Return to David Packham's home page

A solid state chemical reaction occurs when a solvent cast film of a blend of masticated natural rubber and chlorinated natural rubber is heated in the presence of air at 150^oC. The thermal behaviour of solvent cast films of chlorinated natural rubber, masticated natural rubber and a 1:1w/w blend (2%w/v in xylene) of these two polymers has been studied using differential scanning calorimetry, infrared spectroscopy, scanning electron microscopy and nuclear magnetic spectroscopy. The results suggest that carbonyl groups are incorporated into the blend on heating and that the vinyl functionality of the isoprene units is modified during this apparent oxidation. Heating for 2 hours at 150^oC results in a material that no longer contains the rubber-like cis-1,4-polyisoprene units.

Keywords: Natural Rubber, Chlorinated Rubber, Adhesion.

The bonding of rubber to steel is widely used to produce vibration management systems such as bridge bearings and motor car suspension units¹. Rubber to metal bonding is usually achieved by injection moulding rubber onto precoated steel at 150^oC to 170^oC. The manufacture of rubber to metal bonded components is technologically well understood however, the scientific basis of rubber to metal adhesion is not clear. This is mainly due the fact that the adhesive systems work so effectively which inhibits access to the interfaces which are formed during rubber to metal bonding. In addition, another obstacle to gaining an understanding of the bonding processes is the use of proprietary primers and bonding agents. There are reports of the analysis of failure surfaces produced by electrochemical (cathodic) treatment² and destructive adhesion tests³. However, these studies do not directly address the actual nature of the interactions occurring during the actual bonding process.

A typical rubber to metal bonded component usually consists of a treated steel substrate, a primer, an adhesive, and rubber. One aspect of the overall bonding process is the interaction between a compounded rubber and the adhesive, in this case Chemlok 220, during injection moulding. The compounded rubber and adhesive are complex formulations, however, our own analysis has confirmed literature² reports that a significant component of the adhesive is chlorinated rubber and the compounded rubber, cis-1,4-polyisoprene. Because of the complexity of these proprietary systems we have undertaken a study of relatively simple model systems in order to try to understand the nature of the interactions that occur during rubber to metal bonding. We have therefore studied the interaction between masticated natural rubber (MNR) and chlorinated natural rubber (CNR) which serves as a model to complement our more direct work on rubber to metal adhesion.

We have reported our preliminary results for the behaviour of these polymers at 170^oC⁴. We now report results of a more detailed study of the interactions of CNR and MNR at 150^oC.

Thin films of the polymers were prepared by casting solutions of MNR (standard Malaysian rubber, grade SMRL), CNR (Pergut S20, Bayer, 59%w/w Cl, M_n 3.9x10⁴ g mol^-1, polydispersity, 1.8), and a 1:1 w/w mixture (2%w/v in xylene) onto glass, allowing the solvent to evaporate, and drying in vaccuo. In order to simulate the vulcanisation process, some of the films were heated at 150°C for up to 2 hours.

Electron microscopy, supported by electron probe microanalysis, was performed using a Joel 6310 equipped with an X-ray analyser. Transmission infrared spectra were recorded in KBr matrices using a Perkin Elmer 1710 infrared spectrometer. Differential scanning calorimetry (DSC) was studied using a TA Instruments 2910 DSC. Solutions of the samples were placed in DSC aluminium pans and the solvent allowed to evaporate. The samples were finally dried in vaccuo. prior to analysis. Isothermal behaviour was investigated in open pans at 150^oC in air and nitrogen (flow rate, 25cm³minute^-1), and glass transition temperatures were measured in sealed pans at a variety of thermal cycles (see text). Solid state ¹³C NMR spectra were recorded at 75MHz (cross polarisation) at the University of Durham. All chemical shifts are relative to tetramethylsilane (TMS).

The polymer samples, both before and after heating, were studied using scanning electron microscopy with electron probe microanalysis, infrared spectroscopy, differential scanning calorimetry and nuclear magnetic resonance spectroscopy.

The materials were studied using scanning electron microscopy (SEM) together with elemental mapping. The surfaces of the homopolymer films were homogenous in appearance and elemental distribution both before and after heating (150^oC/2 hours). However, films of the blend of the two homopolymers exhibited a two phase morphology. The micrographs are shown in Figures 1 and 2. Elemental mapping of the surface indicated that the sphere-like structures contained a chlorine rich material. Heating at 150^oC for 2 hours caused the material to become homogeneous in both morphology and Cl distribution. This behavior is similar to that reported for similar mixtures heated at 170^oC⁴.

Solvent cast thin films of the samples were heated on glass at 150^oC for 2 hours in air. The resulting materials were dispersed in KBr matrices and the transmission infrared spectra recorded. After heating the blend was distinctly orange in appearance which suggests the presence of chromophores arising from conjugation or charge transfer. The infrared spectra of the samples heated for 2 hours at 150^oC are shown in Figure 3. After heating for 2 hours at 150^oC the transmission spectra of the component homopolymers were essentially unchanged and contained no significant absorption bands due to C=O. The infrared spectrum of heated MNR exhibited absorption bands at 837cm^-1, assigned to C=CH deformation, 1480cm^-1, CH₃ and CH₂ deformations, and 1376cm^-1, CH₃ deformation⁵. The infrared spectrum of heated CNR exhibited absorption bands at ca. 650-800cm^-1, the maximum absorbance being at 736cm^-1, assigned to C-Cl stretch, 1385cm^-1, CH₃ deformation, and, 1275cm^-1, assigned to an activated skeletal vibration⁶. The spectrum of the unheated blend was essentially a sum of the spectra of the two component homopolymers. This suggests that oxidation of these polymers to form C=O groups does not occur to any great extent under these conditions. However, the spectrum of the 1:1w/w blend which had been heated for 2 hours at 150^oC in air exhibited an absorption band at 1720cm^-1, which is assigned to C=O functionalities. The absorption band at 737cm^-1, due to C-Cl stretch, is still present whereas there is no evidence of the absorption band due to C=CH deformation at 837cm^-1. There are absorption bands present at 1443cm^-1 and 1090cm^-1 which are presumably associated with carbon/oxygen functionalities, possibly esters.

The infrared spectroscopic study of films of CNR, MNR and a 1:1w/w blend of these two polymers suggests that the homopolymers are relatively stable to oxidation at 150^oC whereas the blend undergoes some form of oxidation. The absence of an absorption band at 837cm^-1 in the heated blend suggest that the vinyl functionality is undergoing chemical modification.

Nuclear magnetic resonance (NMR) spectroscopy was employed to try to identify the multiplicity of the carbon groups and chemical species present in the oxidised blend. Solution state (in CDCl₃) was initially employed however not all of the products from the oxidation studies were soluble in this solvent. Consequently, solid state ¹³C NMR spectroscopy was used to ensure that all of the bulk samples were analysed.

The solid state ¹³C NMR spectra of MNR and CNR suggested that these two materials on their own were relatively inert to oxidation after heating at 150^oC for 2 hours. The spectra of the unheated and heated blend are shown in Figure 4. In Figure 5 the spectra of heated CNR and the heated blend are compared. It can be seen from Figure 4 that the solid state spectrum of the unheated blend that two distinct types of resonances are observed, i.e. sharp and broad. This suggests the presence of a two component system, a rubber phase and a rigid phase. The sharp peaks are due to polyisoprene at 135.1ppm (C(Me)), 125.6ppm (CH), 32.7ppm (CH₂), 27.0ppm (CH₂) and 23.9ppm (CH₃). These values are very similar to those obtained from our solution state ¹³C NMR data. The broader resonances arise from the CNR. The resonances were assigned, with references to our own solution and solid state ¹³C NMR data, to 76ppm (CR₃-Cl and CR₂Cl₂), 62ppm (CHCl), 44ppm and 39ppm (CH₂) and 21ppm (CH₃). The most striking feature of spectrum of the heated blend is that heating has caused the disappearance of the sharp resonances due to the polyisoprene component and the spectrum now more closely resembles that of the heated CNR (see Figure 5). This suggests that a one phase, i.e. rigid, material is now present. However there are broad resonances which can be assigned to polyisoprene at 134.8ppm and 125.7ppm (the vinyl functionality), and in the aliphatic region between 20-35ppm. Additionally, there are resonances above 140ppm which are probably due to C/O functionalities. The resonance at 171ppm can be assigned to C=O.

The ¹³C NMR spectra suggest that when MNR and CNR are heated together in the presence of air a material is produced which is rigid and no longer two phase, i.e., the material does not contain a rubber like component. Solid state ¹³C NMR spectroscopy suggests that C=O functionalities are present in the heated blend.

Samples of the blend and the component homopolymers were studied using differential scanning calorimetry (DSC) isothermally at 150^oC in air and nitrogen in order to elucidate the thermal behaviour if the two polymers at the moulding temperature. The miscibility of the blend was investigated by studying the positions of the glass transition temperatures of the system.

The thermal behavior of the blend and the component homopolymers was investigated in air and nitrogen at an isothermal temperature of 150^oC. In general, no significant thermal events were detected when the samples were heated for up to 60 minutes at 150^oC in flowing nitrogen. The situation for heating in flowing air is very different. The isothermal behaviour of the blend and the component homopolymers is shown in Figure 6. It can be seen from Figure 6 that solvent cast films of CNR and MNR are relativity inert at this temperature whereas the blend exhibits an exothermic thermal event.

These results suggests that a solid state exothermic reaction occurs when solvent cast blends of MNR and CNR are heated in the presence of air.

The miscibility of polymer systems can be studied using DSC. In general, a two component polymer blend system will exhibit two T_g’s if they are immiscible, one T_g if they are completely miscible and three T_g’s if partially miscible. In a miscible system the position of the T_g will be dependent on the composition of the blend.

Samples of solvent cast masticated rubber were dried and sealed in aluminium pans and cooled (quenched) to -100^oC, held at this temperature for 5 minutes, and heated at a variety of rates to 150^oC in flowing nitrogen. Typically, heating rates of 20^oC min^-1 gave a T_g of the MNR of ca. -60^oC. As expected, the transition occurred at somewhat higher temperatures when the heating rate was increased. The behaviour of the solvent cast CNR was then investigated. No reproducible thermal events which could be assigned to a T_g were found. The thermal behaviour of CNR was found to be irreproducible. However, an endothermic thermal event could be detected (on heating) at ca. 60^oC, see Figure 7. This thermal event was not observed when the sample was exposed to a second identical thermal study. This probably reflects the complex irregular structure of CNR .

The thermal behaviour of MNR/CNR blends was compared to that of the single component homopolymers in air. The DSC thermograms are shown in Figure 7. On initial heating of the blends from -100^oC the thermal event ascribed to the T_g of the MNR is observed (this transition shifted to slightly higher temperatures as the proportion of CNR rubber in the blend was increased). No thermal event was detected in the 60^oC region, in contrast to the behaviour of the CNR rubber, but a broad exothermic thermal event was detected from 150^oC to 220^oC.

This behaviour was compared to that of films of MNR/CNR blends which had already been subjected to heating in air at 150^oC for 2 hours. The DSC thermograms are shown in Figure 8. The heated material exhibited no thermal event due to the T_g of the MNR and a exothermic thermal event at 150^oC to 220^oC.

The DSC study suggests that blends of MNR and CNR are initially immiscible but that heating in air causes chemical modification to occur. The disappearance of the polyisoprene T_g suggests that the material no longer contains rubber-like polyisoprene units.

Results from electron microscopy, thermal analysis and infrared and nuclear magnetic resonance spectroscopy are consistent in supporting the following conclusions. A 1:1w/w mixture of MNR and CNR cast from xylene is essentially immiscible. However heating at 150^oC results in a dramatic change in morphology. There appears to be some type of accelerated chemical oxidation in the blend which does not occur in the component homopolymers. The resulting heated material no longer contains rubber-like polyisoprene units. The implications of this work for a possible mechanism for the adhesion of natural rubber to CNR containing bonding agents are undergoing further investigation

We thank EPSRC and Avon Rubber for financial support and David Apperley at the University of Durham for recording the solid state ¹³C NMR spectra. Chlorinated rubber, Pergut S20, was kindly provided by Bayer.

1. G. J. Lake in Handbook of Adhesion edited by D.E. Packham, Longman, 1992, p393.

2. F. J. Boerio, S. J. Hudak, M. A. Miller and S. G. Hong, J. Adhesion, 23, 99 (1987).

3. J. Kurian, G. B. Nando and S. K. De, J. Adhesion, 20, 293 (1987).

4. J. W. Cook, S. Edge and D. E. Packham, J. Mater. Sci. Lett. 16, 445(1997).

5. F. Scholl and D. O. Hummel in Atlas of Polymer and Plastic Analysis, Hanser Verlag,

1988, Vol II, Part a/I, p343.

6. ibid., p353.

7. M. V. Eskina, A. S. Khachaturov, L. B. Krentsel and A. D. Litmanovich,

Eur. Polym. J., 26, 181 (1990).