(Note: “Optimizing adhesion & Cohesion Parameters in Vulcanized rubber-metal bonding” published in Rubber World August 2014 issue, was presented to “the 12th International Conference on the Science and Technology of Adhesion and Adhesives”, York, UK, September 2013.)
“Vulcanized rubber-metal bonding, Optimizing adhesion and cohesion parameters”
By: Mahmoud Kardan*, Steve Glancy, Robert Ferguson; and Rebecca Seitz
Vernay, Yellow Springs, Ohio, USA
* Corresponding author: Mahmoud Kardan, Ph.D.
The study of rubber-metal bonding in molding process are subject to continual review due to its vast variety applications in medical, consumer products, printing , automotive and aerospace industry. The process, where rubber is vulcanized and bonded directly to the metal insert requires a sophisticated technology and understanding the kinetic of the vulcanization process.
The vulcanization process is a thermally induced cross linking step, which imparts strength, resistance to flow, and elasticity to the crude rubber. Although sulfur vulcanization was the first to be discovered, its application has been limited due to some environmental concerns. Furthermore, sulfur cure requires a carbon-carbon unsaturation in the rubber. Peroxide cure, on the other hand allows vulcanization of both unsaturated and saturated rubbers leading to a network with stable covalent carbon-carbon bond with excellent heat and solvent resistance as well as good compression sets as compared to sulfur cured rubbers. These superior cohesion forces, however, could be at the expense of the adhesion forces between the rubber compound and metal substrates.
This investigation concerns the study of the cohesion and adhesion forces in peroxide cure rubber metal bonding. The measured lap shears between the cured rubber and the metal substrate have been correlated with the structural changes associated with the rubber in cure process, detected by spectroscopy and thermal analysis. These results could have a significant bearing on the mechanism of the strength development in adhesion forces of rubber metal bonding to produce a uniform, high quality product that is free from failure.
The chemical changes of polymers make it possible to create numerous new classes of high molecular weight compounds and to modify the properties and the field of application of fabricated polymers over a wide range. An example of such a chemical changes of polymers is the vulcanization of rubber for production of useful rubber articles. This complex chemical and physiochemical process sharply changes the physical and mechanical properties of the rubber. These changes are due primarily to chemical cross linking reactions between polymer molecules.
The vulcanization process is a thermally induced cross linking step, which imparts strength, resistance to flow, and elasticity to the crude rubber. Ever since Goodyear’s first vulcanization experiment in heating natural rubber with small amount of sulfur (1839), this process has been the most practical method for imparting the drastic property changes, not only in natural rubber, but also in the diene synthetic elastomers such as SBR, butyl, and nitrile rubbers. It has been found since early years of the discovery of the vulcanization, however, that neither heat nor sulfur is essential to the vulcanization process. Sulfur cure requires a carbon-carbon unsaturation in the rubber. Peroxide cure, on the other hand allows vulcanization of both unsaturated and saturated rubbers leading to a network with stable covalent carbon-carbon bond with excellent heat and solvent resistance as well as good compression sets as compared to sulfur cured rubbers. Yet, using co-agents in peroxide cured rubbers could result in a significantly higher crosslink density and better mechanical properties such as tensile strength, modulus and hardness. The increased cross link density, the increase of intermolecular forces in the polymer will change its conformation as well as it’s orientation on the metallic substrate.
Studies by Kardan and co-workers 1- 6 have revealed that when a polymer is coated on a substrate, each structural subunit of the polymer has a defined orientation on the external surface. The orientation of the polymer on the surface and its adhesion to the substrate depends on its conformational changes, which are influenced greatly by the interaction of the polymer with the media 7- 11. In peroxide vulcanization, however, when the average number of cross-links rises, the material becomes more rigid and brittle and losses its flexibility for conformational changes to assume a defined orientation on the substrate surface. The superior cohesion forces peroxide cure elastomers , therefore, could be at the expense of the adhesion forces between the rubber compound and metal substrates due to macromolecule rigidity and lose of its conformation in highly cross-linked polymer. In this study, ATR FT-IR is used as a primary technique to analyze the polymer microstructural changes in the FKM peroxide cured composites. The correlation of the infrared results with the measured lap shears for the composites, when the cross-link density is increased can be useful in the elucidation of the peroxide cure mechanism and its relationship with adhesion-cohesion forces in elastomers. The study of the adhesion of vulcanized rubber to metal in molding process is of a great interest because of its application in many fields such as medical, consumer products, printing, automotive and aerospace industry, binding to metal surface is essential.
Blends of two FKM elastomers, both having the same chemical structures with different Mooney viscosities were made to obtain a processable rubber mix. Then composites of this blend were made by compounding fillers, peroxide, and co-agents into the rubber blend using processing aids (Vernay proprietary) under the same exact conditions. The chemical structure of the peroxide used in this study is 2, 5-dimethyl-2,5-di-(t-butyl proxy) hexane. For all prepared composites, the level of rubbers blend and all ingredients , except the peroxide were kept constant . The level of peroxide was varied from 0.5 to 3.5 PPH (Part per Hundred Part of Rubber).
To measure the physical properties, samples of each compounded composite were molded into 6x6x .075” slabs at 200 C for 2 minutes and post cured at 230 C Oven for 4 hours. These molds are then cut in various shapes needed to test for hardness, tensile strength, modulus, elongation, and compression set.
The cure curves or Rheographs for each composite was obtained by plot of torque (force) values against time for 6 minutes at 180 C using Monsanto Rheometer MDR (Moving Die Rheometer) Model 2000. In this technique, the sample was placed between two dies. The lower die oscillates and the upped die is connected to a torque sensor to measure the torque response of the rubber at the deformation.
To measure compression set, for each composite 1 inch diameter cylinders ( ASTM Button) were mold with 0.5 “ thickness prepared by molding at 200 C for 30 minutes and post cured at 230 C Oven for 4 hours.
In order to obtain information about the molecular orientation and conformational changes of the polymer, the Spectra (400-4000 Cm-1 region) of all compound samples before and after cure with peroxide were obtained using Varian 610 ATR FT-IR Microscope spectrometer. The data were stored in computer for further analysis.
DSC (Differential Scanning Calorimeter) was used to study the exothermic reaction of peroxide in each composite. Samples of each compound (10- 20 mg) were run single heat cycle from 23 C to 250 C, ramping 10 C per minute using TA Instruments DSC model Q20-1558.
Steel substrates of 2.54 cm (one inch) wide were rinsed in toluene, autoclaved and dried in oven. To measure adhesion bond strengths to the steel substrate, samples of each composite were molded to 6.45 cm2 (one square inch) of the substrate for 4 minutes at 180 C. In this molding process, a part of the substrate remained unbounded as the substrate tail and part of the molded rubber was remained (one inch wide) unbounded as the rubber tail. All samples were post cured for 30 minutes at 230 C ovens. Using Instron 5565 Tensile Tester, the lap shear was measured for each prepared rubber molded to steel substrate samples by pooling the rubber from the substrate till separation. .
RESULTS AND DISCUSSION
The substitution of fluorine for hydrogen in organic polymers has resulted in materials with remarkable characteristics called flouroelastomers (FKM). One of the main goals of our analysis is the characterization of the structural aspects of FKM in the composites when cured with different amounts of peroxide and how these structural units contribute to metal bonding. Commercially available FKM’s used in this study comprising copolymers of vinylidene fluoride (VDF), hexaflouropropylene (HFP), terpolymers of tetra fluoroethylene (TFE), as well as perfluoromethylvinylether (PMVE) with about 66 to 70% fluorine content. The performance of FKM especially insolubility can be improved by peroxide vulcanizations. Peroxide cure of FKM result in superior resistance to steam, acid, aggressive engine oils, and chemicals.
In addition to peroxide cure, the performance of FKM in aggressive chemicals depends on the nature of the compounding ingredients used for molding the final products. This performance can vary significantly when varying the compounding ingredients in the formula. In this study, keeping all ingredients, processing aids, and process itself constant as being Vernay proprietary, we concentrate on peroxide variation in the process and examine its affect on cohesion and adhesion forces of the FKM when molded.
One of the aiming of this study is to show that the loading level for peroxide is a critical factor in achieving the degree of rubber modification for desired end use application including adhesion to the substrate, The level of peroxide was varied in this study, to examine the adhesion forces in the rubber as a function of the degree of vulcanization.
It is generally accepted that the free radical formation in peroxide is the first step in peroxide vulcanization as initiation step followed by propagation and termination. In propagation step, the free radical produced by peroxide abstracts a hydrogen atom from the polymer chain and creates another free radical on the polymer itself which cause cross-linking, however, it is not clear how structural disorder is introduced when peroxide is compounded into the rubber. Nor is clear what structural differences exist between the rubbers cured with different amount of peroxides. These structural changes are associated with peroxide concentration and its interaction with the media.
Owing to the weak bonding between the oxygen atoms, peroxide is unstable and easily split into reactive free radicals via hemolytic cleavage that is, two electrons that are involved in the bond are distributed one by one to the two species.
These free radicals are highly reactive towards other substances, or even towards themselves. In our case, the primary decomposition of the peroxide, 2,5- dimethyl -2,5-di(t-butyl peroxy) hexane leads to formation of a t-butoxy radical, which may abstract a hydrogen atom to give a t-butanol or decompose into methyl radical ( 120 C) and acetone. This methyl radical can either abstract a bromine atom from the polymer or add to coagent to give a more stable radical. This stable radical in turn can abstract bromine from the polymer and generate polymeric free radical to initiate cross linking using the co agent as a cross linker.
Theoretically, one crosslink should be expected per each peroxide molecule, but in practice it is lower because of side reactions, which consume the free radicals. For instance at high temperatures, the oxygen can couple to the radical in the polymer backbone and form hydro peroxide radical, which leads to polymer degradation instead of cross linking.
The plot of the measured shear vs extension for different part per hindered (pph) of rubber in composite is shown in Figure 1. Form Figure 1, it can be seen that as the pph of peroxide in composite increases, the adhesion to metal surface decreases. This would be in part due to the polymer conformational changes as we discuss it later in this paper . The initiation step in peroxide vulcanization (peroxide decomposition) follows first order reaction kinetics, i.e., the cleavage of peroxide molecule is only proportional to the concentration of peroxide at any time. The dominant feature in Figure 1 is the abrupt change in plot when peroxide level exceeds about 2 pph. Also, It was interesting to note that as pph of peroxide increases to a certain limit, the composite becomes physically brittle and cause cohesion failure in the system. This observation, suggests that when peroxide level exceeds a certain limit in rubber, it may cause a degradation and cohesion failure. These results indicate that It is important to calculate or experimentally obtain an optimum concentration of the peroxide at a cure temperature for the polymer blend at a given composite to minimize the residual peroxide, because the residual peroxide could lead to oxidation and chain termination or formation of additional cross links, depending on temperature during the process.
The DSC exothermic plots for composites used in this study cured with different amount of peroxides are shown in Figure 2. A single exothermic peak for low peroxide levels is due to splitting peroxide into reactive free radicals, and abstracting hydrogen from the polymer creating another free radical that initiates vulcanization. Evidently, as the concentration of the peroxide exceeds a certain level, the second exothermic pack starts formation and increases as peroxide level increase. We discuss the mechanism of formation of the second peak in microstructural analysis later in this paper. This peak is indication of formation of byproducts discussed above and is not in favor of either cohesion forces or adhesion bonds between the rubber and the metal substrate.
Table I summarizes the results for the mechanical properties measured for the composites of the blend of the flouroelastomers cured with different concentrations (pph) of peroxide. Tensile strength at break (the tensile stress at the moment at which a test specimen tears) maximizes around 1.5- 2.0 pph peroxide for the formulation was used. At this point the composite is tough hard. Although for higher concentrations of peroxide, it increases to even higher values, but the composite becomes more brittle and loses its integrity.
Our interpretation of both DSC (Figure 2) and tensile strength at break results shown in Table 1, are consistent with shear results plotted in Figure 1. In Figure 1, the abrupt change in plot at 3.5 pph peroxide level is not due the adhesion failure to substrate, but it was due to cohesion failure, which causes tear in composite long before adhesion failure.
The measured torque value in MDR cure curves for each composite plotted in Figure 3, which is a direct indication of the sample’s shear modulus (resistance to shearing deformation) may not shows such a drastic changes, when peroxide concentration exceeds 2 pph, perhaps, due to the fact that shear modulus is not sensitive to polymer conformation and its orientation on the substrate surface. However, in Figure 3, two important parameters ts1 and ts2, time required for the increase of 1 unit and 2 unit from minimum torque, respectively, are the same as concentration of peroxide exceed 2 pph. These two parameters, ts1 and ts2 are known as scorch time and are the indication of time for the beginning of the process of cross linking. After the maximum torque has been reached, the slight reversion of the plot for excess amount of peroxide (3.5 pph) could be the indication of the some degradation at higher temperatures. This also, is consistent with adhesive shear measurements and microstructural changes that are described in the next section.
In analyzing the microstructural changes by ATR FT-IR Microscopy, our discussions will first center on the structure of the flouroelastomers blend used in this study and then the structural changes, as different amounts of peroxide milled into the rubber composite will be discussed. The task is made easier by analyzing well assigned infrared vibrations, which are localized in nature with well defined transition moments.
Figure 4 shows the comparison of the ATR spectra for the compounded flouroelastomers blend used in this study compared to the same blend when mill compounded in the presence of peroxide. The predominant band at Figure 3 is near 1100 cm-1, which is assigned to C-F and C-F2 stretching vibrations 12, 13 and evidently does not show any significant changes in intensity with polymer packing structure. The electronegativity of the fluorine atom in flouroelastomers implies a strong and very short distance C-F bonds , higher strength of C-C bonds, and also a very strong Van Der Waals forces between hydrogen and fluorine atoms 12, 13 in these compound. This will cause a lack of mobility of hydrogen atoms in these elastomers and consequently, a lack of C-H vibrations that are infrared active. From Figure 4, it is quite evident for the CH stretching region in 2800- 3000 region, substantial intensity differences exist between the polymer and mill compounded polymer in the presence of peroxide due to breakdown of some hydrogen bonds in the polymer.
Although they are not as well understood, the infrared active bands in the high frequency CH stretching region of 2800- 3000 region can be extremely rich in structural information. The relative intensity of the methyl and methylene stretching vibrations gives valuable information on the orientation of C-C sequences 1,4,5,6 on a metallic surface, which affects the adhesion of the polymer to the substrate. Two bands near the 2923 cm-1 and 2852 cm-1 region can be assigned to CH2 asymmetric and symmetric stretching vibrations, respectively, and band near 2954 cm-1 can be assigned to CH3 asymmetric vibrations. Two bands near 1428 cm-1 (CH2 bending) and 1396 cm-1 (CH2 waging) are also sensitive to the polymer packing structure and structural conformational changes of the polymer in the composite. We assign and interpret some other sensitive vibration bands for C-C stretching, CF2 bending, and CH2 twisting vibrations, later in this section.
The intensity changes for the polymer C-H stretching region as cured with different levels of peroxide are shown in Figure 5. Evidently the total methylene and methyl C-H vibrations slow down and have limited motions as the cross link density increases as results of the vulcanization with peroxide and therefore their total intensity decreases. It is obvious from Figure 5, that the total C-H stretching intensity reaches its minimum value at 1.5- 2.0 pph peroxide level for the composite under study. Therefore, we would interpret this level as a maximum and a constructive cross linking for the compound. It is interesting to note that as the peroxide exceeds this level, the total C-H intensity increases again. This is most likely due to byproduct formation and some decomposition of the polymer when peroxide exceeds this level as we discussed earlier in this paper. Further evidence for maximum cross linking at about 2.0 pph peroxide level is the shift of wave number (energy) for CH2 asymmetric vibration from near 2923 to near 2926 at maximum cross linking and shifting back to 2924 as peroxide exceeds the optimum level.
Although, the total C-H stretching vibration decreases as cross link density increase, but there will be relatively more asymmetric vibrations as compared to symmetric vibrations, if any. Therefore, we further interpret the increase of asymmetric/ symmetric CH2 vibrations in infrared being as increase of cross link density in the compound. As can be seen in Figure 6, the asymmetric/ symmetric CH2 ratios increase drastically for the oven post cured compounds at 230 C, when the peroxide level increases to about 1.5- 2.0 pph and then levels off, further indicating that the excess level of peroxide is not in favor of a solid and maximum cross linking and may damage the rubber. These changes less pronounced for the mold cured compounds at 180 C. It is not unlikely that these differences are the result of further decomposition of the polymer at 230 C due to the excess amount of peroxide and in the presence of oxygen. It was interesting to note that the ratio levels off for the same post cured samples were aged for 30 days at room temperature, indicating, and the broken bonds re- cross link again as the cured compound ages. However, the nature of these re-united bonds may not be the same as initial cured bonds.
The decomposition of the polymer at 230 C as peroxide exceeds an optimum level could be due to susceptibility of portion of this elastomeric copolymer to hydrolysis of some double bonds and formation of carbonyl group. The appearance of band near 1743 Cm-1, in Figure 4, is indication for formation of this group in very small amount even when the flouroelastomers is mill compounded in the presence of peroxide without exposing to heat. Here we argue, based on infrared data that most likely, only the PVDF portion of flouroelastomers undergoes dehydrofluorination when the level of peroxide exceeds a certain level and creates double bond. In return, this double bond goes under hydrolysis to form a carbonyl band.
Due to the strong bonds between carbon and fluorine as mentioned earlier, In flouroelastomers, the carbon atoms to which the fluorine are bonded vibrate with considerably smaller amplitude than the carbon atoms bounded to the hydrogen atom . A week band at 854 Cm -1 appearing for high levels of peroxide in the compound in Figure 7, is assigned to the symmetric stretching vibration of C-C bond of PVDF calculated by normal coordinate analysis 12, 13 with small contribution from stretching of C-F bond ( 16%) and deformation of H-C-H angle ( 12 %). We conclude that the dehydrofluorination due to excess amount of peroxide occurs on PVDF and most likely, the other copolymers in FKM will not be affected by these phenomena. The band appearing at and at 765 Cm -1, is calculated at 761 Cm -1 is assigned to CF2, which evidently is not sensitive to dehydrofluorination. Also, the band at 892, assigned to CH2 rocking vibration or the band at 942, calculated at 938 and assigned to the twisting vibration of CH2 remain constant during these changes.
The strength related properties such as cohesion and adhesion forces in peroxide cured composites of flouroelastomers are sensitive to the peroxide level in the composite. Our results suggest that there is an optimum peroxide level to be used to balance the cohesion and adhesion forces to obtain a favorable bonds in flouroelastomers metal bounded systems. These optimum concentrations needs to be calculated or experimentally obtained for each composite based on the amounts of the polymers, their structure , and the amount of other ingredients used in the composite. The conformation of the polymer, the level of the peroxide, and the degree to which, the copolymer can undergo dehydrofluorination as peroxide exceeds an optimum level play important role in the rubber cohesion forces, i.e., the integrity of the composite as well as maximized the rubber metal adhesion bonds.
We have provided some evidence that the PVDF portion of the flouroelastomers copolymers undergo changes as the peroxide concentration exceeds the optimum level in the composite and most likely, the remaining copolymer structures stay unchanged. We also have demonstrated, when peroxide concentration is under the optimum values, the cross linked density is not optimized and the integrity of the polymer is not maximized. .
- Kardan, M., Rubber Chemistry and Technology, 74, 614(2001).
- Kardan, M., GAK Gummi Fasern Kunststoffe, a German publication for polymer industry 7, 435 (2002).
- Kardan, M., GAK Gummi Fasern Kunststoffe, a German publication for polymer industry 8, 520 (2003).
- Kardan, M. Kaito, A., Hsu, S.L., Thakur, R., Lillya, C.P., J.Phys.Chem,19, 1809(1987).
- Kardan, M., Reinhold, B.B., Hsu, S.L., Thakur, R, Lillya, C.P., Macromolecules, 19, 616(1986).
- Kaito, A. Kardan, M, Hsu, S.L., Thakur, R., Lillya, C.P., Polymer Preprints, 28, 11(1987).
- Kardan, M , Rubber and Plastic News, August 6, 2001.
- Kardan, M, Adhesive and Sealant Industry, October/ November 2001.
- Kardan, M, Adhesive Ages, September 2000.
- Kardan, M., Paint and Coating Industry, May 2000.
- Kardan, M., Coating World, September 1999.
- Boerio F. J., Koeing, J.L., JPolymer Sci, A-2, 9 (1971)1517).
Boerio F.J, Koenig, J.L., J.Polymer Sci, A2, 7(1969) 1489