VAPOR-PHASEPHOTOLYSIS OF FORMIC ACID
June, 19G1
1033
TABLE VI TWO-STAGE EXTRACT ON Extraction 1 Phase
-TICI,
mg.-Calcd.
... ...
57 143
Found
KNOa 4gC1
AgCl A N D LEACHWITH KN03 OF THALLIUM CHLORIDE 200 mg. of TlCl in 50 g. KNOs
WITH
-Extraction Phase --*
+
KNO, AgCl
association constant for thallium chloride was so much smaller. Also, the cubic TIC1 seemed to be highly soluble in the silver chloride phase since the distribution constants (light phase/AgCl) were SO low. Another cubic compound which was found to have a high solubility in silver chloride was potassium chloride. I n the series of experiments when potassium chloride was added, it was found that the observed distribution coefficient was 1.2 in the KN03-AgC1 system and 0.4 in the K2S201AgCl system. It was also found that the high distribution coefficients observed for lead chloride (PbC12) were decreased by the addition of large amounts of potassium chloride (the results of these studies will be published later). The possibility of using this method for solvent, extraction processes was shown by carrying out a two step extraction and leach of thallium chloride in the KNO3-AgC1 system. The results are shown in Table VI. Two hundred mg. of TlCl was dissolved in 50 g. of KY03 and 2 g. of KC1, and the mixture was contacted with 50 g. of AgCl. The phases were separated, and the KN03 phase was contacted a second time with 50 g. of AgCl and 2 g. of KC1. Less than 10% of the thallium remained in the KN03 phase after these two ex-
Phase
Leach ---TlCl, Found
KNO3 AgCl
55 146
2 - 7
YTlCI,m S . 7 Found
Calcd.
16
17 40
..
+
m g . 7 Calod.
67 122
IO,
1
I
A
I IO
30
20
'/,cl-is,
4 0
kglrnoie
Fig. 3.-Determination of true distribution constant and association constant in AgCl-KzSzO, system.
tractions. The amount found checked well with the amount calculabed from the equilibrium constants. The two AgCl phases were then leached with 100 g. oE KNOI, and again the amount of thallium in each phase could be calculated from the constants. Thus, distribution experiments in fused salt systems mill shed light on the nature of fused salts and the factors which control solubility of solutes in them. These experiments may also point out new mays to solve separation problems.
VSPOR-PHASE PHOTOLYSIS OF FORMIC ACID' BY RALPHGORDEN, JR.,AND P. AUSLOOS Division of Physical Chemistry, National Bureau of Standards, Washington, D . C. Receibed Februory 8 , 1961
The direct photolysis and the Hg( 3P1)photosensitized decomposition of HCOOH and DCOOH have been investigated in the presence and absence of radical scavengers. The results indicate, in contrast to those of earlier investigations, that hv HzO CO ( I ) and HCOOH besides reactions I and I1 which involve intramolecular rearrangements HCOOH hv -c HI COz (11) radical-producing primary processes must be considered as well. It will be shown that process HCOOH hv +. HCO OH is the most important source of free radicals in the direct photolysis of formic acid. The effect of temperature on the rate of formation of the products was investigated in some detail. hbove 200' three different chain processes were found to occur: (a) H HCOOH HzO HCO, HCO .-,H CO: (b) H HCOOH +. HZ COOH, COOH -+ COZ H; (c) H HCOOH -+ HCOO Hz, HCOO +. H COz.
+
+
+
+
+
+
+
+
+ +
-+
Introduction Earlier investigations using both direct photolysis and mercury-photosensitization indicate that no free radicals are produced from HCOOH. Ramsperger and Porter2 explained the formation of the products CO, COS, H2 and H20a t the wave lengths between 2260 and 2500 A. by two intramolecular rearrangements (1) This research was supported by a grant from the U. S. Public Health Service, Department of Health, Education and Welfare. Part of this work was presented by one of the authors (R.G., Jr.) in partial fulfillment of the M.S. requirements of the American University. (2) H. C. Ramsperger and C. W. Porter, J . Am. Chem. Soc.. 48,
1267 (1926).
+
-+
+ +
HCOOH HCOOH
+
+ hv -+- HZ+ CO? + hv +H20 + CO
+
(36%)
(64yo)
(I) (11)
Gorin and Taylor3 concluded that a t wave lengths between 1900 and 2540 A. the dimer decomposed to yield only COZ and H2. They found no significant evidence for H-atoms in experiments carried out in the presence of parahydrogen. Burton4 drew the same conclusion from experiments based on the mirror techniques. Bates and Taylor6 investigated the mercury(3) E. Gorin and H. S. Taylor, ibid.. 8 6 , 2042 (1934). 14) M. Burton, ibid., 58, 1655 (1936). ( 5 ) J. R. Bates and E.S. Taylor, ibid., 49,2438 (1927).
1034 TABLE 1 PHOTOLYSIS OF HCOOH
Mole/cc. X loe-HCOOH ----Ecav.---
7 -
0.97 .97 .97 .97 .97 .97
None [CZH,] [0218] Kone [Op] [Op]
0 030 ,014
,018 .012
AT
30'
Rate, cc./min. X 1 0 3
7
Hz
CO
Cot
0.098 ,017 ,016 1.29 0.024 ,020
0.17
0.19 .14
.14
.13 5.55 0.79 .38
C~HIO
ClH6
0,036
0.081
1.9 2.2 3.0
Type of irradiation
Direct photolysis with Corning-fiker 9-54 Direct photolysis with Corning-filter 9-54 Direct photolysis with Corning-filter 9-54 ( 3P1)Hgsensitized ( 3P1)Hgsensitized ( T1)Hg sensitized
TABLE 11 EFFECTO F SC,4VENGERS T, 'C
2 2 101 100 203 203 36
35 34 35 70 70
-Mole,'rc x 106---. DCOOH - - i c a ~ --
0 38 0 38 1 20 1 29 1 24 1 27 10 1 0 10 10 3 81 3 81
,--
Hz
HD
Dl
0,0049 0.033 0.014 ?;one ,0030 ,032 ,015 [COa] 6 55 Sone .99 ,085 ,017 .10 ,011 ,038 [O2"] 0 056 None .72 1.5 .41 [OZ13] 034 .042 0.19 .12 [C&] 031 ,0030 ,0098 ,0011 .0016 ,0035 .0089 053 None .lo7 ,031 .26 ,021 .21 [CZEII] 031 ,099 None ,020 ,066 ,0031 [I21 0 47 .0019 .MI16 .00066
Rate, cc./min. X 103 CO CO? Ethane CaHaDz CIHgD
0.23 .23 1.5 1.7 1.8 2.6 0.13 . 11 2.9 3.1 0.055 ,055
photosensitized decomposition of formic acid at 2537 A. and found process I1 to be three times more important than process I. More recently, Kebarle and Liossing6 reinvestigated the Hg( 3P1) photosensitized decomposition of HCOOH at low pressures (10-25 p ) in a specially designed reactor connected to a mass spectrometer. In this study the products were explained exclusively by the intramolecular rearrangements I and 11. The relative distribution of the products was similar t o the one found by Bates and Taylor. In the present study the direct photolysis and Hg(3P1) sensitized photolysis of HCOOH and DCOOH has been reinvestigated in an attempt to establish the importance of free radical production in the primary act. The fact that free radicals are produced in the photolysis of acetic acid7 and methyl formate,G* as well as in the liquid-phase photolysis of formic acidg seemed to justify a more extensive inresi igation. Experimental Apparatus.-The gas phase was photolyzed in a quartz cell Y 10 cm. in length, 5 cm. in diameter) enclosed in a heavy aluminum furnace provided with double quartz windows. The tcmperatuie was automatically controlled within one-half of a degree. A Hanovia S 100 lamp was used in the direct photolysis experiments. In the Hg(3P1) sensitized experiments the source was a flat-spiral, low-pressure mercury arc. Unless otherwise stated, a Corning-filter 9-54, transmitting radiation a t wave lengths greater than 2200 A. was used. Conversions were about 1% in all experiments. A few y-ray rttdiolysis experiments were performed with an experimental arrangement identical to the one described before .lo Analysis.-The analytical section of the vacuum system tbonsisted of a sequence of traps, a modified Ward still, and a - ___(6) P. Kebarle and F. P. Lossmg, Can, J. Chem., 37,389 (1959), (7) P. Ausloos and E. W. R. Steadle, t b t d . , 33, 1530 (1965). ( 8 ) P. Audoos, %bad.,86, 383 (1958). (9) D. Smithies and E. J. Hart, J . Am. Chsm. Soc., 89, 4775 (1960). 110) L. Stref and 1'. Ausloos, J . Phya. Chewl., 66, In press (1961).
0.10 ,
..
-
CIHIO
Type of
irradiation
Ilirect photolysis with Corning 9-54 filter
1.4 1.5 2.4 2.4 ,098 0.017 0.012 0.010 0.0024 ,080 ,017 ,012 ,009 ,0021 .69 (3PI)Hg semi.97 .20 ,074 .12 ,076 tized .15 Radiolysis ,075
Toepler pump. In most experiments the non-condensable products and CO, were removed a t - 130" and subsequently analyzed on the mass spectrometer. In those done in the presence of ethylene, the C,, CO, and Cc fractions were removed separately at -175, -155 and -llOo, respectively. Materials.-HCOOH (98%) was obtained from Eastman Kodak and was distilled in order to remove most of the water present. The fractions used in this investigation contained 0.5 to 1% water. DCOOH was obtained from Merck and Company. One sample was used without any further purification. The major impurity was H20 (about 5%). Distillation of a second sample reduced the water content to 1%. The mass spectrum cracking pattern was very similar to the one reported by Ropp and Melton for DCOOH.ll The 0 2 l 6 used was an assayed sample from Airco. The sample of O P was obtained from Isomet Corporation and had an isotopic purity of 980j0.
Results A. HCO0H.-The products measured in experiments performed at 30" are presented in Table I. In the direct photolysis, oxygen and ethylene reduce the Hz yields considerably, while the CO yields are only slightly affected. I n the Hg(3P1) sensitized decomposition both the Hz and the CO yields are strongly reduced by 02. The CO and COz produced in the HCOOH-O,ls mixtures did not contain 0 ' 8 . B. DCO0H.-The results given in Tables I1 and I11 indicate that in the Jirect photolysis a t low temperatures HD, Dz, CO and CO, are the major volatile products. At 25' a change in conversion from 0.2 to 0.85% had no effect on the rate of formation of these products. Most of the experiments listed were performed with the sample containing 5% HzO. Those performed with the sample containing less than 1% HzO gave very similar results, except for a slightly higher Dz yield. Colorimetric tests for formaldehyde in the direct (11) G. A. Ropp and C. E. Melton, J . Am. Chem. SOC.,80, 3509 (1958).
1ox 5
V-~POR-PHASE PHOTOLYSIS O F I'oriu~cA c ~ n
.June. 1961
TABLE I11 EFFECT OF TEMPERATTJRE T,
OC.
Moleicc. x 105
(DCOOH)
--
Rate, cc./min. X 10HD Dz co
a 8
156 203 257 156 203 257 306 101 160 203 257
0.108 ,108 ,108 .35 I35 .35 .35 1.29 1.23 1.24 1.14
Corning 0.011 .03 .09 ,038 ,096 .41 1.5 0.017 .19 .72 1.9
42 T6 110 153 204 254 2i6 306
0.35 .35 .35
0.053 ,066 ,072
.35
,099
co2
filter 9-54 0.086 0.052 ,099 ,048 .36 .14 .28 .12 .29 .19 .47 1.1 1.6 5.2 0.99 0.085 1.0 .47 1.5 .42 4.3 1.7
0.19 .23 .54 .49 .60 1.4 3.4 1.5 1.6 1.8 4.1
0.17 .25 .64 .48 .76 2.1 7.6 1.4 1.8 2.5 7.3
Full arc
.35 .35 .35 .35
.27 .94 1.6 2.8
1.1 1.2 1.1 1.2 1.4 2.7 4.7 16
0.98 1.1 1.1
.96 1.3 2.1 2.7 6.1
3.2 3.3 2.8 3.0 3 3 4.8 5.5 12
2.0 2.3 2.3 2.5 2.9 5.4 8.5 22 s.0
L 17.
photolysis experiments were negative. At 306 O the pyrolysis of formic acid accounted for a t most 5%, and at all other temperatures less than 20/0, of the over-all decomposition. (a) Effect of Scavengers.-The products obtained in experiments with scavengers have been given in Table I1 and the results of comparison experiments without scavengers, obtained under otherwise identical conditions, are included. The results given in Table I1 may be summarized as follows: 1. Addition of a large concentration of COZ in the direct photolysis a t 2', has no effect on the HD, Dz, or CO yields. 2 . Oxygen strongly inhibits the hydrogen yield at 100 and 203O, but only slightly affects the rates of formation of CO and CO,. 3 . I n the direct photolysis, ethylene reduced the yields of H D and Dz without greatly affecting those of CO and COz. CHZDCH2CH2CH,D was the major constituent of the butane fraction. The ethane fraction consisted of CzHO, CZH~D and CzH~D,. 4. I n the Hg(3P1)-sensitized decomposition only the DZyield is appreciably reduced by CzH4. The large quenching cross section of ethylene may be responsible for the increase in the rate of formation of H,. 5. The rates of formation of Hz, HD, DZ, CO and COS were determined in the yapor-phase radiolysis. 1 2 inhibits strongly the yields of H2, HD and Dz, but that of CO is unaffected. (b) Effect of Temperature.-The effect of temperature on the direct photolysis of DCOOH a t constant incident intensity has been investigated at different concentrations (Table 111). The results of experiments performed with the full arc a t one concentration have been included in the same table and are also presented in Fig. 1. The data of Table I11 may be summarized as follows: 1. At temperatures below 200' the yield of Dz is considerably greater in the experiments a t full arc
I
I
I
I
I
1.9
2.1
2.3
2.5
2.7
I/T
x
29
lo3
Fig, 1.
than in those performed with a Corning filter 9-54. 2 . I n the experiments a t full arc there is only a very small variation of the rates of formation of HD, Dz CO and COZin the temperature range 42 to 204'. 3. Above 200' the rates of formation of all products increase sharply with temperature. C. The Photolysis of CDsCOCD3in Presence of HCO0H.-A few experiments were performed in which CDXCOCD, has been photolyzed in the presence of HCOOH a t 3130 fi. (Corning 0-53) and at constant incident intensity. The data given in Table IV indicate that above 201' there is a pronounced increase in the yields of Hz, HD, CO and COz. ,4t the low temperatures the rate of production of HZ is consideredly smaller than the rate of formation of COz. Discussion A. HCO0H.-The results given in Table I show that the ratios CO/CO, obtained in the direct photolysis and in the Hg(3P1) photosensitized decomposition are comparable to those reported previously. However, the fact that (a) the Hz yield is in all cases smaller than the COz yield, and (b) the production of H z is reduced markedly in the presence of efficient H-atom scavengers, such as CzH4 or 0 2 , clearly indicates that processes other than I and I1 have to be invoked to explain the product distribution. B. DCOOH. (a) The Primary Process.-These primary processes have to be considered DCOOH
+ hv ---+
+ HD + HDO +DCO + OH -+D + COOH ----, DCOO + H +CO
COz
(111) (IV)
or)
(VI) (VII)
YOI. 6.i
RALPHGORDEN, JR.,AND P. AUSLOOS
10.76
TABLE IV PHOTOLYSIS OF CD&OCD3 IN PRESENCE OF HCOOH T. OC.
->Role/cc. [CDiCOCnzl
132 201
0 .'70 .67
225 251
X 101-
I
[HCOOH]
HZ
HD
. 61
.61
0.02 .I3 .61 2.4
0.01 .02
,154
0.70 .67 .64
Da
0.002
.066
.18
1. Intramolecular Rearrangements.-In the direct photolysis the low yields of H D as compared to those of CO and COzin the presence of scavengers (ethylene and oxygen) indicate that process 111 cannot account for more than 5% of the over-all yield. In the Hg(?NP1)sensitized experiments it is somewhat more difficult to estimate the relative importance of 111,in view of the large quenching cross section of ethylene. It can be seen, however, that even in the absence of ethylene the yield of HD is considerably less than that of COZ. Moreover, the yield of Hz is greatly reduced in the sensitized photolyeis of HCOOH in the presence of oxygen, although that of GOz is not diminished. The latter argument is not as conclusive in view of a possible deactivation or reaction of a triplet excited state with oxygen. The importance of process IV is difficult to estimate. The independence of the yield of CO over a wide temperature range and in the presence of oxygen, can either be explained by the occurrence of process IV12 or by the decomposition of an excited formyl radical formed in process V. The large yields of D2 and CO in the full arc experiments as compared to the filtered light runs may be ascribed to the decomposition of excited DCO radicals formed in process V. The reduction of the CO yield in the Hg(SP1) sensitized photolysis of HCOOH in the presence of O2 may be due to a quenching of process IV or V by oxygen. It can be deduced from the radiolysis experiments performed in the presence and in the absence of 12, that a process such as I11 is unimportant in the radiolytic decomposition of formic acid. The independence of the CO yield in the radiolysis can be explained in a way similar to that for the case of the direct photolysis. 2. Dissociative Processes.-In order to establish the relative importance of the dissociation processes V, VI and VII, a complete understanding of the secondary reaction mechanism is required. However, a few observations can be made. The very small yield of Hz in the experiments at low temperature can be considered as an indication of the unimportamce of process VII. This conclusion is based, however, on the assumption that H atoms will abstract to some extent from the hydroxyl group. A rough estimate of the number of atoms scavenged by ethylene can be made by taking into consideration the yields of ethane and butane in the direct photolysie of either HCOOH or DCOOH (12) Step I V ia analogous to the intramolecular rearrangement of formates into CO and the corresponding alcohol. The latter process has been shown to oeour in the liquid-phase direct photolysis' and in the Hgl'Pi) sensitized decomposition in the vapor phase.'
.004 ,006 .007
Rate, cc./min. X 10* CDnH CD4
0.80 1.5 2.2 2.4
0.09 .2 .31 .50
7
Ethane
co
co3
1.89 1.13 1.01 0.95
2.6 2.5 3.1 3.8
0.55 1.1 1.9 4.7
in the presence of ethylene. A maximum value for the hydrogen atom yield can be obtained by assuming that two hydrogen atoms are consumed for each butane or ethane molecule formed. It can thus be deduced from the results given in Tables I and I1 that the expression 2(ethane butane) is closely comparable to the C02 yield corrected for the intramolecular contribution from process III.13 From this it may tentatively be concluded that only one H or D atom is produced for each COz molecule. These results can best be explained by taking into consideration process V followed by the secondary reactions
+
OH
+ DCOOH +HDO + COOH ----f
H20
+ DCOO
(1) (2)
and the further decomposition of DCOO and COOH. From the butane yields obtained in the direct photolysis of DCOOH-C2H4 mixtures it can be deduced that the steady state concentrations ratio (D/H) is equal to approximately 2.5.14 This would imply that reaction 2 is faster than reaction 1, or that DCOO is less stable than COOH. Secondary Reactions.-The sharp increase of all products above 202" indicates that a chain reaction takes place under these condition^.'^ The experiments in which acetone has been photolyzed in the presence of HCOOH lead to the same conclusion. The increase of the rate of formation of CO in the photolysis of DCOOH can be explained by DCOOH H --+ HzO DCO (31 DCOOH
+ + + D -+ DHO + DCO
(4)
followed by a decomposition reaction DCO+D
+ CO
(5)
The increase of the yield of CO in the photolysis of CD3COCDrHCOOH mixtures as well as the fact that Hz < COZin the photolysis of HCOOH is consistent with this reaction mechanism. Reactions 3,4and 5 are comparable to reaction 6 which has been proposedg to explain the results obtained in the liquid-phase photolysis of HCOOH COOH
+ HCOOH +C o t + HCO + HzO
(6)
or for reaction sequence such as (13) The H D yield in the scavenger experiments is here considered as a measure of process 111. (14) This value is based on the assumption that the only fate of the hydrogen atom is addition t o ethylene. The fact that an increase in the ethylene concentration (Table 11) does not greatly affect the butane or ethane yields indicates that all thermal H or D atoms hare been scavenged. (15) It has been shown in earlier studies' that at room temperature the quantum yield for decompositlon in the direct photolysis is close to unity. Consequently. the increase in the CO and COI yields cannot be explained by a decomposition of DCO, COOH or DCOO formed in a primary process.
THERMODYNAMIC PROPERTIES OF METHYL KETENESERIES
June, 1961 II
+ +
+ IiCOOH --+ HC(OHOH
COOH 0 2 --+ COz COOH CzH4 +COz
+ HCOOH --+ HCO + HzO + HCOOH
HC
