In-situ Real Time Spectroscopic Ellipsometry Applied to the Surface Monitoring of Semiconductors
- PDF / 2,716,493 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 16 Downloads / 201 Views
1025 Mat. Res. Soc. Symp. Proc. Vol. 452 01997 Materials Research Society
DESCRIPTION and PERFORMANCES of the RTSE SYSTEM General description of the system The SOPRA RTSE instrument is a rotating polarizer instrument especially designed for easy adaptation to deposition or plasma treatment systems. The polariser and analyser arms can be attached directly to UHV flanges with easy adjustment by an optical collimator. The angle of incidence is fixed by the setup but can be determined accurately during the calibration procedure. A 5mm beam diameter is defined from an Xe lamp with a system of mirrors and passes through a quartz Rochon polarizer. It can be parallel or can be focused on the sample surface with a spot size lower than 1mm2 . The reflected beam passes through another Rochon analyser and is focused onto the entrance of an optical fiber connected to a spectrograph. Multichannel analyser detection The optical fiber introduces the light inside a spectrograph with a silica prism for quasi linear dispersion versus energy. The dispersed light is focused on a low amplified intensifier and detected with a multichannel silicon photodiode aray. After amplication the signals are transmitted to a PC through a Digitial Signal Processor (DSP) card. The DSP allows control of all the steps of the measurement including non linearity corrections, averages and precalculations. During a RTSE measurement the processor of the PC is in fact only used for the transfer of the precalculated data (tan V and cos A or pseudo optical constants ). In this way, data analysis can take place on the previous measurement during the current acquisition. Signal/noise ratio of the intensified photodiode array detector The intensifer in front of the silicon photodiode array is useful to increase the sensitivity in the UV range. Nevertheless, the amplification must be carefully adapted to optimize the signal/noise ratio of the detector. We have measured this parameter for all the pixels of the detector simply by making the same measurement thirty times using the emission spectra of the xenon lamp through the ellipsometer arms in straight line without sample. The noise evaluated by this method is reported versus the signal detected by each pixel in Figure 1. As shown in the figure, the signal/noise ratio is always higher than 25 even for the pixels with the lower signal level. It can reach values of 100 for the highest signals available (saturation at 65536 or 16 bits of precision of the photodiode array ). These results have been obtained with Is of integration time. If the speed of the measurement can be reduced, the signal/noise ratio can of course be improved by simple averaging. Reproducibility and precision of the measurement The precision and reproducibility of the RTSE instrument have been evaluated simply by making different measurements in straight line without sample. The precision is around 0.3% both for tan Psi and cos Delta parameters. The reproducibility is three times better around 0.1% for Is of integration time as shown in Figure
Data Loading...
