A root cause failure investigation was carried out on buffer rubbers. The buffer rubbers, onto which a vibrating screen are mounted, were subject to premature failures within several hours of operation. The vibrating screen separates the ore that it carries into different size ranges. The vibration of the screen is driven by three gearboxes. The rubbers are therefore loaded by the weight of the screen and the weight of the of the ore product that it is separating as well as the load that the gearboxes supply to the rubbers for vibration.
The failure of the rubbers resulted in shut down time for the separation process. Failure of the rubbers means that the screen must be shut down and bypassed which has severe consequences for the process.
Fig 1: The buffer rubbers showing uneven bulging and to a higher extent than is expected.
Three samples were provided, see Figure 2. One unused new rubber, one rubber that showed irreversible bulging and one that was charred from spontaneous ignition.The samples were specified as being composed of natural rubber (polyisoprene); the chemical structure is shown in Figure 3. This was said to be been vulcanised which is the process of crosslinking polyisoprene chains with sulphide bridges, see Figure 4.
Figure 2: Three samples provided. From far left: unused, uneven bulging and charred sample.
The rubber and the reactants are mixed and cured in a mould at high temperatures. The process of crosslinking relies on a chemical reaction between sulphur and the double bonds of the isoprene moieties. It is dependent on the amount of sulphur, accelerator, temperature and time. The crosslinking of the rubber is directly linked to the mechanical properties of the material and therefore its appropriate end application. If the reaction is not allowed to reach completion (curation) then the rubber will not have the full properties predicted. Equally important is the starting mixture which determines the end crosslinking density.
Figure 3: Structure of polyisoprene (natural rubber).
Figure 4: Structure of vulcanised polyisoprene.
To verify the identity of the rubber, in other words the nature of the polymer chains in Figure 3, the following characterisation tests were carried out: FTIR, DSC and TGA. Each of the tests are explained in Section 2.1, overleaf.
Once the polymer identity was established, compression tests of the sample were carried out on the unused rubber to test how much load the sample could handle and measure the distance the rubber deformed/depressed at the specified load 354 460 N (calculation below). Further MDR analysis was carried out to determine if the crosslinking reaction had fully cured. I.e. if the sulphur bridges seen in Figure 4 had formed with the available sulphur.
Full load = 42 166.70 kg
Load per rubber (/12) = 3 513.89 kg
Compression force (x 9.807) = 34 460.72 N
2.1 Verifying natural rubber identity
2.1.1 Fourier Transform Infrared Spectroscopy (FTIR)
FTIR spectroscopy (Fourier Transform Infrared Spectroscopy) is a technique that uses infrared light to observe properties of a solid or liquid. It is used in many different applications to measure the absorption, emission, and photo-conductivity of matter. This is done by shining a narrow beam of infrared light at the matter in various wavelengths and detecting how the different wavelengths are absorbed or transmitted. This response corresponds to the types of bonds and functional groups present. Once the data has been obtained, it is converted into digital information using a mathematical algorithm known as the “Fourier transform” and is used in conjunction with other techniques to identify compounds. The instrument used was a Perkin Elma instrument with ATR capabilities with a diamond crystal.
2.1.2 Differential Scanning Calorimetry (DSC)
Differential Scanning Calorimetry (DSC) is a technique that measures the difference in the heat flow to a sample and to a reference sample as a direct function of time or temperature under heating or cooling. DSC analysis is used to measure melting temperature (Tm), heat of fusion, latent heat of melting, reaction energy and temperature, glass transition temperature (Tg), crystalline phase transition temperature and energy. By comparing two different samples’ DSC curves, it can be established whether the polymers have the same melting points or decomposition points and crystallinity and therefore whether they are the same. The rate of heating was set at 10 °C/min.
2.1.3 Thermogravimetric Analysis (TGA)
Thermogravimetric analysis (TGA) is an evaluation technique that measures mass changes of materials as their temperature is changed or at a constant temperature over a given time. It is used to analyse decomposition and evaporation rates, oxidation, material purity and many other properties. In this instance it was used to evaluate the amount of fillers, activators and other additives in the rubber mixture. The rate of heating was set at 10 °C/min.
2.2 Compression and Hardness Testing
Compression testing is a very common testing method that is used to establish the compressive force or crush resistance of a material and the ability of the material to recover after a specified compressive force is applied and even held over a defined period of time. The rate of compression was set to 1mm/min to minimise any heat generation that could damage the rubber. The Shore A hardness was also measured using a durometer.
2.3 Moving Die Rheometer (MDR) – Cure properties
Moving Die Rheometer (MDR) analysis was carried out to determine whether or not the rubber was fully cured and all the sulphur bridges (shown in Figure 4) were given enough time to form. MDR measures the change in stiffness of a rubber sample as over time at 170 °C (curing temperature).
The sample is compressed between two heated plates and by an applied oscillating force. The torque shows the conversion in physical strength of the rubber, increasing as the cross-linking reaction takes place over time. The cure rate curve is the derivative of torque over time, which is an indication of the extent of the curing reaction.
Approximately 9 grams of the sample was cut and subjected to isothermal curing at 170 °C, with an oscillation angle of 0.5 ° at a frequency of 1.67 Hz (100 rpm), on a D-MonTech RPA 3000. The changes in torque (S’) and cure rate (dNm/min) were monitored.
2.4 Swell Tests – Crosslink Density
To determine the extent of crosslinking in the sample, swell tests were carried out. The rubber is immersed in solvent (toluene) and the mass changes are measured. This, along with known parameters, are used as inputs into the Flory-Rehner equation. The equation describes the mixing of polymer and liquid molecules according to the equilibrium swelling theory which considers entropy and heat changes.
Where:
Vr = Volume fraction of rubber in swollen vulcanizate
Mr = Mass of “dry” rubber (after solvent removal)
Dr = Density of polyisoprene
Ms = Mass of solvent
Ds = Density of solvent (toluene)
Where:
Mc = Average molecular mass of network chains between adjacent crosslinks (g/mol)
Vms = Molecular mass of solvent
χ = Polymer solvent interaction parameter (toluene and polyisoprene)
3.1 Verifying natural rubber identity
3.1.1 Fourier Transform Infrared Spectroscopy (FTIR)
The rubber contains carbon black which has a high index of refraction. This may approach or exceed the ATR crystal (diamond) resulting in distorted spectra, as seen above (Figure 5).1 Carbon black is commonly used as a reinforcing filler which sits in the rubber matrix and improves tensile strength and wear resistance.2,3 The two peaks between 2800-3000 cm-1 indicate the presence of CH2 and CH3 groups typical of any hydrocarbon including polyisoprene, see Figure 3.
Figure 5: FTIR spectrum of unused rubber.
3.1.2 Differential Scanning Calorimetry (DSC)
Figure 6: DSC plot of the rubber sample.
The DSC result in Figure 6 shows a typical polyisoprene pyrolysis point of approximately 370 °C. At this point sulphide bonds and polymer bonds start to be broken. This temperature is a signature property of polyisoprene which is used to characterise the material.
Figure 7: TGA plot of the rubber sample.
3.1.3 Thermogravimetric Analysis (TGA)
The TGA results in Figure 7 show what happens to the sample mass as it is heated over time. The initial loss of 3.53% can be attributed to the loss of volatile compounds such as softeners, activators or residual accelerators. The larger loss of 61.09% is the degradation of polyisoprene (above 370 °C). The remaining mass of 35.38% can be equated to the amount of carbon black and other additives that were added.4 This profile shows the various degradation components of the rubber mixture as the sample is heated. This is a critical analysis which should be repeated on a rubber which has the mechanical properties for comparison.
3.2 Compression and Hardness Testing
Table 1: The results from the compression testing. Showing the maximum compressive load and the distance the rubber was depressed under this load.
| Maximum Compression Load (N) | Distance the Rubber Height decreased/was compressed (mm) |
| 25 890.23 | 30.50 |
The buffer rubber is produced to be able to withstand 34 460 N load (refer back to Section 2 for calculation) with a maximum of 15-20 mm depression of the rubber height. The above results in Table 1 show that the rubber height was decreased almost double the maximum allowance under a load of 25 900, less than what is actually applied in operation. This indicates that the stiffness of the rubber is not sufficient to withstand the load applied in practice.
The Shore A hardness measured between 63 and 71 °A over various locations on the sample. This is within the specification from the supplier of 65 °A but shows some variance across the rubber. The failure of the compression along with the varying hardness speaks to the crosslinking of the polymer which may not be enough for this application.
3.3 Moving Die Rheometer (MDR) – Cure properties
The graphs below show no significant increase or peaks indicating that no additional curing took place. This means that the full extent of the crosslinking made possible by way of the mixture had occurred. The rubber was fully cured and this was not the cause of the failure.
Figure 8: The torque graph for the rubber sample.
Figure 9: The cure rate graph for the rubber sample.
3.4 Swell Tests – Crosslink Density
Using the theory and equations laid out in Section 2.4, crosslink density was determined. The results are laid out in Table 2.
Table 2: Results of the swell tests using the Flory-Rehner equation.
| Property | Value |
| 1/2 Mc | 0.00050 |
| Mc | 997.55 |
The crosslink density depends on both the initial rubber mixture and the curing treatment. The mechanical assessment indicates that the crosslink density is not sufficient. This test should be repeated on a sample with sufficient mechanical strength as a comparison.
- The characterisation tests showed that the samples were indeed polyisoprene rubber (natural rubber). The possibility of the rubber being of the wrong material was eliminated and thus it is not the cause of the failure.
- The compression test showed that the rubber cannot withstand the load that it is subjected to in practice. It deforms more than it should (up to 30.50 mm under 25 890.23 N) when subjected to simulated conditions. This causes excessive bulging over time and results in the sides of adjacent buffers rubbing together. The charred sample is charred from the outside inwards which confirms that the friction of the touching rubbers causes heat to be generated and the polymer to ignite and burn.
- The hardness of the rubber is not consistent across the sample which is an indication of inconsistent crosslink density distribution in the polymer.
- The mechanical failure together with the varying hardness leads to the conclusion that there is insufficient crosslinking present. This is due to the initial mixture containing the sulphur, accelerators, activators and other additives.
- The MDR analysis showed that the extend of crosslinking, made possible by the initial mixture, had occurred and the curing was not the cause of failure.
- Cross-linking density can give an indication of the possible mechanical properties. Since the desired mechanical strength was not achieved, it can be concluded that the crosslinking density was not sufficient, and this should be compared to a rubber that shows the desired mechanical strength.
The cross-linking density of the polymer determines the mechanical properties including tensile and compression strength. It is therefore recommended that the supplier and manufacturer be contacted to ensure all the above factors that affect crosslink density have been optimized for the intended use of the rubber.
It is recommended that full analysis of a rubber with the desired mechanical properties be carried out as a comparison. These tests should include TGA, DSC, compression and hardness testing as well as swell tests to compare the crosslink density.
It is further recommended that any change in supplier or manufacturer is supported by comparative analysis to ensure the quality of the rubbers are matched.
(1) Bradley, M. Carbon Black Analysis Comparison with Diamond ATR and Germanium Crystals. Thermo Electron Corporation 2005.
(2) Gent, A. N. Engineering with Rubber, 3rd ed.; Gent, A. N., Ed.; Hanser Publications, 2012; Vol. 59.
(3) Fukahori, Y. Carbon Black Reinforcement of Rubber (1): General Rules of Reinforcement. Int. Polym. Sci. Technol. 2004, 31 (8).
(4) Matador. Test Methods of Rubber Materials and Products. Matador Rubber S.R.O. 2007.
