JPH11304710A - Light interference correction method by highfrequency glow discharge emission spectroscopic analysis - Google Patents

Abstract

PROBLEM TO BE SOLVED: To correct light interference with a shorter resolution than the fluctuation period of light intensity due to light interference, by approximating the intensity of measurement light with the sinusoidal wave or the cosine wave of a plurality of limited sections. SOLUTION: Measurement data are obtained from the intensity of measurement light being measured by a high-frequency glow discharge emission spectroscopic analysis device 1. When the measurement data of the emission spectroscopic analysis device 1 are continuous, values that are obtained by sampling the continuous signal for a specific time interval are formed into a data train. Then, a data section for performing fitting processing is obtained from the measurement data train. Then, a data center that becomes the center of the measurement data is set, and the number of data points for performing fitting processing is set, thus determining a processing range. Then, a sinusoidal waveform or a cosine waveform is used as a fitting function, thius performing a fitting processing for each processing range. After that, using a plurality of sinusoidal waveforms being obtained by the fitting processing, a sputtering distance, speed, and a correction intensity value are calculated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to high frequency glow discharge emission spectroscopy, and more particularly, to a method for correcting fluctuations in light intensity due to interference light.

[0002]

2. Description of the Related Art When a high-frequency high voltage is applied between electrodes in a low-pressure argon atmosphere, a glow discharge occurs.
Ar ions generated by the glow discharge are accelerated by a high electric field, collide with the cathode side, and strike a substance from the cathode side. The particles (atoms, molecules, ions) emitted by the sputtering emit light having a wavelength specific to the element when returning from the excited state to the ground state in the plasma. In the high-frequency glow discharge emission spectral analysis, the emitted light is spectrally analyzed by a spectroscope to perform elemental analysis or the like. In this high-frequency glow discharge emission spectroscopy, the sample surface is removed by sputtering, and the elemental analysis in the depth direction can be performed by measuring the temporal change in the emission intensity of each element.

In this high-frequency glow discharge emission spectroscopy, as a method of converting a sputtering time to a sputtering depth, a method of converting the measurement time based on the periodic fluctuation of the measurement intensity with respect to the sputtering time and the wavelength of the measurement light is known (see, for example) JP-A-6-43100).

FIG. 8 is an intensity change diagram for explaining a conventional method for converting a sputtering time to a sputtering depth.

When a layered sample having a thin film formed on a substrate is irradiated with light and the reflected light is detected, the reflected light interferes due to a phase shift and a light path difference on the reflecting surface.
Further, since the thickness d of the thin film decreases with time due to sputtering, the optical path difference 2ndcosφ (n is the absolute refractive index of the thin film) also changes, and the detected light intensity is reduced by the sputtering depth as shown in FIG. On the other hand, it fluctuates at a constant cycle T in theory. In the conventional method, the period T of this intensity change is assumed to correspond to the period Ti of the time change of the measured intensity shown in FIG. The strength is determined and the sputtering time is converted to the sputtering depth.

That is, in the conventional method of converting the sputtering time to the sputtering depth, the fluctuation of the optical interference is approximated by the periodic fluctuation, and each period Ti of the optical measurement intensity is adjusted to a theoretical constant period T for each period. This corrects the error due to the interference, thereby calculating the sputtering speed, interpolating each cycle Ti, calculating the overall sputtering speed, and converting it to the sputtering depth.

[0007]

In the conventional conversion method,
Since the light measurement intensity is corrected by approximation with one cycle as a unit, and the phase is the same within this one cycle,
Short-period fluctuation within one cycle (small fluctuation in FIG. 8A)
Cannot be detected and cannot be corrected, so that there is a problem that a fluctuation in light intensity shorter than the period of light interference cannot be accurately obtained. Therefore, it is difficult to accurately detect the concentration fluctuation of the sample.

Therefore, the present invention solves the above-mentioned conventional problems and corrects the light interference with a resolution shorter than the fluctuation period of the light intensity due to the light interference. The aim is to provide a method.

[0009]

SUMMARY OF THE INVENTION The optical interference correction method using high frequency glow discharge optical emission spectroscopy according to the present invention provides a method for correcting optical interference by approximating a measured light intensity with a sine waveform or a cosine waveform of a plurality of limited sections. The light interference is corrected with a resolution shorter than the fluctuation period of the light intensity caused by the light interference.

Therefore, in the present invention, a high-frequency glow for performing elemental analysis on a sample provided with a light-transmitting thin film on a light-impermeable substrate while spectrally detecting atomic emission while sputtering the sample surface by high-frequency glow discharge. In the discharge emission spectroscopy, fitting processing is performed at each measurement point with a sine waveform or a cosine waveform using the measured light intensity of a nearby measurement point including the measurement point. The light intensity fluctuation of the measurement light due to the optical interference can be approximated by combining a plurality of obtained sine waveforms or cosine waveforms, and the resolution of this approximation corresponds to the distance between the measurement points where the fitting process is performed, When fitting processing is performed for each adjacent measurement point, the resolution is the distance between adjacent measurement points, and the resolution of the conventional optical interference fluctuation is higher than the period.

By the fitting process, an approximate sine waveform or cosine waveform cycle and amplitude can be obtained for each measurement point where the fitting process is performed. From the period of the obtained sine waveform or cosine waveform, the sputtering speed and the sputtering depth at the measurement point can be obtained.

Further, the offset component of the obtained sine waveform or cosine waveform can be used as a correction intensity value at the measurement point.

Further, assuming that the light intensity is proportional to the sputtering speed, the corrected intensity value obtained above can be corrected by the sputtering speed to obtain a final corrected intensity value.

In the approximation calculation of the fitting process, the measured light intensity of the measurement point to be subjected to the fitting process and the measured light intensity of the measurement points near the measurement point are used. No fitting process can be performed. Therefore, in the fitting process of the present invention, at the end of the measurement data,
The value obtained at the measurement point closest to the end is substituted.

[0015]

Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram showing a configuration example of a high-frequency glow discharge optical emission spectrometer to which the optical interference correction method based on high-frequency glow discharge optical emission spectroscopy of the present invention can be applied. The high-frequency glow discharge emission spectrometer 1 shown in FIG. 1 has a discharge electrode 3 formed in a hollow shape and a sample S opposed to each other, and performs vacuum evacuation while flowing an argon gas to maintain a low vacuum atmosphere. By supplying high frequency power, a stable argon glow discharge plasma is formed between the sample S and the discharge electrode 3. The high-frequency power can be supplied by the high-frequency power supply 4 and the matching device 5.

The positive ions of the plasma sputter the surface of the sample S uniformly, and cut at a speed of, for example, 10 nm / sec. Sputtered sample S
Are excited in the plasma to emit light peculiar to the element. The high-frequency glow discharge emission spectrometer 1 splits the emitted light with the spectroscope 2 and detects it with the photodetector 23. Reference numeral 21 denotes a slit, and reference numeral 22 denotes a spectroscope such as a diffraction grating. Normally, the high-frequency glow discharge emission spectrometer 1 having the above-described configuration measures the distribution of the elements in the sample S in the depth direction by measuring the temporal change of the light emission.

According to the present invention, the high-frequency glow discharge optical emission spectrometer 1 detects light intensity for each wavelength, determines the composition of the contained element from the magnitude of the measured light intensity, and determines the time of the measured light intensity. The sputtering depth and the sputtering rate are obtained from the period of the change.

The procedure of the optical interference correction method based on high-frequency glow discharge emission spectroscopy of the present invention will be described below with reference to the flowchart of FIG. 2 and the explanatory diagrams of FIGS.

A measurement data sequence is obtained from the measurement light intensity measured by the high-frequency glow discharge optical emission spectrometer 1, and a fitting process is performed using the measurement data sequence. When the measurement data of the high-frequency glow discharge emission spectrometer 1 is a continuous signal, a value obtained by sampling the continuous signal at a predetermined interval can be used as a measurement data sequence. The sampling interval can be arbitrarily determined according to the required measurement resolution.

The curves (solid line and broken line) in FIG. 3 schematically show an example of the measurement data, and crosses indicate a measurement data sequence obtained by sampling the measurement data (step S1). .

A data section [d ₁ ,..., D _m ] for performing fitting processing is determined from the measurement data sequence. This data section can be set by selecting, for example, a section where optical interference occurs in the measurement data. In FIG.
The solid curve portion sandwiched between the thick broken line sections is defined as a section for performing the fitting process, and this section is defined as a data section [d ₁ ,..., D _m ]. Incidentally, d ₁ and d _m represent the measured data of the end of the period, this is the data section within includes m pieces of the measurement data. Hereinafter, d ₁ ,.
It will be described using the m-number of the measurement data d _m (step S2).

In the present invention, a fitting process is performed for each measurement data. Therefore, it is necessary to determine a processing range for performing the fitting processing. Here, this processing range is determined by setting the data center as the center of the measurement data (step S3) and setting the number of data points for performing the fitting process.

FIG. 4 is a diagram for explaining the fitting process, shows a case of performing fitting processing on the measurement data d _j. When performing fitting processing of the measurement data d _j, and the measured data d _j and data center, when the measurement data d (before time and post time on the time axis) both from _j to process the n number of measured data, respectively Range , The total number of data points in the fitting process is (2
n + 1).

[0024] The value of n is also the same between all the measurement data to be fitting process, and can be changed for each measurement data d _j (step S4). Next, fitting processing is performed for each processing range using a sine waveform or a cosine waveform as a fitting function. Hereinafter, a case where a sine waveform is used as the fitting function will be described. Figure 5 is a diagram for explaining the fitting process of each measurement data d _j. In FIG.
The figure shows a case where the measurement data d _j−1 , d _j , and d _{j + 1} are each the data center, and the number of data points to be subjected to the fitting processing is three.

In this case, in the fitting process of the measurement data d _j−1 , an approximated sine waveform (solid line in FIG. 5) is obtained using (d _j−2 , d _j−1 , d _j ). Measurement data d
_In the fitting process of _j , (d _j−1 , d _j , d _{j + 1} )
Obtains an approximate sine wave (one-dot chain line in FIG. 5) was used to further in the fitting process of the measurement data d _{j + 1} is approximated with _{(d j, d j + 1} , d j + 2) A sine waveform (broken line in FIG. 5) is obtained.

As a result, the measured data is approximated by a plurality of sine curves, and the offset components (d' _j-1 , d' _j , d' _{j + 1} ) of these sinusoidal waveforms represent correction values of the measured data. Will be. Therefore, a correction value of the measurement data can be obtained from the approximate offset component value of each sine waveform.

The calculation for performing the fitting process with a sine waveform will be described below. In this fitting process, the fitting function at the measurement point j is defined as (a _j sin (b _j
+ C _j ) + d ′ _j ), the fitting function, the measurement data d _j (data center) at the measurement point j, and (2n) pieces of measurement data d _{j + i} (i is 0) near the measurement point j. The total (2n + 1) pieces of measurement data excluding (−n) to (n) are compared with each other and are represented by the following equation (1).

[0028]

(Equation 1) In equation (1), a _j , at which the sum of the squares of the differences is minimized,
b _j, c _j, performs fitting processing by performing arithmetic operation for obtaining the d _j.

Therefore, as shown in FIG. 5, the sine waveform obtained as a result of the operation of the operation expression (1) is obtained at each measurement point j−
A part of each sine waveform in a section of n measurement points centered at 1, 1, j, j + 1 is combined, thereby _obtaining measurement data d _j-2 , d _j-1 , d _j , d _{j +1} , d
_{j + 2} fitting processing can be performed.

Here, the coefficient a _j represents the amplitude of the corrected light intensity, the coefficient b _j represents the period of the corrected light intensity,
The coefficient c _j indicates the phase of the corrected light intensity, and the coefficient d ′ _j indicates the value (offset) of the center of the amplitude of the interference waveform, that is, the value obtained by removing the interference waveform. Therefore, by this fitting process, the measured data can be smoothed, and the amplitude, cycle, and phase of the intensity fluctuation due to optical interference, and a value from which the intensity fluctuation is removed can be obtained.

The above operation is performed in the data section [d ₁ ,.
.., D _m ] within (dn _{+ 1 to} dm- _{(n + 1)} ). This calculation is performed in the section of (d _{n + 1} to d _{m− (n + 1)} ) because, in the above-mentioned calculation formula (1), n pieces of measurement data are used on both sides from the data center. If the measured data _dj is equal to or _smaller than d _n or equal to or _larger than d _mn , the data used in equation (1) cannot be obtained.

FIG. 6 is a diagram for explaining a data section in which the calculation of the fitting process is performed. In FIG. 6, section A represents a data section in which fitting processing is performed by the above equation (1), and sections B and C represent data sections in which fitting processing by the above equation (1) cannot be performed. In the fitting processing of the sections B and C, the fitting result at the end of the section A adjacent to the sections B and C can be used instead. For example,
For data interval B (d _j ~d _n) is the measured data d
Applying the sine waveform obtained using _{n + 1} , data section C
(D _mn ~d _m) For applies a sine wave obtained using the measured data d _{m- (n + 1).}

In the following, in steps S3 to S5, the processing performed on each measurement data in the data section A will be referred to as a processing step. The waveform is applied to data section B and data section C (steps S5 and S6). Next, the calculation of the sputtering distance u, the calculation of the sputtering speed v, and the calculation of the correction intensity value are performed using the plurality of sine waveforms obtained by the fitting process.

Here, the measurement wavelength is λ, the refractive index of the thin film is n, and the integration time (corresponding to the time between the measurement point j and the measurement point (j + 1)) of the intensity measurement in each processing step is t.
Then, the sputtering distance u at the measurement point j is
_j can be expressed by the following equation (2) which converts the time of one processing step into a distance. u _j = (1/4) · (b _j / π) · (λ / n) (2) The sputtering speed v _{j at the} measurement point j is obtained by dividing the sputtering distance u _j by the time t, and the following formula ( It can be represented by 2).

V _j = (1/4) · (b _j / π) · (λ / n) · (1 / t) (3) Therefore, for the data section A (d _{j to} d _n ), The sputtering distance u _j and the sputtering speed v _j obtained by the equations (2) and (3) are applied.

For the data section B (1 to n), the sine waveform obtained at the measurement point (n + 1) is applied, and the sputtering speed v _j and the sputtering distance u _j are determined at the measurement point (n +
The value is the same as the value obtained in 1).

The interference correction strength d ' _j is calculated at the measurement point (n +
The fluctuation component (a) due to the sine wave in the equation of the interference waveform in 1)
_{n + 1} sin (b _{n + 1} (j− (n + 1)) + c _{n + 1} )) subtracted from measurement data d _j d ′ _j = d _j −a _{n + 1} sin (b _{n + 1} (j − (N + 1)) + c _{n + 1} ) (4) Note that j is 1 to n.

Further, data section C ((m-n) ~m ) measurement points for (m- (n + 1)) by applying the obtained sinusoidal waveform, sputtering rate v _j and sputtering distance u _j is the measurement point The value is the same as the value obtained by (m- (n + 1)).

Further, the interference correction intensity d ' _j is calculated at the measurement point (m−m
(N + 1)), the variation component due to the sine wave in the equation of the interference waveform (a _{m− (n + 1)} sin (b _{jm− (n + 1)} (j− (m− (n +
_{1))) + c m- (} n + 1))) value d 'obtained by subtracting from the measured data d _j a _{j = d j -a m- (n} + 1) sin (b jm- (n + 1) (j − (M− (n + 1))) + c _{m− (n + 1)} ... (5) is used. Note that j is (mn) to m.

[0040] Further, the measured intensity in the data section A (d _j ~d _n) is proportional to the sputtering rate, v
Correction is performed based on _{n + 1} , and the final interference correction intensity value d _j ″ can be expressed by the following equation.

D ″ _j = d ′ _j · (v _{n + 1} / v _j ) (6) Therefore, the interference correction intensity value is calculated in the data section A (d _{n + 1}
To d _{m− (n + 1)} ), d ″ _j expressed by the equation (6) is applied, and in the data section B (1 to n), the sputtering rate is fixed (v _{n + 1} ) and d ″ _{j j} = d ′ _j · (v _{n + 1} / v _{n + 1} ) = d ′ _j represented by d ′ _j is applied, and the data section C ((mn)
-M), the sputtering rate is constant ( _{vm- (n + 1)} ), and d " _j is represented by d" _j = d' _j. ( _{Vm- (n + 1)} / vn _{+ 1} ). _j is applied (step S7).

FIG. 7 is an example to which the method of the present invention is applied. In the analysis of the depth of boron in the BPSG film, FIG.
FIG. 7A shows measurement data before performing interference correction, and FIG.
FIG. 7B shows the state of one processing step, and FIG. 7C shows the correction data after performing the interference correction.

In FIG. 7C, the distance u is plotted on the horizontal axis, and the interference correction intensity value (d ″) is plotted on the vertical axis, using the above calculation result.
_j ), the correction data is sequentially added to the sputtering distance u _j of each processing step, and is displayed at a predetermined position.

[0044]

As described above, according to the optical interference correction method based on high frequency glow discharge emission spectroscopy of the present invention,
The light interference can be corrected with a resolution shorter than the fluctuation period of the light intensity due to the light interference.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration example of a high-frequency glow discharge emission spectrometer to which the method of the present invention can be applied.

FIG. 2 is a flowchart for explaining the procedure of the method of the present invention.

FIG. 3 schematically shows an example of measurement data.

FIG. 4 is a diagram for explaining a fitting process.

5 is a diagram for explaining the fitting process of each measurement data d _j.

FIG. 6 is a diagram illustrating a data section in which a calculation of a fitting process is performed.

FIG. 7 is a graph showing an example to which the method of the present invention is applied.

FIG. 8 is an intensity change diagram for explaining a conventional method for converting a sputtering time to a sputtering depth.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Glow discharge optical emission spectrometer, 2 ... Spectroscope, 3 ... Discharge electrode, 4 ... High frequency power supply, 5 ... Matching device, 6 ... High frequency plasma, S ... Sample.

Claims (1)

[Claims] 1. A high-frequency glow discharge emission spectroscopy for performing an elemental analysis on a sample having a light-transmitting thin film on a light-impermeable substrate by spectrally detecting atomic emission while sputtering the sample surface by high-frequency glow discharge. In each of the measurement points, by performing a fitting process on the measurement light intensity of a measurement point near the measurement point including the measurement point with a sine waveform or a cosine waveform, the light intensity variation of the measurement light due to optical interference can be reduced by a plurality of sine or cosine waveforms. Approximate by combination, determine the corrected sputtering speed and corrected sputtering depth of each measurement point by the cycle of each sine waveform or cosine waveform, and calculate the corrected intensity value of each measurement point by the offset component of each sine waveform or cosine waveform. , Optical interference correction method by high frequency glow discharge emission spectroscopy.