AVS1997 Session VT-WeA: Vacuum Gauging and Partial Pressure Measurements

Wednesday, October 22, 1997 2:00 PM in Room N

Wednesday Afternoon

Start	Invited?	Item
2:00 PM		VT-WeA-1 A Miniature Quadrupole RGA For Process Monitoring At Millitorr Pressures D.H. Holkeboer, T.L. Karandy, F.C. Currier, L.C. Frees, R.E. Ellefson (Leybold Inficon, Inc.) Monitoring of gas composition during processes in the millitorr range is not possible with a conventional RGA unless a pressure reduction pumping system is used. A miniature sensor, having dimensions compatible with the mean-free-path of gas molecules can operate directly at these atmospheres while being much less expensive and compact than conventional apparatus. This paper describes a quadrupole RGA which has been scaled down by a factor of eight to operate at pressures up to 10 millitorr. This sensor requires different construction methods and drive electronics than a traditional RGA. A unique small ion source with high output has been developed. The small sensor has resolution characteristics very similar to the larger model, but sensitivity is reduced in proportion to the cross-sectional area of the quadrupole. Response is linear with increasing pressure to a greater level than would be expected from simple scaling. Because of the high operating pressures, sensor life is more affected by reactions with the gas environment. Extensive measurements of sensor performance stability have been made and will be presented along with millitorr range process data.
2:20 PM		VT-WeA-2 Response Time and Sensitivity Optimization of Faraday Cup Equipped Miniature Quadrupole Mass Spectrometers F.D. Vaught (Ferran Scientific, Inc.) Residual gas analyzers using quadrupole mass spectroscopy in industrial process control applications and gas chromatography often require fast response, wide instantaneous dynamic range, and high scan speeds. Hardware solutions to these challenges, e.g., channeltrons and variants, while able to deliver necessary sensitivity, suffer from high, uncertain gain drift and drastic gain loss when approaching process pressures. While the purposes of some scientific experiments are often served despite such instrument innacuracies, many otherwise desireable quadrupole industrial process control applications are intolerably impacted. Methods are described for circumventing inherent quadrupole response and sensitivity limitations employing signal processing techniques not commonly applied to quadrupoles. Results of applying these methods to a standard Faraday cup equipped Micropole¹ show that the resultant system exhibits the advantages of gain stability, low cost, and the capability of optimization suitable for process control problems. ¹R.J. Ferran and S. Boumsellek, J. Vac. Sci. Technol. A 14(3), May/Jun 1996
2:40 PM		VT-WeA-3 High Resolution Quadrupole Mass Spectrometry Through Operation Using Alternate Mathieu Stability Regions R.E. Pedder, J. Wei (ABB Extrel) The analysis of isobaric (same nominal mass) species requires high mass resolution. For example, carbon monoxide and nitrogen both have a nominal mass of 28 amu, but their exact masses differ by 0.0112 amu. Quantitative measurement of adjacent mass species at high resolution can be quite challenging for species with widely differing concentrations. The figure of merit which best characterizes this suitability is called 'Abundance Sensitivity, typically identified as the relative contribution of a given species to the baseline level of its nearest neighbors. Under unit mass resolution conditions for a quadrupole mass filter, abundance sensitivity can be one part in 10^5, 10^6, or even 10^7, depending on quadrupole configuration and operating frequency. Such a measurement can be translated to mean the contribution of the intensity at mass M bleeding into the peak at mass M-1 is one part in 10^5 of the intensity measured at mass M. Operating a normal quadrupole at high resolution results in dramatically poorer abundance sensitivity at the limits of resolution, where abundance sensitivity approaches one for one (i.e. the peaks are un-resolved). There exist multiple regions in RF-DC space which allow for stable transmission of ions through the quadrupole, in addition to the standard scan line of DC voltage = 1/6 RF voltage. These alternate Mathieu Stability Regions offer dramatic increases in mass resolution, and more importantly, dramatic improvements in abundance sensitivity. In our lab, we have characterized the performance and various figures of merit (resolution, sensitivity, abundance sensitivity...) for quadrupoles operated in the normal first stability region as well as two alternate stability regions. This presentation will provide a brief primer in quadrupole theory and summarize the results of our studies of the separation of various isobaric species measured using a quadrupole mass filter operated under normal and higher resolution alternate stability modes.
3:00 PM	Invited	VT-WeA-4 Advances In Partial Pressure Measurements for Semiconductor Processes R.E. Ellefson (Leybold Inficon) Measurement of residual gases in semiconductor process chambers with a residual gas analyzer (RGA) has been used for diagnosing vacuum problems for a number of years. Monitoring of gas composition at normal process pressures has been done by pressure reduction sampling and measurement of the reduced pressure with a RGA or, for lower impurity detection, with a closed ion source (CIS). A recent development is the miniature RGA that operates at millitorr process pressures allowing direct immersion of the RGA in selected processes for monitoring partial pressures and gas composition. This talk will review partial pressure analysis methods and present the operating range (pressure and compositon), accuracy, speed of analysis and analytical benefit to be expected with each method. Examples of process monitoring for the principle semiconductor processes will be given. The examples include physical vapor deposition, chemical vapor deposition, metal organic chemical vapor deposition and etch processes and process exhaust monitoring. As partial pressure analysis evolves as a tool for characterizing process gas composition, the issue of management of the large amount of data that can be generated must be addressed. Options for data reduction protocols, display and storage will be presented.
3:40 PM		VT-WeA-6 Stable Cancellation of X-Ray Errors in Bayard-Alpert Gauges B.R.F. Kendall (Elvac Laboratories); E. Drubetsky (The Fredericks Company) X-Ray induced emission of electrons from the ion collectors of Bayard-Alpert gauges has long been recognized as a cause of falsely high readings at low pressures. Less well known is the existence of a reverse X-Ray effect leading to a superimposed, but usually smaller, error signal in the opposite direction. First explained by Redhead and others in the 1960s as being caused by photoemission from the gauge envelope, this phenomenon was immediately recognized as potentially useful in cancelling the forward X-Ray effect. Although promising experimental results were obtained with special gauge tubes, these early researchers concluded that the cancellation process was too unstable to be of any practical use. We report the results of an attempt to stabilize the cancellation process by the use of identical materials at the two photoemission sites. Experimental Bayard-Alpert gauge tubes were built with gold coating on the insides of the gauge envelopes and on the surfaces of the ion collectors. The envelope coatings were operated at approximately 15-20 volts below the potential of the ion collector. Stable reduction of the X-Ray errors by a factor of more than 20 has been demonstrated. Useful improvements have also been observed in sensitivity and in stability at high pressures. Results of long-term stability tests at pressures in the 10^-10 Torr range will be described.
4:00 PM		VT-WeA-7 A New Standard for Low-Pressure Gauges C.R. Tilford, A.P. Miiller (National Institute of Standards & Technology) The National Institute of Standards and Technology (NIST) has completed and placed into service a new Ultrasonic Interferometer Manometer (UIM) with a range of 133 Pa (1 torr). The basic operation of this manometer is similar to the mercury UIMs used at NIST for many years, but the new UIM uses a low vapor-pressure oil as the manometer fluid. This offers several advantages for low-pressure measurements and calibrations: 1) After degassing of the oil, the reference pressure for the manometer is effectively zero - below 1 mPa. 2) The use of the low-density oil (compared to mercury) results in high sensitivity - a standard deviation of 0.4 mPa (3 microtorr) for repeated measurements, which is an order of magnitude better than that achieved by the mercury UIMs. 3) The greatest advantage, somewhat unexpectedly, has proven to be the absence of mercury and the elimination of measurement errors caused by incomplete equilibration of mercury vapor. This talk will describe the design and operation of the manometer, the analysis of its uncertainty, and its use as a calibration standard for low-pressure gauges. It will also summarize our experience with the problems caused by incomplete mercury equilibration, which are common to all vacuum measuring devices that use mercury.
4:20 PM		VT-WeA-8 Orifice Flow Calibration of Ion Gages for Industrial Use P.D. Levine, J.R. Sweda (Lockheed Martin Missiles & Space) The Primary Standards Laboratory at Lockheed Martin Missiles and Space in Sunnyvale, CA. is responsible for the calibration of ion gages which support production and development efforts throughout the company. Primary vacuum standards must be maintained on a continuing basis to ensure that quality and traceability requirements are met. In general, this has been accomplished by sending gages to the National Institute of Standards and Technology (NIST) for primary calibration. Given the increased emphasis on overhead reduction and the importance of timely calibration, an in-house system for creating primary reference standards has been developed. An orifice flow system utilizing a constant volume pressure drop flowmeter is used to generate known pressures in a vacuum chamber on which are mounted a variety of vacuum gages. Previous work has verified system function utilizing spinning rotor gages calibrated by NIST. This paper describes the calibration of ion gages for use as primary reference standards from 10^-8 to 10^-3 Torr. Data is presented on both glass encapsulated Bayard-Alpert gages and a newly developed gage designed for high stability. An analysis is included which shows the uncertainty levels to be higher than those associated with NIST calibrated standards. When weighed against practical considerations and user requirements, it is concluded that the system is adequate for the industrial environment.
4:40 PM		VT-WeA-9 Pushing the Spinning Rotor Gauge Sensitivity by a Compression Stage B.E. Lindenau (Forschungszentrum Jülich, Germany) The principle of increasing the pressure sensitivity of a vacuum gauge by gas compression is well known from the Mc Leod gauge. The applicability of this principle to Spinning Rotor Gauges ( SRG ) was now demonstrated by inserting a miniature molecular drag compression stage at the entrance of the SRG sensor tube. The experimental compressor is based on a fast rotating ball with a gas feeding channel halfway encircling its equator. A pressure gradient is build up by the drag rotor and enables continuous gauging of the amplified pressures in the SRG sensor tube. With nitrogen, the gauge sensitivity was enhanced by a factor of two at 300 m/sec equatorial speed of the drag rotor. As the compression is proportional to the molecular weight, the device can likewise be used to determine the mean molecular weight of a gas enclosed from the pressure ratios at distinct drag rotor speeds.
5:00 PM		VT-WeA-10 Application of Calculated Physical Adsorption Isotherms to a Radon Sensor J.P. Hobson (National Vacuum Technologies, Inc., Canada); A. Appleby, I.S. Kim, G.H. Sigel, Jr. (Rutgers University) Recently a radon sensor using a glass scintillator in a fiber bundle structure was developed at Rutgers University¹. This instrument employs a flow of atomospheric air carrying the radon gas and it progeny in trace amounts through the bundle at room temperature. A central problem of interpretation is to distinguish whether the radioactive Rn atoms whose alpha decay is detected in the sensor, are in the gas phase in the narrow spaces between the fibers or are adsorbed on the surface of the fibers. At the 43rd National Symposium of the AVS (1996) Hobson² reported calculated physical adsorption isotherms of radon on a heterogeneous surface, over a wide range of pressures, coverages, and temperatures. The latter explicitly included room temperature (300 K). The question arises whether this calculated isotherm is applicable to Rutgers' sensor. From radioactive considerations³ it is estimated in one experiment that the partial pressure of Rn is near 10^-9 torr. The calculated isotherm at 300 K gives a relative coverage Θ of 10^-11 monolayer, an extremely low value. Appleby and colleagues kept careful track of room temperature variations which fluctuated between 290 and 310 K, finding that their sensor counting rates varied by 30% between these limits, being highest when the temperature was lowest. It is difficult to explain such a large temperature sensitivity without invoking physical adsorption on the glass fibers. The calculations were repeated for 290 and 310 K and were indeed found to provide semi-quantitative agreement with the experimental results. These findings will be discussed in detail. ¹Kim I.S. et al., Nuclear Instruments and Methods in Physics Research A, 356, 1995, 537-543. ²Hobson J.P., Jour. Vac. Sci. Technol. A, in press. ³Kim I.S., Appleby A., and Sigel Jr. G.H., J. of Porous Materials, in press.