JP2003302334A - Analytical method for ultrathin-film two-layer structure using spectral ellipsometer - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method for measuring and analyzing an ultrathin-film two-layer structure by using a spectral ellipsometer. <P>SOLUTION: The method comprises a spectrum measuring stage, a first analytical stage, a second analytical stage and a third analytical stage. At the spectrum measuring stage, the ultrathin-film two-layer structure on the surface of a substrate as a measuring object is measured by using the spectral ellipsometer so as to obtain data on the object. In a first step at the first analytical stage, a plurality of models of the object are set up, and the plurality of models are fitted to a measuring spectrum in a second step, and the results of the models are decided in a third step. In a first step at the second analytical stage, a result obtained at the first analytical stage is set as an initial value of a new model, and an extended BLMC operation is performed in the second and third steps. In a first step at the third analytical stage, a result obtained at the second analytical stage is used, a final fitting operation is performed, and a result obtained by the fitting operation is confirmed at the second step. In the third step, the results are saved. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention uses data obtained by a spectroscopic ellipsometer to calculate a local minimum value (Extend).
ed Best Local Minimum Caluculation, Exte
The present invention relates to a method for analyzing an ultrathin film two-layer structure which is processed by using N.

[0002]

2. Description of the Related Art A polarization change amount of incident light and reflected light is measured by using a spectroscopic ellipsometer, and from the result, the film thickness (d),
The complex refractive index N (N = n-ik) can be calculated. The polarization change amount (ρ) is represented by ρ = tan ψ exp (iΔ) and depends on parameters such as wavelength (λ), incident angle (φ), film thickness, and complex refractive index. Therefore, the relationship is as follows. . (D, n, k) = f (Ψ, Δ, λ, φ)

When the incident angle is fixed, in the single wavelength ellipsometer, for three unknowns of (d, n, k),
Since only two independent variables can be measured, it is necessary to fix any one of d, n, and k as known. The measurement variable increases when the incident angle is changed even at a single wavelength. However, due to the difference in the incident angle (φ) (Ψφ ₁ , Δφ
₁ ) and (Ψφ ₂ , Δφ ₂ ) have a strong correlation,
It is difficult to accurately obtain d, n, k.

[0004]

The amount of change in polarization can be expressed by the volume through which light passes, (phase angle (β) × area of beam diameter). The phase angle (β) is expressed by the following equation. β = 2π (d / λ) (N ₀ ² −N ₀ ² sin ² φ ₀ ) ^{1/2 When the} beam diameter is constant, the polarization change amount is as follows. Polarization change amount ∝ film thickness (d) × (complex refractive index N) × φ Here, φ is the incident angle. The film thickness (d) is thin,
When N is low, the change in the phase angle (β) is small, and the correlation between d and N is strong.

The above-mentioned problem is caused by the ultra-thin two-layer structure.
Further, there is a possibility that a correlation may occur between the layers, and the measurement result (Ψ _E (λ _i ),
It is difficult to obtain n, k, and d of each layer from Δ _E (λ _i ).

As described above, the measured spectrum (Ψ _E (λ _i ), Δ
_E (λ _i )) is the n, k information of the substrate, the n, k of each layer,
It contains all the k and d information, but from this it is not possible to calculate the unique combination of the n, k information of the substrate and the n, k, d information of each layer (except for the substrate only). . The method for finding this unique combination is called analysis of spectroscopic ellipsometer data. When performing the analysis, a model is created using the n, k information of the substrate and the n, k, d information of each layer. N of the substrate and each layer in this,
As the k information, a reference (known table data), a dispersion formula, or an optical constant of a single-layer thin film made of a similar material is used. The dispersion formula is a formula showing the wavelength dependence of the permittivity of a substance. In the near infrared to ultraviolet region, this permittivity ε
(Λ) is determined from the bonding mode of the constituent atoms of the material. As dispersion formulas, there are known calculation formulas based on harmonic oscillators, calculation formulas based on quantum mechanics, empirical formulas, etc.
Contains one or more parameters. All unknowns (thickness of each layer, dispersion formula parameter, mixing ratio, etc.) included in the model are changed and adjusted to the measured data. This is called fitting. As a result of this fitting, the film thickness and mixing ratio of each layer can be found,
From the dispersion parameters, the dielectric constant ε (λ) of the material can be calculated. The dielectric constant and the refractive index of a material have the following relationship. ε = N ^{2 In the} present invention, the effective medium theory (Effective Medium Thorough) is considered by considering the case where the wavelength dependence of the dielectric constant of the film cannot be specified or is difficult with one dispersion formula for various reasons.
effective dielectric constant (Effective Dielectric F
unction) is calculated. Generally, for example, when N substances (guests) having various dielectric constants are mixed in a host material, the effective dielectric constant (ε) is expressed as follows. At this time, ε _h is the permittivity of the host, ε _j is the j-th permittivity, and k is the screening factor. Here, if the host material and the material inside are mixed in approximately the same amount, or if it is not known which is the host or guest, the host material itself becomes the same as the effective medium material, ε = ε _h . . The effective medium theory of this condition is based on the Bruggeman Effective Medium App
roximation is hereinafter simply referred to as EMA) in the present application. The dielectric constant ε when the three spherical substances A, B and C are mixed in contrast is given by the following equation. f _a (ε _a −ε) / (ε _a + 2ε) + f _b (ε _b −ε)
_{/ (Ε b + 2ε) +} f c (ε c -ε) / (ε c + 2ε)
= 0 where, epsilon: the effective dielectric constant epsilon _a to be obtained, ε _b, ε _c: spherical material A, B, C of the dielectric constant _{f a, f b, f c} : mixing ratio of each substance (Volume Fractio
In n below Vf), if _{f a + f b + f c} = 1 film on the substrate is mixed heterogeneous or discontinuous and, some material is sufficiently smaller than the wavelength order,
For a medium composed of a plurality of physically mixed substances, an effective medium approximation (EMA) is used to make a model. The case where the substance A, the substance B, and the substance C are mixed will be described. At this time, the effective medium approximation (EMA) is the mixture ratio of the substance A, the mixture ratio of the substance B, the mixture ratio of the substance C, the film thickness of the mixed layer of A, B and C, the dielectric constant and the dispersion formula or reference data. Etc. are estimated and fitting is performed to evaluate.

An object of the present invention is to set a combination model such as a film thickness and a complex index of refraction, calculate a simulation spectrum of the combination model, and perform fitting of the simulation spectrum and the measurement spectrum with a wide range minimum value calculation method (Extended BLMC). An object of the present invention is to provide an ultrathin film two-layer structure analysis method using a spectroscopic ellipsometer, which determines the ultrathin film two-layer structure by using the above method.

[0008]

In order to achieve the above object, the method of measuring and analyzing the ultra-thin film two-layer structure according to the present invention is basically such that the ultra-thin film two-layer structure is first measured using a spectroscopic ellipsometer. Then, a spectrum is obtained. The new analysis method basically consists of three stages. The purpose of the first stage of analysis is to select a plurality of models that are considered to fit the actual sample and determine the initial value. In the second stage of analysis, Extend is performed based on the initial value obtained in the first stage.
ed BLMC is performed. If necessary, at the third stage of analysis, final fitting, confirmation, and storage are performed.

The minimum value calculation method (BLMC) used in the ultrathin film and thin film measurement method used in the present invention is an ultrathin film and thin film measurement method on a substrate surface to be measured using a spectroscopic ellipsometer by the minimum value calculation method (BLMC). , The measurement spectra Ψ _E (λ _i ) and Δ _E (λ _i ) of the thin film on the surface of the substrate to be measured are the changes in the polarization of the incident light and the reflected light for each wavelength λ _i by changing the wavelength of the incident light. Ψ _E , Δ _E
Spectrum measurement step, and (N ₀ , (n _0, k
₀ )), (d, N (n _, k)) of the thin film on the substrate is assumed using a dispersion formula, and further, a plurality of film thicknesses (d ± mΔd) within the expected range and the expected range Based on the step of setting a plurality of incident angles (φ ± mΔφ) inside and the combination of the incident angle and the film thickness, the dispersion formula (DS
P) parameter (ε _s, ω _t ) fitting step, and each Ψ _M obtained by the fitting.
From (λ _i ) and Δ _M (λ _i ), Ψ _E (λ _i ) and Δ
_The step of selecting the fitting result (DSP _best ) of the model in which the combination of the film thickness (d _best ) and the incident angle (φ _best ) with the smallest difference from _E (λ _i ) is set, and the selection step Using the incident angle (φ _best ) as a definite value, the film thickness (d _best ) and the dispersion formula (DSP _best )
It consists of the steps of fitting.

In order to achieve the above-mentioned object, the contract according to the present invention
Ultra-thin film two-layer structure using the spectroscopic ellipsometer according to claim 1.
The structure analysis method is an ultra-thin two-layer structure on the surface of the substrate to be measured.
For each wavelength λ by changing the wavelength of the incident light._i Incident light and anti
Measurement spectrum Ψ, which is the change in the polarization of the incident light_E (λ_i ) When
Δ_E (λ _i ) Ψ_E , Δ_E Spectrum measurement stage,
Substrate (N₀, (N_0,k₀ )) And each
Possible complex bending of thin film materials (Mat1, Mat2)
Folding rate (N₁, (N_1,k₁ )), (N₂, (N_2,k₂)),each
Layer thickness (d_1,d₂ ) To make several models
The first step, the measurement spectrum for each model
Second step before fitting with tor, and before
As a result of fitting for each model, minimum χ² The value
Model or the maximum and maximum preset film thickness
Χ in the small price² Result of the lowest model of (d
_{1 (best),}d_{2 (best)} ) Consists of the third step
The first stage of analysis and the results obtained in the first stage of analysis are updated.
The first step of setting the initial value of the model
The thickness of either layer (d_{1 (best)} Well
Or d_{2 (best)}) As the center value and multiple points around it
For each layer (d_{2 (best)} Or d_{1 (best)} )
2nd step of fitting using BLMC at
The result of the BLMC performed at the step and the above multiple points
Minimum χ² Value or preset film thickness, minutes
Within the maximum and minimum values of the dispersion parameter and angle of incidence, respectively.
Contains χ² 3rd step to select the lowest model of
The second stage of analysis and the results obtained in the second stage of analysis.
The first step to perform the final fitting using the fruits
Check the results obtained by the above fitting.
2 steps and 3rd step to save the result
And the third stage of the analysis.

According to the present invention, there is provided a method for analyzing an ultrathin film two-layer structure using the spectroscopic ellipsometer according to claim 2.
In the method for analyzing an ultrathin film two-layer structure using the spectroscopic ellipsometer described above, in the second step of the second analysis step, the layer to be fitted using BLMC is determined by the optical constant in the material in the two-layer structure. It is intended to be the layer that is more difficult to understand.

According to the present invention, there is provided a method for analyzing an ultrathin film two-layer structure using the spectroscopic ellipsometer according to claim 3, which is defined in claim 1.
Alternatively, in the analysis method of the ultrathin film two-layer structure using the spectroscopic ellipsometer described in 2, in the material in the two-layer structure,
For those whose optical constants are better known, with the film thickness obtained in the first step and the third step of the analysis as the central value, within a range of several% to several tens% before and after that, for each set film thickness , An analysis (extended BLMC) is performed by combining the analysis second step second step and the analysis second step third step for performing BLMC on the other layer.

[0013] analyzing method according to claim 4 very thin film 2 layered structure using a spectroscopic ellipsometer according according to the invention, a very thin film two-layer structure of the substrate surface to be measured, each wavelength by changing the wavelength of incident light lambda _i Ψ _E , which obtains the measured spectra Ψ _E (λ _i ) and Δ _E (λ _i ) which are the changes in the polarization of the incident light and the reflected light for each
When the Δ _E spectrum measurement step and either one or two layers are non-uniform or discontinuous and some materials are mixed, a model is created using the effective medium approximation and Possible complex refractive indices (N ₁ ,, N ₀ , (n _0, k ₀ )) of the substrate and materials (Mat 1, Mat 2) of each thin film
(N _1, k ₁ )), (N ₂ , (n _2, k ₂ )), mixing ratio (Vf
₁ , Vf ₂ ), the film thickness (d _1, d ₂ ) of each layer is used to make some models, a second step of fitting each model to the measured spectrum, and the result of the fitting of each model, each of the maximum thickness of the mixing ratio set model or in advance with the lowest chi ² values, the lowest model of which chi ² to enter into the minimum result (d 1 _{( best),} d _{2 (best)} ,
An analysis consisting of the third step of determining Vf _(best) 1
And the film obtained in the first step and the third step of the analysis of the layer containing the unknown dispersion equation as the initial value of the new model based on the results of the film thickness / mixing ratio obtained in the first step of the analysis. Based on the thickness value, the thickness ((d ₁ ± mΔ
d ₁ ) or (d ₂ ± mΔd ₂ )) is set, and the film thickness of the other layer is centered on the value obtained in the first step of the analysis and the third step, and a plurality of points ((d ₂ ±
mΔd ₂ ) or (d ₁ ± mΔd ₁ )), and the multiple values (Vf ± mΔVf) around it with the value of the mixing ratio obtained in the first step of the analysis and the third step as the central value. The first step, the mixing ratio multiple points (Vf ± mΔVf) and the film thickness multiple points ((d ₂ ± mΔd ₂ ) or (d ₁ ± mΔd ₁ ) of the layer containing the unknown dispersion formula) ), The second step of the analysis, the second step of performing BLMC for the layer containing the unknown dispersion equation,
From the results obtained for each combination of the mixing ratio and the film thickness of the other layer containing the unknown dispersion formula, the minimum χ ² value or preset film thickness, dispersion formula parameter, mixing ratio , A third step of selecting a combination having a minimum χ ² value falling between the maximum and minimum values of the incident angles, and a second step of the analysis and a third step of the second step of the analysis. Based on these values, fitting of both film thickness, mixing ratio and dispersion formula or fitting of both film thickness and mixing ratio, 3rd step 1st step, confirmation of the result obtained by the fitting It consists of two steps, and a third analysis step consisting of a third step of storing the result.

A spectroscopic ellipsometer according to claim 5 according to the present invention.
A method for analyzing an ultra-thin film two-layer structure using a meter is set forth in Claims.
Ultra-thin film two-layer structure using the spectroscopic ellipsometer described in 1, 4
Structure analysis method, ultra thin film 2 on the surface of the substrate to be measured
Change the wavelength of the incident light by changing the wavelength of each layer λ_i Each incident
Measurement spectrum Ψ, which is the change in polarization of light and reflected light_E (λ
_i ) And Δ_E (λ _i ) Ψ_E , Δ_E Spectral measurement stage
Floor and multiple measurement conditions (Zi) within expected range
For each of the first and second analysis steps of claims 1 and 4,
Of the results obtained under each measurement condition.
² The value or the parameter of the dispersion formula or the mixing ratio Vf is set
Among the combinations between the maximum and minimum values, χ² 
According to the second step and the fourth step of selecting the best value
It is configured.

According to the present invention, there is provided a method for analyzing an ultrathin film two-layer structure using the spectroscopic ellipsometer according to claim 6, which is a method for analyzing an ultrathin film two-layer structure using the spectroscopic ellipsometer according to claim 1, 2, 3, 4 or 5. In the analysis method, the step of selecting the one having the smallest difference in the first, second, and third steps of analysis obtains the mean square error between the fitted one and the measured value, and the one having the smallest mean square error, or The smallest mean square error among the maximum and minimum values of the preset film thickness, dispersion parameter, mixture ratio, and incident angle variable is determined.

[0016]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of a method for analyzing ultrathin film two-layer structure data using a spectroscopic ellipsometer according to the present invention will be described below with reference to the drawings. FIG. 1 is a schematic diagram showing the configuration of an ellipsometer used in the method of the present invention. A part of the sample (sample) 4 to be measured is shown in an enlarged manner.

The spectroscopic ellipsometer shown in FIG. 1 performs the spectroscopic measurement data acquisition step of the method described below. First, the measuring device will be briefly described.
The Xe lamp 1 is a so-called white light source containing many wavelength components. The light emitted from the Xe lamp 1 is guided to the polarizer 3 via the optical fiber 2. The light polarized by the polarizer 3 is incident on the surface of the sample 4 to be measured at a specific incident angle (for example, φ = 75.00 °). Sample 4 is a measurement sample in which two thin films are formed on the surface of the substrate described later. The reflection from the sample 4 is caused by the photoelastic modulator (P
It is guided to the analyzer 6 via the EM) 5. Photoelastic modulator (PEM) 5 phase-modulates to a frequency of 50 kHz to produce linear to elliptically polarized light. Therefore, several meters
Ψ, Δ can be determined with a resolution of seconds. Analyzer 6
Is connected to the spectroscope 8 via the optical fiber 7.
The output data of the spectroscope 8 is captured by the data capturing section 9,
The acquisition step of the spectroscopic measurement data is ended. PE
The position of M5 can be either after the polarizer 3 or before the analyzer 6.

FIG. 2 is a flow chart showing an embodiment of the thin film measuring method according to the present invention. FIG. 3 is a detailed explanatory diagram showing a first stage of analysis in the embodiment. FIG. 4 is a detailed explanatory diagram showing a second stage of analysis in the embodiment. FIG. 5 is a detailed explanatory view showing a third stage of analysis in the above embodiment. The symbols used in the present invention are collectively shown in the next paragraph.

(Definition of Symbols) Sub: Substrate (physical constant is known, can be handled as a bulk) Mat: Thin film material (optical constant of substance) Mat _ij : Material of i-th layer of j-th model (same as above) dj: jth layer thickness dj _(best) : jth layer thickness d _ij obtained by fitting: i-th layer thickness d _{ij (best) of} j-th model: j-th model obtained by fitting Thickness of the i-th layer X2: χ ² value X2 _(j) : χ ² value in the j-th model Void: Material with n = 1, k = 0 Vf _(ij) : i-th layer in the j-th model Mix ratio (Volume fr
action) Vf _{(ij) (best)} : Mixing ratio of the i-th layer of the j-th model obtained by fitting

(Meaning of χ ² as a result of fitting) N
Measurement data pairs Exp (i = 1, 2 ..., N) and calculated data pairs Mod of the corresponding N models of the model
(I = 1, 2, ..., N), the measurement error has a normal distribution, and the standard deviation is σ _i , the mean square error (χ
² ) is given as follows. Here, P is the number of parameters. The small χ ² means that the agreement between the measurement results and the model is large. Therefore, when comparing a plurality of models, the one with the lowest χ ² is the best model.

(Measurement Stage Data Measurement Step) The measurement is performed by the apparatus shown in FIG. 2 on the substrate surface of measurement target 4
A thin film with a layered structure (see the enlarged view in the figure) is used as a measurement spectrum Ψ _E (λ _i ), which is the change in polarization of the incident light and the reflected light at each wavelength λ _i by changing the wavelength of the incident light. Δ _E (λ
Measure the Ψ _E and Δ _E spectra to obtain _i ). (Measurement Stage Data Storage Step) The data measured in the previous step is stored and used as comparison target data (see FIG. 2).

(First Step of Analysis Step 1) In this step, the possible complex refractive index (N) of (N ₀ (n ₀ , k ₀ )) of the substrate and the material (Mat 1, Mat 2) of each thin film is calculated.
₁ , (n ₁ , k ₁ )), (N ₂ , (n ₂ , k ₂ )), and the film thickness (d _1, d ₂ ) of each layer are used to make some models. In this example, the following models (1) to (4)
Is set. The respective models are schematically shown in a portion indicated by 31 in FIG. Model (1) is a board (Su
The first layer (the optical constant of Mat1 and the film thickness d ₁₁ ) and the second layer (the optical constant of Mat2 and the film thickness d ₂₁ ) are formed on b). Model (2) is the first layer on the substrate (Sub).
(Optical constant of Mat1, film thickness d ₁₂ ) and second layer (Mat2
The optical constant + Void and the film thickness d ₂₂ ) are formed. Void is a space having a refractive index of 1. model
(3) is the first layer (the optical constant of Mat1 and the film thickness d ₁₃ ) and the second layer (the optical constant of Mat2 + Ma) on the substrate (Sub).
optical constants of t3, is obtained by forming a film thickness d _23). Note that (Mat2 + Mat3) means a mixture of the material 2 and the material 3 in a certain ratio. Model (4) is the first layer (Mat1 optical constant, film thickness d on the substrate (Sub)).
₁₄ ) and the second layer (amount (Spe
cial reference) + Void, film thickness d ₂₄ ). The reference amount is an optical constant obtained by using a similar sample. Where model (2) ~
From the mixing ratio set in the second layer of (4), the optical constant as a homogeneous film can be obtained using the effective medium approximation theory. In each model, the optical constant of the material of the substrate Sub is known, the optical constant of the first layer is almost known, and
Optical constants of layers and film thickness d ₁ of the first layer, film thickness d _{1 of} the second layer
Assuming that ₂ is unknown (uncertain), the above 4
I have made a model.

(First Step of Analysis Step 2) For each of the four models (1) to (4) selected in Step 1, the measurement data Ψ _E , obtained from the measurement spectrum,
Fit with Δ _E. In the part indicated by 32 in FIG. 3, the fitting target of each model and the χ ² value of the data obtained as a result of the fitting are shown. Model (1)
Then, by fitting the film thickness d ₁₁ of the first layer and the film thickness d ₂₁ of the second layer, the result d _{11 (best)} , d _{21 (best)} and χ
^{Get the binary} value X2 ₍₁₎ . In the model (2), fitting of the film thickness d ₁₂ of the first layer, the film thickness d ₂₂ of the second layer and the mixture ratio of the second layer is performed, and the result d _{12 (best)} , d _{22 ( best)} , Vf
We obtain _{22 (best)} and χ ² values X 2 ₍₂₎ . In the model (3), fitting of the film thickness d ₁₃ of the first layer, the film thickness d ₂₃ of the second layer and the mixing ratio of the second layer was performed, and the result d
_{13 (best)} , d _{23 (best)} , Vf _{23 (best)} and χ ² value X2
_{Get (3)} . In the model (4), fitting of the film thickness d ₁₄ of the first layer, the film thickness d ₂₄ of the second layer and the mixture ratio of the second layer was performed, and the results d _{14 (best)} , d _{24 ( best)} , Vf _{24 (best)}
And the χ ² value X 2 ₍₄₎ is obtained. From the results of (analysis first stage step 3) the plurality of sets of fittings, the step of selecting the results of the model of the lowest chi ² value contained largest minimum chi ² value or thickness that is set in advance, between the minimum Which is shown at 33 in FIG.

(Second Step of Analysis Second Step) 41 of FIG. 4 shows a model (initial value) used in the second step of analysis. In this example, since the optical constant of Mat1 is almost known, the optical constant (Ma
t1) is used as it is. Since the optical constant (Mat2) of the material of the second layer is unknown, the dispersion formula is used here. As the initial value of the film thickness (d), the value obtained in the analysis first step 3 is used. (Second step of analysis step 2) In this step, as shown by 42 in FIG. 4, d _{1 (best)} +10 with the first layer film thickness (d _{1 (best)} ) of the model (initial value) as the center value. %, D
_{1 (best)} + 5%, d1 _(best) , d1 _(best) -5%, d
_{1 (best)} -Change 10% up and down. The thickness (d _{2 (best)} ) of the second layer is ± 10 with respect to the plurality of points (5 points).
BLMC is performed by changing the range of%. The thickness d _2j of the second layer of each model, the optical constant Mat _2j, and the χ ² value X 2 _(j) are obtained. (Analysis second step Step 3) respective maximum results made in Step 2 the lowest chi ² value or a preset thickness with a distributed type parameter, a model of the lowest chi ² values contained in between the minimum Select (FIG. 3, 43).

(Analysis third step Step 1) The optical constants of the first and second layers of the model selected in the analysis second step Step 3 are fixed and the fitting of the film thickness of the first and second layers is performed. Alternatively, the optical constants of the first layer are fixed and the film thicknesses of the first and second layers and 2
The optical constants of the layer are fitted (see FIGS. 5 and 51). (Analysis third step Step 2) It is confirmed whether or not the result of Step 1 is the minimum χ ² value that falls between the maximum and minimum values of the preset film thickness and the dispersion equation parameter (FIG. 5, FIG. 5). 52). When the confirmation of the result is not valid, the procedure returns to the first stage of analysis, a new model is set, and fitting (step 2) is performed. (Third step in the analysis, step 3) If the result in step 2 is valid, it is saved (see FIGS. 5 and 53).

[0026]

EXAMPLE Next, similarly to the above, an example in which the optical constants of the first layer are substantially known and the optical constants of the second layer and the film thicknesses of the first and second layers are unknown will be described. The flow chart shown in FIG. 2 can be used as is. In this embodiment, the substrate Sub is Si, the first layer material is SiO ₂ , and the second layer material is S
iN _X. The sample is measured by the device shown in FIG. The thin film of the two-layer structure on the substrate surface of the measurement target 4 is
Measurement spectra Ψ _E (λ _i ) and Δ _E, which are changes in the polarization of incident light and reflected light for each wavelength λ _i by changing the wavelength of the incident light
Measure the Ψ _E and Δ _E spectra to obtain (λ _i ), and save the measured data as the comparison target data.

At the first stage of the analysis shown in FIG. 3, the material of each thin film (Si
O ₂ , SiN _x ) possible complex refractive index ((N ₁ , (n _1, k
₁ )), (N ₂ , (n _2, k ₂ )) and the film thickness (d _1, d ₂ ) are used to prepare a model. In this embodiment, it is assumed that the following models (1) to (4) are set.

(Analysis first stage step 1) Model (1)
Is a first layer (optical constant of SiO ₂ and film thickness d ₁₁ ) and a second layer (optical constant of Si ₃ N ₄ and film thickness d ₂₁ ) formed on a substrate (Sub: Si). . Model (2) is a board (S
ub: Si) on the first layer (optical constant of SiO ₂ , film thickness d ₁₂ ).
And the second layer (optical constant of Si ₃ N ₄ + Void, film thickness d ₂₂ ) are formed. Void is a space having a refractive index of 1. Model (3) is the first on the substrate (Sub: Si).
A layer (optical constant of SiO ₂ and film thickness d ₁₃ ) and a second layer (optical constant of Si ₃ N ₄ + optical constant of SiN _X and film thickness d ₂₃ ) are formed. Note that (Si ₃ N ₄ + SiNx) means a mixture of Si ₃ N ₄ and SiNx in a certain ratio. Model (4) is the first layer (SiO ₂ optical constant, film thickness d) on the substrate (Sub: Si).
₁₄ ) and the second layer (SiNx optical constant (known reference amount) + V
Oid and film thickness d ₂₄ ) are formed. Here, from the mixing ratio set in the second layer of the models (2) to (4), the optical constant as a homogeneous film can be obtained using the effective medium approximation theory. In addition, each of the above models is a substrate Su
The optical constant of the material of b is known, the optical constant of SiO ₂ of the first layer is almost known, and the optical constant of SiNx of the second layer and
It is assumed that the layer thickness d _{1 and} the second layer thickness d ₂ are unknown (uncertain).

(Analysis first step Step 2) For each of the four models (1) to (4) selected in Step 1, the measurement data Ψ _E , obtained from the measurement spectrum,
Fit with Δ _E. In the portions indicated by 31 and 32 in FIG. 3, the fitting target of each model and the χ ² value of the data obtained as a result of the fitting are shown. In the model (1), the film thickness d ₁₁ of the first layer and the film thickness d ₂₁ of the second layer
And the result is d _{11 (best)} , d _{21 (best)}
And χ ² value X 2 ₍₁₎ is obtained. In the model (2), fitting of the film thickness d ₁₂ of the first layer, the film thickness d ₂₂ of the second layer, and the mixing ratio of the second layer was performed, and the result d _{12 (best)} , d _{22 (best)} , Vf
We obtain _{22 (best)} and χ ² values X 2 ₍₂₎ . In the model (3), fitting of the film thickness d ₁₃ of the first layer, the film thickness d ₂₃ of the second layer and the mixing ratio of the second layer was performed, and the result d _{13 (best)} ,
Obtain d _{23 (best)} , Vf _{23 (best)} and χ ² value X 2 ₍₃₎ .
In the model (4), fitting of the film thickness d ₁₄ of the first layer, the film thickness d ₂₄ of the second layer and the mixing ratio of the second layer was performed, and the result d _{14 (best)} , d _{24 (best)} , Vf _{24 (best)} and χ ² value X2
_{Get (4)} .

(First step of analysis, step 3) From the results of the plurality of sets of fittings, the minimum χ ² value or the minimum χ between the maximum and minimum values of the preset film thickness is obtained.
A step of selecting the result of a ^binary model, which is shown in 33 in FIG. 3.

(Analysis second stage step 1) to (Analysis third step)
Stage step 3) is as described above.

[0032]

As described above in detail, according to the present invention, the data obtained by using the spectroscopic ellipsometer for the thickness and the optical constants of the two layers of the ultrathin film formed on the substrate can be obtained by a wide range minimum value calculation method. It can be analyzed by processing with (Extended BLMC). As described above, the procedure basically including three steps can be used for analysis of various structures having two or more layers.

According to the method of the present invention: Even with an ultrathin film multilayer structure, reliable film thickness (in each layer) and optical constant (at least one layer) can be obtained by using this method. 2. By performing the first step of this procedure, the initial value range of the unknown parameter can be narrowed. 3. By using Extended BLMC, unfavorable local minimum (Local Minim)
um) is dramatically reduced and results are more reliable.

Various modifications can be made to the embodiments described in detail above within the scope of the present invention. For ease of understanding, we have consistently described Ψ, Δ in relation to data acquisition and model setup. Similar measurements and fittings are possible using the following data pairs, which are well known to those skilled in the art, and are within the scope of the present invention. (N, k) (ε _r , ε _i ) (tan Ψ, cos Δ),
(I _s, I _c )

Mat1 and Mat2 on the substrate can be used not only for the ultra-thin film dielectric material multilayer structure but also for various applications such as various thicknesses and materials. Although the photoelastic modulator (PEM) is shown as an example, an ellipsometer other than the PEM can be used.

All or part of the above means may be carried out, and this is also included in the technical scope of the present invention.

As the substrate, besides Si, glass, quartz, compound semiconductors, etc. can be similarly used. Further, regardless of the type of substrate, any flat substrate or existing substrate can be used.

For the dispersion formula, various formulas and parameters can be used in addition to classical (classical mechanics), amorphous (quantum mechanics), empirical formulas, and these are also included in the technical scope of the present invention. .

In the example, it has been described that all all parameters are fitted simultaneously, but there are cases where they are fitted separately, and this is also included in the technical scope of the present invention.

In BLMC, which is a part of Extended BLMC, the incident angle may be fitted as described above. Further, in the procedure, the incident angle and various parameters are simultaneously fitted, but there are cases where they are fitted separately or the incident angle is fixed, and these are all included in the technical scope of the present invention.

The incident angle may be fitted as a general parameter other than BLMC, which is also included in the technical scope of the present invention.

In the example, in the model preparation process of the first step of the analysis, the substrate is Si, the first layer is SiO ₂ , and the second layer is
The layers are (1) Si ₃ N ₄ , (2) Si ₃ N ₄ and Void, (3) Si ₃ N ₄
And SiN _X , and (4) SiNx and Void models (1) to (4) were used, but the type and number of models may change depending on the manufacturing process, etc., and this is also within the technical scope of the present invention. Shall be included.

As described above, the procedure including two steps can be applied to various structures having two or more layers.

Although EMA is used in the examples, other effective medium approximations can be used and are also within the scope of the invention.

[Brief description of drawings]

FIG. 1 is a schematic view showing a configuration of an ellipsometer used in the method of the present invention, in which a part of a sample (sample) 4 to be measured is enlarged and shown.

FIG. 2 is a flow chart showing an embodiment of a thin film measuring method according to the present invention.

FIG. 3 is a detailed explanatory diagram showing a first analysis step in the embodiment.

FIG. 4 is a detailed explanatory diagram showing a second analysis stage in the embodiment.

FIG. 5 is a detailed explanatory diagram showing a third stage of analysis in the embodiment.

[Explanation of symbols]

1 Xe lamp 2 Optical fiber 3 Polarizer 4 Sample 5 Photoelastic modulator (PEM) 6 Analyzer 7 Optical fiber 8 Spectrometer 9 Data acquisition unit 31 Analysis for preparing multiple sets of models First step First step 32 Fitting Analysis to be performed 1st step 2nd step 33 Analysis to select the result of the model with the lowest χ ² value from the result of fitting 1st step 3rd step 41 Analysis to set the result as the initial value of the new model 2nd step 1st step 42 Analysis of performing BLMC on the other layer for each film thickness of one layer 2nd step 2nd step 43 Analysis 2nd step of selecting the model with the lowest χ ² value from the fitting result 3rd step 51 Analysis 3rd step 1st step 52 Final fitting of the selected model 1st step 52 Analysis 3rd step 3rd step Step 53 third stage third step analysis for saving

   ─────────────────────────────────────────────────── ─── Continued front page    F term (reference) 2F065 AA30 BB01 BB17 CC17 CC31                       FF50 GG03 GG24 HH10 HH12                       JJ01 JJ08 LL00 LL02 LL33                       LL34 LL67 NN08 QQ17 QQ18                       QQ41                 2G020 AA03 AA04 AA05 BA02 BA18                       CA15 CB04 CB32 CB42 CB43                       CC01 CD03 CD12 CD36 CD37                 2G059 AA02 AA03 BB08 BB10 BB16                       EE02 EE05 EE12 FF06 GG10                       HH01 HH02 HH03 JJ01 MM01                       MM10

Claims (6)

[Claims] 1. A measurement spectrum Ψ _E (λ _i ), which is a change in polarization of incident light and reflected light for each wavelength λ _i by changing the wavelength of incident light in an ultrathin film two-layer structure on the surface of a substrate to be measured. And Δ _E (λ _i ) to obtain Ψ _E and Δ _E spectrum measurement steps, and (N ₀ , (n _0, k ₀ )) of the substrate having a two-layer structure of an ultrathin film and materials (Mat1, Mat2) of each thin film. Considering possible complex refractive indices (N ₁ , (n _1, k ₁ )), (N ₂ , (n _2, k ₂ )), and the film thickness (d _1, d ₂ ) of each layer, several models are used. The first step, a second step of performing fitting with the measurement spectrum for each model, and a model having the lowest χ ² value as a result of fitting for each model, or
The result of the model with the lowest χ ² within the maximum and minimum values of the preset film thickness (d _{1 (best),} d
_{2 (best)} ) the first step of analysis comprising a third step, and the first step of setting the result obtained in the first step of analysis as an initial value of a new model, any of the set results Thickness of one layer (d
_{1 (best)} or d _{2 (best)} ) as a center value, and the other layer (d _{2 (best)} or d _{2
1 (best)} ) the second step of fitting using BLMC, and the minimum χ ² value based on the results of BLMC performed at the plurality of points, or a preset film thickness, a dispersion formula parameter,
The second stage of analysis, which consists of the third step of selecting the model with the lowest χ ² within the maximum and minimum incident angles, respectively, and the final results using the results obtained in the second stage of analysis A first step of performing a fitting, a second step of confirming the result obtained by the fitting, and a third step of analysis including a third step of storing the result, and a pole using a spectroscopic ellipsometer composed of: Method for analyzing thin film two-layer structure. 2. The method for analyzing an ultrathin film two-layer structure using the spectroscopic ellipsometer according to claim 1, wherein in the second step of the second step of the analysis, the layer to be fitted using BLMC is determined as two layers. A method for analyzing an ultra-thin film two-layer structure using a spectroscopic ellipsometer in which the material of the structure has a layer whose optical constant is unknown. 3. The method for analyzing an ultrathin film two-layer structure using the spectroscopic ellipsometer according to claim 1 or 2, wherein the material of the two-layer structure whose optical constant is better known is analyzed in the first step, the third step. With the film thickness obtained in the step as the central value, within the range of several% to several tens% before and after that,
The second analysis step and the second analysis step in which BLMC is performed on the other layer for each set film thickness.
Analysis that also includes the third step (Extend)
d BLMC) for analyzing an ultrathin film two-layer structure. 4. A measurement spectrum Ψ _E (λ _i ) which is a change in polarization of incident light and reflected light for each wavelength λ _i by changing the wavelength of the incident light in the ultrathin film two-layer structure on the surface of the substrate to be measured. And Ψ _E , Δ _E spectrum measurement step to obtain Δ _E (λ _i ), and 1 or 2 layers are non-uniform or discontinuous, some materials are mixed, and effective medium approximation is used And a model of an ultrathin film two-layer structure (N ₀ , (n _0, k
₀ )) and the possible complex refractive indices (N ₁ , (n _1, k ₁ )), (N ₂ ,, n of the materials (Mat1, Mat2) of each thin film.
_2, k ₂ )), mixing ratio (Vf ₁ , Vf ₂ ), film thickness of each layer (d _1,
d ₂ ), a first step of making several models, a second step of fitting the measured spectrum for each model, and a fitting result for each model, the minimum χ ² value Model or the model with the lowest χ ² within the maximum and minimum values of the preset film thickness and mixing ratio (d
_{1 (best),} d _{2 (best),} Vf _(best) ) is determined in the first step of the analysis, and the result of the film thickness / mixing ratio obtained in the first step of the analysis is calculated as a new model. Based on the film thickness value obtained in the first step and the third step of the analysis of the layer having the unknown dispersion formula as the initial value, the film thickness within the expected range ((d ₁ ± mΔd ₁ )
Or (d ₂ ± mΔd ₂ )) is set, and the film thickness of the other layer is centered on the value obtained in the first step of the analysis and the third step, and a plurality of points ((d ₂ ± mΔd
₂ ) or (d ₁ ± mΔd ₁ )) is set, and multiple points (Vf ± mΔVf) around it are set with the value of the mixing ratio obtained in the first step of the analysis and the third step as the central value. First step, multiple points of mixing ratio (Vf ± mΔVf) and multiple points of film thickness of another layer ((d ₂ ± mΔd ₂ ))
Or in combination with (d ₁ ± mΔd ₁ )),
BLMC analysis for layers containing unknown dispersion equations, second step, second step, results obtained for each combination of mixing ratio and film thickness of the other layer of unknown dispersion equations The lowest χ ² value or a combination having the lowest χ ² value that falls between the maximum and minimum values of the preset film thickness, the dispersion parameter, the mixing ratio, and the incident angle, respectively. Based on the values obtained in the second step of the analysis and the third step of the second step of the analysis, both film thicknesses, mixture ratios, and dispersion-type fittings, or
An analysis third step including a first step of an analysis third step for fitting both film thicknesses and a mixture ratio, a second step for confirming the result obtained by the fitting, and a third step for storing the result. Method for analyzing an ultra-thin film two-layer structure using a spectroscopic ellipsometer composed of. 5. The spectroscopic ellipsometer according to claim 1 or 4.
In the analysis method of the ultra-thin film two-layer structure using The ultra-thin film two-layer structure on the surface of the substrate to be measured
Each wavelength λ_i Change of polarization of incident light and reflected light for each
Is the measured spectrum Ψ_E (λ_i ) And Δ_E (λ _i )
Ψ_E , Δ_E Spectrum measurement stage, For each of multiple measurement conditions (Zi) within the expected range
Performing the analysis second stage to the third step of claims 1 and 4,
The lowest χ among the results obtained under each measurement condition² Value
Or the dispersion formula parameters and the mixing ratio Vf are set to the maximum,
Among the combinations that fall between the minimum values, χ² Maximum value
The second step and the fourth step of selecting a good one
2 layers of ultra-thin film using spectroscopic ellipsometer
Structure analysis method. 6. A method of analyzing an ultrathin film two-layer structure using the spectroscopic ellipsometer according to claim 1, 2, 3, 4 or 5, wherein the difference in the first, second, and third analysis steps is the smallest. The step of selecting is to find the mean squared error of the fitted value and the measured value, and obtain the one with the smallest mean squared error or the maximum of the preset film thickness, dispersion formula parameter, mixing ratio, incident angle variable , A method of analyzing an ultrathin film two-layer structure using a spectroscopic ellipsometer, which is to determine the one having the smallest mean square error within the minimum value.