790
WILLET
F.
~$‘HIT%lOREAND
one to determine the average number of oxygen bridges per chromic ion present in aged or heated solutions of chromic nitrate or chloride. 2. The postulated composition of a sixty-hour refluxed solution of chromic chloride and of a 1/3 basic, sixty-hour refluxed solution given by E. Stiasny“ and co-workers has been shown to require revision. 3. Data are given for the number of Mo04Hgroups that will coordinate to chromium in variously treated chromic salt solutions. These data show that the average number of oxygen
[CONTRIBUTION FROM THE
J. FREDGERECHT
Vol. 7 2
bridges joining chromic ions increases with time on refluxing solutions of chromic nitrate and chromic chloride until a maximum is reached a t sixty ‘hours. Further refluxing does not cause any increase in the number of bridges. The addition of sodium hydroxide before refluxing causes a large increase in the number of oxygen bridges formed as does also the addition of ethyl alcohol. SALTLAKECITY,UTAH RECEIWD‘*AUGUST18, 1949 (18) Original manuscript received August 20. 1947.
DEPARTMENT O F CHEMISTRY,
POLYTECHNIC INSTITUTE OF BROOKLYN]
The Thermal Polymerization of Monoolefinic Compounds‘ BY WILLETF. WHITMORE AND J. FRED GERECHT Numerous studies of the heat bodying of drying oils have failed t o yield conclusive evidence for any of the reaction mechanisms suggested for this process.2 The thermal polymerization of methyl undecylenate had been studied by one of us3 in the hope that a more complete understanding of the thermal reactions of this simple compound might yield information concerning the bodying of the more complex oils. In the present work an allylic free radical theory t o explain the thermal transformations of simple unsaturated compounds is proposed, and some experimental support for this mechanism has been obtained from a study of two terminally unsaturated compounds, methyl undecylenate and 1-octene. Free radicals, which might arise through chain scission or peroxidation, are capable of initiating a chain reaction by attacking a monoalkyl substituted ethylene to form a resonating allylic radical (I),4 which in either resonance form may add to the terminal position6 of another olefin molecule (11). Either of the resulting unsaturated dimeric radicals may then add intramolecularly t o its own double bond t o yield saturated cyclic radicals (111). In an analogous manner the saturated or the unsaturated dimeric radicals may continue this process of radical addition t o give a variety of polymeric radicals (IV). It is also quite possible for the unsaturated trimeric radicals formed in step IV to cyclize by a n addition similar to step I11 t o form polymeric alicyclic radicals.ba Any of these radicals (monomeric, dimeric or polymeric) instead of adding to another olefin (1) From a thesis presented by J. Fred Gerecht t o the Graduate
School of the Polytechnic Institute of Brooklyn, June, 1948, in partial fulfillment of the requirements for the degree of Doctor of Philosophy. (2) Sunderland, J . Oil and Colour Chemists’ Assoc., 96, 137 (1946). (3) Ross, Gebhart and Gerecht, THISJOURNAL, 67, 1276 (1946). (4) Smith and Taylor, 1.Chcm. Phys., 8, 543 (1940). (6) Farmer, 1.SOC.Chem. Ind., 66, 86 (1947). (Sa) We are indebted to one of the refaear for hi8 ruggutien of thb remati*. of the trimuia radical..
molecule can terminate through chain transfer by withdrawing an a-methylenic hydrogen to yield a stable compound and a new resonating allylic radical as in I. I n an attempt to verify such a mechanism, methyl undecylenate was polymerized a t 325O, and separated by distillation into monomeric, dimeric and higher polymeric fractions8 Each fraction was then subjected to performic acid hydroxylations followed by periodate oxidation.’ The oxidation mixtures were examined for the specific degradation products which should result if the unsaturated compounds predicted were present. According t o the mechanism indicated) the monomeric fraction should contain a certain amount of methyl 9-undecenoate, A, resulting from the reaction of one of the resonance forms of the allylic radical with an a-methylenic hydrogen) and hence, both acetaldehyde and formaldehyde should be present in the oxidation mixture. Such was found to be the case, and from the amount of acetaldehyde found8 i t was estimated that 12% of the recovered monomer had undergone a double bond shift. Furthermore, the dimeric fraction should contain dimethyl 8 - eicosene 1,20 - dicarboxylate, dimer B (R = H) ; dimethyl S-ethylene-octadecane-l,lS-dicarboxylate, dimer C (R = H) and methyl esters of saturated acids of 22 carbon atoms containing the cyclopentane ring, D, resulting from the termination reaction of the dimeric radicals with a n a-methylenic hydrogen. Previously the presence of dimer B (R = H) in this fraction has been established, and a saturated dimer believed to be cyclic also has been isolated.8 Although methods for the recognition of either of the
-
(6) (a) Greenopan, Ind. Eng. Chcm.. 89, 847 (1947); (b) Swern, BUlen, Findley and Scanlan, THISJOURNAL, 67, 1786 (1946). (7) Jackson, “Organic Reactions,” Vol. 2, John Wiley and Sons, Inc.. New York, N. Y.,1944, p. 841. (I) Shupe, J , Ailoc. 0 5 ~A. n . Chrmirca, 96, 140 (1948).
79 1
THERMAL POLYMERIZATION OF MONO~LEFINIC COMPOUNDS
Feb., 1950
R-CHe-CH=CHz
I R-C€I-CH=CH2
4
I1
R. > R--CH=CH-CIIY
+ RH
4
III
K-CH-CH=CH~ CH2-CH-CHzR
R-CH-CH I / R-CH~-CH CH~
R -CH-CH
+
I
I
1
CHz CHL CH
CHz-R
Tv
.
CHz-CHCH?R
CHz-CHCHzR R-CH-CH
I
R--CH-CH
1
I
CHt &HZ
CHz-CH-CHzR CHz-CH-CHzR I
1
R-CH?-CH
I
I
CH,
\/
CHz
\Ci
I1
R-CH=CH-CHz R-CH2-CH-CHz
I
I
CHz
I
IV1
111
*
\/
\/
R-CH-CH=CHz
4
R-CH2-CH=CH2
R-CHZ-CH=CHZ
IV
+ R-CH=CH-CHz I
RCH~CHCH~
I
RCHzCH-CH,
I
CHzR
expected cyclopentane dimers, I), are lacking, McKinley, Stevens and Baldwing have shown that one of the dimers resulting from the thermal polymerization of isobutylene is 1,1,3-trimethylcyclopentane. The formation of this compound may be explained more satisfactorily by a n allylic radical mechanism than by any of the several sequences proposed by these authors. The presence of dimer C (R = H) would be indicated by the formation of formaldehyde in the oxidative cleavage. This aldehyde was demonstrated to be present, and from the quantity foundlo it was inferred that C comprised 12% of the dimer. CH~-C€I=CH-(CHz),-COOCHa CH-(CHz)r-COOCHd
(A)
CH-CH~-CHZ-CH-(CHZ)~-COOCH~
(B)
I/
I
R CHz=CH-CH-(CHz)r--COOCH~
I
CHz-CH-(CH*)a-COOCH
