Method for conducting optical measurement usingfull Mueller matrix ellipsometer

Abstract

A method for conducting optical measurement using a full Mueller matrix ellipsometer, which belongs to the technical field of optical measurements. The optical measurement method comprises: constructing an experimental optical path of a full Mueller matrix ellipsometer; conducting complete regression calibration on the full Mueller matrix ellipsometer; placing a sample to be measured on a sample platform, and obtaining an experimental Fourier coefficient of the sample to be measured; and according to the experimental Fourier coefficient of the sample to be measured, obtaining information about the sample to be measured. Since a calibration method for the full Mueller matrix ellipsometer is not only simple in operation process, but also makes full use of data of the full Mueller matrix ellipsometer measured at the same time, the introduced error is relatively small and the parameter obtained by calibration is more accurate, so that the measurement result is more accurate when the sample to be measured is measured. Thus, the process of the optical measurement method is simplified.

Claims

1. A method for conducting optical measurement with a full Mueller matrix ellipsometer, comprising the following steps: constructing an experimental optical path of the full Mueller matrix ellipsometer, the experimental optical path of the full Mueller matrix ellipsometer including a light source, a polarizer, a first phase compensator, an analyzer, a second phase compensator, a spectrometer, and a sample stage; performing a total regression calibration on operating parameters of the full Mueller matrix ellipsometer; placing a sample to be tested on the sample stage, and obtaining experimental Fourier coefficients of the sample to be tested with the full Mueller matrix ellipsometer; obtaining information of the sample to be tested based on the experimental Fourier coefficients of the sample to be tested; wherein a method for calibrating the full Mueller matrix ellipsometer comprises the following steps: setting rotational speeds of the first phase compensator and the second phase compensator; setting a frequency of the spectrometer for measuring light intensity data, so that the spectrometer measures the light intensity data every T/N time, wherein a total of N sets of light intensity data are acquired, where N≥25, and T is a period of measurement; acquiring the light intensity data measured by the spectrometer; obtaining respective experimental Fourier coefficients α′.sub.2n, β′.sub.2n from N relation formulas between the light intensity data and experimental Fourier coefficients formed by the N sets of light intensity data, based on the light intensity data acquired by a data acquisition module of the spectrometer; obtaining respective theoretical Fourier coefficients α.sub.2n, β.sub.2n based on the respective experimental Fourier coefficients, an initial polarization angle C.sub.s1 of the first phase compensator and an initial polarization angle C.sub.s2 of the second phase compensator which have been calibrated; obtaining, by a phase retardation calculation module for the first phase compensator, a phase retardation δ.sub.1 of the first phase compensator based on the respective theoretical Fourier coefficients, a polarization angle P.sub.s of the polarizer and a polarization angle A.sub.s of the analyzer which have been calibrated, on the basis that a reference sample is isotropic and uniform; obtaining, by a phase retardation calculation module for the second phase compensator, a phase retardation δ.sub.2 of the second phase compensator, based on the respective theoretical Fourier coefficients, the polarization angle P.sub.s of the polarizer and the polarization angle A.sub.s of the analyzer which have been calibrated, on the basis that the reference sample is isotropic and uniform; obtaining accurate values of all operating parameters (d, θ, P.sub.s, A.sub.s, C.sub.s1, C.sub.s2, δ.sub.1, δ.sub.2) of the full Mueller matrix ellipsometer through least square fitting according to the relation formulas between the theoretical Fourier coefficients and the operating parameters, with (d, θ, P.sub.s, A.sub.s, C.sub.s1, C.sub.s2, δ.sub.1, δ.sub.2) being as variables, and with the initial polarization angle C.sub.s1 of the first phase compensator, the initial polarization angle C.sub.s2 of the second phase compensator, the polarization angle P.sub.s of the polarizer, the polarization angle A.sub.s of the analyzer, the phase retardation δ.sub.1 of the first phase compensator and the phase retardation δ.sub.2 of the second phase compensator, which have been calibrated, being as initial values, where d is a thickness of the reference sample, and θ is an angle at which light is incident on the reference sample.

2. The method of claim 1, further comprising the following steps: obtaining respective θ.sub.2n=tan.sup.−1(β′.sub.2n/α′.sub.2n) based on the respective experimental Fourier coefficients α′.sub.2n, β′.sub.2n; obtaining the initial polarization angle C.sub.s1 of the first phase compensator based on the respective θ.sub.2n; obtaining the initial polarization angle C.sub.s2 of the second phase compensator based on the respective θ.sub.2n; obtaining the polarization angle P.sub.s of the polarizer based on the respective θ.sub.2n; and obtaining the polarization angle A.sub.s of the analyzer based on the respective θ.sub.2n.

3. The method of claim 1, wherein N=25, and an experimental Fourier coefficient calculation module directly obtains the respective experimental Fourier coefficients α′.sub.2n, β′.sub.2n based on N relation formulas between the light intensity data and experimental Fourier coefficients formed by the N sets of light intensity data.

4. The method of claim 1, wherein N>25, and an experimental Fourier coefficient calculation module obtains the respective experimental Fourier coefficients α′.sub.2n, β′.sub.2n through least square method based on N relation formulas between the light intensity data and experimental Fourier coefficients formed by the N set of light intensity data.

5. The method of claim 1, wherein the light source is a broad spectrum light source, and the number of wavelengths of light which can be generated by the light source is N′, and the number of relation formulas between the theoretical Fourier coefficients and the operating parameters is 24×N′.

6. The method of claim 5, wherein the number of the reference samples which are isotropic and uniform is m, and the number of relation formulas between the theoretical Fourier coefficients and the operating parameters is 24×N′×m.

7. The method of claim 1, wherein the reference sample is a silicon dioxide film with silicon as a substrate.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 is a diagram showing an experimental optical path of a full Mueller matrix ellipsometer constructed in a method for conducting optical measurement with a full Mueller matrix ellipsometer according to an embodiment of the present invention;

(2) FIG. 2 is a logic block diagram of a method for conducting optical measurement with a full Mueller matrix ellipsometer according to an embodiment 1 of the present invention; and

(3) FIG. 3 is a logic block diagram of a method for conducting optical measurement with a full Mueller matrix ellipsometer according to an embodiment 2 of the present invention.

DETAILED DESCRIPTION OF THE INVENTION

(4) The present invention will be described in detail below in conjunction with the drawings and specific embodiments for the in-depth understanding of the invention.

Embodiment 1

(5) A method for conducting optical measurement with a full Mueller matrix ellipsometer according to Embodiment 1 of the present invention may comprise the following steps:

(6) step 1: referring to FIG. 1, constructing an experimental optical path of a full Mueller matrix ellipsometer. The experimental optical path may include a light source 1, an annular mirror 2, a pinhole 3, a first off-axis parabolic mirror 4, a polarizer 5, a first phase compensator 6, a first plane mirror 7, a sample stage 8, a second off-axis parabolic mirror 9, a third off-axis parabolic mirror 10, a second plane mirror 11, a second phase compensator 12, an analyzer 13, a fourth off-axis parabolic mirror 14, a spectrometer 15 and a terminal 16. An isotropic and uniform reference sample is carried on the sample stage 8. The experimental optical path of the full Mueller matrix ellipsometer, which can be self-calibrated through total regression, may have the following optical process:
S.sub.out=M.sub.AR(A′)R(−C.sub.2)M.sub.c2(δ.sub.2)R(C′.sub.2)×M.sub.s×R(−C′.sub.1)M.sub.c1(δ.sub.1)R(C′.sub.1)R(−P′)M.sub.pR(P)S.sub.in
that is,

(7) $\begin{array}{cc}\left[\begin{array}{c}{S}_A\\ S_B\\ 0\\ 0\end{array}\right]=\left[\begin{array}{cccc}1& 1& 0& 0\\ 1& 1& 0& 0\\ 0& 0& 0& 0\\ 0& 0& 0& 0\end{array}\right]\left[\begin{array}{cccc}1& 0& 0& 0\\ 0& \cos 2A& \sin 2A& 0\\ 0& -\sin 2A& \cos 2A& 0\\ 0& 0& 0& 1\end{array}\right]\hspace{1em}\left[\begin{array}{cccc}1& 0& 0& 0\\ 0& \cos C_2^{\prime }& -\sin C_2^{\prime }& 0\\ 0& \sin C_2^{\prime }& \cos 2C_2^{\prime }& 0\\ 0& 0& 0& 1\end{array}\right]\left[\begin{array}{cccc}1& 0& 0& 0\\ 0& 1& 0& 0\\ 0& 0& \cos {\delta }_2& \sin {\delta }_2\\ 0& 0& -\sin {\delta }_2& \cos {\delta }_2\end{array}\right]\hspace{1em}\left[\begin{array}{cccc}1& 0& 0& 0\\ 0& \cos C_2^{\prime }& \sin C_2^{\prime }& 0\\ 0& -\sin C_2^{\prime }& \cos 2C_2^{\prime }& 0\\ 0& 0& 0& 1\end{array}\right]\left[\begin{array}{cccc}{M}_{11}& M_{12}& M_{13}& M_{14}\\ M_{21}& M_{22}& M_{23}& M_{24}\\ M_{31}& M_{32}& M_{33}& M_{34}\\ M_{41}& M_{42}& M_{43}& M_{44}\end{array}\right]\hspace{1em}\left[\begin{array}{cccc}1& 0& 0& 0\\ 0& \cos C_1^{\prime }& -\sin C_1^{\prime }& 0\\ 0& \sin C_1^{\prime }& \cos 2C_1^{\prime }& 0\\ 0& 0& 0& 1\end{array}\right]\left[\begin{array}{cccc}1& 0& 0& 0\\ 0& 1& 0& 0\\ 0& 0& \cos {\delta }_1& \sin {\delta }_1\\ 0& 0& -\sin {\delta }_1& \cos {\delta }_1\end{array}\right]\hspace{1em}\left[\begin{array}{cccc}1& 0& 0& 0\\ 0& \cos C_1^{\prime }& \sin C_1^{\prime }& 0\\ 0& -\sin C_1^{\prime }& \cos 2C_1^{\prime }& 0\\ 0& 0& 0& 1\end{array}\right]\left[\begin{array}{cccc}1& 0& 0& 0\\ 0& \cos 2P& -\sin 2P& 0\\ 0& \sin 2P& \cos 2P& 0\\ 0& 0& 0& 1\end{array}\right]\hspace{1em}\left[\begin{array}{cccc}1& 1& 0& 0\\ 1& 1& 0& 0\\ 0& 0& 0& 0\\ 0& 0& 0& 0\end{array}\right]\left[\begin{array}{cccc}1& 0& 0& 0\\ 0& \cos 2P& \sin 2P& 0\\ 0& -\sin 2P& \cos 2P& 0\\ 0& 0& 0& 1\end{array}\right]\left[\begin{array}{c}1\\ 0\\ 0\\ 0\end{array}\right]& 2.0\end{array}$

(8) Step 2 of the method may comprise: performing a total regression calibration on the full Mueller matrix ellipsometer.

(9) Step 3 of the method may comprise: placing the sample to be tested on the sample stage, and obtaining experimental Fourier coefficients of the sample to be tested with the full Mueller matrix ellipsometer.

(10) Step 4 of the method may comprise: obtaining information of the sample to be tested based on the experimental Fourier coefficients of the sample to be tested.

(11) The experimental Fourier coefficients have relation to Mueller elements of the sample, an azimuth angle P.sub.s of the polarizer, an azimuth angle A.sub.s of the analyzer, azimuth angles C.sub.s1 and C.sub.s2 of the two phase compensators, and phase retardations δ.sub.1 and δ.sub.2 (refer to Harland G. Tompkins, Eugene A. Irene, Handbook of ellipsometry, 7.3.3 Dual Rotating Compensator 7). However, the Mueller elements of the sample are related to optical constants n, k of a material of the sample, a thickness d, an angle θ at which light beams are incident on the sample, and a wavelength λ of the light beams. Therefore, after the experimental Fourier coefficients of the sample are measured, the Mueller elements of the sample can be obtained according to the above relationships, and then the information of the sample can be obtained.

(12) In an embodiment, the process of performing a total regression calibration on the full Mueller matrix ellipsometer comprises the following steps:

(13) step 21: setting rotational speeds of the first and second phase compensators;

(14) step 22: setting a frequency of a spectrometer for measuring light intensity data, so that the spectrometer may measure the light intensity data every T/N time, and a total of N sets of light intensity data are acquired, wherein N≥25, and T is a period of measurement;

(15) step 23: acquiring the light intensity data measured by the spectrometer;

(16) step 24: obtaining respective experimental Fourier coefficients α′.sub.2n/β′.sub.2n from N relation formulas between the light intensity data and the experimental Fourier coefficients formed by the N sets of light intensity data, based on the light intensity data acquired by a data acquisition module of the spectrometer;

(17) step 25: obtaining respective theoretical Fourier coefficients α.sub.2n, β.sub.2n according to the respective experimental Fourier coefficients, an initial polarization angle C.sub.s1 of the first phase compensator and the initial polarization angle C.sub.s2 of the second phase compensator which have been calibrated;

(18) step 26: obtaining, on the basis that a reference sample is isotropic and uniform, by a phase retardation calculation module for the first phase compensator, phase retardation δ.sub.1 of the first phase compensator based on the respective theoretical Fourier coefficients, a polarization angle P.sub.s of the polarizer and a polarization angle A.sub.s of the analyzer which have been calibrated; and

(19) obtaining, on the basis that the reference sample is isotropic and uniform, by a phase retardation calculation module for the second phase compensator, phase retardation δ.sub.2 of the second phase compensator based on the respective theoretical Fourier coefficients, a polarization angle P.sub.s of the polarizer and a polarization angle A.sub.s of the analyzer which have been calibrated;

(20) step 27: obtaining accurate values of all operating parameters (d, θ, P.sub.s, A.sub.s, C.sub.s1, C.sub.s2, δ.sub.1, δ.sub.2) of the full Mueller matrix ellipsometer through least square fitting according to the relation formulas between the theoretical Fourier coefficients and the operating parameters, with (d, θ, P.sub.s, A.sub.s, C.sub.s1, C.sub.s2, δ.sub.1, δ.sub.2) being as variables, and with the initial polarization angle C.sub.s1 of the first phase compensator, the initial polarization angle C.sub.s2 of the second phase compensator, the polarization angle P.sub.s of the polarizer, the polarization angle A.sub.s of the analyzer, the phase retardation δ.sub.1 of the first phase compensator, the phase retardation δ.sub.2 of the second phase compensator, which have been calibrated, being as initial values, where d is a thickness of the reference sample, and θ is an angle at which light is incident on the reference sample.

(21) A corresponding Mueller matrix of the reference sample that is isotropic and uniform is:

(22) $M_s=\left[\begin{array}{cccc}{M}_{11}& M_{12}& 0& 0\\ M_{21}& M_{22}& 0& 0\\ 0& 0& M_{33}& M_{34}\\ 0& 0& M_{43}& M_{44}\end{array}\right]$

(23) Taking N=36, the ratio of the rotation speed of the first phase compensator 6 to the rotation speed of the second phase compensator 12 equaling 5:3 as an example, at this time, the first phase compensator 6 and the second phase compensator 12 are respectively in a rotating state, and the ratio of the rotation speed of the first phase compensator 6 to the rotation speed of the second phase compensator 12 equals 5:3. At this time, C′.sub.1=5(C−C.sub.s1), C′.sub.2=3(C−C.sub.s2), and the time during which the first phase compensator 6 rotates 5 turns or the second phase compensator 12 rotates 3 turns is a period T, and where:

(24) −C.sub.s1, an angle of a fast optical axis of the first phase compensator 6 at a time t=0,

(25) −C.sub.s2, an angle of a fast optical axis of the second phase compensator 12 at a time t=0,

(26) C=ωt, a rotation angle by which the first phase compensator 6 and the second phase compensator 12 rotate at a fundamental physical frequency ω.

(27) $\begin{array}{cc}{S}_j={\int }_{\frac{\left(j-1\right)T}{36}}^{\frac{\mathrm{jT}}{36}}{I}_0^{\prime }\left[1+\underset{n=1}{\overset{16}{.Math.}}\left({\alpha }_{2n}^{\prime }\cos 2n\omega t+{\beta }_{2n}^{\prime }\sin 2n\omega t\right)\right]\mathrm{dt}=\frac{\pi I_0^{\prime }}{36\omega }+\underset{n=1}{\overset{16}{.Math.}}\frac{I_0^{\prime }}{n\omega }\left(\sin \frac{n\pi }{36}\right)\left[{\alpha }_{2n}^{\prime }\cos \frac{\left(2j-1\right)n\pi }{36}+{\beta }_{2n}^{\prime }\sin \frac{\left(2j-1\right)n\pi }{36}\right]\left(j=1,2\mathrm{.Math.}36\right)& 2.1\end{array}$

(28) where ω=π/T.

(29) With the acquired S1, S2, S3 . . . S36, 36 equations containing 25 unknowns can be obtained through the above formula (n=9, 12, 14, 15, the primed Fourier coefficients α′.sub.2n=0 and β′.sub.2n=0). Through a nonlinear least square method, a total of 24 primed Fourier coefficients α′.sub.2n and β′.sub.2n can be obtained.

(30) The transformation relationship between the theoretical Fourier coefficients α.sub.2n and β.sub.2n and the experimental Fourier coefficients α′.sub.2n and β′.sub.2n is expressed in formulas 2.7 and 2.8:
α.sub.2n=α′.sub.2n cos ϕ.sub.2n+β′.sub.2n sin ϕ.sub.2n  2.7
β.sub.2n=−α′.sub.2n sin ϕ.sub.2n+β′.sub.2n cos ϕ.sub.2n  2.8
where:

(31) ϕ.sub.2=12C.sub.s2−10C.sub.s1; ϕ.sub.4=10C.sub.s1−6C.sub.s2;

(32) ϕ.sub.6=6C.sub.s2; ϕ.sub.8=20C.sub.s1−12C.sub.s2;

(33) ϕ.sub.10=10C.sub.s1; ϕ.sub.12=12C.sub.s2;

(34) ϕ.sub.14=20C.sub.s1−6C.sub.s2; ϕ.sub.16=10C.sub.s1+6C.sub.s2;

(35) ϕ.sub.20=20C.sub.s1; ϕ.sub.22=10C.sub.s1+12C.sub.s2;

(36) ϕ.sub.26=20C.sub.s1+6C.sub.s2; ϕ.sub.32=20C.sub.s1+12C.sub.s2

(37) The theoretical Fourier coefficients α.sub.2n and β.sub.2n can be obtained from the formulas 2.7 and 2.8.

(38) For a sample which is isotropic and uniform, M.sub.13=M.sub.31=M.sub.14=M.sub.41=M.sub.23=M.sub.32=M.sub.24=M.sub.42=0. Further, according to theoretical principles of the Mueller ellipsometer, the following theoretical expressions for the theoretical Fourier coefficients α.sub.2, β.sub.2, α.sub.10, β.sub.10, α.sub.6, β.sub.6, α.sub.14, β.sub.14, α.sub.22, β.sub.22, α.sub.26, β.sub.26 can be obtained:

(39) $\begin{array}{cc}{\alpha }_6=M_{43}{\cos }^2\frac{{\delta }_1}{2}\sin {\delta }_2\sin 2P_s\sin 2A_s& 2.9\\ {\beta }_6=-M_{43}{\cos }^2\frac{{\delta }_1}{2}\sin {\delta }_2\sin 2P_s\cos 2A_s& 2.10\\ {\alpha }_{14}=-0.5M_{43}{\sin }^2\frac{{\delta }_1}{2}\sin {\delta }_2\cos 2\left(P_s-A_s\right)& 2.11\\ {\beta }_{14}=-0.5M_{43}{\sin }^2\frac{{\delta }_1}{2}\sin {\delta }_2\sin 2\left(P_s-A_s\right)& 2.12\\ {\alpha }_{26}=0.5M_{43}{\sin }^2\frac{{\delta }_1}{2}\sin {\delta }_2\cos 2\left(P_s+A_s\right)& 2.13\\ {\beta }_{26}=0.5M_{43}{\sin }^2\frac{{\delta }_1}{2}\sin {\delta }_2\sin 2\left(P_s+A_s\right)& 2.14\end{array}$

(40) From formulas 2.10 and 2.12, it can be obtained:

(41) $\begin{array}{cc}\frac{{\beta }_{14}}{{\beta }_6}=\frac{0.5\sin 2\left(P_s-A_s\right)}{\sin 2P_s\cos 2A_s}\times {\tan }^2\frac{{\delta }_1}{2}{\delta }_1={\cos }^{-1}\left[\frac{1-\frac{2{\beta }_{14}\sin 2P_S\cos 2A_S}{{\beta }_6\sin 2\left(P_S+A_S\right)}}{1+\frac{2{\beta }_{14}\sin 2P_S\cos 2A_S}{{\beta }_6\sin 2\left(P_S+A_S\right)}}\right]\mathrm{Where}\left(P_s+A_s\right)\ne \frac{n\pi }{2}\mathrm{and}{P}_s\ne \frac{n\pi }{2}\mathrm{and}{A}_s\ne \frac{\pi }{4}+\frac{n\pi }{2}& 2.15\end{array}$
(n is an integer) (the Fourier coefficients must be guaranteed to be non-zero)

(42) Similarly, the phase retardation δ.sub.1 of the compensator can also be calculated through formulas 2.9 and 2.12, formulas 2.9 and 2.11, formulas 2.10 and 2.11, formulas 2.9 and 2.13, formulas 2.9 and 2.14, formulas 2.10 and 2.13, formulas 2.10 and 2.14.

(43) The phase retardation δ.sub.2 of the second compensator is calibrated below.

(44) $\begin{array}{cc}{\alpha }_2=0.5M_{34}\sin {\delta }_1{\sin }^2\frac{{\delta }_2}{2}\cos 2\left(P_s-A_s\right)& 2.16\\ {\beta }_2=-0.5M_{34}\sin {\delta }_1{\sin }^2\frac{{\delta }_2}{2}\sin 2\left(P_s-A_s\right)& 2.17\\ {\alpha }_{10}=-M_{34}\sin {\delta }_1{\cos }^2\frac{{\delta }_2}{2}\sin 2P_s\sin 2A_s& 2.18\\ {\beta }_{10}=M_{34}\sin {\delta }_1{\cos }^2\frac{{\delta }_2}{2}\sin 2P_s\cos 2A_s& 2.19\\ {\alpha }_{22}=-0.5M_{34}\sin {\delta }_1{\sin }^2\frac{{\delta }_2}{2}\cos 2\left(P_s+A_s\right)& 2.20\\ {\beta }_{22}=-0.5M_{34}\sin {\delta }_1{\sin }^2\frac{{\delta }_2}{2}\sin 2\left(P_s+A_s\right)& 2.21\end{array}$

(45) From formulas 2.16 and 2.18, it can be obtained:

(46) $\begin{array}{cc}\frac{{\alpha }_2}{{\alpha }_{10}}=-\frac{0.5\cos 2\left(P_s-A_s\right)}{\sin 2P_s\sin 2A_s}\times {\tan }^2\frac{{\delta }_2}{2}{\delta }_2={\cos }^{-1}\left[\frac{1+\frac{2{\alpha }_2\sin 2P_s\sin 2A_s}{{\alpha }_{10}\cos 2\left(P_s-A_s\right)}}{1-\frac{2{\alpha }_2\sin 2P_s\sin 2A_s}{{\alpha }_{10}\cos 2\left(P_s-A_s\right)}}\right]\mathrm{where}\left(P_s-A_s\right)\ne \frac{\pi }{4}+\frac{n\pi }{2}\mathrm{and}{P}_s\ne \frac{n\pi }{2}\mathrm{and}{A}_s\ne \frac{n\pi }{2}& 2.22\end{array}$
(n is an integer) (the Fourier coefficients must be guaranteed to be non-zero)

(47) Similarly, the phase retardation δ.sub.2 of the compensator can also be calibrated through formulas 2.16 and 2.19, formulas 2.17 and 2.18, formulas 2.17 and 2.19, formulas 2.18 and 2.20, formulas 2.18 and 2.21, formulas 2.19 and 2.20, formulas 2.19 and 2.21.

(48) The transformation of the experimental Fourier coefficients and the theoretical Fourier coefficients can be realized by the formulas 2.7 and 2.8. Meanwhile, the theoretical Fourier coefficients have relation to Mueller elements of the sample, an azimuth angle P.sub.s of the polarizer, an azimuth angle A.sub.s of the analyzer, azimuth angles C.sub.s1 and C.sub.s2 of the two phase compensators, and the phase retardations δ.sub.1 and δ.sub.2 (refer to Harland G. Tompkins, Eugene A. Irene, Handbook of ellipsometry, 7.3.3 Dual Rotating Compensator 7). Mueller elements of the sample are related to optical constants n, k of a material of the sample, thickness d, an angle θ at which light beams are incident on the sample, and a wavelength λ of the light beams. The experimental Fourier coefficients α′.sub.2n and β′.sub.2n are related to (n, k, d, θ, λ, P.sub.s, A.sub.s, C.sub.s1, C.sub.s2, δ.sub.1, δ.sub.2), θ is an angle at which light beams are incident on the sample. For a reference sample with known optical constants n, k under a single wavelength, there are 24 α′.sub.2n and β′.sub.2n in total obtained by measurement in an experiment, and correspondingly 24 equations can be obtained, which are related only to (d, θ, λ, P.sub.s, A.sub.s, C.sub.s1, C.sub.s2, δ.sub.1, δ.sub.2). The P.sub.s, A.sub.s, C.sub.s1, C.sub.s2, δ.sub.1, δ.sub.2 obtained above by calibration can be used as initial values, and the wavelength corresponding to the measurement in the experiment are known; 24 equations obtained according to experimental Fourier coefficients have relation to (d, θ, P.sub.s, A.sub.s, C.sub.s1, C.sub.s2, δ.sub.1, δ.sub.2). Thus, the remaining operating parameters (d, θ, P.sub.s, A.sub.s, C.sub.s1, C.sub.s2, δ.sub.1, δ.sub.2) of the Mueller ellipsometer can be obtained by the least square fitting. The reference sample may be a silicon dioxide film sample with silicon as the substrate, the optical constants n and k of which can be consulted in the literatures. Taking the wavelength of 632.8 nm as an example, the optical constants of the reference sample are n=1.457, and k=0.

(49) When N=25, an experimental Fourier coefficient calculation module directly obtains the respective experimental Fourier coefficients α′.sub.2n, β′.sub.2n based on N relation formulas between light intensity data and experimental Fourier coefficients formed by the N sets of light intensity data.

(50) When N>25, the experimental Fourier coefficient calculation module obtains the respective experimental Fourier coefficients α′.sub.2n, β′.sub.2n by the least square method according to N relation formulas between light intensity data and experimental Fourier coefficients formed by the N sets of light intensity data.

(51) The light source may be a broad spectrum light source. The number of wavelengths of light which can be generated by the light source is N′, and the number of relation formulas between theoretical Fourier coefficients and operating parameters may be 24×N′.

(52) The number of the reference samples which are isotropic and uniform may be m, and the number of relation formulas between theoretical Fourier coefficients and operating parameters may be 24×N′×m.

Embodiment 2

(53) Referring to FIG. 3, the difference between the total regression self-calibrated full Mueller matrix ellipsometer according to the embodiment 2 of the present invention and the total regression self-calibrated full Mueller matrix ellipsometer according to the embodiment 1 of the present invention lies in that: the total regression self-calibration method of the total regression self-calibrated full Mueller matrix ellipsometer according to the embodiment 2 of the present invention may further comprise the following steps:

(54) obtaining respective θ.sub.2n based on respective experimental Fourier coefficients α′.sub.2n, β′.sub.2n, where θ.sub.2n is an intermediate parameter defined for the convenience of calculation;

(55) obtaining an initial polarization angle C.sub.s1 of a first phase compensator based on the respective θ.sub.2n;

(56) obtaining an initial polarization angle C.sub.s2 of a second phase compensator based on the respective θ.sub.2n;

(57) obtaining a polarization angle P.sub.s of a polarizer based on the respective θ.sub.2n;

(58) obtaining a polarization angle A.sub.s of an analyzer based on the respective θ.sub.2n;

(59) where
θ.sub.2n=tan.sup.−1(β′.sub.2n/α′.sub.2n)  2.2

(60) Using the method available in the literature (R. W. Collins and JoohyunKoh Dual rotating-compensator multichannel ellipsometer: instrument design for real-time Mueller matrix spectroscopy of surfaces and films Vol. 16, No. 8/August 1999/J. Opt. Soc. Am. A 1997 to 2006), which corresponds to the following formulas 2.3 to 2.6, the initial polarization angles Cs1 and Cs2 of the compensators, as well as the polarization angles P.sub.s and A.sub.s of the polarizer and the analyzer can be calibrated.

(61) $\begin{array}{cc}{C}_{s1}=\frac{{\theta }_{14}-{\theta }_4}{10}& 2.3\\ C_{s2}=\frac{{\theta }_2+{\theta }_4}{6}& 2.4\\ P_s=\left(\frac{{\theta }_4+{\theta }_6}{4}\right)-5C_{s1}& 2.5\\ \mathrm{As}=\left(\frac{{\theta }_{16}-{\theta }_4}{4}\right)-3C_{s2}& 2.6\end{array}$

(62) On the basis of calibrated P.sub.s, A.sub.s, C.sub.s1, and C.sub.s2, in the case that the compensators are not disassembled from the experimental stage or equipment for separate measurement, the method we proposed can calibrate phase retardations of both compensators under different wavelengths in one experiment. The calibration process is accurate and simple.

(63) The above embodiments describe the objects, the technical solutions and advantages of the present invention in detail. However, it should be appreciated that the foregoing is only specific embodiments of the present invention rather than limiting the invention. Therefore, any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.