USE OF THERMO-MECHANICAL ANALYZERS(TMA)
FOR DILATOMETRIC MEASUREMENTS

(Technical Note #58)

A Thermo-Mechanical Analyzer (TMA) is usually a part of a thermal analysis group of instruments (TGA, DTA, etc.) that are aimed more toward the chemical behavior than physical properties of materials. A thermo-mechanical analyzer, as its name implies, was developed for mechanical properties determination of plastics, such as load bearing when heated, penetration, bending, etc. It uses a small sample and it usually comes with a variety of shaped tips to do the above. At some point it was an obvious extension to try measuring expansion coefficients as well, after all the device has a sensitive transducer. This is reasonably possible with plastics where the expansion coefficient is in the 100-300 X 10^-6/C range, and a few mm thick sample will expand quite a bit. However, indiscriminate extension of its use to include essentially any material is inappropriate. Metal samples, for example, will expand about 1/10 of plastics, and some ceramics only 1/100. This means that a TMA that has a ±1% accuracy for a plastic sample can have a ±100% inaccuracy for a ceramic sample. This can neither be calibrated out nor answered with higher amplification of the signal.

Calibration is usually done by mechanical means, such as gage blocks, micrometer screws, etc., and these devices have their own limitations in the small displacement region. A micrometer, for example, can be read only to a certain limit even with a vernier scale, and this limit may be 1 to 10% of the entire displacement range of the transducer. This will produce a very high uncertainty in the numbers. Any uncertainty in calibration will carry over in a direct proportion to the expansion coefficient that is measured. The scatter, due to poor resolution, will produce an identical scatter in the final coefficient values.

A TMA nearly always employs an LVDT (linear variable differential transformer) or a capacitive transducer, both of the high sensitivity, short stroke type. These devices have a narrow range within which they are linear to some degree. Both types are prone to drift caused by amplifiers and thermal effects, and therefore they require frequent recalibration. Dilatometers usually use longer stroke LVDTs which are somewhat easier to calibrate and are more stable, or in the case of Unitherm™ dilatometers, absolute digital displacement transducers are employed which do not require periodic calibrations at all.

One major difference between a TMA and a dilatometer is the size of the sample. Normally, dilatometers employ 25 to 50mm long samples as opposed to the 1 to 2mm for TMA, thus the amount of expansion for a dilatometer sample is twenty-five times more. This contributes substantially to the measurability and reliability of the data and therefore to the overall accuracy of the measured coefficient.

In conclusion, a TMA is not a dilatometer, and it should not be used to generate thermal expansion coefficient data on materials other than polymers.