. nat. Rubb. Res., 7(1), 1-13
Ozone Cracking and Protection of Elastomers at High and Low Temperatures
G.J. LAKE* AND P.O. MENTE*
Ozone attack on elastomers varying in their glass transition temperatures (8G) by some 50° C was studied over a range of temperatures from about ~20°C to + 70°C. If the ambient temperature is low, the rate of crack growth due to ozone may be reduced by the high internal viscosity of the material. However, diffusion of chemical antiozonants or waxes which confer protection by forming a layer on the elastomer surface, is also slowed down. It appears more difficult to obtain good protection at low temperatures. High ambient temperatures do not greatly increase the rate of ozone crack growth and might be expected to improve antiozonant and wax protection for elastomers of high 9G as a result of increased rates of diffusion. This was observed to some extent. Ozone can cause cracking in vulcanised and this paper describes work that is in elastomers subject to constant or repeated progress to try to remedy this. The protec- loading. The ozone concentration in the tive systems examined include chemical outdoor atmosphere at ground level is antiozonants or waxes alone and wax/ typically about 1 p.p.h.m. (part per hundred antiozonant combinations, used in elasto- million of air by volume) although levels mers of various glass transition temperatures an order of magnitude or so higher in tests over a wide range of ozone have been observed for limited periods in concentrations. polluted areas. Indoors, or in enclosed spaces, ozone levels may be an order of magnitude lower unless there are electrical THEORIES FOR LAYER FORMATION or other discharges, when much higher levels may result. The recommended ambient Chemical Antiozonants temperature for standard ozone tests is now 40°C. This is far above operating When a chemical antiozonant is present temperatures for many uses of elastomers, in an elastomer, layer formation occurs only although it may be too low, intermittently, on exposure to ozone and is believed to be for articles subject to direct sunlight, primarily the result of direct reaction particularly in tropical regions. between the antiozonant and ozone. In earlier work it was shown for natural The effects of ozone concentration and rubber that under some circumstances, the ambient temperature on the basic ozone rate of layer formation was consistent with cracking process are well understood, as is control by diffusion of antiozonant across the effect of the glass transition temperature the already-formed layer to react with ozone of the elastomer-2'3. However, when protec- at its outer surface4 (Figure la). Under these tive agents - chemical antiozonants or waxes circumstances, assuming the layer to be - are present, these effects are much less well formed entirely of reacted antiozonant and understood. High or low temperatures, in the concentration of unreacted material particular, have been relatively little studied within it to be small, the mass per unit area
*Malaysian Rubber Producers1 Research Association, Brickendonbury, Hertford SO 13 8NL, United Kingdom
Journal of Natural Rubber Research, Volume 7(1), 1992
a) Concentration
Rubber Layer
Distance Rubber Layer surface surface
b) Concentratio n
Laye r
S C Rubber L S
Distance Rubber Layer surface surface Figure 1 Diffusion models for antiozonant layer formation a) rate governed by diffusion of antiozonant within the layer only, b) rate governed by both diffusion in the layer and in the rubber
ML, of layer formed after time / is given by Consistent with Equation 1, the rate of layer formation on natural rubber was observed M = 1 to be independent of ozone concentration L over a range from about 25 p.p.h m to where C is the concentration of antiozonant 0 750 p p h m although it increased at higher in the bulk of the rubber concentrations
P the density of the layer material L In deriving Equation 1, it is assumed that £,. the diffusion coefficient the concentration of (unreacted) antio- zonant in the layer is small Under these SL the solubility, relative to that m the rubber, of the antiozonant in the circumstances the concentration gradient layer in the layer will be essentially linear (cf Figure la) If, m addition to diffusion across If the product of the latter two quantities is the layer, diffusion within the rubber is also put as important, a similar equation to Equation I
applies with Co replaced Cf the concentra- PL = SLDL 2 tion at the rubber surface, which is lower PL is in essence the relative permeability of than Ca (Figure Ib) Cs is a function of C0 the layer material to the antiozonant and the relative permeabilities of the rubber COPYRIGHT © MALAYSIAN RUBBER BOARD G.J. Lake and P.O. Mente: Ozone Cracking and Protection of Elastomers at High and Low Temperatures and the layer to the antiozonant, and with G the shear modulus of the rubber the assumptions made is given by 9 the absolute temperature
e - ... 3 F is a function involving the wax dis- tribution and other parameters. co where Thus the rate of protective layer for- mation with a wax is not influenced by the ...4 ozone concentration but it is generally sensitive to temperature. This is not so 4 DC, much because 6 appears in Equation 5 (this small effect would in fact be balanced by the and D is the diffusion coefficient of the dependence of G on 9 according to the antiozonant in the rubber. statistical theory of rubber elasticity) but because of the variation of Dw and SH with Waxes temperature. In practice, most waxes used The mechanism of layer formation by are blends of components having different waxes is quite different from that by molecular weights and melting points, often chemical antiozonants. Waxes are relatively over quite a wide range. (Apart from inert and blooming occurs when the amount providing more consistent blooming be- of wax incorporated in the rubber exceeds haviour over a range of temperatures, the the solubility. This is possible because bloom obtained by this means provides the melting points of the waxes used are better protection, as pure waxes tend to give large crystals that form a very uneven generally within the range from about 45°C 6 to 70°C. Thus at processing temperatures, bloom .) which are normally well above the upper limit, the waxes are liquid, with relatively EXPERIMENTAL PROCEDURE high solubility, whereas at operating Ozone tests were carried out using a com- temperatures they are crystalline solids with 7 much lower solubility. Blooming is not mercially-available test cabinet for concen- simply due to super-saturation, however, trations within the range 50 - 200 p.p.h.m. but is believed to involve stress-induced while an apparatus similar to that described by Farlie1 was used for concentrations of diffusion due to the elastic forces acting 4 5 on crystals formed inside the rubber. about 10 to 10 p.p.h.m. In this apparatus, Quantitative solution for the rate of layer ozone was generated from cylinder oxygen using an ultra-violet discharge. Intermediate formation involves, although it is not 2 3 sensitive to, the distribution of crystals in concentrations from about 10 to 10 the rubber. The dependence of the mass of p.p.h.m. were obtained by diluting the ozonised oxygen from the 'high' concentra- wax layer per unit area of surface (Mn) on time (r) can be approximated by tion apparatus with a suitable flow of air. Tests in the commercial cabinet were carried F out at various temperatures with a range M = from about 8°C to 70°C. Tests in the laboratory apparatus were carried out at about 23°C or - 17°C. where Dw is the diffusion coefficient Ozone cracking tests on surfaces were Sn the solubility of the wax in the rubber carried out using strip or dumbbell test- pieces held at an essentially constant strain CH the concentration of wax incorpo- which was normally within the range 10% to rated, 100%. The test-pieces were examined inter- Journal of Natural Rubber Research, Volume 7(1), 1992 mittently during the course of exposure to out with di(methylethylpentyl) paraphenyle- ozone to determine whether or not cracking nediamine (DOPPD - 'UOP88', Universal had occurred. In some cases, an estimate of Oil Products, Illinois) which was used at the rate of crack growth was made by levels from 2 p.h.r. to 20 p.h.r. (parts measuring the crack length (in a direction per hundred of rubber by weight). Various perpendicular to a major surface) at the end wax/antiozonant combinations were also of a test. employed, generally with 6 p.h.r. wax and 3 p.h.r. antiozonant. Measurements of the rate of growth of a single crack by ozone were made using a Diffusion coefficients of the antiozonant test-piece in the form of a strip containing an were estimated from transfer measurements initiating edge cut, as described by Braden in which a test-piece containing DOPPD and Gent1. Protection, by a coating of inert was pressed into contact with a test-piece grease or other means, was used to prevent of a similar vulcanisate containing, initially, formation of subsidiary cracks. no DOPPD (Figure 2). The amount of antiozonant that had diffused after various The elastomers used were natural rubber times was estimated from the weight changes (NR, glass transition temperature, 6G ca of the test-pieces, enabling the diffusion — 67°C), epoxidised natural rubber with coefficient to be calculated according to about 50% epoxidation (ENR 50, 9G ca standard diffusion theory8. — 21 °C) and two acrylonitrile-butadiene copolymers (34% acrylonitrile, NBR 34, The weight of reacted antiozonant layer 9G ca -2TC, or 38% acrylonitrile, NBR 38, formed after exposure to ozone was deter- BG ca — 23°C). Accelerated sulphur-vulcani- mined from the weight change after removal sates of each elastomer were prepared in of the layer from the surface of the rubber. the form of 2 mm thick sheets from which The layer could be removed by scraping with test-pieces were die-stamped. The sheets a blade, swabbing with tissue containing a were stored at ca 0°C to minimise ageing suitable solvent or repeated application of and blooming effects. When a chemical adhesive tape. The latter method was mostly antiozonant alone was incorporated, nearly used for the chemical antiozonant layer, all the experiments described were carried whereas scraping was found to be best for
Direct Across the layer
Layer
Figure 2. Arrangements for transfer measurements: direct; across the layer.
4 G.J. Lake and P.O. Mente: Ozone Cracking and Protection of Elastomers at High and Low Temperatures wax layers. In experiments with waxes, the various amounts of DOPPD at room elastomer sheets were vulcanised between temperature for high ozone concentrations aluminium foil or polyester sheets to mini- of about 104 p.p.h.m. or 105 p.p.h.m. The mise blooming during storage. Test-pieces rates of growth in the unprotected NR, for bloom formation measurements were ENR 50 and NBR 34 vulcanisates are very kept at a constant temperature of either similar and are reduced to a similar extent 20°C or 40°C for a fixed period. The weight by the incorporation of DOPPD. Consistent of the bloom was then determined from the with earlier observations1'9, NBR 38 shows weight change after its removal. somewhat lower rates of growth for the unprotected vulcanisate but rather smaller Preliminary experiments to estimate the reductions when DOPPD is added. Bearing permeability of the antiozonant layer ma- in mind the variability, the results for terial to the (unreacted) antiozonant have both protected and unprotected vulcanisates been carried out by forming a layer on the are broadly consistent with a rate that is surface of a test-piece by exposure to ozone essentially proportional to the ozone con- and then carrying out a transfer measure- centration. Observations at much lower ment through the layer to a control test- ozone concentrations and different tempe- piece pressed into contact (Figure 2). In this ratures, made during the course of threshold case, the permeability was estimated by strain experiments, also appear consistent assuming the layer to act as a membrane . with this. Only a single time could be used in each experiment as the layer was disrupted when Threshold Effects the test-pieces were separated. Certain protective agents have the ability effectively to increase the threshold energy RESULTS or threshold strain. By contrast with the Earlier work has indicated that two basic reduction in rate of growth, these increases parameters govern the growth of a single, tend to be strongly influenced by ozone isolated crack by ozone: i) a threshold concentration. Figure 4 shows results for energy below which the crack will not NBR 34 vulcanisates containing 0, 5 or grow; and, ii) the rate at which the crack 10 p.h.r. DOPPD, at ozone concentrations grows when the threshold is equalled or ranging from 100 p.p.h.m. to 1500 p.p.h.m. exceeded1. For cracking in surfaces, a The exposure strain is plotted against the threshold strain corresponds to the attain- time at which cracks were first observed (T,), ment of the threshold energy at the largest on logarithmic scales. In order to allow flaws that are present in the surface. In the results for different ozone concentrations present work, the single crack growth rate readily to be compared, T, is multiplied by and the threshold strain are being used to the ozone concentration (this will super- assess the ozone resistance of the various impose results where the rate of crack elastomers and protective systems studied. growth is proportional to the ozone con- centration and T, is determined mainly by Single Crack Growth Studies the rate). As can be seen this superimposi- tion works quite well for the results for the Earlier work has shown that the rate nitrile vulcanisates in Figure 4 bearing in of ozone crack growth is essentially mind particularly that considerable scatter proportional to the ozone concentration, in the results at shorter times arises because provided that the test temperature is not the test-pieces were inspected only intermitt- too close to the glass transition temperature ently). The displacement of the curves for of the elastomer '2'3'9. Figure 3 shows single different DOPPD levels is broadly consistent crack results for different elastomers with- with the differences in rate of growth out added protective agent and including observed in Figure 3. These results show 3 Natural rubber * ENR 50 *• NBR 34 * NBR 38
5 0.1 - ca 10 p.p.h.m.
0.01
— - * *'X--^ ca 10. „"! p.p.h.m , .
0.001 10 15 20 DOPPD (p.h.r.) Figure 3. Rate of single crack growth (logarithmic scale) versus antiozonant level at the ozone concentrations indicated on the graph for vulcanisates of natural rubber, ENR 50, NBR 34 and NBR 38. Each point represents an average of about three results.
NBR 34 vulcanisates
1 000 r o Unprotected « + 5 p.h.r. DOPPD • + 10 p.h.r. DOPPD Ozone concentration 100 10 102 103 104 i,x[03 p.p.h.m. x days Figure 4. Time to first cracking (ij for NBR 34 vulcanisates. Each point carries a flag to denote the ozone concentration used. GJ. Lake and P.O. Mente: Ozone Cracking and Protection of Elastomers at High and Low Temperatures little or no increase in the threshold strain - crystallisation10. With 5 p.h.r. DOPPD, no even with 10 p.h.r. DOPPD the indicated cracking occurred in natural rubber at all threshold strain is no more than 10% which strains tested even at an ozone concentration is only slightly above the level expected for of 1500 p.p.h.m. the unprotected vuicanisate. With ENR 50, similar behaviour was By contrast, the results for natural rubber observed to that with NBR 34. With in Figure 5 show evidence of appreciable DOPPD at the 5 p.h.r. level, no significant increase in threshold strain even with change in threshold strain was detected, as only 2.5 p.h.r. DOPPD. The effect is was the case with 10 p.h.r. at an ozone dependent on ozone concentration and is concentration of 1500 p.p.h.m., although a not significant at 1500 p.p.h.m. However, at slight increase was indicated at 100 p.p.h.m. 100 p.p.h.m. the results show no cracking at (Figure 6). Even with 20 p.h.r. DOPPD, a strain of 20% after 1000 p.p.h.m. days - only a slight increase in threshold was once a time of this order has been survived, obtained at 1500 p.p.h.m. although cracking the likelihood that cracking will occur was prevented at all strains at 100 p.p.h.m. subsequently is greatly reduced. Thus a Antiozonant Diffusion threshold strain in excess of 20% is indicated. An interesting feature of these The above results show that while results is that whereas cracking occurred at DOPPD can confer a similar reduction in 100% strain, no cracks were observed after a rate of growth in nitrile or epoxidised natural much longer time at 150%, indicating a rubber to that in natural rubber, its ability to second threshold strain (above which no increase the threshold condition is much cracking occurs) in this vicinity. This second less. Since the latter ability is known to be threshold is believed to be associated with associated with the formation of a protective 1 000 No cracking 100 100 p.p.h.m. 10 [O3] = 1500 p.p.h.m. 10 102 103 104 T, x [O3] p.p.h.m. x days Figure 5. Time to first cracking (ij for an NR vuicanisate containing 2.5 p.h.r. DOPPD at the ozone concentrations indicated on the graph. 1 Journal of Natural Rubber Research, Volume 7(1), 1992 1 000 r No cracking lOOp.p.h.m. 100 - c _o c W 10 [O 10 103 103 104 T. X [OJ p.p.h.m. x days Figure 6. Time to first cracking for ENR 50 vulcanisates containing the levels of antiozonant and at the ozone concentrations indicated on the graph. layer on the surface of the elastomer by the experiments), the diffusion coefficient of antiozonant following reaction with ozone4, DOPPD in NBR 34 or ENR 50 is only which must involve diffusion, it is tempting about 1% of that in NR. At elevated to conclude that the poorer protection temperatures, the diffusion coefficient of of NBR and ENR is associated with slower DOPPD increases for all the elastomers, diffusion of the antiozonant in these although the increase is greater for the elastomers, which would be expected high glass transition temperature materials because of their higher internal viscosities. (Table 1). Of interest is 40°C since it is the Studies of the diffusion of DOPPD in the recommended temperature for standard different elastomers were made by transfer ozone tests (although of limited relevance measurements, as described earlier. Results for practical purposes), while at 70°C the are shown in Table I. At 23°C (the diffusion coefficient in ENR 50 or NBR 34 temperature of the earlier ozone cracking is similar to that in NR at 23°C. TABLE 1. EFFECT OF TEMPERATURE ON DIFFUSION COEFFICIENTS (10 p.h.r. DOPPD) 2 Temperature Diffusion coefficient (cm /s) in a vulcanisate of NR ENR 50 NBR 34 C ca -69°C) (0Q ca -24°C) (0G ca -28 C) -17 6.0x10 " ca 10"" 23 4.6 xKT9 5.1 xlO" 4.9 x 10" 40 2.4xlO~8 6.2x10" 7.2x10" 70 UxKT7 8.8 x 10" 9.7x10" G.J. Lake and P.O. Mente: Ozone Cracking and Protection of Elastomers at High and Low Temperatures Ozone Cracking Experiments at Elevated contrast, only the antiozonant/lower melting Temperatures point wax combinations gave significant protection at 20°C, whereas most of the Results of ozone cracking tests at 40°C systems gave considerably better protection are shown in Table 2. With 5 p.h.r. or at 40°C. 10 p.h.r. DOPPD, the threshold strain for vulcanisates of ENR 50 or NBR 34 is higher With waxes, temperature affects the at 40°C than at 20°C but is still well below blooming rate through both the diffusion that for NR with 5 p.h.r. DOPPD at 20°C. coefficient and the solubility in the rubber, However, the diffusion coefficients, al- which increases rapidly as the melting point though increased by more than an order of of the wax is approached. These parameters magnitude at 40°C, are appreciably lower also enter directly into the blooming than that for NR at 20°C (Table 1). Initial equation (cf Equation 5) whereas in the case cracking test results at 70°C, (where the of DOPPD, the diffusion coefficient of diffusion coefficients are slightly higher than the antiozonant in the rubber only enters for NR at 20°C) indicate the protection of somewhat indirectly into the equation for ENR 50 or NBR 34 still to be inferior to layer formation (Equation 1) via Equations 3 that of NR at 20°C, although various and 4. Thus a rather stronger effect of difficulties have been encountered with temperature on protective action might be these tests and further experiments are in expected for protective systems involving progress. waxes compared with those involving a chemical antiozonant alone. The results in Rather similar behaviour was observed Table 2 appear consistent with this. with wax or wax/antiozonant protective systems for which some results are also Cracking Experiments at Low Temperatures shown in Table 2. For NR, all of the protective systems gave threshold strains in Experiments at 8°C and 50 p.p.h.m. excess of 100% except for the paraffin wax ozone showed little or no protective action alone at 40*C. For ENR 50 or NBR 34, by of a wax/antiozonant system in NBR 34 or TABLE 2. COMPARISON OF THRESHOLD STRAINS AT 20°C AND 4CTC Threshold strain (%) for an ozone concentration of 50 p.p. i.m. at the test temperature indicated for vulcanisates of Protective system NR ENR 50 NBR 34 20LC 40°C 203C 40°C 20CC 40°C 5p.h r DOPPD >100 >100 < 10 ca 15 < 10 ca 15 lOp.h r. DOPPD - - ca 15 ca 50 ca 50 >100 3p.h r 6PPD + 6 p.h.r. Wax A >100 >100 ca 10 ca 30 ca 10 ca 50 3p.h r IPPD + 6 p.h.r. Wax B >100 >100 ca 10 >100 ca 30 ca 100 3p.h r IPPD + 6 p.h.r. Wax C - - ca 50 >100 - 6 p.h r WaxC >100 ca 10 < 10 ca 10 ca 10 ca 10 6PPD is a di(methyl butyl) phenyl substituted ppd (Santoflex 13) IPPD is an isopropyl phenyl substituted ppd (Nonox ZA) Wax A was Antilux 600 (quoted melting point range 62°C to 65°C) Wax B was Antilux 500 (quoted MP range 56nC to 60'C) Wax C was a paraffin wax (MP range 48°C to 50CC) + Failure sometimes occurred by mechanical crack growth in NBR 34 at 40CC Journal of Natural Rubber Research, Volume 7(1), 1992 TABLE 3. THRESHOLD STRAINS AT 8^ Threshold strain (%) for an ozone concentration of Protective system 50 p.p.h.m. at 8°C for vulcamsates of NR ENR50 NBR 34 lOp.h.r. DOPPD > 100% ca 50% ca 50% 3 p.h.r. IPPD + 6 Wax B > 100% - - 3 p.h.r. IPPD + 6 Wax C > 100% ca 10% ca 10% Waxes B and C are as in Table 2. ENR 50 even with the lowest melting point materials (Table 2). Similar layer formation wax (Table 3). With 10 p.h.r. DOPPD, behaviour with a wax/antiozonant system is however, protection of either elastomer was illustrated in Table 4. In this case, test-pieces similar to that obtained at 20°C. Thus the C 8 C results are consistent with the expected TABLE 4. EFFECT OF TEMPERATURE ON temperature effects for the two systems, LAYER FORMATION WITH A WAX/ as discussed above. Reasonable protection ANTIOZONANT" PROTECTIVE SYSTEM of NR was obtained with either system. Mass/unit area of layer (mg/cm2) The rate of crack growth in the high 1 2 glass transition temperature elastomers was Elastomer after 7 months (N/t = 70h ' ) at noticeably slower at 8°C, as expected on 20°C 40^0 visco-elastic grounds2'\ NR 0.88 1.16 3 Experiments at — 17 C were carried out ENR 50 0.02 0.39 at a high ozone concentration of about NBR 34 0.05 0.84 105 p.p.h.m. In order to prevent cracking in natural rubber at this concentration at a3 p.h.r 6PPD and 6 p.h.r. Antilux 600 room temperature, about 15 p.h.r. of DOPPD is required"; similar protection s were stored in the absence of ozone at was obtained at 10 p.p.h.m. at — 17°C with either 20°C or 40aC. As can be seen, the 20 p.h.r. DOPPD. rates of layer formation are again higher for NR than for the other elastomers, with Protective Layer Formation the differences being much greater at 20°C than at 40CC (as expected from the diffusion Measurements of the rates of layer mechanism discussed earlier). NBR 34 again formation for the different elastomers shows somewhat faster layer formation correlate with threshold strain observations than ENR 50. These features parallel the for either DOPPD or wax/chemical anti- differences in threshold strain illustrated in ozonant protective systems. Figure 7 shows Table 2. results for vulcanisates containing 10 p.h.r. DOPPD after exposure to 200 p.p.h.m. of Figure 8 compares layer formation results ozone at 40°C. The rate of layer formation at — 17°C with those at room temperature for NR is some four times that for the for an NR vulcanisate containing 10 p.h.r. other materials, paralleling the much better DOPPD. As can be seen the rates of layer protection obtained in NR, while NBR 34 formation at the two temperatures are has a slightly higher rate than ENR 50, virtually identical. The diffusion coefficient again in accord with the relatively small of DOPPD in NR at - 17°C is very similar differences in the threshold strain for these to that in ENR 50 or NBR 34 at 23CC 10 G J Lake and P G Mente Ozone Cracking and Protection of Elastomers at High and Low Temperatures 03 NR Natural rubber O ENR 50 A NBR 34 02 01 ENR 50 10 15 20 25 30 V Time (h12) Figure 7 Dependence of the mass of antiozonant layer formed on the square root of the time oj exposure to an ozone concentration of200pph m at 40 C for vulcamsates of natural rubber, ENR 50 and NBR 34 (cf Table 1) Thus it appears that factors the same order, leading to a corresponding other than slow diffusion of the antiozonant decrease in the rate of layer formation (cf in the elastomer must be involved in the Equation /) A check on this can be made by much lower rates of layer formation with the transfer measurements across the layer chemical antiozonant alone in ENR 50 and (Figure 2) to obtain the permeability of the NBR 34 at 23°C layer material to the unreacted antiozonant 4 (PL in Equation 2} Earlier measurements in- B 2 One factor that may be relevant is the dicated a value for Pr of about 7 x 10 cm /s partitioning of the antiozonant between the for the layer on natural rubber If it is elastomer and the layer material - S{ in assumed that the layer material is the same Equation 1 Swelling and other measure- irrespective of the elastomer (as is possible if ments indicate the solubility of the (un- it is comprised primarily of reacted anti- reacted) antiozonant in the elastomers to ozonant) then differences in P{ for different vary according to NR 11 Journal of Natural Rubber Research, Volume 7(1), 1992 1.0 r- sC* 0.5 - 10 20 30 V Time (h"2) Figure 8. Mass of antiozonant layer formed versus square root of the time of exposure to W5 p.p.h.m. ozone for an NR vulcanisate containing 10 p.h.r. DOPPD at 23°C and 17CC. amount of material transferred is very small. removal of the layer material and repeated The results are therefore subject to consider- swabbing with acetone. A cross-section able scatter but initial results with ENR 50 cut perpendicular to the surface showed the do not indicate any major difference in PL dark colour to extend a small distance compared with NR. (ca 0.1 mm) below the surface. Furthermore, the extent of the penetration was observed to A further factor that may affect layer increase with the passage of time, approx- formation with DOPPD is the fact that imately as the square root of the time that ENR 50 and NBR 34 are more polar had elapsed from the start of the ozone ex- materials than NR. The reacted antiozo- posure. These kinetics indicated a diffusion nant layer is also polar in nature, dissolving coefficient of about one sixth that of very readily in polar solvents such as unreacted DOPPD in ENR 50 (cf Table I). acetone. There may thus be a greater Thus it appears that a somewhat higher tendency for the layer material to diffuse molecular weight reaction product of back into ENR 50 or NBR 34 instead of DOPPD tends to diffuse back from the remaining on the surface to give protection. layer into ENR 50. Similar behaviour has Evidence indicating this to happen has been not been observed with NR but cannot be obtained with a light-coloured vulcanisate of entirely ruled out as NR vulcanisates are ENR 50 containing DOPPD. After exposure generally darker in colour than those of to ozone the surface of the elastomer ENR 50, making any colour change more remained darker in colour even after difficult to detect. 12 G J Lake and P G Mente Ozone Cracking and Protection of Elastomers at High and Lo\v Temperatures CONCLUSIONS REFERENCES 1 BRADEN, M AND GENT, AN (1960) The The work described shows that the test Attack of Ozone on Stretched Rubber Vulcdni- temperature can have marked effects on the zates I The Rate of Cut Growth J appl ozone resistance of vulcanisates containing Polym Si i 3,90 and 100 chemical antiozonant or wax/antiozonant 2 GENT, AN AND McGRATH, JE (1965) protective systems The effects are mainly Effect of Temperature on Ozone Cracking of associated with differences in threshold Rubbers J Polym Sa 4 3, 1473 strain rather than rate of crack growth and 3 GENT, AN AND HIRAKAWA, HJ (1967) vary according to the glass transition Effect of Temperature on Ozone Cracking of temperature of the elastomer and the Butyl Rubbers J Polym Sci A-2 5, 157 protective system adopted Elastomers with 4 LAKE, G J (1970) Ozone Cracking and Protec- high glass transition temperature, such as tion of Rubbers Rubb Chem Technol 43, epoxidised natural rubber or nitnle rubber 1230 are more readily protected at temperatures 5 NAH, S H AND THOMAS, A G (1980) above room temperature, when the protec- Migration and Blooming of Waxes to the tive agents are more mobile Protection of Surface of Rubber Vulcamzates J Polym Set such materials at lower temperatures is Polym Phys Ed 18, 511 correspondingly more difficult but a dialkyl 6 OWEN, P J Private Communication ppd appears better than wax antiozonant 7 HAMPDEN TEST EQUIPMENT LTD Model systems, for reasons that can be understood No 603 from differences in the diffusion mechanisms 8 CRANK, J (1975) The Mathematics of Diffusion Reasonable protection of natural rubber was Oxford Clarendon C obtained at 8 C with a wax/antiozonant 9 BRADEN, M AND GENT, AN (1962) The system and down to — 17°C with a dialkyl Attack of Ozone on Stretched Rubber Vulcan- paraphenylenediamme izates Part 3 Action of Antiozonants J appl Polym Sti,6, 449 10 LAKE, GJ The Determination of Threshold Strain of Rubber in Ozone Resistance Tests Polymer Testing in the press 11 LAKE GJ (1973) Relation entre essais de Date of receipt August 199J laboratoire et duree de vie du caoutchouc en Date of acceptance December 1991 service Re\ Gen Caouich 50, 287 13