EFFECT OF SUBSTRATE ON ADHESION IN RUBBER MOULDING
D. KIROSKI, D. BURKE & D.E. PACKHAM
School of Materials Science,
University of Bath,
Bath, BA2 7AY, England.
ABSTRACT
Different alloys are associated with different levels of adhesion to nitrile rubber moulded against them. D.S.C. and mechanical property measurement on thin films show that the alloy affects the crosslink density of the surface regions of the rubber. Alloys giving higher crosslink density give higher adhesion.
Many of the objects around us, whether natural or artificial, comprise different substances joined or, in the broad sense of the word, bonded to one another. A major aim of the science of adhesion is to rationalise the phenomena associated with the adherence of these different substances. Some insight can often be obtained by reduction of the problem to an investigation of the properties of the individual phases in isolation and of the interface per se. There are, however, many phenomena in adhesion which can only be understood by treating the object as a whole, and recognising that the properties of one phase may be altered as a result of its being part of a composite whole in which it is bonded to other phases.
There are a number of interesting examples in the study of the adhesion of polymers to metals where the substrate exerts an influence on the properties of the polymer near the interface and, through this, on the strength of the adhesive bond. Transition metals are known to catalyse the autoxidation of many hydrocarbon polymers: this oxidation is often associated with enhanced adhesion[1]. Copper has been shown to affect the phase structure developed in toughened epoxy resins, and consequently to alter the strengths of adhesive bonds involving these materials[2,3]. The enhanced adhesion of natural rubber to brass is a consequence of the reactivity of the alloy to the sulphur in the rubber compound[4]. It is common, in such examples, for the bond to fail cohesively within the polymer, consequently a change of substrate can lead to different adhesion values, all associated with cohesive failure in the same polymer.
This paper addresses a particular example of this phenomenon observed during the moulding of nitrile rubber (NBR - acrylonitrile butadiene copolymer). Change of mould alloy can alter the adhesion of the same rubber compound. This alteration is technologically significant and scientifically interesting.
Mould adhesion of NBR
Nitrile butadiene rubber is produced by emulsion polymerisation[5]. The polymerised rubber, despite washing, is likely to contain residues from the polymerisation process, such as emulsifiers and coagulants. After compounding the NBR is moulded and heated to effect a cure. The adhesion between rubber and mould which develops at this stage is generally low, but varies according to the base NBR selected, the rubber formulation chosen and the nature of the mould alloy. During cure an interfacial layer forms consisting of residues from the polymerisation and derivatives of compounding ingredients, especially of fatty acids and zinc-sulphur complexes[6,7]. When the moulding is removed, parting takes place within this layer. Differences between adhesion of different base rubbers and of different rubber compounds can largely be understood in terms of the different residues which form part of this interlayer. Why can a change of mould alloy lead to different values of adhesion to an identical rubber compound? This is the question that this communication sets out to answer.
Mould adhesion and mould alloy
In working on mould adhesion of N.B.R., we have studied a large number of different formulations based on a large number of base rubbers. The same relative order of adhesion for different alloys has been observed with all these compounds, almost without exception. In particular EN8, a medium carbon tool steel shows higher adhesion than Stavax-420, a steel containing 13% chromium, v. Table 1. Typical values of mould sticking index might be 215 and 153kPa respectively [8].
It was first thought that this difference might be a consequence of a difference in the composition of the residues in the interfacial layer. Different surface energies, for example, might lead to different species being adsorbed on the alloy surface. The surfaces exposed on parting the rubber from the mould have been routinely examined by X.P.S. Different base rubbers and differently formulated compounds lead to consistent differences in this layer[6], but no consistent differences were observed associated with the different alloys.
Alloys and cure kinetics
It was considered possible that the presence of different alloys exerted an effect on the cure properties of the rubber, at any rate in layers adjacent to the interface, and that this would alter the adhesion. In order to study this possibility, rubber was mixed with finely divided samples of the releveant alloys and cured in a differential scanning calorimeter (D.S.C.). Brass is well known to adhere stringly to sulphur-cured rubbers, so alloy this was evaluated along with EN8 and Stavax-420. All alloys were found to accelerate the cure reaction, but brass had the strongest effect followed by EN8, Table 2[9]. In other words, the greater the acceleration of the cure associated with an alloy, the higher the adhesion during moulding.
Rubber adjacent to the interface
These results raised the question: What effect are the alloys having on the rubber? In attempting to answer this question it was necessary to measure properties of the region of the rubber formed close to the appropriate alloy surfaces. In most mouldings the vast bulk of the rubber cures remote from the mould surface, so thin layers of rubber were made by compression moulding in a heated press between platens of the chosen alloys. It was argued that if the samples were thin enough differences in properties in the surface regions would not be swamped out by bulk rubber. Significant differences could not be found in samples ca. 1mm thick moulded between different platens. However, when conditions were altered to produce rubber of ca. 0.25mm thickness differences were observed. The results discussed below were obtained with such samples.
Experimental details
The rubber formulation is shown in Table 3. The dynamic mechanical properties were measured at sub-ambient temperatures with a Rheovibron dynamic mechanical analyser (model DDV II) cooled with liquid nitrogen and operated at a frequency of 11Hz. Tensile tests on samples 50mm × 5mm were conducted at room temperature on an Instron model 1195 with a 500g load cell at a crosshead speed of 20mm/min. The gauge length was 20mm. A Wallace optical extensometer was used.
Discussion of results
The main loss peak of N.B.R., corresponding to the glass transition, is in the region -10 to -20 °C. The dynamic mechanical test results show considerable differences according to the mould alloy used, Table 4. Both the temperature and magnitude of the maximum in loss tangent are affected, showing that the alloy has a marked effect on the structure of the adjacent rubber. Alloys which accelerate the cure most (Table 2) give highest glass transition temperatures. This implies that the crosslink density is greatest for the rubber moulded against brass and least for that moulded against Stavax-420.
The tensile results were analysed according to the predictions of the statistical theory of rubber elasticity[10]:
f /A0 = NvkT (l - 1/l2) (i)
where f is the tensile force, A0 the unstrained cross sectional area, Nv the number of subchains per unit volume, k Boltzmann's constant, T absolute temperature and l the extension ratio. Thus by plotting f /A0 against (l - 1/l2) a straight line should be obtained from which Nv (which is proportional to the crosslink density) can be calculated. Of course, this only applies to gum rubbers, and the rubber used here was reinforced, Table 3. However is is considered that values of "Nv" calculated in this way can still be taken as proportional to the crosslink density, as the formulation and cure conditions were kept constant.
The values of Nv given in Table 4 are averages based on measurements of nine gradients. They show the same trend as the previous results: brass produces rubber with the highest density of crosslinks, then EN8, then Stavax-420. The scatter in these Nv values is indicated in the confidence limits quoted. Taken together with the other results, the trend in Nv is considered meaningful.
Conclusions
These results give a clear indication that the alloy affects the rate of cure and crosslink density of the rubber formed close to its surface. The higher mould adhesion associated with EN8 compared with Stavax-420 is associated with higher crosslink density in interfacial regions of the rubber. This will give a locally higher modulus (cf. equation (i)) and probably higher work of fracture. These will give a higher failure stress in an adhesion measurement[11].
References
1. D.E. Packham in Developments in Adhesives-2, (ed. A.J. Kinloch), Applied Science 1981, p. 315.
2. P J Hine, S El Muddarris and D.E. Packham. J. Adhesion Sci. Technol. 1, 69 (1987).
3.F.J. Boerio and D.J. Ondrus, J.Colloid Interf. Sci. 138, 446(1990).
4.W.J. van Ooij, Rubber Chem. Technol. 51, 52(1987).
5. P.W. Milner, in Developments in rubber technology - 4, (ed. A. Whelan and K.S. Lee), Elsevier Applied Science 1987, p. 57.
6. M. Lotfipour, D.E. Packham & D.M. Turner, Surface & Interf. Anal. 17, 516(1991).
7. L.A. Reeves and D.E. Packham, J.Physics D Appl. Physics 25, A14(1992).
8. R.K.Champaneria, B. Harris, M. Lotfipour and D.E. Packham, Plastics Rubber Process. Appln. 8, 185 (1987).
9. L.A. Reeves, M. Lotfipour, D. Kiroski and D.E. Packham, Kautschuk & Gummi Kunststoffe 45, 369(1992).
10. L.R.G. Treloar, The Physics of Rubber Elasticity, 3rd edn., Oxford University Press, 1975
11. R.J. Good, J. Adhesion 4, 133(1972).
TABLE 1
Composition of the mould alloys
(manufacturers' specifications: wt %, balance Fe)
C Mn Cr Si V S P
S-420 0.38 0.5 13.6 0.8 0.3 - -
EN8 0.4 0.6 - 0.05 to - 0.06 0.06
(080M40) to 0.35
1.0
TABLE 2
D.S.C. results showing the effect of adding various alloys on the time taken for the cure rate on N.B.R. compounds to reach a maximum[9]
Alloy Time (min.) None 9.92 S-420 9.60 EN8 9.28 Brass 8.22
TABLE 3
N.B.R. formulation (parts by weight)
Polymer§ 100
GPF black 30
ZnO 2.5
Stearic Acid 1.0
Santocure NS* 0.7
MC Sulphur 1.5
§ Krynac S3422
* N-cyclohexylbenzothiazyl sulphenamide
TABLE 4
Properties of N.B.R. compound cured adjacent to alloy indicated.
The number of sub-chains was calculated* from Nv , as described in the text
S-420 EN8 Brass
Max of tan d: temp. -14°C -10°C -9°C
Max of tan d: value 0.59 1.68 1.40
No. sub-chains 2.50 2.67 3.09
90% confidence ±0.28 ±0.13 ±0.38
limits
* values shown are ×1017 mm-3
______________________________________________________________________________
Click here to return to David Packham's home page
Revised 12.ix.96