Sputter-Ion Pumps
Applications: These pumps were originally developed for ultrahigh vacuum (UHV) systems and are admirably suited to this application, especially if the system is rarely vented to atmospheric pressure. Their main advantages are as follows.
1. High reliability, because of no moving parts.
2. The ability to bake the pump up to 4008C, facilitating outgassing and rapid attainment of UHV conditions.
3. Fail-safe operation if on a leak-tight UHV system. If the power is interrupted, a moderate pressure rise will occur; the pump retains some pumping capacity by gettering. When power is restored, the base pressure is normally reestablished rapidly.
4. The pump ion current indicates the pressure in the pump itself, which is useful as a monitor of performance.
Sputter-ion pumps are not suitable for the following uses.
1. On systems with a high, sustained gas load or frequent venting to atmosphere.
2. Where a well-defined pumping speed for all gases is required. This limitation can be circumvented with a severely conductance-limited pump, so the speed is defined by conductance rather than by the characteristics of the pump itself.
Operating Principles: The operating mechanisms of sputter-ion pumps are very complex indeed (Welch, 1991). Crossed electrostatic and magnetic fields produce a confined discharge using a geometry originally devised by Penning (1937) to measure pressure in a vacuum system.A trapped cloud of electrons is produced, the density of which is highest in the 10¯4 torr region, and falls off as the pressure decreases. High-energy ions, produced by electron collision, impact on the pump cathodes, sputtering
reactive cathode material (titanium, and to a lesser extent, tantalum), which is deposited on all surfaces within line-of sight of the impact area. The pumping mechanisms include the following.
1. Chemisorption on the sputtered cathode material, which is the predominant pumping mechanism for reactive gases.
2. Burial in the cathodes, which is mainly a transient contributor to pumping. With the exception of hydrogen, the atoms remain close to the surface and are released as pumping/sputtering continues. This is the source of the ‘‘memory’’ effect in diode ion pumps; previously pumped species show up as minor impurities when a different gas is pumped.
3. Burial of ions back-scattered as neutrals, in all surfaces within line-of sight of the impact area. This is a crucial mechanism in the pumping of argon and other noble gases (Jepsen, 1968).
4. Dissociation of molecules by electron impact. This is the mechanism for pumping methane and other organic molecules.
The pumping speed of these pumps is variable. Typical performance curves show the pumping of a single gas under steady-state conditions.
The pumping speed of hydrogen can change very significantly with conditions, falling off drastically at low pressures and increasing significantly at high pressures (Singleton, 1969, 1971; Welch, 1994). The pumped hydrogen can be released under some conditions, primarily during the startup phase of a pump. When the pressure is 10¯3 torr or higher, the internal temperatures can readily reach 5008C (Snouse, 1971). Hydrogen is released, increasing the pressure and frequently stalling the pumpdown.
Rare gases are not chemisorbed, but are pumped by burial (Jepsen, 1968). Argon is of special importance, because it can cause problems even when pumping air.The release of argon, buried as atoms in the cathodes, sometimes causes a sudden increase in pressure of as much as three decades, followed by renewed pumping, and a concomitant drop in pressure. The unstable behavior
is repeated at regular intervals, once initiated (Brubaker, 1959). This problem can be avoided in two ways.
1. By use of the ‘‘differential ion’’ or DI pump (Tom andJames, 1969), which is a standard diode pump in which a tantalum cathode replaces one titanium cathode.
2. By use of the triode sputter-ion pump, in which a third electrode is interposed between the ends of the cylindrical anode and the pump walls. The additional electrode is maintained at a high negative potential, serving as a sputter cathode, while the
anode and walls are maintained at ground potential. This pump has the additional advantage that the ‘‘memory’’ effect of the diode pump is almost completely suppressed.
The operating life of a sputter-ion pump is inversely proportional to the operating pressure. It terminates when the cathodes are completely sputtered through at a small area on the axis of each anode cell where the ions impact. The life therefore depends upon the thickness of the cathodes at the point of ion impact. For example, a conventional triode pump has relatively thin cathodes as compared to a diode pump, and this is reflected in the expected life at an operating pressure of 1x10¯6 torr, i.e., 35,000 as
compared to 50,000 hr. The fringing magnetic field in older pumps can be very significant. Some newer pumps greatly reduce this problem.
A vacuum chamber can be exposed to ultraviolet and x radiation, as well as ions and electrons produced by an ion pump, so appropriate electrical and optical shielding may be required.
Operating Procedures: A sputter-ion pump must be roughed down before it can be started. Sorption pumps or any other clean technique can be used. For a diode pump, a pressure in the 10¯4 torr range is recommended, so that the Penning discharge (and associated pumping mechanisms) will be immediately established. A triode pump can safely be started at pressures about a decade
higher than the diode, because the electrostatic fields are such that the walls are not subjected to ion bombardment (Snouse, 1971). An additional problem develops in pumps that have operated in hydrogen or water vapor. Hydrogen accumulates in the cathodes and this gas is released when the cathode temperatures increase during startup. The higher the pressure, the greater the temperature; temperatures as high as 900ºC have been measured at the center of cathodes under high gas loads (Jepsen, 1967).An isolation valve should be used to avoid venting the pump to atmospheric pressure. The sputtered deposits on
the walls of a pump adsorb gas with each venting, and the bonding of subsequently sputtered material will be reduced, eventually causing flaking of the deposits. The flakes can serve as electron emitters, sustaining localized (non-pumping) discharges and can also short out the electrodes.
Applications: These pumps were originally developed for ultrahigh vacuum (UHV) systems and are admirably suited to this application, especially if the system is rarely vented to atmospheric pressure. Their main advantages are as follows.
1. High reliability, because of no moving parts.
2. The ability to bake the pump up to 4008C, facilitating outgassing and rapid attainment of UHV conditions.
3. Fail-safe operation if on a leak-tight UHV system. If the power is interrupted, a moderate pressure rise will occur; the pump retains some pumping capacity by gettering. When power is restored, the base pressure is normally reestablished rapidly.
4. The pump ion current indicates the pressure in the pump itself, which is useful as a monitor of performance.
Sputter-ion pumps are not suitable for the following uses.
1. On systems with a high, sustained gas load or frequent venting to atmosphere.
2. Where a well-defined pumping speed for all gases is required. This limitation can be circumvented with a severely conductance-limited pump, so the speed is defined by conductance rather than by the characteristics of the pump itself.
Operating Principles: The operating mechanisms of sputter-ion pumps are very complex indeed (Welch, 1991). Crossed electrostatic and magnetic fields produce a confined discharge using a geometry originally devised by Penning (1937) to measure pressure in a vacuum system.A trapped cloud of electrons is produced, the density of which is highest in the 10¯4 torr region, and falls off as the pressure decreases. High-energy ions, produced by electron collision, impact on the pump cathodes, sputtering
reactive cathode material (titanium, and to a lesser extent, tantalum), which is deposited on all surfaces within line-of sight of the impact area. The pumping mechanisms include the following.
1. Chemisorption on the sputtered cathode material, which is the predominant pumping mechanism for reactive gases.
2. Burial in the cathodes, which is mainly a transient contributor to pumping. With the exception of hydrogen, the atoms remain close to the surface and are released as pumping/sputtering continues. This is the source of the ‘‘memory’’ effect in diode ion pumps; previously pumped species show up as minor impurities when a different gas is pumped.
3. Burial of ions back-scattered as neutrals, in all surfaces within line-of sight of the impact area. This is a crucial mechanism in the pumping of argon and other noble gases (Jepsen, 1968).
4. Dissociation of molecules by electron impact. This is the mechanism for pumping methane and other organic molecules.
The pumping speed of these pumps is variable. Typical performance curves show the pumping of a single gas under steady-state conditions.
The pumping speed of hydrogen can change very significantly with conditions, falling off drastically at low pressures and increasing significantly at high pressures (Singleton, 1969, 1971; Welch, 1994). The pumped hydrogen can be released under some conditions, primarily during the startup phase of a pump. When the pressure is 10¯3 torr or higher, the internal temperatures can readily reach 5008C (Snouse, 1971). Hydrogen is released, increasing the pressure and frequently stalling the pumpdown.
Rare gases are not chemisorbed, but are pumped by burial (Jepsen, 1968). Argon is of special importance, because it can cause problems even when pumping air.The release of argon, buried as atoms in the cathodes, sometimes causes a sudden increase in pressure of as much as three decades, followed by renewed pumping, and a concomitant drop in pressure. The unstable behavior
is repeated at regular intervals, once initiated (Brubaker, 1959). This problem can be avoided in two ways.
1. By use of the ‘‘differential ion’’ or DI pump (Tom andJames, 1969), which is a standard diode pump in which a tantalum cathode replaces one titanium cathode.
2. By use of the triode sputter-ion pump, in which a third electrode is interposed between the ends of the cylindrical anode and the pump walls. The additional electrode is maintained at a high negative potential, serving as a sputter cathode, while the
anode and walls are maintained at ground potential. This pump has the additional advantage that the ‘‘memory’’ effect of the diode pump is almost completely suppressed.
The operating life of a sputter-ion pump is inversely proportional to the operating pressure. It terminates when the cathodes are completely sputtered through at a small area on the axis of each anode cell where the ions impact. The life therefore depends upon the thickness of the cathodes at the point of ion impact. For example, a conventional triode pump has relatively thin cathodes as compared to a diode pump, and this is reflected in the expected life at an operating pressure of 1x10¯6 torr, i.e., 35,000 as
compared to 50,000 hr. The fringing magnetic field in older pumps can be very significant. Some newer pumps greatly reduce this problem.
A vacuum chamber can be exposed to ultraviolet and x radiation, as well as ions and electrons produced by an ion pump, so appropriate electrical and optical shielding may be required.
Operating Procedures: A sputter-ion pump must be roughed down before it can be started. Sorption pumps or any other clean technique can be used. For a diode pump, a pressure in the 10¯4 torr range is recommended, so that the Penning discharge (and associated pumping mechanisms) will be immediately established. A triode pump can safely be started at pressures about a decade
higher than the diode, because the electrostatic fields are such that the walls are not subjected to ion bombardment (Snouse, 1971). An additional problem develops in pumps that have operated in hydrogen or water vapor. Hydrogen accumulates in the cathodes and this gas is released when the cathode temperatures increase during startup. The higher the pressure, the greater the temperature; temperatures as high as 900ºC have been measured at the center of cathodes under high gas loads (Jepsen, 1967).An isolation valve should be used to avoid venting the pump to atmospheric pressure. The sputtered deposits on
the walls of a pump adsorb gas with each venting, and the bonding of subsequently sputtered material will be reduced, eventually causing flaking of the deposits. The flakes can serve as electron emitters, sustaining localized (non-pumping) discharges and can also short out the electrodes.
Posts
No comments:
Post a Comment