Charge conduction through silicon dioxide during ion implantation

• Main Author: Broughton, Carl
• Published: University of Surrey 1989
• Subjects: 621.31042; Integrated circuits
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.329673

Ion implantation is used to dope silicon substrates during the manufacture of integrated circuits. Insulating films, inevitably present on the wafer surface during a typical metal-oxide-silicon process, will prevent the charge introduced by the ions from being conducted away. The resultant charge accumulation will produce localised electric fields which can lead to breakdown of the insulator and damage to the devices. In this work an investigation into the underlying charging and charge leakage mechanisms during ion implantation of silicon MOS structures was undertaken, concentrating on charge conduction in silicon dioxide under ion bombardment. A detailed theoretical study of the phenomena that occur as a result of ion implantation indicated that photoconduction, space charge limited current injection, impact ionisation and secondary electrons all have a role in charge conduction through oxide. To distinguish between these various possible types of conduction, X-ray, electron and ion radiations were used for the experiments in this work. The X-ray yield from ion implantation into silicon was measured. From these results and the data in the literature it was deduced that X-ray generated photoconduction in oxide during ion irradiation is insignificant. Electron beam induced conductivity was measured as a function of applied field with various electron energies, electron energy deposition rates, oxide thicknesses and doses. The results of these experiments confirmed the charge conduction mechanisms proposed, i. e. that photoconduction and space charge limited conduction are the main methods of charge conduction through oxide under irradiation. Under ion irradiation the voltage acquired by an aluminium pad on oxide on silicon was measured in real time. The development of the pad potential was measured with various oxide thicknesses, ion species and energies, beam current densities, pad geometries and dose. The major factors determining the pad voltage proved to be the pad area to perimeter ratio and the ions' projected range compared with the oxide thickness. Secondary electrons were also found to contribute to pad potential.