Why do we need both thermodynamic and practical temperature scales? (FAQ - Engineering Measurements)
Although the concept of temperature is familiar from everyday life, and various kinds of thermometer have been used in science, industry, meteorology and medicine, etc. for over 300 years, it was only with the development of thermodynamics in the mid nineteenth century that an understanding of temperature and its relationship with heat was fully developed. Specifically, it was realised that a 'thermodynamic' temperature could be defined independent of any particular material properties, such as the expansion of mercury or the resistance of platinum wires. Fundamental physical laws such as the gas laws and the laws of thermal radiation relate observable properties (pressure, radiant intensity, etc.) to this temperature, and it follows that these laws can be used to produce a 'thermodynamic' scale which has universal applicability in science and engineering.
While this has real physical significance, unfortunately the measurements are extremely difficult and time-consuming, and they are only rarely done even at National Measurement Institutes like NPL. For practical purposes temperature measurements are based on the International Temperature Scale of 1990, ITS-90. This uses more convenient thermometers such as platinum resistance thermometers, which are capable of very high precision and reproducibility, better than 0.001 °C for much of the range. The numerical values assigned in the ITS-90 were chosen using the best thermodynamic measurements made at NMIs, so that it gives temperatures which are very close to the corresponding thermodynamic values. The ITS-90 can therefore be used as a substitute for thermodynamic measurements, but it also provides the high precision and reproducibility which are needed for standards and in calibration services.