Characterisation of vulcanised natural rubber behaviour by monotonic and in situ cyclic X-ray scattering tests

Abstract

The effect of curing and loading conditions on the mechanical response of natural rubber are investigated by monotonic and in situ X-ray cyclic tensile tests. Tests are conducted on four samples, which differ by vulcanisation conditions. Samples are subject to two strain rates (2.7 × 10^− 3 s^− 1 and 16.66 × 10^− 3 s^− 1), and numerous imposed elongation levels range from 450 to 900%. The coupling between the strain rates and the elongation levels on the stress softening evolution resulting from strain induced crystallisation is investigated. In situ thermomechanical tensile cyclic test is performed in order to withdraw the effect of the strain induced crystallisation on the maximum stress decrease. The experimental results analysis shows that an optimum vulcanisation condition (150°C, 30 min) enhances the hardening process in the monotonic loading due to the strain induced crystallisation. However, under optimum curing conditions, cyclic loading induces a large hysteresis loss, a high stress softening and a high degree of strain induced crystallinity. The material softening sensitivity is controlled by coupled effect of strain ranges and elongation levels. This panoply of experimental measurements present a key information for material parameters identification that are useful to predict the lifetime of engineering components made of natural rubber such as racks, laminated rubber bearings and tires.

Keywords

Crystallisation Rubber WAXD Behaviour Cyclic softening

Introduction

Among various types of elastomers, natural rubber (NR) is one of the most used in industrial and household applications. The interest of NR based materials comes from their better mechanical properties. Indeed, NR is known to be a tough elastomer, which can undergo large strains. It exhibits high tensile stress, large hysteresis loss¹ and crystallisation upon stretching.²

The larger hysteresis loss associated to first mechanical cycles is due to the Mullins effect. While many studies have been done on this purpose since the last four decades, the topic is still interesting and subject of investigations. The physical interpretations of the Mullins effect were based on bond rupture,³ chains disentanglements,⁴ network rearrangement, filler cluster,⁵ molecule spilling⁶ and chain retraction as reported in numerous works.

As for the so called strain induced crystallisation (SIC), which is a result of the molecular orientation generated by the stretching process, many researches based on volume change measurements,⁷ nuclear magnetic resonance (NMR) techniques,^{8, 9} X-ray diffraction,^{9, 10} infrared spectroscopy,¹¹ birefringence,¹² stress relaxation and dilatometry were performed. These investigations are performed either in sequential or in simultaneous manners. However, there is a crucial controversy on the draw ratio at which crystallisation occurs.¹³

It has been established that crystallisation upon stretching considerably affects the mechanical behaviour of rubber in both monotonic and cyclic loading. In fact, crystallites act like natural reinforces, which enhance the resistance to fracture and failure.¹⁴

Since this dissipative behaviour is extensively related to different molecular mechanisms and rearrangements during mechanical loading, one proposes to master the length of macromolecular chains, which depends on the distances between the crosslinking knots using different curing times and mixture temperatures. The sample compounds as well as the curing times and conditions are given lately. It should be noted that it is beyond the scope of this paper to cover the determination of crosslink topology and structure. Interested readers could refer to numerous pioneer works that mention the swollen rubber method,¹⁵ gas chromatography¹⁶ and NMR analysis.¹⁷

The purpose of the study is to better understand the dependence of strain rates and vulcanisation conditions on the kinetic of SIC and on the mechanical response of NR. This should give details on the contribution of each process on the dissipative response of NR and therefore simplify the identification strategy of material parameters required to model this behaviour.

Experimental

Material

The materials are obtained by sulphur vulcanisation of NR. All samples, called afterwards NRI, II, III and IV, have the same concentration of ingredients (stearic acid, zinc oxide, antioxidant, sulphur and accelerator), but the curing times and temperatures are different. Formulations of rubber compounds as well as curing time and conditions are listed in Table 1. The vulcanisation kinetic is controlled by means of an MDR 2000 Moving Die Rheometer according to ASTM D5289 standards for the two imposed vulcanisation temperatures 150 and 170°C.

Table 1

Recipe of compounds

		Samples
Ingredients	Units	NRI	NRII	NRIII	NRIV
NR	g	100	100	100	100
Stearic acid (activator)	g	2	2	2	2
Zinc oxide (activator)	g	5	5	5	5
Antioxidant (6PPD)	g	1	1	1	1
Sulphur (vulcanising agent)	g	1.2	1.2	1.2	1.2
Accelerator (TBBS)	g	1.2	1.2	1.2	1.2
Curing temperature	°C	150	150	170	170
Curing time	min	15	30	15	30
Network chain density ν	× 10^− 5 mol g^− 1	5.3	10.7	7.8	6.2

Figure 1 displays the vulcanisation curve at the two given vulcanisation temperatures. This provides an insight into the vulcanisation stage according to chosen curing time. Three regions can be spotted on the vulcanising curve. The first region is the induction stage or scorch time, which is the time elapsed before crosslinking starts. The second region is the curing reaction period or crosslinking reaction. The third region is the maturation stage where the crosslinking network is mature. The end of the maturation period is recorded at the maximum torque value (plateau). Some overcuring reversion may occur when the curing is too long.

Vulcanising curve [torque (dN m) versus time (min)] for 150 and 170°C curing temperatures

The network chain density of each sample could be estimated on the basis of classical theory of rubber elasticity using the relation hereafter¹⁴: where ν is the network chain density, k is the Boltzmann constant, T is the absolute temperature and λ = l/l₀ is the elongation (l is the elongated length and l₀ is the initial length of the specimen) (Table 1).

The specimens (provided by S.T.I.P CO industries) are prepared from a 3.6 mm thickness rubber sheet. A very sharp cutting edge and the same cutting method are used for all the dumbbell shaped tensile test samples. The specimen geometry and dimensions according to standards of thermoplastics and thermo sets (ASTM D638) are illustrated in Fig. 2.

Tensile specimen geometry and dimensions according to ASTM D638

For wide angle X-ray diffraction measurements, thinner specimens (2 mm) with sizes 6 mm × 33 mm were prepared.

Monotonic and cyclic tensile test conditions

The monotonic and cyclic tension tests are conducted on a uniaxial tension machine Instron. The elongation is determined from the distance between the clamps during the deformation. The initial distance between the clamps is 30 mm. For different imposed deformation levels and rates, all measurements are conducted at room temperature. Each specimen is continuously elongated up to the predetermined strain range and then retracted to the original length. The experiment is performed three to five times consequently non-stop during the exposure until stress–strain response is stabilised. The results of the tensile stress–strain measurements are reported as a function of the retraction force F(N) and the elongation λ. Each test was repeated up to eight times to ensure reproducibility of results.

In situ X-ray scattering cyclic tests

Stresses and elongations are measured during the cyclic test using a homemade stretching machine mounted on an X-ray rotating anode generator. The X-ray measurement conditions are summarised in Table 2.

Table 2

X-ray scattering measurements conditions

Cu K_α wavelength	X-ray generator voltage	X-ray generator intensity service	CCD exposure time
0.1542 nm	40 kV	40 mA	10 s

The incident X-ray beam is focused on the sample by a doubly curved graphite monochromator, and a collimator [Fig. 3, (4)] reduces the diameter of the beam to ∼0.8 mm. The samples are fitted at the collimator exit to ensure minimal air scattering, and the diffraction patterns are recorded with an indirect illumination CCD camera [Fig. 3, (2)]. An optical camera is used to position and centre the sample. It may also serve to compute the sample elongation using a precise measure of the sample width. An infrared sensor [Fig. 3, (1)] measures the temperature at the surface of the sample.

Experimental set-up: (1) infrared sensor, (2) XR-CCD camera, (3) load cell, (4) collimator, (5) camera shutter, (6) optical camera, (7) X monochromator

A photodiode is fixed within the beam stop in order to measure the transmitted beam intensity I_t(λ). It serves to correct the diffraction data and evaluate the sample elongation by a precise measure of the thickness d(λ) using equation (2) where μ is the absorption coefficient and I₀ is the incident beam intensity.

The load sensor [Fig. 3, (3)] that is fixed on a translation stage records the retraction force [F(N)] at each elongation λ. A thermomechanical cyclic test at a fixed temperature of 80°C is performed by blowing air into a polycarbonate housing that ensures temperature homogeneity. At this given temperature, SIC does not take place in the present conditions (drawing speed and maximal applied elongation). This test is prominent for comparison and later for identification.

X-ray scattering analysis enables us to access the local segmental orientation of the polymeric chains within the amorphous phase.¹⁸ Recent studies provide a reliable evaluation of the behaviour of the segmental order parameter using X-ray data.^{19, 20} The segmental order parameter of the remaining molten fraction is defined by equation (3) where I_am(φ) is the radially averaged intensity of the amorphous halo that is adjusted by the function A+B cos²(φ); P₂(φ) is the second Legendre polynomial; and φ are the angular scans that include the zone that comprised two dotted circles as shown in Fig. 4.

Diffraction pattern obtained after SIC onset

Results

Monotonic behaviour

All samples were stretched at the same strain rate (1.11 × 10^− 3 s^− 1) until failure. In the most industrial components using NR, elongations do not exceed 1000%. Therefore, it will be interesting to limit this investigation up to this elongation (Fig. 5). NRII has a relatively high stress hardening compared to the other samples. This behaviour is attributed to the curing conditions.

Monotonic tensile test: retraction force–elongation curve of samples at 1.11 × 10^− 3 s^− 1 strain rate

Indeed, the kinetic of sulphur vulcanisation is composed of three stages: induction, crosslinking and maturation (Fig. 1). During the main crosslinking reaction, the sulphur bonds are composed of mono-, di- and multisulphur molecules. The proportion of the three type of sulphur bonds depends on the curing time and temperature and on the accelerator/sulphur ratio.²¹ According to the formulation given in Table 1, the vulcanisation process is a semiefficient (SEV) curing system. According to Akiba and Hashim²² investigations, an optimum curing process with SEV curing system leads to a proportional amount of poly/disulphur crosslinks. However, when the maturation period is followed by reversion, which consists of the extension of the cure beyond the time required to obtain the desired optimised balance of vulcanisate properties, the thermal degradation of unstable polysulphide crosslinks occurs,²² which is the case of sample NRIV (as detected by the moving die rheometer) (Fig. 1). The curing time of NRI was not sufficient. This ascertainment is provided by the vulcanising curve (Fig. 1). For sample NRII, the crosslinking rate has reached its maximum within 30 min of vulcanisation, which can be explained by an optimal diffusion of sulphur into the structure. Then, materials NRII and NRIII are retained for the cyclic loading investigations.

Cyclic loading and in situ X-ray measurements

The in situ X-ray scattering cyclic tests are conducted on samples NRII and NRIII. The crystallinity index is based on the most intense reflections. The data treatment is detailed elsewhere.¹⁹

Effect of vulcanisation conditions

As shown in Fig. 6a and b, the intensity of the slope of the hardening phase as well as the beginning of the crystallisation phase is quite different for NRII and NRIII for the same strain rate (Fig. 6c and d). The first crystallisation cycle is different from the subsequent cycle.

Cyclic loading at 2.7 × 10^− 3 s^− 1 strain rate: a retraction force F(N)– elongation cycle for NRII; b retraction force (N)–elongation cycle for NRIII; c crystallisation–elongation cycle for NRII; d crystallisation–elongation cycle for NRIII

This could be explained by the difference between the macromolecular chain lengths. In fact, a number of short chains reaching their limit of extensibility will break during the first cycle and will no longer take part of the stretching process for the subsequent cycles. This is clearly emphasised in Fig. 6a and b by the discrepancy between the maximum reached stresses of the first and the second cycle.

The stress softening is well marked in sample NRII. This discrepancy is quite proportional to the maximum crystallinity evolution for the first and the second cycle (Fig. 6c and d). The crystallinity increases faster in NRII sample due to the topology of crosslinking molecules, as NRII sample is more crosslinked. The onset of crystallisation and melting is dissimilar for samples NRII and NRIII. This is again due to the chain rupture during the first stretching cycle that induces an increase of the local stresses and an early crystallisation onset for NRIII sample.

Illustrative test of the mechanical cyclic loading, the crystallinity evolution and the segmental order parameter is given in Fig. 7 for NRII and NRIII samples.

Stabilised cycle at 2.7 × 10^− 3 s^− 1 strain rate: a mechanical cyclic loading (stretching recovery) of NRII and NRIII; b crystallinity index evolution for NRII and NRIII; c evolution of segmental order parameter for NRII and NRIII (λ₀, λ₁ and λ_M are elongations at SIC onset and melting)

Three macroscopic elongations relevant for the analysis—λ₀, λ₁ and λ_M—are the elongations at SIC onset and melting as detected at experimental time scale. The values of λ₀ and λ_M are determined by linear extrapolation. At crystallisation onset, a plateau in evolution appears. During recovery, the crystallisation index decreases linearly until λ₁ and then a change of the crystallinity index slope occurs. This observed behaviour is similar with evolution. The elongation at complete melting is λ_M. Physical explanation of these elongations is detailed elsewhere.¹⁹

A summary of test data is presented in Table 3. It should be noted that these elongations (λ₀, λ₁ and λ_M) are kinetic parameters (function of strain rates).

Table 3

Macroscopic elongations and crystallinity index by in situ X-ray analysis

Network chain density ν	Strain rate	λ/%	λ₀/%	λ_I/%	λ_M/%	ξ/%
10.7 × 10^− 5 mol g^− 1	2.7 × 10^− 3 s^− 1	300	…	…	…	…
		500	380	340	205	7.4
		600	450	400	300	11
	16.66 × 10^− 3 s^− 1	300	…	…	…	…
		450	337	350	225	6.01
7.8 × 10^− 5 mol g^− 1	2.7 × 10^− 3 s^− 1	300	…	…	…	…
		600	440	350	230	7.9
	16.66 × 10^− 3 s^− 1	300	…	…	…	…
		500	345	370	230	4.7

Effect of strain range on stress softening

In this section, cyclic loading tests are conducted with different imposed elongation levels for two strain rates of 2.7 × 10^− 3 s^− 1 and 16.66 × 10^− 3 s^− 1. Figure 8 illustrates the cyclic stress softening through the amplitude of maximum stress decrease versus the number of cycles needed to reach a steady state. It has been found that for the same imposed elongation level, material sensitivity to the strain rate is different. The imposed deformation levels are then split into three ranges. The first range is for elongations that do not exceed 450%. The second range is for elongations between 450% and 600% (Fig. 8a), and the last one is for elongations higher than 600% (Fig. 8b).

Cyclic stress softening of NRII sample: effect of strain rates and of imposed deformation levels on stress softening: a second range of elongations; b third range of elongations

For the first range of elongations, when the strain rate increases, almost the viscous effect is the genesis of internal friction. This is clearly superimposed in Fig. 9, where the crystallinity index is insignificant compared to higher elongations.

Effect of strain rates on SIC evolution

In the case of the second range of elongations (450–600%) at low strain rate of 2.7 × 10^− 3 s^− 1 with increasing imposed deformation level, the cyclic stress softening increases, in contrast to the strain rate of 16.66 × 10^− 3 s^− 1 where the increase of imposed elongation level reduces the stress softening as illustrated in Fig. 8a. At this given strain rate, both the entanglement points and the crystallisation upon stretching act as ‘crosslink points’. This leads to more chain segments being oriented, which promote the hardening process for 600% imposed elongation level. Such an explanation seems to be appropriate also for the case of 600% imposed elongation with strain rates of 2.7 × 10^− 3 s^− 1 and 16.66 × 10^− 3 s^− 1, where both the viscous effect genesis of internal frictions and SIC enhance the chain orientation and promote the hardening process. This competition between the viscous friction and the kinetic of crystallite formation depends strongly on the mechanical loading conditions (elongation levels and strain rates). Such ascertainment is only applicable for the sulphur vulcanised NR, whereas for peroxide vulcanised NR there is a very small dependency of the stress–strain evolution for the same varied stretching speed used in the present study.²³

As for the third range of elongations (750–900%) (Fig. 8b), the effect of strain rate on the cyclic stress softening is negligible. This seems logical for higher imposed elongations as the relaxation process between the amorphous chains and the formation of newly crystalline phase is enhanced. Owing to this relatively long relaxation time, the strain rates have little effect on the stress softening. This ascertainment is in accordance with Zhang et al. work²⁴ made on the biopolymer poly(L-lactic acid).

Effect of SIC on stress softening

In order to better understand the contribution of the SIC in stress softening under repeated loading, an in situ thermomechanical test at 80°C was performed. It is confirmed that hysteresis loss is mainly due to the SIC. This is clearly emphasised in Fig. 10 where, for the same imposed elongation level of 600% and strain rate of 16.66 × 10^− 3 s^− 1, no significant stress softening is recorded in the absence of SIC (at 80°C) in comparison with the test conducted at room temperature. The grey area drawn in Fig. 10 represents the impact of SIC on the stress softening.

Effect of SIC on cyclic stress softening of NRIII at 16.77 × 10^− 3 s^− 1 strain rate

The maximum stress decrease could be related to the two coupled sliding mechanisms associated to the amorphous and crystalline phases. The duality between the phenomena of stress softening and internal friction, which subsequently induces crystallisation by stretching (see grey area in Fig. 10), could be modelled and identified. Based on the present experimental finding, such analysis is currently under progress.

Conclusions

It has been established on the basis of the experimental results that: 1.

Rubber vulcanisation at 150°C for 30 min leads to an optimum hardening process in monotonic stretching.

Rubber vulcanisation at 150°C for 30 min produces a large cyclic hysteresis loss, a high cyclic stress softening and a high SIC amount.

The higher is the imposed elongation level, at low strain rates, the more important is the cyclic softening.

The effect of strain rate on the cyclic stress softening is insignificant for high imposed elongation levels.

For elongations between 450% and 600%, with increasing the strain rate, the stress softening decreases.

An in situ thermomechanical test confirms that hysteresis loss is mainly due to the SIC.

The experimental data will be used in near future for the microphysical modelling and the prediction of mechanical behaviour of NR compounds subject to monotonic and cyclic loading.

References

Harwood

J. A. C.

and Payne

A. R.

: ‘Stress softening in natural rubber vulcanizates’, Part V. The anomalous tensile behavior of natural rubber. J. Appl. Polym. Sci., 1967, 11, (10), 1825–1834.

Mandelkern

: ‘The role of elastomers in the study of polymer crystallization’, Rubber Chem. Technol., 1993, 66, (3), G61–G75.

Blanchard

A. F.

and Parkinson

: ‘Breakage of carbon-rubber networks by applied stress’, Ind. Eng. Chem., 1952, 44, (4), 799–812.

Hanson

D. E.

, Hawley

, Houlton

, Chitanvis

, Rae

, Orler

E. B.

and Wrobleski

D. A.

: ‘Stress softening experiments in silica-filled polydimethylsiloxane provide insight into a mechanism for the Mullins effect’, Polymer, 2005, 46, (24), 10989–10995.

Kraus

, Childers

C. W.

and Rollmann

K. W.

: ‘Stress softening in carbon black-reinforced vulcanizates’, Strain rate and temperature effects. J. Appl. Polym. Sci., 1966, 10, (2), 229–244.

Bueche

: ‘Molecular basis for the mullins effect’, J. Appl. Polym. Sci., 1960, 4, (10), 107–114.

Le Cam

J. B.

and Toussaint

: ‘Cyclic volume changes in rubber’, Mech. Mater., 2009, 41, (7), 898–901.

Lin

, Bian

, Yang

and Chen

: ‘Strain-induced crystallization of natural rubber as studied by high-resolution solid-state 13C NMR spectroscopy’, Polymer, 2004, 45, (14), 4939–4943.

Rault

, Marchal

, Judeinstein

and Albouy

P. A.

: ‘Chain orientation in natural rubber’, Part II: 2H-NMR study. Eur. Phys. J.E, 2006, 21, (3), 243–261.

10.

Toki

, Hsiao

B. S.

, Amnuaypornsri

and Sakdapipanich

: ‘New insights into the relationship between network structure and strain-induced crystallization in un-vulcanized and vulcanized natural rubber by synchrotron X-ray diffraction’, Polymer, 2009, 50, (9), 2142–2148.

11.

Siesler

H. W.

: ‘Rheo-optical Fourier transform infrared spectroscopy of polymers’, 10: strain-induced crystallization of sulfur-crosslinked natural rubber during cyclic deformation. Appl. Spectrosc., 1985, 39, (5), 761–765.

12.

Mukherjee

D. P.

: ‘Stress-strain and stress-birefringence studies on natural rubber, isomerized natural rubber, and synthetic poly(isoprene)’, Rubber Chem. Technol., 1974, 47, (5), 1234–1240.

13.

Toki

, Fujimaki

and Okuyama

: ‘Strain-induced crystallization of natural rubber as detected real-time by wide-angle X-ray diffraction technique’, Polymer, 2000, 41, (14), 5423–5429.

14.

Treloar

L. R. G.

: ‘The physics of rubber elasticity’, 3rd edn; 2005;. Oxford: Oxford University Press..

15.

Nakauchi

, Naito

, Kato

and Inoue

: ‘Analysis of cross-linking structure by compressive property of the swollen rubber’, Nippon Gomu Kyokaishi, 1987, 60, (5), 273–279.

16.

Yamada

, Okumoto

, Ohtani

and Tsuge

: ‘Characterization of vulcanized EPDM rubbers by high-resolution pyrolysis gas chromatography’, Rubber Chem. Technol., 1991, 64, (5), 708–713.

17.

Zaper

A. M.

and Koenig

J. L.

: ‘Solid state carbon-13 NMR studies of vulcanized elastomers’, II. Sulfur vulcanization of natural rubber. Rubber Chem. Technol., 1987, 60, (2), 252–277.

18.

Mitchell

G. R.

: ‘A wide-angle X-ray study of the development of molecular orientation in crosslinked natural rubber’, Polymer, 1984, 25, (11), 1562–1572.

19.

Albouy

P.-A.

, Guillier

, Petermann

, Vieyres

, Sanseau

and Sotta

: ‘A stroboscopic X-ray apparatus for the study of the kinetics of strain-induced crystallization in natural rubber’, Polymer, 2012, 53, (15), 3313–3324.

20.

Vieyres

, Pérez-Aparicio

, Albouy

P.-A.

, Sanseau

, Saalwächter

, Long

D. R.

and Sotta

: ‘Sulfur-cured natural rubber elastomer networks: correlating cross-link density, chain orientation, and mechanical response by combined techniques’, Macromolecules, 2013, 46, (3), 889–899.

21.

Leblanc

J. L.

: ‘Rubber-filler interactions and rheological properties in filled compounds’, Prog. Polym. Sci., 2002, 27, (4), 627–687.

22.

Akiba

and Hashim

A. S.

: ‘Vulcanization and crosslinking in elastomers’, Prog. Polym. Sci., 1997, 22, (3), 475–521.

23.

Amnuaypornsri

, Toki

, Hsiao

B. S.

and Sakdapipanich

: ‘The effects of endlinking network and entanglement to stress-strain relation and strain-induced crystallization of un-vulcanized and vulcanized natural rubber’, Polymer, 2012, 53, (15), 3325–3330.

24.

Zhang

, Schneider

, Liu

, Chen

, Brüning

, Wang

and Stamm

: ‘Structure variation of tensile-deformed amorphous poly(l-lactic acid): effects of deformation rate and strain’, Polymer, 2011, 52, (18), 4141–4149.