Preventive maintenance measures for contamination control

12/01/2005

High concentrations of corrosive and toxic gases can be emitted from metal etch chambers and downstream pipelines during preventive maintenance. Particle bursting incidents can also occur after pipes are disconnected from plasma-enhanced chemical-vapor deposition (PECVD) chambers when tools are cleaned. Airborne molecular contamination from these sources presents health risks to fab workers and causes wafer defects and process tool corrosion. In performance evaluations, installing a local ventilation hood and vacuum line sufficiently controlled harmful gases in metal etch chambers and pipes, while applying a heating tape to PECVD pipelines eliminated the particle bursting hazard.

By Shou-Nan Li, Hui-Ya Shih, Kuang-Sheng Wang, Kan Hsieh, Yin-Yung Chen, Y.-Y. Chen and James Chou

Workplace hygiene in semiconductor manufacturing remains an ongoing concern since studies identified higher health risks for wafer fab workers, including restrictive lung abnormality and lung cancer [1, 2]. However, most gas samples taken inside cleanrooms show that the time-weighted average (TWA) concentrations of airborne contaminants are below the legally permissible limits (e.g., threshold limit value, TLV) [3, 4]. Thus, some suspect that the intermittent short-term peak exposures during accidents and maintenance activities may contribute more significantly to the observed health effects [5-8].

For the short-term emitting sources, in addition to their adverse health effects, the release of contaminating gases and particles can also induce wafer defects [9-12]. For example, HCl (hydrogen chloride) of a concentration >28 parts-per-billion (ppb) has been observed to cause wafer corrosion defects [12]. Based on the guidelines published in the International Technology Roadmap for Semiconductors ITRS), the concentrations of the airborne molecular contamination (AMC) and particles are required to decrease as feature sizes shrink to 90nm and below (e.g., DRAM pitch, printed gate length) [13]. This means that AMC and particle controls will be even more crucial for new processes and wafer fabs.

In theory, locating the contamination source and then mitigating the emitting origin are the most effective ways to control the contaminants. Based on this logic, a series of measurements were conducted to determine and then control the emitting gases from the preventive maintenance (PM) activities of process chambers and downstream forelines. These measures are expected to better protect worker health and wafer quality, and to enhance wafer yield.

Materials and methods

In this study, a movable Fourier transform infrared (FTIR) spectrometer was used to instantly quantify the emitting gases from PM activities because of its proven effectiveness in detecting the hazardous gases inside the cleanroom [14]. Immediately before PM, the Teflon sampling tubing of the FTIR (optical path length = 10m) was inserted into the process chamber for ~3 min to determine the potential contaminating sources. Then the equipment engineer opened the process chamber and used a cloth saturated with IPA, H2O, or C2H5OH to clean the inner surfaces. In the whole PM process (~30 min), the sampling tubing was placed directly above the process chamber to sample the emitting gases.

Figure 1. Rotating local ventilation hood designed to draw away residual gases trapped inside process chambers during preventive maintenance.Click here to enlarge image

Upon detection of hazardous gases during PM activities, a customer-designed ventilation hood (Fig. 1) or a 2-in. dia. flexible local vacuum line was plugged in to control the emissions. To evaluate the effectiveness of the control measures, a SF6 flow with a rate of 5L/min and a concentration of 1000 parts-per-million (ppm) was introduced as a tracer gas into the process chamber. The capturing efficiency (CE) of the ventilation hood or the house vacuum line was then calculated by the following equation [15-17]:

CE = C1/C0      (1)

where C1 equals the concentration of the target gas (e.g., SF6) sucked into the ventilation hood or the house vacuum line with the SF6 introduced into the process chamber, and C0 equals the concentration of the target gas sucked into the ventilation hood or the house vacuum line when directly introduced into those areas.

Figure 2. Schematic of the setup for evaluating the capturing efficiency of the local ventilation system.Click here to enlarge image

As shown in Fig. 2, the SF6 flow was first introduced into the house vacuum line (Point 0) to obtain the C0 value, which was measured by the downstream FTIR (about 4m from the process chamber). Then, the SF6 flow was shifted to the bottom of the process chamber (Point 1) to obtain the C1 value when the PM activity was performed. To improve flow uniformity, the SF6 was released at the process chamber bottom through circular 1/4-in. Teflon tubing with 14 holes of 2mm dia.

In addition to monitoring the PM activities of the process chambers, the PM emissions of the downstream forelines were also measured by the FTIR. When a foreline was disconnected for PM, two ends of the foreline were quickly capped, then the inside gas sample was introduced into the gas cell of the FTIR for identification and quantification.

Results and discussions

Table 1 lists the peak concentrations of the residual gases measured right before the process chambers were opened for PM tasks. High concentrations of HCl (343ppm) and HCN (95ppm) would emit from metal etch chambers and contaminate the cleanroom if there were no control measures adopted. HNO3, SiF4, HNO2, HBr, NO2, HCOOH, NF3, and NH3 could also release into the cleanroom from the PM activities of dry etch and CVD chambers. Generally, the PM of a metal etcher would release many more contamination gases than those of other process (e.g., CVD, furnace) tools.

Click here to enlarge image

For a typical PM process of a metal etcher, the measured concentration variations of the contaminating gases are shown in Fig. 3. In the first 2 min when the process chamber was closed (right before PM), the chamber was full of polluting gases (e.g., HCl, HCN). As the chamber was opened for PM (at 9:39 as shown in Fig. 3), the contaminating gases emitted from the process chamber and diffused into the cleanroom. Without control means, the emitting acid gases could cause corrosion on the wafers and process tool surfaces [12], dissolve the dopant material (e.g., boron) of a particle filter to induce wafer defects [10, 18], and influence the respiratory function of exposed workers [1].

Figure 3. Concentration variations of contaminating gases from the preventive maintenance activity of a metal etcher.Click here to enlarge image

To prevent the PM gases from contaminating the cleanroom, a customer-designed ventilation hood (seen in Fig. 1) was mounted on the chamber lid as soon as the chamber cover was opened for PM. To make the hood convenient for PM tasks, the top surface of the hood was rotated by using 5-10 steel balls 6mm in dia. When the ventilation hood was applied, no gas was detected (by FTIR) around the process chamber during all PM activities.

Figure 4. SF6 concentration variations for determining the capturing efficiency of a local ventilation hood.Click here to enlarge image

To accurately determine the CE of the hood, the SF6 flow was introduced as a tracer gas in the PM process. Combining the measured SF6 results shown in Fig. 4 with Eqn. 1 shows that the applied ventilation hood has a CE value of 97.5%. For the process tool (e.g., P5000 or TCP system from Applied Materials) with a window as a side view, as shown in Fig. 2, the local vacuum line of 2-in. dia. was directly inserted into the process chamber through the window to suck away the contamination gases. Similarly, the SF6 techniques demonstrated that the house vacuum line alone (without a hood) had a CE value of 98.8%. The sucking flow rates (Q value) for the house vacuum line and the hood were measured to be ~6250L/min, which was also estimated by the introduced SF6 flow (Q = 1000ppm/0.8ppm × 5L/min = 6250L/min).

Click here to enlarge image

For the PM activities of the following pipelines behind the metal etcher, the measured results by the FTIR were similar to those shown in Table 1 except for HCN. Before the trapped particles of the foreline came into contact with H2O vapor, as shown in Table 2, there was no HCN detected. However, when a water-wetted air stream was introduced into the same foreline, high-concentration HCN with a peak value of 147ppm was measured. The results demonstrated that the HCN hazards inherent in the metal etch chamber and downstream pipeline could not be effectively removed by merely purging the chamber and pipeline with a “dry” N2 flow. In this study, to prevent residual gases from contaminating the cleanroom air, the disconnected (stainless-steel) pipelines were quickly capped and then transported to a well ventilated water booth for cleaning.

During this study, a particle bursting incident occurred when a PECVD foreline was disconnected for PM, although the foreline had been N2-purged for 3.5 hr. The spreading reddish-brown particles then triggered the alarm of a nearby VESDA particle sensor (from Vision Systems Ltd. in Australia). Similar to this phenomenon, Gu and Hauschulz mention that an explosion hazard might be inherent in the exhaust system for the PECVD process using SiH4 and NH3 as process gases [19]. They explain that the explosion hazard may be caused by the SiH4 gas trapped inside the particles. In this study, to investigate the particle bursting incident, a bellow pipeline of the PECVD process was capped and the trapped gas samples tested.

Click here to enlarge image

As shown in Table 3, at room temperature, only low concentrations of SiH4 (2.9ppm) and N2O (8.9ppm) were detected. However, when the same pipeline was heated at 100°C by a heating tape for 5 min, the measured SiH4 and N2O concentrations increased by more than 10×. In addition, high concentrations of other hazardous gases were also found: NO (>760ppm), NO2 (>2196ppm), SiF4 (>4155ppm), HF (1.92ppm), HNO3 (53ppm), and HNO2 (13.7ppm). After the heating, the color of the trapped particles inside the bellow pipeline turned from reddish brown to light brown. It appears that the amount of SiH4 gas trapped is related to the other chemicals present on the particles (e.g., HNO3, NO2). Further study is required to clarify this argument.

To control the gas hazards inherent in the PECVD process, heating the foreline at 100°C by a heating tape proved to be an effective way to expel trapped gases, especially SiH4, from the particles. By using a heating tape operating at 100°C, the particles deposited on the pipeline can also be greatly reduced due to the thermophoresis mechanism [20]. Both the expelled gases and particles will then be effectively removed by a downstream local scrubber (e.g., TPU, BOC Edwards) [21].

Conclusion

Both a local ventilation hood and a flexible house vacuum line appeared to be effective ways to prevent the emitting gases of PM activities from contaminating the cleanroom air quality. Without appropriate control measures, the releasing acid gases (e.g., HCl) from the PM of metal etchers might cause corrosion on the wafers and process tool surfaces. By applying a heating tape operating at ~100°C, the gas and particle hazards inherent in the foreline of the PECVD process using SiH4 and NH3 would be mitigated. For the fab operation, the control measures on the PECVD forelines could also reduce the defects caused by emitting particles and avoid operation interruption resulting from the VESDA alarm.

Acknowledgments

The authors thank Gen-Hou Leu in the Center for Environmental, Safety, and Health Technology at the Industrial Technology Research Institute (ITRI) for programming the quantification software of the FTIR spectra.

References

1. J.C. Luo, K.H. Hsu, L.L. Hsieh, C.J. Wong, M.J.W. Chang, “Lung Function and General Illness Symptoms in a Semiconductor Manufacturing Facility,” J. Occupational and Environmental Medicine, Vol. 40, No. 10, pp. 895-900, 1998.
2. D.M. McElvenny, A.J. Darnton, J.T. Hodgson, S.D. Clarke, R.C. Elliott, J. Osman, “Investigation of Cancer Incidence and Mortality at a Scottish Semiconductor Manufacturing Facility,” Occupational Medicine, Vol. 53, pp. 419-430, 2003.
3. L. Scarpace, M. Williams, D. Baldwin, J. Stewart, D.V. Lassiter, “Results of Industrial Hygiene Sampling in Semiconductor Manufacturing Operations,” American Conference of Governmental Industrial Hygienists: Hazardous Assessment and Control Technology in Semiconductor Manufacturing, pp. 47-52, Chelsea, MI, Lewis Publisher, 1989.
4. S.R. Woskie, S.K. Hammond, C.J. Hines, M.F. Hallock, E. Kenyon, M.B. Schenker, “Personal Fluoride and Solvent Exposures, and Their Determinants, in Semiconductor Manufacturing,” Appl. Occupational and Environmental Hygiene, Vol. 15, No. 4, pp. 354-361, 2000.
5. H. Pastides, E.J. Calabrese, D.W. Hosmer, D.R. Harris, “Spontaneous Abortion and General Illness Symptoms Among Semiconductor Manufacturers,” J. of Occupational Medicine, Vol. 30, No. 7, pp. 543-551, 1988.
6. M.B. Schenker, E.B. Gold, J.J. Beaumont, B. Eskenazi, S.K. Hammond, B.L. Lasley, et al, “Association of Spontaneous Abortion and Other Reproductive Effects with Work in the Semiconductor Industry,” American Journal of Industrial Medicine, Vol. 28, pp. 639-659, 1995.
7. M.F. Hallock, S.K. Hammond, E. Kenyon, T.J. Smith, E.R. Smith, “Assessment of Task and Peak Exposures to Solvents in the Microelectronics Fabrication Industry,” Appl. Occupational and Environmental Hygiene, Vol. 8, No. 11, pp. 945-954, 1993.
8. S.A. McCurdy, D. Pocekay, S.K. Hammond, S.R. Woskie, S.J. Samuels, M.B. Schenker, “A Cross-sectional Survey of Respiratory and General Health Outcomes Among Semiconductor Industry Workers,” American Journal of Industrial Medicine, Vol. 28, pp. 847-860, 1995.
9. F. Belanger, W. Goodwin, D. Ruede, “Concern Grows over AMC in Lithography Bays,” Solid State Technology, Vol. 47, No. 6, June 2004.
10. M. Forsland, S. O’Reilly, “Prevention of Airborne Molecular Contamination,” Semiconductor Fabtech, Vol. 12, pp. 129-135, June 2000.
11. R.P. Donavan, Particle Control for Semiconductor Manufacturing, New York, Marcel Decker Inc., 1990.

12. J.K. Higley, M.A. Joffe, “Airborne Molecular Contamination: Cleanroom Control Strategies,” Solid State Technology, Vol. 39, No. 7, pp. 211-214, July 1996.
13. SIA, The International Technology Roadmap for Semiconductors, 2004 edition, International Sematech, Austin, TX, 2004; http://public.itrs.net.
14. S.N. Li, Y.Y. Chen, H.Y. Shih, J.L. Hong, “Using an Extractive Fourier Transform Infrared (FTIR) Spectrometer for Improving Cleanroom Air Quality in a Semiconductor Manufacturing Plant,” American Industrial Hygiene Association Journal, Vol. 64, pp. 408-414, May/June 2003.
15. M.J. Sik, R.B. Lewis, D.J. Eveleigh, “Assessment of a Scavenging Device for Use in Paediatric Anaesthesia,” British Journal of Anaesthesia, Vol. 64, pp. 117-123, 1990.
16. S.H. Li, S.N. Li, H.Y. Shih, S.F. Chiou, H.D. Yi, “A Local Scavenging System to Remove Waste Anesthetic Gases During General Anesthesia,” ACTA ANAESTHESIOL SIN 2004, Vol. 42, pp. 61-67.
17. B.R. Beamer, J.L. Topmiller, K.G. Crouch, “Development of Evaluation Procedures for Local Exhaust Ventilation for United States Postal Service Mail-processing Equipment,” J. Occupational and Environmental Hygiene, Vol. 1, pp. 423-429, 2004.
18. F.A. Stevie, E.P. Martin, P.M. Kahora, J.T. Cargo, A.K. Nanda, A.S. Harrus, et al., “Boron Contamination of Surfaces in Silicon Microelectronics Processing: Characterization and Causes,” J. Vac. Sci. Technol. A, Vol. 9, No. 5, pp. 2813-2816, Sept./Oct. 1991.
19. Y. Gu, D. Hauschulz, “Comprehensive Downstream Effluent Management,” Solid State Technology, Vol. 41, No. 5, pp. 89-96, May 1998.
20. J.S. Lin, C.J. Tsai, C.P. Chang, “Suppression of Particle Deposition in Tube Flow by Thermophoresis,” J. Aerosol Science, Vol. 35, pp. 1235-1250, 2004.
21. S.N. Li, J.N. Hsu, H.Y. Shih, S.J. Lin, J.L. Hong, “FTIR Spectrometers Measure Scrubber Abatement Efficiencies,” Solid State Technology, Vol. 45, No. 7, pp. 157-165, July 2002.

Shou-Nan Li is the department manager of the Center for Environmental, Safety, and Health Technology (CESH) at the Industrial Technology Research Institute, 11F, Bldg. 51, 195-10 Sec. 4 Chung-Hsing Rd., Chutung, Hsinchu, Taiwan 310, R.O.C.; e-mail [email protected].

Hui-Ya Shih is an associated researcher in ITRI’s CESH.

Kuang-Sheng Wang is an associated researcher in ITRI’s CESH.

Kan Hsieh is a department manager at Winbond Electronics Corp., No. 9, Li-Hsin Rd., Science-Based Industrial Park, Hsinchu, Taiwan 300, R.O.C.

Yin-Yung Chen is an engineer at Winbond Electronics.

Y.-Y. Chen is a senior engineer at Taiwan Semiconductor Manufacturing Co. Ltd. (TSMC), 121, Park Ave. 3, Science-Based Industrial Park, Hsinchu, Taiwan 300, R.O.C.

James Chou is a section manager at TSMC.