The Effects of Plasma Cleaning on Carbon and Hydrocarbon-contaminated Samples and Quantification of Decontamination in the Scanning Electron Microscope

Plasma cleaning of specimens prior to electron microscopy has proven to be a highly effective method of mitigating specimen-borne contamination [1]. In plasma processing, oxidative or reductive gases react chemically with different materials. An oxygen plasma generated from air or oxygen/inert gas mixtures is highly effective in removing hydrocarbon contamination. The disassociated oxygen radicals created by the plasma chemically react with the hydrocarbon residues present on the object being processed, converting hydrocarbons to CO, CO2, and H2O, volatile species which are evacuated by the vacuum system. Previously hydrocarbon contamination has been monitored by means of quartz crystal microbalance experiments [2] and residual gas analysis [3] to generate quantitative measures of cleaning efficiency in test chambers at XEI Scientific. Nanoflight® movies recorded the sequence of hydrocarbon removal for a 1000 nm-thick layer of contamination in a SEM [4] with an estimated rate of removal of 1nm per 9 seconds. In this report we present quantitative data obtained from experiments in a SEM for real world cleaning rates for hydrocarbon decontamination and an exploration of the effect of plasma cleaning on graphite grids. Experiments were performed in a Zeiss Supra VP40 SEM equipped with an Evactron E50 E-TC® alternate gas plasma cleaner (Figure 1A). For a hydrocarbon standard, bare silicon squares were spin-coated with polymethyl methacrylate (PMMA) to a thickness of ∼250 nm as shown in Figure 1B. The squares were placed at fixed distances from the plasma source, facing towards or facing away from the plasma source. Oxygen was used as the source gas and samples were cleaned at 50 Watts for 5 minutes. Carb-N-Grids™ (Figure 1C) made of graphite were cleaned with oxygen plasma at 50 Watts for 6 cycles of 5 minutes each. The results of experiments to plasma clean PMMA are shown in Figure 1D. The rate of hydrocarbon removal is strongly influenced by the distance of the specimen from the plasma source, generating removal rates from 0.75nm/sec at the closest point to 0.22 nm/sec at the most distant point when facing the plasma source. Sample distance has negligible effect on PMMA layer thickness reduction when facing away from the source, generating a relatively constant rate of 0.22 nm/sec. The images of graphite grids before (Figure 1E) and after (Figure 1F) plasma cleaning showed removal of surface water and organic material but minimal effect on the basic structure of the grids. These results will be compared with previous studies and Monte Carlo simulations to understand the contributions of line-of-sight and scattering of the reactive oxygen species to effective decontamination of complex interior surfaces of vacuum chambers and specimens. Typically, vacuum chambers can be plasma cleaned at turbo molecular pressures of 10-2 to 10-3 Torr with cleaning times of 2 - 10 minutes to maintain pristine conditions, with chambers returning to normal operating pressures in < 20 minutes. Current users of Evactron plasma cleaners report significant reduction in pump down time as well as easier maintenance of the pristine state of cleanliness of SEMs, CDSEMs, FIBs, TEMs HV and UHV chambers. (A) A Zeiss Supra VP40 SEM chamber rendering was created to exactly position the sample in relation to the Evactron during the experiments. The purple beam represents the aperture opening of the Evactron plasma system. The sample is directly facing the Evactron. (B) Silicon squares spin-coated with 250 nm PMMA are hydrocarbon standards. (C) A strip of graphite grids examined under a stereomicroscope. The grids are 3 mm in diameter and 70 nm thick. (D) Hydrocarbon removal rates as a function of distance from the plasma source and orientation to the plasma source. (E) Comparison of a graphite grid before plasma cleaning and (F) after plasma cleaning show removal of surface contamination.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.