Radical Cyclopolymerization of Divinyl Ethers. the Polymerization Kinetics and the Polymer Structure.*

Polymer Journal, Vol. 13, No. 7, pp 657-669 (1981)

Radical Cyclopolymerization of Divinyl Ethers. The Polymerization Kinetics and the Polymer Structure.*

Mitsuo TSUKlNO and Toyoki KUNITAKE**

Department of Chemical Engineering, Kitakyushu Technical College, Kokura-minami, Kitakyushu 803 and Department of Organic Synthesis, Faculty of" Engineering, Kyushu University, Fukuoka 812, Japan.

(Received November 29, 1980)

ABSTRACT: The radical polymerizations of divinyl ether (DVE), cis-propenyl vinyl ether

(PVE) and 2-methylpropenyl vinyl ether (CH3 -PVE) were carried out with AIBN initiator. The polymers were composed of five-membered monocyclic units with pendent unsaturated groups and [3,3,0]bicyclic units. The bicyclization was favored at low monomer concentrations and with methyl-substituted monomers. The microstructures of the polymers were determined by 13C-NMR spectroscopy, through extensive use of the model compounds. A common cyclopolymerization process has emerged from the data obtained. The monomers react exclusively at the unsubstituted vinyl group, and the trans ring closure produces five-membered monocyclic radicals which then propagate intermolecularly or cyclize to give trans-fused bicyclic units. KEY WORDS 13C-NMR Spectroscopy I Cyclopolymerization I Poly- (divinyl ether) 1 Poly(cis-propenyl vinyl ether) 1 Poly(2-methylpropenyl vinyl ether) I

Divinyl ether (DYE) has been known to give CH3-PYE. Some kinetic studies were also carried soluble polymers with highly cyclized structures by out. radical polymerization, and the structure of the H H cyclized unit has been inferred from the model CHz=Cft CH=CH2 CHz=Cft )>C( 0 0 CH experiments and from the kinetic data. 1 - 3 We 3 DVE PVE recently examined the polymer structure by means H, CH3 of 13C-NMR spectroscopy and concluded that the CH-CH C=C/ T 'cf 'cH cyclopolymerization process involved a five 3 CHrPVE membered ring intermediate which would either propagate intermolecularly or cyclize to a bicyclic unit,4 •5 as shown in Scheme l. Interestingly, the cyclization process was highly stereoselective. EXPERIMENTAL In the present study, we carried out a structural Materials study on the cyclopolymers of cis-propenyl vinyl The purification of DYE has been described. 5 ether (PYE) and 2-methylpropenyl vinyl ether PYE was prepared by the isomerization of allyl (CH3-PYE), and compared the effect of the poly vinyl ether obtained by the exchange reaction of n• merization condition on the polymer structure for butyl vinyl ether and allyl alcohol in the presence of each of three related monomers: DYE, PYE, and Hg (OAc)z according to the procedure of Watanabe et 6 : 6 * Contribution No. 611 from Department of Organic a/. bp 66-67°C, lit bp 66-67°C. The isomer Synthesis. ization to PYE (cis-isomer) was performed with ** Correspondence should be sent to this author at the reference to the preparation of cis,cis-dipropenyl Fukuoka address. ether7 in the presence of potassium tert-butoxide in

657 M. TSUKINO and T. KUNITAKE

stereoselective O-CH=CH CH cyclization I • 2 II 2 M-CH,-CI-:!._ _...CH /CH . Z CH2 '0

lntermolecular propagation

stereoselective cyclization Scheme 1.

dimethylsulfoxide at room temperature for 10 days. Gelation occurred at conversions of ca. 3% in the PVE was separated by distillation under reduced bulk polymerization of DYE, but no gelation occur pressure, washed with alkali and water, and re red for bulk PVE and CH3-PVE up to conversions distilled: bp 61.5-62.SOC (lit6 61-62°C), of 30--40%. 0. 7899. The isomerization was virtually quanti Poly(PVE) tative, but the trans isomer (5%) was also formed. Anal. Calcd for C5 H80: C, 71.33%; H, 9.51 %. Methallyl vinyl ether was similarly prepared by Found: C, 70.93%; H, 9.52%. the exchange reaction of methallyl alcohol and Poly(CH3-PVE) methyl vinyl ether at 30°C in the presence of Anal. Calcd for C6 H100: C, 73.47%; H, 10.20%. molecular sieve 4A and a catalytic amount of Hg Found: C, 73.06%; H, 10.19%. (OAc)z, according to the procedure of Yuki et al. 8 : bp 89.0-89.SOC (lit. 87-88°C,6 90°C8). The iso Hydrolysis merization was carried out quantitatively in a way The pendent propenyloxy groups of poly(PVE) similar to that mentioned above for 90 h at 90°C, to and poly(CH3-PVE) were converted to hydroxy give pure CH3-PVE, bp 85.6-87.SOC, 0.7940. groups by hydrolysis of the polymers in mixtures of The structure and purity of this monomer were methanol and hydrochloric acid, as already done confirmed by gas chromatography, 1 H-NMR spec for poly(DVE). 5 The complete removal of the pen troscopy and elemental analysis. dent group was confirmed by NMR spectroscopy

Anal. Calcd for C6 H100: C, 73.47%; H, 10.20%. and the NMR data were consistent with the elemen Found: C, 73.01 %; H, 10.42%. tal analysis. Azobisisobutyronitrile (AIBN) was recrystallized Poly(PVE) (hydrolyzed) from ethanol. Solvents were purified by the usual Anal. Calcd for (C5 H8 0)0 .74(C2 H4 0)0 .26: C, procedure. 68. 75%; H, 9.46%. Found: C, 67.91 %; H, 9.52%. Poly(CH3-PVE) (hydrolyzed) Polymerization Anal. Calcd for (C6 H 100)0 .7iC2 H4 0)0 .26: C, Given amounts of monomer, AIBN, and ben 70.88%; H, 10.05%. Found: C, 69.49%; H, 9.95%. zene, when necessary, were placed in ampoules and subjected to the freeze-pump-thaw cycle several M icellaneous times. The ampoules were then sealed in vacuo and The amount of the unsaturated pendent group in immersed in a constant temperature bath. The polymer was determined by 1 H-NMR spectroscopy polymer was recovered by precipitation in meth (Hitachi R-24B, 60 MHz) using the peak area of the anol, purified by reprecipitation from benzene vinyl methine proton or the propenyl methine pro and methanol, and freeze-dried. The polymers were ton (5.4-6.6ppm). 13C-NMR spectroscopy was white powders, soluble in CHC13 , CC14 , etc. obtained under noise decoupling with a Bruker

658 Polymer J., Vol. 13, No. 7, 1981 Cyclopolymerization of Divinyl Ethers

"6c 0 3 .0 30 :iS"' ::J 0 2 20 -DVE PVE ---•--- CHfPVE

0 2 4 6 8 10 12 1JL--l..__l..__..L__..L__.L.J [M) mol/1 3 Figure L Content of the pendent double bond (PDB) in poly(divinyl ether)s. Polymerization conditions: C6 H6 solvent, 60°C. PDB is 100% provided that each mo 2 nomer unit contains one pendent unsaturation.

2 4 6 8 10 WH-90 (22.63 MHz) instrument and a JEOL FX- [M] mol/1 100 (25 MHz) instrument. The molecular weights of Figure 2. Plots of 1/f; against the initial monomer the polymer were determined by gel-permeation concentration: (I) DYE; (2) PVE; (3) CH3-PVE. chromatography: Toyo Soda Co., Model 802 UR. Three columns of different molecular-weight ranges of variation in PDB. This means that the unit were connected in series. Tetrahydrofuran was used structures remain the same and that the PDB as the eluent, and molecular weight was calculated variation is caused by change in the extent of bi using a calibration curve of monodisperse poly cyclization. The extent of bicyclization is greater styrenes. for poly(PVE) and poly(CH_1-PVE) than for poly (DVE). The rates of the intermolecular propagation and RESULTS AND DISCUSSION the intramolecular cyclization from the monocyclic Polymerization Kinetics intermediate (RP and Rc, respectively) are given for The monocyclic and bicyclic units in poly(DVE) DVE by, are formed through a common intermediate.5 The RP = 2kp[M][M ·] (1) same situation arises for poly(PVE) and (2) poly(CH3-PVE). As will be discussed below, the Rc=kc[M·] pendent groups were invariably substituted vinyl were kP and kc are the respective rate constants, and groups (propenyl and methylpropenyl groups), and [M] and [M ·] are the concentrations of the mono single five-membered intermediates were formed mer and the monocyclic radical intermediate, stereoselectively which produced the corresponding respectively. monocyclic and bicyclic units. The propagation Then, occurred preferentially from the vinyl side. Figure 1 shows the relationship between pendent (3) unsaturation in polymer and monomer concen tration. In all cases, the content of the pendent The extent of bicyclization is given by, double bond (PDB) (i.e., the fraction of the mono J- Rc (4) cyclic unit) increased with increasing monomer c- Rp+Rc concentrations, reaching constant values at initial Therefore, monomer concentrations from 3-5 moll- 1 . The 1 kp conversions were kept below 10%. The peak pat -=1+2-[M] (5) tern of 13C-NMR spectra did not change, in spite fc kc

Polymer J., Vol. 13, No. 7, 1981 659 M. TSUKINO and T. KUNITAKE

6 4 (1) (1)

8 (2) 3 6 L (2) ("') I 0 Q_ ::4 0: 0 c 0 ::E 2 0 0 (3) (3) 'I' 4

>< 00 c 2 2 4 6 8 10 12 ::E [M] mol/1 Figure 3. Rate of polymerization of divinyl ethers 2 4 6 8 10 plotted against the initial monomer concentration: [M] mol/1 Polymerization conditions: C6 H6 solvent, 60' C; AIBN Figure 4. Molecular weight vs. initial monomer con 2.5x for DYE (1), 2.5x for centration: (1), DYE; (2), PVE; (3), CH3-PVE.

PVE (2), 4.0x for CH3-PVE (3). Polymerization conditions are the same as those of Figure 3.

In the case of PVE and CH3-PVE monomers, high monomer concentrations may be also attri (6) buted to the solvent effect of the monomer. The since only the vinyl group is involved in the molecular weight of the polymers increased linearly propagation. Therefore, with the monomer concentration as shown in Figure 4. The molecular weight of poly(DVE) was 1 kp -= 1 +-[M] (7) in the range of 5 x I 03 to 4 x 104 under the con kc ditions used, but those of poly(PVE) and

Figure 2 shows the plots of I I against [M]. poly(CH3-PVE) were 2000-8000 and 1000-4000, Linear relationships are not observed against the respectively. Methyl substitution decreases the mo theoretical prediction, and 1/ fc tends to saturate at lecular weight considerably. high monomer concentrations. It is suspected that the solvent effect of the monomers becomes in Structure of Poly( DVE) fluential at higher concentrations. The kc/kP values The structure of Poly(DVE) was determined in were estimated tentatively from the initial slopes: 2 our previous paper5 by comparing the 13C-NMR 1 moll-' for DYE and 3 moll- for PVE and CH3- spectrum of the polymer with the chemical shifts PVE. These values indicate that the bicyclization calculated using model compounds (methyl process is more favorable when the pendent vinyl cyclopentanes and 2,5-dimethylcyclopentane). group is substituted. Recently, there were published 13C-NMR data of Figure 3 shows the relationship between the more appropriate model compounds.9 These new initial rate of polymerization and the monomer data on substituted tetrahydrofurans were com concentration. The polymerization rate shows max bined with previously used data on cyclopentanes10 ima at [M] = 4-6 moll- 1 . The rate decrease at and cyclohexanols11 in order to estimate the chemi-

660 Polymer J., Vol. 13, No. 7, 1981 Cyclopolymerization of Divinyl Ethers

"·'IJY,r;-c 77.0 85.3 c-c-c \ojc-c-c c--;. Q A c +s.'Y(' c-c 87.8 78.4 81.9 79.8 82.8 43.6 40.0 40.7 ?-72.9 H D c c 86.1

c-e--o } (obsdl c-c-rnc-c-c?lt'6 85.5 41.2 4t0 Figure 5. Comparison of the observed chemical shift of the hydrolyzed monocyclic unit of poly(DVE) with c-qc-? those estimated for 2,5-dipropyl-3-hydroxytetrahydro furans. OH d--.z'OH Scheme 2. 80.5 78.4 80.0 c-c-cb-c-c-c- c-c-c\;o-f-\.3y_or c-c-c cal shifts of the model compounds shown in Scheme 41.4 2. In step 1 of this scheme, the chemical shift 80.1 40.9 411 differences between the dimethyl- and trimethyl c-c--c 1'-.c-c-c cyclopentanes are added to the observed chemical 40.9 81.6 80.0 shifts of trans-2-ethyl-3-methyltetrahydrofuran so Sss-3 Sss-4 as to estimate the chemical shift of the ring carbons 78.1 86.4 of the corresponding trisubstituted tetrahydrofuran.

The effect of the substitoent elongation is con 41.1 " 78.1 sidered in steps 2 and 3. Step 4 is concerned with the 82.6 Sss-6 conversion of methyl to hydroxyl groups. Four configurational isomers are present for 2,5- 78.1 84.5 78.1 di-n-propyl-3-hydroxytetrahydrofuran, and the r-c-c c- c-c l;O-t"'"\_e-- c-c 4o\---l-o1 chemical shifts of the ring carbon of these isomers were estimated by procedures similar to that of Scheme 2. The results are summarized and com pared with the observed chemical shifts in Figure 5. The calculated chemical shifts for isomers A and B (trans ring closure units) are in good agreement with the observed values. The agreement with the Figure 6. Comparison of the observed chemical shift experiment is better in the newly estimated set of of the bicyclic unit of poly( DYE) with those estimated chemical shift than the previous estimation, par for the model compounds. ticularly for ring carbon 4. The chemical shifts of the ring carbon of the bicyclic unit were estimated using the new data between calculation and experiment were obtained obtained by the procedure described in Scheme 3 of for S55-7 and S55-8: both of these structures are ref 5. Eight configurationally isomeric units are products of the trans ring closure. The bicyclic present for the bicyclic unit, and the chemical shifts structure is formed from a cyclic radical inter estimated for the ring carbon are compared with the mediate common to the monocyclic unit. 5 There observed values in Figure 6. The best agreements fore, bicyclic unit S55-7 and S55-8 must be form-

Polymer J., Vol. 13, No.7, 1981 661 M. TSUKINO and T. KUNITAKE

8.4(a)

(k)

l44.3(n)

102.l(m)

80 40 ppm from TMS

Figure 7. 13C-NMR spectra ofpoly(PVE) and its hydrolysis product. The peak shift upon hydrolysis is

indicated by arrows. (l) poly(PVE) obtained at [M] =4.7 moll- 1, in C6 H6 solvent at 60"C; [PDB] =26.5%; the spectrum measured in C6 D6/C6 H6 at 42 wt%. (2) the hydrolyzed polymer; the spectrum measured in C6 D6/CH30H at 47 wt%. ed from monocyclic intermediates corresponding The five-membered monocyclic unit may be to A and B of Figure 5, respectively. formed by the propagation process shown in Scheme 3. There are four kinds of propagation Structure of poly( PVE) Figure 7 is a 13C-NMR spectrum of poly(PVE) and a partial spectrum of the hydrolyzed product. These peaks can be assigned in a way similar to that used for poly(DVE). The peaks of the pendent unsaturation are composed of those of propenyl methyl carbon (a), {3-carbon (m), and c>:-earbon (n). These peaks disappear completely upon hydrolysis, as expected. The hydrolysis causes an up-field shift of peak i and down-field shifts of peaks e and k. Therefore, peak i is assigned to the IX carbon with respect to pendent group and peaks e and k to the f3 carbon with respect to the pendent group. These assignments are consistent with monocyclic struc ture I.

The six-membered ring with the pendent propenyl group (II) cannot be compatible with these shift data, since the {3-carbon in the six-membered ring should be located at ppm, and the hydrolysis does not cause any peak shift in this region. Scheme 3.

662 Polymer J., Vol. 13, No.7, 1981 Cyclopolymerization of Divinyl Ethers routes which produce monocyclic radical interme be determined for their hydrolyzed structures by diates III, IV, V, and VI. 1 H-NMR and 13C-NMR correcting the results of Figure 5 for the effect of spectral data indicate the absence of the pendent additional methyl substitutents. However, before vinyl group and, therefore, intermediates IV and calculating the correction, the location of methyl VI are excluded. Thus, the resulting monocyclic substitution must be determined on the basis of the structure is represented by VII or X of Scheme 4. mode of connection of the monocyclic and bicyclic The steric structure of these monocyclic units can units. In the 70-95 ppm region (carbon next to the ether oxygen) of Figure 7, if peaks i, k, and f are assigned, as mentioned above, to the ring carbons of the monocyclic unit, the remaining peaks g, h, j, and I will be attributed to the ring carbons of the bicyclic unit. If it is assumed that there exist only one kind of bicyclic unit, four kinds of carbon adjacent to the ether oxygen should be present (see below). This is consistent with the spectral datiL Furthermore, the spectral pattern does not vary, when the content of the pendent double bond is changed under different polymerization conditions. These results strongly suggest that a unique monocyclic intermediate (III or V of Scheme 4) is formed and that this in termediate gives monocyclic unit VII or X by intermolecular propagation or cyclizes to either of two possible bicyclic structures (VIII or IX from III and XI or XII from V). Thus, if VII is present as the monocyclic unit, the coexisting bicyclic unit must be VIII or IX. Figure 8 shows all possible connections of the monocyclic (VII) and bicyclic (VIII and IX) units in polymer. The content of the pendent double bond determined by 1 H-NMR spectroscopy indicates that the ratio of the monocyclic to bicyclic units varies from 7 : 3 to 5 : 5, depending on the monomer concentration.

cr: c • \.T I! / C-i-C-C, / 1 : " : HI VII,VIII VII VILVIII,IX IX VI I VILVIIL IX E F c1 -tO-e c: c, \ I , 0 I ' / \. I : cCt;0 , o- , / / ! o ! / i o 1 c-c, VILVIII VIII VII,IIIILIX VII IX VII,VIIf.IX H

-c, : c-+c-tCrt' /o- -0/ : o : IX IX I Figure 8. Linkages of the structural units of poly(PVE).

Polymer J., Vol. 13, No. 7, 1981 663 M. TSUKINO and T. KUNITAKE

c-c-c o c-c-c 40.3 78.1 OH c c -2.1 c-c-c-o c-c-cn c-c-c 40.3 76.0 OH +5.5 c } c-Lc -2.4 86.4 c-c-c c } 75.7 40.3 OH c-c '\ o "·'.a J +0.2

92.3 c-c-c \ o v ;..c-c 4"74.3 4o.6 H Scheme 6. K

75.2 90.9 r compatible only with the presence of the single combination for the monocyclic and bicyclic units, i.e., E and G. (obsd) Now, the chemical shifts for the ring carbon in Scheme 5. combination E (hydroxy derivative) are calculated by starting from the data of Figure 5. An example of Therefore, there must be monocyclic-to-monocyclic the calculation procedure and the resulting chemical and bicyclic-to-bicyclic connections in addition to shifts for the four configurational isomers are given the monocyclic-to-bicyclic connection. in Scheme 5. Among the four structures, J and K Combination E is produced by placing structures give reasonable agreement with the observed data. VII or VIII before VII and structures VII, VIII, or Also given in this scheme is the calculated chemical IX after VII. When VII and VIII are the only shift for structure N which corresponds to X in structural units present, E and G are the only com Scheme 4 (the configuration of N is the same as that binations surrounding the monocyclic and bicy of J). Structure N does not give any satisfactory clic units, respectively. If, however, IX is present in agreement. It may be concluded that in the prop place of VIII as the bicyclic unit, combinations E, F, agation of propenyl vinyl ether, intermediate III, H and I will be formed. These two possibilities are which possesses the configuration of J or K, is discriminated by looking at the number of peaks involved and leads to the configurationally unique ascribable to the ct-carbon of the five-membered bicyclic unit VIII. ring: C 1 to C5 in E to I (Figure 8). Two of these The steric structure of VIII can be inferred by the methine carbons (C 1 and C3) are found when the chemical shift difference estimated from the data polymer is composed of structures VII and IX. On given in Figure 6 for poly(DVE). An example of the the other hand, if VIII is replaced by IX, four of the calculation procedure is shown in Scheme 6 and the methine carbons can be present: C1, C2 , C4 , and C5 , calculated chemical shifts are given in Figure 9. corresponding to structures E, F, H, and I. The We can now complete the assignment of the carbon chemical shift is probably different between observed spectrum (Figure 7). The assignments of

C1 and C2 and between C4 and C5 since the ether the broad peak d (38.8 ppm, carbon f3 to the ether oxygen in the y position causes a down-field shift of oxygen) and peak g (76.3 ppm, carbon IX to the ether ca. 2.5 ppm.9 ·10 oxygen) are made fairly readily, as shown in Figure It was mentioned above that there are three major 9. The remaining ct-carbon peaks h, j, and I at 82.6, peaks for the ring carbon adjacent to the ether 85.2, and 92.0 ppm cannot be assigned specifically, oxygen in the monocyclic structure and four major and structures in which these assignments are in peaks for the ring carbon adjacent to the ether terchangeable may be sought. Then, cis-fused struc oxygen in the bicyclic structure. The NMR data are tures such as S55-9 and 8 55-10 give calculated

664 Polymer 1., Vol. 13, No. 7, 1981 Cyclopolymerization of Divinyl Ethers

c-f-

77.7 4'-4 "79.6 sss-9 s55-1o

76ffi085J c-Lc oJY.Lc-c 42.1 1\0 90.6 r"\ J. -51 86.6 86.3 L_0 rc-c sss-11 s55 -12 c 76ill\094/ vr T 89l e-t-c c-c-c C-C-C -.;;;C-C-C 4,_1 l\0 '•o. 1 41.1 t\ 0 87.3 83.3 80.5 sss-13 'ss-14 o c-c c- r 087/ o c-c- c c-c-c poly (DVE) (obsd) p 40.6 t\84.7 90.6 ' . 84.5 y s55-1s s55-16 -C-Q:'>; c-cy 2 "·zcc- 0 3 · OH (obsd)

Scheme 7.

r 76kBJ\@ 14 and 8 55-15, and 8 55-11 and 8 55-16 remain as -c-c t-c- Q).@. 0, 82.6, 85.2, 92.o probable bicyclic units. 8 55-11 is formed from in 38.8 1\0 \Q) 0 termediate K of scheme 5 and 8 55-16 from (obsd) J, and it is concluded that poly(PVE) Figure 9. Comparison of the observed chemical shift consists predominantly of combinations of J and

of the bicyclic unit ofpoly(PVE) with those estimated for 8 55-16 or of K and 8 55-11. the model compounds.

Structure of Poly(CH3-PVE) chemical shifts which are 4-5 ppm away from the Figure 10 shows a 13 C-NMR spectrum of observed shifts, and may be excluded. Among the poly(CH3--PVE) and its hydrolysis product. The trans-fused structures, the following give fair agree pendent unsaturation is due to 2-methylpropenyl ment: 8 55-11, 8 55-12, 8 55-14, 8 55-15, 8 55-16. group: peaks a and b are assigned to the methyl On the other hand, the chemical shift for exocy carbon, peak p to the {:!-carbon and peak q to the clic carbon can be calculated based on that of methine carbon. On hydrolysis, these peaks disap poly(DVE) as shown in Scheme 7, by taking into peared completely, and three other peaks f, I, and m account the configuration of the adjacent bicyclic shifted. There were found seven peaks in the 70- ring. The influence of the steric difference in the 100 ppm region for the carbon adjacent to the ether monocyclic unit seems negligible, because the oxygen. If peak i is assigned to the five-membered methyl carbons of cis- and trans-2,5-dimethyl ring carbon along with peaks I and m which shifts tetrahydrofuran show very similar chemical shifts. 12 upon hydrolysis, the bicyclic ring carbon will be Structure P gives a better agreement with the ob associated with peaks j, k, n, and o, and only one served values than structure 0. In addition, if three kind of major bicyclic unit• will be present as in adjacent substituents on a five-membered ring are in poly(PVE). Considering the fact that the pendent the cis position, the central methyl substituent group is 2-methylpropenyl and that only one kind becomes, subjected to a strong stereocompression each of the monocyclic and bicyclic units is present, effect and an unusual high-field shift ( -7.5 ppm) the propagation process may be represented by would be observed, 10 contrary to experimental ob Scheme 8. The discussion on the combination of the servation. These considerations exclude 8 55-12, S55- cyclic unit performed for poly(PVE) is similarly

PolymerJ., Vol. 13, No.7, 1981 665 M. TSUKINO and T. KUNITAKE

18.6(b)

(f)' (g) r------::;.;!>i' -9.6 ' ' 14.4(a) (i) : 22.8(c) I 1(1)

(m), (n) 34.5(d) l38.9(q)

ll0.9(p)

93.B(n) i O(ol

140 90 80 40 30 20 ppm from TMS

Figure 10. 13C-NMR spectra of poly(CH3-PVE) and its hydrolysis product. The peak shift upon hydrolysis is indicated by arrows.(!) poly(CH3-PVE) obtained by bulk (8.1 moll- 1) polymerization at

60"C; [PDB] = 26.1 %; the spectrum measured in C6 D6 (C6 H6 at 42 wt%. (2) hydrolyzed polymer, the spectrum measured in C6 D6(CH30H at 38 wt%.

78.1 40.3

c-c'() } 85.3 c;

t ..: c- -c o c-c-c 4o. H1: R 2.1 (73.¥15.1( c-c-c o t-e-e 1: 40.6 c40.7 t s uli .

{obsd) applicable to poly(CH3-PVE) and leads to the Scheme 9. conclusion that the major route of propagation involves cyclic radical intermediate XIII (but not sequently. By a procedure similar to that of Scheme XVII) which produces monocyclic unit XIV and 7, the connection of XIV with XIV or XV was bicyclic unit XV. considered and the chemical shift of the ring carbon The steric structure of XIV was examined sub- was calculated for the expected structure using the

666 Polymer J., Vol. 13, No. 7, 1981 Cyclopolymerization of Divinyl Ethers

78.4

cr cr 80.5 t 41.4 c-c-ciP-r'\ t 39.9 8\_7 s -2o 55 } 80.5

("yG c +8.\ } v = t 0 78 4 cr 95.6 1_' c 0 · 35.4y c-c-c Le-e c-c-c c;-c.-c 41. 0 1:. 41.1 t t c c 83.1 95.6 88 8 s55-24 . +0.5 .,_, "'·6c cJJ-·' c-c-c Le-e c 41. o \c 78.9 98.0 Scheme 10.

addition to those used previously.9 -JJ An example of the estimation procedure is shown in Scheme 10. (obsd) The observed peaks for the bicyclic ring carbons Figure 11. Comparison of the observed chemical shift were assigned as shown in Figure 11, by referring to of the bicyclic unit of poly(CH3-PVE) with those esti the calculated value. Assignments for the Cl. carbon mated for the model compounds. (with respect to the ether oxygen) can be made relatively readily. However, the #-carbons could not procedure of Scheme 5. The procedure and the be assigned definitely, and the subsequent assign results are shown in Scheme 9. The chemical shifts in ment was made by alloting peaks at 43.8 and structures T and U, produced by the cis ring closure, 44.8 ppm to carbons 1 and 2 interchangeably. do not agree with the observed shifts. Structures R Since the bicyclic structure is formed from and S (trans ring closure) give better agreement monocyclic intermediates (with trans ring closure) although the chemical shifts of the ring methylene corresponding to structures R and S, the cis-fused carbon are ca. 3 ppm apart from the experimental bicyclic structures cannot exist. At data. Therefore, the monocyclic unit in the polymer the same time, their calculated chemical shifts do must· have the steric structure corresponding to not agree with observation. The trans-fused bicyclic either R or S. structures (855-23 855-26) do not show partic Structure V in Scheme 9 corresponds to one of the ularly good agreement, but 8 55-24 shows the small monocyclic structures produced through inter est discrepancy with the observed chemical shifts. mediate XVII of Scheme 8. The calculated chemical Although the agreement is not necessarily sufficient shifts for V are quite different from the experimental due to the reason discussed below, 8 55-24 or 8 55-25 data, and the propagation cannot proceed via XVII. becomes the main bicyclic structure when the These results point to structure XV as the most monocyclic radical is related to R, and 8 55-23 or probable bicyclic unit. Thus, the chemical shifts of 8 55-26 becomes the main structure when the the ring carbon of all the conceivable configu monocyclic intermediate is related to S. rational isomers of XV was calculated as shown in Figure 10 contains several peaks in the exocyclic Figure 11. This calculation was carried out by carbon region at ppm. They are attributed taking into account the connection of XV with XIV to the main-chain methylene carbons and to the or XV, as was also done for poly(DVE) and gem-dimethyl carbons. The specific assignment of poly(PVE). The model compounds involve tetra these peaks is difficult because of the considerable hydrofurans with gem-dimethyl substituents13 in variation in the mode of ring connection and steric

Polymer J., VoL 13, No.7, 1981 667 M. TsUKINO and T. KuNITAKE structure. The error is relatively large for the bulky diiso In the case of poly(DVE); the calculated chemical propyl derivative even in the case of this very sim shifts for the bicyclic and monocyclic units agreed ple conversion. satisfactorily with the observed data. The agreement become less as the number of methyl substitution Cyclopolymerization Process increased from poly(PVE) to poly(CH3-PVE). The Scheme II summarizes the cyclopolymerization conformation of the five-memoered ring is not fixed, process of divinyl ethers (DYE, PVE, and CH3- unlike that of the six-membered ring. In particular, PVE). The major course of propagation is common the conformational flexibility of tetrahydrofurans among these three monomers. The monomers react and I ,3-dioxolanes which contain flexible ether with growing radicals always at the unsubstituted oxygen as the ring member has been confirmed by side. The head-to-tail addition gives radical XXI their physical constants14 and by 1 H-NMR14 and which further reacts with monomer in the head-to 13C-NMR9 •15 spectroscopies. It is conceivable that tail manner to produce radical XXII. This radical these flexible rings assume various conformations exclusively cyclizes in the head-to-head manner to depending on the degree of substitution. Our calcu give radical intermediate XXIII. The ring closure lation of the chemical shift was carried out by occurs in the trans manner, irrespective of the simple addition of the chemical shift difference due monomer structure. This cyclized radical undergoes to introduction or conversion of substituents on the either intermolecular propagation (formation of chemical shift of simpler model compounds. It is XXIV) or head-to-head cyclization (formation of assumed in these calculations that the ring confor XXV). The bicyclization process is again stereose mation is fixed. The limitation of this procedure is lective, and the trans-fused structure is formed. The illustrated in the following examples where the ease of cyclization (kc/ kP) is larger for substituted calculated values are derived by corrections due to vinyl groups. The rate of polymerization and the repeated conversions of methyl to isopropyl molecular weight of the polymer become smaller, as substituents. the number of methyl substituents increases. It is interesting that the monocyclization proceeds c- -c c- -c always via the trans ring closure. In the bicyclization process, the steric mode of ring closure is trans for H""" 83.2 \..-J,o.s

,.o... /Rl radic

stereoselective \ y1 h-h cyclization rC01 -CHz-CH'c ,...-(:!{ CH=C, -CH2 2 /Rl H2 c/ R2 o-CH=CR xxu xxm 2 trans-form /(}... CH=C, R2 k intermolecular propagation

kc stereoselective h-h cyclization

trans-fused form Scheme 11.

668 Polymer 1., Vol. 13, No. 7, 1981 Cyclopolymerization of Divinyl Ethers

zation of DVE and CH3-PVE, since the cyclizing REFERENCES carbon is symmetircally substituted. The predomi nance of the trans ring closure is observed in the I. C. Aso and S. Ushio, Kogyo Kagaku Zasshi ( J. Chern. Soc. Jpn., Ind. Chern. Sect.), 65, 2085 (1962). homopolymerization of 1,4-dienes (this study) as 2. C. Aso, S. Ushio, and M. Sogabe, Makromol. Chern., well as in their copolymerization.16 This is in con 100, I 00 (1967). trast with the cis ring closure observed for monocy 3. M. Guaita, G. Camino, and L. Trossarelli, clization of I ,6-dienes. 15·17·18 The five-membered Makromol. Chern. 130, 243, 252 (1969); 131, 237 ring is formed in both cases, but the stereochemistry (1970); 149, 75 (1971). of ring closure is reversed. Further study is required 4. T. Kunitake and M. Tsukino, Makrornol. Chern., 177, 303 (1976). for a unified interpretation of these results. 5. M. Tsukino and T. Kunitake, Macromolecules, 12, The foregoing discussion is restricted to the major 387 (1979). structural units. 13C-NMR spectra of poly(PVE) 6. W. H. Watanabe and L. E. Conlon, J. Am. Chern. and poly(CH3-PVE) contain many small and/or Soc., 79, 2828 (1957). broad peaks, in contrast to those ofpoly(DVE). The 7. C. C. Price and W. H. Snyder, Tetrahedron Lett., 69 presence of minor structures cannot be denied for (1962). these polymers. One possible route leading to the 8. H. Yuki, K. Hatada, T. Emura, and K. Nagata, Bull. Chern. Soc. Jpn., 44, 537 (1971). minor structures is the hydrogen abstraction by the 9. E. L. Eliel, V. S. Rao, and K. M. Pietrusiewicz, Org. propagating radical in either the intramolecular Magn. Resonance, 12, 461 (1979). (back-biting) or intermolecular (branching) man 10. J. Q. Adams and L. P. Lindeman, Prepr., Am. Chern. ners. The complex spectra in the exocyclic carbon Soc., Div. Petrol. Chern., 17, C4 (1972). (25-35 ppm) and methyl carbon (10-20 ppm) II. M. Christl, H. J. Reich, and J. D. Roberts, J. Am. regions found in Figures 7 and 10 point to these Chern. Soc., 93, 3463 (1971). 12. M. Fujiki, unpublished results in these laboratories. possibilities. Another possibility is the presence of 13. A. Accary, J. Huet, and Y. lnfarnet, Org. Magn. sterically and/or geometrically isomeric strucrures. Resonance, 11, 287 (1978). The increased methyl substitution certainly renders 14. W. E. Willy, G. Binsch, and E. L. Eliel, J. Am. Chern. the situation complex. Soc., 92, 5394 (1970). 15. M. Tsukino and T. Kunitake, Polym. J., 11, 437 Acknowledgement. The authors are grateful to (1979). Dr. Koichi Hatada of Osaka University for provid 16. T. Kunitake and M. Tsukino, J. Polyrn. Sci., Polym. Chern. Ed., 17, 877 (1979). ing us with the preparation procedure for methallyl 17. S. R. Johns, R. I. Willing, S. Middleton, and A. K. 13 vinyl ether. We also appreciate the C-NMR Ong, J. Macromol. Sci., Chern., AIO, 875 (1976). measurement at elevated temperatures made by Dr. 18. J. E. Lancaster, L. Baccei, and H. P. Panzer, J. H. Mamezuka of Kyushu University. Polyrn. Sci., Polym. Lett. Ed., 14, 549 (1976).

Polymer J., Vol. 13, No. 7, 1981 669