TMA has already been considered in conjunction with glass transition measurements in applications such as degree of cure or heat-distortion temperatures of epoxy resin castings. The technique has proved useful for many other applications, some of which are described below.

• Investigation of extrusion defects

Accordingly, prior to the DSC experiments described above, the samples were placed in DSC sample pans, and the lids rested on the samples, but not crimped, and the TMA load then applied. This allowed the TMA experiment to take place on the uncross-linked sample and then the same sample in the same pan to be used for the DSC experiments where composition investigation was of more interest. Data were obtained by heating at 10 K/min between ambient and 200◦C and the TMA furnace purged with nitrogen throughout. A load of 110 mN was applied to the DSC pan lid for each measurement. The height was zeroed using an empty pan and lid.

It will be recalled that no contaminant could be demonstrated to be present, based on the DSC curves, and that the sample had not been subjected to any cross-linking process. The data in Figure 1 show that the matrix yielded freely at the melting point but that the sample containing the defect did not. Since the defect was concluded not to be a contaminant, based on the DSC evidence, it was possible to conclude that it was a pre-cross-linked gel and thus able to resist the deformation that the relatively light probe force could impart to a fully molten sample of uncross-linked matrix.

Figure 1. TMA response of defect (1) and matrix (2) for samples giving the DSC response

In the case of the second fault, in which DSC indicated a contaminant, no significant difference in TMA response was found between matrix and fault position, and this led to the conclusion that this fault was not a consequence of pre-cross-linking.

• Cross-linking

The cross-linking of materials such as polyethylene, ethylene propylene rubber and some other rubbers used for cable insulation and sheathing is conventionally measured for quality purposes by means of a test of elongation under a specified load at elevated temperature of the permanent set measured when the load is removed and the sample allowed to cool. This hot-set test is very good when it is possible to cut appropriately sized tensile test dumb-bells for the measurement. Where samples are irregular and small, as may be the case in work of a more investigative nature, it is much more difficult to apply, and TMA has been applied, albeit in a fairly qualitative way, to the problem of finding out if a specimen is cross-linked.

The data of Figure 1 are suitable to illustrate this point; the uncross-linked matrix yielded completely in this experiment, whereas that containing the pre-cross-linked defect did not yield completely. Under the experimental conditions described, well-cross-linked materials used for cable manufacture yield only slightly at the crystalline melting point and so the TMA measurement yields a useful qualitative result, with some possibility of commenting on the gross degree of cross-linking. The ability to observe this yield point in the TMA is however useful on occasion, and TMA experiments on material known to be cross-linked, carried out with greater loadings, have proved extremely useful in this respect.

It would be possible to obtain more quantitative data by measurement of gel fraction by solvent extraction of the material in order to effect calibration but this is a time-consuming process and for a single measurement it would render the TMA measurement redundant. If the TMA data are considered likely to prove fit for purpose, they can be obtained easily and quickly, an advantage in an environment where information may be required on an urgent basis.

• Material identification

Some of the materials commonly used in cable making do not give very clear crystalline melting points in the DSC experiment. This was remarked upon with respect to the data of Figure 1. It will be observed that the softening points in Figure 1 are very clear, so penetration measurements have proved very useful for melting point determination of materials with low crystallinity and with heavy filler loadings and also for gaining additional confidence with very small samples of the type described earlier.

As remarked earlier, greater loadings are required for cross-linked samples than for the corresponding uncross-linked material in order to achieve sufficient penetration to adequately detect the crystalline melting point with confidence. TMA data would typically be used in conjunction with some or all of FTIR spectra, DSC responses and also data from thermogravimetric analysis (TGA) as appropriate, in order to more fully characterise an unknown.

A knowledge of the properties of the materials available, or provision of reference samples where some material fault is suspected, greatly facilitates this type of work. It is a great advantage, therefore, to log all thermal analysis experiments and to archive all data for future reference. These will, at some time, amply repay the analyst in interpretation of another investigation and may save much time in collecting reference thermograms, a factor of considerable importance in matters of production or customer support.

• Extrusion studies

Cable production involves extrusion of insulation and sheaths in long, continuous lengths. For technological reasons, both electrical and mechanical, as well as aesthetic considerations, these extrusions have to be of good quality and smoothness is important. TMA studies have been valuable into providing insights into the differences in behaviour of material as a support to extrusion trials and rheological measurements. As an example, consider the data illustrated in Figure 2, which were obtained from two fire-retardant cable sheaths.

Figure 2. Penetration data for two grades of cable sheath using 1mm diameter probe and 110mN loading. Heating rate 10 K/min with nitrogen purging throughout.

It is clear that these materials are likely to have differing extrusion properties and will, to achieve satisfactory finishes, require different processing conditions, such as extrusion head pressure. A higher viscosity may affect output since it will need to be run at a slower speed than the lower viscosity material if increasing extrusion pressure is ineffective or not possible. It should be noted that the extrusion temperature will need to be kept well below the onset temperature of any polymer or fire-retardant decompositions, otherwise there is a risk that even a small degree of decomposition will produce gas bubbles and hence porosity into the extrudate. Such porosity is extremely deleterious to the final product properties and cannot be tolerated.

This type of TMA experiment can be valuable in screening out experimental formulations which are unlikely to succeed, prior to moving to other material evaluations and, more importantly, to production trials. Such screening can lead to time saving and considerable cost reductions for the manufacturing plant, since any failed trial produces only scrapped material. This approach has also been used with some success in combined evaluations and will be discussed in more detail later.

• Fire-retardant mineral insulations

Some types of cables are designed to give continued service under fire conditions, so that services such as fire-detection systems and emergency lighting are maintained until evacuation is completed. Common approaches to the provision of such a cable include that of using silicone rubber, which burns to leave an insulating residue, or mica paper backed with woven glass fabric and bound in resin (‘mica tape’) as at least one layer of the insulation. The behaviour of these materials at elevated temperatures is therefore of considerable interest.

One investigation centred on the comparative performance of different grades of mica tape and the melting and sintering temperatures of the mica and glass was of particular interest.

To facilitate the investigation the mica tapes were solvent extracted to remove the binder and then dried. The mica and glass were separated by hand and cut up into squares approximately 4 mm on each side. Small piles of these squares were placed into copper DSC pans and a second, smaller, lid placed over the samples. These precautions were necessary to avoid adhesion of the sample to the quartz TMA probe and sample support. The height reading for the TMA probe was zeroed with the pans in place but without the sample, and a blank experiment, without sample, was run to verify that the pans would not contribute spurious transitions to the experiment. The TMA was run from ambient to 900◦C with an air purge and heating rate of 10 K/min; the applied load was 100 mN. The curves given in Figure 3 are representative of those obtained.

Figure 3. Deformation curves for glass-woven tape (1) and two different varieties of mica paper.

Inspection of Figure 3 shows that the glass tape yields somewhat up to 300◦C and begins to melt at 650–700◦C. The two micas, which have different compositions, do not melt but one begins to expand at around 700◦C, whist the other expands minimally and then begins to contract at about 850◦C. These temperatures are those at which the two types of micas are reported to sinter and are therefore of interest in terms of the fire process because it is at these temperatures that they will ceramify well, further stabilising the high-temperature insulation.

Silicone rubber burns in air and, if the formulation includes appropriate fillers, will produce a hard insulating ash. A TMA experiment to simulate burning yields interesting results when carried out with a 1-mm-diameter probe, a load of 100 mN, air purge and a heating rate of 10 K/min. Figure 4 shows the result of one such experiment.

Figure 4. Burning of silicone rubber showing sample movement as the temperature increases.

As expected, the rubber expands up to about 300◦C and then begins to contract until, at about 400◦C, there is a rapid expansion which can be correlated with the onset of decomposition processes observed by thermogravimetry. As the rubber continues to decompose and burn, it eventually contracts again and then slowly stabilises. This behaviour is of interest in terms of understanding the interaction of the silicone with other materials in the cable as it begins to burn and also in terms of the eventual development of the mineralised insulation required for fire service.