Titanium contacts on N-type silicon

• Main Author: Wilkinson, John Malcolm
• Published: Middlesex University 1974
• Subjects: 537.622
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.568459

The current transport mechanism usually assumed to be valid for metal silicon Schottky barriers is thermionic-emission. However, Crowell and Beguwala, using a thermionic-diffusion model, suggested that significant deviations from the behaviour predicted by the thermionic-emission theory shonld be observed on low barriers, especially those formed on silicon of low impurity concentration. The barrier height of titanium on n-type silicon is 0.1)0 volt which is lower than most other metals, and should make the effects predicted by the thermionic-diffusion theory more important for titanium contacts. Titanium contacts were prepared on n-type silicon with impurity concentratlon from 2 x 10²⁰ m⁻³ to 3 x 10²¹ m⁻³ • Most of the diodes showed nearly ideal behaviour at low applied voltages and the current-voltage characteristics could be represented by the relationship [handwritten equation impossible to transcribe] with n values as low as 1.01. It was concluded that tunnelling interfacial layer and surface effects were insignificant for such diodes. However, at higher current densities, many of the same diodes exhibited deviations from ideal behaviour which were equivalent to n values as high as 1.25, or which could be interpreted in terms of a rapidly decreasing saturation current IS. Similar effects were observed on magnesium and aluminium contacts on silicon, but at bigher applied voltages, corresponding to the higher barrier heights of 0.55 and 0.72 volts. The main features of the experimental results agreed well with the predictions of the thermionic-diffusion theory for band bending between β = 9 and β = 2. At the upuer value of β the predictions of the thermionic-emission and thermionic-diffusion theories were almost identical and the diode behaviour was closest to ideal with n = 1.01. The lower value of β represents the limit of agreement between the measurements and tho thermionic-diffusion theory. Two possible mechanisms are outlined which could explain the discrepancy below β = 2. These are phonon scattering of electrons between the barrier maximum and the metal, and the effect of the reserve layer on the shape of the potential barrier at very low band bending. The results demonstrate the concli tions under which the thermionic-diffusion theory rather than the thermionic-emission theory should be applied, and suggest a practical lower limit on β for the range of applicability of the thermionic-diffusion theory.