Po(l'mer Degradation and Stability 2 (1980) 277-293

POLYVINYLCHLORIDE STABILISATION WITH ORGANO- TIN COMPOUNDS--PART 1: MECHANISM OF THE

REACTION OF THIOGLYCOLATE DERIVATIVES WITH ALLYLIC CHLORINE MODEL COMPOUNDS

A. MICHEL & A. GUYOT

CNRS Laboratoire des MatOriaux Organiques, 2, avenue Einstein. 69626 Filleurbanne Cedex, France

&

D. NOLLE

Mffnich University. Meiserstrasse l, 8 Munich 2, West Germany

(Received: 24 October. 1979)

A BSTRA C T

Mixtures of the isomers 4-chloro-hex-2-ene and 2-chloro-hex-3-ene of different compositions were prepared in order to study their reaction with organo-tin thioglycolate. Whatever the initial composition of the mixture, the same composition of the isomeric products is obtained. This favours an ionic mechanism in which the rate-determining step is the formation of an allylic carbocation. The substitution reaction with free thioglycolate esters is more rapid than with the tin derivatire. Thus, a second possible route for the stabilisation reaction can be proposed.

INTRODUCTION

It is now well recognised t after the original work of Frye and co-workers 2" 3,~ that the substitution of the allylic chlorine atoms which were present in the pol3Ttaer or formed during thermal degradation is one of the main reactions in the thermal stabilisation of PVC. Using chlorbut- 1-ene as a model compound, Ayrey et al. 5 have shown that the substitution reaction with organo-tin compounds is accompanied by allylic rearrangement of both the substitution product and the unsubstituted chlorobutene. Considering the distribution of isomers, they suggested that the reaction proceeds through an allylic cation. More recently, Wirth and co-workers 67 proposed a mechanism in which a complex between organo-tin chlorides and the allylic chloride is an intermediate between the two isomers.

277 Polymer Degradation and Stability 0141-3910/80/0002-0277/$02.25 ~ Applied Science Publishers Ltd, England. 1980 Printed in Great Britain

278 A. MICHEL, A. GUYOT, D. NOLLE

In previous publications, 8 we showed that the thermal degradation of 4-chloro- hex-2-ene, a better model than chlorobutene for allylic chloride structures in PVC because the double bond is internal to the chain and not at the chain end, consists of a reversible dehydrochlorination to give diene. This proceeds through an ionic mechanism involving an intermediate carbocation delocalised on carbon atoms 2, 3 and 4. A coloured charge transfer complex between HCI and the diene acts as a catalyst for the reaction. ZnC12 is also a catalyst for the dehydrochlorination and this involves an ion pair consisting of the above carbocation and a ZnCI 3 counter ion. The substitution reactions of zinc or calcium carboxylate and also of various organic co-stabilisers (epoxy, phosphites, aminocrotonates, ct-phenylindole...) are also catalysed by ZnCI,, the rate-determining step being generally the formation of the carbocation which subsequently undergoes either nucleophilic substitution (stabilisation reaction) or competitive elimination (dehydrochlorination). Owing to the Lewis acidic character of the organo-tin chlorides, we consider the same kind of mechanism to be highly probable.

The 4-chloro-hex-2-ene model is actually a mixture of two isomers, the second being 2-chloro-hex-3-ene. If isomerisation proceeds through the carbocation mechanism an equimolar amount of the two isomers is to be expected, both for the chlorohexenes and for their substitution products. On the other hand, if an intermediate complex is operative, the distribution of the isomeric products would reflect the distribution of the initial chlorides. Further, if the carbocation mechanism is operative, an elimination reaction might take place, leading to hexadiene. However, the diene may not only react with HCI because of the reversibility of the elimination reaction but may also add the thiol compound formed in the reaction of HCI with organo-tin thioglycolates.

RESULTS AND DISCUSSION

It has been shown recently by Buruiana et al . ° that the chlorohexene prepared by the reaction of dry HCI with hex-2-ene-4-ol is an approximately equimolar mixture of two isomers.

and

C H 3--C H--C H~----C H--C H 2--C H 3 I Cl

(1)

C H 3---C H~-------C H--C H--C H 2CH 3 I

C1

(2)

PVC STABIL1SATION WITH ORGANO-TIN COMPOUNDS--PART 1

TABLE I NMR SPECTRAL ASSIGNMENTS

279

bt c t d t e t a~ b: d 2 c 2 e 2 a 2

C H 3--C H - - C H ~ C H---C H z ~ H 3 C H 3 ~ H~---C H---C H ~-----C H . , ~ H 3

C1 CI

b 3 c 3 d t % a 3 b.~ d.~ c, e.~ a.~

CH 3 ~ H - - C H~-------C H---C H 2--C H 3 C H ~---CH~---C H - - C H - - ~ H 2---C H 3 I b S- -CH 2- -COOCH ~ S - -C H2- -C OOC H 3

B3 C3 B3 C3

(CH 3h--Sn(SCH:COOCH3) 2 (CH 3).,--SnCI(SC H zCOOCH ~)

At B t C l A , B 2 C 2

(CH3),SnC12 HSCH,COOCH~

A 3 D B C

The composition of the mixture can be analysed quantitatively by means of the N MR spectra of the terminal methyl groups (Table 1): for the first (2-chloro-hex-3- ene) the doublet (b t on Fig. l(b)) due to the methyl group, adjacent to the CHCI group, is shifted by 1-51 ppm with respect to TMS (the coupling constant J = 7 Hz) while, for the second (4-chlor-hex-2-ene), the corresponding doublet (b,_ on Fig. l(b)) is at 1-65 ppm (J = 5 Hz).

As shown in Fig. l(a), a mixture of chlorohexenes of different isomeric composition is obtained by treating at - 5 °C hex-2-ene-4-ol with SOCI 2 in ethanol solution in the presence of quinoline (Q) according to the method proposed by Airs et al. I° In this case, the ratio of (1)/(2) is approximately 4. This reaction probably involves a six-membered ring complex:

C H 3 ~ CH C H , C H 3 / - C H 3 C H C H 2 C H 3

\ \ / CH CH soo:

I

OH

CH CH

"S"

CI OH

Q

C H 3 ~ C H C H 2 C H 3 / \ /

CH CH I

CI

280 A. MICHEL, A. GUYOT, D. NOLLE

b I

a) dl

f t l I I ( ~ . . . . . . e~2

c I

c~ 1

L

Fig. 1.

'j • 1/bl ~ . . . . ~1,2

c 1

I I ~ I I I I

6 5 4 3 2 1 0 ( p.p.m. )

N MR spectra (60 MHz) of mixture ofchloro-2-hexene-3 (1) and chloro-4-hexene-2 (2). (a) Ratio (1)/(2):4. (b) Ratio (1)/(2): I. (Spectral assignments are shown in Table 1.)

The simultaneous occurrence of an ionic mechanism ( through a carbocat ion) might explain the format ion of the minor amoun t of (2). It should be noted that no hexadiene is formed using this method, con t ra ry to the situation when Smith et al's

method ~1 is used, with dry HCI. The 4/1 mixture kept in the presence o f quinoline can be heated up to 150°C, without noticeable decomposi t ion, and can be distilled. It is isomerised in the presence o f a Lewis acid such as ZnCI 2 or (CH3)2SnCI 2, slowly

P V C STABILISATION WITH ORGANO-TIN COMPOUNDS--PART 1 281

at 25 °C and more rapidly at 60 °C. In the latter case, there is no NMR indication of the existence of the intermediate complex suggested by Wirth and co-workers 6'7 for the chlorobutene isomerisation.

In the presence of dilute acids, hydrolytic elimination of HCI takes place readily. An equimolar amount of the two isomeric alcohols is obtained, but comparable isomerisation of the initial chloride is also observed.

The reaction of the 4/1 mixture of the chlorohexenes with dimethyl tin di(methyl thioglycolate) does not take place at room temperature. It becomes rapid only at 100 °C with constant acceleration and to completion. N MR spectra of the reaction mixture are shown in Fig. 2. Again, isomerisation of the initial mixture of chlorohexene is observed when the reaction occurs. The addition of HCI accelerates the initial reaction. The same is true for the addition of the free thioglycolic ester, HSCHzCOOCH 3. When (CH3)zSn(SCH2COOCH3): is replaced by the mono- chloride (CH3)2SnCI(SCH2COOCH3), the reaction begins at lower temperature and is more rapid.

All these results support the ionic mechanism. The rate-determining step is the formation of the carbocation, which is accelerated by the introduction or the progressive formation of a Br6nsted (HC1) or a Lewis acid. (CH3)2SnCI 2 might be efficient in this connection. However, its formation is delayed because of the immediate and complete exchange between dichloride and dithioglycolate to give monochloride, which is less acidic.

The effect of the addition of HSCH2COOCH 3 leads us to study its direct reaction with chlorohexene. In addition, it was a direct route for the preparation of the substitution product. No reaction takes place at room temperature. It starts at 90°C, after a short induction period, and becomes very rapid. However. its rate decreases before completion. Again, it leads to equimolar amounts of the two product isomers and isomerisation of chlorohexene is observed. These data can be explained easily because the reaction gives rise to HCI which accelerates the reaction. However, as HCI accumulates the reverse reaction takes place. A final equilibrium is obtained, as shown in Fig. 3.

The reaction of HSCH2COOCH 3 with 2-4-hexadiene is very slow and is observed only above 150 °. The addition of HCI causes the reaction to start at 120°C. Using the trans-trans isomer as a starting material, one again obtair,.s an equimolar mixture of the isomeric products; with the cis-trans isomer, the composition is a little different at the beginning, but is again rapidly equilibrated.

Although it can be distilled under vacuum (62 °C at 0.1 torr), the addition product is not very thermostable and some decomposition takes place at 200°C. giving an increase of the thiol concentration.

All the NMR data about the various products of that study are reported in Table 2.

The fact that reaction of free thioglycolate with chlorohexenes is more rapid than that of the tin derivative might suggest an indirect route for the subsitution reaction.

8)

b)

A 1

tB 1

%2 )~ / / "

o2 L IC3

b 1

A 2

II~ C3 l A 2

B3 bl a

d1'2 d3'4 c12 ~ b24 ~ t

I

A 3

b3 4

I I I I I

6 5 4 1 0

3 B3

b 4

I i 3 2

(p.p.m) Fig. 2. NMR spectra ofreaction mixture ofchlorohexene and dimethyl tin di(methyl thioglycolate). (a) Initial equimolar mixture or after heating to 90°C. (b) Mixture (a)---after 15 min reaction at 110°C. (c) After addition to mixture (b) (heated at I10°C) of chlorohexene, the chloro tin compound in equimolecular amounts at room temperature. (d) Mixture (c)--after 5 min reaction at 100°C. (Spectral

assignments are shown in Table 1.)

P V C STABILISATION WITH ORGANO--TIN COMPOUNDS--PART 1 283

a)

d 1 C

i b,

D

b 1

22

Fig. 3.

b)

C

c,t/c 2

I I Z

6 5 4

I C b23 ' b 3 [ ~1,23.4

B

B D

c3/ I ~]23,4 I I [

3 2 1 (p.p.m]

1

0

NMR spectra of the reaction mixture of chlorohexene with methyl thioglycolate. (Spectral assignments are shown in Table 13

284 A. MICHEL, A. GUYOT, D. NOLLE

T A B L E 2 N.~R (60 MHz) DATA FOR HEXENYL DERIVATIVES AND HEXADIENE

bl b2 cl c2 doublet doublet quintet quartet J = 7 J = 5 J = 7 J=6"5

C1 1-51 1"65 4-23 4"17 O H - - 1 "64 - - 3-89 S C H . , C O O C H 3 1"21 1"69 3"43 3"41 hexad iene - - 1.67 - - - -

C H 3----C H X - - C H ,~---~ H---C H.,---C H ~ C H 3---C H~-----C H---C HX---C H 2----C H 3

b l CI b2 C2

C H 3--42 H ~----C H---C H~----C H---C H 3

b2 b2

Minute amounts of HC1, produced by a small amount of elimination after formation of the carbocation, might be enough to start the process because HCI reacts immediately with tin thioglycolate to give thiol, which would, in turn, react with chlorohexene with regeneration of HCI. A more complete kinetic study is needed to clarify this point.

The fact that the composition of isomer mixture, either for chlorohexene or its substitution products, is always the same, whatever the initial product composition may be, is a strong argument in favour of the carbocation mechanism. This mechanism, which is well documented in the case of stabilisation of PVC by metal soaps, is most probably very general for PVC stabilisation. It suggests that the ability of a metal chloride to catalyse the substitution reaction will be dependent on its Lewis acidity. Also, according to that mechanism, the competition between the substitution reaction and the elimination reaction will be a constant feature of the organometallic stabilisation of PVC. Work is in progress in our laboratory to substantiate these statements.

EXPERIMENTAL

Materials Alkyltin chlorides and thioglycolic acid were supplied by Ciba-Geigy and Rh6ne-

Poulenc. Thioglycolic acid was distilled at reduced pressure. Anhydrous sodium carbonate

was supplied by Merck.

NMR studies of the preparation of alkyltin thioglycolates and monochlorides 60 MHz 1H NMR spectra were obtained on a Varian DP60 spectrometer using

PVC STABILISATION WITH ORGANO-TIN COMPOUNDS--PART 1 285

CHC13 as a solvent and tetramethylsilane as a reference. Ethyl alcohol is an impurity in C H C I 3.

By mixing methylchlorotin (CH3).SnCI~_., and thioglycolic esters (HSR), no reaction takes place even after heating at 80°C. When SnCl.~ is used. a white precipitate appears which is a complex. N MR spectra of the organo-tin compound methyl group remain unaltered; however, slight shifts of the band of thioglycolic ester are observed, depending on the Lewis acidity of tin chloride: the SH group triplet becomes larger and finally vanishes, although the S--4~H:-group doublet merges to a singlet. These effects are typical for an acid-base interaction.

If anhydrous NaCO 3 is added, immediate reaction takes place with evolution of

TM5

a)

I

p.p.m 4

Fig. 4.

B

C 5

A 4

~S

A S

I I I

3 2 1 0

(A 4 ) 4.

Me 5n C t3 / HSCH2CO2Me

O ¢_)

Od

I /Cs Me 5n (5CH2CO 2 ,'~e) 3

NMR spectra of a stoichiometric mixture of methyl tin trichloride and methyl thioglycolate before (a) and after (b) reaction in the presence of NazCO a.

286 A. MICHEL, A. GUYOT, D. NOLLE

CO2. If a sufficient amount of carbonate is added, reaction:

R4_.SnCI . + nHSR' + nNazCO z

_ /'/ O /7 _ R4 ,Sn(SR'), + ~ C . , + nNaCl + ~NazCO,. H20

is complete and the water produced is totally fixed. Heating up to 80 °C may be needed if R and R' are heavy alkyl groups (n- or iso-

octyl). Figures 4 to 6 show changes in N MR spectra caused by reaction of mono, di and

trimethyl tin compounds, respectively. Reaction with an excess of tin chloride would be expected to give partially

substituted products. However, the excess of NazCO 3 needed to keep the mixture dry might cause formation of Sn--O--Sn bonds. If no excess of carbonate is used,

TM5

p.p.m.

Fig. 5.

a)

b)

C

B

J& A 3

C 1

A 3

l l

3 2 1

A 3

N

4 0

(A 3) ,L Me 2 5nCI 2 /

2 HSCH~COoMe

(D) (B) (C]

0

0,1 : /

Me25n (~CH2 CO 2 Me)

N MR spectra o f a s to i ch iometr ic mixture o f d imethy l tin d ichlor ide and methyl th iog lyco la te before (a) and after (b) reac t ion u p o n addi t ion o f an excess o f NazCO 3.

PVC STABILISATION WITH ORGANO-TIN COMPOUNDS--PART l 287

a)

b)

Fig. 6.

I C

. . . .

I

TM5 p

Me3$nCl /

H SCH.~CO.~ Me (D) (B) (c)

o

'7

(A71 (B 7) (C7) 1, .L

Me3SnSCH2CO 2 Me

I 1. I p.p.m. 4 3 2 1 0

N MR spectra of a stoichiometric mixture of trimethyl tin chloride and methyl thioglycotate before (a) and after (b) reaction.

hydrolysis may take place, and cyclic carboxylate may be formed through the following reaction"

S ,,,~:co, / \

) (CH3)2--Sn CH 2 (CH3)2SnCI2 + HSCH2COOH -:~cr I [

O C

O

The cyclic compound leads to an insoluble polymer. For these reasons, syntheses of alkyltin thioglycolate ester chloride are carried out

using alkyltin thioglycolate compound, either by addition of HCI (obviously in the absence of carbonate) or through an exchange reaction with alkyltin chlorides.

As shown by the spectra in Fig. 7, quantitative formation of free thioglycolate and of organo-tin chloride occurs upon addition of HCI to dimethyl tin dimethyl thioglycolate.

The HSCH2COOCH 3 spectrum is clearly visible at a Cl/Sn ratio of 0-25 (Fig. 7(b)). When this ratio reaches 1 (Fig. 7(d)) the initial spectrum (Fig. 7(a)) for

288 A. MICHEL, A. GUYOT, D. NOLLE

h)

CI/Sn 2 .0

g) CI/Sn

1.8

r) CI/Sn

1.6

e) Cl/Sn 1.3

-4

C A 3

B D A3 A 3

C

y 'B2 B D ~2/3

I B IA2/3

TM5

c%) Me-.s/el

Me / "Cl

(D) (B) (C) HSCH2CO2Me

d) CI/Sn

i.O

c) Cl/5n 0.6

c21[c ~B2 I A2

b, c/cl/l Cl/5n C2 2 B 0.25 ^

A 2

Me "S~ Cl Me/• "5%H2CO2 ~ e

(A 2 ) (B 2) (C 2 )

a) Cl/Sn

0.0

~.p.m. Fig. 7.

BI Me / "5CH2CO2Me • r 'f t

(~) (B 1) (Cl)

I I I

4 3 2 1 0

N MR spectra of reaction products ofdimethy] tin di(methy[ thioglycolate) (a) upon addition of increasing amounts of HCI, CI/Sn from 0.25 (b) to 2 (h).

PVC STABILISATION WITH ORGANO-TIN COMPOL.~DS--PART | 289

TMS

c~ ~1 (~11 (c¢ Me s~SCH 2 CO2~e

I A1 Me" "%CH2co2Me a) B1 I

i Cl/5n = O.0

- . 4 . . . . . . .

. C, A1 [ r

b) C2 A2 Cl/Sn = 0.2S I'

c~

n .g c) ~

A2

d) M~'s~"cl M~" "scH2co 2 ~

~ l ~ CAt2 ) CB2) (C2)

I l I I p.p.m. 4 3 2 1 0

Fig. 8. N MR spectra of the reaction products ofdimethyl tin dilmethyl thioglycolate) upon addition of increasing amounts of dimethyl tin dichloride from O (a) to an equimolar mixture (d).

290 A. MICHEL, A. GUYOT, D. NOLLE

dimethyl tin dithioglycolate has vanished and the spectrum corresponds to (CH3)2SnCISCH2COOCH 3. Upon addition of more HCI. dimethyl dichloride is formed in quantitative yield up to a C1/Sn ratio of 2 (Fig. 7(b)). Broadening of the methyl band indicates scrambling of the ligands of mono- and dichloride compounds. In the same spectra, a singlet is superimposed on the SH group triplet (D); the free SH group may be partly co-ordinated to the tin atom and such co- ordination has to be associated with the scrambling effect.

Chlorotin thioglycolate can be prepared also by disproportionation of dichloride and dithioglycolate compounds according to

CI /

R2SnCI 2 + R2Sn(SCH2COOR')2 ~ 2R2Sn \

SCH2COOR' As shown by the spectra in Fig. 8, the reaction is quantitative and, upon addition

of R2SnCI 2, both di- and monothioglycolate compound spectra are clearly distinguished. Correspondingly, the carbonyl band maximum in the ir spectra shifts from 1738 to 1686cm -~, in favour of the Hutton and Burley interpretation. ~2 A number of compounds with R being methyl-, butyl- or n-octyl and R' being either methyl- or iso-octyl, were prepared. The characteristics of their NMR spectra are given in Table 3.

The present results confirm Hutton and Burley's statement ~2 that reaction between R2SnCI ~ and HSCH2COOR' does not take place easily in the absence of an HCI acceptor. On the contrary, the equilibrium

C1 /

R2SnC12 + HSCH2COOR' ~ HCI + R2Sn \

SCH2COOR'

is totally displaced towards the left because, upon addition of HC1 to tin thioglycolate compounds, tin chlorides are immediately and quantitatively formed.

We also confirm the data of Parker and Carman ~3 showing that exchange between R2SnCI 2 and R2Sn(SCH2COOR')2 gives quantitative )ields of the monochloride compound. Reaction takes place probably through an intermediate complex

CI /

R2Sn \

CI

+

R e

I S R CI SR' \

S n R 2 ~ R--Sn Sn--R /

S C1 S R I I

R' R'

R I

C1 S / \

~ - - - R 2 S n + S n R 2 \ /

Sn Cl

TABLE 3 NMR DATA OF ORGANO-TIN TIIIOGLYCOLATF IN CHCI 3 SOLUTION

(60 M H z - - p p m from TMS)

.< (3

t"

fi t H/pl, m .. _ _ S C I I 2 CO 2 R R = i-oct.vl R = f_'ll 3 C I I j S n ( X )

- - C t l 2 0 - - C l l 2 S - - C t l z S C'1130 iJ i i i / p p m J ( l l z ) C'II3Sn(C'I) S I I

6 l l l / p p m d ( l l z ) fi l l l / IUml

.-4

"Z

H--SCH~CO2R 4-02 3.23 3.27 3.73 Me4Sn MeaSnSCH2CO2R 4.00 3.22 3.27 3.70 Bu3SnSCH2CO2R 4.00 3.22 MezSn(SCH2CO2R)2 4-04 3.39 3.40 3.71 Bu2Sn(SCH2CO2R)2 4.04 3.39 Oc2Sn(SCH2CO2R)2 4.04 3-38 MeSn(SCH,,CO2R): 3 4.05 3.56 3.55 3.74 BuSn(SCH2CO2R)3 4.04 3.55 Me2SnCI(SCH2CO2R ) 4.15 3.59 3.64 3.85 l~Iu 2SnCI(SC 112CO2 R ) 4.15 3.58 Oc2SnCI(SCH.,CO 2R ) 4.15 3.57 BuSnCI(SCH2CO2R) 2 4-13 3.68 BuSnCIz(SCH2CO2R ) 4.25 3-81 CI~SnSCH2CO2R 3-86 4.04

0.07 54 0.07 54 0-47 45 0.66 58

0.84 67 1.22 71

1-29 81 1.66 100

0.97 72 1.22 71

2"03 ,...4

O 7~

>- 7.

,'4

¢.-.) o

7.

I

70 ,.q

i.,.J

292 A. MICHEL, A. GUYOT, D. NOLLE

The fact that the equilibrium is totally displaced towards the right shows that monochloro compounds are stabilised by extra energy AE, which might correspond to an enhanced interaction between Sn--S and SnCI g/Tz bonds. The value of AE is probably low because the monochloride compound cannot be distilled (even for methyl derivatives). Upon heating, equilibrium may be reversed and R2SnC12 is distilled while the dithioglycolate compound is decomposed. As shown by Parker and Carman, equilibrium remains totally shifted to the right with benzyl mercapide

3.38

3.41

3.42

0.87

3.32

3.33

L_

0.9'

3 3 5

I I I p.p.m. 4 3 2

Fig. 9.

0.94

M 5

O,J

=) O

t -

O

,.,.m o - -

1 0

Me25n CI ( S C H 2 C O 2 M e l

C I / $ n = 1 .0

('kJ ¢.J

N

CI/$n = 1.2S

Cl~n = 1.7

NMR spectra of the reaction products of dimethyl tin (chloride) (methyl thioglycolate) upon addition of increasing amount of dimethyl tin dichloride.

P V C STABILISATION WITH ORGANO-TIN COMPOUNDS--PART 1 293

derivatives, but not with tin carboxylate derivatives. This exchange reaction is a general one, but the rate of exchange is dependent on the ligands. Spectra in Fig. 8 show that the rate of exchange between chlorine and thioglycolate groups remains slow when the amount of thioglycolate groups is larger than that of chlorine atoms. In the reverse case, the exchange rate increases because the RzSnCI _, and RzSn(C1)SR' compound spectra merge (Fig. 9). Parker and Carman observed the same results and indicate that these two spectra are separated at - 10°C.

Examination of Table 3 shows that. upon substitution of chlorine atoms, regular shifts occur; these shifts not only concern the CH 3 and SCH 2 groups directly attached to tin atoms, but also O- -CH, , groups attached to the carbonyl group. There is therefore a co-ordination between the carbonyI group and the tin atom, in addition to the Sn--421 bond inductive effect, which was expected for the first two groups. Hutton and Burley's tz ir data indicate that only one ester group is co- ordinated. N M R spectra show that there is a rapid co-ordination exchange, because of coalescence of co-ordinated and non co-ordinated group spectra. Resonances of non-chlorinated compounds are only slightly shifted as compared with free HSCH2COOCH2R. The degree of the coalesced shift is a measure of the proportion of co-ordinated groups. When only one ester group remains, the shift is maximum and, as shown by iT, practically all ester groups are co-ordinated. When R groups are heavy (iso-octyl), the exchange rate for co-ordination becomes slower, as shown by broadening of resonances.

REFERENCES

I. G. AYREY, B. C. HEAD and R. C. POLLER, J. Polymer Sci., Macromolecular Reviews. 8, I (19743. 2. A. H. FRYE and R. W. HORST, J. Polymer Sci., 40, 419 (19593. 3. A. H. FRYE and R. W. HURST, J. Polymer Sci., 45, I (1960). 4. A. H. FRYE, R. W. HORST and M. A. PALIOBAGtS, J. Polymer Sci., A2, 1765, 1785, 1801 (1964). 5. G. AYREY, R. C. POLLER and 1. H. SIDDtQUI, J. Polymer Sci., B8, 1 (1970). Al , 725 (19723. 6. H. O. WroTH and H. ANDREAS, Pure and Appl. Chem., 49, 627 (19773. 7. H. O. WIRrH, H. A. MULLER and W. WEHNER. Paper presented at the 2nd International Symposium

on Degradation and Stabilization, Dubrovnik, October 4--6 (19783. 8. T. V. HOANG, A. MICHEL and A. GUYOT, Fur. Polymer J., 11,469,475 (19753. 12, 337. 347, 357

(1976). J. Macromol. Sci. (Chem.), A12, 411 (1978). 9. E.C. BURUtAYA, G. ROBILA, E. C. BEZDADEA, V. T. BARBINTA and A. A. CARACULACU. Eur. Polym.

J., 13, 159 (19773. 10. R.. S. AIRS, M. P. BALFE and J. KENYON, J. Chem. Soc,, 18 (1942). 11. L. F. SMITH, H. E. HUGNADE, W. M. LAUER and R. M. LEERLEY, J. Am. Chem. Soc., 61,3079 (19393. 12. R. E. HuT'toN and J. W. Bt;RLEY, J. Organometal. Chem., 105, 61 (1976). t3. A. G. PARI,:.ER and C. J. CARMAN, Advances in Chem. Ser., 169, 363 (19783.