Choosing the right ionization gauge for high-vacuum processes
03/01/2005
The selection of high-vacuum gauge technologies for measurements in pressure-dependent semiconductor processes can affect overall cost and manufacturing yields depending on the application and the environment in tool chambers. This article examines the characteristics and developments in two primary choices for high-vacuum measurement technologies based on hot-cathode or cold-cathode ionization gauges.
In wafer fabs, vacuum gauges are often used to validate the start of a pressure-dependent process cycle, such as physical vapor deposition (PVD). These gauges also are widely used to measure the vacuum pressure during ion implantation. Many scientific and analytical applications rely on repeatable high-vacuum measurements as well. While high-vacuum gauges show up in a wide range of systems, measurement requirements are unique among various applications. It is crucial to select the right gauge technology based on the needs of an application as well as overall cost targets [1].
For decades, both hot-cathode and cold-cathode ionization gauges have been employed in high-vacuum measurement applications. Ionization gauges determine vacuum by measuring ion current. In both hot- and cold-cathode gauges, energetic electrons ionize gas molecules, causing them to have a net positive charge. These ionized molecules arrive at the collector electrode as ion current, which is measured by an electronic circuit. The rate at which electrons interact with gas molecules is proportional to the density of gas molecules; thus, the ion current is proportional to the gas density, or pressure [2].
Hot-cathode ionization gauges use a heated wire to provide a source of electrons, which are accelerated toward a grid where their controlled energy is appropriate to ionize neutral gas molecules and atoms (Fig. 1). The created ions are in an electric field that directs them toward an ion collector electrode. The electrode is attached to circuitry, which can measure very small electric currents. The current is characteristic of the molecular and atomic density of a gas in the gauge. In other words, the current is characteristic of the pressure at a specific temperature. In this type of gauge, the electron emission current is carefully controlled. The most common hot-cathode ionization gauge is the conventional Bayard-Alpert (B-A) gauge.
 Figure 1. Basic elements of hot-cathode gauge (left) and cold-cathode gauge (right). |
In a cold-cathode gauge, gas molecules are ionized by a circulating space charge current of electrons trapped by crossed electric and magnetic fields as seen in Fig. 1. The source of electrons in these gauges comes from an initial startup sequence using a cosmic ray, a radioactive source, or the momentary operation of a hot cathode built into the same gauge [3]. The continuing supply of electrons comes from the ionization events and those ions striking the ionization volume enclosure. The ionizing electron current cannot be controlled, and it depends in part on the state of the surface of the cold cathode. The ions are directed toward the cathode, where they are measured as a very small electric current characteristic of the gas density in the gauge, or pressure. Several varieties of cold-cathode gauges exist, including the Penning, the magnetron, and the inverted magnetron [2].
Gauge designs are evolving
In the past decade, improved hot-cathode ionization gauges have been introduced, offering increased accuracy and stability compared to conventional B-A ion gauges [4]. The newer hot-cathode gauge designs and technology provide precision electrodes and careful control of the physical electronics inside the gauges. Older conventional B-A gauges are known to have lower accuracy and stability characteristics due to variations in electrode geometries, changes in emission density distribution along the hot cathode, and potential changes in the glass enclosures of these gauges [5].
Several design changes have improved accuracy and stability in hot-cathode ionization gauges. For example, tensioned filaments and precision wound anodes are now used to ensure proper electrode geometries in hot-cathode gauges, such as the Granville-Phillips Stabil-Ion series (Fig. 2). The enhancements provide a fixed electrode geometry, which maintains constant trajectories of ions and electrons over time. These features - combined with an all-metal construction instead of glass enclosures - result in long-term stability and greater consistency in measurement performance between gauges, compared to traditional B-A gauges [6].
Cold-cathode gauges, based on the inverted magnetron and double inverted-magnetron designs, offer improved performance over the Penning gauge. These gauges depend on both electric and magnetic fields to generate very long electron trajectories, providing a higher probability of ionization/electron than the hot-cathode gauge [3].
Multiple factors to consider
When determining the right gauge for a vacuum application, consideration must be given to overall cost-of-ownership (CoO). This takes into account not just the initial purchase price, but also gauge stability, lifetime, maintenance costs, pressure measurement range, corrosion resistance, accuracy, and reproducibility (gauge-to-gauge). Some applications may involve vapor chemistry that can result in a range of chemical interactions with materials used in gauges. Additionally, IC processes can result in numerous types of film deposits on many of the gauge’s components. Consideration must be given to the gauge design as it relates to these factors. Users should ask some questions: How stable or repeatable are gauge measurements over time? How often can a system be shut down for periodic maintenance or gauge replacement? Will the gauge design help reduce the buildup of deposited layers from semiconductor processes?
Gauge performance can affect process throughput, equipment uptime, and overall cost in a number of ways. If an ionization gauge reads too low due to output drift or if a replacement gauge performs differently, the true pressure in a system may not be low enough to provide the base pressure required for a specific process step or environment. This problem can result in the production of poor-quality semiconductors. If a gauge reads too high, the process base-pressure set point will require extended vacuum pumping beyond what is actually needed. This wasted pumping time will extend the process cycle, reducing throughput. Wasted pumping time can be difficult to identify, but it will contribute to higher overall costs. Both equipment uptime and overall product quality play a factor in total CoO. Consequentially, gauge selection criteria should include required process uptime and the impact of product quality costs associated with gauge performance drift over time.
Stability over time and reproducibility of measurements (gauge-to-gauge) are two performance factors for gauges that influence process uptime and product quality. Cold-cathode gauges can experience a slow degradation in measurement accuracy over time, especially in dirty process environments where materials might be deposited inside the sensor. Instability over time also can be a problem in hot-cathode gauges, such as traditional glass B-A units. To address these problems, sophisticated computer models are used to study the electrical fields in and around hot-cathode gauge electrodes. Computer modeling also is applied to the trajectories of electrons for ionization and the positively charged ions resulting from ionization. The result has been new designs for more robust, stable ion gauges.
To control the quantity of collected ions, which indicate the density of gases (or pressure at constant temperature), the quantity and paths of electrons for ionization must be controlled. Attention also must be given to the collection factor of ions created. In Stabil-Ion gauges, for example, the design provides for electron paths independent of emission density along the hot filament. This addresses the one parameter in hot-cathode gauges that is not controllable. This gauge design also provides for a very high collection fraction of ions created by means of the shape of the electric fields interior to the anode [6]. These same design aspects, along with special construction of the electrodes that retain their shape and location, achieve gauge-to-gauge reproducibility and measurement repeatability.
The function of a high-vacuum gauge is to determine the pressure in a working chamber. However, a high-vacuum gauge reports only on the pressure in the gauge itself. Gauge design and installation (location and flange seal material) are key parameters to ensure that gauge measurements accurately represent the actual chamber pressure. A performance characteristic that opposes this objective is gauge pumping, which can occur when a conductance restriction exists between the gauge and the chamber. When this happens, the ionization gauge can pump its own internal volume to a lower pressure than what exists in the chamber. The cold-cathode gauge has a faster pumping speed than hot-cathode gauges, and may have more conductance restriction.
Another distinguishing feature to consider in gauge selection is turn-on time. Hot-cathode gauges establish electron emission with filament activation and will read pressure very rapidly. This is not the case with cold-cathode gauges, however, which normally require much longer start-up times in ultrahigh-vacuum applications unless additional hardware is included in the gauge [3].
Gauge lifetime and unscheduled downtime
Two gauge performance factors that can adversely affect the overall performance of vacuum-based processes are failure to operate and pressure readout being so far from true pressure that the measurement is no longer useful. The end of usefulness for the hot-cathode ionization gauge occurs with the loss of the thermionic emission electrode (the filament). The selection of proper filament materials, along with control over heating currents and venting in hot-cathode gauges, can minimize unscheduled downtime.
Most hot-cathode ionization gauges are now available with either oxide-coated iridium or bare tungsten filaments. An oxide-coated iridium cathode is a good choice for processes that chemically react with tungsten. Tungsten is the choice for some process chemistries that degrade the electron emission qualities of an oxide-coated filament. Tungsten filaments are more susceptible to failure when exposed to a high partial pressure of air or water vapor. Hot cathodes are known to last for many years when the filament material is compatible with the process chemistry (see table). Many hot-cathode gauges have a backup filament. In these designs, the failure of one filament will not normally cause a process shutdown.
 |
Hot-cathode gauges also have the ability to clean themselves of deposits from process gases. A degassing function is initiated by the gauge controller, which provides sufficient heating to drive off deposits from the surface of components inside the gauge. The degas function heats the grid by electron bombardment or by resistive heating, depending on the gauge grid design. The ion collector is heated by radiated heat from the grid. The hot-cathode temperature rises during the degas process. When allowed to proceed past several minutes, the degassing process will increase the gauge wall temperature to the point that contaminants are also driven off this surface.
Cold-cathode gauges become useless when pressure readouts stray far from true pressure conditions in chambers. This phenomenon is due to the cathode surfaces changing their emission characteristics because of deposit or coating buildup from processes and vacuum chamber environments. Cold-cathode gauges often need to be removed from the chamber and mechanically cleaned to restore their vacuum measuring capability. The realignment of the interior parts and the exterior magnets is critical when cleaning these gauges for reuse in chambers.
Conclusion
Selecting the proper ionization gauge for an application will depend on a number of factors, including long-term accuracy or stability, gauge-to-gauge reproducibility, corrosion resistance, maintenance costs, reliability, and useful life. All these factors contribute to the overall cost-of-ownership of the device. Careful consideration of the process environment, required performance over time, allowance for routine maintenance, and expected throughput will facilitate the selection of the appropriate gauging technology.
Acknowledgments
Granville-Phillips and Stabil-Ion are registered trademarks of Helix Technology.
References
- P. Burggraaff, “Vacuum Gauges: Do They Rule the Fab?” Semiconductor International, p. 58, Nov. 1993.
- P.H. Singer, “Trends In Vacuum Gauging,” Semiconductor International, p. 78, March 1992.
- M. Hablanian, High-Vacuum Technology, A Practical Guide, pp. 444-445, Marcel Dekker Inc., 1997.
- P.C. Arnold, S.C. Borichevsky, “Nonstable Behavior of Widely Used Ionization Gauges,” J. Vac. Sci. Tech., pp. 568-573, March/April 1994.
- D.G. Bills, “Causes of Nonstability and Nonreproducibility in Widely Used Bayard-Alpert Ionization Gauges,” J. Vac. Sci. Tech., pp. 574-579, March/April 1994.
- P.C. Arnold, D.G. Bills, M.D. Borenstein, S.C. Borichevsky, “Stable and Reproducible Bayard-Alpert Ionization Gauge,” J. Vac. Sci. Tech., pp. 580-586, March/April 1994.
Paul M. Rutt is the chief scientist for Granville-Phillips vacuum instrumentation with Helix Technology Corp.
Steven Smith is the product marketing manager for Granville-Phillips vacuum instrumentation with Helix Technology Corp., 6450 Dry Creek Parkway, Longmont, CO 80503; ph 303/652-4536, fax 303/652-4484, e-mail [email protected].