Logo der Physikalisch-Technischen Bundesanstalt

4,4'-Azoxyanisole for temperature calibration of differential scanning calorimeters in the cooling mode - yes or no?


It is generally accepted that when you are performing differential scanning calorimetry (DSC) experiments in the cooling mode, for reaching the best possible measurement uncertainty it is advisable to perform the calibration of the instrument also in the cooling mode. Liquid crystals have found some interest for the calibration of DSCs in the cooling mode. It has been stated that some of the transitions within their phase systems are of higher order and are, thus, not subject to the phenomenon of supercooling. This is a desirable feature of calibration substances for use in the cooling mode. The problem is to unequivocally distinguish phase transitions of first order from those of higher order by the methods usually available to thermoanalysts. A further requirement imposed on the combination of instrument and calibration substance is a sufficiently good resolution. This means that in case where the substance shows two transitions close to each other the instrument must be able to separate the two peaks completely and reach the baseline between them.

A thermodynamic criterion for the supposed order of the transition is the heat capacity, coefficient of expansion and compressibility vs. temperature function. 4,4'-Azoxyanisole has been one of the earliest well characterised and interpreted liquid crystal with nematic to isotropic phase transition. The shape of the curves published show a Lambda like appearance. The conclusion should be that this phase transition is of higher order. Investigations of this phase transition by DSC, however, show a first order behaviour: By measuring the phase transition peak at different heating rates the peak height increased by a factor of  Ö2 when doubling the heating rate. This is a clear indication of a first order phase transition which is accompanied by a phase transition enthalpy. With phase transitions of higher order, this factor should be 2 because without enthalpy change the apparent heat capacity of a sample should double when doubling the heating rate. On the contrary, the Landau theory yields an order of the transition which can only be described as "weakly first order".

4,4'-Azoxyanisole has been recommended as a temperature calibration material for the cooling mode. The published peak onset temperatures show a nearly perfect linear dependence on the heating and cooling rates, each exhibiting the same slope and virtually no step at zero heating rate.

It is clear from the literature and our results that the nematic to isotropic liquid transition of 4,4'-Azoxyanisole is not a phase transition of purely first order but a so called weakly first order transition with a specific transition enthalpy Dtrsh of approx. 3 J/g. The experiments show that there is nearly a linear dependence in the peak onset temperatures on higher heating and cooling rates, resp. The step at zero heating rate is almost negligibly small. However, the slopes of the two straight lines differ from experiment to experiment.

These findings emphasize that phase transitions in liquid crystals are often more complex than expected. It is, thus, recommended that the choice of any liquid crystal as one perfect material for temperature calibration of DSC on cooling has always to be done with great care. Similarly, the theory behind the operation of DSCs seems more complex than often anticipated and, obviously, still deserves further elucidation. The same applies to the evaluation algorithms and procedures hidden in the software of the manufacturers of the calorimeters.