The Use of Thiol-Amine Chemical Probes in Network Characterisation

Return to David Packham's home page

The Use of Thiol-Amine Chemical Probes in Network Characterisation

of NBR Vulcanizates

D.Kiroski*, J. Sims*, A. L. Gregory^# and D. E. Packham*

*School of Materials Science, University of Bath, Bath, BA2 7AY, U.K.

^{# Trowbridge College, Trowbridge, Wiltshire, BA14 0ES, U.K.}Der Gebrauch von Thiol-Amin chemischen Sonden bei der Netzwerkbestimmung von Nitril-Butadiene Gummivulkanisierern.

Zusammenfassung
Die Anwendung chemischer Thiol-Amin-Sonden zur Bestimmung der Vernetzungsstruktur in NBR-Schwefelvulkanisaten wurde untersucht. Disulfid- und Polysulfid-Netzbrücken konnten gespalten werden, indem man sie Hexan-1-thiol (1M) in Piperidin 72 h lang in Stickstoff bei 23°C aussetzte; demgegenüber konnten Polysulfid- Netzbrücken gespalten werden, indem man sie Propan-2 thiol (0.4M) und Piperidin (0.4M) in Toluol 2h lang aussetzte. Spaltung vor Netzbrücken kann zur Folge haben, daß bedeutende Anteile des NBR sich vom Netzwerk lösen: zieht man dies nicht in die Berechnung mit ein, verfälscht es die Werte für die Netzwerke. Bei normaler Schwefel- und Beschleunigerkonzentration führt eine Zunahme des Schwefel: Beschleuniger-Verhältnisses zu einer Zunahme an der Gesamtvernetzungsdichte und dem Verhältnis von Disulfid- zu Polysulfidbindungen. Ein Ansteig der Vulkanisationstemperatur ergibt Netzwerke mit geringerer Vernetzungsdichte, die einen höheren Anteil an Monosulfidbindungen besitzen.

Schlüsselwörter: Nitrilkautschuk; NBR; Netwerkstruktur; chemische Sonden;Thiolsonden; Abbau.

Abstract
The application of thiol-amine chemical probes to determine the crosslink structure in sulphur-cured gum vulcanizates of nitrile butadiene rubber (NBR) has been studied. Disulphide and polysulphide crosslinks could be cleaved by treating with hexane-1-thiol (1M) in piperidine for 72 h under nitrogen at 23°C while polysulphide crosslinks could be cleaved by treating with propane-2-thiol (0.4M) and piperidine (0.4M) in toluene for 2 h. Crosslink cleavage can result in freeing significant amounts of the NBR itself from the network: unless this is accounted for a misleading value of crosslink density will result. Within practical sulphur and accelerator loading an increase in sulphur:accelerator ratio results in an increase in total crosslink density and the proportion of di- and polysulphide crosslinks. An increase in vulcanization temperature gives rise to networks of lower total crosslink density with a higher proportion of monosulphide crosslinks

Keywords: Nitrile rubber; crosslink structure; chemical probes; thiol probes; cleavage.

1. Introduction

The sulphur-based crosslinking of diene rubbers is well established and of enormous industrial importance. It is not just the overall crosslink density that is significant, but the crosslink structure - the proportion of monosulphide, disulphide, polysulphide crosslinks and other structures formed - has an influence on many mechanical and chemical properties[1].

Most of the work reported in the literature on the nature of the sulphidic crosslinks has concerned vulcanizates of natural rubber (NR). Such studies have prepared networks of different crosslink structure through modification of vulcanization system and vulcanization time. Thus in NR networks it is well established that an increase in accelerator:sulphur ratio results in improved utilisation of sulphur leading to networks with a higher proportion of monosulphide crosslinks. The influence of compound formulation and vulcanization conditions on crosslink structure of synthetic rubbers such as nitrile butadiene rubber (NBR) has received less attention.

Crosslink structure is generally studied using chemicals (termed chemical probe reagents) which can be homogeneously imbibed into networks and can react with specific crosslink structures at a controlled rate. Thiol-amine chemical probes in particular have been applied extensively since their introduction in the late 1960s to the study of crosslink structure of NR vulcanizates[2,3]. The reaction between thiol-amine and crosslinks results in crosslink cleavage and measurement of crosslink density before and after treatment allows determination of the concentration of reactive (cleaved) crosslinks. In the chemical probe treatment of vulcanizates, most recent studies have used the reagents and reaction times proposed originally by Campbell and Saville. These authors showed that treatment of NR vulcanizates of thickness 1 mm with hexane-1-thiol (1M) in piperidine for 48 h at 25°C cleaves disulphide and polysulphide crosslinks while treatment with propane-2-thiol (0.4M) and piperidine (0.4M) in n-heptane for 2 h cleaves only polysulphide crosslinks. The reactivity and specificity of these thiol-amine chemical probes is sensitive to factors such as, the structure of the thiol, (whether primary or secondary), concentration of reagent, reaction time, reaction temperature, dimensions of the sample and also presumably to the structure of the rubber, such that it is generally necessary to establish optimum conditions before applying such probes experimentally.

The crosslink structure of synthetic rubbers including NBR has been studied using the thiol-amine chemical probes under the original conditions established by Campbell and Saville for NR[4,5,6]. This is surprising because the reactivity depends upon so many factors.In contrast with the methyl group in NR, the nitrile group in NBR is strongly electron withdrawing and might affect the reactivity of thiol-amine chemical probes thereby invalidating the reaction times of Campbell and Saville[2].

This paper presents the results of a systematic study of the application of the Campbell and Saville thiol-amine probes to vulcanizates of nitrile butadiene rubber, aimed at establishing optimum conditions for obtaining selective cleavage of sulphidic crosslinks. A commercial NBR containing 27% of acrylonitrile was selected, and compounded in different ways expected to produce vulcanizates with a range of crosslink structures. The compounds employed are simple non-commercial sulphur based gumstock formulations. Although such compounds are of little use in practical applications, they were chosen in order to reduce complications arising from polymer filler interaction.

In addition to chemical probe treatment, characterisation of networks also requires measurement of crosslink density. This was made by compression deflection measurement on swollen samples. An advantage of testing swollen samples is the better fit of stress-strain behaviour to elastic theory due to separation of network chains and reduction in entanglement effects[7,8]. The Flory-Rehner approach to was not used because of the uncertainty in the value of interaction parameter, c , for probed samples. These were too weak to be tested by conventional tensile stress-strain measurement which would have provided one measurement of c.

In addition to establishing optimum conditions for the use of thiol-amine probes, the work shows the influence of compound and vulcanization temperature on crosslink structure in NBR

2. Experimental

2.1 Preparation of Samples

The base rubber used was Krynac 27.50 (Bayer). It contained 27 % combined acrylonitrile and had a Mooney viscosity of 52 (ML1 + 4 at 100°C). The vulcanizate formulations (Table 1) were selected to facilitate investigation of the effect of formulation on crosslink composition and thiol-amine reactivity. Filler was omitted so as to simplify determination of crosslink density. Further compounds of the PA series were prepared by variation of sulphur accelerator ratio and accelerator type.

Compounds were mixed on a water cooled 2-roll mill using a two stage mixing procedure. Activators and sulphur were added in a first stage and the remaining ingredients of the formulation in the second mix. Vulcanizate sheets of these compounds, of thickness 1 mm were vulcanized in a compression press At 150°C to Tc 95, (time to 95 % of torque maximum, determined by a Monsanto rheometer) and submerged in a water filled container at room temperature following moulding.

Vulcanizate sheets were extracted prior to chemical probe treatment to remove extra-network material with an azeotropic mixture of acetone, 1,1,1-trichloroethane and methanol (110: 60: 42 parts by volume respectively) using cold Soxhlet extraction (room temperature) for a period of 48 h. Residual solvent was removed by leaving vulcanizates under vacuum at room temperature until of constant weight. The weight loss was highest for A3 containing plasticiser (13%). For materials with antioxident (A2 and PB) is was about 6% and 3 to 4 % for the unstabilised vulcanizates A1 and PA.

Table 1. NBR compound mix formulations.

Ingredient	A1	A2	A3	PA	PB	X
Krynac 27.50	100	100	100	100	100	100
MBTS	1.0	1.0	1.0
TMTM	0.5	0.5	0.5
CBS				1.5	1.5
Dicumyl peroxide						1.9
Permanax BLW*		2.0			2.0
Dioctyl sebacate			10
ZnO	5.0	5.0	5.0	5.0	5.0
Stearic acid	1.0	1.0	1.0	1.0	1.0
MC. Sulphur	1.5	1.5	1.5	1.5	1.5

*Permanax BLW - Diphenyl amine antioxidant

2.2 Chemical Probe Treatment

Specimens of dimensions 6 mm x 6 mm x 1 mm were cut from pre-extracted vulcanizate sheets and treated with the following chemical probe treatments:

i. Specimens were pre swollen in n-heptane under nitrogen for 16 h at 23°C, followed by treatment with propane-2-thiol (0.4M) and piperidine (0.4M) in n-heptane at 23°C under nitrogen for periods of 2 h, 4 h and 6 h. In experiments where toluene was used instead of n-heptane experimental procedure was as above and treatment was carried out for 2 h.

ii. Treatment with hexane-1-thiol (1M) in piperidine was carried out at 23°C under nitrogen for periods of 48 h and 72 h. In this treatment pre-swelling was not carried out.

The chemical probe reagent was removed from the network subsequent to probing by washing with petroleum ether, under nitrogen. This involved repeating four times the step of removing the probe solution plus petroleum ether and replacing with fresh petroleum ether followed by standing for 1 h. Probed samples were kept under vacuum at room temperature until of constant weight.

2.3 Determination of Crosslink Density

Compression deflection measurements were made on samples swollen to equilibrium in toluene at room temperature using a Wallace Smith compression-deflection reticulometer. The instrument has been described in detail elsewhere[9,10]. Crosslink densities were calculated using eqs 1 and 2 discussed below. For each vulcanizate six or seven measurements were made on each of three independently mixed and vulcanised samples. Values of elastic constant, C1 which were used to calculate the network chain density (N_v). were thus reproducible to within ±5 %, (expressed to 95 % confidence).

2.4 FTIR Analysis

Analysis of material extracted during chemical probing was made by FTIR using the specular reflection technique. Samples were prepared by allowing extract solutions to evaporate at room temperature on aluminium substrates. The instrument, (Perkin Elmer, model 1720 FTIR spectrophotometer) was operated using a resolution of 2 cm-1. The number of scans was 100.

3. Results and Discussion

3.1 Crosslink density

Crosslink densities were calculated from the results of compression of swollen samples using the relationships described by Mark [7] and by Melley and Stuckey[8] which are briefly summarised here.

In this treatment the elastic constant C₁ of vulcanizates is calculated from the compression measurements made on swollen samples:

F = 2Ao(l-2-l) C1Vr-1/3 1

(F is the compressive force, Ao is the original cross sectional area and Vr is the volume fraction of rubber in the swollen network and l the extension ratio.)

This equation has previously been successfully used in such work [1] for values of compression ratio in the range 0.05-0.15.

The relationship between the constant C₁ and the network chain density (number of network chains per unit volume) N_v comes from the kinetic theory or rubber elasticity:

Nv = 2C1/kTu2s2/3 2

where, Nv is network chain density, T is test temperature (Kelvin), k is Boltzman's constant and u2s is the volume fraction of polymer in the network.

In determination of crosslink density it was assumed that crosslinks were tertrafunctional and crosslink density = 1/2 Nv. This treatment does not consider the effect of chain ends and values of physical crosslink density so obtained are invariably greater than chemical crosslink density because of a contribution to measured retractive force of physical crosslinks associated with chain entanglements.

3.2 Study of Thiol-Amine Reaction Time

Sulphur based vulcanization introduces a range of chemical structures into NBR. These may be stress-supporting crosslinks or non stress-supporting structures, such as cyclic groups, main chain modifications and pendant groups[11]. Stress supporting chemical crosslinks may be mono-, di- and polysulphide. The possibility of the formation of direct carbon-carbon crosslinks during sulphur-based vulcanization was considered. These links are resistant to treatment by methyl iodide, which cleaves monosulphidic crosslinks. When sulphur-cured samples subject to the hexanethiol treatment were then treated with methyl iodide, it was found that no crosslinks remained, indicating an absence of carbon-carbon crosslinks.

Before using the thiol-amine probes to elucidate the crosslink structure it was necessary to establish whether they had any effect on the NBR itself. A peroxide cured NBR vulcanizate (compound X in table 1), which of course has no sulphidic crosslinking, was treated with hexane-1-thiol (1M) in piperidine solution, the more aggressive of the probes, for 72 h. The result shows that NBR is unreactive towards this treatment, the elastic constant C₁ being 1.67 x105 N/m2 before probing and

1.71 x105 N/m2 after probing.

For determination of the contribution of polysulphide crosslinks to physical crosslink density one should compare crosslink densities of untreated samples with those of samples treated with propane-2-thiol (0.4M) and piperidine (0.4M) in n-heptane. This is shown for compounds A1, A2 and A3 in Figure 1. Also shown in this figure is the effect of reaction time on the concentration of remaining crosslinks. In the original work of Campbell and Saville it was recommended that in treatment of NR vulcanizates of approximate thickness 1 mm reaction be carried out for 2 h to affect cleavage of polysulphides while not appreciably affecting disulphide crosslinks[2].

In Figure 1 it is seen that a plateau is reached after approximately 2 h treatment with propane-2-thiol (0.4M) and piperidine (0.4M) in n-heptane. Previous studies into reaction with low molecular weight di-alkyl sulphide compounds[2] have shown that the reaction of such treatment with dialkyl polysulphide is a thousand times greater than with equivalent disulphides. Thus few disulphide crosslinks would be cleaved by this time. As in NR based vulcanizates it appears that the use of a reaction time of 2 h is also suitable for NBR.

The use of n-heptane as solvent for NBR may however be unwise because, according to data of Bristow and Porter[12], its solubility parameter is 7.4 (cal/cm3)0.5 and therefore significantly different to that of 28 wt. % AN content NBR at 9.25-9.38 (cal/cm3)0.5. In order to assess the effect of solvent, compounds PA and PB were also treated with propane-2-thiol (0.4M) and piperidine (0.4M) in toluene for 2 h, samples being preswollen in toluene. Toluene was chosen because it was successfully used to swell samples for compression deflection measurement and its solubility parameter of 8.91 (cal/cm3)0.5, more closely matches that for NBR. The effect of solvent is shown in Table 2.

Table 2. Physical crosslink densities (moles/cm3 x104) of PA and PB after treating pre-swollen samples with propane-2-thiol (0.4M) plus piperidine (0.4M) in solvent for 2 h.

Vulcanizate	Untreated	in n-heptane	in toluene
PA	0.799	0.531	0.396
PB	0.915	0.547	0.462

Clearly, more crosslinks are cleaved when toluene is used in preference to n-heptane as the carrier solvent presumably because toluene is a better solvent for these systems and swells the network more thereby introducing the probe more rapidly and homogeneously into the network. Bearing in mind the relatively short reaction time needed for propane-2-thiol treatment to achieve discrimination between polysulphide and disulphide crosslinks, the use of toluene for pre-swelling of vulcanizates and as carrier for the probe has greater justification. Piperidine, which was used with hexane-1-thiol treatment also produced rapid swelling of vulcanizates.

The effect of reaction time of hexane-1-thiol (1M) in piperidine treatment is shown in Figure 2. Again the reaction times chosen for this work are based on the work of Campbell and Saville 2, 3. Remembering the 5% precision, the results of the figure show that although only few additional crosslinks are cleaved beyond 48 h, a reaction time of 72 h is more appropriate to achieve more complete cleavage of disulphide and polysulphide crosslinks.

During the chemical probing procedure networks were found to suffer weight losses, (shown in Table 3), the values being dependent on compound and probe treatment. Weight loss is particularly high in hexane-1-thiol treated vulcanizates and can not be explained by the small number of additional crosslinks being cleaved by hexane-1-thiol treatment. FTIR analysis of this extracted material gave rise to a spectrum characteristic of NBR polymer, showing that the extracted material is mainly soluble NBR. Such material was combined into the network and insoluble prior to probing, since it was not extracted during cold Soxhlet extraction with an azeotropic solution of acetone, chloroform and methanol. The average molecular weight of such extracted polymer would be in the region of average molecular weight between crosslinks, (Mc). Higher weight loss in hexane-1-thiol treated samples, particularly in PA and PB. This high loss is consistent with higher values of Mc, (i.e. lower crosslink density) of these materials.

The realisation that material which is extracted during chemical probing is low molecular weight polymer raises the question as to its treatment during measurement of crosslink density. There are few published results on the treatment of secondary extraction of networks during chemical probing or conventional swelling in measurement of crosslink density by stress-strain analysis and equilibrium swelling. However, ignoring this elastically inactive volume element would give values of crosslink density which are erroneously high.

Table 3. Physical crosslink density (CLD) and weight loss of NBR vulcanizates treated with propane-2-thiol (0.4M) and piperidine (0.4M) in toluene for 2 h and hexane-1-thiol (1.0M) in piperidine for 72 h.

	Propane-2-thiol treatment		Hexane-1-thiol treatment
Compound	CLD mol.cm-3 x104	Weight loss %	CLD mol.cm-3 x104	Weight loss %
A1	0.700*	2.3*	0.666	6.6
A2	0.634*	0.5*	0.513	4.3
A3	0.578*	1.8*	0.505	4.9
PA	0.396	1.2	0.136	12.7
PB	0.462	1.2	0.231	15.2

*Earlier results with n-heptane as solvent.

Since the extracted uncombined polymer is elastically inactive it may be considered a diluent and its volume added to total sample volume, in values of crosslink density. This will have the effect of reducing crosslink density, most significantly for samples PA and PB. In swelling measurements values of the volume fraction of rubber in the swollen system, (Vr) will also be affected by this and may misleadingly be interpreted as a change in the value of the polymer-solvent interaction parameter, (c) in the Flory-Rehner treatment. It may be significant that previous workers have reported a change in the value of c following probing [2,3].

3.3 Influence of Compound on Crosslink Distribution

The influence of various variables on crosslink density and distribution of vulcanizates cured to equivalent state of cure was investigated by preparing further variation of compound PA (Table 1), in which sulphur:accelerator ratio was varied. The results of crosslink density and distribution of these networks are shown in Table 4. Additional compounds were prepared by replacing CBS accelerator by MBTS and TMTM at equal weight. Results on crosslink density and distribution are shown in Table 5. The influence of vulcanization temperature on crosslink density and distribution of NBR vulcanizates is outlined in Table 6. (These particular calculations do not allow for the volume loss of polymer by extraction.)

Table 4. NBR compound PA modified to show effect of sulphur/CBS ratio on crosslink density and distribution. Cured at 150°C to Tc 95.

S	CBS	CLD mol.cm-3 x104	S1 (%)	S2-6 (%)
0.5	3.0	0.524	94	6
0.5	2.5	0.538	87	13
0.5	1.5	0.372	29	71
1.0	2.0	0.847	28	82
1.5	1.5	0.799	17	83
3.0	1.5	1.982	13	87

Table 4 shows that an increase in sulphur with a simultaneous decrease in CBS, is associated with an increase in crosslink density, and increase in the proportion of disulphide and polysulphide crosslinks at the expense of monosulphides. Similar results have been reported by Brown, Porter and Thomas for vulcanizates based on NR [1]. Here it is shown that accelerator sulphur ratio has a similar effect in NBR vulcanizates. Vulcanizates using TMTM as accelerator give a higher proportion of monosulphide crosslinks than vulcanizates which are compounded with either CBS or MBTS at the same weight. Total crosslink density is also somewhat higher.

An increase in moulding temperature causes a reduction in total crosslink density and a reduction in the concentration of disulphide and polysulphide crosslinks. This is consistent with increased decomposition of disulphides and polysulphides. It is well known that disulphides and polysulphides are less thermally stable and more liable to decomposition than monosulphides [13].

Table 5. NBR compound PA modified to show effect of accelerator type on crosslink density and distribution. compounds of PA. Cured at 150°C to Tc 95. S1 gives percentage of monosulphidic crosslinks, S2-6 gives percentage of di- and polysulphidic crosslinks.

Accelerator type	CLD mol.cm-3 x104	S1 (%)	S2-6 (%)
C.B.S.	0.799	17	83
M.B.T.S.	0.811	33	77
T.M.T.M.	1.092	81	19

Table 6. NBR compound of PA (unmodified): effect of vulcanization temperature on crosslink density and distribution. Cured to Tc 95.

Vulcanization temperature	CLD mol.cm-3 x104	S1 (%)	S2-6 (%)
150°C	0.799	17	83
165°C	0.656	76	24
180°C	0.595	100	0

4. Conclusions

This work clearly shows that thiol-amine chemical probes are suitable for characterising the crosslink structure of NBR gum vulcanizates. The most appropriate conditions differ from those used for analogous work with natural rubber.

Treatment by propane-2-thiol (0.4M) and piperidine (0.4M) in toluene at 23°C under nitrogen for 2 h was found suitable for cleavage of polysulphide crosslinks. Equivalent treatment with propane-2-thiol (0.4M) and piperidine (0.4M) in n-heptane, which is used for NR, resulted in fewer crosslinks' being cleaved probably due to the large difference in solubility parameters between n-heptane and NBR.

Treatment of vulcanizates with hexane-1-thiol (1M) in piperidine at 23°C under nitrogen for 72 h was shown to be capable of cleaving polysulphide and disulphide crosslinks. This is a longer time than usually employed for NR

It is well known that polymer chains may remain uncombined during vulcanization if a low crosslink density results. During swelling of networks, uncombined polymer chains are readily extracted. The results in this paper show the reduction of crosslink density by chemical probes can lead to loss by extraction of a significant amount of the original polymer. Unless this is allowed for in calculating Vr (volume fraction of swollen rubber), consequent calculations of crosslink density will be unreliable.

An increase in accelerator:sulphur ratio in NBR was seen to increase total crosslink density and the proportion of disulphide and polysulphide crosslinks. Compounds based on TMTM are predominantly monosulphide, (approximately 80% of all crosslinks), the remainder being di-and polysulphide. In contrast CBS and MBTS gave networks which contain mostly di- and polysulphide crosslinks (approximately 80%). An increase in vulcanization temperature of a compound containing CBS from 150°C to 180°C results in a reduction in total crosslink density and elimination of di- and polysulphide crosslinks.

5. Acknowledgements

Financial support from the Defence Research Agency, Holton Heath, Dorset, and valuable discussions with Mr M.R. Bowditch, Dr. J.M. Lane and Mr. B. Ochiltree are gratefully acknowledged.

6. References

1. P. S. Brown, M. Porter and A. G. Thomas, Influence of Crosslink Structure on Properties in Crystallising and Non-Crystallising Polyisoprenes, International Rubber Conference, Kuala Lumpur (1985) paper. 1100.

2. D. S. Campbell and B. Saville, Current Principles and Practices in Elucidating Structure in Sulphur-Vulcanized Elastomers, Proceedings of the First International Rubber Conference, Brighton (1967) 1.

3. B. Saville and A. A. Watson, Rubber Chem. Techn. 40 (1967)100.

4. A. S. Deuri and A. K. Bhowmick, J. Appl. Polymer Sci. 34 (1987) 2205.

5. V. Brajko, V. Duchacek, J. Tauc and E. Tumova, Plasty a Kauchuk, 17, (1980) 166.

6. T. C. P. Lee and S. H. Morrell, J. I.R.I., February (1973) 27.

7. J. E. Mark, Rubber Chem.Techn. 55 (1982) 762.

8. R. E. Melley and J. E. Stuckey, J. Appl. Polymer Sci. 14 (1970) 2327.

9. Technical information: Wallace Smith compression-deflection reticulometer, H.W. Wallace & Co. Ltd., St. James' Road, Croydon, Surrey.

10. J. Sims, Studies on the Crosslink Structure and Mechanical Properties of Natural Rubber and Nitrile Rubber Vulcanizates, M.Phil. Thesis, University of Bath (1988).

11. N. J. Morrison and M. Porter, Rubber Chem. Techn. 57 (1984) 63.

12. G. M. Bristow and W. F. Watson, Trans. Faraday Soc. 54 (1958) 1731.

13. T. Kotani and T. Teramote, Intern. Polymer Sci. Techn, 7, (1080) T18.

The Authors:

Dr. D.Kiroski was a research student in the School of Materials Science, University of Bath, Bath, BA2 7AY, U.K. when this work was done. He is now Team Leader Development Chemist at Avon Rubber Company, Melksham, Wiltshire, U.K.

Mr. J. Sims, was a research student in the School of Materials Science, University of Bath, Bath, BA2 7AY, U.K. when this work was done. He is now at Du Pont Dow Elastomers, S.A., P.O. Box CH-1218, Le Grand-Saconnex, Geneva, Switzerland.

Mr. A. L. Gregory is a member of academic staff in the Polymer Centre at Trowbridge College, Trowbridge, Wiltshire, BA14 0ES, U.K.

Dr. D. E. Packham is a member of academic staff in the School of Materials Science, University of Bath, Bath, BA2 7AY, U.K.

Figure captions

Figure 1. Effect of reaction time of propane-2-thiol (0.4M) plus piperidine (0.4M) in n-heptane treatment on physical crosslink density of NBR compounds A1 (o), A2 (¨) and A3 (×). Note the ordinate readings for zero reaction time give the crosslink density of the untreated material.

Figure 2. Effect of reaction time of hexane-1-thiol (1M) in piperidine treatment on crosslink density of compounds A1 (o), A2 (¨) and A3 (×).Note the ordinate readings for zero reaction time give the crosslink density of the untreated material.