|
Productive and reliable techniques and probes are required for thermometiy of semiconductor crystals in plasma
processes of microtechnology. A technique may be considered as productive when the time one takes to perform a preliminaiy
to the temperature measurement is negligible as compared with the duration of the etching or deposition itself. The reliability
of measurement may be achieved when it is directly measured the temperature of semiconductor wafer and not of holder or
test satellite.
Interferometric technique for temperature measurements of glasses' and semiconductor crystals57 is used in many
laboratories. This technique holds much promise for noncontact thermometry of semiconductors since the temperature
coefficient of refraction n1dn/dJ' is almost two orders of magnitude greater than the thermal expansion coefficient h'dh/dt
(here n is refractive index, h is thickness, T is temperature of wafer). Semiconductor wafer, on being heated to several tens of
degrees, changes its optical thickness nh many times as compared to wavelength ? of sounding light: A(nh)>>A. In
consequence of this, it possible to register numerous interference maxima and minima which are used for calculation of
tempera! changes in temperature. The phase shift of 2t (one interferogram fringe) for 500 pm thick silicon wafer irradiated
by HeNe laser (1. 15 pm) is caused by the temperature change of AT=5.2 K when T=300 K and AT=3.4 K at 650 K.8 The
time it takes for the optical setup to be adjust completely after loading of a wafer into reactor is approximately equals to
20+30 S.9 In this paper fluther advancements of the interferometric laser thermometry towards total elimination of any
adjustment procedure are presented. In addition, we observe a distinct increase in the temperature when a wafer is placed onto
cooled holder and is heated by RF plasma. Furthermore, it has been found and explained an anomalous fringe shift in SF6
plasma.
|
