Diagnostic test for ion implantation dosimetry

Ion implantation has become an accepted tool in the fabrication of solid-state devices. Two reasons for the widespread use of ion implantation are its potential for precise doping control and for reproducibility. However, effects peculiar to ion impact on solids can interfere with accurate dosimetry and thereby jeopardize these primary advantages of ion implantation. Diagnostic techniques are therefore needed to determine the presence and origin of dosimetry errors. The most straightforward method of dosimetry is the integration of the ion current falling on the sample. 1 We present here a simple test that diagnoses errors in such an integration system. From the total charge deposited on the target, one can easily calculate the total number of incident ions that have been implanted if the average charge state of the ions is known. Complications arise, however, because secondary electrons and ions, which are generated by the primary particles impacting the sample and other surfaces, can cause spurious currents which are also integrated. The number and type of secondary particles depends on the incident ions' species, the sample material, and on the surface condition of the sample.2 The problem may be complicated further by the action of suppression biases applied to various electrodes in the vicinity of the target. Frequently, these suppression biases are only partially effective in suppressing the various secondary particles. 3 A diagnostic procedure should reveal the presence and magnitude of spurious current. The test should permit immediate evaluation of modifications and adjustments of suppression biases, and the procedure should require very little in the way of ancillary equipment or ion implantation modification. These criteria are met in our test by preparing target samples which generate a higher yield of secondary particles over one half of the surface while the other half generates the normal current of secondary particles. The current in the sample is monitored by an oscilloscope synchronized to the sweep of the ion beam across the target. The change in the sample current as the primary beam is scanned over the two halves measures the relative magnitude of the secondary error current. We prepared two types of samples and both performed well. The first is a silicon single-crystal wafer upon half of whose face was evaporated 500 A of gold. The gold film has a higher secondary electron yield than silicon. The second type of sample is a 2-in. tantalum disk half of which is covered by a tantalum "roof top," i.e., an inverted "V" of Ta sheet. This sample makes use of the fact that the emission of secondary electrons and ions strongly varies with the angle of incidence of the primary particle. Over one half of the tantalum disk the ions are incident nearly perpendicularly to the surface, while over the other half they penetrate at a grazing angles. The targets are sketched on the upper left sides of Figs. 1 and 2.