JPH03174802A - Method and apparatus for automatic tuning for microwave circuit for generating plasma - Google Patents

Abstract

PURPOSE:To allow the device to cope with a plasma load by applying tuning to a tuning object value to which a constant giving a travelling wave, reflecting wave or standing wave state avoiding the discontinuous point of a load impedance and closest at least to the matching point is set. CONSTITUTION:An automatic tuning device consists of a waveguide 1, standing wave detectors 6a-6c, a tuning controller 7, a stub drive motors M1-M3 and stubs S1-S3. Then the tuning is applied to a tuning object value to which a constant giving a travelling wave, reflecting wave or standing wave state avoiding the discontinuous point of a load impedance and closest at least to the matching point is set. Thus, the tuning is a nonlinear load like a plasma load.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an automatic tuning device for a microwave circuit for plasma generation.

[Problems to be solved by conventional technology and invention] No. 16
The figure is a schematic configuration diagram showing a conventional automatic matching device using a stub used in a microwave circuit. In the figure, 1
2 is a waveguide for passing microwaves, and 2a to 2e are standing wave detectors consisting of five antennas, which are arranged at appropriate intervals in the longitudinal direction of the waveguide 1. 3a-3
d is a stub, in which a first composite stub consisting of a pair of 3a and 3b and a second composite stub consisting of a pair of 3c and 3d are arranged at appropriate intervals, and are inserted into the pipe from the waveguide wall. The drawer is installed so that it can be pulled out freely.

Ml and M2 are stub drive motors for causing the stubs 3a, 3b, 3c, and 3d to see-saw, and 20 is an alignment control device.

In the above configuration, vA-vb-vd and VB-1/2 (Ve+V
a) Since VA and VB of -Vc are orthogonal to each other, these outputs drive the stub drive motor Ml.

The impedance matching of the load is performed by rotating M2 and adjusting the stubs 3a, 3b, 3c, and 3d so that both VA and VB become zero (Japanese Patent Application Laid-Open No. 15502-1982).
issue).

However, when the above-mentioned automatic alignment device is employed in semiconductor manufacturing equipment using plasma, the following problems occur.

In other words, the plasma load within the membrane is non-linear, i.e., as shown in FIG. 17, when the microwave power is increased or decreased, the impedance 1z1 of the plasma load changes, and discontinuities f, , f2 occur at I21 due to hysteresis. occurs. It has been confirmed that this discontinuity point changes variously depending on the type of gas used to generate plasma, the pressure, etc.

Now, assuming that the impedance IZI of the load is the value at point e shown in Fig. 17, the stub starts moving and moves in the direction of matching, so the power supplied to the load gradually decreases. When the power increases and reaches a power at which a discontinuous point f occurs, a phenomenon occurs in which 121 jumps at this point. At this time, since 121 suddenly changes, the reflected power of the microwave power increases, and as shown by the dotted line in FIG. 17, the power supplied to the load decreases, for example, returns to point e, and this Power is supplied at the point.

Subsequently, the above-mentioned operation is performed again, but from then on, this operation is simply repeated, and the state in which matching cannot be achieved continues.

[Means for solving problems]

In order to solve the above-mentioned problems, the present invention does not perform matching, but focuses on traveling waves and reflected waves detected between a microwave power source and a load of a microwave circuit, or standing waves. In an automatic tuning method in which impedance matching is performed by driving a stub using a signal obtained from a wave, discontinuities in the load impedance that occur in the plasma load are avoided, and at least the traveling wave, reflected wave, or standing wave closest to the matching point is used. It is characterized by tuning to a tuning target value set with a constant that gives the wave state.

[Effect]

The present invention is particularly applicable to plasma loads by avoiding discontinuities in load impedance and tuning to a tuning target value with constants that provide traveling and reflected waves or standing wave conditions that are at least closest to the matching point. can do.

〔Example〕

The first embodiment corresponds to claims 1 and 2 of the present invention.

FIG. 1 is a schematic configuration diagram showing a first embodiment of an automatic tuning device that implements the method of the present invention. In the figure, 1 is a waveguide, 6a, 6b, and 5c are standing wave detectors capable of detecting voltage standing wave amplitude, and three antennas are arranged at intervals of, for example, 1/6 of the average pipe wavelength λg. They are arranged sequentially in the longitudinal direction of the waveguide 1. S1. S2. S3
are stubs located on the load side of the standing wave detectors 6a to 6c, and the three stubs are 1/4 of the average tube wavelength λg.
. M
l, M2. M3 is a stub drive motor that drives each stub 5l-33, and 7 is a computer-based tuning control device.

Next, an example of automatic tuning of a plasma generation microwave circuit performed using the above-mentioned stub will be explained using the flowchart shown in FIG. 2, and each step will be explained using FIGS. 3 to 13. I will explain. First, when operating the standard standing wave detector as shown in FIG.
(r'-dX reflection coefficient), and the phase at which the voltage standing wave amplitude is p at an arbitrary position Po, and the phase at which the voltage standing wave amplitude is detected by the standard standing wave detector 6a is a. Let θ be the declination angle.

(1) Give a predetermined arbitrary reflection coefficient Γa and deflection angle θa, and set the conductance Ga and susceptance Ba obtained at that time to the tuning target values, for example (Ga, jBa) shown in FIGS. 6 and 11. Step A),
As shown in FIG. 7, the conductance Gb and susceptance Bb (Gb, jBb) are determined by phase-shifting the conductance Ga and susceptance Ba by 180'' (step B).

(2) The standing wave detectors 6a to 6c detect the standing wave amplitude a,
b, detect C (step C).

(3) Installation point P a -PC of standing wave detectors 6a to 6c
The voltage standing wave amplitudes a, b, and c correspond to a, b, and C on the waveform showing the voltage standing wave distribution in FIG. Considering that,
The result is as shown in the vector diagram of FIG.

Therefore, a, b, c, and d, r”, θ
Expressions (1) to (3> hold true between .

a22=Γ′ 2+d2−2r” dXeos(π−
θ) ... (1) b2-r" 2+d
2-2r”dxcos(π-θ+4/3yr
) −(2)22 c −r” 1d2−2r” d xcos(π−θ+2/3π) ・・・(3)(
The simultaneous equations 1) to (3) are calculated to determine the reflection coefficient F with the declination angle θ and the incident voltage amplitude d-1, and the distance between the detector 6a and the stub S3 is simply set to 1 of the average tube wavelength λg. /2
If the wavelengths are made to be an integral multiple of , the wavelengths will be in the same phase, so θ and r at this time will be those at the position of the reference stub, for example S3 (step D).

(4) As shown in Figure 9, the coordinates determined by r and θ are
Stub S2. In step E), it is determined whether the stubs S2. When in the tunable region X by S3, the initial susceptance Bo at the position of the stub S3 is determined as follows.

The reflection coefficient F at any point on the line is expressed by equation (4) using a complex number.

r" −l r' l e jθ−u+jv
...(4) And the admittance Y, which is the reciprocal of the line impedance, is expressed by the following equation (5).

Y-(12=Γ')/(1+r')-c+j13=
(1u-jV)/(1+u+jv)...(5) Here, G: conductance, B: susceptance, (
Equations (6> and (7>) are obtained from Equation 5>, respectively.

G-(1-u2-v2)/[(1+u) 2+v2]・
...(6) B=-2v/[(1+u) 10v2] -(7)
Equations (6> and (7) can be transformed into equations (8) and (9>), respectively.

[u+G/ (G+1)] 2+v2 - [1/ (G+1) Ko 2
... (8)(u+1) +(
v+1/B) 2 - (1/B) 2...(9
) In other words, equations (8) and (9), as shown in FIG. jv) is (G, j
It can also be expressed as A stepwise tuning operation is performed. Therefore, the initial susceptance Bo at the position of stub S3
It is necessary to find (Step F).

Here, the conductance G at this time is set as Go for convenience, and the coordinates r'(u, jv) are set as r'(uo).

j　vo)−(Go, j　Bo).

(5) When the susceptance Bo is determined, the stub S2.
The susceptance B2.corresponds to the insertion length of S3. Find B3 as follows.

First, as shown in Fig. 6, find the coordinates (u2.jv2) of the intersection of the circle symmetrical with respect to the origin O of the circle G-Gb (the circle G'-Gb) and the circle G-Go. (Step G), then from this coordinate (u2.jv2), by formula <7),
The susceptance B2' is determined (step H), and the initial susceptance Bo is subtracted from the susceptance B2' to obtain the susceptance B3 corresponding to the insertion length of the stub S3.
(-82'-Bo) is determined (Step I).

(6) Furthermore, as shown in FIG.
That is, the susceptance B2' at the position of the stub S2
Step J) Subtract the susceptance B2' from the susceptance Bb of the tuning target value to obtain the susceptance B2 (-Bb) corresponding to the insertion length of the stub S2.
−82”) (step K).

(7) Susceptance B2. When B3 is determined, stub S2. Drive S3 as follows.

By measuring in advance the susceptance B with respect to the insertion length of the stub, the relationship between L and B is as shown in FIG. 8, so that the insertion length of the stub with respect to a certain susceptance B can be determined. Therefore, the susceptance B2. The insertion length L2°L3 of each stub with respect to B3 is calculated and determined (step L), and the insertion length L2°L3 of each stub S2°S3 is determined as shown in FIG.
2. By driving to L3, tuning to the tuning target value is completed (step M).

However, in reality, as soon as the stub starts moving, the power supplied to the load gradually increases until the desired insertion length is reached, and the load impedance changes. Because there is a difference between the load impedance of
A deviation occurs from the tuning target values Ga and Ba shown in the figure.

Therefore, since the output of the standing wave detector constantly changes with each sampling, it is necessary to repeat the steps C to M described above, and by repeating this, the tuning target value is finally reached and the tuning is completed, and the shaded area in FIG. 9 becomes the tunable region X.

(8) In step E described above, stub SI.

When in the tunable region Y by B2, Γ,
The initial susceptance Bo at the position of the stub s3 determined from θ is determined by the stub interval λ as shown in FIG.
Initial susceptance B at the position rotated by g/4-180", that is, the position of the second reference stub s2
o'(Go',jBo') (step N
).

(9) When the susceptance Bo′ is determined, the stub Sl
, S2 corresponding to the insertion length Bl.

Find B2 as follows.

First, as shown in Figure 11, the circle of G=Go' and the circle of G=
The coordinates (u
l, jvl) (step o), and then calculate this coordinate (
ul, jvl) by equation (7) to find the susceptance Bl' (step P), and then calculate the susceptance B
By subtracting the initial susceptance Bo' from l', we get the susceptance B2 (-Bl
'-Bo') is determined (step Q).

(10) Furthermore, as shown in FIG. 12, the susceptance Bl' is rotated by /4-180'' to the stub spacing λ, that is, the susceptance Bl' at the stub Sl.
Step R), the susceptance Bl' is subtracted from the susceptance Ba of the tuning target value to obtain the susceptance Bl (=Ba
-Bl') is determined (step S).

(11) When the susceptances Bl and B2 are determined, the stub S1. B2 is driven as follows.

Similar to step L described above, the susceptances B1 and B2
The insertion length of each stub for L1. L2 is calculated and obtained (step T), and the stubs Sl and S2 are set as L1.L2, respectively, as shown in FIG. By driving to L2, the tuning to the tuning target value is completed.Step U
), the shaded area is the tunable region Y.

In addition, regarding the setting of the tuning target value, measure the discontinuous point of the plasma load under each condition, avoid this point, and set r closest to the matching point.

All you have to do is input θ. In addition, in the above explanation, the interval between the standing wave detectors was set to 1/6 of the average tube wavelength λg, but λg
/6 (excluding integral multiples of λg/2). For example, if λg/3 is used, the above-mentioned equations (1) to (3> become as follows.

2 a 2=Γ′ +d2−2r′ dXcos(
π−θ) 2 b 2=Γ′ +d2−2r′ dXcos(
π−θ+ 8/:1(r) 2 c 2=Γ′ +d2−2r′ dXcos(
r-θ+4/: (r) Also, the distance between the detector 6a and the stub S3 may be any value, and the declination angle θ' at that time can be easily obtained from λg/2-2π, so the detection Changes in specifications for the spacing between detectors, spacing between stubs, and spacing between detectors and stubs can be easily accommodated by partially changing the software.

The second embodiment corresponds to claims 1 and 3 of the present invention.

FIG. 14 is a schematic configuration diagram showing a second embodiment of an automatic tuning device that implements the method of the present invention. In the figure, two stubs 3a and 3b are placed at the center of the wide side of the waveguide 1, approximately at the center of the frequency band used, and separated by a distance of (1/4) of the channel wavelength. It shows a set of composite stubs that are differentially and sequentially varied such that when the insertion length is maximum, the insertion length of the other is zero, and in between, an increase in one is a decrease in the other. , two sets of such composite stubs 3a, 3b and 3d are arranged with a distance of (1/8) of the 3C% average pipe wavelength. 2a to 2e are 5
This is a standing wave detector consisting of two antennas, each of which is placed at the center of the wide side of the waveguide at a distance of (1/8) the average guide wavelength. At this time, the detected voltage amplitude corresponding to the waveguide input power is set to Vi, and the reflection coefficient when looking at the load from the central antenna 2C is set to r. Now, if we ignore errors due to frequency changes, the output voltage of each antenna will be as shown in the following equation.

Ve-kl Vtl 2(1+I r' l 2-2
l r' l cos(θ-π)]Vd-kl Vtl
2 c+++ r' l 2-21 I' l
cos (θ-π/2)) Vc=kl Vt l 2(
1+l r' l 2-21 r' I co
s θ]Vb-kl Vt + 2 (1+l
r l 2-2 l r' I cos(θ+
π/2)] Va-kl Vtl 2 [1+l r'
l 2-2 I r' l cos(θ+ π
)] Now, the antenna 2b is connected to the input terminal of the differential amplifier 10a.

Adding the output of 2d and taking the difference voltage output, VA=
Vb-Vd-4k l Vt l 2I r l si
n El, and this output V^ and the output Sa of the first tuning target value setter A are input to the differential amplifier 11a. Also, (1/
2) and the output voltage of the antenna 2c, the differential amplifier 1
When calculated using 0b, VB-1(Ve+Va)-Vc-4k l Vt l
21 r' I cos θ, and this output and the output sb of the second tuning target value setter B are input to the differential amplifier 11b.

Note that since VA and VB are orthogonal, differential amplifier 1
Since the outputs of 1a and llb are also orthogonal to each other, these outputs are applied to power amplifiers 12a and 12b, and these outputs rotate stub drive motors M1 and M2, which drive composite stubs 3a, 3b and 3c, respectively.

When 3d is adjusted, when both differential amplifiers 11a and llb become zero, automatic tuning is performed to avoid discontinuities in the load impedance and to a point closest to the matching point.

The third embodiment corresponds to claims 1 and 3 of the present invention.

FIG. 15 is a schematic configuration diagram showing a third embodiment of an automatic tuning device that implements the method of the present invention. In the figure, a traveling wave power component incident from the power supply side is extracted by a directional coupler 13, and its output is divided into two by a signal dividing circuit 15 and applied to composite detectors 17a and 17b, respectively.

On the other hand, the reflected wave component reflected from the load is extracted by the directional coupler 14, and this output is sent to the 90 degree component generator 16.
The signal is divided into three signals having the same phase and a phase difference of 90 degrees, and is applied to the other input terminals of the composite detectors 17a and 17b. This composite detector performs square-law detection and synthesizes the sum and difference voltages of the inputs at both terminals, and the output voltage of one composite detector is VA = 4kla Rl r' 1cosθ, and this output VA and the first tuning The output Sa of the target value setter A is input to the differential amplifier 18a.

In addition, the output voltage of the other composite detector is VB-4kla 121 because one of the input signals has a phase difference of 90 degrees.
r l5inθ, and this output VB and the output sb of the second tuning target value setter B are connected to the differential amplifier 18b.
Enter. Since the outputs of the differential amplifiers 18a and 18b are orthogonal to each other as in the second embodiment, these outputs are applied to the power amplifiers 19a and 19b, and these outputs are used to drive the stub driving motor M1. Rotate M2,
When the composite stubs 3a, 3b and 3c, 3d are respectively adjusted, when the differential amplifiers 18a, 18b both become zero, the load impedance discontinuity point is avoided and automatic tuning is performed to the point closest to the matching point. Ru.

Regarding the setting of the tuning target value in the second and third embodiments, the point of discontinuity of the plasma load under each condition is measured, and r, sInθ, or All you have to do is input r' and cos θ.

〔Effect of the invention〕

As described above, according to the present invention, by tuning to a tuning target value set with a constant that avoids discontinuities in load impedance and at least provides a standing wave state closest to a matching point, it is possible to It can handle such nonlinear loads.

[Brief explanation of the drawing]

Fig. 1 is a schematic configuration diagram showing a first embodiment of an automatic tuning device that implements the method of the present invention, Fig. 2 is a flowchart showing an example of operation based on the present invention, and Fig. 3 is a voltage standing wave distribution. A waveform diagram showing voltage standing wave amplitudes a, b,
A vector diagram showing c, FIG. 5 is a Smith diagram showing a circle passing through the initial susceptance Bo at the position of stub S3,
Fig. 6 is a Smith diagram showing how to obtain susceptance B3 from the intersection of the circle G'-Gb and the circle G-Go, and Fig. 7 shows the conversion and tuning target value to susceptance B2' at the position of stub S2. A Smith diagram showing a 180" phase conversion, FIG. 8 is a diagram showing the relationship between susceptance B and the insertion length of the stub, and FIG. 9 shows the locus of stubs S2 and S3 and their tunable region X. Smith diagram, Figure 1O is a Smith diagram showing the conversion to the initial susceptance Bo' at the position of stub S2, Figure 11 is the susceptance calculated from the intersection of the circle G'-Ga and the circle G-Go'. A Smith diagram showing how to obtain B2, Fig. 12 is a Smith diagram showing conversion to susceptance Bl' in stub Sl, and Fig. 13 is a Smith line showing the trajectory of stubs S1 and S2 and its tunable region Y. Figure, 14th
15 and 15 are schematic configuration diagrams showing second and third embodiments of an automatic tuning device implementing the method of the present invention, respectively. FIG. 16 is a schematic configuration diagram showing a conventional automatic tuning device using a stub, and FIG. FIG. 17 is a diagram showing impedance changes of plasma load. 1... Waveguide, 2 a to 2 e, 6 a to 6
c Constant wave detector, 3a, 3b, 3c, 3d--Composite stub, 7,...Tuning control device, 13.14
... directional coupler, A... first tuning target value setter, B... second tuning target value setter,
Ml-M3... Stub drive motor, S1 to S3.
··stub.

Claims (3)

[Claims] 1. In an automatic tuning method in which impedance matching is achieved by driving a stub using signals obtained from both traveling waves and reflected waves or standing waves detected between the microwave power source and the load of a microwave circuit, plasma load A microwave circuit for plasma generation that avoids a discontinuous point in load impedance that occurs at least at a matching point, and tunes to a tuning target value set with a constant that provides both a traveling wave and a reflected wave or a standing wave state closest to a matching point. automatic tuning method. 2. In an automatic tuning device for a microwave circuit for plasma generation using a stub tuner used in a microwave circuit, there is a waveguide 1 through which microwaves pass, and a waveguide 1 having a diameter of 1/6 of the average internal wavelength λg in the longitudinal direction of the waveguide 1. Three standing wave detectors 6a, 6b, 6c arranged sequentially at intervals
and an average tube wavelength λg at a position different from the standing wave detector.
three stubs S1, S2, and S3 sequentially arranged at intervals of 1/4, 1/6, and 1/8 of the stub S;
Stub drive motors M1, M2, M that drive 1 to S3
3, and each of the voltage standing wave amplitudes a, b, and c detected by the detectors 6a to 6c are input, and the incident voltage amplitude at the position Pa of the standing wave detector 6a as a reference is d, Let the apparent reflection coefficient be Γ′ (=d×reflection coefficient Γ),
When the polarization angle between the phase where the voltage standing wave amplitude is p at an arbitrary position Po of the waveguide and the phase where the voltage standing wave amplitude is a at the position Pa is θ, a^2=Γ'^2+d^2-2Γ'd ×cos(π-θ) b^2=Γ'^2+d^2-2Γ'd ×cos(π-θ+4/3π) c^2-Γ'^2+d^2-2Γ 'd × cos(π-θ+2/3π)
Find the reflection coefficient Γ that is set to 1, and then find the conductance G and susceptance B from Γ and θ at this time, and calculate the conductance Ga and susceptance Ba found when giving a predetermined arbitrary reflection coefficient Γa and deflection angle θa. Set the conductance G and susceptance B to conductance Ga and susceptance Ba as tuning target values.
Find the susceptances B1 to B3 at the positions of the stubs S1 to S3 necessary to change the stubs S1 to S3, and find the insertion lengths L1 to L3 of each of the stubs S1 to S3 necessary to obtain the susceptances B1 to B3. and each stub from L1 to L
for plasma generation, comprising: a tuning control device that inputs a signal for inserting only 3 into a stub drive motor that drives the stub, and feeds back output signals of the detectors 6a to 6c that change at that time; Automatic tuning device for microwave circuits. 3. A standing wave detector or directional coupler and stub are placed between the microwave power source and the load, and the reflection coefficient |Γ| In an automatic tuning device for a plasma generation microwave circuit, which obtains first and second calculation outputs proportional to the product of the sine sin θ and the cosine cos θ, and thereby drives a stub to match the load, a first tuning target value setter that sets an arbitrary reflection coefficient |Γ|, sine sinθ, and an arbitrary reflection coefficient |Γ
｜, a second tuning target value setting device for setting the cosine cos θ, two stubs are set up in the longitudinal direction of the waveguide 1 at an interval of 1/4 of the average tube wavelength, and one and the other are activated. two sets of these composite stubs are arranged at an interval of 1/8 of the average tube wavelength, and the difference output between the first calculation output and the first tuning target value setting device output and the second calculation An automatic tuning device for a plasma generation microwave circuit that drives first and second composite stubs to tune to a tuning target value based on a difference output between the output and the output of a second tuning target value setter.