JPH04303599A - Diagnostic method of plsme - Google Patents

Abstract

PURPOSE:To carry out quantative diagnosis magnetron discharge plasma by meauring the illumination intensity of plasma from two waay directions, and reconstructing the illimination distribution obser inside observation surface by the CT technology method with the use of measured data. CONSTITUTION:Indide a vacuum chamber 1, parallel electrode plates 2, 3 are arranged and observation apertures 4, 5 are aet in normanl direction to each other. A mirror box 12 capable of rotating is disposed at positions corresponding to the apertures 4, 5 to decide an optical path and connected to obtical fiber. The box 12 is on a stage capable of moving in the lingitudinal direction and in parallel with the surface of apertures 4, 5, the stage is used for CT measurement. In order to grasp the space dischage structure of magnetron discharge plasma, the three dimensional distribution of illumination with the application of CT technology method os measured from the scanning measurement from two observation positions 4, 5. As a result, it is found that the measurement is much concerned with the distrubution of magnetic fieled in association with magnetron. It is thus possible to perform plasma control so as to improve the yielding rate of products.

Description

[Detailed description of the invention]

[Purpose of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to discharge plasma, which has been widely used industrially in recent years.

0002

[Prior Art] Dry process is a process technology that uses so-called weakly ionized plasma in a plasma where the net positive and negative charges are balanced, and the field benefiting from this is semiconductor processing. There are many technologies, and their applications are diverse, such as plasma CVD and microfabrication technology. One way to generate weakly ionized plasma is to
There is a plasma that uses a magnetron, which has an electrode structure capable of applying a magnetic field, or magnetron discharge plasma.

Magnetron discharge plasma is also generated in a vacuum chamber, and it is difficult to observe it without changing its internal state. Emission spectroscopy is a well-known method for observing the light emitted when a particle is dissolved, and is a method for measuring the distribution of excited particles.

[0004] Therefore, emission spectrometry of weakly ionized plasma is performed at only one arbitrary point in a plane parallel to the parallel polar plates, and in most cases, a condensing optical system consisting of an optical lens or a condenser lens is used. Using this system, we have only been able to investigate the characteristics of plasma through measurements with relatively low spatial resolution, such as focusing on the center axis of parallel polar plates.

[0005] In addition, process plasmas are required to treat large areas, and for diagnosis, there are two types of diagnosis: one that uses plasma light emission but does not require positional resolution, and one that requires plasma positional resolution. Furthermore, if
A lens focused on an observation point within the plasma may be used.

[0006]

[Problem to be solved by the invention] Even in magnetron discharge plasma, which is widely used industrially, it is essential to diagnose the plasma, confirm the discharge structure, and theoretically analyze it. It is desirable to do so.

[0007] The present invention has been made in view of the above circumstances, and is particularly aimed at quantitatively diagnosing magnetron discharge plasma.

[Configuration of the invention]

[0009]

[Means for Solving the Problems] The first feature of the plasma diagnosis method according to the present invention is that the emission intensity of the plasma is measured from two directions within the observation plane, and the measured data is computed using CT.
The aim is to reconstruct the luminescence distribution within the observation plane using techniques.

[0010] The second aspect of the invention is characterized in that it is equipped with a measuring device that is equipped with a detector that detects only the light emitted within a linear optical path with a minute cross-sectional area. The fourth aspect of the present invention is to add a measurement device that measures plasma over a wide range within the observation plane using a viewing window attached to the chamber by moving the angle and position of the detector within the same plane. The feature of the invention is that the detected light is divided into specific wavelengths and the emission of excited species or ion species of a specific gas within the observation plane of the plasma is measured.

[0011]

[Operation] Generally, in discharge plasma, a non-uniform state is formed, so it is essential to investigate the three-dimensional spatial structure. Therefore, in the present invention, plasma is generated by supplying direct current or radio frequency (RF) power to electrodes provided in a vacuum chamber. Furthermore, CT (Computer Tom) is used to effectively measure the spatial structure of non-uniform plasma.
The three-dimensional spatial distribution of luminescence was measured by performing spatial spectroscopic measurements using the following technology: This method is
The feature is that the light emission state over the entire discharge space can be observed by scan measurement from a few observation positions.

0012

Embodiments An embodiment of the present invention will be described with reference to FIGS. 1 to 6. That is, a vacuum device that performs the discharge essential to the method of the present invention, which uses magnetron discharge plasma as an example, has parallel pole plates 2 and 3 each having a diameter of 100 mm arranged in a vacuum chamber 1, as shown in FIGS. 1 and 2. Observation windows 4 and 5 are provided on an extension line of this substantially central line and in directions perpendicular to each other, and a looper (not shown) is installed inside for the purpose of preventing Al from accumulating.

Among the parallel plates 2 and 3, the one located above and functioning as an anode electrode is grounded and electrically connected to a DC power supply 6 that applies a negative bias to the lower one. Further, on the back side of the parallel pole plate 3 that functions as a cathode, permanent magnets, that is, magnetic field devices 7... are arranged radially with the N pole facing outward to generate a magnetic field between the electrodes. Indicated. In FIG. 1, only a simplified cross-sectional shape of the magnetic field device 7 is shown.

In FIG. 3, the horizontal axis indicates the radial direction in centimeters, and the vertical axis indicates the axial direction in centimeters.
It is clear that the lines of magnetic force are concentrated toward the center.

The vacuum chamber 1 includes a decompression mechanism 8 (not shown).
An exhaust hole 9 connected to a pressure reducing mechanism such as a turbo molecular pump is installed.
Exhaust route 1 to the rotary pump, etc.
0, of course, a vacuum measuring instrument such as a Pirani gauge (Pirani gauge) is used.
aniGauge) and ionization vacuum gauge (Ionizatio)
nVacuum Gauge) and check the degree of vacuum in the vacuum chamber 1.

A gas such as Ar gas whose flow rate is adjusted by a mass flow controller (not shown) is introduced into the vacuum chamber 1 from gas inlet ports 11 and 11 installed above the exhaust hole 8.
flows into the vacuum chamber 1, and the degree of vacuum within the vacuum chamber 1, which is maintained at several tens of mTorr or less, is measured using a Pirani gauge or the like. Further, the ionization vacuum gauge was used to measure relatively low pressures, such as 1 Pa or less.

Vacuum chamber 1 corresponding to observation windows 4 and 5
On the outside, rotatable mirror boxes 12, 12 are mounted at right angles to each other to determine the optical path, which is guided to an optical fiber 13. The optical fiber 13 enters the spectrometer through an exit fiber collimator attached to the entrance of the spectrometer to select a wavelength. Light with the selected wavelength enters the photomultiplier shown in FIG. 1, converts it into an electrical signal, and becomes a photon signal. Next, signal components are extracted by an amplifier and a discriminator, and then counted by a photo counter and a multi-channel analyzer. Further, a multi-channel analyzer is a counter memory having more than 1,000 memory channels, and each channel functions as a counter of about 1 million.

The mirror box 12 includes a rotatable mirror 14, a collimator (not shown), and a second mirror 15.
A pinhole 16 is formed in the wall surface of the mirror box 12 located between the two, and is connected to the optical fiber 11 via a connector for the optical fiber 11 (not shown). That is, with few observations,
It has become possible to measure information over a wide space with a small number of measurements. Further, the observation optical path can be scanned in a fan shape by controlling the angle of the collimator, and almost the entire discharge light emission region can be observed.

Further, the size of the optical path is limited by the pinhole 16 to determine the resolution ability of the approximate position, and as described above, the light from the pinhole 16 is guided to the optical fiber 13 by the collimator. Furthermore,
The collimator can condense parallel light, improving the resolution of the device, and the diameter of the pinhole 16 and that of the core of the optical fiber 13 are the same.

A control unit (not shown), that is, a galvano scanner using a DC drive system is attached to the collimator. This is usually used to control the pointing of semiconductor laser spots, such as in laser printer heads, and is capable of very high precision and high-speed control, and can even follow sine wave inputs of several KHz. ing. This is because the scanner is not controlled by a pulse motor, and the deflection angle is proportional to the analog input voltage. Therefore, in order to operate the analog DC voltage, a D/A converter was used for actual driving of this scanner, and the deflection angle was controlled by a personal computer.

The 12-bit digital signal output from the computer is converted from -10 to + by the D/A converter.
It is output as a 10V analog signal, and this output is sent to the EXT. of the galvano scanner control unit. It is input to the IN terminal and converted to a ±5V signal for the galvano scanner. The scanner is proportional to this voltage ±25°
It can be swung to an angle of .

The control unit is also equipped with a terminal for monitoring the deflection angle of the scanner, and outputs a voltage of ±5 V depending on the actual deflection angle of the mirror, and the voltage is measured by a digital voltmeter.

The mirror box 12 is installed on a Z stage that can move vertically and an X stage that can move parallel to the planes of the observation windows 4 and 5, and the Z stage has parallel polar plates 2 and 3.
The X stage is used for CT measurement.

Originally, it is desirable to measure the emission of the magnetron discharge plasma to be observed while moving the optical path in parallel, but from the observation windows 4 and 5 of the vacuum chamber 1,
It is not possible to observe all the light emission between the parallel polar plates 2 and 3. Therefore, by controlling the angle of the collimator and performing fan-shaped measurements, the observation range can be expanded to about the size of the parallel polar plates 2 and 3. However, since the mirror box 12 is located away from the center of the parallel polar plates 2 and 3, the observation angle becomes very small, making control difficult. For this purpose, if we install an 2.3 Logical observation points can be found near the center (see Figure 5)
. Further, the X stage also allows slight movement of the mirror box 12 position in the lateral direction.

Before starting measurements using the vacuum apparatus described above, the following preliminary experiments were conducted. First, in the observation system, many control operations are linked by a computer, so the question of which devices to control and in what order is a problem, so we conducted an automatic control test and an interlock test to confirm the operation.

After calibrating the mirror control unit to determine whether the angle control characteristics of the mirror had linearity sufficient to withstand the measurement requirements, measurements were made using a laser or a mercury lamp to determine whether the measurement system had the characteristics as designed. Furthermore, we performed emission spectroscopy, longitudinal position-resolved spectroscopy, and CT spectroscopy.
The ART algorithm was verified.

The ART algorithm is a basic reconstruction image calculation method, and calculation results have been shown in various research results. However, there is no example in which reconstruction is performed from observation data in only two directions as in the present invention. Therefore, it is necessary to verify to what extent reconstruction using the ART algorithm is possible from data in only two directions. Here, for each of the stepped distribution, linear distribution, and Gaussian distribution that rises like a δ function, we create an axis-symmetric two-variable function of a rotating body in which the distance from an appropriate axis is taken as a variable, and then we create a virtual axisymmetric two-variable function. We obtained the data measured in 2018 and verified whether it is possible to reconstruct the original distribution using the ART algorithm. The results are shown in FIG. 5, where the Gaussian distribution is shown as a solid line and the reconstructed values are shown as a plot.

The test data has a Gaussian distribution with a variance of 10, and is ring-shaped with a peak located 30 mm away from the center.

By the way, there are countless charged particles in a plasma, and the properties of the plasma are determined by their movement and interaction. Lorentz force is applied. Next, considering the motion in a space where an electric field and a magnetic field are applied, positively charged particles are accelerated downward by the electric field and rotate with a gradually larger Lamor radius because they are influenced by the magnetic field. The charged particles whose direction of motion has been changed in this way now perform a swirling motion while being decelerated upward by the electric field. As a result, the particles move in the ExB direction while rotating, and this is a phenomenon called ExB drift.

[0030] It is expressed as vd =c(E×B)/B2,
Although it does not depend on the mass or charge of the particles, in plasma it is necessary to consider the effects of collisions in addition to this. However, in discharges at very low pressures such as magnetron discharge plasma, the effect of E×B drift is large, and many particles are trapped in areas where there are many parallel electric components of the magnetic field, and the parallel polar plate In the mold, it can be expected that a ring-shaped region with a high density of charged particles will be formed.

In the method of the present invention using the emission CT technique, a reconstruction technique is applied without going through the so-called Radon transform theory. As a reconstruction method, Al
Gebraic Reconstruction
Technique (ART) method was used. First, after considering a coordinate axis with the origin at the center of the cathode on a plane zmm from the cathode, the area where light emission is observed is divided into a mesh of fine pixels. An initial value f (xiyi) is set for each pixel pixel (xiyi). In addition, the observation optical path is set as follows: ρ is the distance from the origin of the mirror box, and β is the angle from the X axis (β = 0, π/2).

[0032]

It is expressed as [Equation 1]. The measurement data is the line integral of the emission intensity along this optical path,

[0033]

[Math 2]

[0034]

[Math 3]

[0035]

[Math 4]

[0036]

[Equation 5] Here, in (3), each optical path is pixel(xi,
It is a weight proportional to the length across yi). This (4)
and the measured data R(β, θ) are distributed again along the optical path to create (5). By repeating this process, the emission intensity distribution on this cross section can be determined. By determining profiles on a plurality of cross sections using z as a parameter, a three-dimensional light emission distribution in space can be determined.

CT spectroscopy was performed under the following conditions in a vacuum apparatus in which various parts were installed in the vacuum chamber 1. Pure Ar1 (spectrum symbol of atomic beam) was introduced into the vacuum chamber 1 at a flow rate of 3 sccm,
The pressure of the vacuum chamber 1 is set to 23 mTorr.

[0038] Also, the voltage is 254V, the current is 20 to 30mA
, Power Density1.
As a result of lowering the power to 4 mW/cm to avoid deposition on the observation windows 4 and 5 as much as possible, the curve shown in FIG. 6 was obtained.

As is clear from FIG. 3, a static magnetic field is applied between the parallel pole plates 2 and 3, and the parallel pole plate 2
, 3 at a position of 32 to 34 mm from the center of parallel plates 2 and 3.
A magnetic field is applied almost parallel to , and the E×B drift is maximum here.

FIG. 5 shows Ar1 (λ=
419.8) was clarified. That is, the emission is z=8m from parallel plates 2 and 3.
There was a peak at position m, and CT spectroscopy was performed using the reconstruction method at points A, B, C, and D shown in FIG. 5, and an emission image of ArI (not shown) was taken. This video has ε
An emission image corresponding to the spatial density distribution of high-energy electrons of 14.57 eV was obtained, and a ring-shaped ring like an emission peak was formed at a position approximately 34 mm from the center of the parallel polar plate, and this peak was in the electromagnetic field region. matches.

[0041] In this example, an example was shown in which magnetron discharge plasma is generated by applying a magnetic field and an electric field using parallel polar plates. It should be noted that the method of the present invention can be used for the diagnosis of. It is further noted that plasma generation can be generated by a single electrode and chamber rather than by parallel plates.

Furthermore, other means that can be applied in addition to the means described in the above embodiments are shown below. First, as a means for observing light emission in plasma, excited light emission by collisions of electrons and ions and light emission by laser can be used.

Methods for measuring luminescence include measuring the intensity of light from one direction within the surface to be observed, and measuring the total amount of luminescence within the measurement path.Scanning means include mirrors, fibers, and chambers. be.

The grating method, prism method, and filter method can be applied as the spectroscopic method, and the means for converting optical signals into electrical signals include photomultiply, photocell, and CC.
D is applicable.

Furthermore, a photo counter and an A/D converter can be used for conversion into a digital signal.

Last but not least, the reconfiguration is A in the above embodiment.
FBP that uses Radon transformation although it uses rt method
(Filtered Back Project
It should be noted that the on) method can also be used.

[0047]

[Effects of the Invention] The three-dimensional spatial distribution in magnetron discharge plasma was measured for the first time using CT technology, and it was found that it is closely related to the distribution of the magnetic field accompanying the magnetron. In various fields including the semiconductor process field,
Plasma control becomes possible using theoretical rather than empirical parameters, which greatly contributes to improving product yield.

[Brief explanation of drawings]

FIG. 1 is a sectional view showing the main parts of a vacuum apparatus used to carry out the method of the present invention.

FIG. 2 is an enlarged cross-sectional view of a part of the vacuum device.

FIG. 3 is a diagram showing a magnetic field line distribution of a magnetic field device attached to a vacuum device.

FIG. 4 is a diagram showing the relationship between an actual observation point and a logical observation origin in measurement using a vacuum device.

FIG. 5 is a diagram clarifying the distribution of reconstruction in the method of the present invention.

FIG. 6 is a curve diagram showing the effect of the method of the present invention, with the emission integration line and the distance from the cathode taken as both axes, and showing the relationship between the two.

[Explanation of symbols]

1: Vacuum chamber, 2, 3: Parallel plate, 4, 5: observation window, 6: DC power supply, 7: Permanent magnet, 8: Exhaust hole, 9: Decompression mechanism, 10: Exhaust route, 11: Gas introduction hole, 12: Mirror box, 13: Optical fiber, 14: Mirror 15: Second mirror, 16: Pinhole.

Claims (4)

[Claims] [Claim 1] The emission intensity of plasma is measured at 2 points in the observation plane.
A plasma diagnosis method that consists of measuring from the direction and reconstructing the luminescence distribution within the observation plane from the measured data using CT techniques. 2. A method for diagnosing plasma, comprising a measuring device equipped with a detector that detects only light emitted within a linear optical path with a minute cross-sectional area. 3. Addition of a measuring device that measures the plasma over a wide range within the observation plane using a viewing window attached to the chamber by moving the angle and position of the detector within the same plane as the plasma observation plane. Characteristic plasma diagnostic method 4. A method for diagnosing plasma, which comprises dividing the detected light into specific wavelengths and measuring the emission of excited species or ion species of a specific gas within the observation plane of the plasma.