JPH01263A - Method and apparatus for providing high resistance layer by cathode sputtering - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention provides a method according to the preamble of claim 1, in which a conductive target material is made reactive under reduced pressure using a magnetron cathode supplied with a direct current voltage. A high resistance layer is formed on the substrate by cathode sputtering under continuous supply of reactive gas in an atmosphere.
In this case, the substrate is held or guided in a plane parallel to the target surface, and a shielding device is present between the target surface and the plane of the substrate, the shielding device being positive or self-contained with respect to the cathode potential. The present invention relates to a method of providing a highly resistive layer on a substrate, which is maintained at a potential that is adjustable.

[Conventional technology]

While spraying low-resistance metal layers presents only minor technical problems, the production of high-resistance layers is relatively problematic. This is because, as the operating time progresses, the metallic device parts, which must have a certain potential difference with respect to one another, are coated with increasingly more resistive layer material, and this coating then becomes larger in area. This results in a flow of process parameters that makes the coating process unstable, so that the products produced by the process have different properties as the working time increases.

A "high resistance layer" refers to a layer with a resistivity above 5.times.10@-4 ohm/GrrL, although this limit value is not critical and may be evaluated differently on a case by case basis. A layer of this type already has the properties of an insulating layer on the condensation surface for the dispersed particles.

The reactive gas may be a pure reactive gas, such as, for example, oxygen or nitrogen, or a mixture of at least one of these gases and an inert gas, such as, for example, argon, required for the sputtering process. .

In this case, in order for the sputtering method to be able to work with direct voltage, the target must be electrically conductive, and it generally consists of a metal that forms the reactive component of the layer to be produced and other The reactive components of are derived from the reactant gas.

The cooperative action of the metal vapor flow and the reactant gas results in the deposition of the desired compound on the substrate. However, the reactant gases act on the target surface and create an insulating coating on the more or less large turbine) f surface, which reduces the sputtering rate and compromises the stability of the process. In this case, what is largely free from the insulating coating is the typical plasma discharge in magnetron cathodes, where the plasma discharge is confined.
There is only an erosion groove running through a tunnel of magnetic field lines. At the bottom of the erosion groove, the sputtering rate is so high that in practically usable method embodiments no unfavorable insulating coating can be formed. However, in this case a series of other method parameters must still be taken into account.

- To obtain sufficient process stability, one generally works with high decoupling between substrate and target. At high decoupling, the effect of the reactant gas on the substrate is relatively high compared to the effect on the target, due to the area ratio of substrate area to effective target area (area of the erosion groove on the target). High decoupling is achieved inter alia by a large distance between target and substrate, which is typically about 100H.

Another means of decoupling is the separation of the working gas (e.g. aldene) and the reactant gas at the inlet point, where the working gas is introduced near the target and the reactant gas (or mixture) is introduced near the substrate. A relatively high method stability is achieved if a constriction is additionally introduced between the gas inlet points and sufficiently separates the region near the substrate and the region near the target. This constriction once acts as an additional blanket surface for the reactant gas in the vicinity of the target. This reduces the introduction of reactive gas into the area near the target and at the same time increases the action of the reactive gas on the substrate. In particular, when working with a diaphragm when applying an insulating layer, the discharge current must be maintained by means of an additional anode, which is arranged on the side of the diaphragm facing the substrate (see German Patent Application No. 3).
331707 specification).

However, the known solutions have technical drawbacks.

For example, the resulting deposition rate is significantly reduced due to the target-substrate distance, in particular because about 60% of the sprayed target material impinges on the aperture. Furthermore, the density of the plasma is significantly reduced in the vicinity of the substrate due to the aperture.

This necessitates working with a large excess of reducing gas, which in itself is contrary to the desire to increase process stability by moving the reactant gas away from the target. Furthermore, the known solutions are also particularly unsuitable for coating with ions, since the specific properties of the layer can only be achieved with strong plasma action. Furthermore, known devices are prone to operational failures. This is because the layers grown on the striations, especially in the case of insulating layers, tend to peel off.

[Problem that the invention seeks to solve]

Therefore, the problem of the present invention is to
It is an object of the invention to provide a process of the type mentioned at the outset, which allows high rate deposition with great process stability.

[Means for solving problems]

The solution to the problem set out is achieved according to the invention by the means specified in the characterizing part of claim 1, namely: a) directing the reactant gas to the target at a distance as small as possible from the target a-plane, but viewed in plan; b) the distance “a” between the target surface and the substrate surface is 601Ei
Keep it smaller than I, especially 40-50mm, C
) by combining the means of selecting the distance "d" between the surface of the shielding device facing the target and the substrate to be less than half the distance "a".

These measures, in contrast to the prior art, result in extremely high coupling between the reactant gas and the target on one side and between the beam target and the substrate on the other side, which is especially true at the gas inlet in the vicinity of the target. arrangement and short substrate distance. However, in solving the problem, obtaining a high deposition rate also plays an important role, and this is achieved in practice by working without a drawstring. in this case,
The aperture in the shielding device should be sized such that a large portion of the particle stream can pass through the aperture. The opening in the shielding device practically no longer acts as an aperture if it approximately corresponds to the width of the target or is even larger than this.

It is therefore advantageous to choose the distance "X" between the inner edges of the shielding device between 0.6 and 2.0 times the lateral dimension "C" of the tarquet.

For disc-shaped targets, the lateral dimension M CI+ corresponds to the diameter of the target and the distance "x" corresponds to the inner diameter of the shielding device. This dimension can easily be transferred to rectangular targets, in which the inner edges of the shielding devices extend parallel to each other at a distance "X'.

According to the stated sizing provisions, at least 60% of the particle flow
% can flow unhindered towards the substrate.

In this case, it is particularly advantageous in certain configurations for the shielding device, which also performs an anode function, to be arranged in close proximity to the substrate surface. In this case, it is particularly advantageous if the distance "d" between the side of the shielding device facing the target and the substrate corresponds at most to the thickness of the shielding device plus the gap width 5111IE. This allows the plasma to act practically unhindered and with high density on the substrate or on the layer forming on the groove.

In the method according to the invention, the discharge current cannot reach the normally existing dark shield or plasma shield, but is guided in the vicinity of the substrate.

Thus, the plasma density in the vicinity of the substrate is increased. This makes effective use of the reaction gas. The result is that much less reactant gas is required to obtain a particular degree of reactivity than in methods with lump decoupling. In contrast to all previous inventions, by creating a potential state of this type in the area of the shielding device or the anode, it is surprisingly possible to achieve an increase in the basic stability of the discharge.

A very important advantage of the method according to the invention is that the coating parameters and thus the layer properties remain constant over time. When chemical compounds or high-resistance layers are deposited by reaction, it is inevitable that an insulating layer is deposited on the shielding device or other device parts that are located in the vicinity of the reaction zone. Since the device parts described are generally electrically conductive metal parts, the potential distribution changes during the course of coating. This actually causes the effective anode to move over time from near the cathode to near the substrate. Depending on the deposition rate and layer thickness, this travel time can range from a few minutes to over an hour. The consequences of such changes in potential conditions on layer properties are intolerable.

In fact, the method according to the invention also results in the deposition of coating substances on the shielding device, but the method according to the invention makes it possible to ensure that this effect does not have a detrimental effect on the long-term stability of the method parameters. .

This means that from the beginning at 0.degree. C. an optimum defined potential state is achieved for the coating process.

In this case, on the one hand, the suction force of the vacuum pump device that determines the reduced pressure, and on the other hand, the amount of reaction gas supplied per unit time, from the reaction zone between the target and the substrate, the amount b of the reaction gas supplied, It is particularly advantageous to arrange for a smaller amount than the lumps to be discharged unconsumed.

This measure also contradicts the known technology, in which normally a much larger amount of the reactant gas supplied, in most cases even several times more, is discharged unconsumed by means of a vacuum pump.

In this case, the ratio of the amount of reactant gas supplied to the amount drawn off unconsumed can be determined in a particularly simple manner by selecting the gap width between the inner casing surrounding the cathode and the substrate. Accordingly, the difference between the total pressure inside the inner casing and the total pressure inside the vacuum chamber surrounding the inner casing is a factor of 2.
It can be made smaller towards.

These measures, as already mentioned, differ from the prior art, but the prior art always states that the method stability is improved on the one hand by increasing the suction power of the vacuum pump, and on the other hand by increasing the feed flow of the reaction gas. The reason for this is due to the high throughput of the reaction gas, which results in only a small portion of the reaction gas being converted into the desired compound.

However, when increasing the suction power of a vacuum pump, there are other difficulties,
This means that in layers with high reactivity, there is a condensation rate limitation or, for a given deposition rate, a reactivity limitation. The reason appears to be that as the reactant gas flow rate increases, the gE strike rate, which is necessary to adjust a specific reactivity, must also increase.

The reduction in the suction power of the vacuum pump equipment was also accompanied by a reduction in the amount of unconsumed reactant gas discharged, which made it possible to achieve not only a high process stability but also a higher reactivity than before. I was able to observe that.

This allows the reactant gas to flow from an entry point near the target almost completely into the dense plasma region in front of the target surface, from where it diffuses towards the substrate, where a small portion is deposited there. Most of it is consumed in chemical reactions before being sucked through the gap. Thus, the density of excited reactive gas particles six times in front of the substrate increases obviously. This again reduces the impact number ratio required for the required reactivity adjustment. For this reason, a large amount of metal per unit time can be reacted with a high degree of reactivity if the reaction gas flow is equal or even smaller. This is because when applying the method of the present invention, introducing a reactive gas near the target, especially the narrow bond itself between the target and the substrate, is expected to reduce the basic stability of the discharge. It also explains how the scales operate stably.

The invention also relates to a device for carrying out the inventive method according to the preamble of claim 10.

The solution to the same problem is achieved by the features in the characterizing part of claim 10.

Further advantageous developments of the subject matter of the invention are apparent from the further claims.

〔Example〕

Next, embodiments of the present invention will be described in detail with reference to FIGS. 1 to 3.

FIG. 1 shows a magnetron cathode 1, which as such belongs to the state of the art. Further details of the principle structure of such a magnetron cathode can be seen from FIG. The magnetron cathode 1 is connected by a conductor 2 to a DC voltage source 3 capable of emitting a negative DC voltage of between about 400 and 1000 volts. The magnetron cathode 1 is coated with a substantially congruent target 4 of conductive material which provides one component of the layer material to be formed. The target has a surface 4a, which is the so-called sputtering surface.

The magnetron cathode or target 4 is surrounded at its periphery by a distribution device 5 for the reaction gas, in which case the spacing is chosen to be as small as possible; Look, they don't cross each other. The outlet holes of the distribution device 5, which are not shown in detail, point downwards. Target 5 has horizontal dimension ″′
C”. In the case of a circular target 4, the dispensing device 5
is also shown as an annular conduit, which is not shown in the drawings for clarity.

A substrate 6 to be coated with reaction products from the target material and the reaction gas passes past the magnetron cathode 1 aligned in a plane parallel to the Ft plane 4a. The distance from the target surface 4a to the substrate 6 has a dimension Na". This distance may be between 40 and 601111, and the lateral dimension "C" of the target is not particularly important and has a normal dimension. This distance"a", there is a shielding device T in the immediate vicinity of the substrate 6, and its inner edge 7a has a mutual distance nx"
, and this distance lies between 6 and 2.0 times the lateral dimension IIc" of the target 4. The variability of this distance "X" is dominated by the dotted extension of the shielding device γ. ing.

The shielding device 7 is arranged insulatingly, which is represented by the support insulator 8 shown on the left. moreover,
The shielding device is connected by a conductor 9 to a DC voltage source 10 supplying a DC voltage between 0 and +200 volts. The surface 7b of the shielding device facing the target 4 has a distance "d" from the substrate 6 which is less than half the distance Na".

In this embodiment, the distance "d" is at most equal to the thickness of the shielding device 1 plus the gap width of approximately 5n.

In the right part (part A) of Fig. 1, there is a magnetron cathode 1.
The device is shown having a relatively large distance from the wall 12 of the vacuum chamber beside which it lies laterally. This distance is bridged by another wall 13 of the vacuum chamber, both of which are at ground potential. In this case, the distance of the vacuum chamber wall 12 from the central axis of the cathode 1 has a dimension "b", which is b
=2c. In this device, the chamber walls 12 and 13 with layers 14 and 15 are at least partially made of an insulating material, so that the chamber walls gradually lose their anodic function compared to the starting state.

In the left-hand part of FIG. 1 (part B) a magnetron cathode with narrowed spacing is shown. Tsu-! The walls 13 of the vacuum chamber at ground potential are significantly shorter. In this case, in order to prevent changes in the potential state as the coating time progresses, the immediately adjacent side walls 1T of the vacuum chamber are suspended insulatingly, which is represented by a support insulator 18. As a result, the wall 17 is itself set to a freely selected intermediate potential.

Both configurations according to FIG. 1 are particularly suitable for depositing layers with a certain residual conductivity, such as 5102 for example.

In this case, the part acting as an anode is connected to earth or to a positive potential with respect to earth and is placed directly in front of the substrate in the particle stream. In this case, the layer deposited on the part that acts as an anode still has sufficient electrical conductivity to receive the discharge current. An insulating layer can easily be deposited on the other side. If the shielding device γ, which acts as an anode, is at zero potential, it can be short-circuited with the wall 12 (FIG. 2, part A). Usually, however, the shielding device T acting as an anode is separated by surrounding walls acting as a plasma screen (for example if these walls can be "floating" like wall 11). It is located.

This is because these walls are suspended insulatingly (FIG. 1, section B-).

Both configurations according to FIG. 2 correspond in principle and arrangement to the corresponding configurations A and B according to FIG. 1, so that the same reference numerals are used. However, as already mentioned above, the arrangement according to FIG. 2 is designed for the production of insulating layers. In this case, an insulating layer 19 is also deposited on the shielding device T, so that the shielding device can no longer perform the function of an anode, at least after a while. In order to ensure the anode function here, a special anode 20, which is designed as a water flow tube, is arranged behind the shielding device 7 in the sputtering device. The anode 20 is connected by a conductive wire 21 to a DC voltage source 22, which can apply a positive voltage of about 40 to 120 μm to the anode. In this case, the anode 20 is set back in the gap 11 relative to the edge 7a, so that only a very small portion of the anode can be hit by the insulating particles. The shielding device 7 together with the walls 12 and 13 is either connected to earth potential (FIG. 2, part A) or is automatically kept at a freely adjustable potential by means of supporting insulators 8 (second Figure, part B).

In each case, the shielding device 1 also has a "C-shielding function" for the special anode 20.

As can be seen from FIG. 2, both the anode 20 and the shielding device 7 are arranged in close proximity to the substrate and in such a way that the proportion of the coating material reaching the substrate 6 is not actually changed by the shielding device 7. There is. The potential state in the entire plasma space, in particular in the vicinity of the substrate, remains temporally and spatially stable if an insulating layer is generated or is already present on the surface. However, in this case the above-mentioned parts must be electrically conductive. This is because otherwise these parts would not be able to fulfill their task of suppressing unfavorable lateral discharges within the area of the cathode.

In order to obtain operating conditions that are stable from the beginning, it can also be provided that the surface is intentionally provided with an insulating layer with a chemical phase before the actual coating of the substrate.

In the embodiment according to FIG. 6, the vacuum chamber 23, which has been omitted from FIGS. 1 and 2, is also schematically illustrated.

This vacuum chamber is connected by a suction pipe 24 to a set of vacuum pumps, not shown. Inside the vacuum chamber 23 is arranged an inner casing 25 which surrounds the magnetron cathode 1 and the distribution device 5 as closely as possible. However, in this case the cathode 1 with its target 4 is shown in some detail υ, and here the magnetic poles N and S are shown, which form a closed field line tunnel above the target surface 4a, Plasma P is confined within this. In this case, the plasma follows the course of the magnetic pole surface; in the case of a circular magnetron cathode, it has approximately the shape of a nodule. The magnetron cathode 1 still has a cathode base body 1a. The conduit 2 is here constructed as a supporting tube and penetrates the wall of the vacuum chamber 23 by means of a feedthrough insulator 26 . The sides, back and conduit 2 are surrounded by a dark room shielding device 27 which is also at ground potential.

In this case, the shielding device 7 is part of the inner casing 250 and is therefore at ground potential. Between the inner edges 7a, there is a first
Similarly to FIG. 2, an outlet 30 for the coating material is formed with dimension "x". However, the shielding device T and the substrate 6
A surrounding gap 11 with a defined @W and a defined length S is formed on the periphery of the inner casing 25 between. This creates a constriction effect at constant degrees Celsius, which allows gas exchange by flow and diffusion. The distribution device 5 supplies the pure reaction gas or the working gas (
A reactant gas mixed with argon) can be supplied.

Outside the inner casing 25 and inside the vacuum chamber 23,
A further distribution device 28 is arranged, with which, for example, pure working gas (argon) is supplied.

An inner casing 25 delimits a reaction zone R between the target 9 and the substrate 6 (similar to the wall 1γ in FIGS. 1 and 2), and a gap 11 allows unconsumed gas to be removed from the reaction zone R. An equilibrium state of gas movement is achieved such that a smaller amount than the ridge can be discharged by the suction pipe 24. The gap 29 on the back surface of the substrate 6 is not necessary. This gap is formed by a plate 30, which can be used in a plasma process if the substrate 6 consists of, for example, a series of individual plate-shaped substrates, which pass through the reaction zone R at intervals. This is important.

By means of the above-mentioned means, the reactant gas flows (through the distribution device 5) close to the target into a dense plasma region almost completely in front of the target surface and from there diffuses into the substrate;
Most of it is consumed here in chemical reactions. Simply, the remaining amount above is only 1! l! 11. In this way, the excited reaction gas is directly drawn in by the vacuum pump and removed with it from the reaction to the compound at the substrate. Thus, the density of excited reactive gas particles in front of the substrate surface increases.

This again results in a smaller impulse number ratio needed to adjust the required reactivity.

As illustrated in FIG. It consists in the selection of the effective suction power of the vacuum pump.

In this case, two cases are distinguished: in addition to the reaction gas, a working gas (argon) is also introduced into the inner housing 25, or the reaction gas alone is introduced into said casing. The flow rate as the sum of the flow rates of the reaction gas and of the working gas at the selected total pressure in the inner casing 25, which in the first mentioned case is determined by the dimensions of the gap, is determined by the total pressure in the vacuum chamber 23 such that the effective suction of the vacuum pump Must be adapted to the force. In the second case, the working gas is introduced into the vacuum chamber by means of a distribution device and passes by diffusion into the inner casing 25, where the reaction process proceeds. In this case, the gap 11 is dimensioned essentially only for the reaction to be attracted to the current. But in both cases,
It is common that the partial pressure difference between the inner casing 25 and the space within the vacuum chamber 23 is small in order to ensure that a small reaction through the gap 11 generates only an electric current.

In the device according to FIG. 6, the effective suction force is 1001 /
Adjusted to s. The magnetron cathode had a length of 500 mm, and a stable current supply device supplied 1000 W of power. The total pressure inside the vacuum chamber 23 is 4 x 10-3 mIJ bar, and the distance between the target W surface and the substrate is "a".
was 40 mm. In this case, the distribution device 5')
A flow of pure oxygen of 350 sec to the inner casing 25
A flow of 2505 ccIrL of argon was introduced by the distributor 28. The pressure difference obtained on both sides of gap 11 was 1.7 x 10-3 mbar.

The substrate holding and guiding device is not shown in FIGS. 1-6. The problem with this is, for example, 1
substrate holders that rotate about two axes or are otherwise mounted stationary, or plate-like or frame-like substrate holders that can run on rollers or rails in a vacuum chamber.

Particularly for the production of insulating layers, the substrate holder has a self-regulating potential (
"float, endis potenzial")
It is necessary to exist in For such layers, the substrate holder, which is at ground potential, has a strong influence on the potential state around the cathode. This is because conductive surfaces within this space can affect the effectiveness of the anode and thereby lead to layer irregularities and arc discharges. Even for conductive layers, the floating potential of the substrate holder is not disadvantageous in any case.

[Brief explanation of the drawing]

The accompanying drawings show embodiments of the invention, FIG. 1 being a schematic cross-sectional view of two configurations of an apparatus for producing layers of still considerable electrical conductivity, and FIG. FIG. 3 is a schematic cross-sectional view showing two configurations of an apparatus for producing discernible layers; FIG. . DESCRIPTION OF SYMBOLS 1... Magnetron cathode, 2... Conductive wire, 3... DC voltage source, 4... Target, 4a... Tarpit surface, 5... Distribution device, 6... Substrate, γ... - Shielding device, 7a... Inner edge, 8... Support insulator, 9...
・Conducting wire, 10... DC voltage source, 11... Gap, 12
... Wall, 13... Wall, 14... Layer, 15... Layer, 13... Wall, 17... Wall, 18... Supporting insulator, 19... Layer, 20 ...Anode, 21...Conductor, 2
2... DC voltage source, 23... Vacuum chamber, 24... Suction pipe, 25... Internal casing, 26... Penetrating insulator, 2γ... Dark part shielding device, 28... Distribution Device,
29...Gap, 30...Plate.

Claims (1)

[Claims] 1. Using a magnetron cathode supplied with a DC voltage, a conductive target material is formed into a high resistance layer on a substrate by cathode sputtering in a reactive atmosphere under reduced pressure and with continuous supply of a reactive gas. , the substrate is held or guided in a plane parallel to the target surface, and a shielding device is present between the target surface and the plane of the substrate, the shielding device being positive with respect to the cathode potential. A method of providing a highly resistive layer on a substrate, which is maintained at a potential that is self-regulating or self-regulating, in which: a) the reactant gas is introduced at as small a distance as possible from the target surface, but outside the contour of the target when viewed in plan; b) the distance “a” between the target surface and the substrate surface is 6
by cathodic sputtering, characterized in that: c) the distance "d" between the surface of the shielding device facing the target and the substrate is selected to be smaller than half the distance "a"; A method of providing a high resistance layer. 2. The method of claim 1, wherein the distance "x" between the inner edges of the shielding device is selected between 0.6 and 2.0 times the lateral dimension "c" of the target. 3. The method of claim 1, wherein the distance "d" corresponds at most to the thickness of the shielding device plus the gap of 5 mm. 4. The method of claim 1, wherein the shielding device is at a potential such that at least a % of the discharge current is directed onto the shielding device. 5. The method as claimed in claim 1, wherein the cathode additionally located behind the shielding device in the sputtering direction is at a potential such that at least 70% of the discharge current is conducted onto the anode. 6. The method of claim 5, wherein the shielding device is at ground potential. 7. The method of claim 5, wherein the shielding device is insulated and self-regulating. 8. On the one hand, the suction pipe of the vacuum device determines the reduced pressure, and on the other hand, the amount of reactant gas supplied per unit time is set to 1/3 of the reactant gas supplied from the reaction zone between the target and the substrate.
2. The method according to claim 1, wherein the method is adjusted such that a smaller amount is discharged unconsumed. 9. The method according to claim 8, wherein the ratio between the amount of reactant gas supplied and the amount drawn in unconsumed is determined by the width of the gap between the inner casing surrounding the cathode and the substrate. 10. A vacuum pump, a magnetron cathode carrying a target, a shielding device with a substrate holding or guiding device and having a definable or self-regulating potential, which is arranged between the plane of the target surface and the plane of the substrate; 2. A device for carrying out the method according to claim 1, comprising a reactant gas supply and distribution device, characterized in that: a) the reactant gas distribution device (5) is located on the target surface (4);
a) at the smallest possible distance from but seen in plan,
a shielding device arranged outside the contour of the target, b) the distance “a” between the target surface (4a) and the surface of the substrate (6) is less than 60 mm, and c) facing the target; (7) An apparatus for providing a high resistance layer by cathode sputtering, characterized in that the distance "d" between the surface of (7) and the substrate (6) is less than half the distance "a". 11. Distance "d" is maximum, thickness of shielding device (7) and 5
11. The device according to claim 10, which corresponds to the sum of the gaps in mm. 12. The distance “x” between the inner edges (7a) of the shielding device (7) is 0.6 times the lateral dimension “c” of the target (4) and 2.0
11. The apparatus of claim 10, wherein the apparatus is between 1 and 2 times. 13. The reaction zone (R) between the target (4) and the substrate (6) as well as the distribution device (5) for the reaction gas is located in the vacuum chamber (
surrounded by an inner casing (25) arranged inside the substrate (23), which casing is surrounded by a substrate (6) on one side;
) is connected to the interior space of the vacuum chamber by at least one gap (11), which gap is dimensioned in such a way that less than one-third of the unconsumed reaction gas can be discharged from the reaction zone. 11. The device according to claim 10, wherein: