Measurement of sample charging during sputtering of III–V materials and devices

Abstract

A common problem in the analysis of non conducting solids using ion bombardment (sputtering) is charging of the sample, which may affect the electrostatics of a mass spectrometer's ion source, and hence ion transmission efficiency and reproducibility. The voltage changes induced on samples of III–V materials and devices during sputtering are measured in real time via the Stark effect in a resonance ionization mass spectrometer. High-lying electronic (Rydberg) levels of Ga are spectrally shifted by the strong electric field (> 14 kV cm ) present in the instrument. The Stark-shifted lines are a beneficial diagnostic to probe the ion source's local electric field, which depends on the sample position and potential. Sample charging and edge effects modify the instrument's ion transmission and reproducibility. Changes corresponding to 10 V in the sample's bias can be detected spectrally. Continuous and pulsed sputtering produce different voltage shifts under a variety of incident primary ion types (Xe + and O 2 +), energies and doses. Charging of semi-insulating GaAs is significant enough at 2 keV to deflect the ion beam out of the analysis area in the ion source. At 6 keV, the beam is not moved as much even though the amount of charging is about the same.