Spreading resistance

Fig. 2. Spreading resistance profiles from (a) Au-diffused n-type Si, (b) same sample after hydrogen plasma-exposure for lh at 200°C and (c) another position of the sample in (b). The bar represents the range of initial resistance values of the Si prior to Au diffusion (Mogro-Campero et al., 1985). 

A comparison of the deuterium profile measured by SIMS and the spreading resistance profile obtained on deuterated samples is shown in Fig. 6. The region over which there is a reduction in thermal donor concentration matches well with the depth of deuterium incorporation. There is an excess of deuterium over the amount needed to passivate all the oxygen-donor centers. This is frequently observed in hydrogenation experiments and indicates there is hydrogen present in several states. 

Fig. 6. Spreading resistance and SIMS profile from deuterated CZ Si containing an initial uniform concentration of 6 x 1016 cm-3 thermal donors. The high, near-surface concentration is due probably to deuterium molecule formation (Pearton et al., 1986). 

Figure 2 shows that samples with boron concentrations ranging over four orders of magnitude can be effectively treated. The samples were hydrogenated at 122°C the upper three samples for one hour, and the most heavily doped sample was hydrogenated for four hours. Note that the spreading resistance increases toward the surface. The value of resistivity... 

Fig. 2. Spreading resistance profile of four B-doped samples of (100) Si hydrogenated at 122°C. The three higher resistivity samples were hydrogenated for one hour the lowest resistivity sample was treated for four hours. The resistivities were obtained from a calibration curve. Note the greater penetration depth of atomic hydrogen as the boron concentration decreases.

Heating the B-doped samples above 200°C in vacuum to dissociate the hydrogenated complex results in flat spreading resistance (Rs) at the original bulk value. A second exposure to Hj restores the increased Rs at the surface. These are crucial experiments that demonstrate that hydrogen is involved and that the process is reversible and reproducible. 

Fig. 3. Spreading resistance profiles for 10-0 cm, B-doped (100) Si hydrogenated as shown.

In related experiments by Johnson (1985), atomic deuterium was used instead of Hx to neutralize boron in Si. Similar results on spreading resistance were obtained. Furthermore, the distribution profile of D was measured by secondary-ion mass spectrometry (SIMS), as shown in Fig. 4. The distribution profile of D reveals 1) that the penetration depth of D is in good agreement with the resistivity profile and 2) that the D concentration matches the B concentration over most of the compensated region. In another sample, the B was implanted at 200 keV with a dose of 1 x 1014 cm-2, the damage was removed by rapid thermal anneal at 1100°C for 10 sec., and then D was introduced at 150°C for 30 min. As shown in Fig. 5, it is remarkable that the D profile conforms to the B profile. 

The ability of atomic hydrogen to neutralize acceptors other than boron is demonstrated in Fig. 10. Samples of silicon doped with Al, Ga, and In were exposed to Hx at 125°C for one hour. They all exhibit an increase in spreading resistance at the surface by at least one order of magnitude. 

Fig. 10. Spreading resistance profiles for silicon doped with indium, gallium, and aluminum.

Fig. 12. SIMS concentration profiles of B, H, and O in silicon. The inset shows the spreading resistance of this sample before (B) and after (A) hydrogenation. The hole concentration at the surface is 2.0 x 1019 cm-3 before and 1.4 x 1018 cm-3 after hydrogenation.

Spreading resistance—the resistance between the deep interior of a semiconductor and a very sharp metal point pressed on the surface— measures the local resistivity on a scale of the order of the contact radius (Ehrstein, 1974). It thus measures the amount of hydrogen taking part in donor or acceptor passivation, whether this occurs by complex formation or by compensation. However, some methods of preparing samples for a spreading resistance measurement may involve heating above room temperature, and this may cause redistribution even of hydrogen bound in some types of complexes (Mu et al., 1986). 

Capacitance-voltage (C-V) measurements on Schottky diodes orp-n junctions can provide information similar to that yielded by spreading resistance, namely the net charge density of fixed centers that are ionized in the... 

Fig. 17. Advance of a passivated region into silicon uniformly doped with 5 x 1018 boron atoms per cm3, after exposure for 30 min. at about 150°C to atomic deuterium from a plasma source (Johnson, 1985a). (a) Spreading resistance profile, (b) Depth distribution of total deuterium and of boron from SIMS.

The ambient temperature sensor in the IRT 3000 is a KTY type spreading resistance sensor. The resistance of the sensor can be written as a second order polynomial function... 

The doping densities calculated from the slope of the C 2 versus V plots, as shown for example in Fig. 10.2, agree well with values measured by other methods [Otl]. Deviation between results obtained by spreading resistance measurements and electrochemical CV measurements are usually found to be below 20% for doping densities between 1012 and 1018 cnT3 [Pe3]. 

G. Mutter, S. Kalbitzer and G. N. Greaves, Ion Beams in Amorphous Semiconductor Research J. Boussey-Said, Sheet and Spreading Resistance Analysis of Ion Implanted and Annealed Semiconductors... 

---