To ensure the highest quality gas for use in semiconductor fabs, research efforts have focused on the materials used in gas distribution systems.
In addition to meeting requirements for ultra-high purity (UHP) the distribution system for specialty and corrosive gases must be corrosion resistant and exhibit no catalytic behavior. Vereecke(1) et al. reported that reaction between reactive gases and construction materials used in gas delivery systems can result in the formation of corrosion by-products, which may contaminate the gas stream as both particles and volatile metal complexes.
The SEMI test method for evaluation of particle contribution from gas system components exposed to corrosive gas service provides a means to compare gas handling components for potential particle generation in corrosive gas service. In this particular study, Vereecke used the SEMI test sequence to compare the corrosion resistance of a gas filter assembly employing nickel media with a gas filter assembly employing 316L stainless steel media.
Traditionally, most employ 316L stainless steel to constructing subcomponents for use in reactive gas distribution systems. The proper selection of the stainless melt can improve the corrosion resistance in UHP gas systems.
In the case of high-purity semiconductor gas filter assemblies, specifying the proper chemistries, grain size, and inclusions is important.(2) In addition, the suitability of the material to be formed, mechanically polished, and electropolished must also be evaluated. The careful selection of the 316L stainless steel utilized in the fabrication of filter assemblies is critical, because materials of similar composition can perform very differently.
Wang et al. noted variability in the corrosion behavior of steel alloys with bulk compositions are virtually identical when exposed to moist HCl.(3) Smudde et al.(4) observed that when moisture is below 1 ppmv, bromine from HBr is not incorporated beyond the native oxide of 316L stainless steel, and no macroscopic degradation of the metal occurs. Fine et al.(5) confirmed the latter observations by investigating the effect of moisture content on the extent of HBr corrosion for 316L electropolished stainless steel.
The scanning electron microscopy and x-ray emmision spectroscopy analysis of the exposed sample coupons indicated no effect upon exposure to HBr containing less than 0.5 ppm of moisture. A moisture content of 10 ppm resulted in bromide incorporation and the onset of corrosion. Fine noted the formation of corrosion pits when increasing the moisture level in the HBr to 100 ppm. Researchers noted a dense bromide scale at a moisture concentration of 1,000 ppm.
Corrosion is typically quantified by such techniques as trace gas analysis, change of surface morphology, particle shedding, and leakage. Wang et al.(6) estimated the lifetime of electropolished 316L stainless steel tubing in HCl service, containing 1 ppm of moisture, to be on the order of 2 to 3 years, based on particle shedding due to corrosion. The latter value is in agreement with field experience.
Several maintenance practices are recommended to prevent the introduction of moisture into the corrosive gas distribution system.(7) When adequate purge and evacuation procedures are followed to remove corrosive gases (such as HBr), exposing electropolished 316L stainless steel to moist air without diminishing the initial surface quality is possible. However, if the purge and evacuation procedures are not followed, iron- and bromine-rich crystalline deposits form on the surface.
To maintain higher purity in corrosive gas service, researchers are investigating new construction materials. One such material is nickel, which is corrosion resistant in aggressive environments. However nickel is also a reactive material and is commonly used as a hydrogenation catalyst.
 |
Researchs at Tohuku University investigated the thermal decomposition characteristics of active specialty gases on various metal surfaces.(8) The metal surfaces investigated included nickel, oxygen-passivated 316L stainless steel, chromium-passivated 316L stainless steel, and 316L stainless steel with an electropolished surface.
Researchers monitored the thermal decomposition of the active specialty gases with the aid of a gas chromatograph and a Fourier transform infrared spectrometer. A spectrometer monitored the specialty gas concentration exiting the test sample (0.25-in. diameter, 1-m-long tube). In the case of phosphine, 100 ppm of phosphine in argon passed through the tube at a flow rate of 5 sccm. The nickel sample exhibited a strong catalytic effect on the phosphine decomposition.
In addition to the metal surfaces listed above, researchers investigated the catalytic effect of a nickel filter on thermal decomposition of silane. They passed 1,000-ppm silane gas in argon through the nickel filter at a flow rate of 5 sccm. Initially the gas was introduced at room temperature. No silane was detected at the outlet of the filter for 150 minutes. After 150 minutes researchers detected silane and increased the silane level to approximately 800 ppm.
The fact that they did not detect 1,000 ppm of silane after 330 minutes suggests that the catalytic decomposition of silane continued to occur. Upon increasing the temperature to 100 degrees C, researchers observed a large spike in the hydrogen concentration due to further silane decomposition. Further testing of silane with nickel filters at flow rates typical of actual use conditions suggests that the nickel surface is readily saturated and that minimal thermal decomposition of the gas occurs(9) after the surface saturation occurs.
Given these results, due consideration must be given to the corrosion resistance and the catalytic behavior of nickel when selecting materials for use in reactive gas service.
Stainless Steel vs. Nickel Media
The SEMI document “Test method for evaluation of particle contribution from gas system components exposed to corrosive gas service” (Draft Doc. # 2572) provides a method to compare gas handling components for particle generation in corrosive gas service. The document is intended as a practical means of generating performance data for a group of components to be compared in a selection process. The test sequence was used to compare the corrosion resistance of a gas filter assembly employing nickel media with one employing 316L stainless steel media.
The all-metal point-of-use filter assemblies employed in this study are constructed of an EP 316L stainless steel housing and a media pack (nickel or 316L stainless steel) of sintered metal fiber of a specified diameter. The housing of the assemblies has an internal surface finish of < 7 microinch Ra and a chromium-enriched surface layer.(10)
The gas filter assemblies in question were initially subjected to a particle cleanliness test. The particle cleanliness test was based on the Sematech Semaspec #93021511A-Std, “Test method for determination of particle contribution by filters in gas distribution systems.” This test consists of a 45-minute steady flow test, followed by a 45 minute dynamic test, and finally a 10-minute impact test.
The nitrogen test flow rate was 10 lpm at an inlet pressure of 30 psi. The particle cleanliness testing was conducted in a Class 1,000 cleanroom. The particle levels downstream of the test filter assemblies were monitored with a Condensation Nucleus Counter (CNC, Model 3025, from TSI, Inc., Minneapolis, MN), sampling at a flow rate of 0.01 scfm for particles larger than 0.003 micron in size.
After completion of the particle cleanliness test, the test component was removed from the aerosol test stand and capped in a glove bag under nitrogen purge. The capped test component was then transferred to the HCl exposure test stand. All components and tubing employed in the test stand were constructed of EP 316L stainless steel surfaces. Where possible, VCR fittings were used. The test filter assembly was installed on the exposure test stand employing a glove bag, while maintaining a positive nitrogen purge. The exposure test stand (between V5 and V1) was baked for two hours at 70 degrees C while maintaining the purified nitrogen purge (< 1 ppb moisture) at a test flow rate of 500 cc/min.
After overnight cooling under nitrogen purge, the HCl exposure was commenced. The HCl exposure was run once per hour for six hours. After completion of the sixth exposure, the test filter assembly was held in static HCl for 18 hours (SEMI method specifies 16 hours). The exposure cycle described was repeated eight times. Then it was suspended over the weekend and a dry nitrogen purge was maintained.
The HCl exposure was resumed at the start of the next work week. The moisture content was monitored with the aid of a Meeco Aquamatic + moisture analyzer (Meeco, Inc., Warrington, PA) to ensure that a constant moisture content of 100 ppm in the nitrogen stream was maintained. The HCl was purified to < 100 ppb of moisture by the use of a point of use resin-based purifier (GLPV2HCLVMM4).
The probability of particle generation by a test component when subjected to the corrosive exposure sequence described above is supported by previous reports of particle generation by EP 316L stainless steel tubing after exposure to moist HCl (100 ppm) for 8.5 days. As previously reported, Wang et al.(6) estimated the lifetime of EP 316L stainless steel tubing based on the time required to generate 10 particle/scf at a flow rate of 3.531 scfm (100 slpm). In addition, Fine et al.(5) noted that 200 ppm of moisture is required for condensation of liquid HBr, while the onset of corrosion is noted at 100 ppm.
The test filter assemblies were subjected to a 2-hour dry nitrogen purge following completion of the eight-day HCl exposure. The test stand was subjected to a cycle purge for 30 cycles to remove any traces of HCl from the test component. The cycle purge consisted of 20 seconds of rough vacuum followed by 10 seconds of pressurization with dry nitrogen to 80 psig.
The test filter assemblies were subjected to a final 24-hour purge with dry nitrogen prior to removing the test component from the exposure stand. The final purge with dry nitrogen was performed at a low flow rate to ensure that no particles generated by the exposure were removed.
 The sequence of exposure of gas handling components to corrosive service and the subsequent determination of the particle contribution. |
null
The capped test component was then transferred back to the cleanroom for particle cleanliness testing in purified nitrogen. After completion of the particle cleanliness testing, the exposed filter assemblies were sectioned and subjected to microscopic analysis to determine the extent of corrosion of the 316L housing, 316L stainless steel media pack and the nickel media pack. The effects of corrosion were analyzed in terms of surface roughening, change in fiber diameter, pitting, and localized corrosion at the weld interface between the assembly housing and the media pack.
Cleanliness Testing
During the particle contribution testing, the particle levels detected downstream of the stainless steel media filter (GLFF1100VMM4) and the nickel media filter (GLFN1100VMM4) immediately after the HCl exposure were within background level. The latter result indicates that no particle corrosion by-products were generated during the HCl exposure in the case of the nickel media and the 316L stainless steel media.
The absence of particle corrosion by-products for the nickel media is expected as nickel is a corrosion-resistant material. The results obtained for the 316L stainless steel media suggests that the material is corrosion resistant under the conditions typically experienced in a semiconductor UHP specialty gas distribution system. Previous reports have indicated that all-metal point-of-use filter assemblies with sintered 316L stainless steel media remain suitable for use in semiconductor gases after more than two years of service in BCl3, HBr and Cl2.(11)
Microscopic Inspection
In addition to monitoring the extent of corrosion in terms of particle shedding, the exposed samples were sectioned and subjected to microscopic inspection to determine changes in surface morphology.
The internal housing surface finish of the upstream and downstream sections of the HCl exposed filter assemblies, including inlet and outlet bore, were measured to be < 7 microinch Ra. The measured surface finish of the filter assemblies exposed to moist HCl is within the maximum internal surface finish specification of the all-metal filter assembly. The fine surface finish exhibited by the exposed assemblies suggests that the internal surfaces of the filter assemblies were not subjected to corrosion. The fine surface finish is expected as the all-metal filter assemblies evaluated possess a chromium-enriched surface layer that enhances the corrosion resistance of electropolished 316L stainless steel surfaces.
Previous reports have indicated that electropolished stainless steel is susceptible to aggressive corrosion caused by halogen containing gases, specifically in weld areas.(12) The most significant types of corrosion are intergranular corrosion, general surface corrosion and pitting corrosion. In the case of weld areas, the corrosion typically occurs in the heat affected zone or downstream of the weld with respect to the purge gas flow. The extent of corrosion is strongly influenced by the number of inclusions, presence of volatile impurities (manganese and copper) and the levels of carbon and sulfur.(13)
The weld regions of the all-metal filter assembly with sintered stainless steel media (GLFF1100VMM4) and the all-metal filter assembly with nickel media (GLFN1100VMM4) revealed no evidence of corrosion (see Figures V and VI). The absence of any particle corrosion by-products, as evidenced in the particle cleanliness testing, is in agreement with the corrosion-free weld regions observed. These regions illustrate that the proper selection of raw materials (chemistry and processing technique) combined with refined manufacturing techniques strongly influences the corrosion resistance of a stainless steel component.(2)
 |
The internal surface area of the assembly housing is less than 1 percent of the total BET surface area of the filter assembly. The higher surface area of the medium pack suggests that corrosion is most likely to occur at the medium pack. The upstream and downstream sintered woven 316L stainless steel mesh support layers and the nickel mesh support layers employed in the point-of-use gas filter assemblies subjected to the HCl exposure revealed no corrosion.
The diameter of the sintered 316L stainless steel fibers of the media pack from the filter assembly exposed to HCl was determined using a scanning electron microscope (SEM) and compared to a filter assembly supplied from manufacturing stock. The diameter of the 316L stainless steel fibers from the returned filter assemblies is identical to that of the filter assembly removed from manufacturing stock (see Figures VII and IX). The photomicrographs revealed some contamination (scaling), but no degradation of the sintered fibers. The particle cleanliness observed for the point-of-use all-metal filters after the HCl exposure is further evidence that no degradation of the 316L stainless steel media pack occurred during exposure to the moist HCl.
 |
Similarly, the diameter of the sintered nickel fibers of the media pack from the filter assembly exposed to HCl gas was determined with the aid of a SEM and compared to a filter assembly supplied from manufacturing stock. The diameter of the nickel fibers from the returned filter assemblies is identical to that of the filter assembly removed from manufacturing stock — see Figures VIII and X. The SEM photomicrographs revealed no degradation of the sintered nickel fibers. The particle cleanliness observed for the point-of-use all-metal filter assembly with the nickel media after the HCl exposure is further evidence that no degradation of the nickel media pack occurred during exposure to the moist HCl.
Conclusions
The point-of-use all-metal filter assemblies with 316L stainless steel media and nickel media evaluated in moist HCl gas revealed no evidence of the generation of particles due to corrosion. The particle cleanliness of the assemblies after exposure to moist HCl gas was observed to be within background levels, even under dynamic and impact test conditions. In addition, the assemblies displayed a high surface finish and no evidence of corrosion or degradation of the media packs.
The latter observations indicate that the 316L stainless steel media pack and the nickel media pack are corrosion resistant under the normal service conditions experienced in a semiconductor corrosive gas distribution system. In the case of active specialty gases, such as arsine, phosphine and silane, however, consideration must be given to the catalytic behavior of various metal surfaces. Through our experience, with this study and a previous study where no degradation was noticed at a period of more than two years for returned filter assemblies employed in corrosive gas service,(11) we suggest that, in an effort to maintain good housekeeping practices, gas filters should be changed every two years.
Joseph O'Sullivan, Ph.D., is a technical director in the scientific and laboratory services department at Pall Corp., in Port Washington, NY. In this post he has provided technical support to the semiconductor industry for more than seven years. The author of several articles on filtration and contamination control, O'Sullivan received his bachelor of science degree in chemistry and his doctorate in physical chemistry from University College Cork, Ireland. He is a member of the IEST and serves on various SEMI task forces and committees.
References
- “Generation and Transport of Corrosion Products in HCl Gas Lines,” G. Vereecke, M.M. Heyns, N. Anderson, C. Elsmore and P. Espitalier-Noel, Future Fab International, p. 283, Issue 2, Volume 1, 1997.
- “Selection of 316L Stainless Steel For High Purity Semiconductor Gas Filter Assemblies,” W. Murphy and B. Gotlinsky, Solid State Technology, Contamination Control Supplement, May 1995.
- “Using Atomic Force Microscopy to Evaluate Alloys for Corrosive Gas Service,” H.W. Wang and S. Chesters, Microcontamination, June 1994.
- “Materials Selection for HBr Service,” G.H. Smudde, W.I. Bailey, B.S. Felker, M.A. George and J.G. Langan, Corrosion Science, Vol. 37, p. 1931, 1995.
- “The Role of Moisture in the Corrosion of HBr Gas Distribution Systems,” S.M. Fine, R.M. Rynders and J.R. Stets, J. Electrochemical Society, Vol. 142, No. 4, April 1995.
- “Estimating the Lifetime of Electropolished Stainless Steel (EPSS) Tubing in Corrosive Gas Service,” H.W. Wang, G. Doddi and S. Chesters, Journal of the IES, p. 28, July 1994.
- “The Effects of Corrosive Gases on Metal Surfaces,” P.M. Clarke. R.A. Hogle and S.M. Lord, in Microcontamination ’93 Conference Proceedings, p. 433, Canon Communications, Santa Monica, CA, 1993.
- “The Evaluation of Thermal Decomposition Characteristics of Active Specialty Gases on Various Metal Surfaces Using FT-IR Method,” T. Ohmi, Y Shirai, S. Lee and S. Miyoshi, Proceedings of the Institute of Environmental Sciences, p. 17, 1995.
- “Compatibility of Silane with Nickel Filter Media,” R. Binder, Productronica Exhibition, 1995.
- “Filter Products for the Semiconductor Industry,” p. 106-109, Second Edition, Pall Corp., 1996.
- “Recommended Change-out Schedule For Electronic Grade Filters–Case Study,” J. O’ Sullivan and B. Gotlinsky, STR-PUF 32.
- “Investigating the Corrosion Resistance of Heat-Affected Zones in CrP Tubing,” S. Krishnan, S. Grube, O. Laparra and A. Laser, p. 37, Microcontamination, May 1996.
- “Effect of Stainless Steel Melt Methods on the Corrosion Resistance of Welds in UHP Gas Lines,” C. Burton, Workshop on Gas Distribution Systems, SEMI 1997.