INVESTIGATION OF THE INFLUENCE OF STEARIC ACID ON RUBBER–BRASS ADHESION
Abstract
The influence of stearic acid loading on the adhesion of rubber to brass-plated steel wires was investigated. The so-called squalene method was employed to investigate the adhesion interface built up during the vulcanization reaction. Variation of the stearic acid loading has a direct influence on the bonding interface and at the same time also has a strong influence on the rubber properties. The surface of the sulfidated wires was analyzed using optical, focus variation, and scanning electron microscopy coupled with energy dispersive X-ray spectroscopy. Increasing amounts of stearic acid accelerated the sulfidation reaction. Furthermore the focus variation microscopy data were used to calculate roughness parameters of the sulfidated wires. Rubber properties and adhesion values were measured for natural rubber compounds with variable amounts of stearic acid. In most cases the adhesive strength exceeded the cohesive strength of the rubber.
INTRODUCTION
Brass-coated steel cords and wires are extensively used as reinforcement materials in various rubber products such as hydraulic hoses, tires, and handrails. Therefore the mechanism of the adhesion process has been the subject of many studies since the 1960s, which were summarized in many reviews.1–5 It is generally accepted that adhesion is achieved through the reaction of a thin brass layer (around 200 nm) with the rubber compound during the vulcanization process.1,2 According to the mechanism described by van Ooij, an active sulfurating species reacts with the brass layer, and an adhesion interface between the rubber and the steel wire is formed.6 The adhesion interface consists of a mixture of nonstoichiometric copper sulfides (CuxS) and zinc sulfides. It is not totally clear how the interaction of the sulfide layer with the rubber produces a stable bonding, but it is postulated that adhesion is achieved through the mechanical interlocking of rubber with the rough adhesion interface.1
The reaction and hence the adhesion is affected by many factors, such as the brass composition and thickness7–9 as well as the compound composition.10–15 For good bonding a copper sulfide layer must form in an optimal thickness. If the interface is too thin no effective bonding occurs, if it gets too thick it gets brittle and may detach from the brass.16 In both cases the adhesion strength would be rather low. Therefore, the understanding of the influence of compound ingredients on the adhesion layer is very important to control the adhesion in rubber products.
The purpose of this investigation is to determine the effect of stearic acid loading on the rubber adhesion. Stearic acid is normally added to rubber compounds to work in combination with ZnO as an activator.17,18 Stearic acid reacts with ZnO to create the soluble zinc stearate. A zinc-accelerator complex is built up, and stearic acid or amines work as ligands. This complex reacts with S8-rings of the sulfur in the compound to create a polysulfide complex, which is responsible for sulfur transfer.19 By this process, the effectiveness of the accelerator is enhanced. For this reason stearic acid loading affects not only the rubber properties but also the adhesion of rubber to brass-coated steel.
The effect of stearic acid on the adhesion was investigated by Jeon et al.,13 who studied the effect of stearic acid contents in the range of 0 to 10 phr. The best adhesion values were achieved with a mixture containing 3 phr stearic acid for unaged compounds. Furthermore, they concluded that very high stearic acid loading deteriorates the adhesion properties.13 Therefore, the present study investigates the influence of small variations in the stearic acid content in a range (0–2 phr) that is typically employed in rubber compounds.
A known problem concerning the analysis of rubber to brass adhesion is the question of how to make the interface visible without damaging it. As a result, the concept of using a liquid model compound (squalene) instead of rubber has been developed.20 Squalene is a low molecular weight analogue of natural rubber (NR) and has therefore been extensively used to simulate reactions in rubber compounds.21 It has also been used to study the effects of various compound formulations on the sulfidation reaction of brass-plated steel during vulcanization.10,14
In this study a simplified squalene mixture was used to analyze the adhesion interlayer built up during the vulcanization reaction. But this approach cannot be used to investigate all aspects that concern actual adhesion. For this reason, a rubber compound, which is analogous to the squalene mixture, was used to determine rubber properties and adhesion values depending on the stearic acid loading. The characterization of the adhesion layer was done by optical microscopy, focus variation microscopy, and scanning electron microscopy (SEM) coupled with energy dispersive X-ray analysis (EDX).
EXPERIMENTAL
Brass-coated steel wires (67.5% Cu, 32.5% Zn, diam 0.71 mm) were provided by Semperit Technical Products GmbH.
Squalene (98%) was purchased from Sigma Aldrich (Vienna, Austria). All other materials (naphtenic oil, zinc oxide, stearic acid, cobalt stearate (9.3–9.8% cobalt), sulfur, N,N-dicyclohexylbenzothiazole sulfonamide (DCBS) and N 550 carbon black) were provided by Semperit Technische Produkte GmbH (Wimpassing, Austria) and were used without further purification.
Squalene experiments were performed according to Hamed and Paul.22 Formulations are given in Table I. All mixtures contained squalene, naphthenic oil, zinc oxide, sulfur, and DCBS and varying amounts of cobalt stearate and stearic acid. Co stearate has the function to act as an adhesion promoter. In sample AS no Co stearate was used to obtain samples that are totally free of stearic acid.
Before the reaction, brass-coated wires were cut to the required length and washed with toluene. Squalene was heated with naphthenic oil to 160 °C. Then zinc oxide and stearic acid were added and stirred for 1 min. Subsequently cobalt stearate was added and stirred for another minute followed by DCBS and sulfur. After another minute of stirring the wires were immersed into the squalene mixture for 20 min. We were careful to retain a turbulent stirring during the reaction to get a uniform reaction on the wire surface. Then the sulfidated wires were removed, washed with toluene, and stored in vials under a nitrogen atmosphere until characterization.
Specimens for pull-out testing were prepared according to ASTM D 1871. Compound formulations are identical to the formulations used in the squalene experiments, which are given in Table I. A second testing series with carbon black (see Table II) was also produced to get a better comparability with real rubber compounds. In samples A and AC no Co stearate was used to obtain samples that are totally free of stearic acid. The compounds were vulcanized for 20 min at 160 °C with embedment lengths of 10 mm. Cure rate data (scorch time t05, optimum curing time t90, minimum torque ML, maximum torque MH) of the rubber compounds were obtained according to DIN 53529/3 (Table III). Additionally the specimens (A–F, AC–FC) were thermally aged for 4 h at 150 °C. These conditions should lead to an accelerated aging and thus should give information on the long-term stability of the specimens.
CHARACTERIZATION
Optical microscopy was used for a first study of the wire surface after the squalene experiments. The data were obtained with an Olympus BX60 microscope.
In addition, the morphology of the wire surface was studied with focus variation microscopy. The data were obtained with the infinite focus microscope from Alicona Imaging GmbH. A region of 145 × 110 μm was measured and subsequently planarized (the cylindrical geometry of the wire was subtracted from the three-dimensional image to reduce it to its surface structures), and for better visualization the images are stretched 10-fold in z-direction.
The focus variation data were used to calculate two different surface texture parameters, the mean surface roughness (Sa) and the skewness (Ssk). Sa is defined by the average roughness of a surface topography data array, and it is calculated from
Ssk, also known as “third moment,” represents the skewness of a surface topography data array. Based on a histogram of the heights of all topography array points, the deviation from a normal distribution is represented by Ssk. It is calculated as follows:
Microanalysis was performed by using a Tescan Vega3 scanning electron microscope with an energy dispersive X-ray spectrometer (Oxford Instruments, INCAx-act) attached to it. The electron energy used for the analysis was set to 20 keV. All elements starting from boron can be detected. Elements were normalized proportional to Cu, which was used as an internal standard.
Pull-out testing was performed on a Zwick/Roell Z2.5 universal testing machine. T-test specimens were used to evaluate the adhesion performance by pulling out the wires at constant rate (=100 mm/min) applying a preload of 50 N. Rubber coverage was rated from 0 to 3 (0 = 0%, 1 = 1–49%, 2 = 50–99%, 3 = 100% rubber coverage). The adhesion was also measured after a thermal aging treatment (4 h at 150 °C, data marked with prime symbol (′)).
Elongation at break and tensile strength were measured according to DIN 53504, and tear strength was determined following the procedure described in ISO 34-2.
RESULTS AND DISCUSSION
The so-called squalene method was used to analyze the adhesion layer built up during the vulcanization reaction. This method uses squalene as a low-molecular weight model compound for natural rubber. Brass-plated wires are immersed into a mixture of squalene and all essential curing components at 160 °C for 20 min. After the reaction the wires are washed with toluene, and the newly built up layer on the wire surface can be characterized without further pretreatment.
Optical microscopy images and SEM images of squalene-treated wires are shown in Figure 1. The images show the wire surface after reaction in squalene mixtures with increasing amounts of stearic acid. Optical microscopy images and SEM images in the middle of Figure 1 are of the same magnification, SEM images on the right side are of higher magnification. As shown in the optical microscopy images, the surface changes depending on the amount of stearic acid used. The untreated brass-plated wire is dominated by drawing lines from the wire production process, which can be easily seen in the optical microscopy image as well as in the SEM images. Sample AS (without cobalt stearate and stearic acid) has a yellowish color, but sample BS is bluish. If the amount of stearic acid is increased the color changes from greenish, over yellow to orange. Not only has the color of the wire surface changed, the surface structures have changed as well. This can also be seen in the SEM images. By increasing the amount of stearic acid, increasing amounts of surface structures can be detected. Moreover, the structures seen in the samples are different depending on the amount of stearic acid used. Samples with lower stearic acid content (AS–CS) show drawing lines from the wire as the most dominant structures, in samples with higher stearic acid content dark spots can be seen. Sample CS shows both pronounced drawing lines and dark spots. In samples with a medium amount of stearic acid (CS), the observed dark spots are bigger but fewer than the ones observed in samples with high stearic acid loading (FS). The structures observed in the optical microscopy can also be observed in the SEM images. Drawing lines can be easily seen in the images of the brass-plated wire. Samples with no or low stearic acid content (AS–CS) still show drawing lines but also some spotlike structures, which increase with higher amounts of stearic acid. The diameters of the observed structures are the biggest for samples CS–ES and decrease again with higher stearic acid loading.
Citation: Rubber Chemistry and Technology 85, 2; 10.5254/rct.12.88940
It can be assumed that the stearic acid content has a direct influence on the reaction of the brass with the compound. With low stearic acid content the reaction of the wire seems to be less since the drawing lines can still be seen. If the amount of stearic acid is increased, a spotlike structure can be identified. These spots are fewer but more pronounced on samples with a medium stearic acid amount in the squalene mixture (CS) than on wires, which have been treated in squalene mixtures with higher stearic acid content (FS).
The obtained data suggest two possible mechanisms on the brass surface that explain how the stearic acid loading affects the sulfidation. Firstly, the zinc oxide layer is partly dissolved by the stearic acid and, therefore, the reaction of the brass is accelerated. The second explanation is that because of the higher stearic acid content, more active sulfurating accelerator complexes are created and thus the brass reaction is accelerated.
To prove these assumptions, additional characterization was done by EDX analysis of the wires treated in squalene (wire, AS–FS, see Figure 1). In most cases the specimens consist of a steel substrate with two layers on it: a brass layer, which may contain some additional elements like oxygen, sulfur, and phosphor, and a mainly carbonaceous layer on top. At the electron energy used for the analysis (20 keV) the penetration depth of the electrons is bigger than the thickness of these layers, and in the X-ray spectra there is also the signal from the steel substrate present. Therefore, the analysis volume is not homogeneous, additionally the surface is not flat and therefore any calculated concentrations of elements are roughly approximated values rather than accurate results. Additional errors can result from the fact that some elements, like oxygen or sulfur, can be bound to two layers. For example, oxygen can be bound in the ZnO layer as well as in the carbonaceous layer. But values can be evaluated in comparison with each other to analyze a general trend.
According to the EDX analysis (see Table IV) the surface of the untreated wire consists mainly of Cu, Zn, and O. Fe of the steel core is also detected. Before characterization, the wire was washed with toluene. Nevertheless a carbon signal was detected that probably stems either from residual lubricant or from a contamination caused by handling during the SEM-EDX analysis. Wires treated in squalene mixtures have an additional S peak because of the sulfide layer created during the sulfidation reaction. With increasing amounts of stearic acid, increasing amounts of S are detected, thus confirming the observations made by the optical characterization of the wires treated in squalene mixtures: stearic acid accelerates the sulfidation reaction on the wire surface. The amount of Zn is almost the same for all samples. The oxygen signal is constant within the experimental inaccuracy. With increasing stearic acid content, a growing amount of carbon was detected. We attribute this increase to immobilized organic residues caused by the better interlocking of the cross-linked squalene with the peak shaped surface structure. Sample FS shows increased amounts of Fe. This is caused by a thinner brass layer (probably due to drawing lines) and therefore more steel is detected.
The focus variation microscope is a powerful tool to analyze surface structures. It enables the recording of an entirely sharp three-dimensional representation of a surface structure despite the wire geometry and it provides a good visualization of surface structures as can be seen in Figure 2. Drawing lines as well as new peaklike structures can be seen very well. The comparison of the real color and the color coded focus variation images (Figure 2) shows that the dark spots seen in the optical microscopy images are actually peaks, which is in accordance with Buytaert.23
Citation: Rubber Chemistry and Technology 85, 2; 10.5254/rct.12.88940
The focus variation data were used to perform numerical calculations of the surface roughness. The results are shown in Figure 3. For each compound formulation three similar samples were investigated. As can be seen in the plot, an overall trend can be observed.
Citation: Rubber Chemistry and Technology 85, 2; 10.5254/rct.12.88940
The mean surface roughness (Sa) represents a two-dimensional measure of the texture comprising the surface. It is in analogy to the Ra parameter, which represents the averaged roughness of a one-dimensional profile scan. Sa carries significantly more information about a given surface texture than the Ra parameter. Therefore, outliers have minor influence on the Sa values, and its reproducibility is also better. However, it does not discriminate deep valleys from high peaks, and it is not recommended to use Sa in case of height symmetry texture features. In this case Sa may provide misleading signals. The Ra parameter has been used in the field of mechanical engineering for decades.24 The more powerful Sa parameter, which needs more sophisticated instrumentation for its determination, is about to be established in technical science as well. Overall, the Sa parameter is a good choice for detecting deviations in the texture characteristics.
The calculated Sa parameter (see Figure 3) is the highest for compounds CS and DS. This can be explained by the results obtained through the optical characterization. The structures in the untreated wire and compound BS are mainly dominated by drawing lines, while compounds ES and FS lead to finely distributed peaks. In compounds CS and DS both type of structures, drawing lines and peaks, can be found. The existence of drawing lines and peaks at the same time leads to greater differences in the surface profile and therefore increases in the surface roughness.
Another proof for the results just mentioned is the analysis of the skewness (Ssk). Ssk is a parameter that relates the symmetry of the surface heights to the mean plane. The sign of Ssk shows the predominance of peaks (Ssk > 0) or valleys (Ssk < 0). If the value for Ssk is 0, the surface heights are symmetrically and normally distributed. Since Ssk involves the higher order powers of the surface heights, a considerable amount of measurement is needed to provide statistically significant values. It also needs proper filtering to eliminate erroneous peaks and valleys.25 In Figure 3 Ssk values start from the negative for the untreated wire and turn to the positive for compounds with increasing amounts of stearic acid. Positive values show a tendency for peak-dominated surfaces, while negative values are indicative for valleys at the surface. For that reason, it can be concluded that for the untreated wire the drawing lines are the dominant structures. For BS, CS, and DS, the values are around zero or maybe slightly positive, which means that drawing lines as well as peaks are present as mentioned before. For wires treated in compound ES and FS the value is very high, which means a strong tendency to peak-dominated surface structures.
These roughness calculations explain clearly how the rough surface, which is necessary for mechanical interlocking, forms during the vulcanization process. Furthermore, it is an additional explanation for the fact that there is an optimal thickness of the sulfide layer. As can be seen in Figure 3, the roughness has a maximum and decreases again with stronger reaction of the brass.
However, not all aspects concerning the adhesion between rubber and brass can be investigated by squalene experiments, since the created interface is only an artificial one. To correlate findings from the squalene experiments with actual values in rubber products, two series of analogous rubber compounds (containing natural rubber instead of squalene) were prepared. The first one had the same composition as the squalene mixtures (compounds A–F), and the second one additionally contained carbon black (AC–FC) to get better comparability with real rubber compounds. Tensile strength, elongation at break, tear strength (see Table V), and pull-out forces (see Table VI) of these compounds were tested and compared with the results obtained from the squalene experiments. Thermally aged values for tear strength could not be obtained for compounds without carbon black because the specimens were not mechanically stable enough after the aging treatment.
Carbon black is used in rubber products as reinforcing filler. Therefore, it is not surprising that compounds containing carbon black have higher values for tensile strength and tear strength.26 Values of unaged compounds are higher than aged ones (see Table V). For compounds containing carbon black (unaged and aged), tensile strength slightly decreases with increasing amounts of stearic acid. Compounds without carbon black show a contrary trend for the unaged specimens: the tensile strength increases with increasing amounts of stearic acid.
Elongation at break for compounds without carbon black does not show a consistent trend with the loading amount of stearic acid. For compounds that contain carbon black, the elongation at break decreases with increasing content of stearic acid. This effect can be attributed to a higher cross-link density due to the activating effect of stearic acid.17,18 Values for tear strength are approximately of the same dimension within a testing series (A–F, AC–FC, AC′–FC′) but do not show a consistent trend with stearic acid loading in any of the compound formulations.
Pull-out forces and rubber coverage are shown in Figure 4 and Table VI. No adhesion data could be obtained for thermally aged compounds without carbon black, because the specimens were not mechanically stable after the aging treatment. Compounds containing carbon black yielded higher values than unfilled compounds. In most cases level 3 coverage was observed, which means 100% rubber coverage of the wire. Therefore, it can be concluded that in these cases the adhesive strength exceeded the cohesive strength of the rubber and that the adhesion values obtained were only affected by the rubber properties. Filled unaged rubber compounds (AC–FC) have coverage levels between level 2 and 3 (level 2 = 50–99%, level 3 = 100% rubber coverage). This means that adhesion partly fails in cases with level 2 coverage. Therefore, these pull-out forces might be seen as real adhesion values. Aged filled compounds (AC′–FC′) have all coverage level 3. Thus it can be concluded that it is mostly the rubber that is the weak part, whereas the adhesion interface does not break.
Citation: Rubber Chemistry and Technology 85, 2; 10.5254/rct.12.88940
Compounds without Co salt (A, AC, AC′) showed lower values than comparable compounds with Co (B, BC, BC′). For compounds containing carbon black (aged and unaged), the highest values are obtained by compounds with low stearic acid contents (BC, BC′). These compounds contain a small amount of stearate due to the Co stearate, which acts as an adhesion promoter.
In the case of the unfilled compounds, highest pull-out forces were achieved with compound E (1.5 phr stearic acid). Coverage for all values was level 3. The comparison of the pull-out forces with tensile strength values shows no correlation, but the comparison with tear-growth data shows a similar trend. The tear-strength values for unfilled compounds are quite low compared with the filled compounds. It seems as if tear growth is the weak point in these compounds. Therefore it is rather the rubber that fails than the adhesion layer, which is also shown by a rubber coverage level of 100%.
Data obtained by pull-out testing suggest that very small amounts of stearic acid are good for the adhesion. Unfilled compounds have the best adhesion values for samples with 1.5 phr stearic acid, but coverage is 100%, and hence it is more likely that this results from the rubber properties of the sample. Unfortunately it is very likely that in most cases the adhesion values observed are due to the rubber properties because the adhesive strengths exceed the cohesive strengths of the rubber compounds. Thus it is not possible to create a correlation between the data obtained through pull-out testing with the results from the squalene experiments.
CONCLUSION
The loading amount of stearic acid has a great influence on the sulfidation reaction of brass-plated steel wires with rubber compounds during vulcanization. Increasing amounts of stearic acid accelerate the reaction of the rubber compounds with the brass-plated steel wires, which can be easily visualized by optical microscopy. The enhanced sulfidation with increasing amount of stearic acid can be attributed to two different effects: first a partial dissolution of the ZnO layer on top of the brass surface, which leads to a faster reaction of the brass, and second an activation effect. The higher concentration of formed zinc stearate leads to a higher concentration of the active zinc-accelerator complex, and thus the sulfidation reaction is enhanced.
Investigation of roughness parameters showed an increasing amount of peaks and a vanishing of drawing lines with higher loading of stearic acid. Furthermore, it was shown that there is an optimal amount of stearic acid (1.0 phr stearic acid) to achieve maximum surface roughness, which is important for good mechanical interlocking.
Variation of the stearic acid loading has a direct influence on the bonding interface but at the same time also a strong influence on the rubber properties. Therefore a direct comparison of pull-out forces with rubber properties (tensile strength, elongation at break, tear strength) can help to separate these effects. Best pull-out forces were achieved with compounds with low stearic acid content. In most cases the rubber coverage was level 3; therefore, it was concluded that the measured adhesion values can be attributed to the rubber properties of these compounds, as can be seen by comparison with tensile strength and tear strength. As a result it is not possible to directly correlate the results from the adhesion testing with the results obtained by the squalene experiments. For future work, it would be helpful to investigate a rubber system where the cohesive forces of the compound exceed the adhesive forces between rubber and brass.
Optical microscope images (left) and SEM images (middle and right) of untreated wire and wires after the squalene treatment in compounds AS–FS.
Planarized focus variation microscope image in real color (left) and color coded (right) of wire after squalene treatment in compound CS.
Mean surface roughness (Sa) and skewness (Ssk) of the untreated sample and samples after squalene treatment in compounds B–F.
Pull-out force of rubber compounds with different stearic acid contents.
Contributor Notes