Position sensor microsystem based on Hall devices

Project description :
The test structures we have designed consists in a wheatstone bridge in the n-type substrate, patterned with deep p-type diffusion (Figure 1). Two out of four resistances are covered with metal, and therefore light insensitive. A current is driven in the bridge (0.5 to 1 mA) and the offset voltage is measured after having left the wafer in the dark for 15 mn. Light is then shone upon the wafer for 10 seconds, and the offset is then measured again, a few seconds after the light has been switched off. The change in offset voltage is proportionnal to the change in conductivity of the two resistances in the bridge that are not covered with metal. Since free excess carriers are recombined very quickly after the light has been switched off, the persistent photoconductivity we are then measuring should be related to the deep electronic states with very low emission and capture timeconstants.

This test structure was fabricated on wafers with a doping level of 2 x1014 cm-3. Both CZ and FZ wafers were processed. The results of the experiment described above, performed on both CZ and FZ wafers, are displayed on Figure 2. The difference between the two types of wafers is very clear : The average value of the offset change is 30 times bigger for the CZ wafer, which is known to contain more oxygen (about 1018 cm-3 for CZ and 1016 cm-3 for FZ).

The results of Figure 2 clearly validate our assumption concerning the relationship between persistent photoconductivity and oxygen concentration in the wafer. It is a straightforward experiment which can be performed directly at the wafer level, and can be implemented in a production testing environment. In order to use this technique to evaluate the quality of the denuded zone instead of the whole CZ wafer, one has to decrease the distance between the n+ contacts. This would force the current lines to remain in the shallow region of the wafer. The structure would then be sensitive to changes in conductivity in the top 5-10 µm of the wafer. The change in offset voltage obtained should therefore be similar to the one of FZ wafers. The experimental verification of this last concept is under way.

Figure 1: Layout of the Persistent Photoconductivity Wheatstone bridge test structure. The typical dimensions are 300 X 300 µm.

Figure 2: A typical result of PPC measurement on CZ and FZ-based structures. The current used for driving the bridge was 500 µA. The average change in offset voltage is 140 µV for the CZ and 5 µV for the FZ.

[2] Y. Haddab, D. Manic, R. S. Popovic, Persistent photoconductivity as a tool for monitoring oxide clusters in silicon wafers, proceedings of the MIEL'97, Nis, Yougoslavia, september 1997.