JPH04256845A - Plasma measuring method - Google Patents

Abstract

PURPOSE:To obtain a highly accurate plasma measuring method wherein the disturbance of plasma is suppressed to the minimum degree in the measurement of the plasma for a semiconductor process. CONSTITUTION:Parallel paired cables 40 are penetrated through plasma 40. A microwave is propagated through the parallel paired cables 40 from a microwave source 1 through a coaxial cable 43, an isolator 2 and a matching device 31. The microwave through the matching device 32 is detected with a microwave detector 3. The plasma parameters are determined based on the changes in phase and in strength of the microwave for measurement affected by the plasma.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a plasma measurement method using microwaves used in the manufacturing process of various devices such as semiconductor integrated circuits.

0002

[Prior Art] In the manufacturing process of various devices such as semiconductor integrated circuits, RF plasma using high frequency discharge and electron cyclotron resonance using microwave excitation are used as plasmas for thin film formation and etching.
CR plasma is used. In particular, it has been revealed that ECR plasma has the characteristics of low gas pressure, high ionization rate, and high activity, and that it can exhibit excellent processing characteristics by selecting the introduced gas and controlling the ion energy (Japanese Patent Application Laid-Open No. 1983-56).
155535, Japanese Patent Publication No. 58-13626), it is used for etching fine patterns and forming thin films.

FIG. 6 shows the basic configuration of a plasma processing apparatus using ECR plasma. In the figure, 61 is a plasma generation chamber, 62 is a sample chamber, and 63 is a microwave introduction window provided in the upper part of the plasma generation chamber 61, which is made of, for example, a quartz glass plate. 64 is a rectangular waveguide for introducing microwaves. 69 is a plasma extraction window for guiding the plasma generated in the plasma generation chamber 61 to the sample 71 placed on the sample stage 70. A magnetic coil 72 is used to generate a magnetic field strength that satisfies electron cyclotron resonance conditions in a suitable region inside the plasma generation chamber 61, and is also used to form a diverging magnetic field to form a plasma flow 68. 65 is a cooling water system for cooling the plasma generation chamber. 66 and 67 are gas introduction systems, which are used alone or in combination depending on the processing purpose. Necessary gases are introduced from the gas introduction systems 66 and 67, and a magnetic field (875 Gauss when the microwave frequency is 2.45 GHz) that satisfies the electron cyclotron resonance condition is applied to a suitable region inside the plasma generation chamber 61 by the magnetic coil 72. The microwave is generated and introduced into the plasma generation chamber 61 through the microwave introduction window 63 to cause electron cyclotron resonance. Plasma is generated when electrons accelerated by electron cyclotron resonance collide with gas and are ionized. The generated plasma forms a plasma flow 68 due to the divergent magnetic field of the magnetic coil 72 and reaches the sample 71 .

[0004] Semiconductor integrated circuits are becoming increasingly highly integrated and miniaturized, and the diameter of samples to be processed is also becoming larger. More precise plasma control is required. To this end, it is essential to have a thorough understanding of plasma generation and transport based on accurate plasma measurements.

Langmuir probe method and microwave interferometry are known as low-temperature plasma measurement methods used in semiconductor processes. In the Langmuir probe method, a metal electrode is inserted into the plasma, a voltage is applied, and the electron current flowing into the probe is measured.The voltage-current characteristics are analyzed assuming the electron velocity distribution, and the plasma electron temperature,
This is a method for determining plasma parameters such as electron density (Nobuyuki Tsutsui, Basic Plasma Engineering, Chapter 3 (Rokaku Uchida, 1986)). This method has the advantage of high positional resolution because measurements can be made at the location where the probe is placed. However, plasma disturbance due to insertion is unavoidable, making accurate measurement difficult. In addition, when calculating the electron temperature and electron density, it is necessary to assume the electron velocity distribution and the sheath shape of the electrode surface, so especially when measuring non-equilibrium plasma such as semiconductor process plasma, The error becomes larger. Furthermore, when there is a magnetic field as in ECR plasma, the conditions for electron inflow are not constant, leading to large errors and making it difficult to use. From a practical standpoint, contamination of the plasma device due to the insertion of metal electrodes is also a problem.

Microwave interferometry uses microwaves with a higher frequency than the plasma frequency, detects the attenuation and phase shift of the microwaves caused by the plasma, and determines the refractive index of the plasma. This is a method of determining plasma parameters such as plasma density and electron temperature by utilizing the fact that the rate is determined by the plasma density and the electron collision angular frequency ν, as will be described later. Figure 7
The basic configuration of plasma measurement using microwave interferometry is shown in {Nobuyuki Tsutsui, Basic Plasma Engineering, Chapter 4 (Uchida Rokaku, 1986)}. In the figure, 1 is a microwave source, such as a klystron oscillator. 21 and 22 are magic tees, 23 and 24 are horn antennas, 12 is a variable attenuator, 13 is a variable phase shifter, and 3 is a microwave detector. Microwaves emitted from the microwave source 1 are guided to the magic T21 by a waveguide and divided into two. One side is sent into the plasma via the horn antenna 23, propagates through the plasma, is collected by the horn antenna 24 on the opposite side, and reaches the magic T 22. The other one passes through the variable attenuator 12 and the variable phase shifter 13 and reaches the magic T22. The microwave detector 3 detects the intensity of the microwaves synthesized by the magic T22. A variable attenuator 12 and a variable phase shifter 13 are installed in advance so that the intensity of microwaves reaching the microwave detector 3 becomes zero when there is no plasma.
The variable attenuator 12 and the variable phase shifter 13 are adjusted so that the intensity of microwaves reaching the microwave detector 3 becomes zero even when plasma is present. The attenuation and phase shift of the microwave due to the plasma are calculated from the amount of adjustment, and the plasma parameters are determined.

This method is used to measure large volume plasma such as nuclear fusion plasma, and has the advantage that the combination of microwave and horn antenna causes little disturbance to the plasma. However, the microwaves that exit the horn antenna spread and propagate with an angular distribution determined by the horn antenna, so if the plasma is in contact with a wall, such as in a small volume plasma for semiconductor processing, some of the microwaves may be It reaches the wall and creates a complex reflection. Furthermore, when the plasma is far from the wall, some of the microwaves are reflected at the plasma boundary and propagate outside the plasma, or cause complex reflections on the walls of the container. All of these factors are added to the microwave detected by the horn antenna, so the microwave attenuation and phase shift include attenuation due to plasma, phase shift, and attenuation due to the reflections listed above. It also includes a phase shift, making accurate measurement impossible. Furthermore, since two horn antennas must be installed facing each other, the locations where measurements can be made with the plasma device are limited. Furthermore, since a waveguide is used, the frequency of the measurement microwave cannot be changed significantly.

Furthermore, when there is a magnetic field, such as in ECR plasma, the refractive index of the plasma with respect to microwaves is determined by the propagation direction of the microwaves, the strength of the magnetic field, the direction of the microwave electric field, etc. However, in microwave interferometry, since microwaves are introduced into the plasma using a horn antenna, the direction of microwave propagation in the plasma and the direction of the microwave electric field cannot be determined, making accurate measurement impossible. It is. Thus, conventional microwave interferometry cannot accurately measure relatively small volumes of plasma used in semiconductor processes.

[0009]

With the miniaturization of semiconductor integrated circuits, the importance of ECR plasma devices, which have the characteristics of low gas pressure, high ionization rate, and high activity, is increasing. Further improving the performance of ECR plasma devices requires highly accurate plasma control based on accurate understanding of ECR plasma generation and transport, and for this purpose, highly accurate plasma measurement methods are required. However, as described above, with conventional measurement methods, it is theoretically impossible to accurately measure plasma in a magnetic field, such as a relatively small volume plasma used in a semiconductor process or an ECR plasma.

The present invention was proposed in order to solve the above-mentioned drawbacks of the conventional method in principle, and its purpose is to use relatively small volume plasmas used in semiconductor processes, especially ECR plasmas, etc. The objective is to provide a highly accurate measurement method for plasma in a magnetic field.

[0011]

[Means for Solving the Problems] In order to achieve the above object, the present invention injects measurement microwaves into the plasma to be measured, and detects changes in the phase and intensity of the measurement microwaves received from the plasma. The gist of the invention is a plasma measurement method for determining parameters, characterized in that the measurement microwave is propagated through the plasma to be measured using a parallel pair of cables.

0012

[Operation] The present invention makes it possible to limit the microwave propagation mode, propagation path, and region of electromagnetic field in the plasma by propagating measurement microwaves through the plasma using parallel pair cables. Main characteristics. As a result, in the measurement of plasma for semiconductor processing, disturbance to the plasma can be minimized and plasma can be measured with high precision.

[0013]

[Example] Next, an example of the present invention will be described. Note that the embodiments are merely illustrative, and it goes without saying that various changes and improvements can be made without departing from the spirit of the invention.

(Embodiment 1) FIG. 1 shows a first embodiment of the present invention. In the figure, 1 is a frequency variable microwave source;
40 is a parallel pair cable, 3 is a microwave detector such as a spectrum analyzer, 43 is a coaxial cable, 2 is an isolator for preventing damage to the microwave source 1 due to reflected waves, 31 and 32 are a coaxial cable 43 and a parallel pair cable 40 62 is a sample chamber, 68 is a plasma flow, 69 is a plasma extraction window,
70 indicates a sample stage.

The measuring microwaves emitted from the microwave source 1 propagate through a coaxial cable 43, pass through an isolator 2 and a matching box 31, and are guided into the plasma by a parallel pair cable 40. The microwave propagated through the parallel pair cable 40 is
The signal is guided through the matching box 32 to the microwave detector 3 via a coaxial cable 43 and detected. Plasma measurement with this configuration is performed in the following steps. First, the transmitted intensity of microwaves in the absence of plasma is measured by changing the frequency of the measurement microwave. Next, we will measure the transmitted intensity of microwaves when plasma is present. The two are compared to determine the amount of attenuation caused by the plasma, and plasma parameters such as plasma density and electron temperature are determined.

FIG. 2A shows a cross section of the parallel pair cable 40. 80 is a transmission line, 84 is an insulator, and 85 is a state of electric lines of force of a microwave electric field. Parallel pair cable 40 is 2
It consists of parallel transmission lines 80 that are close to each other, and each transmission line is covered with an insulator 84 such as quartz to prevent direct contact with the plasma.
covered with. The basic propagation mode in such a configuration is T
It is an EM wave. In the experiment, the outer diameter of the insulator 84 was 1.2
A mm quartz tube was used. Since the cross-sectional area of the parallel pair cable 40 can be made small in this way, disturbance to the plasma caused by insertion into the plasma can be suppressed to a small level. Further, since the microwave is propagated in the plasma by the parallel pair cable 40, it is possible to eliminate influences such as reflection on walls and wraparound of the plasma due to spread of the microwave in the plasma.

A configuration using three transmission lines as shown in FIG. 2(B) as the parallel pair cables is also conceivable. In this case, if the transmission line 81 and the transmission line 82 are connected to have the same potential, the microwave electric field will be as shown in 85, as shown in Fig. 2(A).
Compared to the case of parallel pair cables shown in Figure 1, the area where the plasma and microwave electric field interact increases, improving measurement accuracy. By shielding a portion of the parallel pair cables shown in FIGS. 2(A) and 2(B) from the plasma, the positional resolution in the direction along the parallel pair cables can be increased.

Generally, the fundamental mode of microwaves propagating through a parallel pair cable is a TEM wave, and its propagation characteristics are the same as those of a plane wave. Therefore, in the absence of a magnetic field, the analysis results of the propagation characteristics of microwaves in plasma using plane wave approximation can be applied to the propagation characteristics of the fundamental mode of microwaves propagating in plasma along the parallel pair cables.

An explanation will be given of the analysis results regarding the propagation of microwaves in plasma using plane wave approximation. The refractive index n of plasma for microwaves in the absence of a magnetic field is
When the collision angular frequency ν of the electron is sufficiently small compared to the angular frequency ω of the microwave and can be ignored,

[0020]

[Math 1]

It is known that Here, ω is the angular frequency of the microwave, ωp is the plasma angular frequency, and if the plasma density is np, the electron charge is -e, and the electron mass is me, then ωp = (e2 np / εo
me ) 1/2. The electric field E of a microwave propagating in a medium with a refractive index n is E~exp{-i(ωt-nko z)} when the propagation direction is the z direction.
It can be written as (2). However, ko
is the wave number in vacuum, and if the wavelength in vacuum is λ, then ko =
2π/λ. Therefore, from equations (1) and (2), the plasma angular frequency ωp becomes the cutoff, and when the microwave angular frequency ω is smaller than the plasma angular frequency ωp, the refractive index n becomes an imaginary number, so the microwave The intensity of the microwave reflected by the plasma and transmitted through the plasma is attenuated. When the microwave angular frequency ω is larger than the plasma angular frequency ωp, the refractive index n becomes a real number,
Microwaves can propagate in plasma.

The refractive index of plasma for microwaves in the presence of a magnetic field varies depending on the propagation direction of the microwaves and the direction of the microwave electric field. Table 1 shows the refractive index of plasma for microwaves in the case of plane waves. In the case of microwaves (normal waves) where the propagation direction of the microwave is perpendicular to the magnetic field and the electric field is parallel to the magnetic field, the refractive index is the same as when there is no magnetic field.

[0023]

[Table 1]

Since the electric field of the microwave propagating through the parallel pair cable 40 is strong in the direction that connects the two transmission lines as shown in FIG. By connecting the two transmission lines parallel to the magnetic field as shown in Figure 2 (A), the electric field of the microwave and the external magnetic field can be made almost parallel, making it possible to arrange the microwave in the same way as for normal waves. . At this time, the refractive index of the plasma becomes the same as equation (1), and the plasma angular frequency ωp becomes the cutoff. Due to the peculiarity of plasma when there is a magnetic field, waves other than normal waves are the fundamental mode of parallel pair cables (TEM waves).
cannot be compatible with

Therefore, if the propagation direction of the measurement microwave is made perpendicular to the magnetic field using a parallel pair cable, and the electric field of the measurement microwave is made parallel to the external magnetic field, the plasma angular frequency ωp can be adjusted regardless of the presence or absence of the magnetic field. is the cutoff. With the configuration shown in FIG. 1, the plasma density can be determined by measuring the transmission characteristics and directly determining the cutoff frequency. In this example, the oscillation frequency of the microwave oscillator is 2 to 18 GHz, so the range of measurable electron density is 5 x 1010 cm-3 to 4 x 1012 cm-3, but by changing the oscillation frequency, The range of measurable electron density can be changed.

FIG. 3 shows the measurement results of ECR plasma according to this example. The vertical axis of Figure 3(A) is the intensity of the transmitted microwave (in dB), and the horizontal axis is the frequency of the microwave (GH
z unit). □ indicates the intensity of the transmitted microwave when there is no plasma, and ■ indicates the intensity of the transmitted microwave when there is plasma. In ECR plasma, the gas pressure is low, so the condition ν<<ω holds, so the refractive index of the plasma is (1
), and ωp is the cutoff.

The cutoff frequency is determined as follows. Microwave attenuation logarithm ΔI (
db) (=without plasma - with plasma), where d is the thickness of the plasma and c is the speed of light.

[0028]

[Math 2]

It can be written as [0029]. From equation (3), (ΔI)2 is a linear function of ω2, and the frequency at which (ΔI)2 = 0 is ω
It turns out that p. Therefore, the plasma angular frequency ω
For p, ■ Calculate ΔI from the measured data, ■ Indicate (ΔI) on the vertical axis.
)2, ω2 is plotted on the horizontal axis, and ■ω<ωp
Determine the equation of the straight line using the method of least squares from the data,
■It can be determined by finding the intersection with the horizontal axis.

The above is a case where there is no reflection at the plasma boundary, but in actual measurement, the measurement microwave is reflected at the plasma boundary due to the difference in refractive index between the plasma and the vacuum. The case where there is reflection will be explained below.

The amplitude reflectance and amplitude transmittance at the boundary when the microwave enters the plasma are ΓA, TA, and the amplitude reflectance and amplitude transmittance when the microwave exits the plasma are ΓB, TB.
Then,

[0032]

[Math 3]

It can be written as [0033]. However, n is the refractive index of plasma. Therefore, if the thickness of the plasma is d, the amplitude A of the microwave transmitted through the plasma is:

[0034]

[Math 4]

[0035] A0 is the amplitude of the incident microwave, and if ω<ωp, the refractive index of the plasma is

0
036]

[Math 5]

Therefore, the power transmittance of the measurement microwave is

[0038]

[Math 6]

[0039] If ωp d/c >> 1, then
The second term in the denominator is larger than the other terms,

0040
]

[Math 7]

[0041] Therefore, the attenuation logarithm ΔI of the microwave is

[0042]

[Math. 8]

[0043] When ωp d/c >> 1,

004
4]

[Math. 9]

##EQU1## which is consistent with equation (3). Therefore, the thickness d of the plasma irradiated onto the parallel pair cable is

004
6]

[Math. 10]

In the case of ω, use equation (3) to accurately calculate ω
p can be determined. When the plasma frequency is 2 GHz or more (5×10 10 cm −3 or more when converted to plasma density), d may be set to 5 cm or more. Note that if this approximation cannot be used, ωp can be determined using exact equation (3)' without using equation (3).

FIG. 3(B) shows the results of the above data processing of ■ to ■ based on the data of FIG. 3(A), and the vertical axis is (Δ
I)2, the horizontal axis is ω2, and □ is the measurement point. In addition, the straight line was determined using the least squares method from data below 3 GHz. From this figure, the cutoff frequency is 3.7GHz,
The plasma density can be determined to be 1.7 x 1011 cm-3. When the gas pressure is high and electron collisions cannot be ignored, the refractive index n of the plasma with respect to microwaves is expressed as follows, where ν is the electron collision angular frequency.

[0049]

[Math. 11]

Therefore, in the measured transmission characteristics of the microwave, the cutoff is blunted, and microwave attenuation occurs even where the angular frequency ω of the microwave is greater than the plasma angular frequency ωp. Therefore, the collision angular frequency ν of the electron can be determined from the rounding near the cutoff or the amount of attenuation of the microwave when ω>ωp. Furthermore, the electron temperature can be estimated from the gas pressure, the electron collision angular frequency ν, etc.

(Embodiment 2) FIG. 4 shows a second embodiment of the present invention. Microwaves emitted from a frequency variable microwave source 1 pass through a coaxial cable 43, pass through an isolator 2, a coupler 10, a matching device 31, and are introduced into the plasma by a parallel pair cable 41. 42 indicates a dummy resistor. The microwave reflected by the plasma 68 is transmitted to the matching box 3.
1, is branched by a coupler 10, and reaches a microwave detector 3. The parallel pair cable 41 has the same configuration as that used in the first embodiment except that one end of the parallel pair cable 40 used in the first embodiment is terminated with a dummy resistor 42 corresponding to the impedance of the parallel pair cable.

The refractive index of plasma for microwaves is (
1) Since it is expressed by the formula, when the angular frequency ω of the microwave is larger than the plasma angular frequency ωp, the refractive index becomes a real number, and the microwave propagates through the plasma, reaches the dummy resistor 42, and is absorbed. . Conversely, when the angular frequency ω of the microwave is smaller than the plasma angular frequency ωp, the refractive index becomes an imaginary number, so the intensity of the microwave reflected by the plasma and detected by the microwave detector 3 increases. Therefore, with the configuration shown in FIG. 4, the plasma frequency can be measured from the frequency dependence of the reflected wave, and the plasma density can be determined. In this case, since there is nothing connected to the terminal end of the parallel pair cable, there is no restriction on the installation location in the plasma, and there is an advantage that the degree of freedom in measurement is very large. As a modification of this embodiment, reflection characteristics can be specified in combination with the measurement of transmission characteristics in Example 1 to improve measurement accuracy.

(Embodiment 3) FIG. 5 shows a third embodiment of the present invention. In the figure, 1 is a microwave source, and the frequency of the microwave is selected to be higher than the plasma frequency. 43 indicates a coaxial cable, and 40 indicates a parallel pair cable. 2
is an isolator, 10 and 11 are couplers, 31 and 32 are matching devices, 3 is a microwave detector, 12 is a variable attenuator, 13
indicates a variable phase shifter. The microwaves leaving the microwave source 1 reach the coupler 10 via the isolator 2 and are divided into two. One of them passes through the matching device 31 and propagates in the plasma by the parallel pair cable 40, and then reaches the coupler 11. The other reaches the coupler 11 via a variable attenuator 12 and a variable phase shifter 13. The intensity of the microwaves combined by the coupler 11 is detected by the microwave detector 3. The variable attenuator 12 and the variable phase shifter 13 are adjusted so that the intensity of the microwave detected by the microwave detector 3 becomes zero when there is no plasma, and it also becomes zero when there is plasma. Then, the phase difference and attenuation due to the presence or absence of plasma are determined from the adjusted amount. From the phase difference and attenuation amount (1
) or (4), the plasma angular frequency ωp
, calculate the electron collision angular frequency ν, and determine the plasma parameters from them.

Unlike the method using the horn antennas 23 and 24 shown in FIG. 7, the propagation path of the microwave in the plasma and the direction of the electric field of the microwave are limited by the parallel pair cable 40, so the area is small. Accurate measurement of plasma for semiconductor processing is possible.

[0055]

[Effects of the Invention] As described above, the present invention propagates measurement microwaves in plasma using parallel pair cables with a small cross section, so the microwave propagation path is limited and the microwave electric field is Because the direction is defined, it has the advantage of minimizing disturbance to the plasma and achieving highly accurate plasma measurements when measuring plasma for semiconductor processing, including plasma placed in a magnetic field such as ECR plasma. .

Further, since the parallel pair cable does not cause a cutoff with respect to the microwave frequency, it has the advantage that the microwave frequency can be changed significantly and measurement can be performed with high precision.

Furthermore, it has the following effects. ■The effects of complex reflections from the walls of the container can be removed. ■Since the region where the microwave electric field exists can be limited, measurement with high positional resolution is possible. ■The parallel pair cable can be easily moved within the container, so there are no restrictions on the measurement location. ■Since the direction of the microwave electric field can be specified, highly accurate measurement is possible even in the presence of a magnetic field.

It is clear that the present invention can be applied to measurements of plasmas other than semiconductor process plasmas.

[Brief explanation of the drawing]

FIG. 1 shows a first embodiment of the invention.

[Figure 2] Cross section of a parallel pair cable, (A) shows two transmission lines.
In the case of books, (B) shows the case of three books.

[Fig. 3] Diagrams showing the effects of the present invention, showing the results of measuring ECR plasma; (A) is the relationship between frequency and intensity; (B)
represents the relationship between frequency and dB.

FIG. 4 shows a second embodiment of the invention.

FIG. 5 shows a third embodiment of the invention.

FIG. 6 shows the basic configuration of a plasma processing apparatus using ECR plasma.

FIG. 7 shows the basic configuration of plasma measurement using microwave interferometry.

[Explanation of symbols]

1 Microwave source 2 Isolator 3 Microwave detector 10, 11 Coupler 12 Variable attenuator 13 Variable phase shifter 21, 22 Magic T 23, 24 Horn antenna 31, 32 Matching box 40, 41 Parallel pair cable 42 Dummy resistor 43 Coaxial cable 61 Plasma generation chamber 62 Sample chamber 63 Microwave introduction window 64 Rectangular waveguide 65 Cooling water system 66, 67 Gas introduction system 68 Plasma flow 69 Plasma drawer window 70 Sample stage 71 Sample 72 Magnetic coil 80, 81, 82, 83 Transmission line 84 Insulator 85 Electric lines of force

Claims (4)

[Claims] 1. In a plasma measurement method in which measurement microwaves are introduced into a plasma to be measured and plasma parameters are determined from changes in the phase and intensity of the measurement microwaves received from the plasma, the measurement microwaves are A plasma measurement method characterized by using parallel pair cables to propagate through the plasma to be measured. 2. In the plasma measurement method according to claim 1, a microwave oscillator with a variable oscillation frequency is used as the measurement microwave source, and the frequency of the measurement microwave is varied to adjust the intensity of the microwave transmitted through the plasma. A plasma measurement method characterized by detecting with a microwave detector and measuring the transmission characteristics of microwaves. 3. In the plasma measurement method according to claim 1, a microwave oscillator with a variable oscillation frequency is used as the measurement microwave source, and the frequency of the measurement microwave is varied to adjust the intensity of the microwave reflected by the plasma. A plasma measurement method characterized by detecting with a microwave detector and measuring the reflection characteristics of the microwave. 4. In the plasma measurement method according to claim 1, the microwave from the microwave source is divided into two, one of which is propagated in the plasma, and the other is propagated through a separately provided reference path. A plasma measurement method characterized by combining two microwaves and detecting the phase difference and intensity difference between the two microwaves.