Improving Throughput in SEM/FIBs

Vacuum-based processes display lower performance in the presence of adventitious hydrocarbons that are volatilized from various sources, such as solvents and oils, as well as chamber and sample surfaces¹. For instance, in electron microscopy, the presence of hydrocarbons (HC) causes unwanted effects such as image blur, “black square” and “black frame” formation during prolonged beam exposure times². These and associated issues have directed the need for RF-driven plasmas to decontaminate vacuum chambers via production of excited-state species (usually oxygen radicals) that mildly remove contamination. It is crucial to note that these systems produce electrically neutral cleaning species that flow from the plasma into the chamber so the decontamination is realized by mild chemical etching.

The oxygen radicals formed in the plasma oxidize carbon compounds, producing CO₂, CO, and H₂O, which are evacuated from the instrument. Quantum chemistry rules regarding energy loss state that these oxygen atoms do not react with diatomic molecules in two body collisions but require a third body to kinetically remove excess energy. Oxygen radicals also react on solid surfaces such as metals where they can react or recombine with hydrocarbons. Several studies performed at XEI Scientific using pre-contaminated quartz crystal microbalances (QCM) to measure cleaning rates have demonstrated that Evactron cleaning is highly effective at removing hydrocarbons^{3, 4}. Faster decontamination rates were noted in FIBs and SEMs fitted with turbo molecular pumps.

The present generation of Evactron plasma cleaners includes models such as the EP illustrated in Figure 1 and the new E50 in Figure 2. The EP and E50 models are engineered for high efficiency cleaning with lesser cost by simplification of both software and hardware. The compact design of the model EP makes it a versatile solution for small sample preparation chambers and SEM/FIB loadlocks.

Figure 1. The Evactron Model EP De-contaminator system (a) includes a desktop controller and the KF clamp together plasma radical source. (b) The small size PRS fits on a FESEM column with numerous analytical accessories.

Figure 2. The Evactron Model E50 De-contaminator system (a) includes a desktop controller and the newly designed external hollow cathode plasma radical source for 50 Watt operation. (b) The Evactron E50 PRS mounted on a SEM port has a clean, compact footprint.

Technology

Invented in 1999, the Evactron plasma cleaner is based on RF hollow cathode excitation (RFHCE) plasma which is uniquely available from XEI Scientific. Other plasma cleaners use other excitation techniques such as inductively coupled plasma (ICP) that use more power and produce more heat for similar cleaning rates. RFHCE forms more neutral radicals for chemical etching and fewer energetic ions that can result in sputter damage to the surfaces being cleaned

The original Evactron plasma source used an internal plasma electrode that was built to be used on chambers at pressures between 1 Torr and 200 m Torr attained by roughing pumps. Lower pressures were avoided because of the then common oil diffusion pumps used to acquire high vacuum (down to 10^-7 Torr). The conversion to turbo molecular pumps on new SEMs has made it possible to encompass plasma cleaning down to high vacuum in the 10^-4 Torr range. At these low pressures, there is a balance between input gas flow and pump speed that offers a pressure in 10^-2 to 10^-4 Torr. In the new Evactron E50 model, gas input down to 6 sccm will develop a plasma with a chamber pressure in the 10^-4 T range, but with a loss of cleaning rate because of lower radical production (Figure 3). In this pressure range, there is a trade-off between low pressures for low flow and higher cleaning rates from more input gas flow.

Figure 3. At low-pressure cleaning rates begin to drop as fewer oxygen molecules are available for radical production. Higher RF power increases the cleaning rate and less flow lowers it.

Materials and Methods

To prove the effect of Evactron Turbo Plasma Cleaning on pump down time and HC contamination removal, a research was done using two models of Evactron plasma cleaners:

1. A Model Evactron EP Plasma De-Contaminator was positioned on a large, extremely oil-contaminated 50 L vacuum chamber fitted with a 450 L/second turbo molecular pump, MKS 972B dual range pressure gauge, 14 CFM scroll pump, and a residual gas analyzer (RGA). All pump down curves are from atmosphere where the chamber is vented for 10 minutes to accept configuration changes. A baseline pump down curve from atmosphere was taken and RGA data taken when the measured pressure was 1E-7 Torr. To contaminate, a two-inch nipple was contaminated with two drops of pump oil and was installed on the chamber. The nipple was wrapped in bake tape and foil, and then it was heated to 120 °C while being pumped on with only the roughing pump. This bake process lasted for ~ 18 hours. Pump down curves and RGA data were captured before and after plasma cleaning.

Figure 4. A 50-liter test chamber was used for measuring pump-down times of a pristine chamber with datapoints before heavy contamination, with contamination, and after plasma cleaning.

RGA spectra of contamination (Figures 7-9) and pump down curves (Figure 11) were collected before and after plasma cleaning so as to show the effects of Evactron Turbo Plasma cleaning on pump downtime and hydrocarbon removal. The chamber is said to be in pristine condition when the HC peaks on the RGA spectrum are less than a partial pressure of 2 × 1^-10 Torr.

An E50 model Evactron De-Contaminator was mounted on a 22 L vacuum chamber to measure cleaning rates as a function of distance from the plasma source for the data charted in Figure 10. For the E50 system, cleaning rates were measured by means of QCMs placed 25 cm and 0 cm from the mounting nipple with the plasma radical source operating at 20-70 W.

Results and Discussion

With Evactron plasma cleaning technology, short cleaning times of five minutes or less are quite sufficient to eliminate all hydrocarbon contaminants as illustrated in this before and after image pair (Figures 5 and 6).

Figure 5. Before cleaning.

Figure 6. After 5 minutes of cleaning.

Such results are established by the comparison of RGA spectra in Figure 7, a pristine chamber before plasma cleaning, Figure 8 after contamination, and Figure 9, after 10 minutes of plasma cleaning. This rapid cleaning allows vacuum chambers to be kept pristine, and SEMs/FIBs to have more uptime for imaging and analysis.

Figure 7. Pristine chamber before contamination shows all HC peaks are absent from RGA scan, leaving residual water and atmosphere.

Figure 8. RGA scan after contaminating the chamber with hydrocarbons. Characteristic hydrocarbon peak series are present. To assure a pristine chamber, six 5 minute plasma cleans at 20 watts using an Evactron EP were executed with a 2 minute wait time between cleans.

Figure 9. The RGA scan after 10 minutes shows that the cleaning byproducts including atmosphere and water are still being detected but all signs of the hydrocarbon contamination are gone.

The length of the pump downtime differs based on HC contamination levels in FIBs and SEMs. Thus the pump downtime could be used as a sign of the cleanliness of the vacuum system. Figure 10 shows that efficient HC decontamination rates of hundreds of angstroms/minute are realized at normal SEM/FIB working distances. The data in Figure 11 illustrates that Evactron plasma cleaners can considerably reduce both the pump downtime of FIBs and SEMs as well as HC contamination, thus enhancing sample processing throughput without sacrificing the quality of the analysis.

Figure 10. The Evactron E50 Combination of higher power and low pressures offers good cleaning rates throughout larger vacuum chambers.

Figure 11. The plasma cleaned chamber reached a vacuum of 3E^-6 Torr in 20 minutes pump down whereas the contaminated chamber did not reach this level after 44 minutes.

Remote plasma cleaning can be done on FIBs and SEMs at pressures below 75 mTorr (10 Pa) during direct pumping with a turbo molecular pump. Earlier studies⁵ revealed that low chamber pressures boosted the rate of cleaning and the distances at which cleaning was observed. Low pressures increase cleaning efficiency by increasing the mean free paths and decreasing the recombination rates of the oxygen radicals by three-body collisions. This result along with the new hollow cathode design at low pressures allows more effective cleaning strategies for large instrument vacuum chambers.

Ignition at low pressure also allows the plasma to be switched off after cleaning a short time (1-5 minutes), causing a quick return to base pressure with the turbopump to remove reaction products, and then restarting the plasma to perform the cleaning cycle again. While at base pressure any unseen hydrocarbons can degas and redistribute within the chamber because of long free path molecular flow. The redistributed products can be eliminated in the following plasma cycle. Cyclical plasma cleaning is demonstrated to be a highly effective way to attain a pristine chamber using this new plasma cleaning technology at a high vacuum. Typical turbo molecular pump input flows of 20 sccm offer both good cleaning and do not overheat the pump.

Conclusions

As the need for superior quality data and higher throughput of samples in Scanning Electron Microscopes (SEMs) and Focused Ion Beam (FIB) systems increases, so does the requirement to shorten pump downtimes between loading samples. Industry demands SEM/FIB systems to be working round-the-clock and preferably maintained in pristine condition with uncompromised image quality. Frequent venting of the FIBs or SEMs to load samples adds contamination and moisture into the vacuum chamber, resulting in much longer pump downtimes and lower efficiency. Evactron cleaning removes this contamination effortlessly. Samples can be imaged more speedily in a clean environment, and if contamination is detected, it can be swiftly removed, the SEM pumped back to operating pressure quickly and, with only minutes of cleaning delay, imaging and analysis are restarted. The result is increased productivity.

The new Evactron^® Turbo Plasma™ De-Contaminators remove hydrocarbon (HC) contamination from FIBs, SEMs, and other analytical tools using a gentle, down-stream plasma afterglow process. At turbopump pressures, Evactron cleaning becomes faster and spreads all through the chamber. This is because of longer mean-free-paths that cause less recombination of oxygen radicals in the required three-body collisions and reduced scattering to chamber walls. In the majority of cases, short plasma cleaning cycles are adequate to remove contamination and considerably reduce pump downtime, allowing for high throughput for sample processing and analysis.

The Evactron E series of plasma cleaners offer quick, effective, and powerful cleaning over a broad range of pressures, enabling superior quality, artifact-free images, and better efficiency of sample analysis. The ‘Fastest Way to Pristine’ is a slogan that translates into the fastest productivity for your laboratory.

References and Further Reading

1. Sullivan, N. et al., (2002). Microsc. Microanal. 8(2), 720.
2. Joubert, L.-M. (2013). Microsc. Anal. 27(4), 15.
3. Gleason, M.M. et al., (2007). Microsc. Microanal. 13(2), 1734.
4. Morgan, C.G. et al., (2007). Microsc. Microanal. 13(2), 1736.
5. Vane, R. et al., (2016) Microsc. Microanal. 8(2), 720.

This information has been sourced, reviewed and adapted from materials provided by XEI Scientific.

For more information on this source, please visit XEI Scientific.