ROHSTOFFE UND ANWENDUNGEN RAW MATERIALS AND APPLICATIONS Rubber
Cross-link distribution
Network heterogeneity
AFM
Mechanical properties

Cross-link Clusters: reality or fiction?

One of the intriguing questions has always been how chemical rubber-networks really look like. One possible explanation for the fundamental differences in properties between sulphur and peroxide cured networks, is that the cross-links formed by either of the two tend to cluster leading to what we will call network heterogeneity. The paper gives an overview of past and recent attempts to characterize this heterogeneity of vulcanisation networks. Recent data generated with Atomic Force Microscopy on sulphur, peroxide and radiation cured EPDM shed new light on this fundamental question. Grainy structures are observed, as were seen before with TEM-microscopy, which can tentatively be interpreted in terms of heterogeneities of the pertinent cross-linked networks.

A review of the state of the art, enlarged with 1 recent AFM-data

Kautschuk Netzwerke: RealitaÈt oder Fiktion? UÈberblick uÈber den Stand der Forschung, erweitert durch neue AFM-Daten Gummi
Netzknotenverteilung
NetzwerkheterogenitaÈt
RKM (Rasterkraftmikroskopie) / AFM (Atomic Force Microscopy)
Mechanische Eigenschaften Eine der herausragenden Fragen ist seit langem wie ein chemisches Kautschuknetzwerk wirklich aussieht. Eine moÈgliche ErklaÈrung fuÈr die Unterschiede in Eigenschaften zwischen schwefel- und peroxidvernetzten Kautschuken ist die NetzwerkheterogenitaÈt, wodurch man die Aggregationstendenz der Vernetzungspunkte versteht. È bersicht uÈber Dieser Beitrag gibt eine U bekannte und neuere Versuche, HeterogenitaÈten von vulkanisierten Netzwerken zu charakterisieren. Die grundlegende Frage der NetzwerkheterogenitaÈt von schwefel-, peroxid- und strahlenvernetztem EPDM erscheint durch unsere neuesten mittels Rasterkraftmikroskopie erhaltenen Ergebnissen einem neuen Licht. Granulare Strukturen, wie zuvor schon mit TEM beobachtet, werden mittels Rasterkraftmikroskopie beobachtet. Diese koÈnnen als HeterogenitaÈten der Netzwerke interpretiert werden.

426

E. W. Engelbert van Bevervoorde-Meilof, D. van HaeringenTrifonova, G. J. Vancso, L. v. d. Does, A. Bantjes, J. W. M. Noordermeer, Enschede (The Netherlands)

A considerable amount of work has been spent over the years to predict the effects of network structure on the mechanical properties. The network structure is not defined only by the amount of cross-links but also by the type and spatial distribution of the cross-links. An uneven or heterogeneous cross-link distribution may be the result of a sort of clustering of cross-links for either physical or chemical reasons, resulting in locally high crosslink densities, see Fig. 1. A vast amount

1

Presented at the Kautschuk-Herbst-Kolloquium '98, Deutsches Institut fuÈr Kautschuktechnologie e.V., Hannover, BRD, 15 ± 17 October 1998

of literature exists, quoting various causes for cross-link clustering. A few representative examples of these causes may be referred to in this context for the purpose of clarification. Mark [1] indicates that the addition of a peroxide radical to a double bond may give rise to a polymerisation reaction between adjacent double bonds. This occurs in a relatively small volume until termination takes place. This results in the formation of small and intensely crosslinked polymer spheres in a much less firmly cross-linked polymer matrix. Gehman [2] proposed the opposite, an autocatalytic reaction during sulphur vulcanisation. In the vicinity of a polysulphidic cross-link higher localised concentra-

Fig. 1. Crosslink cluster

KGK Kautschuk Gummi Kunststoffe 53. Jahrgang, Nr. 7-8/2000

Cross-link Clusters: reality or fiction? . . .

tions of sulphur exist, because of higher solubility of sulphur in these regions. In addition, during vulcanisation also desulphuration of polysulphidic cross-links occurs. The autocatalytic vulcanisation would lead to the formation of adjacent cross-links, thus to clustering of sulphidic cross-links. Vilgis [3] more recently postulated that the dynamics of a chain in the neighbourhood of a cross-link is slower compared with un-cross-linked regions. Thus it is more probable that such a chain is cross-linked once more; i.e. a cross-link will appear near another one. There is a marked difference in properties between sulphur and peroxide cured networks: high tensile strength and elongation at break for sulphur compared to peroxide vulcanisates. Besides the difference in cross-link type: sulphur bridges versus carbon-carbon cross-links, network heterogeneity is also quoted to cause the difference between sulphur and peroxide, albeit that opinions differ as to which of the two gives the most homogeneous network, as seen before [1 ± 2].

strength and larger elongation at break are obtained. l A heterogeneous network has a lower Young's modulus than a homogeneous network, because a cross-link cluster can be considered to act as a multifunctional cross-link. This multi-functionality does not fully compensate the net lower cross-link density as expressed in number of (cluster) crosslinks per unit volume. l From the overall shape of the stressstrain curve no evidence can be derived relating to network heterogeneity and consequently to the effect of the latter on ultimate properties. The overall conclusion to be drawn is, that simple stress-strain measurements of rubber vulcanisates by themselves cannot be used to derive evidence about network heterogeneity and its consequence for ultimate properties. This evidence has to be gained from other sorts of experiments that measure local properties instead of bulk properties and can only indirectly be related to observed differences in ultimate properties measured in tensile tests.

Theoretical considerations on network heterogeneity

Existing experimental evidence for network heterogeneity and its influence on ultimate tensile properties

Polymer networks are commonly characterised by macroscopic parameters, like the Young's modulus and the degree of swelling in good solvents. By reference to the theories of rubber elasticity such data are analysed to give values for molecular parameters, such as the degree of cross-linking. Such analysis does not provide information about possible ± even likely ± occurrence of spatial variations in the degree of cross-linking. The kinetics of network formation can easily cause a non-random spatial distribution of cross-links. It is only recently that theoretical attempts have been made to include the effects of this non-random spatial distribution of cross-links into the theories of rubber elasticity [3 ± 6]. Some of the conclusions and predictions as to the effects of network heterogeneities on the vulcanisate properties are summarised here: l A heterogeneous network allows for a more effective distribution of stress due to relaxation effects of the softer areas. It can therefore accommodate larger deformations: higher tensile

Light Scattering of Swollen Networks The first to realise the possibility of effects of network heterogeneity were Stein [7], Bueche [8], Wun and Prins [9]. They found an increase of factors as large as 100 in the light scattering of swollen networks of a variety of polymers, dependent on the degree of crosslinking and corresponding to the degree of swelling. The higher the swelling the more light scattering counterparts. This was assigned to non-uniform swelling of the networks, as a result of the inhomogeneous character of the location of the cross-links in the networks. Strongly

cross-linked areas will swell less than lightly cross-linked, which gives rise to refractive index differences and thus extra light-scattering. Wun and Prins [9] and Stein et al. [10] derived a so-called correlation distance which serves as a measure for the average size of the regions of cross-link heterogeneity. Typical correlation distances for the systems investigated by these authors are given in Tab. 1. The length scales of ca. 500 nm as found by Stein et al. compare to the average molecular length between cross-links of 5 nm, and are therefore a factor 100 larger. It should be realised though that light scattering can only determine heterogeneities of the order of the wavelength of light, not much smaller than the 500 nm. No direct correlations were laid with ultimate tensile properties of these networks. Freezing Point Suppression of Solvents in Swollen Networks Freezing point depression of solvents, imbibed in swollen polymer gels has been known for long. Kuhn and Majer [11] ascribed the freezing point depression to a limited crystal growth in the gel, resulting from a mesh formed by the chain segments in the polymer network. A higher vapour pressure can be expected for microcrystals due to their high surface to volume ratio, rendering a decrease in freezing point. Polymer chains subdivide the solvent in the gel, creating zones too small for nucleation to occur at the freezing point of the solvent. The higher the cross-link density, the larger the freezing point depression. A linear relationship between freezing point depression and volume fraction of rubber in the swollen network is expected for a randomly crosslinked network. A heterogeneous swollen gel will contain domains with higher and with lower cross-link densities than the mean cross-link density. Deviations from non-linearity are indicative of heterogeneity of cross-link density in the network.

Tab. 1. Typical correlation distances of crosslink heterogeneity as observed with light scattering of swollen networks Polymer system

Technique

Length scale (nm)

Reference

Poly(2-hydroxyethyl methacrylate) cis-BR

Swollen light scattering

1000 ± 2000

3

Sw. light s.

ca. 500

4

KGK Kautschuk Gummi Kunststoffe 53. Jahrgang, Nr. 7-8/2000

427

Cross-link Clusters: reality or fiction? . . .

Grobler and McGill [12] studied polyisoprene networks formed by four different curing systems: Conventional sulphur, sulphur donor, peroxide and radiation curing. They found evidence for increased network heterogeneity in the order: radiation curing peroxide curing sulphur donor conventional sulphur

most homogeneous > j ? most heterogeneous

Thus, tentatively they concluded that an increase in network heterogeneity would parallel an increase in ultimate tensile properties: tensile strength and elongation at break, as indicated before. Bimodal networks In an extensive series of papers recently reviewed, Mark [13] and many co-workers investigated model networks of end-linked polydimethylsiloxanes with bimodal chain length distributions: a large number of very short chains linked to much longer chains. These bimodal networks had better mechanical properties (high tensile strength in combination with a high elongation at break) than unimodal networks. The interpretation of the phenomena given by Mark is the limited extensibility of the short network chains, which if included as a non-Gaussian contribution in the theory of rubber networks elasticity does indeed predict such a behaviour of the stress-strain relationship. The short chains add to the strength because of their limited extensibility, but do not negatively influence the elongation at break as long as their relative content in the matrix remains limited. During deformation at first only the long chains will deform, the short chains do not contribute until close to rupture. Jacobi et al. [14] prepared heterogeneous networks of polyisoprene by a two-step vulcanisation. During the first step densely cross-linked IR molecules were created in dilute solution by intramolecular cross-linking with various amounts of a bis-triazoline-dione as cross-linking agent. After removal of the solvent, dicumyl peroxide was added and a second vulcanisation was performed to create intermolecular cross-

428

links. The sum of cross-links of both steps was kept equal throughout the series. With increasing relative amount of the triazolinedione versus peroxide the network heterogeneity increased. The outcome of their experiments was, that with increasing network heterogeneity the Young's modulus decreased, the elongation at break increased and the tensile strength essentially remained constant. Both sorts of bimodalities give evidence, that network heterogeneity parallels increased ultimate properties. However, as to the particular influence on either tensile strength or elongation at break the two sorts of networks just show opposite effects! Others related techniques Various other techniques have been or are being tried to shed more light on the question of cross-link heterogeneity, like birefringence measurements, neutron scattering and lately NMR analysis: 1H T2 relaxation and 13C cross polarisation with magic angle spinning [15]. In any case, these are rather complicated techniques from which it seems difficult to draw solid conclusions.

New evidence by atomic force microscopy In an attempt to visualise an eventual spatial heterogeneity of cross-links, phase imaging was performed, an atomic force microscopy technique (AFM) in the so-called tapping and force modulation mode. Herewith it is possible to detect surface topography, the size, shape and spacing of different material domains in for instance polymer blends. Also differ-

ences in surface modulus or viscoelasticity of a sample produce phase contrasts. Therefore, with phase imaging local differences in cross-link density resulting in differences in the local modulus should also be detectable. Experimental Three types of non-reinforced EPDM vulcanisates were investigated: sulphur, peroxide and electron beam cured, and compared with an unvulcanised compound and a green solvent-cast EPDM sample. Compound recipes are given in Table 2. The compounds were vulcanised in a press at 200 bar at 160 8C and 170 8C for the sulphur and peroxide compounds, respectively. Electron beam irradiation was performed at the Interfaculty Reactor Institute of the Delft University of Technology with a 3 MV Van de Graaff electron accelerator in nitrogen atmosphere at room temperature. The rubber sheets were turned after irradiation with half the radiation dose to obtain a uniform dose density. Solvent cast films of the green EPDM were prepared from a 10% solution in toluene by putting a drop of the solution on mica and drying in air. The sample surfaces were microtomed at ÿ 90 8C with a diamond knife. The surfaces were left to relax prior to examination. The AFM experiments were performed with a NanoScopeâ III (Digital Instruments) set up. The images were obtained under ambient conditions while operating the instrument in the tapping and force modulation mode. Commercial Si cantilevers were used.

Tab. 2. Recipes of EPDM-compounds for AFM purposes (phr) Curing technique

Sulphur 1

2

EPDM Keltan 514 ZnO Stearic acid CBS TMTD Sulphur Dicumyl peroxide Radiation dose (kGy) Total (phr)

100 5 1 0.85 0.45 0.45

100 5 1 2.7 1.35 1.35

Peroxide

Electron Beam Radiation 1 2

100

100

100

100 100

300 100

2 107.75

111.4

102

KGK Kautschuk Gummi Kunststoffe 53. Jahrgang, Nr. 7-8/2000

Cross-link Clusters: reality or fiction? . . .

Results

Fig. 2. Height (left) and phase image of the unvulcanised compound

All images of the tapping mode experiments revealed the same structure. Figs. 2 ± 5 show height (left) and phase images (right) of the unvulcanised, a sulphur cured, an electron radiation cured and the peroxide vulcanisate respectively. Each image shows a grainy structure: domains of various shapes with sharp boundaries. Force modulation AFM mode revealed the same grainy structures, but the image quality was poorer than that of the tapping mode. This was not further pursued. AFM tapping mode experiments performed with different forces showed a change in the observed phase shift or relative contrast between the grains and boundaries. This means that the pattern visible both in the height and phase images is not just due to surface topography. They should be attributed to differences in local stiffness, viscoelasticity or adhesion. The observed change in the phase shift with the change in tapping conditions suggests that there is a difference in the properties between the grains and the boundaries.

Discussion of the AFM-results

Fig. 3. Height (left) and phase image of sulphur vulcanisate 1

Fig. 4. Height (left) and phase image of electron beam vulcanisate 1

430

Several explanations can be given for the nature of this grainy structure. It is important to note, that similar grainy structures were observed in TEM micrographs of swollen vulcanisates and swollen vulcanised blends of natural rubber by Shiibashi [16] and Tinker [17]. Shiibashi found that the grain size decreased with increasing network chain density and felt that he saw the actual cross-link-network. Cook et al. [18] pointed out objections to his interpretation in relation to the experimental set-up. It is highly interesting that AFM also registers these grains. It seems to support the interpretation of Shiibashi. The grain boundaries seen with AFM would then correspond to other viscoelastic properties ± e. g. higher modulus due to the cross-linking ± compared to the unvulcanised matrix in between. Typical grain sizes for all samples of our investigation are given in Tab. 3, together with the network chain densities obtained from equilibrium swelling. Although the grain size of the unvulcanised compound is markedly higher than that of the vulca-

KGK Kautschuk Gummi Kunststoffe 53. Jahrgang, Nr. 7-8/2000

Cross-link Clusters: reality or fiction? . . .

Fig. 5. Height (left) and phase image of the peroxide vulcanisate

decrease in grain size. However, AFM measurements of a solvent cast film ± which had not been prepared with a microtome ± did not show any signs of a grainy structure. These are contradictory results, which do not support the explanation of micro-cracks. But the grainy structure is related to the way the samples are prepared. As another explanation, Kilian suggested that the grainy structure is a visualisation of an aggregate structure of liquids [19]. Atomic Force images of glassy layers of polymethyl methacrylate revealed similar heterogeneous structures. They reflect the many equivalent microstructures a liquid is running through constantly. The microstructures or aggregates are considered as dynamic subsystems (reversible aggregation) with a broad and asymmetric size distribution. Rubbers have much lower glass transition temperatures than glasses, but at the high frequencies of the AFM tapping mode experiments, quasi-stationary

nised compounds, there is no clear relation between grain size and network chain density for the others. Neither is there a clear difference in grain size between the various vulcanisation methods. The linear grain size observed by Shiibashi ranged between 5 and 15 nm. Although smaller than our values, still rather large for coiled polymer chains. Furthermore, also the unvulcanised sample showed a grainy structure. For those reasons it is indeed not very likely that the grainy structure is a representation of the cross-link-network. Another explanation raised is the preparation method of the AFM-samples. Cooling of the samples in liquid nitrogen might result in microcracks throughout the sample, which could become visible as boundaries between the grains. To check this, it was investigated whether such cracks would heal upon annealing of the samples, in particular for the unvulcanised sample. On the contrary, annealing for 1.5 and 7.5 hours only leads to a

Tab. 3. Typical mean grain sizes (nm) for the samples investigated, before and after annealing for 1.5 and 7.5 hours. Also mentioned network chain density m Compound

m Eq. Swelling (mol/cm3  10ÿ4)

Unannealed

Annealing time 1.5 hrs

at 80 8C 7.5 hrs

Unvulcanised Sulphur 1 Sulphur 2 Peroxide Radiation 1 Radiation 2

ÿ 1.55 3.42 3.61 1.44 3.61

79 47 56 49 53 59

56  3 56  3 ÿ ÿ ÿ 51  2

39  2 36  2 ÿ ÿ ÿ 28  1

     

2 2 2 2 3 3

KGK Kautschuk Gummi Kunststoffe 53. Jahrgang, Nr. 7-8/2000

structures of the rubber surface should become visible. According to Kilian it might be expected that chemical crosslinks are formed preferentially in the boundaries between the grains and might therefore be heterogeneously distributed. In that case, the AFM tests do see a true heterogeneity of the networks resulting from the high modulus of the boundaries where all cross-links are concentrated. This explanation would correspond with the fact that the solvent-cast film did not show a grainy structure, but cannot explain why the unvulcanised compound did show grain as well. Another somewhat speculative explanation might be that the grains reflect striae created during the mixing process of the compounds. The boundaries constitute masticated sheared rubber, surrounding globules of non-masticated material. While all shearing and mixing thus is concentrated in the boundaries, also the vulcanisation ingredients tend to concentrate in the boundaries. Their transport into the globules is controlled by diffusion. By the same token as with explanation 3, this would mean that cross-links are preferentially formed in the boundaries. Or the striae are merely showing a viscoelastic behaviour different from the matrix, because the material in the striae is masticated and the matrix not. The latter would explain why all compounds show grain and the solvent-cast film not.

Conclusion None of the theoretical and experimental works on topological network heterogeneities has given a conclusive explanation for the difference in ultimate properties of sulphur versus peroxide vulcanisates. AFM measurements shed new light on this question. Grainy structures are observed with AFM, like seen before with TEM-microscopy. The fact that these grains show up in two fundamentally different techniques indicates that it is a real phenomenon. The cause of these grains needs further study. Some of the ± still speculative ± explanations point in the direction of heterogeneities in the distribution of vulcanisation ingredients and consequently, heterogeneous cross-link distributions. Even then, it is still a long way to link this to the differences in ultimate properties between sulphur and peroxide vulcanisates.

431

Cross-link Clusters: reality or fiction? . . .

References [1] H. F. Mark, Rubber Chem. Techn. 61 (1988) G73. [2] S. D. Gehman, Rubber Chem. Techn. 42 (1969) 659. [3] T. A. Vilgis, Macromolecules 25 (1992) 399. [4] T. A. Vilgis and G. Heinrich, Kautsch. Gummi Kunstst. 45 (1992) 1006. [5] A. Onoki, J. Phys. II (Paris) 2 (1992) 45. [6] T. A. Vilgis and G. Heinrich, Macromol. Theory Simul. 3 (1994) 271. [7] R. S. Stein, Polymer Letters (J. Polym. Sci.) 7 (1969) 657. [8] F. Bueche, J. Coll. and Interf. Sci. 33 (1970) 61. [9] K. L. Wun and W. Prins, J. Pol. Sci., Polym. Phys. Ed. 12 (1974) 533. [10] R. S. Stein, R. J. Farris, S. Kumar and V. Soni: Elastomers and Rubber Elasticity, J. E. Mark and J. Lal Eds., ACS Symposium Series 193, ACS Washington DC (1982).

[11] W. Kuhn and H. Majer, Angew. Chem. 68 (1956) 345. [12] J. H. A. Grobler and W. J. McGill, J. Polym. Sci.: Part B: Polym. Phys. 32 (1993) 287. [13] J. E. Mark, Acc. Chem. Res. 27 (1994) 271. [14] M. M. Jacobi, M. Bandeiry, E. Birnfeld, J. Rohrmann and L. Porto Lusa, International Rubber Conference, NuÈrnberg BRD (1997) 39. [15] V. Litvinov in E. W. Engelbert van Bevervoorde, PhD Thesis Twente University, June (1998). [16] T. Shiibashi, International Polymer Science and Technology 14 (1987) T/33. [17] A. J. Tinker, Blends of Natural Rubber, A. J. Tinker and K. P. Jones, Eds., Chapman & Hall, London (1998). [18] S. Cook, P. E. F. Cudby and A. J. Tinker, Am. Chem. Soc., Rubber Div., Nashville (1992). [19] H.-G. Kilian, B. Zink and R. Metzler, J. Chem. Phys. 107 (1997) 8697.

Acknowledgement This research was sponsored by the Ministry of Economic Affairs of the Netherlands, Program IOPPCBP and TNO Industry.

The authors Mrs. E. W. Engelbert van Bevervoorde-Meilof is presently employed as a Materials Engineer at Philips Competence Centre Plastics B. V., Eindhoven, the Netherlands; Mrs. D. van Haeringen-Trifonova and Prof. G. J. Vancso are affiliated to the Twente University, Enschede, the Netherlands, research group of Material Science and Technology of Polymers; Dr. L. van der Does, Prof. A. Bantjes and Prof. J. W. M. Noordermeer are affiliated to the Twente University, research group of Rubber Technology. Corresponding author University of Twente, Faculty of Chemical Technology, Prof. Dr. J. W. Noordermeer, Rubber technology, P.O. BOX 217, NL-7500 AE Enschede

VERANSTALTUNGEN EVENTS 18. ± 22. 9. 2000 Hannover

Produktorientierte Kautschuktechnologie, Fortbildungsseminar fuÈr Meister/Techniker S

Deutsches Institut fuÈr Kautschuktechnologie, Eupener Str. 33, D-30519 Hannover, Tel: +49(0)511/84201-0, Fax: +49(0)511/8386826

25. ± 27. 9. 2000 Halle

Polymerwerkstoffe' 2000 L

Institut fuÈr Polymerwerkstoffe, Martin-Luther-UniversitaÈt, Geusaer Straûe, D-06217 Merseburg

27. ± 28. 9. 2000 Aachen

Kautschuktechnologie: Prozessauslegung und Produktentwicklung S

Institut fuÈr Kunststoffverarbeitung in Industrie und Handwerk, RWTH Aachen, Pontstr. 49, D-52062 Aachen, Tel: 0241/803806, Fax: 0241/8888262

27. ± 28. 9. 2000 Weinheim

Schadensanalyse an Elastomerbauteilen S

VDI-Gesellschaft Werkstofftechnik, Postfach 101139, D-40002 DuÈsseldorf, Tel: 0211/6214556, Fax: 0211/6214160

5. ± 8. 10. 2000 Kuala Lumpur, Malaysia

M-PLAS 2000 Messe DuÈsseldorf Asias Pte Ltd., International Plastics and Rubber Trade Fair 5 Temasek Boulevard, #05-05 Suntec Tower Five, for Malaysia Singapore 033985, Tel: (65)332 9620, M Fax. (65)332 9655

9. ± 11. 10. 2000 Berlin

Berliner Polymerentage 2000 Sy

Humboldt-UniversitaÈt, Institut fuÈr Physik, Invalidenstr. 10, D-10115 Berlin, Tel: 030/2093-7788, Fax: 030/2093-7632, E-mail: [email protected]

10. ± 11. 10. 2000 Berlin

High Performance Elastomers Conference C

Rapra Technology LTD., Shawbury, Shrewsbury, Shropshire, SY4 4NR UK, Tel: +44(0)193925-0383, Fax: +44(0)193925-1118

10. ± 11. 10. 2000 Aachen

Fluidinjektionstechnik- Verfahren und Anwendungen S

Institut fuÈr Kunststoffverarbeitung in Industrie und Handwerk, RWTH Aachen, Pontstr. 49, D-52062 Aachen, Tel: 0241/803806, Fax: 0241/8888262

11. ± 12. 10. 2000 Kaiserslautern

IVW-Kolloquium K

Institut fuÈr Verbundwerkstoffe GmbH, Erwin-SchroÈdinger-Str., Geb. 58, D-67663 Kaiserslautern, Tel: 0631/2017-303, Fax: 0631/2017-199

M ± Messe, T ± Tagung, Ko ± Kongress, A ± Ausstellung, S ± Seminar, Sy ± Symposium, L ± Lehrgang, W ± Workshop, C ± Conference, Co ± Courses, K ± Kolloquium

432

KGK Kautschuk Gummi Kunststoffe 53. Jahrgang, Nr. 7-8/2000