Extreme Silicon

Monitoring organics on wafer surfaces using thermal desorption GC-MSD/AED

Peng Sun and Martin Adams, formerly of MEMC Electronic Materials; and Terrie Bridges, MEMC Electronic Materials

By combining mass spectrometry and atomic emission detectors, an analytical technique can provide both qualitative and quantitative data.

Because their effects on semiconductor devices are numerous and destructive, organic contaminants on wafer surfaces must be monitored and controlled. It has been reported that the presence of organic contamination can change surface hydrophobicity, lower breakdown voltage, form silicon carbide, affect oxide growth and quality, cause unintentional doping, contribute to degradation haze formation, and generate post-CVD defects.^1–4 To prevent such effects, various techniques have been used to analyze organic contaminants on silicon wafer surfaces, including Fourier transform infrared spectroscopy, ion mobility spectrometry, time-of-flight ion mass spectrometry, x-ray photoelectron spectroscopy, and thermal desorption gas chromatography–mass spectrometry.^5–11 However, all these methods have limitations. They either lack the capability to identify specific contaminants or are unable to quantify total organics accurately.

This article describes a new technique that uses thermal desorption gas chromatography (TD-GC) coupled with a mass spectrometry detector (MSD) and an atomic emission detector (AED) for organic contaminant analysis. By combining the superior separation ability of GC, the accurate compound-identification ability of an MSD, and the sensitive elemental-analysis ability of an AED, TD-GC-MSD/AED can provide simultaneous qualitative and quantitative measurements of organic contaminants on silicon wafers. The article also discusses the use of this combined technique to monitor organic contamination in wafer manufacturing processes and environments at the MEMC Electronic Materials facility (St. Peters, MO), detailing the detection procedures and their capabilities.

The TD-GC-MSD/AED Technique

Apparatus. The TD-GC-MSD/AED system used in the work described in this article consists of a Model 6890 GC interfaced with a Model 5970 MSD and a Model G2350 AED, all from Hewlett-Packard (Palo Alto, CA). The MSD is employed for the identification of organic contaminants, and the AED is used for the quantification of total organic carbon, total organophosphorus, and other elements of interest. Samples are introduced into the GC separation column via an ATD 400 automated thermal desorption unit from PerkinElmer (Wellesley, MA). An HP-5 capillary column (30 m x 0.32 mm x 0.5 µm) was used for the GC separation. A simple schematic of the system is provided in Figure 1.

 

Figure 1: Schematic of the TD-GC-MSD/AED organic analysis system.

Sample Preparation and Analysis. To analyze its surface contaminants, the wafer sample is first cut into small strips and placed into a wafer desorption tube. This tube is then attached to a sample collection/desorption tube that is packed with graphitized carbon black materials, after which the wafer desorption tube is placed into a furnace, where it is heated to 275°C and held at that temperature for 30 minutes. The organic contaminants that have desorbed off the wafer surface during the heating process are transferred into the sample collection/desorption tube by a flow of high-purity nitrogen. Next, the sample collection/desorption tube is loaded into the automated TD unit and the organic molecules are injected from that unit into the GC column for separation. A T-shape column splitter at the end of the GC column ensures that a portion of each separated sample goes to the AED and the remainder goes to the MSD. For total carbon analysis, n-hexadecane (n-C^₁₆H^₃₄) is used as the standard, and tributyl phosphate ((C^₄H^₉)^₃PO^₄) is used as the standard for organophosphorus quantification.

Comparison of AED and MSD Capabilities. Mass spectrometry is a compound-dependent detection technique that is based on the different fragmentation behaviors of dissimilar compounds. The structures of unknown organic contaminants are identified by comparing their unique mass spectra with standard spectra stored in a mass spectrum library. Individual sample components must be quantified using standards of the various compounds of interest. Because of the technique's compound-dependent nature, errors tend to be large when an MSD is used to perform total organic analyses with a single standard. Analyses performed by an AED, on the other hand, are compound independent. The signal intensity of an element of interest is proportional to the total quantity of the element in the compound and is independent of the structure of the compound. This feature makes it possible to conduct accurate total organic analyses by using a single standard compound.

Table I compares AED and MSD quantification results for 19 different compounds using n-hexadecane as the only standard. The data show that errors were small for most compounds analyzed with an AED (only one compound showed an error >10%). When an MSD was used, however, the errors ranged from –58 to 106%. The relative standard deviations shown in the table indicate that the AED and the MSD have comparable reproducibilities.

 

Compound AED MSD RSDa (%) Error (%) RSDa (%) Error (%) C8H18 C10H22 C12H26 C14H30 C16H34 C18H38 C20H42 C10H20 C10H16 C6C12D4 C10H30O5Si5 C12H8O C16H30O4 C17H34O2 C16H22O4 C12H27O4P C17H36 C8H10NOPS C9H15C16O4P 2.9 2.20 0.60 0.80 1.30 1.60 4.00 1.40 2.00 1.90 1.30 2.00 4.50 4.10 1.20 1.60 0.60 4.20 5.60 –4.00 –0.50 1.60 –0.40 –0.80 –2.40 –2.30 –0.60 –9.20 –4.60 14.90 –7.70 0.20 –7.20 –6.60 –0.40 6.20 –1.00 6.9 0.60 0.80 0.80 0.80 0.30 1.30 7.30 2.20 1.60 2.10 0.50 2.20 2.80 3.10 4.10 2.80 3.10 2.80 4.9 –23.70 –22.30 –15.60 –10.60 –2.10 –3.60 –4.90 –35.60 –47.60 –22.90 48.60 –50.90 –4.00 –21.10 –29.10 –58.00 –11.50 –13.00 105.8 aN = 5

Table I: Comparison of AED and MSD quantification capabilities using n-hexadecane as the single standard.

Figure 2 demonstrates the relative accuracy of the AED and the MSD by comparing the calculated carbon quantities and the actual carbon quantities (in nanograms) of different compounds introduced into the GC instrument. In this experiment, n-hexadecane was again used as the standard for quantification. The AED achieved a very good match between the calculated values and the actual injected values (y = 0.9891x – 0.399, R² = 0.9941), while the MSD did not (y = 0.311x + 39.322, R² = 0.5553).

 

Figure 2: Comparison of carbon amounts calculated by an AED and an MSD versus actual carbon amounts injected into the GC instrument (calculations were done using n-hexadecane as the single standard).

Monitoring Organic Contaminants

Two basic capabilities are required of the analytical techniques used to monitor silicon wafer surface organic contaminants: They must be able to identify individual contaminants and to quantify the total amount of organic contaminants or of specific compounds of interest. The TD-GC-MSD/AED technique meets both of these criteria. As explained above, when an AED and an MSD are interfaced with a GC system, it is possible to acquire qualitative and quantitative information simultaneously.

An example of these capabilities is provided in Figure 3, which presents four chromatograms of a wafer sample contaminated with organics. The total ions chromatogram (a) was produced by the MSD and shows a major contaminant peak at ~18 minutes. The other three panels (b, c, and d) are AED chromatograms for the carbon 193, sulfur 181, and phosphorus 178 channels, respectively. The phosphorus chromatogram also has a major peak at ~18 minutes. Based on the AED calculation, the organophosphorus contamination level was 5.2 x 10¹² atoms/cm², which is significantly higher than the maximum acceptable surface level. By comparing the mass spectrum of the unknown peak at 18 minutes in the MSD chromatogram to the mass spectrum library, this organophosphorus contaminant was identified as Fyrol PCF (tri-(ß-chloroisopropyl phosphate)), a commonly used flame retardant that can cause unintentional n-type doping on silicon wafers.¹² The matching spectra are shown in Figure 4.

 

Figure 3: Chromatograms of a wafer sample that had been contaminated with organics: (a) total ions MSD chromatogram, (b) AED chromatogram for carbon 193 channel, (c) AED chromatogram for sulfur 181 channel, and (d) AED chromatogram for phosphorus 178 channel.

Figure 4: Identification of Fyrol PCF by a comparison of mass spectra: (a) unknown peak, (b) spectrum of unknown peak, and (c) spectrum of first peak in Fyrol PCF.

Wafers can become contaminated by organic compounds from a variety of sources, such as cleanroom air, process chemicals (solids, liquids, and gases), wafer carriers and packaging boxes, cleanroom construction materials or components (e.g., sealant used in ULPA filters), process equipment materials, and fab personnel. Organic contaminants commonly seen on normal production wafers include plasticizers (e.g., TXIB, DBP, DOP), antioxidants (e.g., BHT and its derivative compounds), hydrocarbons, low levels of siloxanes, and low levels of organic amines. Organic contaminants detected on contaminated wafers include organophosphates (e.g., triethyl phosphate [TEP], Fyrol), surfactants used in packaging and cleaning, glycol compounds (e.g., solvents in floor cleaning solutions), high levels of siloxanes, and high levels of organic amines. Because high levels of these latter compounds on the wafer surface tend to cause problems during device fabrication, it is important to monitor surface contamination at various process steps. Figure 5 presents monitoring data that illustrate the trend of total surface organics on wafers from two different processes. Although the test wafers from process A generally had lower levels of surface organic contamination than those from process B, monitoring detected an organic contamination upset in process A (labeled in the figure). Through the use of spectra matching, the cause was identified as the surfactant used in cleaning the wafer packing boxes.

 

Figure 5: Trend chart of total surface organics on wafers from two different processes.

Since airborne organics are a major source of wafer surface contamination, it is also important to monitor cleanroom manufacturing environments. At MEMC's manufacturing facility this is accomplished with a witness wafer exposure method. A thermally treated wafer (free of surface organic contamination) is exposed to the ambient of the area of interest for 24 hours and then analyzed by the TD-GC-MSD/AED method for organic contamination. Figure 6 shows examples of the air-monitoring data from two cleanroom areas. Area A, which had chemical filters installed in the air supply system, consistently had lower levels of airborne organics than area B, where no chemical filters were used.

 

Figure 6: Trend charts of airborne organics in two cleanroom areas: (a) area A and (b) area B.

Conclusion

The TD-GC-MSD/AED technique has been found to be a powerful tool for monitoring organic contaminants on silicon wafer surfaces. The method combines the identification power of MSD and the accurate elemental quantification capability of AED, thereby providing qualitative and quantitative information on surface organic contaminants simultaneously. This method is being used at MEMC Electronic Materials to identify and eliminate organic contamination in silicon wafer manufacturing processes and in the cleanroom environment.

References

01.W Kern and DA Poutinen, RCA Review 31 (1970): 187–206.

02.P Smith and PM Lindley, "Analysis of Organic Contamination in Semiconductor Processing," in Characterization and Metrology for ULSI Technology: 1998 International Conference, ed. DG Seiler et al. (Woodbury, NY: American Institute of Physics, 1998), 133–139.

03.M Tamaoki et al., "The Effect of Airborne Contaminants in the Cleanroom for ULSI Manufacturing Process," in Proceedings of the IEEE/SEMI Advanced Semiconductor Manufacturing Conference (Piscataway, NJ: IEEE, 1995), 322–326.

04.EJ Mori, JD Dowdy, and LW Shive, "Correlating Organophosphorus Contamination on Wafer Surfaces with HEPA-Filter Installation," Microcontamination 10, no. 11 (1992): 35–38.

05.S Ojima et al., "Establishment of Complete Cleaning Technology for Hydrocarbon Contamination on Si Wafer Surface," in Proceedings of the 41st Annual Technical Meeting of the IES (Mount Prospect, IL: Institute of Environmental Sciences, 1995), 441–446.

06.KJ Budde et al., "Application of Ion Mobility Spectrometry to Semiconductor Technology: Outgassing of Advanced Polymers under Thermal Stress," Journal of the Electrochemical Society 142, no. 3 (1995): 888–897.

07.A Licciardello et al., "Effect of Organic Contaminants on
the Oxidation Kinetics of Silicon at Room Temperature," Applied Physics Letters 48 (1986): 41–43.

08.J. Lee et al., "Direct Surface Analysis of Organic Contamination on Si Wafers," in Microcontamination Conference Proceedings (Santa Monica, CA: Canon Communications, 1994), 464–475.

09.M Camenzind, "Identification of Organic Contamination in Cleanroom Air, on Wafers, and Outgassing from Gloves and Wafer Shippers," in Semiconductor Pure Water and Chemicals Conference Proceedings (Sunnyvale, CA : SPWCC, 1996), 352–373.

10.K Saga and T Hattori, "Identification and Removal of Trace Organic Contamination on Silicon Wafers Stored in Plastic Boxes," Journal of the Electrochemical Society 143, no. 10 (1996): 3270–3284.

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 (Bellingham, WA: SPIE, 1997), 118–127.

12.JA Lebens et al., "Unintentional Doping of Wafers Due to Organophosphates in the Clean Room Ambient," Journal of the Electrochemical Society 143, no. 9 (1996): 2906–2909.

Peng Sun, PhD, is a senior microcontamination lab engineer at Intel (Santa Clara, CA). Previously he was a senior staff engineer in the technology department of MEMC Electronic Materials in St. Peters, MO, where he was responsible for developing analytical methods for silicon wafer surface contamination analysis, cleanroom airborne contamination control, and cleanroom material 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 GC-MS. Sun is the principal author of more than 15 research papers in the separation sciences. He received his PhD in analytical chemistry from the State University of New York at Binghamton in 1994. (Sun can be reached at 408/765-8784 or peng.sun@intel.com.)

Martin Adams is a customer service representative in St. Louis for the chromatography division of Varian (Palo Alto, CA). Previously he worked at MEMC for four years in the analytical lab of the technology department. He was responsible for the routine operation and analytical development of capillary electrophoresis, ion chromatography, and TD-GC-MSD/AED equipment. He attended Oglethorpe University in Atlanta and the University of Missouri at Rolla. (Adams can be reached at 314/939-3155 or martin.adams@varianinc.com.)

Terrie Bridges has been at MEMC for 12 years and works in the analytical lab of the technology department. She is responsible for atomic force microscopy, capillary electrophoresis, ion chromatography, and TD-GC-MSD/AED. She received a BA in chemistry from Southeast Missouri State University (Cape Girardeau) in 1986. (Bridges can be reached at 636/474-7460 or gbridges@memc.com.)

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