JP5586286B2 - Plasma processing equipment - Google Patents

Description

The present invention relates to a plasma processing apparatus and a plasma processing method, and more particularly to a plasma processing apparatus and a plasma processing method capable of suppressing the generation of foreign matters due to reaction products.

As a material to be etched used in the field of semiconductor device manufacturing, DRAM (Dyn
For example, volatile materials such as Si, Al, and SiO 2 are used for amic random access memory) or logic circuit ICs. FRAM (Ferroelectric Random Access Memory) or M
Nonvolatile materials such as Fe are being used for RAM (Magnetic Random Access Memory) and the like.

The nonvolatile material is difficult to etch because the melting point of the reaction product generated during the etching is high. Moreover, the vapor pressure of the reaction product after etching is low, and the vacuum container (vacuum processing chamber)
Since the coefficient of adhesion to the inner wall is high, the inner wall of the vacuum vessel is covered with deposits of reaction products only by processing a small number (several to several hundreds) of samples. Further, when this deposit is peeled off, foreign matter is generated.

When the reaction product is deposited, the coupling state between the induction antenna and the plasma in the reaction vessel changes, and the etching rate, the uniformity thereof, the etching perpendicularity, the adhesion state of the reaction product to the etching sidewall, and the like change with time.

Specific examples of the non-volatile material include Fe, NiFe, PtMn, IrMn, DRAM capacitor and gate portions, and FRAM capacitor portions used as ferromagnetic or antiferromagnetic materials for MRAM or magnetic heads. MRAM TMR (Tunneling Magne
to Resistive) Pt, Ir, Au, Ta, and Ru, which are noble metal materials used in the element portion. In addition, high dielectric materials such as Al2O3, HfO3, Ta2O3, and ferroelectric materials such as P
ZT (lead zirconate titanate), BST (barium strontium titanate), SBT (
And strontium bismuth tantalate).
Similarly, in the field of semiconductor device manufacturing, as a semiconductor device manufacturing process, Si
A technique of forming a SiO 2 or SiN film by plasma CVD is often used. In this technique, a polymerizable gas such as monosilane is incident on plasma to form a film on a wafer. At this time, a large amount of the polymerized film adheres to the inner wall of the reaction vessel other than the wafer, which hinders mass production stability. That is, if the polymer film is deposited too thickly on the inner wall of the reaction vessel, the polymer film is peeled off from the inner wall surface and adheres to the wafer as a foreign substance as in the case described above. For this reason, it is necessary to perform plasma cleaning using a violent special gas such as NF3 or manual cleaning performed by opening the reaction vessel.

In the field of semiconductor device manufacturing, a plasma dry etching process of SiO2 is frequently used. In this etching, carbon fluoride such as C4F8, C5F8, CO, CF4, and CHF3 is used as an etching gas. The reaction product produced by the reaction of these gases in the plasma contains a large amount of free radicals such as C, CF, C2F2, etc., and when these free radicals are deposited on the inner wall of the reaction vessel, as in the case described above. It may cause foreign matter. Further, when free radicals are re-evaporated from the deposited film into the plasma, the chemical composition in the plasma changes, and the etching rate of the wafer changes over time. As a conventional plasma processing apparatus, an induction type plasma processing apparatus in which a coiled antenna is provided on the outer periphery of a vacuum container, a plasma processing apparatus for introducing a microwave into the vacuum container, or the like is known. In any of these processing devices,
Since the countermeasure for deposits on the inner wall of the vacuum vessel is not sufficient when etching the nonvolatile material, the manual cleaning with the air release is repeated. When the manual cleaning is started, it takes 6 to 12 hours to start the processing of the next sample, so that the operation efficiency of the apparatus is lowered.
For example, in Patent Documents 1, 2, and 3, a plasma is generated in a processing vessel by an induction method, and a Faraday shield is provided between the induction antenna provided on the outer periphery of the vacuum vessel and the plasma, and a high-frequency power source is provided to the Faraday shield. A plasma processing apparatus is shown that can reduce the adhesion of reaction products to the inner wall of the vacuum vessel or can clean the inner wall of the vacuum vessel by connecting and supplying power.

This apparatus is effective for a portion of the vacuum vessel that is formed of a non-conductive substance, that is, ceramic or quartz, and a portion where the electric field due to the Faraday shield is sufficiently reached. However, it is not effective for portions formed of other non-conductive materials or portions formed of conductive materials.

JP-A-10-275694 Japanese Patent Laid-Open No. 11-61857 JP 2000-323298 A

As described above, when the reaction product is excessively deposited on the inner wall of the vacuum vessel, the deposited film is peeled off from the inner wall surface and adheres to the wafer as foreign matter. In addition, in the plasma processing apparatus using the induction antenna, the coupling state between the induction antenna and the plasma in the reaction vessel changes, and the etching rate, the uniformity thereof, the etching verticality, the adhesion state of the reaction product to the etching sidewall, etc. Change. Further, when cleaning the inner wall of the vacuum vessel, it takes time to start the processing of the next sample, so that the operation efficiency of the apparatus is lowered. In addition, by providing a Faraday shield between the induction antenna provided on the outer periphery of the empty container and the plasma, and connecting the high-frequency power supply to this Faraday shield and supplying power, the adhesion of reaction products to the inner wall of the vacuum container is reduced. Alternatively, the effective range of the plasma processing apparatus that can clean the inner wall of the vacuum vessel is limited.

The present invention has been made in view of these problems, and provides a plasma processing apparatus that controls a deposited film deposited on the inner wall of a vacuum vacuum vessel and is excellent in mass production stability.

  In order to solve the above problems, the present invention employs the following means.

A mounting table disposed at the lower part of the vacuum processing chamber and on which the sample is mounted; a bell jar made of an insulating material that forms the upper part of the vacuum processing chamber and covers the plasma generation space above the mounting table; and an outer periphery of the bell jar A coiled antenna for supplying a high frequency electric field for generating plasma in the plasma generation space in the vacuum processing chamber, and a Faraday shield disposed between the antenna and the bell jar and provided with a high frequency bias voltage, A metal ring-shaped member that is disposed below the lower end of the bell jar and the lower end of the Faraday shield and forms a part of the vacuum processing chamber, and the inner peripheral side of the ring-shaped member and the bell jar in the vacuum processing chamber open the gap to the wall, the ring-shaped lower end portion of the bell jar electric field by the Faraday shield does not sufficiently reach Disposed over them over the upper end of the timber, equipped with a conductor made of a plate-shaped member which is at a predetermined potential.

ADVANTAGE OF THE INVENTION According to this invention, since the deposited film deposited on the inner wall of a vacuum vessel is controlled, the plasma processing apparatus and plasma processing method excellent in mass production stability can be provided.

It is a figure which shows the plasma processing apparatus concerning embodiment of this invention. It is a figure which shows a Faraday shield. It is a figure explaining the method of optimizing FSV. It is a figure explaining feedback control of FSV. It is a figure which shows the example of attachment of the Faraday shield with respect to a bell jar. It is a figure explaining the attachment structure of an adhesion prevention board. It is a figure which shows the example of the heat calculation of an adhesion prevention board. It is a figure explaining the support structure of an adhesion prevention board. It is a figure explaining the deposit countermeasures adhering to a bell jar inner wall. It is a figure explaining adhesion of the deposit in the adhesion board vicinity. It is a figure which shows the structural example of an adhesion prevention board. It is a figure which shows the other structural example of an adhesion prevention board. It is a figure which shows the other structural example of an adhesion prevention board. It is a figure which shows the structure of the sample holding part containing a mounting base. It is a figure which shows the substrate bias circuit containing a susceptor surface. It is a figure explaining the relationship between the thickness of a susceptor and the bias voltage which generate | occur | produces on the susceptor surface. It is a figure explaining the adhesion situation of the deposit to a thin-walled susceptor. It is a figure explaining the adhesion situation of the deposit to a thin-walled susceptor. It is a figure which shows the example which sprayed the metal film on the susceptor lower surface. It is a figure which shows the example which sprayed the metal film on the susceptor lower surface. It is a figure which shows the example which embedded the metal film in the susceptor. It is a figure which shows the example which embedded the metal film in the susceptor. It is a figure which shows the example which applied the susceptor provided with the metal film to the mounting base made from a ceramic dielectric. It is a figure which shows the example which applied the susceptor provided with the metal film to the mounting base made from a ceramic dielectric. It is a figure which shows the connection structure of the electrode for bias application. It is a figure explaining the means to adjust the high frequency bias voltage applied to the susceptor surface. It is a figure explaining the structural example of the means to adjust the high frequency bias voltage applied to the susceptor surface. It is a figure explaining the example which supplies a high frequency bias to a susceptor using another power supply. It is a figure explaining the example of an electrode structure in the case of supplying a high frequency bias to a susceptor using another power supply. It is a figure explaining the method of optimizing a susceptor bias voltage. It is a figure explaining the susceptor bias application circuit with a feedback circuit. It is a figure explaining the susceptor bias application circuit with a feedback circuit. It is a figure explaining each area | region inside a vacuum processing chamber.

A first embodiment of the present invention will be described below with reference to the drawings. In the first embodiment, a method for suppressing deposition of reaction products on the inner wall of a vacuum vessel during processing will be described by taking an example of etching processing when a sample to be plasma processed is a nonvolatile material.

FIG. 1 shows a cross-sectional view of the plasma processing apparatus according to the present embodiment. The vacuum vessel 2 includes a bell jar 12 made of an insulating material (for example, a non-conductive material such as quartz or ceramic) that closes an upper portion thereof to form a vacuum processing chamber. Inside the vacuum vessel, there is provided a mounting table 5 for mounting a sample 13 to be processed, and plasma 6 is generated in the processing chamber to process the sample. The mounting table 5 is formed on the material holding unit 9 including the mounting table.

On the outer periphery of the bell jar 12, a coiled upper antenna 1a and lower antenna 1b are arranged. A disc-shaped Faraday shield 8 that is capacitively coupled to the plasma 6 is provided outside the bell jar 12. The antennas 1a and 1b and the Faraday shield 8 are connected in series to a high frequency power source (first high frequency power source) 10 via a matching unit (matching box) 3 as will be described later. Further, a series resonance circuit (variable capacitor VC3 and reactor L2) capable of changing the magnitude of impedance is connected in parallel between the Faraday shield 8 and the ground.

A processing gas is supplied into the vacuum vessel 2 through a gas supply pipe 4a. The gas in the vacuum vessel 2 is evacuated to a predetermined pressure by the exhaust device 7. Vacuum container 2 from gas supply pipe 4a
In this state, the processing gas is turned into plasma by the action of an electric field generated by the antennas 1a and 1b. A substrate bias power source (second high frequency power source) 11 is connected to the mounting electrode 5. Thereby, ions present in the plasma 6 can be drawn onto the sample 13.

As the high frequency power supply 10, for example, 13.56 MHz, 27.12 MHz, 40.
By using a high frequency power in the HF band such as 68 MHZ or in a VHF band having a higher frequency, the high frequency power is supplied to the inductively coupled antennas 1 a and 1 b and the Faraday shield 8 to generate plasma in the vacuum vessel 2. Can be obtained. At this time, by using the matching unit (matching box) 3, the impedance of the inductively coupled antennas 1 a and 1 b can be matched with the output impedance of the high-frequency power source 10, thereby suppressing power reflection. As the matching unit (matching box) 3, for example, as shown in the figure, a capacitor in which variable capacitors Vc1 and Vc2 are connected in an inverted L shape is used.

As shown in FIG. 2, the Faraday shield is made of a conductor having a vertically striped slit 14 and is disposed so as to overlap with a ceramic vacuum vacuum container (Berger 12). The voltage applied to the Faraday shield 8 can be adjusted by the variable capacitor (VC3 shown in FIG. 1). The voltage (shield voltage) applied to the Faraday shield 8 may be set to an arbitrary value corresponding to the processing recipe for each wafer or corresponding to the cleaning processing recipe.

The principle of cleaning the inner wall of the vacuum vessel by the Faraday shield is that a high-frequency voltage applied to the Faraday shield generates a bias voltage inside the vacuum vessel (the inner wall of the bell jar), thereby drawing ions in the plasma into the vacuum vessel wall, The drawn ions are bombarded against the vacuum vessel wall to cause physical / chemical sputtering, thereby preventing reaction products from adhering to the vacuum vessel wall.

Faraday shield voltage (Farday shield voltage)
aday Shield Voltage (FSV). The optimum FSV is affected by the high frequency power supply frequency, the vacuum vessel wall material, the plasma density, the plasma composition, the entire vacuum vessel configuration and the material of the workpiece, the processing speed, and the processing area. Therefore, the optimum value of the FSV needs to be changed for each process.

FIG. 3 is a diagram for explaining a method for optimizing the FSV, and shows the relationship between the emission intensity (light quantity) of the FSV and the vacuum vessel wall material (for example, aluminum or oxygen constituting alumina when the wall material is alumina). It is a thing. As shown in the figure, the emission of the wall material becomes stronger when the FSV becomes higher than a certain FSV (point b in FIG. 3). This shows that deposits (depots) are deposited on the wall in the FSV below the b point, and not only the deposit is sputtered and deposited in the FSV above the b point, but the wall material itself is also sputtered. Is shown.

The FSV optimum value is the voltage at point b, but depending on the process, point a may be the optimum value. For example, when the wall material is released into the gas phase by sputtering of the vacuum wall material, the processing reaction of the object to be processed or the reaction in the gas phase deviates from what is assumed, and the desired process cannot be executed. To do. That is, by setting the FSV to point a, deposition of a small amount of deposit is recognized on the inner wall of the vacuum vessel, thereby preventing the wall material from being sputtered at all. This makes it possible to prevent process failures due to the release of the wall material. However, before the deposit is sufficiently deposited on the inner wall of the vacuum vessel, it is necessary to clean the inner wall of the vacuum vessel using a process dedicated to cleaning (here, FSV is set higher than the point b).

In contrast to the above, there is also a process for setting the point c to an optimum value. For example, if any reaction product accumulates even a little on the inner wall of the vacuum vessel, foreign matter is generated, or high frequency power for generating plasma is absorbed by the deposit, and the characteristics of the plasma change to stabilize the target process. It may not be possible to execute. In this case, the FSV is set to the point c as described above. That is, the inner wall may be slightly cut, but the reaction product may be set to a condition that does not accumulate at all. In this case, there is a disadvantage that the exhaustion of the vacuum container increases, but the number of cleaning of the inner wall can be reduced.

If neither the inner wall scraping nor the reaction product deposition is desired, the FSV is set at the point b. At this time, it is important to improve the reproducibility of the set voltage of the FSV. This is because it is necessary to suppress a change with time when the same process is performed in different apparatuses or when the same process is continuously performed in the same apparatus. For this purpose, feedback control of the FSV is important.

FIG. 4 is a diagram for explaining feedback control of FSV. As shown in the figure, the output of the plasma generating high-frequency power supply 10 is applied to the Faraday shield 8 via the impedance matching devices (VC1, VC2) and the antennas 1a, 1b. The FSV is divided by capacitors C2 and C3 to form a small signal. After removing harmonics and other frequency components through the filter 15, the detector 16 detects the signal, converts it to a DC voltage, and amplifies it by the amplifier 17. In this way, a DC voltage signal proportional to FSV is obtained. This signal is compared with a preset value or a set value set by the comparator 18 in the recipe output of the main unit control unit 20, and a variable capacitor for controlling the motor via the motor controller 19 to determine the voltage of the FSV. Rotate VC3.

As a result, the FSV can be controlled to a value set by the main device control unit 20, and for example, even when different processes or the same apparatus processes the same process continuously, the FSV value can be controlled to be constant. . In addition, it is possible to suppress disparity between devices and changes with time.

The Faraday shield is capacitively coupled to the plasma through the wall (belger) of the dielectric vacuum vessel. As a result, the FSV is divided by the capacitance between the Faraday shield and the plasma and the capacitance by the ion sheath formed on the wall, and the voltage after the division is applied to the ion sheath. As a result, the ions are accelerated and the inner wall of the vacuum vessel is subjected to ion sputtering. For example, when the wall thickness of an alumina vacuum vessel is 10 mm, the FSV is 500 V.
The voltage applied to the ion sheath is about 60V.

It is useful to increase the voltage applied to the ion sheath with a low FSV. High FSV
This is because the handling becomes more difficult because abnormal discharge is likely to occur. In order to increase the voltage applied to the ion sheath with a low FSV, the capacitance of the ion sheath is uniquely determined by the plasma characteristics of the process used, so the capacitance between the Faraday shield and the plasma is made as small as possible. It is effective to do. In order to realize this, the dielectric constant of the material of the dielectric vacuum container is high, and the thickness of the wall of the dielectric vacuum container is made as thin as possible. As a material suitable for this, typical alumina can be adopted as a material having high strength and high dielectric constant.

When a vacuum vessel with a thin wall thickness is made of a material having a high dielectric constant such as alumina, the problem is a gap between the Faraday shield and the wall (belger) of the vacuum vessel. Since the dielectric constant of alumina is about 8, when the wall thickness is 10 mm, the thickness is 10 /
8 = 1.25 mm. Here, assuming that the gap between the Faraday shield and the vacuum vessel is 0 to 1 mm, the distance between the Faraday shield and the plasma is 1.
It will change nearly 25 times from 25 to 2.25 mm. This means that the voltage applied to the ion sheath changes from about 33 V to 60 V under the above conditions.

In this way, when the voltage applied to the ion sheath changes greatly, deposits (depots) adhere to parts of the inner wall of the vacuum vessel, and deposits do not adhere to other parts, and FSV is applied. By doing so, the effect of suppressing deposit adhesion is reduced. In order to prevent this, it is necessary to make the gap between the Faraday shield and the vacuum container constant, or to manufacture the Faraday shield with a thin film and make it closely contact with the vacuum container.

It is easy to fabricate a Faraday shield by processing a metal plate, but the wall of the vacuum vessel (
It is not realistic to make the gap with Berja) to be 0.5 mm or less. However, a conductive elastic body, for example, a conductive sponge, can be attached under the Faraday shield, and the gap between the Faraday shield and the vacuum vessel wall can be filled with the sponge.

FIG. 5 is a view showing an example of attaching the Faraday shield to the bell jar. FIG.
Is an example in which there is a gap between the Faraday shield 14 and the bell jar 12, and deposits are likely to be deposited on the inner surface of the vacuum vessel where there is a gap. On the other hand, deposits do not accumulate near the bottom of the gap. FIG. 5B is a diagram showing an example in which the gap is filled with an elastic conductor 12a, for example, a conductive sponge. Thereby, the Faraday shield 14 can acquire the same effect as having adhered to the bell jar 12. In addition, since conductive sponge has a large elasticity, it can fill a large and small gap flexibly.

FIG. 6 is a view for explaining the attachment structure of the deposition preventing plate, and FIG. 6 (a) shows the gas outlet 23 formed in the bottom of the bell jar 12 and the gas ring below it. When the plasma treatment is continued in this configuration, deposits are deposited in the portions indicated by A and B in the figure. Deposits can be prevented from being deposited on the inside of the bell jar and above the portion B in the figure due to the effect of ion sputtering by FSV. The problem here is the portion indicated by A and B. The portion A is the periphery of the gas ejection port 23. When deposits adhere to the gas ejection port 23, the deposits easily peel off due to the effect of the gas flow, and the peeled deposits are on the wafer to be processed. It becomes a foreign object and becomes an obstacle to the process. The part B is the inner wall of the bell jar 12.
The Faraday shield 14 is far from the inner wall of Berja. For this reason, the ion sheath voltage by FSV falls and the deposit adhesion inhibitory effect by ion sputtering does not fully work.

  FIG. 6B is a diagram illustrating a structure in which the gas outlet 23 is covered with the deposition preventing plate 22.

The portion A is near the gas outlet, and deposit adhesion to this portion needs to be reduced as much as possible. In order to reduce deposit adhesion to the gas blowing port 23, the plasma 6 region that can be seen from the gas blowing port 23 through the hole of the deposition preventing plate 22 is reduced, that is, the prospect angle for the plasma is reduced, and the gas blowing port 23 Do not look directly at the wafer, ie
The central axis of the gas outlet 23 needs to be set in the direction of the plasma generation space above the sample so that the sample is included outside the expected angle.

FIG. 6C is a diagram showing a detailed example of the relationship between the deposition preventing plate and the gas ejection port. In this example,
The prospective angle is reduced to about 30 degrees, and the wafer is made invisible directly from the gas ejection port.

At this time, it is effective to leave a gap between the deposition preventing plate 22 and the gas ejection port 23. The size of the gap is desirably 0.5 mm or more. This gap creates several advantages. First, even if the gas passage hole formed in the deposition preventing plate has the same size, the prospective angle to the plasma can be reduced by opening the gap, and the amount of deposits adhering to the gas ejection port 23 can be reduced. Further, when the gas is blown out from the gas ejection port 23 into the vacuum container, a large pressure drop occurs, and the flow moves from a viscous flow to an intermediate flow and finally becomes a molecular flow. Here, the gas pressure is still relatively high in the vicinity of the gas outlet 23 and is in an intermediate flow state. When deposits adhere to the gas outlets 23, the deposits are easily peeled off due to the force from the gas flow. By forming the gap, the gas flow in the vicinity of the deposition preventing plate 22 becomes a molecular flow, and the gas flow has less force to peel off the deposit attached to the deposition preventing plate, and the peeling of the deposit can be reduced. Furthermore, as will be described later, the temperature of the deposition preventing plate 22 can be increased efficiently, and the amount of deposits adhering to the deposition preventing plate 22 can be reduced.

FIG. 7 is a diagram illustrating an example of heat calculation of the deposition preventing plate. From the results of the process treatment, it has been found that materials such as Fe and Pt hardly adhere to members having a temperature of 250 ° C. or higher. Therefore, the deposition preventing plate was designed so that the temperature of the deposition preventing plate was 250 ° C. or higher. In the thermal design, the heat balance of the three was calculated: heat input from the plasma, heat escape from the protection plate support, and radiation heat escape from the entire protection plate. The result of this heat calculation is shown in FIG.

When the adhesion preventing plate is SUS (stainless steel), it can be seen that the equilibrium temperature exceeds 250 ° C. when the radio frequency (RF) input to the plasma is about 500 W. When the deposition preventing plate is made of Al (surface anodized), the deposition preventing plate equilibrium temperature is 250 ° C. or more when the RF input is 1000 W or more. The structural features of each part in the calculation will be described below.

Since the plasma heat is diffused isotropically in the reactor, RF input to the plasma ×
It is calculated by the area of the protection plate / the total area of the plasma contact. In the adhesion plate specification designed this time, if the RF input to the plasma is 1200 W, the heat input to the adhesion plate is 260 W.

The heat radiation escape from the adhesion-preventing plate can be suppressed to a low level because the surface radiation rate can be reduced to about 0.2 by using SUS as a material and having a mirror-finished surface. When Al (surface alumite treatment) is used for the adhesion-preventing plate, the alumite surface emissivity is about 0.6, so that the heat radiation escape becomes large.

FIG. 8 is a view for explaining the support structure of the deposition preventing plate. The protection plate supports the entire circumference at three points so that the heat transfer from the support portion is reduced, and the contact area with the gas ring main body is almost point contact as shown in FIG. Suppressed. As a specific example, the contact portion radial length is 3 mm, and the contact portion circumferential length is 1 mm. Even if the contact thermal resistance is estimated to be an excessive value of about 3000 [W / (m 2 · K)], the deposition plate calculated by the contact area × the contact heat transfer coefficient × (the deposition plate inner surface temperature−the gas ring temperature). Heat transfer from the support is only about 10W.

Actually, a protection plate was prototyped and the surface temperature was measured. The material of the deposition preventing plate used is Al (surface anodized). With an RF input of 1200 W, the surface temperature was about 250 ° C., and it was confirmed that the value was almost as designed.

As described above, the adhesion preventing plate cannot completely eliminate the adhesion of the deposit even if it is maintained at a high temperature. For this reason, it is important to stably deposit the deposits attached to the deposition preventing plate. For this reason, it is desirable that the surface of the deposition preventing plate has some unevenness in order to improve the adhesion of the deposit mechanically. According to the inventors' experiment, it is known that the surface roughness is desirably 10 μm or more.

However, when the deposit starts to adhere, the deposited deposit gradually increases in thickness from the state of the thin film. For example, the unevenness of 10 μm provided on the deposition preventing plate has an anchor effect for deposits having the same film thickness. However, the anchor effect decreases as the thickness of the deposited deposit increases. Therefore, in order to make the anchor effect work effectively from an initial state where the amount of deposits is small to a state where the amount of deposits has increased to some extent, two types of unevenness, for example, 10 μm unevenness and 10 μm unevenness
It is desirable to have 0 μm irregularities on the surface at the same time. As a processing method for forming such irregularities, for example, knurl processing can be used for forming irregularities of 100 μm, and blasting can be used for forming irregularities of 10 μm.

As described above, it is desirable that the surface of the deposition preventive plate be mirror-finished in order to raise the temperature of the deposition preventing plate, and that the surface be unevenly formed in order to stably deposit the deposit. Therefore, in reality, the surface of the deposition plate to which deposits adhere (surface facing the plasma) is uneven, and the surface to which deposits do not adhere (for example, between the deposition plate and the gas ejection port). The surface facing the gap can be mirror-finished. In addition, in order to reflect the heat radiated from the deposition preventing plate, it is desirable that the surface of the gas ring where the deposits do not adhere is mirror-finished.

  The size of the deposition preventing plate is desirably the minimum size for covering the gas outlet.

This is because deposits are inevitably deposited on the protective plate, and a thermal history is generated on the protective plate in order to increase the temperature in order to reduce the amount of deposit attached. This is because the deposits are easily peeled off due to the difference in the amount of expansion and contraction due to heat between the deposits and the protective plate material.

Further, it is desirable that the deposition preventing plate is made of an electrically conductive material and grounded. This is because the discharge becomes more stable when the ground contact area with respect to the high frequency for generating plasma is increased. In addition, when the deposit is charged, the deposit is easily peeled off by the repulsive force between the deposits due to the Coulomb force, so that the deposit is also prevented from being charged as much as possible.

The structural design and the thermal design as described above were carried out to produce an adhesion-preventing plate having a surface roughness of 10 μm and 100 μm, and the performance was examined by etching 500 platinum Pt continuously. As a result, almost no deposit was found on the gas outlet. Moreover, the deposit adhering to the adhesion preventing plate was stable, and no peeling of the deposit occurred.

9 is a portion B of FIG. 6 (a portion where the Faraday shield of the inner wall of the bell jar 12 is far from the inner wall of the bell jar. Therefore, a portion where the ion sheath voltage due to the FSV is lowered and the deposit adhesion suppressing effect due to the ion sputtering does not sufficiently work) It is a figure explaining the countermeasure against the deposit adhering to.

In part B of FIG. 6, the distance between the inner wall of the bell jar 12 and the Faraday shield 14 is long.
This is a region where ion sputtering by FSV is hardly effective. Therefore, by extending the deposition preventing plate 22 and covering this portion, it is possible to reduce the deposit adhesion amount and stabilize the deposit. This structure is shown in FIG. 9 (a). When this was used to test the deposit adhesion, it was found that the deposit adhered to a region of about 15 mm width on the inner wall of the bell jar centering on point C in FIG.

FIG. 9B is a modification of FIG. As shown in the figure, the bell jar 12 is formed so that the inner surface thereof is substantially continuous with the inner surface of the gas ring 4, and the bell jar 12 is disposed on the gas ring 4 to form a vacuum processing chamber.

According to this configuration, the deposition preventing plate can be continuously formed on the inner surface of the bell jar and the inner surface of the gas ring. Thereby, the region where the ion sputtering by the FSV is difficult to work can be effectively protected by the deposition preventing plate.

FIG. 10 is a diagram for explaining adhesion of deposits in the vicinity of the deposition preventing plate. First, the dotted lines indicate plasma isodensity lines. When attention is paid to the point C, the point C corresponds to a corner portion made of an adhesion preventing plate and a bell jar, and the plasma density in this portion is slightly lower than the surrounding area.

This is because the plasma is less likely to go around the point C due to the thickness of the deposition preventing plate. For this reason, at point C, the number of ions sputtered per unit area of the inner wall of the bell jar is small, and it is considered that deposits are difficult to remove. There is another possible cause. In other words, the deposition plate is electrically conductive, and the ion sheath generated on the deposition plate has no FSV.
The DC voltage of about 15 to 20 V determined by the plasma characteristics is applied to the ion sheath. On the other hand, in the region where FSV is effective, the ion sheath formed on the inner wall of the bell jar is applied with a high-frequency voltage of about 60 V, for example, on the DC voltage determined by the plasma characteristics. Sputter. In other words, around the point C,
This corresponds to a transition region from a low voltage ion sheath generated on the deposition plate to a high voltage ion sheath formed on the inner wall of the bell jar, and the voltage of the ion sheath increases as the distance from the deposition plate increases near point C. It becomes a region where ion sputtering becomes effective gradually.

Due to the two causes described above, it is considered that a weak sputter region due to FSV is formed in the vicinity of point C as shown in FIG. In this region, it is considered that the deposit adheres because the deposit adheres more preferentially than the sputtering by FSV.

11, 12, and 13 are diagrams showing examples of the structure of the adhesion preventing plate, respectively. As shown in FIG. 11, in order to remove the cause of the decrease in the plasma density, which is one of the causes of the formation of the weak sputter region, a knife edge-shaped deposition preventing plate was manufactured and tested. As a result, as shown in FIG.
It was confirmed that the weak sputter area was reduced and the deposit adhesion area was reduced. Therefore, in order to remove another cause, as shown in FIG. 12, when the upper portion 22a of the deposition preventing plate is changed to an insulator (in this case, alumina), as shown in FIG. Sediment areas could be matched and there was almost no deposit adhesion. Since the surface of alumina cannot be knurled, irregularities were processed on the surface by blasting. Also, quartz or aluminum nitride can be used as the insulator material.

In order to prevent deposit adhesion more thoroughly, as shown in FIG. 13, it is sufficient that the strong sputter region is slightly larger than the deposit adhesion region. Therefore, a gap is provided between the deposition preventing plate and the bell jar so that the plasma can enter between the deposition preventing plate and the bell jar. In order for the plasma to enter, the gap interval needs to be sufficiently larger than the ion sheath, and the gap interval needs to be 5 mm or more. On the other hand, if the size is too large, the deposit is wrapped around by diffusion, so that the effect is reduced. The maximum value of the gap to prevent the deposit from wrapping around is determined by the material / gas type of the deposit and its pressure, and therefore varies depending on the treatment process. However, the approximate result of the test is 15 mm. As a result of producing and testing the adhesion-preventing plate having the structure shown in FIG. 13, adhesion of deposits to the bell jar could be completely suppressed. In the case of this structure, the upper part of the protective plate does not need to be an insulating material, and the performance does not change even if it is made of an electric conductor.

The upper part of the susceptor, which is the cover of the mounting table 5, also causes foreign matter generated on the wafer when deposits adhere. Therefore, a high frequency bias was also applied to the susceptor, and physical and chemical ion sputtering was caused to prevent deposits from adhering.

FIG. 14 is a view showing the structure of the sample holder 9 including the mounting table. As shown in the figure, a mounting table to which the substrate bias power supply 11 is connected is mounted on the ground base 36 and the insulating base 35. As the material for the mounting table, aluminum or titanium alloy is generally used. A dielectric film is formed on the part on which the object to be processed (sample 13) is mounted at the upper part of the mounting table so that the object to be processed can be electrostatically adsorbed. Although the dielectric film is a sprayed film in the drawing, it may be formed of a polymer material such as epoxy, polyimide, or silicone rubber. Ceramic materials formed by thermal spraying include alumina, aluminum nitride, PBN (P
yrolytic Boron Nitride). Further, in FIG. 14, in order to prevent high frequency power from escaping into the plasma from the side surface of the mounting table 5, the grounding base 36 and the insulating cover 3
7 shows a structure for shielding by using 7. The susceptor is generally made of quartz or alumina, and covers the electrode portion other than the surface on which the sample of the mounting table is mounted to prevent damage by plasma.

FIG. 15 is a diagram showing a substrate bias circuit (equivalent circuit) including the susceptor surface. The substrate bias power supply 11 is mixed with the direct current voltage for electrostatic attraction supplied from the electrostatic attraction power supply in the impedance matching unit (MB) 32 and then supplied to the mounting table. Here, the substrate bias power supply 11
The high frequency passes through the susceptor 34 from the mounting table 5 and is also supplied to the upper surface of the susceptor. At this time, the susceptor 34 forms a capacitor using a susceptor material as a dielectric. In FIG. 15, the capacitor formed in this way is represented as a capacitor C (33).

The inventors first examined the deposit adhesion when the susceptor thickness was 5 mm as shown in FIG. As a result, it was found that a large amount of deposits adhered to the upper surface of the susceptor.

Therefore, the relationship between the thickness of the susceptor 34 and the bias voltage generated on the susceptor surface was theoretically examined. The result is shown in FIG. It has been found that when the voltage generated on the inner wall of the bell jar is about 60 V or more, deposit adhesion can be suppressed. Further, in the tests of the inventors, the bias voltage (peak-to-peak) Vpp at the time of the test was often set to a range of about 400 to 500 V. A susceptor thickness of 4 mm was selected so that a voltage could be generated.

FIGS. 17 and 18 are diagrams for explaining the state of deposit attachment to a thin-walled susceptor (for example, 4 mm thick). As shown in FIG. 17, the entire surface of the susceptor was set to a thickness of 4 mm, and the deposition state of the deposit was tested. As a result, it was confirmed that there was no deposit adhering in the range indicated by the arrows in the figure (depot adhesion limiting region). Thereby, it turned out that deposit adhesion can be suppressed in the part which is in direct contact with the mounting table. However, in the configuration of FIG. 17, deposits adhere to the outer peripheral portion of the upper surface of the susceptor, and there is a concern that this may become a foreign substance on the object to be processed and hinder processing. Therefore, except for the insulating cover 37 formed on the side surface of the mounting table, the mounting table and the susceptor are in contact with each other over the entire upper surface of the susceptor and the side surface of the susceptor. This configuration is shown in FIG. Using the structure shown in FIG. 18, the state of deposit adhesion was experimentally investigated in the same manner as described above. As a result, deposits were not adhered on the susceptor upper surface and the susceptor upper surface in contact with the mounting table. However, it has been found that if the susceptor is repeatedly attached and detached, the deposit may not be removed sufficiently even under the same conditions. In addition, it was found that when this deposit could not be removed sufficiently, the deposit was deposited with an uneven distribution, and in particular, the deposit was likely to remain on the side surface of the susceptor.

The deposits were deposited in a distribution with a bias, and the reason why the deposits could not be fully removed was estimated as follows. That is, the material of the susceptor is alumina, the thickness thereof is 4 mm, and the dielectric constant is about 8, which corresponds to about 0.5 mm when converted to an air layer. Here, if the gap between the susceptor and the mounting table is 0.1 mm, for example, the thickness of the dielectric forming the capacitor C in FIG. 15 is 0.5 mm for the susceptor and 0 to 0.1 mm for the gap. Total, varies from 0.5 to 0.6 mm (20%). This fluctuation causes a bias in the high-frequency voltage generated on the susceptor surface, thereby causing a bias in how to remove the deposit. However, it is difficult to manufacture the susceptor and the mounting table in close contact with each other with an accuracy of 0.1 mm or less, which is not practical.

In order to solve this, as shown in FIG. 19, a metal film is sprayed on the lower surface of the susceptor 34 to form a metal spray film 39. Tungsten was used as the sprayed metal because it is known that it has good adhesion to alumina. The metal film need not be tungsten if it has electrical conductivity and good adhesion to the susceptor, and gold, silver, aluminum, copper, etc. are also possible. Also, the metal film manufacturing method does not have to be thermal spraying, and any method can be used as long as it can form a thin film such as plating, sputtering, vapor deposition, printing, coating, and thin film adhesion. By adopting this structure, if the metal film and the mounting table 5 are in contact with each other at one place, the same voltage as the mounting table is generated on the entire metal film, so the problem of the gap between the susceptor and the mounting table can be avoided.

As a result of investigating the adhesion state of the deposit by an experiment using the apparatus configuration of FIG. 19, the adhesion of the deposit could be eliminated with good reproducibility within the deposit adhesion restricted region indicated by the arrow. The advantage of this method is
If the metal film and the mounting table are in contact with each other, the same voltage as the mounting table 5 is generated across the metal film,
A uniform high frequency voltage can be generated on the surface of the susceptor 34. Therefore, as shown in FIG. 20, even when there is another structure such as the insulating cover 37, a uniform high frequency voltage can be generated on the surface of the susceptor in an arbitrary range by expanding the spraying range of the metal sprayed film. . In addition,
In the configuration of FIG. 20, it was experimentally confirmed that deposit adhesion can be eliminated with good reproducibility in the deposit adhesion limited region indicated by the arrow.

From the above results, it was found that by using a metal film such as metal spray, a high-frequency voltage can be uniformly generated on the susceptor surface, and the deposit adhesion can be suppressed uniformly. When this technique is used, the same effect can be obtained by embedding a metal film in the susceptor even when the thickness of the susceptor must be increased structurally. This structure is shown in FIGS.

As shown in these figures, a position at a predetermined depth from the surface of the susceptor 34 (in the case of the figure, approximately 4 mm).
The metal sprayed film 39 is embedded, contacts are made from the metal sprayed film 39 to the mounting table 5 to ensure electrical continuity, and the same high frequency voltage as that of the mounting table 5 is generated on the metal sprayed film 39.

The mounting table on which the sample 13 is mounted includes ceramic dielectrics such as aluminum nitride and alumina in addition to the type of forming an electrostatic adsorption film on a metal mounting table by thermal spraying as shown in the drawings. There is a type in which a metal electrode is embedded in a manufactured mounting table, and electrostatic adsorption or high frequency bias is applied from the metal electrode. Even in the case of this kind of mounting table, it is possible to manufacture a susceptor having exactly the same function by forming a metal film on the susceptor.

Examples of this are shown in FIGS. FIG. 23 shows a case where a metal film is formed on the back surface of the susceptor 34. The mounting table 5 is made of aluminum nitride, and an electrostatic adsorption / high frequency bias application electrode 40 made of tungsten is embedded in the mounting table 5. From this electrode, a metal sprayed film 39 is formed.
The conductive pattern (the collar conductive patterns 41, 42, 43) is embedded toward the electrode 40, whereby the electrode 40 and the metal sprayed film 39 are electrically connected. Thereby, the metal sprayed film 3 on the back surface of the susceptor
9 can generate the same high-frequency voltage as the tungsten electrode. As a matter of course,
The ability to control deposit adhesion to the susceptor surface by this structure can be made exactly the same as described above.

FIG. 24 shows an example in which a metal sprayed film 39 is embedded in the susceptor 34. The conductive pattern described in FIG. 23 (trunk conductive patterns 41, 42, 43) is extended and the mounting table 5 is embedded with contacts. If the metal sprayed film 39 buried in the electrode 40 and the susceptor 34 is connected, the function can be exactly the same as in the case of FIG.

In the case of the mounting electrode 5 shown in FIGS. 23 and 24, a pattern for supplying a high frequency to the metal sprayed film 39 from the electrostatic adsorption / high frequency bias applying electrode 40 embedded in the electrode 5 is formed inside the mounting electrode 5. One example is shown in FIG. 25.

In FIG. 25, the collar conductive pattern 41 that is in parallel with the tungsten electrostatic adsorption / high frequency bias applying electrode 40 embeds a tungsten thin film in the mounting electrode in the same manner as the tungsten electrode. After these embedded tungsten thin films are molded,
It can be connected by drilling the necessary part and brazing the penetrating terminal.

In the bias application method for the susceptor described so far, when the high-frequency voltage of the mounting table is a certain value (here, 400 V), the deposit adhesion on the susceptor upper surface is just suppressed. However, when the voltage of the mounting table is increased, the high-frequency voltage on the upper surface of the susceptor becomes too high, and there is a disadvantage that the susceptor is scraped and the component life is shortened. This drawback can be solved by introducing means for adjusting the high-frequency bias voltage applied to the susceptor surface from the outside, as shown in FIG. FIG. 26 shows a circuit for adjusting the voltage of the metal film of the susceptor with a variable capacitor VC attached to the outside. This is shown in FIG. 27 as an actual structure.

A ceramic coating 50 is formed on the surface of the mounting table in contact with the susceptor by thermal spraying or the like so that the susceptor metal sprayed film 51 and the mounting table 5 do not directly contact each other. This ceramic coating 50 forms a capacitor C ′ shown in FIG. 26 and has a function of transmitting a part of the high-frequency voltage applied to the mounting table 5 to the susceptor metal sprayed film 51. Then, the high frequency voltage applied to the mounting table 5 is transmitted to the susceptor metal sprayed film 51 by another external variable capacitor VC. Since the high-frequency voltages transmitted by the two capacitors have the same phase, they are simply added, and the high-frequency voltage generated on the susceptor surface is determined by the voltage. For example, assuming that the susceptor thickness is 4 mm, the surface area of the susceptor metal sprayed film is 400 cm 2, the ceramic film is 300 μm of alumina, and the maximum capacity of the variable capacitor VC is 8000 pF, the variable capacitor VC is When the capacitance of the susceptor is varied, the susceptor surface voltage can be varied in the range of about 30 to 100V. Thus, the high frequency voltage generated on the susceptor surface can be controlled by appropriately selecting the susceptor thickness, ceramic coating, metal spray coating surface area, and variable capacitor VC. Further, although the susceptor metal sprayed film at this time is not shown, it can be incorporated in the susceptor as long as it can be connected to the variable capacitor VC.

In order to vary the bias applied to the susceptor, a high frequency power source 1 for supplying a high frequency to the mounting table
It is also possible to use a high frequency power source different from 1. This is shown in FIG. Here, a susceptor bias power supply 11a for supplying a high frequency to the susceptor metal film is used separately from the substrate bias power supply 11 for supplying a bias to the mounting table. The electrode structure in this case is shown in FIG. What is important here is that an insulation and ground shield (grounding base 36) is provided between the mounting table 5 and the susceptor metal sprayed film 51 so that the high frequency voltage applied to the susceptor metal sprayed film 51 is not affected by the high frequency voltage of the mounting table 5. It is necessary to incorporate. Thus, although there is a disadvantage that the susceptor bias power supply 11a is necessary, the bias applied to the susceptor is the same as that of the sample 13.
It becomes possible to control completely independently of the high-frequency voltage applied to. Further, the susceptor metal sprayed film 51 at this time is not shown, but can be incorporated in the susceptor as long as it can be connected to the susceptor bias power supply 11a.

FIG. 30 is a diagram for explaining a method of optimizing the susceptor bias voltage. FSV mentioned above
As with, the susceptor bias voltage has an optimum value. This voltage is influenced by the frequency of the bias power source, the susceptor material and thickness, the plasma density, the plasma composition, the entire vacuum vessel configuration and the sample material, the processing speed, and the processing area.

Therefore, the optimum value of the susceptor bias voltage needs to be changed for each process. As in the case of FIG. 3, when the susceptor bias voltage increases with a certain value (point b in FIG. 30) as a boundary, the susceptor material emits light. When the susceptor bias voltage is lower than the point b, the deposit is deposited on the susceptor, and when the susceptor bias voltage is higher than the point b, the deposit is sputtered and not deposited, and the susceptor material itself is sputtered. Is shown.

The optimum susceptor bias voltage is the voltage at point b, but it may be at point a depending on the process. This corresponds to a case where the susceptor material is released into the gas phase by sputtering of the susceptor material, and the sample processing reaction or reaction in the gas phase deviates from what is assumed, so that a desirable process cannot be performed. That is, by setting the susceptor bias voltage to point a, deposition of a deposit is recognized although it is slight on the susceptor, and the susceptor material is not sputtered at all. This is to prevent process failure due to the release of the susceptor material. Instead, it is necessary to clean the susceptor by a cleaning-dedicated process (here, the susceptor bias voltage is set higher than the point b) before the susceptor deposits are sufficiently deposited.

On the other hand, if even a small amount of deposit adheres to the susceptor, the target process may not be stably executed due to the generation of foreign matters. In this case, the optimum point of the susceptor bias voltage can be set to the point c, and the susceptor may be slightly cut, but may be set to a condition in which no deposit is attached. In this case, there is a disadvantage that the consumption of the susceptor increases, but there is an advantage that the cleaning of the susceptor may be small.

If both the susceptor scraping and deposit adhesion are desired, the susceptor bias voltage is set to the point b. At this time, it is important to improve the reproducibility of the set voltage of the susceptor bias voltage. This is because it is necessary to suppress a change with time when the same process is performed in different apparatuses or when the same process is continuously performed in the same apparatus. For this purpose, feedback control of the susceptor bias voltage is important.

31 and 32 are susceptor bias applying circuits with feedback control circuits corresponding to FIGS. 26 and 28, respectively. In both circuits, the voltage of the susceptor metal sprayed film is detected through an attenuator and a filter 52 and converted to a DC voltage. Thereby, this DC voltage signal becomes a signal proportional to the susceptor bias voltage. By comparing this signal with a preset value or a set value set by a recipe or the like of the main body device control unit 57, in the case of FIG. 31, the motor that rotates the variable capacitor VC that determines the susceptor bias voltage is controlled. In the case of FIG. 32, the output of the susceptor bias power supply 11a is controlled. By using this method, the susceptor bias voltage can be controlled to the value set in the main unit, and the susceptor bias voltage value can be kept constant when processing the same process continuously in different devices or the same device. It is possible to control, and to suppress the difference between devices and the change with time.

The above is a region in which the deposit is controlled not to be deposited or adhered, that is, the bell jar 1.
2. The method and structure of the gas outlet 23 and the susceptor 34 have been described. As long as the reactive organisms coming out of the sample 13 and the substances synthesized in the gas phase are volatile components having a high vapor pressure, these substances are exhausted from the discharge part and the periphery of the object to be processed by the exhaust device, and the electrodes Although some deposit in the lower part and the exhaust duct, most of them are exhausted.

However, highly deposited substances, that is, vapor pressure is low and the adhesion coefficient to solids is close to 1 (almost trapped when contacted with solids) are used as sample reaction products or in the gas phase. If synthesized, these substances are deposited on the vacuum vessel wall including the bell jar, susceptor or gas outlet around the sample and are hardly exhausted.

In such a situation, if the deposits are controlled not to adhere to any part in the vacuum vessel, these highly depositing substances lose their deposition sites. For this reason, the density in the gas phase of a highly depositing substance increases, and the driving force for the deposition increases. As a result, the material is forcibly deposited on a bell jar or a susceptor.

That is, controlling the deposits not to adhere to the bell jar, the susceptor or the like produces the effect by preparing a place where the deposits are deposited in some amount. Then, by increasing the amount of deposits that can be deposited, or by quickly depositing from the gas phase, the deposit amount control capability of the deposits in the bell jar and the susceptor can be increased.

In other words, a region in which a large amount of deposit is quickly deposited from the gas phase around the object to be processed in which a highly depositable reaction product is generated or the plasma region (deposit trap region).
It is necessary to provide. The deposition preventive plate functions as a cover that suppresses deposits from adhering to the gas outlet, but it is also a kind of trap because it is assumed that deposits accumulate on itself.

  FIG. 33 shows the inside of the vacuum vessel divided into regions including the deposit trap.

First, the bell jar region and the wafer (sample) / susceptor region are regions that are controlled so as not to deposit deposits. All other areas in contact with the plasma are deposit trap areas, and the deposit trap area (1) is an area including the deposition plate and the lower part of the gas ring, and can be seen directly from the wafer (can be seen). It is an area. These bell jar region, wafer / susceptor region and deposit trap region (1) are all the regions that can be directly viewed (viewed) from the wafer, are regions that generate plasma, and from the wafer, Alternatively, a region having a strong deposition property formed in a plasma gas phase is most likely to adhere. If deposits are deposited in an uncontrolled state in these regions, foreign matter may be caused on the wafer or plasma may change over time. Therefore, in those areas that can be viewed directly from these wafers, deposit deposition must be controlled as completely as possible.

In the present invention, when the structure shown in FIGS. 12 and 13 is used for the deposition preventing plate, 100% of the area that can be seen from these wafers is in a state in which the deposits are controlled. Further, even if the structures of FIGS. 6, 9, and 11 are used, the surface area of the region that can be viewed from these wafers is 90%.
% Or more of the deposits must be controlled.

In addition, as described above, the deposit trap region performs a sufficient function to increase the deposit suppression function of the bell jar region and the wafer / susceptor region. Therefore, the surface area of the bell jar region and the susceptor region is as small as possible. It is desirable to make the surface area of the sediment trap region (1) as large as possible. When reactive organisms with strong deposition properties are generated from the wafer, assuming that the surface area of the wafer is SW, the surface area S1 of the deposit trapping region (1) is S1 <0.
. It has been found through experiments by the inventors that the deposit suppressing function in the bell jar region and the wafer / susceptor region decreases at 5 SW. Therefore, in order to quickly deposit reactive organisms in the sediment trap, a relationship of S1> = 0.5S1 is necessary, and S1> = S1 is desirable.

The sediment trap region (2) is called a ring cover, and the sediment trap region (1)
At the bottom of. Although this region cannot be desired directly from the wafer, a highly depositable material is transported by diffusion to adhere to the upper portion of the region. The deposit trap region (3) is a cover on the side of the electrode, which is also not directly desired from the wafer, but like the deposit trap region (2), a large amount of deposit is deposited on top of it. Since these deposit trap regions (2) and (3) are regions that cannot be directly expected from the wafer, it is unlikely that deposits attached to these will become foreign matter on the wafer or cause changes in plasma over time. These sediment traps are important for efficient cleaning work when the apparatus is opened to the atmosphere. In other words, since reactive organisms are highly sedimentary, 90% or more of them are deposited trap areas (1) (2)
(3) It can be attached and collected. Therefore, these sediment trap regions (1
) (2) By replacing (3) with a swap kit (replaceable) and replacing all parts that have been cleaned after being released into the atmosphere, the inside of the vacuum vessel can be efficiently cleaned. This requires two conditions for the sediment trap to be lightweight and easy to remove / install. In order to be lightweight, it is important that the material of the deposit trap is a lightweight member such as aluminum.

After opening the vacuum vessel to the atmosphere, remove the traps (1) to (2) and (3) in order, and perform the minimum necessary cleaning operation. The minimum required cleaning place is, for example, the periphery of the wafer transfer opening. A swap kit of cleaned sediment traps can then be installed in reverse order and vacuum can be entered immediately. Thereby, the cleaning operation can be performed in a minimum time. Performing the cleaning operation in such a procedure not only shortens the cleaning time, but also shortens the time required for evacuation. The reason is that opening to the atmosphere for the minimum necessary time can minimize the moisture in the atmosphere that is adsorbed by non-vacuum components, and the minimum required amount of cleaning solvent (pure water, alcohol, etc.) This is because the amount of the solvent remaining in the vacuum vessel can be minimized by using. The removed sediment traps (1), (2), and (3) are reused as a swap kit for the next air release / cleaning operation after cleaning. The area to be swapped as a sediment trap need not be limited to the area shown in FIG. Although it depends on the process and the material to be handled, it is efficient to make the entire region where the deposit adheres as a deposit trap. For example, when deposits adhere only to the upper half or more of the electrode cover, the upper half of the electrode cover is made into a swap kit. On the contrary, under the condition that the deposit adheres to the exhaust duct, the inner wall of the exhaust duct is also used as the deposit trapping region, and it becomes efficient if a swap kit is used.

1a Upper antenna 1b Lower antenna 2 Vacuum processing chamber 3, 32, 32a Matching box (matching unit)
4 Gas ring 4a Gas supply pipe 5 Mounting table 6 Plasma 7 Exhaust device 8 Faraday shield 9 Sample holder 10 High frequency power source (first high frequency power source)
11 Substrate bias power supply (second high frequency power supply)
11a susceptor bias power supply 12 bell jar 12a elastic conductor 13 sample 14 slit 15 filter 16 detector 17 amplifier 18 comparator 19 motor controller 20 main unit control unit 21 bell jar restrainer 22 deposition plate 23 gas outlet 24 flange 31 electrostatic adsorption Power supply 33 Capacitor 34 Susceptor 35 Insulating base 36 Grounding base 37 Insulating cover 38 Sprayed film 39 Metal sprayed film 40 Electrostatic adsorption / high frequency bias applying electrode 41, 42, 43 Saddle conductive pattern 50 Ceramic coating 51 Susceptor metal sprayed film 52 Attenuator and Filter 53 Detector 54 Amplifier 55 Comparator 57 Main Unit Control Unit VC1, VC2, VC3 Variable Capacitor

Claims (5)

A mounting table disposed at the bottom of the vacuum processing chamber and on which the sample is mounted;
A bell jar made of an insulating material that forms the upper part of the vacuum processing chamber and covers the plasma generation space above the mounting table;
A coiled antenna that is arranged on the outer periphery of the bell jar and supplies a high-frequency electric field for generating plasma in the plasma generation space in the vacuum processing chamber;
A Faraday shield disposed between the antenna and the bell jar and provided with a high frequency bias voltage;
A metal ring-shaped member disposed below the lower end of the bell jar and the lower end of the Faraday shield and constituting a part of the vacuum processing chamber ;
A gap is formed between the ring-shaped member and the inner peripheral side wall surface of the bell jar in the vacuum processing chamber, and the lower end portion of the bell jar and the upper end portion of the ring-shaped member where the electric field due to the Faraday shield is not sufficiently reached. over are placed over them, the plasma processing apparatus and a conductor made of a plate-shaped member which is at a predetermined potential. Keep plasma processing apparatus according to claim 1, wherein the plate-like member, the plasma processing apparatus the distance between the Faraday shield and the inner wall surface of the bell jar covers long region.   3. The plasma processing apparatus according to claim 1, wherein the gap is set to a value between 5 mm and 16 mm.   4. The plasma processing apparatus according to claim 1, wherein an inner wall surface of a lower end portion of the bell jar is disposed so as to be vertically continuous with an inner wall surface of the ring-shaped member disposed adjacent to the bell jar. Plasma processing apparatus. 5. The plasma processing apparatus according to claim 1, further comprising a high-frequency power source that supplies a high-frequency voltage to the Faraday shield. 6.

Images