Extreme Silicon

Demonstrating a contamination-free wafer surface extraction system for use with CE and IC

Peng Sun and Marty Adams, MEMC Electronic Materials

As the semiconductor industry ramps up to 300-mm wafers and faces the evolving defect density requirements of the SIA roadmap, nonmetallic ionic contaminants on wafer surfaces have become an increasingly important concern.

As the semiconductor industry ramps up to 300-mm wafers and faces the evolving defect density requirements of the SIA roadmap, nonmetallic ionic contaminants on wafer surfaces have become an increasingly important concern. Because such impurities can degrade wafer surfaces, particularly in the presence of adsorbed water, they have detrimental effects on device fabrication.^1–3 Unfortunately, methods for measuring nonmetallic ions at the ultratrace levels found on wafer surfaces lag behind the industry's demands. Recent research in this area has focused on using ion chromatography (IC) and capillary electrophoresis (CE).^3–11

Figure 1: Schematic of the wafer-extraction system.

A common approach is to use high-purity deionized (DI) water to extract the wafer surface (or a specific surface location) and then analyze the extraction solution using IC or CE. Through years of experience in surface contamination analysis, the current authors have found that sample preparation and handling are among the most critical processes for the successful measurement of surface ionic contaminants. Airborne gas-phase species such as ammonium, hydrofluoric acid, hydrogen chloride, and sulfur and nitric oxides can cause significant contamination problems during the surface extraction process, especially when it is done in an open environment in which such chemicals are used. This airborne contamination problem has also been reported elsewhere.⁵ Performing extractions in a Class 1 cleanroom cannot totally eliminate the problem, since ULPA filters are not designed to remove airborne molecular species. When a traditional extraction box method is used, an operator has to open the box to collect the extraction solution, and airborne ionic species can contaminate both the solution and the wet wafer surface during this operation. Thus, there is a clear need for a contamination-free wafer surface extraction system to ensure accurate and reproducible measurements of surface nonmetallic ions. This article describes research on such a system that was done at MEMC Electronic Materials (St. Peters, MO), including a study that correlated the results of CE and IC analyses.

Experimental Procedures

Apparatus. The wafer surface extraction apparatus used in this research is shown schematically in Figure 1. This apparatus is a modification of one reported on elsewhere.^11,12 The key feature of the system is that various operations can be performed without opening the extraction box: the extraction box itself can be prerinsed, the extraction fluid added, and the sample solution collected via a series of valves and inlet and outlet lines. The DI-water supply bottles are pressurized with high-purity helium (He) gas to automate the filling of the extraction box and to prevent airborne ionic species from contaminating the DI-water supply. Multiple extractions of one wafer can be performed to check the removal efficiency of the first extraction, and single-side extractions are possible by replacing the double-side extraction box shown in the figure with a single-side extraction box.

Figure 2: Electropherograms for anion analysis of a 10-ppb standard solution (top) and a wafer extraction solution (bottom).

Wafer Extraction. Before the extraction process is begun, the extraction box, DI-water supply bottles, and sample collection bottles must be washed with 18-M‡ DI water, after which the supply bottles are filled with 18-M‡ DI water and the extraction system is connected. The syringes and transfer lines are then rinsed with DI water from the supply bottles, and 40 ml of DI water are introduced into the extraction box by filling syringe A from the supply bottle and switching the selection valve to the inlet line. After the extraction box has been rinsed with this water, it is emptied using the outlet line, syringe B, and the sample collection line. Another 40 ml of water are introduced and left in the box for 15 minutes before being transferred to a sample collection bottle through the outlet line, syringe B, and the sample collection line. This water solution can be used as an extraction box blank. The operator then places a wafer sample in the extraction box and introduces another 40 ml of DI water into the box. The wafer sample is extracted for 15 minutes, during which the operator manipulates the box every 3 minutes so that the DI water is circulated around the wafer surface. The wafer extraction solution is then transferred into a sample collection bottle.

	CE Method	IC Method
Cation analysis
Capillary/column	HP extended light capillary (ID = 50 µm, OD = 375 µm, length = 64.5 cm)	IonPac CS12A 2-mm analytical column with IonPac CG12A guard column
Injection	4.5 kV, 45 sec	750 µl
Buffer/eluent	IonPhor cation Cu electrolyte buffer (pH = 3.0)	20 mM H2SO4
Separation condition	Separation voltage = 25 kV	Flow rate = 0.35 ml/min
Detection	Indirect UV	Suppressed conductivity
Anion analysis
Capillary/column	HP extended light capillary (ID = 50 µm, OD = 375 µm,length = 64.5 cm)	IonPac AS12A 2-mm analytical column with IonPac AG12A guard column
Injection	5 kV, 60 sec	750 µl
Buffer/eluent	IonPhor PMA electrolyte buffer (pH = 7.7)	21 mM NaOH
Separation condition	Separation voltage = 30 kV	Flow rate = 0.35 ml/min
Detection	Indirect UV	Suppressed conductivity

CE and IC Analysis. A 3D CE system (Hewlett Packard; Palo Alto, CA) and a DX500 IC instrument (Dionex; Sunnyvale, CA) were used to analyze extraction solutions for ammonium (NH₄⁺) and anions of interest in the study. All CE buffer solutions were obtained from Dionex, and cation and anion calibration-standard solutions for sample quantification were purchased from Solution Plus (St. Louis) and SPEX (Metuchen, NJ). Instrument operating conditions and parameters are listed in Table I.

Results and Discussion

This article focuses on the analysis of NH₄⁺ and chloride (Cl^–). The authors have observed that these are the two most abundant ions on polished silicon wafer surfaces. A large portion of these surface contaminants come from ambient air in final process areas and from final wafer cleaning processes. NH₄OH is present in the SC-1 cleaning solution and hydrogen chloride (HCl) in the SC-2 solution.¹³

The advantages of using CE include rapid analysis and short instrument equilibration times, small sample volumes (nanoliters versus the milliliters used in IC), low operational costs, and ease of use. On the other hand, IC provides higher sensitivities with the use of preconcentration columns, which allow larger injection volumes, and lower retention time shifts caused by separation column degradation. Both CE and IC are suitable techniques for trace ion analysis in wafer extraction solutions.

Figure 3: Chromatograms for NH₄⁺ analysis of a 10-ppb NH₄ ⁺ standard solution (top) and a wafer extraction solution (bottom).

Figure 4: Correlation between CE and IC results for standard solutions (NH₄⁺ : y = 1.0227x – 0.0971, R² = 0.9955; ClUP>–: y = 1.1243x – 0.3313, R² = 0.9897).

Correlation between CE and IC Results. Experiments were carried out to test the correlation between CE and IC measurement results for surface nonmetallic ions. After test wafers were extracted, solutions were analyzed by CE and IC. Typical electropherograms of the 10-ppb anion standard solution and a wafer extraction solution are shown in Figure 2, while Figure 3 shows typical chromatograms of a 10-ppb NH₄⁺ standard solution and a wafer extraction solution. Figure 4 plots the correlation between the CE and IC results for NH₄⁺ and Cl^– standard solutions, demonstrating that the two techniques yield similar data for these two ions. The squares of the correlation coefficients (R²) were 0.9955 for NH₄⁺ and 0.9897 for Cl^–. Figures 5 and 6 show the correlation results for the wafer extraction solutions. In these correlations, the squared correlation coefficients were 0.9672 for NH₄⁺ and 0.9243 for Cl^–. Levels of other surface anions were at or below the instruments' detection limits.

Figure 5: Correlation between CE and IC NH ₄⁺ results for wafer extraction solutions (y = 0.9607x + 16.735, R² = 0.9672).

Figure 6: Correlation between CE and IC Cl^– results for wafer extraction solutions (y = 0.9985x + 28.44, R² =0.9243).

Comparable detection limits have been obtained using CE and IC for routine surface nonmetallic ion analysis: NH₄⁺ = 20–40 ions/cm², Cl^– = 10 ions/cm², SO⁴⁼ = 5–10 ions/cm², F^–= 50 ions/cm², NO^3– = 30 ions/cm², NO^2– = 40 ions/cm², and HPO⁼ = 40 ions/cm².

Extraction Efficiency. To study the efficiency of the extraction system, three extractions were performed on each of four test wafers and subjected to CE analysis. The extraction efficiency was then calculated using the following equation:

where C₁, C₂, and C₃ are the ion concentrations in parts per billion of the first, second, and third extraction solutions,
respectively.

Test Sample	Extraction Efficiency (%)
	Cl–	SO4=	NO3–	NH4+
Wafer 1	93.75	92.58	100	100
Wafer 2	94.75	96.26	100	95.45
Wafer 3	100a	88.50	N/Ab	100
Wafer 4	94.99	81.91	100	93.48
Average	96.50	89.81	100	98.48
Standard deviation	2.68	6.15	0.00	3.29
a 100% extraction efficiency was obtained when ions in the second and third extraction solutions were below the instrument detection limit. b No NO3– was detected in the three extraction solutions.

Table II presents these results. More than 90% of the water-soluble ions detected (NH₄⁺ , Cl^–, SO₄⁼, and NO₃^–) were extracted from the wafer surface by the first extraction. Because the levels of other ions (such as F^–, Br^–, NO₂^–, and HPO₄⁼) on the test wafers were below the analytical method's detection limits, no extraction efficiency results could be obtained for these ions. It was also observed that ion concentrations in the second and third extraction solutions were very close to those from the extraction box blank solutions, which were collected using the same process but with no wafer sample in the extraction box. This result confirms that the wafer extraction system developed in our laboratory can eliminate extraction solution contamination from airborne ionic species that can occur when extraction is done in an ambient environment or in a nonisolated box.

Conclusion

Experiments have demonstrated that an extraction system can prevent airborne molecular species from contaminating extraction solutions, thereby improving the accuracy of measurements of nonmetallic ions on silicon wafer surfaces. Results of CE and IC analyses correlated well for anion and ammonium standard solutions and wafer extraction solutions. This system can be used to monitor wafer surface contaminants, identify and eliminate contamination sources in wafer production processes, and investigate the mechanisms of surface degradation in the presence of nonmetallic ions.

References

1. JD Sinclair, "Corrosion of Electronics—The Role of Ionic Substances," Journal of the Electrochemical Society 135, no. 3 (1988): 89C–95C.

2. W Kern, "The Evolution of Silicon Wafer Cleaning Technology," Journal of the Electrochemical Society 137, no. 6 (1990): 1887–1891.

3. D Yang et al., "Combating Chlorine Corrosion through Ion Chromatography," Precision Cleaning VI, no. 5 (1998): 17–23.

4. D Rathman, "Trace Analysis of Inorganic Contamination on Si-wafer Surface by Ion Chromatography," The Electronic Society Extended Abstracts 91–2 (October 1991): 816.

5. S Tan, R Liu, and H La, "Monitoring of Ionic Contamination on Silicon Wafer Using Ion Chromatography" (paper presented at the 1997 International Ion Chromatography Symposium, Santa Clara, CA, September 14–17, 1997).

6. E Kaiser et al., "Determination of Trace Anionic Contamination on Silicon Wafer Surfaces by Ion Chromatography" (paper presented at the 1997 International Ion Chromatography Symposium, Santa Clara, CA, September 14–17, 1997).

7. RA Carpio, R Mariscal, and J Welch, "Determination of Boron and Phosphorus in Borophosphosilicate Thin Film on Silicon Substrates by Capillary Electrophoresis," Analytical Chemistry 64 (September 1992): 2123–2129.

8. T Talasek et al., "Determination of Anionic Impurities on Silicon Wafers by Microextraction Electrophoresis," in Proceedings of the 1993 Microcontamination Conference (Santa Monica, CA: Canon Communications, 1993), 773–781.

9. J Boden et al., "Application of Capillary Zone Electrophoresis with an Isotachophoretic Initial State to Determine Anionic Impurities on As-polished Silicon Wafer Surfaces," Journal of Chromatography A 696 (1995): 321–332.

10. T Ehmann et al., "Optimization of the Electrokinetic Sample Introduction in Capillary Electrophoresis for the Ultra Trace Detection of Anions on Silicon Wafer Surfaces," Chromatographia 45 (1997): 301–311.

11. P Sun et al., "Molecular and Ionic Contamination Monitoring for Cleanroom Air and Wafer Surfaces," in Proceedings of SPIE: In-line Characterization Techniques for Performance and Yield Enhancement in Microelectronic Manufacturing 3215 (Bellingham, WA: SPIE, 1997), 118–127.

12. P Sun and M Adams, "Silicon Wafer Surface Ionic Contamination Analysis by Capillary Electrophoresis and Ion Chromatography" (poster presented at the 1998 Pittsburgh Conference, New Orleans, March 1–5, 1997).

13. W Kern, "Overview and Evolution of Semiconductor Wafer Contamination and Cleaning Technology," in Handbook of Semiconductor Wafer Cleaning Technology, ed. Werner Kern (Park Ridge, NJ: Noyes, 1993), 19.

Peng Sun, PhD, is a senior staff engineer in the technology department of MEMC Electronic Materials (St. Peters, MO), where his responsibilities include the development of analytical methods for silicon wafer surface contamination analysis, cleanroom airborne contamination control, and cleanroom materials certification. Before joining MEMC in 1996 he was a senior research fellow at the American Health Foundation (Valhalla, NY), specializing in PAH adducts analysis in human tissues by gas chromatography—mass spectrometry. The principal author of more than 15 research papers in the separation sciences, Peng received his PhD in analytical chemistry from the State University of New York at Binghamton in 1994. (Peng can be reached at 314/279-7460 or psun@memc.com.)

Marty Adams has been on the staff of the MEMC technology department analytical lab for four years. His areas of responsibility include routine operation and analytical development for capillary electrophoresis, ion chromatography, and thermal desorption gas chromatography—mass spectrometry—atomic emission equipment. He attended Oglethorpe University (Atlanta) and the University of Missouri at Rolla.

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