JPS63148108A - Ellipsometer - Google Patents

Abstract

PURPOSE:To take a measurement in a short time with a high S/N and high accuracy by making linear polarized light incident on a sample and finding the amplitude reflec tion coefficient ratio and phase difference of elliptic polarized light, and finding the refractive index of the sample surface and the thickness of a thin film according to the found ratio and difference. CONSTITUTION:Light from a white light source 10 is made incident on a mono chromater 11 to select homogeneous light with wavelength lambda, which is made incident on a linear polarizing element 12 to become linear polarized light in a specific polariza tion direction and then made incident on the sample 1 at an angle psi0 of incidence. Is reflected light is elliptic polarized light and made incident on a phase modulating element 13. The sample reflected light which is polarized and modulated by the element 13 is made incident on a photodetector 15 through an analyzer 14. The output signal of the detector 15 is separated by a signal component separating circuit 16 into a DC component, a omega, and 2omega component, which are processed by an arithmetic unit 17 to find the reflection coefficient rp/rs and phase variation gamma, from which the reflec tion factor of the sample surface and the thickness of the thin film are found. Thus, measurements are taken in a short time with the high S/N and high accuracy.

Description

Detailed Description of the Invention Field of Industrial Application This invention is an ellipsometer that uses the polarization property of light to measure the thickness of a thin film on an object (sample surface) and the refractive index related to the physical properties of the film on the specimen surface. It is related to.

Conventional technology As shown in Fig. 1, if linearly polarized light is incident obliquely from above on the surface of a sample 1 having a thin film 2 on the surface at an incident angle ψ0, the reflected light will vary depending on the thickness and refractive index of the thin film on the sample surface. The polarization state of the light changes, and it is usually reflected as elliptically polarized light.

By measuring the amount of change in polarization of this reflected light and performing analytical calculations, it is possible to determine the thickness and refractive index of the thin film on the sample surface.This is called an ellipsometer, and the device is generally called an ellipsometer. It is called.

In such an ellipsometer, the important parameter of the amount of polarization change of reflected light necessary to determine the thickness and refractive index of a thin film is that the p component wave on the horizontal p coordinate plane and the perpendicular S coordinate plane due to reflection are Changes in the principal axis direction of polarized light due to the phase shift γ that occurs between the S component wave and the difference in amplitude of the image component wave that occurs due to the difference in reflectance between the P component wave and the S component wave. There is.

By the way, conventional ellipsometers can be roughly divided into two types: photometric type and extinction type. In the photometric type, a polarizing prism is continuously rotated and the phase shift γ and change in principal axis direction are calculated from the relationship between the angle ζ and the detected light intensity. On the other hand, in the extinction type, the polarized light that has changed in the sample is returned to its original state by rotating an optical element, and the phase shift γ and the change in principal axis direction 114/ are determined from the compensation angle. Both of these two methods obtain the necessary information by measuring the amount of change in light intensity, that is, the DC output of the photodetector.

Problems to be Solved by the Invention Among conventional ellipsometers, photometric type ellipsometers calculate the phase shift γ, which is a parameter for the amount of change in polarization, and the change in the principal axis direction. There is a problem that measurement accuracy deteriorates in areas close to .

On the other hand, the extinction type directly measures the polarization angle as an angle, so it has the advantage of achieving high measurement accuracy to the extreme of a polarizing prism. There is a long drawback.

A common drawback of both the photometric type and extinction type is that because the light intensity is detected as a DC component, they are directly affected by background light and have a poor signal-to-noise ratio (S/N).

Therefore, an object of the present invention is to provide an ellipsometer that can measure thin films and refractive indexes on objects with high precision and in a short time with high S/N without being affected by background light. .

Means for Solving the Problems In the ellipsometer of the present invention, basically linearly polarized light is incident on the sample, and the amplitude reflection coefficient ratio rp/rs (mi ρ = tanv) and phase difference ( (retardation) γ and its amplitude reflection coefficient ratio rp/
The birefringence of the sample and the thickness of the thin film are determined from rs and the phase difference γ.

In the ellipsometer of this invention, a phase modulation element is inserted into the optical path at +1 mark to modulate the incident light or reflected light to the sample, and the modulation frequency component is synchronized with the modulation frequency from the signal of the photodetector. , a frequency component twice the modulation frequency, and a DC component, and calculate the reflection coefficient ratio and phase difference from the size of each component, and finally calculate the birefringence value of the sample and/or the thickness of the thin film. I decided to find the value by calculation.

Specifically, the first invention provides an ellipsometer that makes linearly polarized light incident on a sample surface and measures a value related to the sample surface from the polarization state of the reflected light. A linear polarizing element is arranged so as to keep the polarization direction of the incident linearly polarized light constant, and a phase modulating element with an amplitude δ0 and a modulation angular frequency ω is placed in the optical path of the reflected light from the sample surface so that its slow axis is the incident plane. At the same time, an analyzer is arranged on the output side of the phase modulation element so that the transmission axis is parallel or perpendicular to the incident plane, and further on the output side of the analyzer. A photodetector for photoelectric conversion of light is arranged, a signal component separation circuit is provided to independently extract the ω component, 2ω component, and DC component of the output signal of the photodetector, and calculations are performed from each of the components. The method is characterized in that the reflection coefficient ratio rp/rs of the sample surface and the phase change γ of the reflected light are determined by the method, and the refractive index of the sample surface and/or the thickness of the thin film on the sample surface are determined based on the results. is.

Further, in the ellipsometer of the second invention, the arrangement of the phase modulation element and the sample is reversed from that of the first invention. That is, the ellipsometer of the second invention is an ellipsometer that makes linearly polarized light incident on a sample surface and measures a value related to the sample surface from the polarization state of the reflected light. A linear polarizing element is arranged to keep the orientation constant, and a phase modulating element with an amplitude δ0 and a modulation angular frequency ω is placed between the linear polarizing element and the sample surface, and its delay axis is set at 45° with respect to the incident plane. An analyzer is placed in the optical path of the reflected light from the sample surface so that the transmission axis is parallel or perpendicular to the incident surface, and the light is photoelectrically converted on the output side of the analyzer. A signal component separation circuit is installed to independently extract the ω component, 2ω component, and DC component of the output signal of the photodetector. This method is characterized in that the reflection coefficient ratio rp/rs and the phase change γ of reflected light are determined, and based on these, the refractive index of the sample surface and/or the thickness of the thin film on the sample surface are determined.

Function Before explaining the function of the ellipsometer of the present invention, the measurement principle of the ellipsometer will be explained first, and then the theoretical analysis of the ellipsometer of the present invention will be explained together with its function.

A: Ellipsometer measurement principle An ellipsometer measures the optical constant (refractive index) of the object itself by observing changes in the polarization state when light is reflected on the surface of the object.
Alternatively, there is a method of determining the thickness and optical constant (refractive index) of a thin film attached to the surface of an object. First, we will explain the principle of measuring the optical constants of the object itself, that is, the underlying material when there is no thin film, and then the principle of measuring the thickness and optical constants of the thin film when there is a thin film.

A-1: Base light 7' number? j! ! I''First, let the optical constant of the sample be 100=n-ik.

where blindness is the complex refractive index, n is the refractive index, k is the absorption coefficient,
1 is an imaginary unit.

The amplitude of the imaging component (P component) parallel to the plane of incidence of a monochromatic parallel light beam incident from vacuum at an angle of incidence ψ0 (Fresnel coefficient) is calculated by the amplitude of the imaging component (S component) perpendicular to the plane of incidence. Reflectance (Fresnel coefficient)′? ′″S, and these are respectively 7′″p = rp eXp(-iφp)
(1)'? ”s = r s eXt) (-iφ
s ) (2). These are functions of the optical constant 'fi=n-ik of the sample and the incident angle ψ0.

In a transparent sample with an absorption coefficient of =o, φP and φS are generally O.
Or π, therefore? Since p and Ts are real numbers, the incident linearly polarized light does not become elliptically polarized light but is reflected as linearly polarized light.

However, in absorber samples such as metals (kf = o), the phase difference (retardation) γ due to reflection, that is, γ = φP - φ5 (3), is continuous from O to π depending on the value of the incident angle ψ0. Generally speaking, '?"p/?"S is a complex number (?p
S and each S is also a complex number).

= tanveXp! γ = ρexp tγ (4), −V
ρ(=rp, "rs>" is called (film width reflectance ratio or amplitude reflection coefficient ratio (real number).

In this manner, since 7'''p/?s is a complex number in the absorber sample, the incident linearly polarized light is reflected as elliptically polarized light.

If we measure two parameters of the elliptically polarized light (for example, the azimuth angle α of the long axis of the ellipse and the ellipticity X), we can then calculate the amplitude reflectance ratio ρ; can. The following relational expression is known between these two quantities tanv and γ· and the refractive index n and absorption coefficient.

n2−2 Therefore, if we know the values of ψ and γ, we can calculate the refractive index of the sample n
Therefore, the absorption coefficient k can be obtained.

A-2: Determination of the fluorescence constant of a thin film As shown in Figure 2, a refractor with a refractive index h1 = n1-ik1 and a thickness d is placed on the base surface (assumed to be known) at 2 = n2-1. Assume that there is an isotropically homogeneous thin film 2, into which linearly polarized light is incident at an incident angle ψ0.

The reflected light R is a combination of the light R1 reflected on the surface of the thin film, the light R2 reflected on the interface between the thin film and the base, and the subsequent light R3 that comes out while reciprocating within the thin film. That is, R=R1+R2+R3+... (7) The amplitude reflectance of the entire surface, taking into account repeated reflection interference within the film, is -'? for the p component and the S component, respectively. "ts+'?"2sexp(-iδ)R3=
=R3e 'φS1+F''ts?''
2SeXE)(-iδ) (9).

Here, (j=1.2> n j−I 5illψj−1= ?'i j Sin
ψj(12)δ=4πn[jcosψ1/λ
(13) And is it JPN? ``JS is the amplitude reflectance (Fresnel coefficient) of the p and S components on the first surface (vacuum-film) when j=1 and on the second surface (film-base) when j=2. Further, δ is a phase difference caused by one round trip of the film width, and λ is a wavelength in vacuum.

Here, Rp/FEs is Fresnel coefficient such as TeJP or δ
Since is a complex number even if it is a real number, the reflected light becomes elliptically polarized light.

The complex reflection coefficient ratio Rp/Rs is calculated as =tanVeXD(i 7 > (
14) Hail.

Here, the values tan v, γ on the right side are the quantities measured with an ellipsometer, while the coefficient ratios on the left side are (6) to (1
1) As can be understood from the equation, Fit (nt , k
t), n2 (n2, to 2>, d.

It is a function of λ and ψ0. That is, γ, V = F (nl, n2.kt, 2.d, λ, ψ0)
Among the parameters on the right side of equation (15), n2, R2, λ,
If ψ0 is known and the measured value γ and force are used, nl and d@ can be solved as unknowns.

For example, if k1=O (transparent film), the unknowns are nl,
d, and its value can be easily determined using a computer. Even in the case of kl>Q (absorbing film), nl, kl, and d can be determined by measuring γ and ψ.

A method of calculating the required quantities n1, kl, and d from the measured quantities γ and ri is well known.

As described above, the thickness d of the thin film on the sample surface, the optical constant refractive index n1, and the absorption coefficient 1 can be obtained from the amplitude reflection coefficient ratio tanv (=ρ) and the phase difference force measured by an ellipsometer. I can do it.

[B; Today's Theoretical Analysis The output of the optical array of the ellipsometer of this invention (see FIGS. 3 and 4) will be analyzed using the Mueller matrix analysis method in the following steps.

First, with respect to the output signal of the photodetector, the components of the same frequency as the modulation frequency of the phase modulation element, the component of twice the frequency, and the DC component are derived. Next, these three components give the reflection coefficient ratio 1) (= rp /
We will derive how rs = tanV> and the phase difference γ are expressed. Based on this analysis, the relationship between the reflection coefficient ratio ρ, the phase difference γ, and the birefringence index is derived. lastly,
The thickness of the thin film is derived based on the above analysis results.

Next, these analysis procedures are divided into sections and described.

B-1: Relationship between each signal component and reflection coefficient ratio phase difference First, each optical element is expressed by a Mueller matrix.

The Mueller matrix A45 of the analyzer with the plane of polarization rotated by 45 degrees is expressed as follows.

Also, the Mueller matrix of the reflective surface of the sample for the 45° polarization plane can be calculated by giving the reflection coefficient ratio rp/r3, S (45
, rp , rs , 7) as S (45, rp
, rs, T). However, T is the difference in phase change (phase difference=retardation), and rp and rs are the reflection coefficients of the p component and the S component, respectively.

Mueller matrix of phase modulator M45. δ(ω) is δ
is the angular frequency (angular velocity) of modulation of the optical phase modulator
, the amplitude is expressed as δ=δo sinωt (19).

The Mueller matrix ■0 of linearly polarized light, roughly Oo, is expressed as .

The output I (d) of the photodetector in this case is (16
) to (20), expressed as a product of Mueller matrices.

I (d) = A+s・S (45, rp, r3.
γ)・M45δ(ω)・■.

Therefore, from equations (16) to (20) and (21), I
When (d) is obtained, the following equation is obtained.

I (d) Here, if S1nδ and COSδ are expanded using the Bessel function, Slnδ=s+n (δ□ s+nωt)=2J1
(δo) Sinωt+2J3 (δo)
s: n3ωt・ (23) Excuse δ-COS
(δO5illωt)=Jo(δo)+2J2(δ
o) cos2ωt+2J4(δo) cos4ωt 10-... is given.

If the electrical signal of the output of the photodetector proportional to the light intensity I (d) is V (d), then V (d) is the DC component VDC-
It can be expressed as follows by harmonic components of ω component V(ω), 2ω component V(2ω), and 3ω component V(3ω).

V (d) = V, c + V (ω) + V (2 ω) + (harmonic peak) In equations (23) and (24), Jo (δ
The term 0) corresponds to the DC component, the term Jr (δ0) corresponds to the ω component, and the term J2 (δ0) corresponds to the 2ω component. Therefore, from equations (22), (23), and (24), each component vDcSv(ω), °v(2ω) of equation (25)
If we calculate, V(DC)=rp2+r32+ (rp2-rs2)
・Jo (δo) (2B)■(ω) =4 rp
rs-s+nγ・Jl(δO) (27)
V(2ω) = 2 (rp2-rs2) ・J2 (δ
o) (28).

Next, the reflection coefficient ratio rp/rs and the value of s+nγ for the phase difference (retardation) γ are determined using equations (26), (27), and (28).

Add and subtract equations (26) from equation (28), equations (29) and (30) from equation (28), and obtain r p2, r
By deriving s2 and taking the ratio of the two, therefore, is derived. In other words, the reflection coefficient ratio rs/rp is a function of the direct current component V (DC) of the output of the photodetector and the 0 frequency component 9V (2ω) twice the modulation angular frequency.
rs/rp is determined from these components.

Next, find Sinγ.

Equation (29) is derived from Equation (27), and Equation (34) is derived from Equation (33). In other words, Si17 is a function of the modulation frequency ω component ■(ω) of the output of the photodetector, the twice its frequency 2ω component V(2ω), and ρ('=rp/rs). From these, the value of Sinγ is determined.

As described above, r's/rp can be calculated from equation (32) by S
1nγ can be determined from equation (35).

Next, the relationship between birefringence 'fi=n-ik is determined.

Then, equation (32) becomes Equation (35) is It can be rewritten as

By the way, for light incident on the sample surface at an incident angle of ψ0 (45°), the amplitude reflectance is different between the oscillating component of the electric field parallel to the plane of incidence (p component) and the oscillating component of the electric field perpendicular to the plane of incidence (S component). , each one-arrow reflectance is given by the Fresnel coefficient determined by the refractive index and angle of incidence of the two media bounded by the surfaces, respectively. Now let p and Fs be the film width reflectance for the p and S components.As mentioned above, these can generally be written as complex numbers as follows.

? p = rp e-'φ0(1) and 3 = r3 e-1φ3(2) In a transparent body, the refractive index is a real number, so φS and φP are O or π, rp and r3 are real numbers, and rp/rs is also a real number.

However, in absorbers such as metals, the refractive index is a complex number of hundreds = n-
Since it is expressed as ik, 7”p, 'F's becomes a complex number, =ρexp(iγ) (40)(40
) is the quantity to be measured, and the p and S components have different amplitude ratios and a relative phase difference γ, so linearly polarized light is reflected as elliptically polarized light.

Assuming that the angle of incidence on the sample is ψ0 and the angle of refraction is ψ1, the blind COSψ1 is calculated as n cosψ1= [(ncosψ1)2]i=[102
(1,-5uiφ1)]坏=[n2-'? 'i2s+dΦ1]i=[n2sat
! ψO] ' (42) Here, the following (43)
Using the law of refraction given by the formula % formula % (43), ′? ′″P Therefore, the theoretical formula was found from E formula (45).

Therefore, the relationship between the reflectance coefficient ratio "?'"p/"i"s, the phase change difference (retardation) T, and the refractive index is determined. Therefore, for convenience of calculation, P and Q defined by the following equation (46) are set.

By setting P and Q in this way, it becomes possible to analytically solve the relationship between P and Q and the refractive index. Therefore, in order to find the relationship between P, Q, and the refractive index, first, from equations (45) and (42), (
46) Find the left side of the equation.

=ρe1γ...(Measured quantity> (
47) 1-ρerγ Sinψ0-ψo
(48) or (51), the real part corresponding to P cos Q and P
When the imaginary parts corresponding to sin Q are respectively derived and divided, 2ρ sin r tllnQ = −(52) 1−ρ2 Therefore, equation (49) is expressed as s+tlψotan(1)o (53).

That is, when the incident angle ψo=45°, p2cos2Q=2 (2 for n2-) -1(56)P
2s+n2Q=-4n K (57)
Therefore, the relationship between the refractive index n1 absorption coefficient and P and Q for the complex refractive index 5 (=n-ik) is expressed by equation (58) and (59)
It was determined by the formula.

However, as already mentioned from equations (50) and (52), P2 and tanQ are the DC component v (DC), ω component v (ω), and 2ω of the output signal of the photodetector from equation (32). Component V (
2ω), r3/rp can be found. −
From equation (35), the ω component V (ω), the 2ω component V (2
If ω) and rs/rp are known, S1nγ can be found.

Zarani r3/rp.

Once s+nγ is determined, P2 and tanQ are determined from equations (60) and (61). If P and Q are found,
The refractive index is determined from equations (58) and (59).

Here, by substituting equation (62) into equation (58) from equation (59), we get 16n4-8(1+P2cos2Q)n2-P4s+t
i2Q=O(63) Therefore, from equation (62), absorption rate k
However, the refractive index n can be found from equation (64), and the complex refractometer can also be found from 100=n-ik. After all, from the DC component V (DC), ω component V (ω), and 2ω component V (2ω) of the output signal of the photodetector, the value of the amplitude reflectance ratio rp/r3 and the phase difference (retardation) γ are determined. t, t
It is clear that the complex refractive index can be determined via the value of mr.

When expressed as r 3 / r p =tanv (65) as another method of expression, n and k can also be obtained using (66) and (67).

B-3 = Calculation of refractive index n1, thickness d of the thin film The amount to be determined for the thin film on the sample surface, that is, the refractive index n1, the absorption coefficient 1, the thickness d of the thin film, the phase difference γ measured by an ellipsometer, and 1 As already mentioned, the relationship with the radial width reflectance ratio tanv is = tanveexp (iγ) = ρ exp i γ (68
) is given. That is, γ and ri are functions determined by the following equation.

Army = f (nl, 1, d, n2. 2.ψ0.λ
)7=f (nl, 1, d, r12.2.ψ0
．． λ) Therefore, nl , kl , d, n2, k2, ψ0
, λ can be given, then ri and γ can be calculated. (68) has already been calculated and charted, and if you create this kind of detailed chart, you can immediately know the refractive index n1 and the thin film thickness d from the measured values by interpolation. I can do it. Furthermore, from the measured values of γ and ri, nl and d can be calculated using a computer.
It is also possible to find.

Embodiment FIG. 3 shows an embodiment of the ellipsometer of the present invention (an example of the first invention).

In FIG. 3, light from a white light source 10 is incident on a monochromator 11 to select monochromatic light with a wavelength λ. polarizer) 12, it becomes linearly polarized light with a predetermined polarization direction, and the incident angle ψ0 (
Usually, the angle of incidence is 45°). As described above, the reflected light from the sample 1 usually becomes elliptically polarized light, and enters a phase modulation element (optical polarization modulation element) 13, for example, a photoelastic modulator 13 such as a Faraday cell. This phase modulation element 13 alternately changes the incident elliptically polarized light into counterclockwise polarized light and clockwise polarized light with an amplitude δ01 and an angular frequency ω, and the delay axis is set at 45° with respect to the plane of incidence. It is located in Furthermore, an analyzer 14 is arranged on the output side of the phase modulation element 13 so that the transmission axis is parallel or perpendicular to the incident plane. A photodetector 15 such as a photomultiplier for photoelectric conversion is arranged.

Therefore, the sample reflected light polarization-modulated by the phase modulation element 13 is incident on the photodetector 15 via the analyzer 14, and a signal corresponding to the incident light is output from the photodetector 15.

The output signal of the photodetector 15 is sent to a signal component separation circuit 16.
is separated into a direct current component V (DC), an ω component (ω), and a 2ω component V (2ω). This signal component separation circuit 16 is constructed using a filter, a synchronous rectifier circuit, and the like. The signals of each component (rJC), V(ω>, V(2ω)) obtained from the signal component separation circuit 16 are input to an arithmetic device 17 such as a computer or a dedicated arithmetic device. is V(DC)
, V(ω), ■(2ω), wavelength λ, etc., the reflection coefficient ratio rp /r3 (=ρ) is calculated using the method described above.
and the difference in phase change (phase difference=retardation) γ are determined by calculation, and roughly based on these, the refractive index n and the thickness d of the thin film are determined.

FIG. 4 shows another embodiment (an example of the second invention) of the ellipsometer of this invention.

In the ellipsometer shown in FIG.
is arranged in the incident optical path to the sample 1. That is, a phase modulating element 13 having an amplitude δ0 and a modulation angular frequency ω is arranged between the linear polarizing element 12 and the sample 1 so that its delay axis is at 45° with respect to the plane of incidence. The other configurations are similar to the example shown in FIG.

In addition, on each side above, the monochromator 1 of the light source part
1. If a device such as a diffraction photon with a variable selection wavelength is used and measurement is performed while scanning the wavelength, the wavelength dispersion of the polarization state can also be determined.

Effects of the Invention The ellipsometer of the present invention performs phase modulation on the sample reflected light or sample incident light at an angular frequency ω to separate the DC component, ω component, and 2ω component of the detected light intensity signal, and calculates the sample This method determines the refractive index and thickness of the signal, and separating the signal components by modulating in this way means alternating current detection, so the influence of background light is much less than in the case of conventional direct current detection. etc. by significantly reducing S.
/N can be made good and the refractive index and thickness can be determined with high accuracy.

[Brief explanation of the drawing]

Figure 1 is a schematic diagram showing the concept of general ellipsometry.
Fig. 2 is a schematic diagram showing the concept of ellipsometry for a thin film on a sample surface, Fig. 3 is a block diagram showing an example of the ellipsometer of the first invention, and Fig. 4 is a block diagram showing an example of the ellipsometer of the second invention. It is a diagram. DESCRIPTION OF SYMBOLS 1... Sample, 2... Thin film, 12... Linear polarization element, 13... Phase modulation element, 14... Analyzer,
15... Photodetector, 16... Signal component separation circuit.

Claims (2)

[Claims] (1) In an ellipsometer that makes linearly polarized light incident on the sample surface and measures values related to the sample surface from the polarization state of the reflected light, the polarization direction of the linearly polarized light incident on the sample surface is determined during the optical path of the incident light on the sample surface. A linear polarizing element is arranged so as to keep constant.
A phase modulation element with an amplitude δ_0 and a modulation angular frequency ω is arranged in the optical path of the reflected light from the sample surface so that its delay axis is 45° with respect to the incident plane, and on the output side of the phase modulation element, An analyzer is arranged so that the transmission axis is parallel or perpendicular to the incident plane, and a photodetector for photoelectric conversion of light is arranged on the output side of the analyzer, and the output signal of the photodetector is A signal component separation circuit is provided to independently extract the ω component, 2ω component, and DC component, and the reflection coefficient ratio r_ of the sample surface is calculated from each of the components.
An ellipsometer characterized in that p/r_s and a phase change γ of reflected light are determined, and based on these, the refractive index of the sample surface and/or the thickness of a thin film on the sample surface are determined. (2) In an ellipsometer that makes linearly polarized light incident on the sample surface and measures values related to the sample surface from the polarization state of the reflected light, to keep the polarization direction of the linearly polarized light constant during the optical path of the incident light on the sample surface. A linear polarizing element is placed between the linear polarizing element and the sample surface, and a phase modulating element with an amplitude δ_0 and a modulation angular frequency ω is placed between the linear polarizing element and the sample surface so that its delay axis is 4
5°, and in the optical path of the reflected light from the sample surface, place an analyzer so that the transmission axis is parallel or perpendicular to the incident surface, and then place the analyzer on the output side of the analyzer. A photodetector for photoelectric conversion of Therefore, the reflection coefficient ratio r_p/r_ of the sample surface
1. An ellipsometer characterized in that the refractive index of the sample surface and/or the thickness of a thin film on the sample surface are determined based on the determined s and the phase change γ of reflected light.