Laser interferometry measures thermal strains and CTE by determining the change in a test sample when heating, by directing a laser beam through the centre of the sample and reflecting the beam off sapphire optical flats on the bottom and top of a sample.
The two reflected beams of laser light are now out of phase (due to the path difference between the two reflective surfaces) and interfere constructively and destructivity to produce a ‘Newton rings’ interference pattern, that is projected onto a screen above the sample.
As the sample is heated and the path difference between the optical flats changes, the interference pattern changes. These fluctuation in the interference pattern are recorded, from which the path difference changes and sample change in length can be measured.
This method can be used to determine the change in sample length as small as 1/100th of the wavelength of light, allowing for very accurate thermal strain characterisation on even very thin samples (as thin as 0.5mm). This approach require high precision sample geometries. The method is also time consuming and therefore more expensive, but is ideal for low thermal expansion or small sample testing (e.g. surface coatings, oxides). For bulk material characterisation push rod dilatometer is likely to be more appropriate.
This method of CTE measurement is relatively simple: a sample is placed in a furnace with one end against a ceramic insert, and a ceramic push rod is held against the other. At the other end of the ceramic rod (outside the furnace) is located a displacement transduce that measures the changes in sample length.
To measure CTE and thermal strains the sample is heated at a prescribed rate (to ensure uniform heating) with the temperature and change in sample length recorded through out. From these measurements CTE and thermal strain values can be determined.
To reduce errors in the system: the furnace chamber is purged with a high purity inert gas to prevent oxidation; the ceramic holder and push rod are made of the same material to reduce the effect of these components expanding when heated; and a calibration run is always performed in the system with a similar size sample, of known CTE, to provide a correction curve for the expansion of the other components in the measurement apparatus.
This approach is relatively cheap to perform and can be done quickly but it does require a sample of at least 15mm in length to allow accurate determination of thermal strains. Smaller sample can be measured but with an increased uncertainty on the results. It is a good technique for bulk material characterisations.