 14,4 45.3 3 5 3 1 5.0 such that it is independent of the surface to volume -1.0 18.4 49.1 4 1 30.0 32.5 ratio of the reaction vessel. An increase of the 14.7 4 2 0.6 2.0 84.7 5 5 S / V ratio by a factor of 16 or 35 has little, if (1) This research was supported in part b y Headquarters, Air Research and Development Command, Office of Scientific Research, under Order No. CS-670-54-9, through Project SQUID. (2) Presented a t 127th meeting of t h e American Chemical Society, Division of Physical and Inorganic Chemistry, March 29 t o April 7 , 1955, in Cincinnati, Ohio. (3) (a) Olin-Mathieson Corp., New Haven, Conn.; ( b j General Electric C o . , Cincinnati, Ohio. ( 4 ) C. J . hI. Fletcher, P m r . R o y . S o r (/.o?ido?!),Al46, 3.57 ( 1 9 3 4 ) . (a) 1, P a t a t and H. %. Sachsse. %. Eleklrochein., 41, 493 ( l $ l 3 > ) . (.i) (6) J . 15. I.onyfield, Thesis, University of Rochester, lY.51 ( 7 ) hl. D. Scheer. "Fifth Symposium (International) on Coinbustion," Reinhold, 1955. (8) These analyses were performed through the courtesy of A. G. Sharkey and R. A. Friedel on a Consolidated Alodel 21.102 mass spectrometer.

9.8 51.2 39.0 3.0 0.2 5.0 CDzO, 96% in D, prepared from a deuterated polymer supplied by Tracerlab, Inc. Parts per 100 of the total hydrogen isotopes produced. Analyzed with Consolidated Model 21-610 mass spectrometer.

6

a

hydrogen isotopes a t the end of one minute gave D2, 13.77,; HD, 0.5%; Hz, 5.8%. The H-atom concentration was determined under the assumpz is a good approximation tions that (1) 1 / 2 P ~(final) in a one minute run, for the time average of PH* ( 2 ) H D arises from the exchange between HS and (9) Dz. purified by diffusion through a palladium thimble, analyzed 99.8% Dz, 0.2% H D .

THEPYROLYSIS O F FORMALDEHYDE

Jan. 5 , 1956

51

with respect to surface to volume ratio requires chain initiation and termination t o be both either in the gas phase or a t the wall. The difference in rate between “aged” vessels and vacuum baked vessels, together with the well established fact that termination in the photochemical decomposition of formaldehyde involves breaking a t the walls, l1 TABLE I1 strongly suggest that initiation and termination COURSEOF REACTION FOR FORMALDEHYDE PYROLYSIS AT in the thermal decomposition are wall effects. 547’ IS THE REGIONOF 140 MM. Fletcher has shown that the initial pressure PCH~OHdependence of the formaldehyde pyrolysis follows (Pco -, an approximately second-order rate law. If the Expt. P C H Z O ( i n i t i a 1 ) ,Ti,me, A P , Pcond.‘ PCO, PHz, PHd. mm. mm. mm. mm. mln. mm. second-order kinetics are maintained throughout mm. KO. the reaction, it can be shown that 6.9 1 . 0 1 2 . 1 108.7 2 4 . 2 17.3 138.6 1

Dz only. The interaction of D with HzCO is neglected. The H-atom concentration so calculated necessarily represents an upper limit. It can be shown that in a mixture of Hz and Dz, the initial rate of H D formation is proportional to the atom concentrationlo.

2 3 4 la 4a a

2.0 142.2 5.0 141.8 141.5 1 0 . 0 1.0 139.7 136.1 1 0 . 0 Reaction products

2 1 . 0 100.4 34.3 79.6 49.2 62.8 2 0 . 3 102.2 87.2 34.4 condensable at

36.6 57.3 75.6 31.9 100.2 -195’.

26.2 39.2 52.3 25.9 88.7

10.4 18.1 23.3 6.0 11.5

Therefore

\dt-)t

+

=

(%)t

-

= 4 0. thermal

x

1

-

1

P H ~ Pco

1

Fo’kt

1 +

where Fo is the initial formaldehyde pressure and k is a second-order rate constant. The linearity of the plot of the reciprocals of AP, P H and ~ Pco with the reciprocal of time, Fig. 1, confirms the hypothesis. It may be noted that because of methanol formation, the straight line plot does not extrapolate to l / F o on the ordinate. 0 09 0 08

+

0.07

0 06

‘i 0 0 5

h

# 004

v

10-3mm./min 0.03

(q)

= 0.5 mm./min.

t -

0 02

0

0 01

This gives an upper limit [(H)

AP

o, thermal

where [ (H) (D) ] refers to the atom concentration in the CH~O-DZmixture and [(H) (D)]thermai refers to the atom concentration in a hypothetical Dz-H2 mixture undergoing purely thermal exchange a t 547”. This mixture is presumed to have the same Dz pressure and one-half the final hydrogen pressure as developed in the c H 2 0 - D ~ mixture.

and

1

--=---s-

+ (D)] = 125 [(H) $. (D)]th~rmal

4.6 X 10-*mm.

0

Discussion

0.25

0 50

0 75

1.0

(min. ) -1. There is little doubt that the decomposition of Fig. 1 -The pyrolysis of formaldehyde at 547” initially at formaldehyde occurs by a chain mechanism. The 140 mm. pressure. evidence is (1) the formation of considerable H D from CHzO-CDzO mixtures in the absence of Methanol has been shown by Longfield to form exchange; (2) the effect of the nature of the in the formaldehyde decomposition by a secondsurface on the reaction; and (3) the radical sensi- order, possibly non-chain, reaction. We have tized decomposition of formaldehyde.6 Table I confirmed these observations indirectly. Methanol, shows the approximate constancy of the ratio measured by PCO - P H ~follows , second-order P~D/PHJ’D about ~ , 3.6, over a 25-fold variation of kinetics since CO and Hz are produced by secondthe initial P C H ~ O , / P C ratio. D ~ O This, together with order processes. The rate of its formaticn is inthe fact that mixed formaldehydes do not exhibit dependent of the extent of surface. If methanol isotopic exchange in pyrolysis, excludes a bi- is formed by a homogeneous non-chain reaction, molecular reaction as the mechanism of decomposi- second order in formaldehyde, less methanol is tion. The latter leads to a value P&D/’(PHJ’D~) expected the faster formaldehyde is consumed in a of one or less depending on whether a bimolecular concomitant chain process. Comparison of 1 and H D forming reaction between CHZO and CDZO l a , and 4 and 4a, Table 11, shows that less methanol would give only H D or HD, HZand Dz. Clearly, is produced in the faster reactions. This conformaldehyde decomposes thermally by a chain stitutes evidence that methanol arises from a process. non-chain, homogeneous process in the thermal The independence of the rate of the pyrolysis decomposition of formaldehyde. (10) R. Klein, M. D . Scheer and L. J. Schoen, THISJ O U R N A L , 78,

47 (1956).

(11) J G Calvert and E W R Steacie, J Chcm P h y s , 19, 17fi (1951).

Table I shows the approximate constancy of the ratio P $ D / ( P H ~ P D , ) over a 25-fold variation of the initial PCH~O/PCD~O ratio. Kinetically this implies, assuming formation of the molecular hydrogen species by interaction of a radical with a formaldehyde molecule, that the radical concentration is directly proportional to the appropriate formaldehyde concentration. That is, if dPan - - - k[(D)(HzCO) -I- ( H ) ( D X O ) l , dt dt

= k[(H)(HaCO)]and

%

There is derived from the mechanism

+

Stoichiometrically, CHzO = CO HT. P H ~ exceeds A P somewhat because of an unspecified condensation reaction. However, as shown in Table 11,it is relatively small. From the mechanism, and with the available data, it is possible to estimate kn

= k[(D)(D,CO)]

then

is a constant equal to 4 only if D and H are proportional to the first power of DzCO and HzCO, respectively. LVe have neglected the difference in rates due to isotope effects, although their inclusion would not alter the argument. X scheme for the thermal decomposition of formaldehyde, consistent with the data and observations, is CH20

+ CHzO- K1 wall

HCO

kz

H

HCOH

+ H?CO-Hz HCO

k3

Kd wall

-+

[COETRIBUTION FROM

+ H + CH20

+ CO + HCO

(I) (11)

k3 = 7 X l o 4 (mm ) - I sec.-l

On the basis of collision theory, an upper limit for the activation energy of the H HzCO --t Hz HCO reaction is calculated as 5.7 kcal. This agrees with the upper limit estimates of other investigators.lT

+

+

(12) K . H. Geib, Ergeb. E x a k l . Saliirwiss., 15, 44 (1936).

products

THE

Theref ore

PITTSBURGH, PA.

COLLEGE O F CHEMISTRY AND

PHYSICS O F THE PENXSYLVANIA

STATE UNIVERSITY]

Kinetics of the Hydrogen Fluoride Catalyzed Alkylation of Toluene by t-Butyl Chloride1 B Y A.

s. GOw, J R . , ~AND J. H. s I M O N S 3 RECEIVED APRIL4. 1955

Further measurements have been made on the reaction of t-butyl chloride a n d toluene as catalyzed by hydrogen fluoride. The reaction rate is greatly retarded by hydrogen chloride. The effect of five promoters was studied. It was found necessary to modify the equations and theory of Pearlson a n d Simons5 to correlate the rate measurements in a n y one experiment. The mechanism postulated for the reaction explains all of the experimental facts. I t is essentially a proton transfer mechanism.

Previous kinetic s t ~ d i e sof~ .the ~ alkylation reaction between toluene and t-butyl chloride as catalyzed by hydrogen fluoride present strong evidence against any ionic or free radical intermediates for this reaction. Two variations of a proton transfer mechanism in an amphoteric medium are presented. The previous work4,jhad shown that the rate of this reaction is proportional to the first power of the concentrations of reactants, and a high power (6-7) of the catalyst pressure. I n addition, i t was found that the rate is greatly increased by the presence of extremely small amounts of organic oxygen-containing compounds. An extension of (1) Taken in part from t h e Ph.D. thesis of A. S. Gow, Jr., T h e Pennsylvania S t a t e University, 1954. (2) Allied Chemical and Dye Corporation Fellow, 1948-1949. (3) Department of Chemistry and Chemical Engineering, t h e University of Florida, Gainesville, Florida. (4) J. W. Spraner and J. H. Simons, THISJ O U R N A L , 64, 648 (1942). (5) W. H . Pearlson and J . H. Simons, ibid , 67, 352 (1945).

these measurements appeared highly desirable. I n the present work the effect of catalyst pressure was studied a t three temperatures. The effect of five promoters also was studied. Experimental Preparation of Materials .-The reactants and catalyst were prepared as previously described.6 Anhydrous hydrogen chloride was prepared by dropping degassed, concentrated sulfuric acid (m.p. 10.0-10.5") onto sodium chloride ( C . P . ) in an evacuated system. The gas was dried by passage through anhydrous calcium chloride a t

-77". The promoters used were 2-fluoropyridine, trimethylacetic acid, trifluoroacetic acid, silver trifluoroacetate and water. 2-Fluoropyridine was prepared by the diazotization of 2-aminopyridine in hydrogen fluoride solution.6 The product was distilled through a n efficient fractionating column and a middle cut of constant index of refraction was stored over anhydrous potassium carbonate. Trimethylacetic acid (Eastman Kodak Co.) was distilled through a n (6) A. E . Chichihabin and RI. D . Rjozancev, J. Rusr. Phys. C h e m . S o c . , 46, 1.571 (1951).