Deposits in semiconductor corona emitters in cleanroom and simulated air
07/01/1998
Deposits on semiconductor corona emitters in cleanroom and simulated air
Charles G. Noll, ITW Static Control and Air Products, SIMCO Static Control and Cleanroom Products, Hatfield, Pennsylvania
Scanning electron microscope (SEM) photographs are presented to illustrate deposits on silicon and germanium emitters after a one-month exposure to positive- and negative-polarity electrical corona. Deposits formed in Class 100 cleanroom air with about 5000 ppm of moisture are relatively independent of polarity and electrode material. The deposits formed in simulated air (80% nitrogen and 20% oxygen, <50-ppm moisture) show features that are characteristic of material and polarity.
The elimination of static electricity is an important component of cleanroom manufacturing, particularly in industries that produce semiconductor devices and assemblies, flat panel displays, and magnetoresistive heads. Static charges can enhance contaminant deposition and result in electrical discharges that damage products and reduce product yield.
Static charges are typically found on insulating materials and conductors that are isolated from ground. The insulating barrier between these charges and ground prevents conduction and charge neutralization.
Electrical air ionizers use controlled electrical discharges, called corona, to produce positive and negative ions in the air. Figure 1 illustrates a typical cleanroom ionizer. The ions produced in the corona move easily in air or flowing gases to the isolated or fixed charges until the charged particles become electrically neutral. Electrical air ionizers are commonly used in the cleanroom manufacturing processes mentioned above.
 |
Figure 1. Typical air ionizer for cleanroom use.
As cleanroom technology advances to serve the electronics industry, manufacturers of air ionizers have sought lower particle generation levels from corona emitters. Electrical corona is a nonthermal plasma process. Such processes are known to be highly reactive, both corroding metallic electrode materials and producing very fine particles from the plasma reaction products. These conductive particles are sometimes incompatible with the cleanroom manufacturing environment. The use of semiconducting silicon and quartz emitters has shown the potential to reduce particle generation.
Recent work has shown the performance of semiconducting germanium needles as corona emitter electrodes for electrical static eliminators [1]. The author and colleagues tested positive- and negative-polarity corona emitters made from semiconducting silicon and germanium in a simulated air environment composed of 80% nitrogen gas from liquid nitrogen and 20% oxygen from bottled gas. The simulated air environment was chosen to eliminate moisture and other contaminants from the test regimen.
In this article, electron microscope photographs and other data are presented to compare corona reaction products and deposits that form in simulated (dry) air and Class 100 cleanroom air. The corona consists of free electrons, ions, and excited species that can etch and form particles above electrode surfaces [2-4]. The active species in air are formed from nitrogen, oxygen, carbon, hydrogen, and traces of other gases and particles. The ionic species change in the presence of water vapor, generating particles [5, 6] that become either airborne contaminants or deposits on emitters [7]. Although the preferred operating mode for low-particle-generating emitters is a clean, dry-air environment, some users demand nonmetallic emitters for applications in room air. Knowledge of the influence of normal moisture levels and other airborne contaminants on emitter corrosion is, therefore, important to practical application of corona emitters.
The cleanroom environment contains about 5000 ppm of water vapor, whereas the simulated environment contains <50 ppm. The present results provide a basis for understanding the deposits that are observed on emitters and their relation to particle generation. These results are preliminary to a broader study of reactions involving moisture and halogen gases over the semiconductor corona emitters.
Samples and experimental arrangement
The germanium emitters are made from >99.999% pure, n-type (antimony doped), polycrystalline germanium. Their electrical resistivity is in the range 5-40 W-cm. The cleanroom tests included unetched needles and others that were exposed to a bright etching solution. The tests in simulated air used only unetched needles. We observed no evidence of tip smoothing by the nitric acid etch over germanium, and there were no significant effects attributable to this etching on the emitter corrosion rates.
The silicon emitters were made from >99.999% pure, optical grade, p-type (boron doped), single-crystal silicon. They were ground and bright etched, and had an electrical resistivity of 40-100 W-cm. All tests were run using continuous DC corona, and exposure times were approximately one month with currents between 1.5 and 8 ?A [1].
Silicon and germanium emitters in simulated air
Positive polarity. Over the course of the exposure, the emitter developed a bluish tint to about 0.5 of the diameter from the tip. The tip (first 1/8 diameter) appeared to be dulled and some fine particles were on the bluish-tinted surface of the needle. There was no significant weight gain or loss with the positive silicon emitters during the tests. Under the SEM, the tips contained a patterned structure with fine holes or channels (Fig. 2a).
 |
Figure 2. a) Deposits on silicon emitter (+); b) deposits on germanium emitter (+).
Deposits - mostly silicon oxide contaminants coming from the negative-polarity silicon emitters (Fig. 2b) [1] - also formed on the germanium emitter tip. The observation of this contamination is important, since it reveals that the presence of intense electric fields at the emitter attracts negative-polarity silicon particles to its surface, even in the presence of the corona. The particles had the appearance, under light microscopy, of a fine brown powder drawing color from the underlying germanium emitter. The SEM images revealed an almost frothy appearance, with typical features on the order of 10 ?m in size. The surface contains pores similar to those observed on the silicon emitter. In this case there are fewer holes, with a less orderly hole pattern. The disordered structure could possibly come from the precipitation of larger particles carried over from the silicon emitters. As with the positive-polarity silicon emitter, the positive-polarity germanium emitter exhibited no significant weight gain or loss.
Negative polarity. The surface finish of the negative-polarity [8] silicon electrodes had a uniform dull gray appearance where it was exposed to the corona (Fig. 3). Under higher magnification, the tip seemed coated with a layer of fine particles that were approximately 1 ?m in diameter and appeared to be composed of agglomerates of finer particles. The texture was grainy, like fine sand. The needle exhibited a measurable weight gain.
 |
Figure 3. Silicon emitter (-) in simulated air.
The deposit on the tip was about 10-20 ?m in thickness. There was no obvious erosion of the tip. If anything, the surface layer was thickened slightly by the formation or deposition of particles. The surface of the tip was bright under electron illumination, the result of an electrically insulating deposit.
The germanium emitters also showed a measurable weight gain and a brownish deposit was observed on their tips (Fig. 4a). The deposits on the negative-polarity germanium emitter consisted of irregularly shaped, germanium oxide, 10-?m-particle clusters. There were occasional pores in the deposit (Fig. 4b).
 |
Figure 4. a) Germanium emitter (-) in simulated air; b) deposits on germanium emitter (-).
Silicon and germanium emitters in cleanroom air
Both the positive- (Fig. 5a) and negative-polarity (Fig. 5b) silicon emitter tips were covered with crusty and cracked deposits. During attempts at Auger surface analysis, the deposits were found electrically insulating. Where data could be obtained, the peaks indicated a composition of silicon dioxide beneath the cake exceeding 2.7 ?m in thickness.
Careful examination of images for the negative-polarity emitter revealed some pores in the deposits, but they were primarily located at the tip and randomly positioned, unlike those observed for the simulated air case. On the face of it, the encrusted deposits arose from surface processes, moisture, and contaminants in the environment. A variety of elemental species was found in these deposits through Auger spectroscopy. Such results have been reported by other researchers, especially those studying deposits over metal emitters in air [9, 10].
 |
Figure 5. a) Deposits on silicon emitter (+); b) deposits on silicon emitter (-).
The germanium needle tips contained deposits that were similar to those observed on the silicon needles. They appeared more uniform and tightly adhering in their coating of the tips (Fig. 6a and b). Since the tests of the germanium and silicon emitters were not simultaneous, we must suggest that these differences had no fundamental significance; rather they had a similar origin in surface reactions, environmental contaminants, and moisture that change over time.
From an electrical standpoint, the ionizers with silicon and germanium performed similarly in the cleanroom environment. This was expected since the semiconducting electrodes have similar physical dimensions.
 |
Figure 6. a) Deposits on germanium emitter (+); b) deposits on germanium emitter (-).
Conclusion
Ionizers using semiconducting silicon and germanium emitters and dry-air purge systems have reduced particle generation to the levels needed for some semiconductor, flat panel display, and magnetoresistive head manufacturing processes. In this paper, information is reported on the deposits and reaction products found on emitter electrodes used in Class 100 cleanroom and simulated dry air. These conditions represent realistic bounds on modern cleanroom environmental conditions where these emitters might be used.
In the Class 100 cleanroom, where the tests took place, there were very few differences among the deposits on the silicon and germanium emitters at positive and negative polarity. It is clear from the SEM photographs that silicon and germanium emitters perform similarly under typical cleanroom moisture and contaminant levels. Both materials accumulate particulate matter and shed particles to the environment.
The use of clean, dry air reduces deposits on the emitters from contaminants and the production of particles from moisture in the air. When moisture is reduced to very low levels, as in the simulated gas case, there is a variety of corrosive deposits on emitter tips. One then might distinguish basic mechanisms for the failure of emitter materials that are otherwise masked in the cleanroom air environment. Results have shown that only the negative-polarity silicon emitters produce measurable particle emissions at low moisture levels [1]. These particles, unfortunately, contaminated the tip of the positive-polarity germanium emitter and influenced our understanding of deposits for that case. The contamination, however, revealed the importance of electrostatic precipitation in the emitter corrosion process. Work is underway toward clarifying the basic mechanisms of emitter failure resulting from changes in the simulated air environment.
Acknowledgments
The author acknowledges the surface characterization and analysis support of the Materials Analysis Group, Philips Semiconductors, Sunnyvale, CA; Materials Analytical Services, Norcross, GA; and the ITW Technology Center, Glenview, IL.
References
1. C.G. Noll, P.A. Lawless, "Comparison of Germanium and Silicon as Emitter Electrodes for Air Ionizers," 19th Ann. EOS/ESD Symp., Santa Clara, CA, pp. 195-204, 1997; accepted for publication in J. Electrostatics, 1998.
2. See special issue on charged dust in plasmas, IEEE Transactions on Plasma Science, 22(2), 1994.
3. K. Nashimoto, "Growth of SiO2 Needles Induced by Positive Corona Discharging," Jpn. J. Appl. Phys., Part 2, 26(7), pp. L1138-1140, 1987.
4. C.R. Gorla, S. Liang, G.S. Tompa, W.E. Mayo, Y. Lu, "Silicon and Germanium Nanoparticle Formation in an Inductively Coupled Plasma Reactor," J. Vac. Sci. Technol. A, 15(3), pp. 860-864, 1997.
5. N.L. Allen, P. Coxen, R. Peyrous, T. Teisseyre, "A Note on the Creation of Condensation Nuclei by Negative Corona Discharges in Air at Low Pressure," J. Phys. D., Appl. Phys., Vol. 14, pp. L207-209, 1981.
6. S. Sakata, T. Okada, "Effect of Humidity on Hydrated Cluster-ion Formation in a Cleanroom Corona Discharge Neutralizer," J. Aerosol Sci., 25(5), pp. 879-893, 1994.
7. K.D. Murray, V.P. Gross, P.C.D. Hobbs, "Clean Corona Ionization," 12th Ann. EOS/ESD Symp., Orlando, FL, pp. 36-40, 1990.
8. The SEM photograph for the silicon sample in this case was secured after a 500-hr exposure to corona.
9. R. Le Ny, "Corrosion of Electrodes Subjected to Corona Discharges in Room Air and its Influence on Discharge Parameters," Inst. Of Phys. (UK), Conf. Ser. 66, pp. 173-178, 1983.
10. M. Blitshteyn, S. Shelton, "Contamination and Erosion of Cleanroom Air Ionizer Emitters," Microcontamination, Vol. 8, pp. 28-32, 1985.
CHARLES G. NOLL received his PhD degree in solid state physics from The Ohio State University. He is VP of research and development for ITW Static Control and Air Products companies, including SIMCO, Herbert Static Control, and Richmond. He has more than 20 years of experience with electrostatic processes for contaminant control and static elimination. SIMCO Static Control and Cleanroom Products, 2257 NorthPenn Rd., Hatfield, PA 19440-1998; ph 215/822-2171, fax 215/822-3795, e-mail [email protected].