G. J. MAIXS) A. 8. XEWTON AND A. F. SCIAMASKA
1286
TABLE I11 F ~ t a c r r o ~INs EQLTITJBRIUM IN THE LIQIJIDPHASE Catalyst
Temp., O C .
3-MH *2-:\IH Mole % 2-MH
ClSOIH H&Oa : ClSO8H AlCla
-33.46 06 207 258 36. 606
65.5 62.8 62 56-59 58.5 59.5
H?SO, (A1Br3: H,PO,) HzSO4
2,3-DMP F! 2,4-DMP Mole % 2,4DMP
i2.8
70.6 60 65-66 65.4 59.6
composition data. If degradation is not controlled, the product distribution is altered in favor of the isomers which best resist alkylation, cracking and isomerizat'ion. The most resistant, isomers are t'hose wbich cannot form t,ertiary carbonium ions and these are (6) C. P. Rlaury, R. L. Burweil. Jr., and R. H. Tuxworth, J . A m . Chem. SOC.,7 6 , 5831 (1954). (7) J . J. B. van Eijk van Voorthuijsen, Rec. trau. chim., 6 6 , 323 (1947). ( 8 ) 8 . K. Roebuck and B. L. Evering, J . A m . Chem. SOC.,7 6 , 1631 (1953).
Yol. 65
preferentially retained in the heptane fraction undergoing degradative reactions. This is illustrated by the data in Table 11 which compare the heptane isomerate a t equilibrium and after extensive cracking. 2,2-Dimethylpentane, 3,3-dimethylpentane and n-heptane build up preferentially under these conditions. Degradation of the tertiary isomers formed from 2,2-dimethylpentane also occurred after 10-30% conversion of this isomer. Since the rate of disappearance of the heptane fraction was approximately equal to the rate of isomerization of 2,2dimethylpentane a t these conversion l e d e , equilibrium could not be reached with this isomer. As noted above, degradation reactions lead to the formation of hydrocarbons containing more and less carbon atoms than the starting molecule. The large molecules are usually highly unsaturated, the light compounds contain isomers with four or more carbons. Isobutane aiid isopentane generally predominate, indicating the tendency of paraffins to crack ionjcally to yield tertiary structures. Possible mechanisms of degradation have been postulated and discussed by several investigator^.^
THE RADIOLYSIS OF BIACETYL VAPOR1 BY
GILBERTJ. A T A I S S , AMOS8.NEWTON'
AND
ALDOF.
SCI.4MA4KXA
Lawrence Radiation Laboratory, University of California, Berkeley, Califorma Received October 7 , 1960
The radiolysis of biacetyl vapor was studied at 25, 120 and 200' with pulsed electrons from a 4.2 Mev. microwave linear accelerator. Effects of pressure, pulse rate and total dose were studied. At room temperature the relative yields of methane and ethane were only slightly dependent on experimental parameters, but a t higher pressures the relative yields were pressure dependent. A free radical mechanism has been proposed to explain the formation of the major products and this is shown to account qualitatively for the experimental observations.
Introduction Preliminary studies of the radiolysis of some simple ketones as liquids and vapors have been but extensive investigations of these systems have not been carried out. Ausloos and Paulson3 have indicated that 85% of the methane produced in the radiolysis of liquid acetone could be explained in terms of a normal abstraction reaction by a methyl radical. The relative yield of methane t o ethane found in the vapor phase radiolysis appears to be too large compared to t'he ratio found in the photolysis of a ~ e t o n e ~t'o- be ~ explained completely in terms of thermal radical reactions. A similar effect was found in the radiolysis of methyl ethyl ketone arid diethyl ketone.3 In view of the results cited, a thermal radical mechanism by itself might be insufficient to account for (1) This work was performed under t h e auspices of t h e U. S. Atomic Energy Commission, (2) Author t o whom requests for reprints are t o be addressed. (3) P. Ausloos a n d J. F. Paulson, J . A m . Chem. Soc., 80, 5117 (1958). (4) J. D. Strong a n d J. G. Burr, ibid., 81, 775 (1959). ( 5 ) J. C. McLennan a n d JT.L. Patrick, Can. J . Res., 6, 470 (1931). (6) P. Ausioos a n d E. W. R. Steacie, ibid., 3 8 , 47 (1955). (7) TV. A. Xoyes, Jr., G. B. Porter and J. E. Jolly, Chem. Revs., 6 6 , 49 (1956). (8) R. K. Brinton a n d E. W. R. Steacie, Can. J . Chem., 38, 1840 (195.5). (91 1'. .-\iisloos, i h t d . . 36, 400 ( 1 9 3 ) .
the product distribution obtained in the vaporphase radiolysis of small aliphatic ketones. In order to investigate this point further, the radiolysis of biacetyl vapor has been studied over the pressure range 5 to 30 mm. a t 25, 120 and 200" by use of a pulsed electron beam current from a 4.2-Mev. microwave linear accelerator. In several experiments the pulse rate was varied, and in two experiments the current was varied. To determine microscopic dose-rate effects at 2.5' a few experiments were carried out with lower electron beam currents from a 2-Mev. Van de Graaff accelerator. Iodine scavenging experiments also have been run a t each of these temperatures.
Experimental Eastman Kodak White Label biacetyl was purified by a method similar to that of Groh.10 I t was dried by allowing the liquid to stand overnight under vacuum and in contact with pre-ignited anhydrous sodium sulfate. After it was dried, the biacetyl monomer was vacuum distilled from the polymer and degassed by trap-to-trap distillation in vucuo. I t was finally distilled onto another sample of pre-ignited anhydrous sodium sulfate in the storage bulb to stand overnight. The biacetyl prepared in this manner was found to be free of impurities by both mass spectrographic and vapor chromatographic analysis. It was stored in an ampoule on the vacuum line a t -80' until used. Aside from the formation of possible traces of photolysis products, it was
August, 1961
THERADIOLYSIS OF BIACETYL Va~oii
found possible to store solid biacetyl a t -80" for more than a month under thest: conditions without re urification. The Pyrex bombardment cells were cygnders, 5.4 cm. in diameter and 60 cni. long. One end of the cell was a thin concave window through which the electron beam was directed; the other end was fitted with a glass break-seal to facilitate analysis. The clean dry cell was evacuated and flamed. When the cell was cool, the biacetyl vapor was admitted to the desired pressure as read on an oil multiplying manometer. The cell was isolated and a sample of the biacetyl vapor checked by mass spectrometric analysis for the absence of air, water and other contaminants. After this check, the biacetyl in the bombardment cell was condensed in a liquid nitrogen cooled finger a t one end of the cell and the cell W ~ sealed B off. Irradiations were made using a microwave linear electron accelerator (Linac) a t the Lawrence Radiation Laboratory. This produces a pulsed beam of electrons with a pulse length of 5 microseconds and a pulse current variable from 10 to 100 milliamperes. The pulse repetition rate is variable. Most irradiations were made at 50 ma./pulse and 30 pulses per second. The electrons have a mean energy of about 4.2 Mev. and hence traverse the entire target cell in gas phase work. Because of :scattering in the windows, the entire cell, except for a small region a t the front, was covered by the beam, though variations in intensity undoubtedly occurred along the length of the cell. During this work it was found that the reproducibility of individual pulses from this accelerator probably was not better than i20Q/0, and variations in dose and dose rate were occurring within these limits. A regulating circuit later was added to the injector and deviations were reduced t o about &5y0. Most of the work reported here was done without the additional regulation. The Van de Graaff accelerator used was a 2 Mev. High Voltage Engineering Company machine a t the California Research Corporation. Owhg to insensitivity of the current metering systems in both the Linac and Van de Graaff installations, measurements of the integrated beam current a t the levels of operation used were only approximate. In early bombardments a t room temperature the cells, cooled only with ai1 air blast warmed to about 35" during the irradiation. In Iran de Graaff experiments the cells were immersed in a water tank to reduce the temperature rise to less than 1', and in later microwave accelerator bombardments the cells were fitted with water jackets and the wat,er temperature controlled to 25". If the bombardment was to be carried out a t elevated temperatures the cell was inserted into a thick wall brass tube, 70 mm. 0.d. and 85 cm. long. This was wrapped with three 300 watt heating tapes connected in parallel, then with asbestos tape and, finally, with glass wool matting. The power to the heaters was controlled manually by means of a variable transformer. The temperature of the system was determined by a thermocouple inserted near the middle of the brass oven. The temperahre variation along the length of the oven was less than 2 a t 200". After some experience, the temperature of the cell could be controlled to f 2 ' during bombardment. In experiments where I? was added, the cell was equilibrated a t 25" with 1 2 vapor, this was then frozen out in a side-arm tube connected with a stopcock, the biacetyl added to the desired pressure, frozen out and the IZ added from the sidearm. If more than 0.31 mm. 1 2 pressure were added, it was achieved by malting multiple equilibrations of the cell and respectively freezing each in the sidearm. Blank experiments in which the 1 2 was tit.rated with sodium arsenite indicated that the system used achieved 88% saturation qf 1 2 a t 25' in t'he two minute time allowed for equilibration. Blank experiments idso showed the thermal reaction between iodine and biacetyl a t 200" to be negligible in one hour of heating. Stopcocks through which Iz was passed were greased with Fluorolube JIG (Hooker Electrochemical Company). Some experiments a t 25' were run with no stopcocks in the loading system. These gave the lowest yields. After bombardment the cell was fitted with a stopcock and break-in device. The cell was then attached t o the inlet of a Consolidated Electrodynamics Corporation Model 21-103A Mass Spectrometer, the connecting line evacuated, the break seal ruptured, and the mass spectrum of the total gas determined. The cell was then removed to a vacuum line where the gaseous products were separated into three frartions, volatile at -l60", volatile at -119", and volatile at, -80". Thr PV of each fraction was measured and the
1287
mass spectrum of that fraction determined. The mass spectrum of the residual liquid also was run. Not all samples were subjected to all these measurements, but in all cases a t least a -160" fraction was analyzed. Most of the gaseous products were found t o be in the -160" fraction, which consisted of Hz, GHz, C Z H ~CO, , C2H, and COz. The -119" fraction always constituted less than 10% of the total products (frequently less than 5%) and was found t o contain mostly ketene (in the highertem erature runs) and lesser amounts of ClH2, CsHd, C Z H ~ , C3$? C3HB,C3Hs, COz and possibly C,H+ Because this fraction was small it was analyzed only for the ketene and Con. Since acetylene forms a complex with ketones and appears partly in the incompletely analyzed - 119" fraction, 10
,
, / . , I
I
i
1288
G. J. MAINS,,4.S. NEWTONAND A. F. SCIAMANNA
Vol. 65
then a single 5 psec. pulse a t 50 ma. current will yield an initial methyl radical concentration of 0.7 X 1OI2 radicals per ml. (assumed equal to the CO yield). Using the rabe constant for methyl radical recombination of 1 0 1 3 . 6 mole-' cc. sec.-1,11 the calculated time to reduce the methyl radical concentration to half its initial value is 22 milliTABLE I seconds. Because methyl radicals also disappear YIELDa OF PRODUCTS FROM THE IRRADIATION O F BrAcwrYL by reaction with other radicals, e.g., with acetyl, VAPORWITH ELECTRONS this probably represents a maximum value. G (product) at temp. specified Temperature 25' 1200 2000 If RcH!/Rc~H~~/' for each pulse rate (at constant Product pressure) is plotted against the time between pulses, H?b 0.55 0.55 0.55 and if each pulse is an individual event, then the CII," .22 1.4 5.9 points should follow a line proportional to t-lh. C& .l5 0.15 0.15 On the other hand, if the pulses are overlapping, C2H4 .06 0.1 0.2 there should be a deviation from this line. In CzH6 2.0 2.2 1.8 Fig. 1, this has been plotted and it is apparent co 7.4 9.2 16 that the deviations from the law start a t about CiHe 0.03 .. ... 30 milliseconds between pulses. Therefore the CaHs .06 .. ... mean radical life in this system must be less than co* .1 0.4 0.6 30 milliseconds. The agreement between the calCHtCWvd .5 0.7 2.0 culated value of 22 milliseconds and the curve is CHaCHO .3 0.4 0.3 sufficient to conclude that the bulk of the methane CHaCOCHa" 1.1 1.1 2.4 and ethane must be formed by thermal radical CHICH~COCOCIII 0.3 0.5 0.7 reactions a t room temperature. A small increase * Values listed are for irradiation with a pulsed beam, 5 in the ratio R c H ~ / R cwith ~ H ~time between pulses Msec./pulse, 50 ma. pulse current, 30 pulses/sec., 90 min. was observed even to times of one second between irradiation, and 20 mm. biacetyl (measured at 25'), and cell geometry described in experimental section. * Values pulses. This observation might be interpreted as normalized on H2 yield which apparently is dependent only indicating the existence of a few very long lived on amount of radiation absorbed. Yields probably are ac- methyl radicals in the system, these disappearing curate t o j ~ 2 0 7 ~ e. Value depends on radiation parameters. mainly by abstracting hydrogen from biacetyl. d Not found in all irradiations at 25 and 120'; behavior is Experiments at Elevated Temperatures.-The erratic. results of ten experiments a t 120" and four experiA few other products are present in small yield ments a t 200" are shown in Table 111. ,4lthough and analyses for them were not feasible nor was the ratio of Ht, CO, CzHO and COz do n o t change the identification certain in all cases. These with pressure, the relative yield of methane ininclude methylacetylene, diacetylene, vinylacetyl- creases with pressure a t both 120 and 200". At ene, butadiene, vinyl methyl ketone, acetic acid 200' the relative yield of ketene increases directly (yield a t 25" about equal to that of acetyl pro- with pressure. Hydrogen and ethane show no pionyl), and a t least two unidentified products temperature coefficient. The No. 88 series of boiling higher than acetyl propionyl. Effects of Experimental Parameters at Room experiments shows a small effect of total dose on Temperature.-The effect of pressure, total dose, the products. Whereas the rate of H2/C0 is and pulse repetition rate, and the effect of a con- constant, the ratio C2H6/CO decreases with total dose and the ratios CH4/CO and CO,/CO increase tinuous beam as compared to a pulsed beam were with dose. This would lead one to infer that a stmudieda t 25". Table I1 summarizes these data. reaction is competing for the ethane preThe outstanding feature of these results is the cursors. product At the and higher beam currents small effect of these experimental parameters on in expts. 87A andlower 87B, the ratios of CH,/CO and the relative yields of the various products. Within C2H6/C0 definitely indicate more methane to experimental error the yield of CZHs/CO, H2/CO result a t 10 ma. than a t 80 ma. beam current. and C02/C0 are constant. The ratio of C H d Iodine Scavenger Experiments.-The results of C2H6 increases slightly with increasing pressure and with total dose. This latter ratio is lower in adding Izvapor as a radical scavenger are shown in the Van de Graaff runs than in the Linac runs. Table IV. These results have been converted to On the other hand the ratio CH4/C2Heis higher in G-values for comparison with Table I. The yield the 12-liter cell where dead space exists. This of CO a t 25" was used as a basis for the conversioii result is expected where diffusion of methyl from the yield of product per unit time of irradiaradicals into areas of low radical density is possible. tion to G-values. The apparent agreement in R c H ~ / R c ~ H ~(here' / ~ ( B ) I n all cases the yield of CHJ is about equal to after designated as 2) between the Linac and the the yield of CO. While the CO does show a small Van de Graaff runs at the same pressure probably is temperature variation in the presence of 1 2 , this is fortiiitous. If a different pulse rate on the Linac not much beyond the variations in CO yield a t had been chosen, the agreement would not have 25O, and is much less than found in the absence of iodine. It is also curious that in several cases h e n as good. The effect of pulse rate gives a measure of radical the I2 uptake was measured and while the variations lifetime in the system. If one considers the dis- were fairly large (*20%), a t 25 and 120" the I, (11) R. Gorner and G. B. Kistiakowsky, ibid., 19, 85 (1951). appearance of methyl radicals to form ethane,
4 MeV. electrons. Such runs were limited to 25". As some product yields are dependent on pulse rate, total dose, pressure, or pulse current, these C-values should be used only as guides to the order of magnitude of yields under other conditions of irradiation.
THERADIOLYSIS OF BIACETYL VAPOR
A2ugust,1961
1280
TABLE I1 VAPORAT Roonr
~ ~ A U I O L Y S IOF S BracwrYL
Electrorl S O L L Ice Variable
Experiment no. Pressure (mm.)' Beam current (ma.) Bomb time (min.) Temp., "C.
TEMPERA'I'uRE -, linear accelerator +--Tutal d~se-----+ GWlll."
-Microwave
--+
I
+--------€'res~~~re----
851, 20
iOB 5
50 120 22.5
50 90 25-35
2 2 . 5 p a b 22.5* 30 30 25-50 25
50b 13.5 25
19.75 7.34 3.64 2.07 65.35 239.G 6.6
11.72 5.0P 1.92 0.81 38.77 144.7 1.29
1.68 0.77 0.26 0.41 7.32 25.8
4.77 1.61 0.49 0.28 15.96 63.1 0.54
7521 6
75H
75C 20
751) 30
20
8!)1i 20
89C
10
50 90 25-36
50 90 25-35
50 90 25-35
50 90 25-35
25 20 22.5
50 40 22.5
7.23 2.73
15.62 6.22
19.50 8.73
..
..
1.80 0.63 0.41 0.27 6.59 22.7 0.33
7.09 2.85 1.49 1.66 27.21 97.0 2.43
HP 4.77 CH, 1.73 Product C2H2d .. yields C2H1 1.02 (pnmoles)
RCRJ R ~ H , ,109 Rc,H,/Rco ,249 RH,/Rco ,076 [ RCH,/ R C ~ H ~ ' ' ~ [B)] X W 3 8.58
0.73 1.25 25.3 53.4 93.0 209.7 1.61 3.07 0.108 .272 .078
...
3.22 69.2 258.5 4.64
0.116 ,255 . 074
5.25 4.18 Irradiation cell was 12 liter spherical flask. ml. Minimum yield. Q
\
0.126 .268 ,075
b
'0
0.105 ,281 .073
,096 ,290 ,079
Van de Graafi
0.112 . 372 ,082
0.131 ,267 ,081
82C 20
T.29 2.20 0.24 1.07 22,d 07.2
...
..
0.105 ,284 ,065
0.098 ,230 ,075
84B 20
.lo1 .253 ,076
3.34 2.55 4.03 3.86 ... 9.70 3.95 5.11 Volume of cells = 1100 =t50 Continuous beam, beam current in panip.
TABLE I11 RADIOLYSIS OF BIACETYL VAPORAT ELEVATED TEMFERATVRE Variable Run no. Temp., OC. Pressure (rum.)" Beam current (ma.) Time bombardment (min.)
HZ Product yields (pmoles)
CHc CIH~* ClHi CnH4
CO
coz
Pressiire----+ 78B 78.4 78C 110-130 110-130 110-130 5 20 30 50 50 50 c -
90 2.62 5.92 0.5 0.5 10.4 52.8
. . .0
CHnCHO CHzCO iCHa)iCO cI&nd
...
CH4/CzIIe C*Ila/CO HdCO
0.57 197
[ R C I I J R C ~ I I ~ I :(~B ) 1 x 1012
...
... ...
11.4 42.6 1.4 1.2 46.3 223 7.9
...
11.0 25 0.8
90 24.3 113.8 4.6 3.2 93.1 489.6 23.5 6.7 39.8 10 6 6
20
40
88.4
4.48 9.11 0.8 0.4 20.6 74.9 2.5 3 5
... ... ...
7.16 17.35 1 5 1.1 32.7 124.7 4.6 7.0
...
10 8.8
88C 120 20 50 120 22.8 59.3 4.3 2.7 89.7 369.2 17.0 13 9
--Ihee 87.1 120 20 10
rate-+ 87B 120 20 80
450
4 ij
15 7 45 6
...
... 53.8 261.5 6.6 10.5
11.6 29.0 2 3 2.4 52.6 203.3 8 9 9.4
...
...
...
43 14.0
16 4.4
2 r, 11.0
,051
0.02 ,208 ,051
1.22 0,190 0 050
0.44 ,275 ,080
0.53 262 , 0.57
0.66 ,243 , OG2
0.85 ,206 ,060
0.55 ,250 ,057
3.6
3.1
3.9
2.10
2.24
2.67
1.36
2.77
Pressure measured at 25'. of the product in these runr. 0
90
120 20 50
Total doe-88B 120 20 50
c -
Vol. cells 1100 =i= 50 cc. in each case. iicetyl propionyl.
used is approx mately equal to the CHJ found. At 200" the 1 2 uptake was about twice the CHaI found, but the results at 120 and 200" are single experimeLtt 9. At room temperature, iodine reduces the yields of H?, CO, CkH2 and COz by about 25.30%. Methane is r c d u r d by 8,570 and ethane by 990/,. The methane rixilt was, however, extremely variable, and apparently trace impurities (unidentified) were able to greatly increase this figure. In one c a w ;t value of methane mas found in the presence of I, which was 2.5 times as large as that found in the absence of 12. Such large variations in the yield of methane were not observed in the absence of Iz, so the effect probably is specific for the presence of iodine. These observations cast a
* Minimum yield.
+
83.4 200 5 50
83'3 200 10 50
00
90
5.17 28.7 1 1 1.9 17.26 103.7 4.4 2.8 0.3 13 4.9 1.66 0.166 0.050 13.5
I'res.iure 83C 200 20 50
6.82 62.1 1.7 3.4 24.1 192.7 7.2 5.4 11.6 33 11.5 2.57 0.125 0.035 12.4
00
15.64 167.7 3.3 5.9 45.7 428.2 12.8 6.3 55.7 68 18.5 3.67 0.107 0.036 12.2
-+
83I) 200 30 50 90 19.76 279. G 9.8 7.5 79.4 709.7 22.6 6 9 131 106 22.7 3.52 0.112 0.028 10.3
Blanks refer to no determination
doubt on whether or not there would be any methane which would not be scavenged by I? in an ideal experiment. At high temperatures the amount of methanc not scavenged by I2increases slightly, but the teniperature coefficient IS small compared to that observed in the absence of iodine. One can thiii conclude that most of the methane and essentially all the ethane is produced by a precursor which i. scavenged by 1 2 . Presumably this precursor i. the methyl radical. The results on iodine uptakrl indicate the acetyl radical either to be not scavenged (which we find difficult to believe), or the acetyl iodide undergoes a further reaction to regenerate iodine, either in the cell or in the titration of the excess iodine after separation.
G. J. MAIXS,A. S.SEWTOS ASD X.1;. S c ~ a ~ s s s a
1290
\.oL 65
TABLE IV thermal radical reaction in z1 proposed mechanism YIELDS OF PRODUCTS FROM THE IRRADIATION OF BIACETYL for the radiolysis of azomethane vapor. Such a VAPORPLUS IODINE VAPORWITH ELECTRONS mechanism must account for the following results. (1) The relative yields of radiolysis products are Temp., "C. 25" 120 200 essentially dose, dose rate and pressure independent 1, pressure, mm. 0.3 0.9 0.9 at 25'. ( 2 ) Iodine scavenger reduces methane Biacetyl pressure, mm. 20.0 20.0 20.0 Product c
r -
G for product-------
Hz CH4 CJI2 CJL C*H6 C0 C3Ht C~HB
0.38 0.35 0.49 .03" .07 ,22 .1 .09 .14 . 03 .02 .05 .02 .02 .01 0.3 5.5 8.4 0,013 .. ... ,006 .. ... co, . 07 0.07 0.16 CHZI 5.9 5.2 7.7 I (atoms used) 6.4 5.3 14.8 Represents average values from 7 experiments. Values a t 120 and 200' are single experiments. 5 and 10 minute irradiations a t 25O, 15 minute at 120 and 200". Beam conditions 50 ma./pulse, 30 pulses/sec. * Result variable. Value listed represents lowest values found. c Scetone, ketene, acetaldehyde and acetyl propionyl were not studied.
Mechanism.-It is apparent from the studies a t 25' that most of the products must arise from a mechanism which is essentially independent of dose rate, tot,al dose and biacetyl pressure. Iodine scavenging shows t.he precursors of ethane and most of the methane to be scavenged a t all temperatures, and presumably these precursors are thermal methyl radicals. Ion molecule reactions for the production of methane do not' appear likely as only two peaks in the mass spectrum of biacetyl appear (from pressure dependence measurements) to have an ion molecule origin. One is a t mass 129 and presumably arises from t'he reaction CHaCO+
+ (CHjC0)z --+ (CH3C0)3+
(1)
whereas the ot'her is a mass 59 and may arise from the reaction CHa+
+ (CHZC0)2 +CH:j(CH3C0)2++ CH,CO + (CHa)&HO+
(21
It is difficult to see how these ion molecule reactions can contribute to methane, though reaction ( 2 ) might contribute to ketene and acetone. Similarly excited molecule reactions might contribute but little is known about react'ions of this type. However, products such as acetylene, diacetylene, methyl vinyl ket'one, and butadiene may arise from highly excited precursors. Except for hydrogen and a small fraction of the methane, t'he result's are indicat'ive of a thermal radical mechanism for the major products. Since such a mechanism has been suggested to account for photolysis and pyrolysis experiment^'"^^ it appears reasonable to modify these mechanisms to account for t'he radiolysis results. Stief and Ausloos'* have used (12) P.dusioos and E. W. R . Steacie, Can. J. Chem., 33, 39 (1955). (13) F. E. Blacet and W. E. Bell, Disc. Faraday Soe., 14,70 (1953). (14) G.F Sheats and JT..i. Noyes, J r . , J . Am. Chem. Soc., 77,1421 (1955). (15) G.F. Sheats and W. A . Noyes, Jr., ibid., 77,4532 (1955). (16) J. Heicklen and W. A . Noyes. Jr., ibid., 81, 3858 (1959). (17) E. IT. R . Steacie, "Atomic and Free Radical Reactions," Reinhold Publ. Corp., Yew York, N. Y.,1954,p. 354. (18) L.J. Stief and P. Ausloos, Paper presented at the 138th meeting of t h e Anier. Cliem. Soc., New York, September, 1960.
by about SO%/,,H2by 20% and ethane by essentially 100%. (3) The rates of hydrogen and ethane production are essentially independent of temperature while the methane shows a temperature coefficient which increases with temperature. The failure to completely scavenge the hydrogen may be caused by hot radical reactions, molecular detachment, or failure of 1 2 as an H atom scavenger. It is not easy to formulate reactions in which molecular detachment can occur to leave stable products. Therefore most of the hydrogen is postulated to occur by a hot radical mechanism and it is possible that a small part of the methane is also formed by a hot radical mechanism. Since ion neutralization might be expected to yield highly excited species, it is not unreasonable that such reactions occur. The following sequence represents the mechanism suggested Radical Formation.-
+ + + + + +
(CHYCO), +CH3 CO CHsCO +CH3CO CH,CO* --+ H* CHICOCOCH~ CHICO* +CH3 CO CH3CO +CHI CO H*+M + H + N
(3) (4)
(5) (6)
(7) (8)
Reactions 3, 4 and 5 should not be regarded a9 elementary kinetic steps since the formation of these species may be the result of ion neutralization as well as direct energy absorption. There is no eridence of reaction 8 in this investigation and reaction 7 becomes of importance only a t elevated temperatures. Sheats and NoyesI4 have shown their photolysis data to be generally consistent with an initial fragmentation according to reaction 3 rather than fragmentation into two acetyl radicals. Product Formation. Hydrogen.H*
+ (CHaCO), +H? + CHzCOCOCH3 H + H + >I +HS + H
(9) (10)
The bulk of the hydrogen must be formed by reaction 9 and its thermal analog or by a molecular process. Since the H atom concentration is low, the absence of a temperature coefficient for Hz production is not necessarily an unequivocal criterion for the presence of hot H atoms. The direct combination of H atoms with a third body is probably of little importance as the yield of Hz is independent of geometry and pressure. Furthermore, carbonyl compounds are known4 to be SCSLVengers for thermal H atoms so they should have a very short lifetime in this system. The ultimate products of a reaction such as reaction 11 are not known. H
+ (CH$O)z +CH3COCOHCH3 (?)
Methane.-
CH3* + (CHaCOz) +CH, CHI (CHIC0)Z +CH4 CH3 CHZCO+CH,
+
+
+ CHzCOCOCH, + CHzCOCOCH3
+ CH&O
(11) (12)
(13)
(14)
PHASEBEHAVIOR AND THERMAL PROPERTIES OF SH,F-HF
August, lMl Ethane. -
1291
tion 14 may be a significant rontrihiitor to methane production even at 200". A t all temperatures the material balance of Carbon Monoxide.-Reactions (3), (6) and (7). CH, 2(C2Hs) is about 5 5 7 , of the CO yield. If Acetone.-acetone arid acetyl propioriyl are added, about SOYo CHB CH&O +CHBCOCHB (16) of the CO yield is accounted in terms of methyl CH1 + IC"yCO'I> +CHjCOCH, + CH3CO (17) radical utilization. The fate of the remaining 20% of methyl radicals is not known. The mechanism Ketene and Acetaldehyde.described does not account for the unsaturated hy2CHiCO +CHZCO CHaCHO (18) drocarbons or for CO?. The latter compound does CH?COCOCHj --+ CHZCO CHaCO (19) not appear to be a secondary product as its formaAcetyl Propiony1.tion is approximately a linear function of total CH, + CH2COCOCHj --+ CHaCH?COCOCHX 120) dose and pressure. In experiments 87A and 87B. the effect of dose 1:xcepi for. the formation of H2 and the hot methyl rcacticin (12), all of these reactions have rate was compared and relatively more CH, was found a t the low dose rate than a t the higher h e n propoied iii photolysis studies. At 25", reaction 14 must account for the hulk dose rate. This is expected from the mechanism of the methaiie, n ith reaction 12 accounting for and the contribution of reactioli I S to methane less than 20% of it and reaction 13 another 20% formation a t 120". The effect of temperature and pressure on the or so if one awumes Xusloos and Steacie's value of 0.G X IO-'' for the function 2 to represent yields of acetone and ketene can be ascribed t o the limiting value of the methane formed by the increased importance of reaction 17 to the proabstraction by thermal radicals. These reactions duction of acetone and reaction 19 to the producthus account for the observation of a low activation tion of ketene at elevated temperatures. It is clear that the basic free radical mechanism energy near room temperature and for the relative which has been postulated to explain the photolysis independence of the methane/ethane ratio with pressure, doie rate, etc The ratio of CH4/C2He of biacetyl, can also be used to explain qualitatively was significantly higher in the 12-liter flask where the major products in the radiolysis of biacetyl. the radical could escape to a region of lorn radical Small contributions to methane a t room temperaconcentration where reaction 13 would become ture and most of the hydrogen production may be significant. Fieaction 14 has been shown to he of ascribed to "hot" radicals, though other mechaimportance in the room temperature photolysis nisms have not been completely eliminated. No mechanism has been proposed for the formation of biacetyl12 and acetone.6 At elevated temperatures, reaction 13 is expected of unsaturated hydrocarbons and carbon dioxide, to become the major methane source. The evidence which products are not reported in the photolysis indicates it t o do so. The activation energy for of biacetyl. Acknowledgments.-The authors wish to thank methane formation was an average of 4.7 kcal. between 25 arid 120", and 7.0 kcal. between 120 N r . William Everette for aiding in the electron and 200". This latter figure agrees with the irradiatioiis with the microwave linear accelerator, photolysiq resiilts of Blaret and Bell,', but is lower and to Dr. K. L. Hall and Mr. Norman Shields of than those of Ausloos and Steacie. This, together the California Research Corporation for use of with the high values of the function Z a t 120 and their Van de Graaff accelerator and for aid in the 200" compared to those of Ausloos and Steacie12 irradiations with it. The authors also wish to and Sheats and Xoye715 indicate that under the thank Dr. Peter Ausloos of the National Bureau of high iiitwiiity conrlitioiis of our irradiations, reac- Standards for helpful suggestions and criticisms. CHa
+ c& +C2Hs
(15)
+
+
+
+
PHASE BEHAVIOR ,4SD THERMAL PROPERTIES OF THE SYSTEM KHdF-HF' BY ROBERTD. EULER
EDGAR F. WESTRUM, JR.
Deportment qf Chemistry, University of Michigan, Ann Arbor, Michigan Received October 12, 1960
The system NHIF-HF \?-as studied by thermal analysis between the limits XHIHFP and HF. S o indication of the composition NHilH2F3previously reported by Ruff and Staubc was found. Low t'emperature heat capacity measurements on four conipositions approximating N H ~ H B Fconfirmed , and extended the thermal analysis and revealed the existence of a solid solution at this composition. Several thermal anomalies were found between 180°K. and the melting point. The decomposition pressure of YHdHZF, \vas also determined.
Introduction Ruff alld St:iube? investigated the KH,F-HF sYs(1) From the dirmertation of R. D. Euler submitted in partial fulfillment of the requirements o f the Doctor of Pllilosophy Degree at tlie University of hlichigan. (2) o. ~ u f and f I,. stauk z. anorg. allgem. Chem.. 219, 399 (1933).
tern over a composition range from 50 to 85 mole yo HF. Their data indicated the occurrence of the congruently melting compounds XH4HF2, SH4HzF3, NH4H3F4and KH,HZFg, as well as a' solid phase invariance occurring st - 3" extending over the entire composition range investigated.
