Synthesis and ring-opening polymerization of some aryl substituted polycyclic alkenes

Fur. Polym.J. Vol. 27, No. 1, pp. 27-33, 1991 Printed in Great Britain.All rights reserved

0014-3057/91 $3.00+ 0.00 Copyright© 1991 PergamonPress plc

SYNTHESIS A N D RING-OPENING POLYMERIZATION OF SOME ARYL SUBSTITUTED POLYCYCLIC ALKENES W. JAMESFEAST Chemistry Department, Durham University, South Road, Durham DHl 3LE, England LAMIESA. SHAHADA Chemistry Department, University of Qatar, P.O. Box No. 2713, Doha-Qatar (Received 17 April 1990)

Abstract--Three aryl norbornene derivatives have been readily polymerized via ring-opening at the vinylene bond in the presence of the two-component ring-opening metathesis polymerization initiator WCI6/(CHs)4Sn, to give atactic polymers.

INTRODUCTION

RESULTS AND DISCUSSION

This work expands earlier investigations [1, 2] which established that monomers carrying aryl substituents could be polymerized by a catalyst based on tungsten hexachloride activated with a tetraaryl or tetraalkyl tin. In this work, we have extended the range of aryl substituted monomers which can be polymerized by the same technique. Monomers I, II and III were polymerized readily using the two component initiator system WC16/(CH3)4Sn to give atactic polymers. It was observed that polybenzonorbornadiene derived from MoC15/CH3Sn initiation was not very soluble, although it was swollen sufficiently by CDC13 to allow 13C N M R spectra to be recorded [1, 3]. In view of this experience and the relatively low solubility of the polymers obtained from monomer I and II using WC16/Me4Sn initiation, MoCIs/Me4Sn initiation of monomers I, II and HI was not studied at the time of this investigation. The relatively low solubility of samples of polybenzonorbornadiene prepared using well defined titanium initiators has also been noted [4]; however, recent work with a well defined molybdenum based initiator has given soluble polybenzonorbornadiene suggesting that the observed insolubility may be due to "initiator induced" cross linking rather than an inherent property of the material.

(I)

(II)

Monomers

Diazotization of anthranilic acid or 3-amino-2napthoic acid yields an arene diazonium-2-carboxylate which can be pyrolysed [7] or photolysed [8] to produce benzyne or naphthalyne; such species are reactive as dienophiles in cyclo-addition reactions. In earlier work [l] we adopted the method reported by Mich [10] for preparing benzonorbornadiene, since this avoids the risks involved in isolating explosive diazonium salts. In this study 2,3-naphthobicyclo[2.2.1]hepta-2,5diene (11) was prepared via two routes using either naphthalyne addition to cyclopentadiene or benzyne addition to triene (IV) [11]. An extension of this procedure allowed the synthesis of 2,3-anthracenobicyclo[2.2.1]-hepta-2,5-diene (HI) via naphthalyne addition to triene (IV). These syntheses are summarized in Scheme 1. The structures of compounds I and II follow in a straightforward manner from the mode of preparation and were confirmed in a routine manner by elemental analysis, i.r., mass, IH N M R and 13C N M R spectroscopy, the N M R data being the most informative (see Tables 1 and 2). These compounds have not been reported previously. The assignments recorded in Tables 1 and 2 are based on analogy with the spectra of compounds with related structures [1]. Monomer IlI was prepared by the route outlined in Scheme 1, naphthalyne readily added to triene IV to give the 1 : 1 Diels-Alder adduct (V). In practice it proved impossible to isolate a pure sample of compound V because it was so easily oxidized to yield the fully aromatized monomer HI. The easy oxidation of compound V is not surprising because the hydrogens removed during oxidation are both allylic and benzylic and the product is aromatic. The initial product from the addition of naphthalyne to triene IV was a pale yellow solid which was shown to be a mixture of compounds V and IH by a combination of spectroscopic analyses. Thus, mass spectroscopy indicated a

(III) 27

28

W. JAMESFEASTand LAMmSA. Srt~rIADX

~

NH2

N

+ C5 H11 ONO COOH

H

2

+ C5 H1t ONO COOH

~

Monoglyme

l Monog l y me

N2 CO0-

CO0-

~

I -N2

-N2 C02

- C02

-

I ll~(IV

(IV) CsCL40z )

>

(I)

(II)

(V)

(III)

Scheme 1 relatively strong parent peak at m/e 244 with (P + 1) and (P + 2) peaks of appropriate intensity for the formula C19H16 (V), this parent loses 2 amu easily to give a base peak of m/e 242, C19HI4 (III). The u.v. spectrum of the crude reaction product showed band multiplicities and profiles consistent with a mixture of both naphthyl and anthracyl residues; however, it appeared that the naphthyl bands were bathochromically shifted whereas the anthracyl bands were hypsochromically shifted. These observations are consistent with the expected effects of the structural features of compounds V and III on these chromophores. Thus, in the case of compound llI, the hypsochromic shift of the anthracyl bands is consistent with the increased strain due to the bridged bicyclic unit often associated with a shift to shorter wavelength; for example 2max for cyclopentadiene is 239 nm whereas for cyclohexadiene it is 256 nm. In

view of the oxidative sensitivity of compound V, it seemed of little interest as a monomer and so no serious attempt was made to obtain a pure sample but the whole reaction product was refluxed with chloranil in xylene for 24 hr in order to complete the dehydrogenation. Ill was obtained as a crystalline white solid of low solubility which was characterized by elemental analysis, i.r., mass and 13C N M R (see Table 3).

Polymerizations Monomer I and II were found to polymerize easily in chlorobenzene using the WC16/(CH3hSn initiator system to give, respectively, PI and PII. The product polymers were soluble in chloroform; the former dissolved readily while the latter was considerably less soluble. A soluble fraction of PII was obtained by refluxing with an excess of chloroform for about two days, the mixture was filtered and the solvent

A r y l substituted polycyclic alkenes

29

Table 1. tH N M R shifts for compounds I and II measured at 60 MHz relative to internal TMS in d6-aeetone

H6 ~. ~

H5 H4 H4' H2 ~

~

H

Shift' (integrated intensity) HI

H H

H

H

(I ) H4

H3

H

H

H2

7.15 (4) m

H f/H 6

6.80 (2) m

Hl

3.55 (2) bb

H2

3.40 (4) bb

H 4/H 4,

2.02 (2) bb

H3/Hy

7.6-7.4 (6) m

(If)

Hy H 4, H s

6.7 (2) s

Hn

3.93 (2) s

H2

6A 2.21

H

Assignment

JAB = 8 Hz

fib 2.37

H, Hb

aChemical shifts refer to centres of broad bands (bb) or partially resolved multiplets (m) unless otherwise described.

was evaporated from the filtrate producing a viscous solution which was added dropwise to a vigorously stirred excess of methanol. The material which precipitated was recovered by filtration and dried under vacuum. Attempts to polymerize compound Ill in chlorobenzene solvent using both WCI6/(CH3)4Sn and MoCIs(CH3)4Sn catalysts were frustrated by the insolubility of the product polymer Pill. Thus, in all of many attempts at different dilutions the polymer Pill precipitated from solution almost

immediately on addition to the active catalyst solution. This precipitated material was always obtained as a white powdery product which was resistant to all attempts to dissolve it, even prolonged refluxing with aromatic solvents. It was clear from a comparison of the i.r. spectra of monomer III and polymer Pill (Fig. 1) that extensive reaction had occurred. The spectrum of the product, presumed to be HIP, shows the line broadening often associated with polymeric samples and evidence of extensive oxidation (broad Table 3. t3C N M R shifts for compound Ill measured at 75.47 MHz in CDCI 3 relative to internal TMS

Table 2. t3C N M R shifts for compounds I and II measured at 22.64 MHz in CDCI~ relative to internal TMS Compound

5 2

7 8

~

1

(I}

~

7

I

5

a

(If)

2

Shifts

Assignments

144.21 142.39 134.47 129.01 125.77 71.33 52.23 31.05

C4

C~ C6 Cs C7 C3 C2 C5

148.88 141.86 131.99 127.96 125.10 119.26 66.51 49.49

C4 C~ C6 C8 C7 C5 C3 C2

+ HC~CH

Shifts

Assignments

1

147.53

C8

2 4

141.44

C~

131.70

C6

131.18

C4

125.68

C9

124.87

C7

118.91

Cs

64.40

C3

49.38

C2

9

(Ill)

=~HC~CH #

(PII)

(PI)

Scheme 2

30

W. JAMESFEAST and LAM~S A. SHAI-tM)X

~

CH:'~- CH-~n

(PII)

(III) Scheme 3

25

40

30

(III) ~

4000

50

~

SSO0

3000

2500

60

TO

80

90

12 14 161820 25303540 I

10

~

2000 1800 1600 1400 1200 1000

800

600

400 250

Fig. 1. The i.r. spectrum o f monomer (III) and polymer (Pill).

Table 4. Ring-opening polymerizations of !, II and III using WCIJMe4Sn in chlorobenzcne at room temperature Expt No.

Monomer

1

(

2

~

~l[

I1

]

TM

Monomer m. mol

WC16 (MoCI~) m. tool

1.546

0.0079

7.32

M¢Sn4 m. tool

Chlorobenzene (ml)

Time of reaction

Yield

2

2

1

33

0.0367

9.99

6

I

53

4.167

0.02085

5.6

80

2

80

1.488

0.0149

4

15

1.736

(0.0174)

4.7

35

Precipitated powdery material (insoluble product)

(l)

3

[~/~,~ v

V

~

~,~ ~

(IZ)

4 5 (II Z )

Aryl substituted bands in 1700-1750 cm- ~region). Precipitation of the product occurred before any exposure to air and we ascribe these observations to the formation of Pill which is either of an inherently low solubility and precipitates at a relatively low DP, or to catalyst promoted cross linking reactions, the observed oxidation occurring during work up. The polymerization conditions and yields are summarized in Table 4. Polymer characterization

Poly(4, 5 - benzotricyclo [6, 2,1,027] - undeca - 2(7), 4, 9triene) (PI) and poly(2,3-naphthotricyclo[2.2.1]hept2,5-diene) (PII) were characterized by elemental analysis, i.r., ~H N M R and ~3C N M R spectroscopy which confirmed the assigned structures. Elemental analyses were low on carbon PI (found C, 90.16; H, 9.22 calculated C, 92.73; H, 7.21%) while for PII (found C, 89.75; H, 6.24% calculated C, 93.75, H, 6.25%); the deviations between found and calculated values can probably be attributed to oxidation of the sample prior to analysis and/or to solvent contamination. The i.r. spectra of these polymers showed the expected C - - H aromatic absorptions above 3000 cm- ~, aliphatic C - - H absorptions in the region between 2800-2980 cm -~, and carbon-carbon stretching in the 1585 to 1400 cm -~ region. The out-of-plane C - - H deformation modes have been used extensively to identify cis and trans vinylene units, the trans unit giving rise to an absorption in the region 960-980 cm-~ and the cis to an absorption between 665-730

polycyclic alkenes

cm-1, bands are present in these regions in all spectra, consistent with the presence of cis and trans vinylene units in the polymer as would be expected for this initiation system, but the evidence cannot be taken as conclusive because of possible interference from arylene absorptions. The tH N M R data are collected in Table 5. The aromatic-H and vinylic-H resonances for polymers derived from monomer I and II were observed as single broad bands. The allylic and methylene hydrogens for polymer PI were resolved into two bands consistent with the presence of c/s and trans vinylene units. Table 6. Chemical shift in the ~3C N M R of PI 3

--(CH~---CH~

PI PII

--CH 21.3, 1.6 2.8

~SH 3.3, 3.4 4.1

--CH~---CH-5.4 5.8

CH'--CH}n

~8 Peak No. 1

2 3 4

Table 5. Polymer No.

31

5 6 7

AryI-H

Chemical shift*

Assignment

29.88 5t 30.1 5c 38.68 3tt 39.0 3tt 39.26 3ct = 3tc 39.5 3cc 45.19 2cc 50.52 2tt 50.59 2tc 125.78 8 128.87 7 133.86-'~ 134.27)~--~ 4,6,1(c + t) /

/

7.15 7.62

8

Integral

CO

35.39) 19.95) . . . . . . . . 0.36 4.96 7.94 7.62 . . . . . . . . 0.41 7.74 24.12 11.50 17.28 8.736 49.966 Overlap between vinylic acid quaternary carbons

135.1.~

>5 >7 6

50

40

50

Fig. 2. ~3C NMR spectrum of PI recorded at 90.56 MHz, upper spectrum DEPT.

32

W. JAMESFEASTand L ~ m s A. SHAHADA

~3C N M R spectroscopy almost invariably proves to be the most powerful tool for establishing the structure and the cis/trans content of unsaturated polymers; in favourable cases the microstructure may be deduced in detail. The 13C N M R spectra of polymers PI and PII prepared using WCI6/Me4Sn in chlorobenzene were obtained; the assignments for PI follow fairly straightforwardly from the background data of polybenzonorbornadiene presented earlier [1]. Spectrum of PI, Fig. 2, shows two regions in which resonances occur, viz. the alphatic carbon resonances from 25 to 55 ppm and the olefinic and aromatic carbon resonances from 120-150 ppm. The DEPT spectrum, in which CH2 carbons are inverted, CH are unchanged and quaternary signals vanish, allows us to assign the resonances at ca 30 and 39 ppm to the methylenes labelled C5 and C3 respectively; some fine structure is almost resolved but the resolution is insufficient to make unambiguous assignments, see Table 6. The other aliphatic carbon resonances are due to the g-carbon, C2 and occur at 45.19 C2 (allyl carbon adjacent to cis vinylene) and 50.52 and 50.59 C2 (allyl carbons adjacent to trans vinylene) giving a cis vinylene content ac = 0.46; although it is not clear in Fig. 2, the methylene resonance at ca 30 is assigned by the N M R instrument's computer to two signals at 29.88 and 30.10ppm and, if we assume the splitting of these signals is due to cis/trans vinylene effects and assign the signals as 5t and 5c respectively, the computer integration generates a value of crc= 0.36 which is in rather poor agreement with the a¢ value derived from the allylic carbon signals. The methylene at C3 would normally be expected to appear as at least a three line signal [3(tt):3(tc=ct):3(cc)] but four lines are observed in the recorded spectrum (Fig. 2, Table 6). A provisional assignment of these lines is based on the assumption that the 3(tt) line is split, possibly as a result of syndio and isotactic sequences; using this assignment, a ¢~ value of 0.41 was computed. The instrument computer also splits the allylic carbon signal adjacent to trans vinylenes into two resonances at 50.59 and 50.52, although this is not apparent in C3

C5

C7

Fig. 2, whereas the allylic carbon signal adjacent to cis vinylenes at 45.19 appears as a fairly narrow line.

The significance of these differences with respect to microtacticity is not certain but it may indicate that cis dyads have only one tacticity (presumably syndiotactic) whereas trans vinylenes occur in both syndio and isotactic dyads; however, in the absence of other evidence, secure assignments of these multiplicities cannot be made since they may result from long range cis/trans vinylene effects and/or tacticity effects. The aromatic, C7 and Cs, carbon resonances are assigned as shown in Table 6; unfortunately the quaternary carbons at C4 and C6 overlap with the vinylene resonances as is clearly demonstrated by comparison of the normal and DEPT spectra in which quaternary carbon resonances vanish. In the DEPT spectrum the vinylene resonances are split into two roughly equal signals giving a ~rc value of 0.37 which is consistent with the earlier computations. We can conclude that the polymer PI produced by WCIt/Me4Sn catalyst appears to have the expected structure and an ca 40:60 distribution of cis and trans vinylenes, but unfortunately S/N and resolution are insufficient to say anything positive concerning the tacticity of the polymer, although the indications are that the polymer is essentially atactic. The 13C N M R spectrum of polymer PII produced using WCIt/(CH3)4Sn is shown in Fig. 3; the shifts and assignments are given in Table 7. Polymer PII was difficult to dissolve; it dissolved slowly in a large excess of chloroform and was recovered, after concentration to give a viscous solution, by precipitation in a large excess of cold methanol. The limited solubility of this polymer required a very long accumulation time ( = 66 hr), which precluded recording a DEPT spectrum. By comparison with the assigned ~3C N M R spectrum of polymer PI, the two bridging quaternary aromatic carbons, C4 and C6 are associated with the lower intensity signals in the low field part of the spectrum. The broad signal at ca 145 ppm is assigned to C6, and the signals due to C4 appear as a doublet with cis C4 (130.74 ppm) downfield from the trans C4 (128.79 ppm). The vinylene carbon

C8

4

-an

~5 C2

C3

,7 C4 C6

W

140

430

120

Fig. 3. ~3C NMR spectrum of Pll recorded at 90.56 MHz.

r

I

Aryl substituted polycyclic alkenes Table 7. Chemicalshifts in the 13CNMR of Pll

(-CH-'CH~

Peak No.

1

2 3 4 5 6 7 8

3

"1

ell--"CH-)n

Chemical shift Assignment Integral 42.61 42.75 43.08 43.51 43.92 47.33 47.51 122.27 125.25 127.67 128.79 130.74 145.11 133.15 133.39

3tt 3 t c ~ et

133.68 133.86

lc

3cc 2c 2t cs c~ c5 4t 44: c6 It

22.08 18.42 6.49 13.93 17.02 19.97 11.24 65.57 78.33 68.13 13.66 9.48 30.36 49.63 21.76

CONCLUSIONS 0.34

0.36

0.41 --

10.01 31.03

signals occur as two overlapping multiplets centred at 133.86 and 133.15 ppm, the multiplicity possibly being due to cis/trans sequence or tacticity effects. In the aromatic region, we can observe three fairly sharp signals at ca 122.27, 125.25 and 127.67ppm due to the aromatic carbons Ca, C7, C5 respectively. The methine carbons are both allylic and naphthylic and consequently their resonances occur at lower field than the methylene resonances. In the spectrum of polymer PII, the methine carbons appear as an unresolved signal at 43.92 assigned to the cis while the trans signal appears as a doublet at 47.33 and 47.51 ppm. The methylene peaks at 43.51 and 43.08 ppm are assigned to C3 (cc), while the peaks at 42.75 and 42.61 are assigned to the 3ct = 3tc and 3tt respectively. The justification for these assignments is that the proportion of cis double bonds calculated using the various sets of signals assigned as above are in quite

EPJ 27/1~

good agreement. Thus from the vinylene carbons, ac was computed as 0.37; from the methine carbons, a value of ac = 0.36 was obtained and from methylene carbons the ere value was 0.34. Using the bridging aromatic carbon at C4 ~ = 0 . 4 1 , the discrepancy between this value and those derived from Ct, C2 and C3 may well be due to the fact that the spectrum phase appears to be marginally out in the C4 region.

ac

0.37 9

33

Arylnorbornenes undergo metathesis ring-opening polymerization (ROMP) when exposed to the twocomponent initiator system WCIt/(CH3)4Sn. The polymerization is non-selective giving rise to atactic polymers with both cis and trans vinylenes in the main chain. Polymers derived from anthracylnorbornadiene were insoluble. REFERENCES

1. I. F. A. F. EI-Saafin and W. J. Feast. J. moiec. Catal. 15, 61 (1982). 2. W. J. Feast and L. A. H. Shahada. Polymer 27, 1289 (1986). 3. L. A. H. Shahada. Ph.D. Thesis, Durham University (1984). 4. L. F. Cannizzo and R. H. Grubbs. Maeromolecules 21, 1961 (1988). 5. J. S. Murdzek and R. R. Schrock. Organometallics 6, 1373 (1987). 6. G. C. Bazan, E. Khosravi, R. R. Schrock, W. J. Feast, V. C. Gibson, M. B. O'Regan, J. K. Thomas and W. M. Davis. (Submitted for publication). 7. M. Stiles and R. G. Miller. J. Am. chem. Soc. 8 2 , 3802 (1960). 8. R. S. Berry, G. N. Spokes and M. Stiles. J. Am. chem. Soe. 82, 5240 (1960); 84, 3750 (1962). 9. M. Stiles, R. G. Miller and U. R. S. Burckhardt. J. Am. chem. Soe. 85, 1792 (1963). 10. T. F. Mitch, E. J. Nienhouse, T. E. Farina and J. J. Tufariello. J. Chem. Edu. 45, 272 (1968). 11. K. Alder, S. Hartung and O. Netz. Chem. Bet. 90, 1 (1957). 12. K. J. Ivin. Olefin Metathesis. Academic Press, London (1983).