121s
EMERSON L. ~VITTBECKEK, H. K. HALL,J K . ,
Since we had concluded that hydrogen crowding is the major factor causing cyclic monomers to polymerize,' i t was of interest to examine a few of these compounds in the 1450 cm.-l region to see whether a correlation existed between these frequencies and polymerizability. Unfortunately, as the results in Table IV show, a multiplicity of absorption frequencies was observed and no basis for choosing particular ones was apparent. Therefore, this type of measurement cannot a t present be used as a criterion of polymerizability.
AND
Tor) IV. C.\mwxL
1'01. S"
Experimental The absorption spectra were obtained from 0 1 or 0.01 % solutions in carbon tetrachloride with a Perkin-Elmer single-beam double-pass spectrometer equipped with a calcium fluoride prism. Cyclobutanone was obtained from the Aldrich Chemical Co. and was used as received. The other munocyclic ketones were redistilled commercial samples. The bi- and tricyclic ketones were the samples prepared in the preceding article .lo .___
~~
(IO) H. K. Hall, Jr , THISJ O U R N A L , 82, 1209 (19GO).
WILMINGTON 98, DEL.
CONTRIBUTION FROM THE P I O N E E R I U C RESEARCHD I V I S I O N , TEXTILE FIBERS DEPARTMEST, E. I. DU P O N T DE CO.,
P\I'EMOURS
&
INC.]
Synthesis and Polymerization of Bridged Bicyclic Ethers' BY
EXERSON
L.
\%rITT13CCKER,
H. K. HALL,JR.,
AXD ' r 0 D
ItT. CAMPBELL
RECEIVEDJULY 20. 1959 Polycyclohexylene ether has been prepared by a n ionic polymerization of 7-oxabicyclo [2: 2 : 11heptane using catalysts suitable for the polymerization of tetrahydrofuran. This high melting (450') polyether was obtained as a white powder which was insoluble in most solvents except phenols. The effect of catalyst concentration, temperature, polymerization time and copolymerization with tetrahydrofuran was investigated. endo- and cxo-2-methyl-7-oxabicyclo [2 :2 : 11heptane polymerized to polyethers melting at 247 and ZOO", respectively. Several other oxides of various ring systems were examined for polymerizability.
In the present paper we report an extension to bicyclic ethers of earlier s t u d i e ~of ~ ,the ~ polymerization of atom-bridged bicyclic monomers. Synthesis of Bridged Ethers.-Dehydration of suitable diols by alumina was used in almost all cases. 7-Oxabicyclo [ 2 :2 : 11heptane (I) was readily prepared by literature methods from hydroquinone. 2-Methyl-7-oxabicyclo [ 2 :2 : 11heptane was prepared similarly from 2-methylhydroquinone and the endo- I11 and em-isomers 11were separated by fractional distillation. Failures to prepare B-oxabicyclo [3: 1:11heptane (IV) similarly have been recorded on three occasion^^-^ and we were also unsuccessful. The adducts of butadiene and cyclopentadiene with quinone were hydrogenated and dehydrated similarly to provide in low yields 1,4-endoxodeca(VI), relin(V) and 1,4-endoxo-5,S-methanodecalin spectively. 2,3-Benzo-T-oxabicyclolieptane(VIT) was prepared by adding furan to benzyne and hydrogenating the a d d ~ c t . ~ Dehydration of 3- and 4-hydroxycyclohexancinethanols gave other products than the desired 6-oxabicyclo [3:2 : 11octane (IX) and 2-oxabicyclo[2 : 2 :21 octane (VIII), respectively. 7-Oxaspiro [5 :31 nonane (X)I was not obtained by dehydration of 1,l-diniethylolcyclohexane over (1) Presented in p a r t a t t h e 120th A.C.S. Meeting, Dallas, Tex.. April, 1956, p. 8R of abstract; see also J. P. Wilkins, U. S. P a t e n t 2,764,539 (1956). (2) H. K. Hall, Jr., THISJOURNAL, 80, 6412 (1958). (3) H. K. Hall, Jr., ibid., 82, 1209 (1960). (4) M. F. Clarke a n d L. N. Owen, J . Chem. Soc., 2103, 2108 (1950). (5) R. B. Clayton and H. B. Henbest, ibid., 1982 (1957). ( 6 ) F. V. Brutcher, el al., Chemistry & I n d u s t r y , 1295 (1957). (7) G . Wittig a n d L. Pohmer, Chem. Ber., 89, 1334 (1956). (8) This compound was recently reported b y S. Searles, E. F. 1,utz and hI. Tamres, Abstracts of Spring, 1959, A.C.S. Meeting, Bostrau, Mass., p. 63-0.
alumina, ketones and alcohols forming instead, but could be obtained in moderate yield by the alkaline hydrolysis of the corresponding ditosylate. S-Oxabicyclo [4 : 3 : 01nonane (XI) was obtained from 1,2-cyclohexanedimethanol. Dehydration of 1,3-~yclohexanedimethanolreadily gave 3-oxabicyclo [3:3 : 11nonane (XII), but similar treatment of the l,4-isorner failed to give 3-oxabicyclo [3 : 2: 21 nonane (XIII). 6-Oxatricyclo [3 : 2 : 1: 1 3 r 8 ] nonane (XIV) was prepared from bicyclo [ 2:2 : 11heptene5-methanol. In a t least two cases oxides have been obtained from chlorohydrins, but not by dehydration of the corresponding diol. Clarke and Owen l5 prepared I X in this way and we obtained ether X similarly. Thus the dehydrohalogenation procedure may be more generally useful though less convenient. Attempts to reduce two bridged lactones directly 0
0
,'
to ethers failed. Lithium aluminum hydride gave exclusively diol, while copper chromite and hydrogen did not react. Polymerization Results 7-Oxabicyclo-[ 2 : 2 : 1]heptane.-Alkyl, acyl, oxonium and hydrogen ions have been used to polymerize tetrahydrofuran.'O These initiators were applied to 7-oxabicyclo [2 : 2 : 11heptane to give a high melting (450') polyether. Phosphorus pentafluoride". was also effective though slightly irreproducible and was much better than boron trifluoride. Much higher concentrations were re(9) H. A. Bruson a n d T.W. Riener, U. S. P a t e n t 2,440,220 (1948). (10) Summarized by K. I i a m a n n , Axgew. Chcm., 63,236 (1951). (11) T. W.Campbell, U. S. P a t e n t 2,831,825 (1958).
hlarch 5, 1960
URIUGELIBICYCLICETHERS TABLE I
1219 TABLE I1
POLYMERIZATION O F 7-OXABICYCL0[2 : 2 1]IIEPTANE AT
I
I1
I11
CH3
V
I\'
kQ IX
VI11
CataMole lyst % FeCla 0.015 FeCla .0225 FeCld .03 FeClr .OR FeCla .12 SbCl3 ,l SbCls .2 BFa 4.0 RF3 0.2 SbCls .1 SbCla .2 SbClj .3 PFa 12.0 PF: .. SnClr 2 0 FeCli 0.04 FeCla .OY FeCla .03 FeClr .06 FeCla .03
Mole, Time, O0 hr. 0.62 70 23 0.93 SOCl? SOCI: 1.24 18 Socle 2.48 7 3.96 5 SOCIz Epichlorohydrin 0.2 336 Epichlorohydrin .2 336 Epichlorohydrin ,l 72 Propylene oxide .2 336 Propylene oxide .2 40 Propyleneoxide .2 40 Succinic anhydride 1 . 0 72 " ,, 102 b , . 92 Acetyl chloride 1 0 120 SnClr 0.1 72 Succinyl chloride 1 . 2 4 1168 Succinyl chloride 0.62 45 Succinyl chloride 1.24 48 Succinylchloride 0.62 132 Cocatalyst
SOCln
Yield,
0'
[?inhl
(TCE-
(lo phenol)
45.4 33.2 36.6 41 6 29.8 15.8 18.2 4.4 1.0 22.0 12.4 29.4 58.8 81.0 5.6 18.0 51.2 38.2 46.8 32.6
a 3 3 % solution in ethylene dichloride at -30'. solution in nitrobenzene a t -30'.
0.56 .55 .63 .51 .51
.6l
.50 .2%
..
.OO
.45 .t(2 .63 1.04
..
0.M .50
.50
.5R .5l
* 33%
TABLE I11 EFFECT OF TEMPERATURE Catalyst was 0.03 mole % of FeCla and 1.24 mole % of SOCll Temp.,
OC.
-
XI1
XI11
h
h
m
21 0 18
Time, hr.
Yield,
%
Intrinsic viscosity
0.5 22 312
i5.8 27.0 25.6
0.23 .53 .82
The rate of polymerization is dependent on catalyst concentration and temperature, but if these are kept constant the effect of time can be determined (see Table IV). The molecular weight as indicated by intrinsic viscosity reached a maximum in a relatively short time with a very low yield. The yield increased with time, while the intrinsic viscosity decreased very slightly. TABLE IV RATEO F POLYMERIZATION O F 7-OXABICYCL0[2: 2: 11HEPTANE Constant temp., 0'; constant catalyst, 0.03 mole % Feel3, 1.24 mole T i m e , hr.
quired for satisfactory rates of polymerization. Catalysts other than Lewis acids were completely ineffective. The following variables were investigated : Catalyst concentration affected the rate and degree of polymerization (see Table 11). As the catalyst concentration was increased, the polymerization time was decreased, but the intrinsic viscosity was little affected. Reproducibility was good in-these experiments. The temperature of the polymerization also affected the rate and degree of polymerization. Side reactions were decreased as the temperature was lowered and, therefore, higher molecular weights were obtained. The polycyclohexylene ether with the highest intrinsic viscosity was polymerized a t -18'. The polymerization rate, however, was so slow t h a t it was impractical to use temperatures much below 0". The effect of temperature on the polymerization of 7-oxabicyclo [2 : 2 : 11heptane is summarized in Table 111.
2 5 8 24 32
96
Yield,
yo SOClz Yo
171
2.8 12.0 18.6 48.2
0 73 .61 .62
74.0 74.0
56 56
.x
The fact that the molecular weight is not a function of conversion means that a chain reaction is occurring. Properties of Polycyclohexylene Ether.-Polycyclohexylene ether melted between 430 and 450°, as observed both visually and by X-ray diffraction. It did not dissolve in o-cresol, anisole, dioxane, tetramethylene cyclic sulfone, tetrachloroethane, benzyl benzoate] o-dichlorobenzene, nitrobenzene, Dowtherm A, formic acid or other common solvents. The polymer is soluble in a mixture of 66 parts tetrachloroethane and 100 parts phenol. It is very slightly soluble in diphenyl ether and in phenol. An attempt to cast n film from tetrachloro-
ethane-phenol gave an incoherent structure with the appearance of frosted glass. The molecular weight of a sample with an inherent viscosity of 0.5swas calculated from the chlorine content (l.02yc) to be 3800, assuming both terminal groups of the polymer chain to be C1. An osmotic pressure determination indicated that the molecular weight of a sample with an inherent viscosity of 0.72 in TCE-phenol was in the 5,00010,000 range. This inherent viscosity is comparable to those shown by polyamides ol similar molecular weight.12 I n this case stiffness of the polymer chain, rather than hydrogen-bonding is the main factor in raising the inherent viscosity. Copolymers.-In order to obtain lower melting, more soluble polyethers, 7-oxabicyclo [2 : 2 : 11heptane was copolymerized with tetrahydrofuran. Copolymers melting below 250' were waxy while the higher melting ones were brittle. For example, starting with a 2-to-1 inole ratio of i-oxabicyclo[2 : 2 : 11heptane and tetrahydrofuran, a 39% yield of a polymer with an inherent viscosity of 0.94 was obtained after polymerizing for 30 days a t -18". The polymer melted over the range 235-275". Other copolymers prepared are listed in Table V. The melting points of copolymers prepared are listed in Table T.'. The melting points of copolymers were obtained by observing the powder in a capillary tube while heating it in a stirred liquid bath. C O P O L Y M E R S OF
TABLE 1. ;-OXABICYCLO [2 : 2 : 11H E P T A S E
The methylcyclohexylene ethers were soluble in phenol-tetrachloroethane (100 : 66 by wt.), 771cresol, trifluoroacetic acid, chloroform and chloroform-formic acid (60:40 by vol.). They were insoluble in 90 and 09C/b formic acids. Ethers V and IrI gave colored solutions with PFa in ethylene dichloride, but no viscosity change or other indication of polymerization was noted. Ether VII, 0.2 g., in 1.0 ml. of toluene at -SO" polymerized when a trace of PFj was introduced. The solid white polymer, which was probably of low molecular weight, melted a t 50" to a viscous liquid. Other Ethers.--2-Oxaspiro [3 :3 Jnonane polymerized readily when treated with PF6; XI gave a thick sirupy polymer when treated with fluorosulfonic acid, a catalyst concentration of 3-127. giving a 4050y0 yield of polymer of molecular weight 37009800 in 72 hours; XI1 and XI\' failed to undergo polymerization. Discussion Mechanism.-.% direct displacement mechanisn~ as Initiation:
Propagation: 0
A S D TETRA-
Inherent
Epoxy-
cyclohexane, mcile
'l'etrahydrofuran, mole
Temp OC.
,
T i m e . Yield, hr c?c
VIS-
cosity
CIS? ll.p., 'C.
4 4 . 0 0 03 149-170 42.4 .5.5 220-2i5 0 77.2 47 286-305 003 0 50 8 60 320 025 -18 180 38 8 . 9 4 233-275' .09 ,025 0 113 6 5 . 8 . 7 5 250-294 Catalyst used was 0.03 mole yo of FeCls and 1.24 mole l~'; of SOClz. h Polymerization was run in fi-dichlorobenzcnc. X-ray m.p. 260". 0.0: 05 0.5 05 05
0.05 025 0125
19 0
c1so--,
0
HYDROFURAN~
375 96 00 06
0
Fe( 1;
'12) (a) G H . l'aylor, T m I O U R N A L , 69, 633 (1947); (b) J. R. Schaefgen and P. J. Flory, ~ b t 70, L 2709 (1948).
-Or
FeCI;
seems reasonable for the polymerizations described here. Termination involves collapse of ion pair to chloroether, since the polymer contains chlorine Termination :
(1
Other 7-Oxabicycloheptanes.-endoand exo-2methyl-7-oxabicycloheptanes polymerized well when treated with PF6 in nitrobenzene or ethylene tlich1orid.e solution a t -30". In Table VI are given the best examples of these experiments. Again some difficulty with reproducibility was encountered, as is often the case with cationic polymerization. Addition of protonic compounds such as water, acetic acid, sulfuric acid or hydrofluoric acid, suppressed polymerization completely. The methylcyclohexylene oxides were white polymers which, on heating. gave clear melts from which short fibers could be drawn. Films were meltpressed from material having Tiilh of 1.3 or more. They were cloudy and only moderately tough. A copolymer of v i n h 2.6 (TCE-phenol), cited in Table VI, was melt-spun to a clear bright monofilament. However, it had very low tenacity and couid not be drawn.
0
_ ,
0- FeCld
-halogous mechanisms have been proposcd for polymerization of other cyclic ethers.I3 Polycyclohexylene ether is undoubtedly an alltrans polymer, as evidenced by its extremely high melting point. This fact strongly supports the above mechanism. A carbonium ion process would give a mixed cis-tuans polymer. Polymerizability and Conformation.- Conformational analysis has been used previously to account for the poIymerizabilit>- of cyclic carbonyl compouiids.23YIt also is useful in the case of cyclic ethers. Cyclopentane is known to be strained to the extent of G kcal. per mole owing to H-H crowding. Tetrahydrofuran with almost the same bond angles is also strained by 2.9 kcal.,l4which ac(13) J . B. Rose, J . C h e m . Soc., 346 (1956). (14) R. C . Cass, S.E . Fletcher, C . T, hlortimer, 13. D. Springall and T. R. White, ibid., 1406 (1958).
POLYMERIZATIONS O F
Monomer, g.
a
1221
BRIDGEDBICYCLIC ETHERS
March 5, 1960
endo, 26.0 3.0 endo, 15 unsubst., 1 5 exo, 26.0 e m , 15.0 exo, 1 5 . 0 exo. 15.0 Catalyst, gaseous PFj.
TABLE VI 2-METHYL-7-0XABICYCLO[2 :2 1 ]HEPTANES'
Solvent, ml.
Ethylene dichloride, 50 Nitrobenzene, 9 Ntrobenzene, 45 Ethylene dichloride, 30 Ethylene dichloride, 15 Ethylene dichloride, 30 Ethylene dichloride, 60
T e m p . , 'C.
25 0 - 30 - 30 - 30 - 30 - 30
Time
Yield, %
11 p , OC
[tllinh (TCEphenol)
4 22 120 90 100 100 100
98.4 90.7 45.7 66.5 78.0 37.4 16.9
234 247 261 2'10 200 200 2rr0
0.53 0.73 2.64 0 67 0 83 1 06 1.05
Cyclopentanediol, b.p. 115-133" (11 mm.) (lit.I9 b.p. 115' ( 6 mm.)), was prepared by hydrogenation of cyclo pentenediol (Columbia-Southern Co.) over ruthenium ciioxide. The infrared spectrum showed a trace of carbonyl, owing to ready autoxidation, but n o olefinic linkage. counts for its polymerizability. Anal. Calcd. for C;HloO?: C, 55.8; H, 9.9. Found: 7-Oxabicycloheptane exists in the boat form C, 58.3; H, 9.7. 1,5-hTaphthalenedio1 was recrystallized twice from nitromethane and hydrogenated over ruthenium dioxide in ethanol. The product crystallized well: crop 1, m.p. 163205' (mostly a t 186'); crop 2, m.p. 136-163" (mostly at 145'). Anal. Calcd. for CloHls02: C, 70.5; H , 10.7. Found: Here angle strain as well as H-H crowding is re- C, 70.7, 70.8; H, 10.5, 10.8. lieved on polymerization. A single methyl or benzo Monocpclopentadienequinone, m.p. 73-T4" (lit.20 m.p. substituent does not suppress polymerization but a 77-78"), and monobutadienequinone, m.p. 56-5i0 (litZ1 m.p. 5S0), were prepared by literature methods. Hydrofused cyclohexane or bicycloheptane ring does so. genation over ruthenium dioxide in ethanol proceeded Ethers VIII, I X and XI11 should be polymeriz- smoothly. Filtration and evaporation left the diols as able. Ether XII, however, with two fused chair colorless viscous glasses. forms, should be stable as observed. Bicyclohept-5-ene-2-rnethanol, b.p. 95-97' (14 mm.), Acknowledgments.-We are indebted t o the n% 1.4981 (lit.22b.p. 92-95' (13 mm.)), was obtained by addition of allyl alcohol to cyclopentadiene. following and their associates: Dr. J. P. Wilkins theOxides.-Dehydrations of the diols were accomplished for. the original suggestion, Mr. I. D. Plank for by heating with an equal weight of alumina (Alcoa Actithe microanalyses, Dr. H. E. Cupery for the hy- vated Alumina, Grade F-20) to 250-320' until distillation drogenations, Mr. H. Thielke for vapor-phase ceased. For most small-scale reactions this required 1-2 chromatograms, Dr. R. Zbinden for infrared hours, while the large-scale preparations were allowed to distil overnight. Pressure was maintained a t 1.50 mm. in spectra, Dr. P. W. Morgan for helpful encourage- the distillation of the high boiling fused-ring oxides. The ment, and Mr. C. bIardecz and Mr. G. Elechko for distillate was dried and subjected to a preliminary vacuum distillation through a Claisen head, and was then fractiontechnical assistance. ated in a spinning band column. 7-0xabicyclo[2:2:llheptane, obtained in 347, yield, Experimental had b.p. 119.0-119.5° (lit.*3 b.p. 117-118'), n z j ~1.4409 Diols.-Cyclohexane-l,4-diol and 2-methylcyclohexane- (iit.24 nzoD 1.4477). Similarly, a mixture of exo- and endo-2-methyl-7-osabicy1,4-diol were prepared by hydrogenation of the corresponding hydroquinones in ethanol solution over ruthenium diox- cl0[2:2: 11heptanes was obtained by heating 2-methylcyclohexane-l,4-diol with alumina. These were separated ide. in a 125-cm. Podbielniak column by Dr. H . Eatough and 3-Hydroxycyclohexanemethanol was prepared by lithium aluminum hydride reduction of 6-oxabicyclo[3:2 : I ]octan- the purity of each was established as
