JPS609101B2 - How to generate glow discharge - Google Patents

Description

[Detailed Description of the Invention] The present invention relates to a method of generating an electrode type glow discharge,
More specifically, the present invention relates to a method of generating glow discharge for use in sputtering materials.

Various electrode type glow discharge methods have been devised and used. Although methods in this field are constantly being improved, modified, and developed, the ideas of this invention, including their differences, are believed to represent novel improvements, resulting in increased efficiency and
Provided is a glow discharge generation method with greater stability and reliability. In accordance with the ideas of this invention, an improved apparatus can be designed for depositing coatings or coatings, such as thin films of materials, by cathodic sputtering techniques. Further, in accordance with the ideas of the present invention, the method is efficient in producing radiation having wavelengths associated with electronic transitions in ionized gas species.
A low pressure gas discharge method can be designed that includes a light source as well as a radiation source. (preferably less than about 10-3)
An inventive method that can maintain a strong glow discharge on the target surface at low pressure is desired. Conventionally, for example, a suppression discharge has been proposed that eliminates the coupling between a magnetic field that restricts the movement of ionizing electrons and the generation of ionizing electrons and the sputtering process. However, this type of device has the disadvantage of being flat in shape, and it is known that in the case of a flat shape, it is difficult to maintain a uniform ion current distribution, that is, the ratio of sputtering to the cathode (target) surface. It is being If a magnetic field is used in a device with a cylindrical cathode so that the cathode's cathode line is parallel to the magnetic field lines, further performance improvement can be expected. Their motion is restricted, allowing them to remain close to the cathode surface until most of their energy is dissipated in ionizing collisions with the surrounding gas. At the same time, the geometrical symmetry Therefore, the cathode current can be made uniform in all directions.The outer surface of the cylinder may be used as the cathode surface, or the inner surface may be used.In practice, in either geometric form, It was found that the primary electrons, still retaining a significant portion of their energy, migrate axially and escape from the vicinity of the cathode.By progressively lowering the operating pressure and progressively increasing the diameter-to-length ratio of the cathode, As the size increases, these terminal losses become increasingly important.In addition to the problems mentioned above, in plasma physics, ``the ordinary DC sputtering method for electrically conductive target materials can be used for sputtering electrically insulating target materials. It is well known that this is not possible.

This is because charge build-up in the insulating target material limits the flow of bombarded ion current to values that are too small for practical use. The difficulties encountered in ejecting material from electrically insulating targets by direct current sputtering have generally been overcome by using radio frequency sputtering techniques. In this method, a conductive plate is placed in close proximity behind the dielectric target to be sputtered, and the plate is biased with a high frequency voltage, typically in the megacycle range. For this reason, the radio frequency dielectric current passes through the insulating target, and if any charge is accumulated on the target surface, it is effectively removed, allowing continuous sputtering by ion bombardment. Most conventional radio frequency sputtering devices use a planar target shape, often using a combination of radio frequency and direct current auxiliary discharges, and some sputtering devices use externally applied magnetic fields. Both conventional radio frequency devices have a number of drawbacks that reduce their suitability for various glow discharge and sputtering applications.

These problems include high cost, complexity, inadequacies, low sputtering yields, sputtering from undesirable surfaces in the equipment, small available volume area, high voltage requirements, substantial coupling between maintaining the plasma discharge and controlling the rate of sputtering, resulting in significant limitations on the operating range of the device, and high gas pressures. lack of flexibility, radio frequency leakage, dependence on the distance between target and substrate, contamination of the deposited coating, and plasma field exposed to the target. presence of ion density gradients in Losses can cause sputtering to vary axially, limit performance at low pressures, etc. With this invention, the energy cost per ion produced is minimized. A similar efficient glow discharge is created.

Furthermore, the ions are created in a region that is aligned relative to the target such that a very large portion of the ions are used for sputtering. Furthermore, the plasma discharge is distributed over the target surface in a controlled manner so that the source of sputtered material has the desired geometric distribution. For this reason, sputtering characteristics can be either uniform or intentionally non-uniform. This invention creates a plasma trap that restricts the movement of primary electrons emitted from a cathode target in both the radial and radial directions, and ``these primary electrons transfer most of their energy to the additional plasma.'' It stays close to the target surface until it is consumed as effective work in the ionizing collisions that generate it.

In this specification, complete containment by a trap means that substantially all, or the majority, of the primary electrons that still have sufficient energy to generate additional ions are contained. . That is, electrons that are not accommodated in the traps do not perform substantial ionization work. Plasma trapping is achieved in an environment of low gas pressure by a combination of reflection of electrons from a surface at cathode potential and deflection of electrons by a magnetic field.

The plasma is completely contained within the discharge device of the present invention by means of a trap that blocks all six sides of a volume surrounded by a straight line, and when viewed in cross section of the plasma trap, four of the six sides are closed. It is limited by the magnetic field lines and the electrode plane, and the remaining two sides are closed by the square symmetry of the device. This axial symmetry is obtained through the magnetic field and/or electrode structure. When the primary electrons are thrown into the trap, their energy is used to create additional ions within the plasma.

Once the primary electrons have used nearly all of their useful energy to create ions, they are no longer of value for ionization purposes, so they maintain the required current flow in the overall electrical system. used to do. This current flow is created by electrons, ions, and a combination of both. Therefore, the present invention also provides an efficient anode contact at the trap so that the electrons can easily complete the current path once their energy has been extracted. The anode is preferably placed in a position that blocks the trap so that it efficiently collects low energy electrons without collecting high energy electrons. Briefly, embodiments embodying the ideas of this invention may utilize a flanged cathode format that produces high cathode current densities at moderate to low voltages as well as low ambient gas pressures. Two embodiments using cylindrical cathodes are described. A plasma is generated on the outside of an outwardly flanged cylinder. '. }Plasma is generated in a hollow cylinder flanged inward. Both configurations use a magnetic field, the lines of which are generally parallel to the axis of the cylinder or cathode body. By adjusting the magnetic field strength, the plasma can be localized close to the wall of the fuselage, essentially within a volume confined by the cylindrical surface of the fuselage on the one hand and by the circumference of the flange on the other hand. Therefore, a large number of plasma ions are formed near the cathode layer along the cathode surface. Thus, such cathodes essentially operate by simultaneously using magnetic and electrical trapping effects of fast electrons, the traps being caused by the fact that individual magnetic field lines are separated from the cathode sheath.
formed as a result of intersections at points. This intersection occurs when the plane of the sheath and the magnetic field lines are approximately perpendicular. Roughly three sides of the electron trap are formed by the cathode sheath, and the fourth side is formed by the cathode sheath.
The edges are formed by the magnetic field. Due to the symmetry of this structure,
The plasma uniformity is good. For radio frequencies, a dual electrode format is used.

This also produces high cathodic current densities at moderate to low voltages and relatively low ambient gas pressures. In normal radio frequency operation in a glow discharge, layer, each electrode in a pair of electrodes is anode and cathode to each other, but for most of the radio frequency cycle, both electrodes behave as cathodes to the plasma. It can be shown that In this respect, the electrode becomes anode to the plasma only for very short periods of alternating half-cycles. However, the plasma and ions within the cathode sheath behave essentially at all times as if they were in a direct current glow discharge environment, with both electrodes acting as cathodes. [a] plasma is generated outside a pair of cylindrical electrodes, a plasma is generated inside a pair of hollow electrodes, and a plasma is generated outside a pair of cylindrical electrodes. Three main types of electrodes generated adjacent to a surface will be described.

Both types use cooperating magnetic fields, using field coils outside or inside a particular electrode structure. The magnetic field is shaped in conjunction with a particular electrode structure to define at least one plasma containment trap with the aforementioned axial symmetry imposed by the magnetic field and/or electrode structure. The shape and strength of the magnetic field can be adjusted to confine the plasma within the trap and eliminate end losses. A single plasma trap can be defined by a pair of electrodes and a magnetic field working together. Alternatively, each electrode can cooperate with the magnetic field to define a trap, thus creating a separate trap from that defined by another electrode together with the magnetic field. Furthermore, both types of traps can be used simultaneously in a hybrid format where a separate trap is defined by a magnetic field and each electrode, and both electrodes cooperate to define the trap along with the magnetic field. Furthermore, the present invention contemplates defining multiple plasma traps, each with its own individual electrode and magnetic field. For prismatic and hollow cathode configurations, end wings or flanges may be utilized in some configurations to help confine the plasma trap;
Alternatively, if the trap boundaries can be completely defined by the remaining electrode structure and magnetic field, the blades can be completely occluded.

When using wings, make them large enough to effectively confine the trap. Otherwise, the magnetic field strength would have to be increased so much that it would reduce the economic benefits of the device. In this point,
The wing dimensions are directly related to the space required for the primary electrons to lose their useful energy to create additional ions. The idea of the invention is to generate a plasma with such efficiency that the replenishment requirements for maintaining the plasma in the trap are minimized to the extent that the loads normally imposed on the target as a cathode do not interfere with the sputtering process in any way. You can think of it as a way to provide a trap.

Therefore, the process of maintaining a strong plasma discharge across the target surface is only weakly coupled to the sputtering process itself. In this regard, all that is required is a slight adjustment of the magnetic field strength to easily obtain a wide range of sputtering voltages for a given working gas pressure and sputtering current. It is therefore an object of this invention to provide a new and improved glow discharge method. The above and other objects and advantages of the present invention will become apparent from the following detailed description of the drawings. FIG. 1 shows a conventional vacuum chamber 10 with an electrode discharge device 12 according to the invention installed therein.

Discharge device 12 includes a spool-shaped winged or flanged cathode 14 mounted within chamber 10 and held by support means 15 .

The cathode 14 has an upper wing or flange 4a and a lower wing or flange 4b. An annular anode 16 is disposed around the upper flange 4a of the cathode 14 and is attached to the cover of the vacuum chamber 10. By using two field coil windings, a magnetic field is created in the area on the surface of the cathode.

One coil winding is placed inside the cathode 14.
The other coil winding 21' is arranged around the outer wall 22 of the vacuum chamber 10. The magnetic field due to the outer magnetic field coil 21 is
Typically, it extends from slightly above to slightly below the cathode flanges 4a, 14b. By means of these coils a suitable composite magnetic field is created, as shown by magnetic field lines 23, with magnetic field lines close to and parallel to the body of the cathode. The cathode 14 and the anode 16 are evacuated to create a low-pressure environment.
are both hollow, and a source of cooling water supplied from appropriate piping 25a to 25c is coupled to the apparatus to supply cooling water to both the cathode 14 and the anode 16.A power supply 27 for both high voltage and magnetic fields is shown. It goes without saying that typically two separate power supplies are used. This voltage source 27 is connected to the cathode 14 to provide a negative voltage to the anode 16 when grounded, and and 21. At the lower end of the vacuum chamber 10 there is a line 29 which can be connected via a valve to a vacuum pump (not shown) in the conventional manner. A vacuum chamber 1 is installed so that a coating or plating can be sputtered on
It is shown that it is located within 0.

In the embodiment of FIG. 1, optional upper and lower collars 31 and 32 of magnetic metal may be provided within the vacuum chamber 10.

The collars 31 and 32 are used to straighten the lines of magnetic force, but when an inner magnetic field winding is provided inside the cathode 14,
These may not be necessary. The discharge device 12 shown in FIG. 2 is shown attached to and depending from the upper cover 18 of the vacuum chamber 10.

Vacuum chambers are typically made of non-magnetic metal such as aluminum. The cathode 14 is attached to support means 15 fixed to the cover 18. Considering first the structure of the cathode 14, it includes a cylindrical tube or body 40 made of or plated with the material to be sputtered. One example of a suitable material is a non-magnetic material. Stainless steel, the outside of the cathode body is chrome plated for sputtering the chrome onto the substrate. The cathode 14 also includes upper and lower wings or flanges 4a, 14b, each of which typically has a two-part structure for ease of disassembly.The upper flange 14a includes a lower part 42 and an upper part 43.

These parts are connected by a number of fixing screws 44,
It is sealed by an O-ring 45 and becomes the upper cylindrical wing or flange 14a. The flange 4a has an internal cavity 47 for receiving a cooling fluid such as water. The lower flange 14b has a similar structure, with upper and lower parts 52
, 53 included. These two parts have a large number of retaining screws 5
4 and sealed together by an O-ring 55. A similar cavity 57 is formed between sections 52, 53 to receive cooling water. Preferably the upper portions 42, 52 are made of a non-magnetic metal such as stainless steel plated with chrome (if chrome is to be sputtered) and the lower portions 43, 53 are made of a magnetic metal such as stainless steel. The wings 14a and 14b are connected to a large number of connecting rods 59
The four parts of the fuselage are fixed by the screws, and the O-rings 60 and 61 are sealed.

Inside the fuselage 40 is an annular chamber 62 that receives cooling water through an annular hole 63 in the lower part 42 of the upper flange 14a. A similar annular hole 65 is provided in the upper part 5 of the lower flange 14b.
It is set in 2. The upper part 43 of the upper flange 4a has a pair of annular smoother rings 67a, 67b,
○ ring 67c placed in the groove formed between them
Although the reason will be explained in detail later, glass,
An upper insulator 70 made of Pyrex, ceramic, quartz or other suitable material is connected to an annular gap or groove 71.
It is appropriately spaced from the upper surface of the upper part 43 of the upper wing 14a of the cathode, as shown in FIG. Turning now to internal field assembly 20, it has a spool 76.

This spool can be made from magnetic stainless steel. The spool has a cylindrical body 77 and upper and lower wings 77a, 77b. Wings 77-a and 77
b is fixed to the body 77 by a large number of fixing screws 79 and 801. Plastic rings 82 and 83 are placed around the fuselage 77 near the ends thereof, and a coil 85 having a number of turns of wire is placed around the fuselage 77 between the rings 82 and 83. This wire may be an insulated copper wire, and one end (not shown) is electrically connected to the fuselage 77. The other end 86 passes through a hole in the lower end of the body 77. Preferably, the tube 86 is movable and passes upwardly through a central interior chamber 87 in the body 77 and through an aperture 88 in a support tube 89 of the support means 15. Internal magnetic field winding 8
Preferably, 5 has a greater number of turns near its ends, ie near rings 82 and 83, creating a more uniform magnetic field. The internal spool 76 is connected to the upper wing 7 of the spool 76.
7a at several locations 92 (only one location shown) on the inner surface of the upper portion 43 of the upper flange 14a of the cathode.
It is supported within the cathode 14.

The spool 76 is attached to the lower part 5 of the flange 4b on the lower side of the cathode.
It is also supported and attached by a set screw 94 centered on the inner surface of 3. This structure (particularly that the retaining screw 94 connects the lower part 53 to the wing 77b) ensures that the internal pressure within the cathode is high (the cooling fluid is above atmospheric pressure) and that the vacuum chamber 1
The low external pressure within zero prevents the cathode wings from bending outward. Lower wing 7 of spool 76
7b has a number of holes 95 through which cooling water can return from the central chamber 87 to the outside of the entire assembly, as will be explained later. Next, the structure of the anode 16 will be described. This anode includes upper and lower annular portions 100, 1 which can be made of non-magnetic stainless steel.
01, which are joined together by a number of set screws 102 and 103 located along the section.

These parts are attached to the inner and outer ○ rings 105, 106.
an internal annular cavity 10 which is hermetically sealed and receives cooling water;
form 8. The lower surface of the lower part 100 has a number of annular grooves 1
have 10. The anode 16 is connected to the cathode 14 as shown at 112.
is close to, but spaced from, the upper flange 14a. The anode 16 is held in place by two supports 17 connected to the upper cover 18 of the vacuum chamber 10.

Only one of these supports is shown. Both supports are similar and provide a ground connection of the anode to the cover 18;
One side will be the inlet of the cooling water, and the other will be the outlet of the cooling water. The support 14 includes a joint 114 fixed to the upper part 101 of the anode 16 by a set screw and sealed by an O-ring. Furthermore, the support 17 includes a tube 115. This tube has a flange 1 connected to the baser 114 by means of a number of retaining screws and sealed thereto by means of an O-ring.
16 can be attached with silver solder. The upper end of the tube 115 passes through a hole in the cover 28 of the vacuum chamber 10 and includes a threaded compression fit 119 (which may include an inner circle, not shown), an O-ring 120, and a cover plate 121. and one or more set screws 122 to secure and seal the cover 28'. Although the support 17 can be insulated from the vacuum chamber 10 so that the anode 16 can be operated at voltages other than ground potential, as shown and described in this case, it is assumed that the anode 16 is at ground potential. Operate. The support group 7 has an internal passageway base 24 that makes it possible to supply cooling water to the cavity 108 in the anode 16, and the return flow of water is dependent on a second support (not shown). Additional support for the anode turtle 6 can be provided by a plurality of metallic arms 126 (eg four, only one shown).

One end of this arm is connected to the upper cap arm 27 of the support means 15.
The other end of the anode 16 is connected to the lower flange 27b and the other end thereof is attached to the upper part of the anode 16 via a spacer 128. These arms also provide a ground connection to the anode. Turning now to the support means 15, it includes the upper cap 127 mentioned above.

The cap includes an upper flange 27a secured to the underside of the upper cover 28 of the vacuum chamber 10 by a number of set screws 131. The cap 127 is connected to the lower flange 27
b, and the upper insulator 70 is connected to this lower flange 27b.
and the ring 67a on the upper surface of the upper part 43 of the upper flange 4a of the cathode? 67b and the ○ ring 67c, and the flange 127b and the ○ ring 132.
sealed by. Cap 127 and support tube 89 can be made of non-magnetic stainless steel. An inner stem or lead wire insulator 134, which may be made of a plastic such as that sold under the trade name Lexan, is mounted within the cap 127, with its cylindrical lower end extending into the hole 135 in the insulator 70. engaged and jammed. The support tube 89 has a lower end threaded onto the upper wing 77a of the internal spool 76 and has a flange 36 that engages and is sealed with the upper surface of the upper portion 43 of the upper flange 14a of the cathode.

Support tube 89 extends upwardly through insulator 134 and has a threaded upper end secured to main nut 140 and nut 141. The thrust bearing 144 and the arcuate spring washer 145 are
It is attached around the upper end of the support tube 89 between the top surface 146 of the insulator 134 and the bottom surface of the main nut 140 .
The spring washer 145 is used to absorb expansion and contraction due to heat. As is clear from Fig. 2, main nut 1
40 fixes the cathode 14 with the intermediate upper insulator 70 to the support means 15.

In this way, the threaded lower end of the support tube 89 is connected to the spool T.
The upper wing 14a of the cathode is combined with the upper wing 77a of the cathode.
, upper insulator 70, upper cap 70 and insulator 134
compress. In this way, the lower cylindrical shoulder 149 of the insulator 134 abuts the upper inner edge of the lower flange 27b of the cap 127 and the upper ring 67a, 67b of the upper wing 14a of the cathode and the lower part of the upper cap 127. Provides a primary retaining stress that causes the lower surface of side flange 27b to abut snugly and sealingly with each surface of upper insulator 70. The ○ ring 67c and the tatami 3 on the upper wing 14a of the cathode and the lower flange 27b of the upper cap 127
2 seals it. Insulator 134 is not exposed to the vacuum within chamber 10 . The upper mounting part 152 is attached to the rising part of the support tube 89 by means of a large number of retaining screws 53 attached to the upper nut 141, and is sealed by an O-ring 154.

The upper end of the inner magnetic field coil conductor 86 passes upwardly through the fitting 152 and is sealed by the compression fitting 155. Mounting portion 152 has an inlet chamber 156 for receiving cooling water. This chamber 156 is an inlet chamber 157 in the support tube 89.
And Ren suits. The inlet chamber 157 has at its lower end a number of holes 15 which fit into the cavity 47 in the upper flange 4a of the cathode.
I have 8. The pipe 89 is connected to the outlet passage 16 of the mounting portion 152.
0 and has an exit passageway 88 suitable for the lotus. Appropriate sealing of the inlet and outlet passages 156 and 16 to passages 157 and 88 in tube 89 may be provided by respective seals at the junctions of the respective passages. Cooling water flows into the inlet pipe 25a, from the inlet passage 156 of the fitting 152 through the passage 157 in the tube 89 to the cavity 47 in the cathode upper flange 4a, and around the upper wing 77a of the spool 76. can be changed.

The cooling fluid passes through the annular hole 63 to the upper flange 1 of the cathode.
4a, enters the chamber 62 in the cathode body 40 and enters the cavity 57 in the flange 4b on the underside of the cathode. Fluid flows around the lower wings 77b of the spool 76 and then through the holes 95 and into the central chamber 87 in the body 77 of the spool. Wings 77a and 77b provide baffles that direct the water to properly cool cathode flanges 4a and 14b. The fluid then flows through outlet passage 88 of tube 89 to outlet passage 160 of fitting 152 . Finally, the second
The illustrated assembly includes a bottom with a lower insulator 164 made of the same material as the upper insulator 70, for example.

The insulator 164 is sandwiched between the non-magnetic lower cap 165 and the bottom 53 of the lower flange 14b of the cathode. A plastic insulator 166 made of the same material as insulator 134 is screwed into lower part 53 and cap 1
65 is screwed into this insulator 166. o Rings 167 and 168 form a seal with the insulator 164, and the illustrated device has an upper insulator 70 and an upper flange 14 of the cathode.
A long and narrow annular gap 171 is formed, similar to the gap 71 between the top part and the upper part of a. Next, the insulator 70 and the insulator 164
To explain, where there is contact between the metal and the insulator when the metal is at or near the potential of the cathode (or at least at a considerable distance from the potential of the anode),
A long narrow slot or gap is provided, such as gap 71 in insulator 70 and gap 171 in insulator 164, which gap is perpendicular to the magnetic field lines. This configuration prevents the sputtered material from taking its path along the gap and coating the side of the insulator facing the gap, ultimately creating a conductive path and causing a short circuit. It will be appreciated that the magnetic field lines are shown at 172 and that the annular gaps 71 and 171 are perpendicular thereto. The magnetic field lines in gaps 71 and 171 are parallel to magnetic field lines 172. The perpendicularity of the magnetic field lines prevents electrons from entering the gap and causing ionization therein. During operation, there is a shed 173 near the cathode body 40.

This sheath extends virtually to the annular slot 112 between the flange 14a and the anode 16. Using relatively low pressure,
A large cross-sectional area can be obtained.

That is, since the mean free path, or average length of collision-free line-of-sight motion, of the sputtered material is inversely proportional to pressure, it is possible to operate at relatively low pressures. In the device, the base body faces the line of sight of the above-mentioned coating material even at a large radius. According to this invention, magnetic field lines 17 in FIG.
2, the cathode body 40 and the flange 14
A good trap is formed between a and 14b.

This is accomplished, for example, by adjusting the currents in the inner and outer field coils and by selecting magnetic and non-magnetic metals.
Another important advantage is that the dimensions of the device can be varied.

That is, the diameter and especially the length of the cathode can be easily varied without essentially altering satisfactory operation. Cathode body 40 is a quasi-ray source and therefore can be easily extended in length with minimal other modifications. Another advantage is that
It is possible to operate with relatively low voltages between the cathode and the anode, resulting in two advantages.

First, the increase in sputtering yield (the number of sputtered atoms per incident ion)
When the energy is large, it becomes smaller than linear. That is, for a given energy input, more sputtering occurs when operating at lower voltages than when operating at higher voltages. Second, low voltage, high current operation allows for high levels of energy to be delivered to the sputtering equipment without costly insulation problems in the power supply as well as in the design of the sputtering equipment itself. means. Another advantage is that the operating current is proportional to some power of the operating voltage. That is, current 1 is proportional to Vn. In this), n is about 5 to 7. Therefore, small changes in voltage result in large changes in current. - Another main advantage of this invention is that a good anodic connection can be provided.

This is due to several aspects of the invention, including the flanged cathode structure and the ability to operate at relatively low pressures. A good anode connection means that when the electrons hit the magnetic field lines that do not intersect the flange, the electrons move radially outward to a position such that they do not intersect the cathode flange as quickly as possible (or after they have moved radially outward to a position such that they do not intersect the cathode flange). refers to positioning the anode 16 relative to the flange 4a so that it is screened by the anode (nearly). As a result, there is no voltage drop from the plasma to the anode, so that almost all of the applied voltage is used advantageously. Another advantage is that heating of the substrate is minimized since the bombardment of the substrate by the radiation as well as by the plasma is negligible.

This is also because the plasma is contained around the cathode. Therefore, ``this device can also be used for substrate materials that are easily melted or easily gasified, and with this device there is no noticeable change in the structure of the substrate material due to electron bombardment.On the other hand,
If desired, it is also possible to perform electron bombardment by interrupting the magnetic field or by intentionally bombarding the substrate with plasma. Another advantage is that excellent water cooling of both the cathode as well as the anode can be easily achieved.

This allows the device to have very high current densities. Another advantage is that the insulator 70 and tortoise 64 are expected to last for an indeterminate period of time, primarily due to the provision of problems 7i and 71 and their arrangement.

For example, in the apparatus shown in FIG. 2, the only materials exposed to the chamber vacuum are metal and insulators 70 and 164, typically made of Pyrex, ceramic, quartz, or the like. Therefore, no crooked elastomer or plastic is exposed to the vacuum leading to gasification problems. Another advantage is the ability to provide a stable current density over a large area.

For this reason, the sputtering material has good uniformity over a large range of current density.The operating voltage V in the sputtering device can be considered to be an index of discharge efficiency, and can be expressed as follows: .

Ei V=7 Terako) where Ei is the average energy consumed by electrons to create one ion, y is the ion-electron secondary emission coefficient due to ion bombardment on the cathode material, and r is the electron energy. Of the ions lost in the trap, 2 is the ions that reach the cathode.

Ei is a function of the characteristics of the gas used. For example, for argon this is approximately 30 electron volts (eV)/ion. y for an argon environment is approximately 0.1. That is, one out of ten ions hitting the cathode produces one electron that is emitted into the plasma, and the rest are lost within the metal of the cathode. These emitted electrons are accelerated in the cathode sheath and acquire an energy approximately equal to the applied voltage V. When these electrons become trapped and lose energy, they create a plasma, which creates more ions, which in turn sputters to create more electrons. If the trap obtained by the magnetic field is very efficient, then D is 1 and the electrons will be trapped until they lose all their energy. Some ions will drift out of the plasma and some will reach the cathode. It is pulled.For this reason, Togo2 is usually not 1.Do2 is an ion,
This is the amount that is pulled to the cathode, and varies from a minimum value of 0 to a maximum value of 1. In a perfect trap, the product of 2 and 2 is 1, in which case the operating voltage would be 300 volts for the complete device.In the example device, the operating voltage would be approximately 800 volts, so the The product of 2 and 2 is approximately 0.3, which means that the efficiency is 30%.In the hollow cathode device, which will be explained later, the ``wrapping'' shape of L is advantageous in increasing the height of the lattice. be. Therefore, the typical operating voltage is about 400 volts, and the product of d, x d is 0.
．． It is 75. Explaining the structure of the type shown in FIG. 2 again, it is of course desirable that 0 and 02 be large and close to 1.

This is accomplished, as previously discussed, by providing an efficient trap and locating the trap as close to the body of the cathode as possible. That is, if an efficient discharge is created, , becomes close to 1, and if ions are created as close to the cathode as possible so that a large number of them come near the cathode, then , becomes close to 1. As mentioned earlier, H is sensitive to the gas environment used, and for argon it is approximately 0.1. Ei varies depending on the gas used, but typically there are approximately three
It is Noion. In conventional conventional sputtering equipment, the applied voltage was approximately 3000 volts, requiring great care to provide adequate insulation. In contrast, lower voltages (
(e.g., 800-850 volts), insulation problems for both the cathode and its power supply are substantially minimized. The device shown in Figure 2 has both inner and outer magnetic fields, but one or both can be used, especially when appropriate materials are selected for the parts, with or without a magnetic plate that acts on the magnetic field lines. Needless to say, it is possible.

As previously mentioned, it is particularly desirable for the magnetic field lines to be close to and essentially parallel to the cathode body.
This is best achieved by using both an inner field and an outer field, in which case both can be adjusted to obtain the desired composite field. When it is necessary to adjust the magnetic field strength, it is possible to adjust both fields together while maintaining the proper position of the magnetic field lines relative to the cathode. Turning now to Figures 3a-3d, another embodiment and variation thereof is shown which utilizes the ideas of the present invention, but which utilizes a flanged hollow cathode assembly.

Although this is in some respects the opposite of the cathode 14 of FIG. 2, the operation is exactly the same. Figure 3a shows the assembly in its entirety, Figure 3b is an end view thereof, and Figures 3c and 3d detail preferred deformations for the ends of the assembly. Figure 3c is a partial view showing a portion of the left end of the assembly, which can be used for both ends.
Figure 3d is a partial view of the right end of the assembly, but this can also be used for both ends of the assembly. In FIG. 3a, the hollow cathode is connected to cathode body 214 (and end flanges 214a and 2).
14b.

The body is cylindrical and the flanges 214a, 214b are in the same shape of annular rings. These flanges may have cooling water passages 214c as shown in the variant of Figure 3d. The cathode body and flange may be made of the material to be sputtered, or may be made of a non-magnetic material such as aluminum or stainless steel, with the material to be sputtered plated or plasma sprayed on the inner surface.
Typical thicknesses for such licked or plasma sprayed coatings range from 10 to 100 mils. A replaceable cathode body 214 can be welded to an outer cylindrical housing 215 as shown in FIG. 3a, forming an annular cavity 215a for receiving cooling water. As shown in the variations of FIGS. 3c and 3d, the body 214 is closed to the sealing ring 2.
16b and 215c are preferably captured inside the housing 215 to form an annular cavity 215d for receiving cooling water. Annular anodes 216a and 216b are provided and insulated from the cathode by ring-shaped insulators 218 and 219 made of the same material as insulator 70 of FIG.

Grooves 216c are shown on the anode surface. As discussed with respect to FIG. 2, grooves and/or beat blasts are provided to improve adhesion of sputtered material incident on the anode surface. The anodes 216a and 216b have respective cooling water cavities 223 and 224 and are attached to respective anode end plates 221 and 222 to form respective cooling water cavities 223a and 224b, as shown in FIGS. 3c and 3d. is preferred. By removing the end plates 221, 222, the cooling water cavity can be cleaned periodically. An elongated and narrow annular gap 2 between the cathode, anode and insulators 218, 219
28 and 229 (Figures 3a, 3c and 3d) are provided.

These gaps are the gaps 71 and 1 of the assembly shown in FIG.
71 and is perpendicular to the magnetic field lines generated by the magnetic field coil 234. As shown in Figures 3c and 3d, a series of holes 228a and 228
By adding b to help eliminate gasification from these gaps, the gasification problem described with respect to FIG. 2 can be suppressed. each cathode flange 214a,
214b and the annular skirt 231 of the anodes 216a and 216b.
and 232, the second set of gaps 231, 232 (Fig. 3a) and 231a, 232a (Fig. 3c and 3d).
) is formed. These gaps 231, 232 and 2
31a and 232a are parallel to the lines of magnetic force. In practice, the field coil 234 typically consists of a number of windings on a suitable former of non-magnetic metal (e.g. aluminum), with insulation separating the windings from the cathode outer housing 215. It is provided on the shield 234a (FIGS. 3c and 3d).

These windings are arranged along the assembly shown in FIG. That is, each winding 234 in FIG. 3 represents a plurality of coils in the preferred construction. These coils can be stratified and the coil current can be adjusted to create the desired magnetic field within the cathode. Magnetic end plates 237 and 238 can be attached to the outer ends of anodes 216a and 216b, respectively. The magnetic field lines are passed through the magnetic end plates 237, 238 to allow the magnetic field lines to pass through the anode, thereby improving the anode connection. Preferably, the magnetic end plate is made of iron. A series of bolts 23 through which the anode magnetic end plate assembly is threaded onto sealing rings 215a and 215b. a and 238a (FIGS. 3c and 3d) are attached to the cathode so as to capture the insulators 218 and 219. These sealing rings are polycarbonate. Made of an insulating material such as plastic. A pair of cylindrical insulators 241 and 242 form the magnetic end plate 2
At the outer ends of 37 and 238, respective grounding shields 245 and 246 are also positioned adjacently.

These insulators may be the same material as insulators 218 and 219. The grounding shield allows the plasma discharge to be localized within the hollow cathode when the bias sputtering process is performed on a grounded substrate, i.e. using one power source to connect the anode to the grounded substrate. The second power source drives the cathode to a negative potential of several hundred volts relative to the substrate.
The plasma generated by the resulting discharge resides at or near the anode potential. Therefore, the grounded substrate is negative to the plasma and sputtering occurs from the surface of the substrate. The relative potentials are adjusted such that the sputtering rate from the substrate is less than the deposition rate so that net deposition occurs.This sputtering of the substrate continuously cleans the substrate during the subsequent deposition process. Oribe shields 245, 246 ensure that the positively driven anode does not draw current from the vacuum pump and other grounded elements in the apparatus.Appropriate end fittings 249 and 250 at the ends of the apparatus The end fittings 249, 250, grounding shields 245, 246 and outer insulators 241, 242 are secured to the plates 237, 238 by suitable bolts or tie rods. The connecting rod is connected to the insulator 2 as shown in Figures 3c and 3d.
46a and 246b to prevent the anode from being electrically connected to a grounding shield. End fitting 249
, 25 can be connected to a vacuum device, in which a movable or fixed base 252 is mounted, or a third
It is possible to mount the entire assembly of Figure a in a vacuum chamber of the type shown in Figure 1. Grounding shields 245 and 246 with stainless steel screens 247 and 248
A hole is provided through which the base body 252 passes. An electrical lead 256 is connected to the left anode 216a shown in FIG. 3a, and an electrical lead 257 is passed under the magnetic field winding 234 and connected to the right anode 216b. Winding wire 23 with conducting wire 257
4 to prevent significant transients from being induced in the magnetic field coil assembly 234 in the event of a surge in the plasma discharge stream.The conductors 256 and 257 are It may be connected to ground as in the embodiment, but as mentioned previously for bias sputtering,
It may also be connected to a positive potential. The cathode conductor 259 is connected to the shell 21
5 and a negative voltage source. A contacting set screw 215e, shown in FIG. 3c, ensures good electrical contact between the cathode body 214 and the housing 215. Suitable tubular passageways are provided in the anode cavities 223, 224 and the cathode cavities 21.
5a and 214d, and supplies cooling water to them. It goes without saying that these lines can be interconnected to a source of cooling water in any suitable manner similar to that described for the apparatus of FIG. The conductors of magnetic field winding 234 are also preferably hollow and also water-cooled. Cathode body 2
14 and the cathode flanges 214a, 214b, the anode 2
O-ring seals (not shown) are preferably provided between 16a, 216b and the anode end plates 221, 222, and between the anode end plates 237, 238. Surfaces of insulators 218, 219 and flanges 214a, 214b
○ ring seal 27 between the anodes 216a and 216b
0,273 is used, and the sealing rings 215b, 215c
(Figures 3c and 3d) and cathode assemblies 214-215
○ ring seals 274 and 275 are used between the two.
In the plane of the insulators 2417 242, between the plates 245, 246 and the fittings 249, 250, and when this embodiment is used, the water-cooled body 214 and the end flange 214a of FIG. 3c.
, 214b, a similar seal (not shown) is used. In operation of a device of the type shown in FIG. 3, it has been found that the device can be operated at voltages as low as 300 volts.

This is because the product of Go and Go2 is close to 1. The hollow cathode structure of FIG. 3 allows the operating current to be varied over a wide range at a given operating pressure, similar to the cathode structure of FIG. When using a hollow cathode structure, the deposition rate for the substrate 252 at the center of the cathode is primarily independent of pressure and almost independent of voltage. Therefore, pressure is essentially an independent parameter and can be varied for other reasons as desired, for example to influence the nucleation process on the substrate. It will be appreciated that the hollow cathode embodiment shown in FIGS. 3a-d and described above has essentially the same advantages as the spool type cathode format shown in FIG.

That is, in the hollow cathode format, as previously mentioned, the electron trap exists in a thin annular region extending inwardly from the cathode face to a radius limited by the inner diameter of the anode. This embodiment provides a good anodic connection so that the desired electrons can be transferred across the trap from the edge of the cathode sheath (which extends only a few millimeters from the cathode surface) to the edge of the trap (i.e. the inner diameter of the anode). An energy gradient exists. To knead, it leaves the trap by diffusion inward,
The electrons that pass into the vicinity of the substrate rather than the anode have very low energy, 3 to 1 V. That this is indeed the case was verified by electrostatic probe measurements across the radius of the plasma. Therefore, in the hollow cathode format, the heating of the substrate from the surrounding plasma is necessarily greater than in the spooled cathode format, where the plasma does not extend to the substrate, but in the hollow cathode embodiment described above. ) Since the plasma cools down before reaching the vicinity of the substrate, heating is much less than before. Next, referring to FIG. 4, this figure shows a double electrode type discharge device 31
An apparatus for carrying out the radio frequency glow discharge method according to the invention is shown having a conventional vacuum chamber 311 1 fitted with a vacuum chamber 311 .

Discharge device 312 includes a pair of winged or flanged coaxial electrodes 313, 314 mounted within vacuum chamber 311 and supported by a suitable mounting trunk 315.

Electrodes 313 and 314 are physically separated and electrically isolated from each other by a suitable insulator 316. Electrode 313 has upper wings 313a, whereas electrode 314 has lower wings 314a, which are used for attachment or to help confine the plasma trap. However, as will become clear later, electrode 3
13 and 314 cooperate with the magnetic field, the end wings 313a, 3
The trap can be effectively confined without using 14a. In this case, these wings may be removed from the structure.
Electrodes 313, 314 together constitute a cylindrical radio frequency column cathode assembly. A radio frequency power source 3 1 7, typically operating at a frequency of 1.8 Mc, is connected to a shielded transmission line 318 that extends through the mounting trunk 315 to the electrodes 313, 314. Multiple magnetic field coil windings 318
is disposed around and insulated from the outer wall of the conductive housing 311a of the vacuum chamber 311.

These magnetic field coils create a uniform magnetic field within the room. The magnetic field lines approach and extend between the electrodes 313, 314, as shown by magnetic field lines 319, where a uniform magnetic field is desired. Vacuum chamber 311 contains a suitable working gas at a relatively low pressure, typically 0.5 to 1.0 microns.

Attachment trunk 315 and transmission line 318 pass through a suitable vacuum seal 320 at the top of chamber 311.

A suitable insulator 321 is attached to the trunk 315 and the electrode 31
3 to prevent the mounting trunk from having a cathodic potential and therefore becoming a sputtering target. At the lower end of vacuum chamber 311 is conduit 322 .

This can be coupled in the conventional manner to a vacuum pump (not shown) via any suitable valve arrangement. A conductive screen 323 covers the closure in the conductive housing 311a of the chamber 311 through which the conduit 322 intersects with the chamber. In this regard, the screen 323 and the outer conductive housing 311a cooperate to define a radio frequency shield that is suitably connected to the center tap ground (not shown) of the radio frequency power source to carry circulating ground currents. It seems to be avoided. Additionally, screen 323 shields the vacuum pump and prevents electrical discharges that might otherwise travel to the pump. An end cap consisting essentially of an insulator 324 and a shield 325 is provided on the outside of the electrode assembly adjacent to the wings 314a of the electrode 314 to prevent sputtering from occurring on the end surface of the electrode. . A suitable coolant conduit (not shown) is also connected to the discharge device 312 via the mounting trunk 315 for supplying a suitable cooling fluid, such as water. A substrate 326 is disposed within the vacuum chamber 311. and each electrode 313 is the cathode electrode of this device.
, 314 for sputtering a coating or plating. In Figures 5a to 5d,
Various embodiments of prismatic dual electrode structures are shown.

Each of them includes a conical electrode 313 and a coaxial conical electrode 31.
4, the truncated cones of which terminate adjacent to each other, and as shown in connection with FIG.
16. However, Figures 5a to 5d all show different magnetic field formats.
For this reason, the plasma traps limited in each case are different. As previously mentioned, the end wings 313a, 314a shown in conjunction with electrode 3137314 can be provided for attachment or to help define a plasma G-trap.

If you do not need wings for either of these purposes, there is no need to include end wings. However, the provision of these end wings allows the angle of the conical surfaces defining the electrodes 313, 314 to be selectively adjusted so that the material to be sputtered is thrown in a selected direction. - End wings 313a, 31 to help define the boundaries of the trap.
If using 4a, make them large enough to effectively confine the trap.

Otherwise, the magnetic field strength would have to be unduly strong. In this regard, the wing dimensions are directly related to the space required for the primary electrons to lose their useful energy in creating additional ions. In FIG. 5a, the magnetic field lines 3 extend parallel to the symmetry axes of the conical electrodes 313, 314, so that on each side the magnetic field lines 3
13, 314, thus defining a single plasma trap utilizing both electrodes 313, 314 working together.

Whether wings 313a and 314a help confine the trap depends on the magnetic field strength and the distance from the axis of symmetry required to fully accommodate substantially all of the primary electrons available to do additional ionization work. radial distance.
The magnetic field shown in Figure 5a is typically generated by a magnetic field coil, such as coil 21 in Figure 1, wrapped outside the vacuum chamber. Figure 5b shows an electrode format similar to Figure 5a, but with two sets of magnetic field coils 328, 32'9 provided inside electrodes 313, 314, respectively, with bent coils terminating on each electrode. Generates a magnetic field.

For this reason, the embodiment of FIG. 5b also has electrodes 3 acting in concert.
139 defines a single plasma trap with 314 and confines this trap. In FIG. 5c, there is also a pair of magnetic field coils 330, 331, with each electrode 313,
There is one magnetic field coil inside 314. However, section 5c
In the illustrated embodiment, the magnetic fields generated by each field coil begin and terminate on the same electrode, thus defining two separate plasma traps. Each plasma & trap is limited by a magnetic field 319 and only one electrode. However, due to the continuity of the electrical circuit, some small electron current exists between the two traps. Fifth
Figure d shows the same machine as the embodiment shown in Figure 5c, with two magnetic field coils 332 and 333 in electrodes 313 and 314, respectively.
Use it.

However, in the embodiment of FIG. 5d, the jagged magnetic field is limited so that a plurality of separate plasmas are associated with each electrode. You will get a trap. Although not specifically shown in the drawings, by combining various types of magnetic fields, such as the magnetic field shown in Figure 5b and the magnetic field shown in Figures 5c or 5d, a ``hybrid magnetic trap type'' can be created. It can be realized and used.

In such a hybrid format, "multiple traps are defined by individual electrodes and magnetic fields, and additional plasma traps are defined by magnetic fields and both electrodes acting in concert." Although the magnetic field in the embodiments of Figures 5b to 5d has been described as being generated by the inner magnetic field coil, it is also possible to generate such a magnetic field using a permanent magnet. Needless to say.

Figures 6a to 6e show that to limit the plasma trap in the glow discharge method according to the idea of the present invention,
Figures 5a-5d illustrate various dual electrode post formats suitable for use with any of the magnetic field formats shown in Figures 5a-5d.

Figure 6e is preferably used with a straight magnetic field as shown in Figure 5a. This electrode type can use any of the magnetic field types previously mentioned. Axial symmetry is obtained by the magnetic field facing the electrodes. In all the glow discharge electrode and magnetic field types according to the invention so far described, the plasma trap is characterized by a sea bream symmetry about the axis of rotation.

This symmetry can be achieved by the magnetic field, the electrodes, or both the magnetic field and the electrodes. Referring next to FIG. 7, there is shown a schematic diagram similar to FIG. 4, but illustrating the use of a hollow electrode assembly to generate internal plasma. A pair of hollow conical coaxial electrodes 4
13,414 is an outer conductive cylindrical housing 415;
properly supported within.

The ends of the housing 415 terminate in end shields 416, 417 having ports 41 that can be drawn to a vacuum.
There are 6a and 417a. Openings 416a and 417a are covered by conductive screens 418 and 419. The electrodes 413, 414 are separated by a suitable electrical insulator 419 and are connected to an electrical connection point 422 of each electrode 413, 414 via a shielded transmission line 420 from a suitable power source 421.
, 423 are supplied with radio frequency power. Grounded power source 421, shielded transmission line 420, cylindrical housing 41
5. End shields 416, 417 and screens 418,
419 together constitute a Faraday-type radio frequency shield similar to that provided by the vacuum chamber and screen in the embodiment of FIG.

Communication with the vacuum pump is provided by any suitable conduit 424.

Conduit 424 is preferably separately grounded and insulated from end shield 417 by insulating member 425 . A similar insulating member 426 is provided on the opposite side of the device and is provided with an end cap, another stage of the glow discharge device, or a suitable substrate such as a comb substrate 430 and a feed device for feeding along the central axis of the electrodes 413, 414. connected to. The electrical insulator 427 is the electrode 41
3 and end plate 416, and a similar electrical insulator 428 extends between electrode 414 and end plate 417 to isolate the electrode from the conductive radio frequency shield.

A number of field coils 4 similar to field coil 318 of FIG.
29 is wrapped around the outside of cylindrical housing 415 to create a magnetic field for plasma containment that cooperates with electrodes 413 and 414.

In this regard, when using the hollow cathode method, the magnetic field is always supplied by the outer field coil. Figures 8a-8c illustrate various magnetic field formats that can be used with hollow cathode electrode formats.

Although a conical electrode assembly is shown in each of the embodiments of FIGS. 8a-8c, this is only one example, and it is understood that other hollow electrode formats can be used with the magnetic field format shown. Needless to say. Furthermore, although an insulator 419 is shown separating the electrodes 413 and 414, this insulator can be removed if the electrode itself is an electrically insulating material, such as when the target material to be sputtered is a dielectric. I can do it. In FIG. 8a, all the magnetic field lines 431 are connected to the electrode 413.
to electrode 414, the plasma trap is limited by the magnetic field and both electrodes working together. Accordingly, FIG. 8a illustrates the case of the pillar-type hollow electrode shown in FIGS. 5a and 5b. Figure 8b shows
The magnetic field lines are shown entering and exiting at two locations on the same electrode, thus defining separate plasma traps depending on the magnetic field and each individual electrode.

FIG. 8b therefore corresponds to the case of a hollow electrode of the type shown in FIG. 5c for a columnar electrode. Similarly, FIG. 8c shows a jagged magnetic field format, where a plurality of separate traps are defined by the magnetic field and each electrode.

For this reason, Figure 8c is similar to the magnetic field format previously described with respect to Figure 5d. Figure 8c shows the case of hollow electrodes. Figures 9a to 9e are similar to Figures 8a to 9e, including hybrid cases where the magnetic field shown in Figure 8a can be used in combination with the field type shown in Figures 8b or 8c. 8
Can be used with any of the magnetic field formats shown in Figure 2.
Figure 3 shows various embodiments of the heavy electrode hollow cathode format.

FIG. 9a shows a pair of hollow coaxial cylindrical electrodes 413, 414, optionally separated by an electrical insulator 419.

When the magnetic field type of FIG. 8a is used alone or in conjunction with the magnetic field types of FIGS. 8b and 8c, wings 413a, 414a at the outer ends of electrodes 413, 414, respectively, are used to confine the plasma trap. It is desirable to provide
At other times, a magnetic field is combined with each electrode to generate a separate plasma. If the trap is limited, there is no need to provide end wings. FIG. 9b is similar to FIG. 9a, except that inner end wings 413b and 414b are added to the electrodes 413 and 414, respectively, and these inner wings are similar to the outer end wings 413.
Figure 9c shows a pair of coaxial conical electrodes 413, 414 with inner end wings 413b, 414b, respectively. It shows.

This format allows any of the magnetic field formats previously established for hollow cathode electrodes to be used. Figure 9d is the 9th
Although it is similar to Fig. c, the cylindrical inner surface of the electrode 4 tatami mat 3, 414 has a generalized and complicated shape, so that the sputtering pattern shape can be changed to meet arbitrary requirements.

Figure 9e shows a pair of hollow cathode electrodes consisting of a cylinder with a crack along its long axis. If the electrode is a conductive material - an insulator 419 is provided between the two halves of the cylinder. The electrode format of FIG. 9e can also be used with any of the magnetic field formats previously described. The axis of symmetry of the resulting plasma trap is determined by the magnetic field and not by the symmetry of the electrode structure. Having described various configurations of dual electrode configurations, cylindrical and hollow electrode types, in combination with various magnetic field configurations to define the various different configurations of traps containing the plasma of glow discharge devices, we will now discuss The functions of these traps will be explained.

In all dual electrode radio frequency glow discharge devices described in connection with this invention, both electrodes are sputtering targets.

Glow discharge devices are operated at a combination of pressure and frequency such that nearly all of the energy exchange between the radio frequency electric field and the loaded plasma particles occurs in the cathode sheath region. The operation of such a radio frequency discharge device is similar to that of a direct current discharge device, with each electrode acting as a direct current cathode for most of the radio frequency voltage cycle. At the cathode, each target electrode has an average voltage drop approximately equal to the applied radio frequency voltage from zero to the peak value minus the voltage drop across the dielectric sputtering target. exist. Ions of the working gas drift from the plasma volume into the cathode sheath, where they pass through the sheath potential as they are accelerated toward the cathode target surface. When these bombarded ions collide with the cathode surface, atoms of the target material are ejected (by sputtering) and secondary electrons are emitted from the cathode surface. These electrons are accelerated in the cathode sheath to acquire an approximate totality and voltage, and then enter the plasma volume as primary electrons, creating new ions by an electron-one-atom collision process. A portion of the ions created in the plasma volume by these primary electrons and their secondary electrons eventually reach the edge of the cathode sheath, where they enter the sheath and impact the cathode target surface. The process described above is thus repeated. The discharge is thus self-supporting due to the cathodic process. In such DC or radio frequency glow discharge devices, the operating voltage can be considered as an index of discharge efficiency.

When using an electrically insulating target, in the case of radio frequency, the operating voltage is not the total voltage across the discharge device, including the voltage drop across the dielectric sputtering target, but the zero to peak value across the glow discharge. The operating voltage is up to. Using a pair of electrodes and a suitably shaped magnetic field in combination in a device with parasitic symmetry, a trap is created that restricts both the radial and axial movement of the primary electrons, thus displacing most of the electron's energy. electrons (close to the surface of the target) until they are consumed in ionizing collisions.

Therefore, ions are created adjacent to the target surface. That is, the plasma created by the trapped electrons substantially confines an annulus around the axis of rotation surrounding and in intimate contact with the electrodes that make up the target. This is accomplished by a combination of reflection of electrons from a surface and deflection of electrons by a magnetic field.

A plasma trap is thus formed, confined on one side by the magnetic field and on the other side by the cathode surface, so that end losses are completely eliminated. When talking about confining a plasma trap on all sides, perpendicular to the axis of symmetry, since the actual rectangular volume of the plasma has six sides, two of the sides of the trap are closed due to axial symmetry. It will be understood that it is only necessary to consider the cross-sectional shapes of the four side surfaces.

In the radio frequency dual electrode format, a pair of electrodes is driven by two radio frequency feed lines.

The pairs of electrodes are driven by alternating voltages so that the electrodes alternately become anodes and cathodes with respect to each other. Referring to the two electrodes as A and B, respectively, at some times electrode A is a cathode with respect to electrode B, and at other times it is an anode with respect to electrode B.

Generally, both electrodes are negatively biased with respect to the plasma during most of the radio frequency cycle. That is, electrodes A and B behave as cathodes for plasma. Only for very short periods of alternating half-cycles of the radio frequency voltage are electrodes A and B anode to the plasma. With the above in mind, radio frequency plasma traps can be divided into two types. 1st
In the ant range, the two electrodes A and B act together to partially close the trap, and the magnetic field completes the confinement of the trap. In such a case, magnetic field lines exit from electrode A and enter electrode B. In this case, there is relatively free movement of electrons along the lines of magnetic force from electrode A to electrode B, and the two electrodes of the cathode sheath form a reflective surface for the movement of electrons. Therefore, if we consider the cross section of the device in the plane that includes the axis of symmetry, the two sides of the plasma trap are closed by the magnetic field lines,
Another side is closed by electrode A and the last side by electrode B. In the second case, electrodes A and B do not act together, but are rather completely independent of each other, in which case the magnetic field is one trap adjacent to electrode A, a second trap adjacent to electrode B. Used to complete traps separately.

One example is when magnetic field lines exit electrode A, bend, and then enter electrode A again.

In this case, considering the cross section in the plane containing the axis of symmetry, a trap is formed adjacent to electrode A, one side of this plasma trap is closed by the cathode sheath at electrode A, and the trap is closed on one side by the cathode sheath at electrode A. The remaining three sides are closed by the magnetic field at electrode A. Similarly, a separate trap is created at electrode B, formed by the cathode sheath of electrode B and the magnetic field. In the third case, that is, the hybrid case, by appropriately manipulating the current in the magnetic field generating coil, some magnetic field lines pass from electrode A to electrode B, while other magnetic field lines exit from electrode A and reach electrode A. It is possible to create magnetic field lines that go back and do the same thing at electrode B so that both types of plasma traps, one separate and one bridging between the electrodes, are present at the same time.

When the two previously mentioned electrodes A and B act in conjunction with each other and the magnetic field is configured to confine a single plasma trap, the trap acts to maintain the desired plasma density. The manner in which this is done will be explained next.

For convenience of explanation, consider the moment when electrode A is a cathode with respect to electrode B.

Since both electrodes are cathodes with respect to the plasma, they are both subject to ion bombardment, which causes electron emission from these electrodes in a manner that depends on the secondary emission coefficient yi. Electrons emitted by electrode B are accelerated by the cathode sheath around electrode B and move away from electrode B roughly along the magnetic field lines. Therefore, in the assumed format, the lines of magnetic force extend from electrode B to electrode A, so these emitted electrons move to electrode A. However, since electrode A is a cathode with respect to electrode B, electrons are reflected by the cathode sheath in electrode A and returned to electrode B, where they are reflected again by the cathode sheath of electrode B. The net result of this is that electrons emitted by electrode B are trapped by electrodes A and B. Next, consider electrons emitted from electrode A in contrast to what was previously described for electrons emitted from electrode B.

It was previously mentioned that electrode A is at a more negative potential than electrode B, ie, electrode A is a cathode with respect to electrode B. Electrons are accelerated by the cathode sheath around electrode A, move away from electrode A along magnetic field lines, and finally reach electrode B. Because the gas pressure in the discharge device is low, these electrons lose little or no energy when they reach the vicinity of electrode B, and have enough energy to pass through the cathode sheath at electrode B. , so it actually collides with electrode B. As a result, the trap is free of electrons. However, secondary electrons can be emitted from electrode B as a result of this collision. In contrast to collisions with gas ions, this emission due to electron collisions depends on the secondary emission coefficient ye. From the above explanation, it will be clear that the electron emission from electrode B increases as the electrons initially emitted from electrode A are removed from the trap.

Therefore, although the electrons from electrode A are only trapped for a very short period of time, these electrons enhance the electron emission from electrode B, which substantially traps the electron emission. Therefore, the plasma density of a radio frequency device when both electrodes work together to confine the plasma trap can be determined by the two secondary emission coefficients, i and ye. It should be obvious. Based on a similar analysis,
If the two radio frequency electrodes do not work together and the magnetic field is formed in such a way as to confine separate traps together with the magnetic field, the plasma density is mainly determined by only one coefficient yi, and the second order It can be shown that the emission is mainly induced by ion bombardment, with minimal contribution from electron bombardment.

Referring now to Figures 10a-10d, these figures show a pair of planar electrodes 513, 514 that maintain the discharge and cooperate with the magnetic field to confine the plasma containment trap.
Two embodiments of radio frequency glow discharge methods are shown.

The embodiment shown in FIGS. 10a and 10b and 10c
The main difference from the embodiment shown in Figures 10a and 10d is the shape of the electrodes and the type of plasma trap that is limited. It is shown that the electrodes are letter-shaped and, when taken together, define a circle with a fissure along its diameter.

In FIGS. 10a and 10b, "the axis of rotation is the electrode 513,
514 and a plurality of substrates 52
A toroidal coil 515 located on the opposite side of the electrodes from the target surface facing 0' defines a toroidal magnetic field that cooperates with each electrode to define a hybrid plasma containment trap. Field coil 515 is separated from electrodes 513, 514 by a suitable electrical insulator 516. Each electrode 513, 514 is powered by a suitable radio frequency power source 517 via a shielded bus bar 518, 519, respectively. In Figures 10c and 10d,
Electrodes 513, 514 are in the form of spaced concentric disks, with electrode 513 forming an outer ring-shaped electrode and electrode 514 forming an inner solid disk. Furthermore, a toroidal magnetic field coil winding 515 defines the magnetic field in FIG. 10c, in which case both electrodes 513, 514 act in cooperation with each other and with the magnetic field to define the plasma G trap.FIG. 10a. The axis of symmetry of the plasma trap in the embodiment shown in Figures 10a to 10d is the axis of rotation of the toroidal magnetic field. Although the shape is shown, the electrode is not a planar electrode, but may be a curved sheet-shaped electrode, or may have a shape other than a circular shape.

A glow discharge device of the type shown in FIGS. 10a-10d may include a suitable trunk similar to trunk 315 in any suitable enclosed room environment such as that shown in FIG. 4 for a columnar electrode assembly. It can be installed using

As previously mentioned, one major advantage of the present invention is that relatively low operating pressures can be used.

Using relatively low pressures for the working gas, material to be sputtered from the target is ejected in a substantially straight line and is thus sputtered line-of-sight onto the substrate, in contrast to typical conventional equipment. When using high operating pressures, such as in Furthermore, due to the scattering of the gas, it does not only coat the specific side of the substrate facing the target, but also coats other sides, such as the sides and back side, depending on the orientation of the substrate with respect to the target. The use of relatively low pressures also results in large bond areas.

This is because the mean free path of the sputtered material, ie, the average length of line-of-sight motion without collisions, is inversely proportional to pressure. For this reason, the sputtering apparatus according to the invention can be operated at relatively low pressures, so that the flow of coating material is in line of sight even on substrates arranged at large radii. For this reason, the present invention allows the use of a large room and, accordingly, the use of a large bed area. The invention can be operated at low operating gas pressures such that sputtered material from any point along the length of the target electrode can reach a predetermined substrate location along a line-of-sight trajectory. Therefore, it is possible to analytically predict with considerable accuracy what the effect will be when a barrier plate is placed between the target electrode and the substrate.

Thus, such uniform cutting boards can be used to stack the coating with a controlled gradient of coating thickness, as well as to extend the area of the substrate over which a uniform coating thickness is obtained. Another advantage is that the method of the invention is electrically very stable and reliable with minimal drift in electrical operating parameters even after many hours of operation.

While a particular form of the invention has been illustrated and described, it will be understood from the foregoing that various modifications may be made within the scope of the invention.

The invention is limited only by the scope of the claims.

[Brief explanation of drawings]

Fig. 1 is a simplified side view of an electrode type discharge device according to the idea of the present invention installed in a vacuum chamber, Fig. 2 is a detailed sectional view of the discharge device according to the idea of the invention, and Fig. 3a is a diagram according to the invention. A sectional view of another type of discharge device, FIG. 3b is an end view thereof;
3c and 3d are enlarged partial views showing details of the apparatus of FIG. 3a and variations thereof; FIG. 4 is a double electrode of the invention suitably placed in a suitable vacuum chamber with magnetic field coils; Figures 5a-5d are simplified side views of a cylindrical radio frequency glow discharge device, illustrating embodiments of cylindrical dual electrode structures with a variety of different magnetic field formats that can be used to practice the invention; 6a through 6e illustrate several embodiments of dual electrode construction types that can be used to practice the invention, and FIG. 7 is similar to FIG. 4 but with a hollow cathode assembly. Figures 8a to 8c are diagrams showing the hollow cathode structure as well as the various energy formats that can be used in the present invention, and Figures 9a to 9e are simplified side views showing the case in which the present invention is implemented. Fig. 10a shows an embodiment of the glow discharge device of the present invention using double planar electrodes; Fig. 10b shows some embodiments of hollow cathode structures that can be used in 10a is a bottom view of the electrodes and magnetic field windings of the device shown in FIG. 10a; FIG. 10c is a diagram showing another embodiment of the glow discharge device using double planar electrodes according to the invention; FIG. 10d 10c is a bottom view of the electrodes and field windings of the device shown in FIG. 10c. Explanation of main symbols 10... Vacuum chamber, 14... Cathode,
15... Supporting means, 16... Anode, 17... Hollow spacer (guiding means), 18... Cover, 20...
Inner magnetic field coil, 21... Outer magnetic field coil, 23... Lines of magnetic force, 27... High voltage and magnetic field power supply, 31... Base body,
40... Cathode body, 42, 43... Upper flange, 4
7,57...Cavity, 52,53...Lower flange, 62...Annular chamber, 70...Upper insulator, 71...Gap, 76...Spool, 77...Cylindrical shape fuselage, 77
a, 77b... Upper and lower wings, 85... Coil, 8
7... Central internal chamber, 112... Gap, 164...
Lower insulator, 171... Gap, 172... Lines of magnetic force, 17
3... Cathode sheath, 214... Cathode body, 214a,
214b... end flange, 215... housing, 2
16a, 216b... annular anode, 218, 219...
・Ring-shaped insulator, 228, 229...Gap, 234...
- Magnetic field coil, 252...Base, 256,257...Conducting wire, 311...Vacuum chamber, 312...Double electrode type discharge device, 313,314...Flanged coaxial electrode, 3
15... Mounting trunk, 316... Insulator, 317.
... Radio frequency power supply, 318 ... Shielded transmission line, 31
9... Lines of magnetic force, 326... Substrate, 328, 329,
330, 331, 332, 333...Magnetic field coil, 4
13,414...Hollow conical coaxial electrode, 415...Housing, 419...Insulator, 420...Transmission line, 4
21...Power supply, 429...Magnetic field coil, 431...
・Magnetic field lines, 513, 514... Electrode, 515... Toroidal coil, 517... Radio frequency power supply, 518
, 519... Bus bar with shield. Susugi No F→Nuktaku Futaku Fukuezu Kakeme Make Yuu〆〃Hana〆〃Fuku It's been a long time Meustatakumeusu evening.〆Meusuttaku〃Flower〆〆Meusukui of the flowers Z Meusuku 2〆Iufuku.

Claims (1)

[Claims] 1 In a method of generating glow discharge for forming a thin film on a substrate by cathode sputtering, a method is used in which a vacuum chamber that also serves as an anode and a cathode placed in a low-pressure inert gas atmosphere within the vacuum chamber are used. While applying voltage,
A predetermined magnetic field and cathode structure create a plasma trap that radially and axially restricts the motion of electrons emitted from the cathode, so that most of the energy of these electrons is transferred to additional gas ions. A method for generating a glow discharge characterized in that the glow discharge remains close to a cathode target surface until it is dissipated as useful work in collisions with the cathode.