ULTRAPURE FLUIDS

Evaluating reusable HDPE containers for delivery of high-purity hydrofluoric acid

R. Thomas Talasek, Brian K. Hunt, and David C. Cooke, General Chemical

As semiconductor device geometries continue to shrink, the purity of process chemicals has become even more critical. Nowhere is this more evident than in the specifications for reducing metallic and particulate impurities, especially for chemicals used in such critical cleaning operations as the RCA process¹ and processes using solutions of hydrofluoric acid. Metallic impurities can adversely affect oxide integrity,^2,3 reduce minority carrier lifetimes, provide nucleation sites for stacking faults during film growth, and cause surface microroughness^4—7 and a multitude of other process problems.^3,8,9 Metallic impurities have also been shown to adversely affect the stability of SC-1 (NH_⁴OH: H_²O_²:H_²O) solutions, reducing their effectiveness.¹⁰ Purity specifications of < or = 1 ppb are becoming common for many chemicals, with typical levels significantly lower than the specification.

Failure mechanisms caused by particulate impurities have been understood for years and include disruption of surface geometries, production of defects in thin films by inclusion, or introduction of metallic or nonmetallic impurities. The size of a particulate able to produce a device failure has been historically believed to be one-fourth to one-half of the critical (lateral) geometry or one-half of various vertical dimensions.¹¹

Although the initial chemical's purity is important, the most critical impurity level is the one existing at the point of use. Many additional opportunities for contamination occur after the chemical is manufactured. One of the most significant can be exposure to the polymeric materials used to manufacture chemical containers and the components (tubing, fittings, valves, liners, and the like) used for constructing distribution systems. Several studies have evaluated metallic contaminants from various polymers. Special attention has been paid to various fluoropolymers, with the most attention given to perfluoroalkoxy (PFA) resins.^12—16 High-density polyethylene (HDPE) has received only passing interest, even though it is the most common material used for drums and other large containers used in returnable and one-way service. This is surprising since containers must maintain chemical purity.

Hydrofluoric acid is often used in processes where metallic and particulate impurities must be closely controlled. Because hydrofluoric acid is one of the most aggressive chemicals that require high purity, it was selected for evaluating several commercially available drums; drums suitable for it should also be suitable for less aggressive chemicals.

Filling, Storing the Samples

Three drums each from three drum manufacturers specializing in semiconductor chemical containers were selected for the study. All drums were manufactured from unpigmented HDPE. They were filled with 49% hydrofluoric acid (by weight), which contained <500 ppt of any single metallic impurity. The acid was passed through 0.1-µm filtration prior to entering the container. The containers were filled as received from the manufacturer, without any cleaning or other pretreatment, to create a worst-case scenario. The fill operation was performed in a Class 100 cleanroom, using a closed-fill system, to minimize metallic or particulate contamination from the environment.

Samples were extracted from the drums immediately after filling and analyzed for metallic impurities by inductively coupled mass spectrometry (ICP-MS) and graphite furnace atomic absorption (GFAA). Particle counts were also taken immediately after filling, with the particle counter intake inserted directly into each drum. After the drums were stored at ambient temperature for 7 days, metals and particles were measured again. The procedure was repeated after 21 days.

Comparing Metallic, Particulate Contamination Levels

Figure 1 compares metallic impurity levels, averaging results across the three drums from each manufacturer. Minor variations were observed in common environmental contaminants, such as sodium, potassium, iron, and calcium. However, significant differences were found in contaminants often associated with HDPE-resin manufacturing processes, such as zinc, magnesium, titanium, and barium. Zinc is also a significant environmental contaminant at these levels. Only one of the three drums was capable of supporting a 1-ppb metal specification.

Figure 1: Comparison of metallic impurity levels immediately after drum fill.

Figure 2 compares initial particle counts at 0.5 and 0.2 µm. Significant differences were found in the drums from all three manufacturers, with the ones from manufacturer C demonstrating the best initial performance.

Figure 2: Comparison of particulate impurity levels immediately after drum fill.

Figures 3 and 4 demonstrate the performance of the drums over 21 days. Figure 3 illustrates the typical concentration of the common environmental contaminants mentioned above, with sodium selected as the example element. Contamination levels plateau after 7 days (or less), with the drums from manufacturer C again performing best. Figure 4 shows the typical concentration of resin contaminants, with barium chosen as the example element. When the contaminant is found in measurable quantities, the concentration rises to its maximum level in 7 days (or less), similar to the results seen for environmental contaminants.

Figure 3: Comparison of sodium concentration after 7 and 21 days.

Figure 4: Comparison of barium concentration after 7 and 21 days.

Figure 5 is a time study of 0.2-µm particle levels in the three drum types. Although the drums from manufacturer C show significantly better initial performance, all the drums tended toward a terminal particle level of 200—300 particles/ml. The initial particle levels are thought to be indicative of cleanliness in the manufacturing process, but the final particle level is dominated by particles generated by the resin/additive package used to make the drum; most unpigmented drums use similar mold releases, plasticizers, and other additives that contribute to the final particle level.

Figure 5: Comparison of 0.2-µm particle levels after 7 and 21 days.

Conclusion

Selecting appropriate packaging is critical to maintaining the chemical purity required by the semiconductor industry. This is especially true in the case of metallic impurities, where both resin selection and manufacturing processes play an important role. In the case of particulate impurities, although initial performance is significantly affected by manufacturing processes, the drums studied in this article tended toward similar performance over time. The best drum maintained its superior performance, although deterioration was evident.

Acknowledgment

Portions of this article were presented at the 43rd Annual Technical Meeting of the Institute of Environmental Sciences, Los Angeles, May 1997.

References

1. Kern W, and Puotinen DA, "Cleaning Solutions Based on Hydrogen Peroxide for Use in Silicon Semiconductor Technology," RCA Review, 31:187—206, 1970.

2. Meuris M, Heyns M, Küper W, et al., "Correlation of Metal Impurity Content of ULSI Chemicals and Defect-Related Breakdown of Gate Oxides," in ULSI Science and Technology 1991, Andrews JM, and Celler GK (eds), Pennington NJ, Electrochemical Society, pp 141—161, 1991.

3. Meuris M, Heyns M, Mertens P, et al., "The Combined Effect of Silicon Surface Roughness and Metal Impurity Contamination on Gate Oxide Integrity," in Cleaning Technology in Semiconductor Device Manufacturing, Pennington, NJ, Electrochemical Society, pp 144—169, 1992.

4. Miyashita M, Makihara K, Tsuga T, et al., "Dependence of Surface Microroughness of Cz, Fz, and EPI Wafers on Wet Chemical Processes," Journal of the Electrochemical Society, 139(8):2133—2142, 1992.

5. Meuris M, Heyns M, Mertens P, et al., "Metallic Contamination and Silicon Surface Roughness in Wet Preoxidation Cleaning Treatments," in Microcontamination '91 Conference Proceedings, Santa Monica, CA, Canon Communications, pp 658—665, 1991.

6. Meuris M, Verhaverbeke S, Mertens PW, et al., "The Relationship of Silicon Surface Microroughness and Gate Oxide Integrity in NH^₄OH/H^₂O^₂ Mixtures," Japan Journal of Applied Physics, 31:L1514—1517, 1992.

7. Ohmi T, Miyashita M, Itano M, et al., "Dependence of Thin-Oxide Film Quality on Surface Microroughness," IEEE Transactions on Electronic Devices, 39(3):537—545, 1992.

8. Miyashita M, Itano M, Imaoka T, et al., "Optimizing NH^₄OH/H^₂O^₂ Cleaning Process for Ultra-Cleaning Wafer Surface Preparation," in Extended Abstracts, 179th Electrochemical Society Meeting, Pennington, NJ,Electrochemical Society pp 709—710, 1991.

9. Ohmi T, Imaoka T, Kezuka T, et al., "Segregation and Removal of Metallic Impurities at the Interface of Silicon and Fluorine Etchant," Journal of the Electrochemical Society, 140(3):811—818, 1993.

10. Schmidt HF, Meuris M, Mertens PW, et al., "Effects of Metallic Impurities on the Stability and Performance of Hydrogen Peroxide—Based Cleaning Solutions," in Proceedings of the Institute of Environmental Sciences' 39th Annual Technical Meeting, Mt. Prospect, IL, Institute of Environmental Sciences, p 238—242, 1993.

11. Osborn CM, "Aerosol Control in Semiconductor Manufacturing," in Aerosols, Liu BYH, Pui DYH, and Fissan HJ (eds), New York, Elsevier Science Publishing, pp 673—675, 1984.

12. Miki N, and Ohmi T, "Ultrapure Liquid Chemicals — Hydrofluoric Acid," in Proceedings of SEMICON East 89 Technical Symposium, Mountain View, CA, SEMI, pp 223—242, 1989.

13. Kikuyama H, Takano J, and Miki N, "Clean Technology of Chemicals," in SEMI STEP/Particle '88 Proceedings, Mountain View, CA, SEMI, pp 369—376, 1989.

14. Paulsen PJ, Beary ES, and Bushee DS, "Analysis of Ultrapure Reagents from a Large Sub-Boiling Still Made of Teflon PFA," Analytical Chemistry, 61(8):827—830, 1989.

15. Goodman JB, and Van Sickle PM, "Extracting Ionic Contaminants from PFA Polymeric Materials," Microcontamination, 9(11):21—25, 1991.

16. Talasek RT, Hall LH, Mallini L, et al., "Determination of Leachable Metallic Impurities from Semiconductor Chemical Packaging Material," in Microcontamination '94 Conference Proceedings, Santa Monica, CA, Canon Communications, pp 362—368, 1994.

R. Thomas Talasek, PhD, is director of business operations for General Chemical's electronic chemicals group in Pittsburg, CA. Before joining General Chemical, he was director of laboratories at Texas Instruments' chemical operations in Dallas. He has written more than 60 articles and technical presentations. Talasek is cochair of the SEMI Process Chemicals Guidelines Subcommittee. He holds a PhD in analytical chemistry from the University of North Texas. (Talasek can be reached at 510/458-7345.)

Brian K. Hunt is process engineering supervisor for General Chemical's electronic chemicals group. He received his MS in chemical engineering from the University of Alabama, and his BS in chemical engineering from Pennsylvania State University.

David C. Cooke, PhD, is laboratory manager for General Chemical's electronic chemicals group. Before joining General Chemical, Cooke was quality assurance manager for Commonwealth Technologies. He holds a PhD in analytical chemistry from Duke University and has written eight technical publications.

MicroHome | Search | Current Issue | MicroArchives
Buyers Guide | Media Kit