JPH02297008A - Optical measuring method for thickness of opaque thin film - Google Patents

Abstract

PURPOSE:To execute the measurement extending over a wide film thickness range by allowing a light beam to be made incident on an opaque thin film, measuring a reflected light transmit ting through the thin film and a reflected light containing a surface reflected light in the outside of a characteristic absorption band, calculating the transmittivity by a specific expres sion, and deriving the thickness from a Beer Lambert law. CONSTITUTION:A light beam (intensity I0) is made incident on an opaque thin film, a reflected light transmitting through the opaque thin film and a reflected light (intensity Im) containing a surface reflected light are measured in the outside of a characteristic absorption band, intensity Ir of the surface reflected light contained in the reflected light is derived, transmittivity T is calculated as (Im - Ir)/(I0 - Ir), and thickness is derived from a Beer Lambert law from transmittivity data by wavelength of the outside of the characteristic absorption band. As for a thin film (opaque film 1) containing light scattering particles, an interference phenomenon comes to be scarcely generated. When the optical wavelength becomes larger than the scattering particle diameter, light is scattered more in front of the particles than in the rear thereof, therefore, the light scattering intensity is not influenced so much as long wavelength. Accordingly, it is desirable to use near infrared rays or infrared rays instead of a visible light, that is, a semiconductor lazer whose wavelength is about 1.3mum band or 1.5mum band is used. Also, in the case of applying an interference method to a painting paint sample, it is desirable to utilize about 2 - 2.5mum band and 3 - 5mum band being a wavelength band being free from an infrared characteristic absorption band, namely, having high transparency in an infrared area.

Description

[Detailed description of the invention] (Industrial application field) The present invention relates to a method for optically measuring the thickness of opaque thin films.

(Prior Art and Problems to be Solved by the Invention) Typical optical thickness measuring methods for transparent thin films include optical interference method and infrared absorption method. In an optically transparent thin film, researchers use the phenomenon that interference occurs between the front and back surfaces of the thin film to calculate the thickness from the relational expression between the interference fringes and the thickness. The latter uses the infrared characteristic absorption characteristic of thin film materials to determine the thickness from the relationship between thickness and absorbance, a so-called calibration curve. Each uses a different wavelength blade. The former uses a transparent wavelength band with no characteristic absorption.

The latter cannot be accurately measured unless the interference phenomenon peculiar to optical thin films is removed. In this way, interferometry and absorbance measurement are two sides of the same coin.

By the way, for thin films containing light-scattering particles, for example, opaque films, or painted films containing pigments, both the optical interference method and the infrared absorption method have applicability limits. Light-scattering particles include ordinary particles that absorb and reflect incident light, such as pigments in paint films, as well as particles that absorb visible and infrared incident light, such as carbon particles in paint films. There are some that do not diffuse the incident beam very much, and others, such as metal particles, that reflect the incident light and diffuse it in a macroscopic state close to diffuse reflection.

In a thin film containing such light-scattering particles, interference phenomena are less likely to occur due to the scattering and attenuation of the light beam and the decrease in coherence caused by the light scattering phenomenon caused by the particles in the sample.

Furthermore, since the attenuation due to this light scattering is different from the attenuation due to infrared characteristic absorption, the infrared absorption spectrum is also distorted. moreover,
As the scattering phenomenon becomes stronger, e-stripes are no longer observed. On the other hand, the absorbance is not proportional to the thickness and exhibits an inversion phenomenon. In other words, the Beer-Lambert law breaks down. For this reason, both methods must solve these problems.
It cannot be applied to measuring the thickness of opaque samples or paint films, and the range of application cannot be expanded.

It is an object of the present invention to provide a method for optical measurement of opaque thin films.

(Means for Solving the Problems) A first method of optically measuring the thickness of an opaque thin film according to the present invention involves injecting light (intensity Io) into the opaque thin film, and comparing the reflected light transmitted through the opaque thin film and the surface reflected light. Reflected light including (intensity ■l1
1) is measured outside the characteristic absorption band, the intensity Ir of the surface reflected light included in the reflected light is determined, and the transmittance T is (II, l-1,)
/(Io-I,), and the thickness is determined from the Beer-Lambert law from the transmittance data at a wavelength outside the specific absorption band. The measurement method involves injecting light (intensity I0) into an opaque thin film, measuring the wave number dependence of the reflected light (intensity Im) that includes the reflected light that has passed through the opaque thin film, and the surface reflected light, and measuring the surface reflection included in the reflected light. The light intensity Ir is separated from the apparent transmittance data outside the characteristic absorption band as a thickness-independent part, and the characteristic absorption in the characteristic absorption band is calculated by interpolation from the transmittance data in a wide wavenumber range including the characteristic absorption band. Obtain the transmittance data assuming that there is no transmittance, and at a certain wavelength within the characteristic absorption band, calculate the transmittance of 100% calculated from the standard of transmittance of 0% obtained from the surface reflected light intensity and the interpolation data. Determine the standard, calculate the transmittance converted from the transmittance 0% standard and 100% standard for the transmittance data at the above wavelength within the specific absorption band, and calculate the thickness from this transmittance using the Beer-Lambert law. It is characterized by

(Function) The first thickness measurement method and the second thickness measurement method according to the present invention are as follows:
In both cases, the reflection of incident light on an opaque thin film is photometered, the reflected light is corrected to remove the surface reflected light, and the measured data is calculated from the true incident light intensity (Io-I, ) and true transmitted light intensity (It-I).
r) and determine the transmittance. The first thickness measurement method uses transmittance data outside the characteristic absorption band to
Determine the thickness using Lambert's law. Furthermore, in the second thickness measurement method, the transmittance when it is assumed that there is no characteristic absorption within the characteristic absorption band is determined by interpolation from data in a wide wavelength range. Then, regarding this data, the transmittance is 0%, which is calculated separately.
The transmittance is determined using the standard and the 100% standard, and the thickness is determined using the Beer-Lambert law.

(Example) Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

(a) Thickness measurement method using interferometry and light scattering method The thickness measurement method using light scattering method according to the present invention uses the same measuring device as the conventional interferometry method, and the thickness measurement method using the light scattering method according to the present invention There are some points in common with the thickness measurement method. Both methods can be used together. Therefore, for comparison, first, the conventional interference method will be explained.

Interferometry is originally a measurement method that can be applied to transparent samples. First, a transparent sample l on a metal reflector plate 2
Consider the interference of

As shown in Figure 1, strength ■. , when a monochromatic light beam of wavelength λ is incident on sample 1 at an incident angle θ, due to the difference between the refractive index of sample 1 and the refractive index of air, at the incident point BI on the sample surface, refracted/transmitted light Ij and surface reflected light Ir
divided into The refracted/transmitted light Ij is reflected at the reflection point D1 on the back surface of the sample and then emitted again from point B on the sample surface, but since it is parallel to the surface reflected light, an interference phenomenon occurs between the two components and the true The sample transmitted light component Ij is intensity-modulated. Therefore, if the wavelength λ of the incident light is scanned or the incident angle θ is scanned, the interference fringes will be f[l! be measured.

FIG. 2 shows the configuration of a conventionally known incident angle scanning interferometry thickness gauge. Constant wavelength λ emitted from light source 3
The light passes through a polarizer 4, a rotating mirror 5 and a concave reflecting mirror 6.
The light is sequentially reflected at and incident on the sample l. sample 1
The reflected light is reflected by the concave reflecting mirror 7 and detected by the detector 8.

The light source 3 is a HeNe laser (wavelength: 0.6328μ).
m), or a semiconductor laser for optical information processing (wavelength 0
．． 7 to 0.8 μm band, etc.) is used. The incident light is unpolarized light or S-polarized light. By rotating the rotating mirror 5, the angle of incidence θ on the sample 1 can be scanned.

This measurement method has the advantage of being able to measure from one side of a transparent film because it is a reflection measurement.

The thickness d of the sample l can be calculated from the relational expression between the interference fringes and the sample thickness d, if the refractive index n of the sample is known in advance. That is, d=(λ/2)/
(β” Sin'(jp n Sin pal
)......(υ λ: wavelength of human light n: refractive index of the sample θ9.θ, +1 = incidence corresponding to the pth and (pal)th interference fringe peaks (or valleys) from the center of the interference fringe In addition, if the refractive index value n of the sample is unknown, conversely, by measuring the interference fringes of a sample of the same type with known thickness, the refractive index value n can be obtained from equation (1), so from now on, You can use this value.

Next, the application of the interferometry method to thin films containing light-scattering particles, such as opaque films, or painted films containing pigments, will be described. In these samples, scattering of the light beam due to light scattering phenomena by particles in the sample;
Interference phenomena become less likely to occur due to the reduction in attenuation and coherence. Attenuation due to this light scattering (hereinafter referred to as scattering/attenuation) is different from attenuation due to infrared characteristic absorption (hereinafter referred to as absorption/attenuation). In addition, even with a black paint film containing carbon particles that are visible and infrared absorbers, the amount of transmitted light ■t
is significantly attenuated and the interference effect is reduced.

In order to expand the application of interferometry to such opaque samples, it is necessary to consider the wavelength of light.

The light scattering intensity depends on the relationship between the scattering particle diameter a and the light wavelength λ. When the wavelength becomes thicker than the particle size, the fourth
As shown in the figure, more light is scattered in front of the particle 1' than in the rear (this is called the Mie effect). That is, long wavelengths are less susceptible to scattering. Therefore, from the viewpoint of light scattering effect, it is better to use near-infrared light or infrared light than visible light. From a practical point of view, it is better to use a semiconductor laser with a wavelength of 1.3 μm or 1.5 μm, which is used for optical communications (however, if a long wavelength is used, it may cause damage to thin samples). (The accuracy of thickness measurement will be lower, so a compromise will need to be made depending on the application.)

Figure 3 is an example of measuring interference fringes of a red paint sample on an aluminum plate when a semiconductor laser with a wavelength of 1.3 μm is used as the light source 3 in the incident angle scanning system shown in Figure 2. . From this measurement data, θ is determined and the thickness d is calculated.

To further expand the application of interferometry to thick films with poor transparency or painted paint samples, in the infrared region,
The 2-2.5 μm band and the 3-5 μm band, which have no infrared characteristic absorption band of the sample, that is, the wavelength band with high transparency, are +
It is preferable to use 1. Furthermore, in the case of a black paint film containing carbon particles which are visible/infrared absorbers, from the viewpoint of the diffraction effect caused by the particles, it is preferable to similarly use a long wavelength.

When the wavelength exceeds 5 μm, the effect is generally small due to the complex infrared characteristic absorption of the sample.

At present, the types of lasers that oscillate in the 3-5 μm band are very limited, and it is difficult to do so manually, so a desired monochromatic light beam is selected from the white light source spectrum using an interference filter and used.

(b) Thickness measurement method using scattering/transmission method using light scattering phenomenon Even if the optimal long wavelength is used from the viewpoint of interference effects, as the light scattering/attenuation of the sample becomes stronger, the contrast of the interference fringes becomes worse. This makes the application of interferometry practically impossible. Next, a thickness measurement method according to the present invention that utilizes a light scattering phenomenon, which can be applied to samples with even stronger light scattering and attenuation, will be described. Hereinafter, in order to clarify the explanation, wavelength bands other than the infrared characteristic absorption band will be limited.

Since light scattering intensity is a function of wavelength, particle density, etc.
If the particle size or its distribution is constant, it is expected to be correlated with the thickness of the sample.

That is, it is possible to quantify the thickness by measuring the scattering intensity. When we change the thickness of sample l using the optical system shown in Figure 1 and examine the spectral intensity of forward scattered and transmitted light using an infrared spectrometer, we find that there is a quantitative relationship with the thickness of the paint sample, which will be explained later. It will be confirmed that More specifically, the optical system of the interferometric film thickness meter shown in FIG. This means that we are measuring the intensity of transmitted light, and we were able to demonstrate that the intensity of forward scattered and transmitted light has a quantitative relationship with the thickness of the sample. In other words, for samples where interference fringes cannot be observed or measured, if the rotating mirror 5 is fixed at a position that provides the optimum angle of incidence, for example 60@, it can be used in the same way as an interferometric film thickness meter. The thickness of the sample can be measured using an optical system. It is advantageous that the range of application of thickness measurement can be expanded with the same optical system.

<Influence of light reflected from the surface of the sample> In principle, thickness can be measured by measuring the amount of forward scattered and transmitted light, but unlike in the case of interferometry, the influence of specular reflection from the sample surface Light becomes an impediment to precision measurement. Even if the surface reflected light (2) does not cause an interference phenomenon, it enters the detector 8 together with the true transmitted and reflected light Ij, and the sum becomes the output signal of the detector 8. If this detection signal is ■, then 1, = It+Ir −・” (2) Normally, ■t and ■ cannot be separated, so this signal lll1
The transmitted light intensity is used to calculate the transmittance T.

That is, T,=1. /Io-(I t+ Ir)/Io--(3
) However, this is an apparent transmittance. FIG. 5 shows the transmittance spectrum obtained from the data of equation (3) for a black paint sample on an aluminum plate.

Dotted line A shows the approximate contribution of surface reflection.

In addition, Figure 6 shows the absorbance spectrum (the transmittance spectrum shown in Figure 5 was converted to an absorbance spectrum), but it is clear that the absorbance is not proportional to the thickness, and the infrared absorption method is It turns out that it cannot be applied as is.

In normal photometry, the surface reflected light ■, has a large value of several percent or more of the incident light io, and as the transmitted light ■【 becomes smaller, for example, ■t is equal to or smaller than ■, , the thickness error exceeds 100%. This is the main reason why the Beer-Lambert law breaks down. The true transmittance T is T = I t/(io-I,) ・
・”=(4).To find the true amount of transmitted light 1j,
It is necessary to remove or correct the surface reflected light I to obtain the transmittance shown in equation (4). On the other hand, when a surface reflected light component occurs, the true effective incident light is also not Io, but a value corrected by subtracting the light reflected light Ir, that is, (I
o-Ir).

Method for Reducing Sample Surface Reflected Light In the case of a transparent thin film, when P-polarized light is incident at a polarization angle, the sample surface reflected light component is reduced or not generated. Therefore, the optical interference phenomenon will be reduced or will not occur (
(See Japanese Patent Application Laid-Open No. 60-224002). Also, in the case of opaque samples, the P polarization/polarization angle incidence method (or using the incident angle that minimizes the surface reflection component)
, it has the effect of reducing surface reflection components. Furthermore, since the light scattering effect of particles and the reflection characteristics of metal substrates depend on the polarization characteristics of incident light, it is important to determine the polarization characteristics of incident light.
It also has the effect of increasing signal reproducibility and stability.

<Correction Method for Sample Surface Reflection Light> Although the surface reflection component can be reduced using the P-polarized light and polarization angle incident method in this way, it cannot be completely removed, so it is necessary to correct it.

In the first correction method, photometry is performed using a thick sample from which only surface reflected light is reflected, and the surface reflected component is removed. To be precise, the surface reflected light Ir is the sum of the sample surface reflected light and the component of the specular reflection direction of the scattered light from the vicinity of the surface, and is within the normal thickness range of the sample to be measured, except for very thin samples. It is considered to be constant regardless of the thickness. There, the incident light is completely absorbed within the sample. That is, if a sample thick enough to satisfy 1j=0 is prepared and photometered, the detection signals (2) and (2) can be considered to be only surface reflected light. That is, ■、゛〜■、
(5) Therefore, it is possible to correct the transmittance by regarding the detection signals ■ and ゛ as ■. That is, the corrected transmittance T0 is T, = (Ill-1°r)/(Io-1',) ~ (1
,-1,)/(io-1ρ=Ii/(Io-1ρ=T
...(6) becomes.

In the second correction method for surface reflected light, in the relationship between the transmittance T1 in equation (3) and the sample thickness, the transmittance decreases exponentially as the thickness increases. Fit the transmittance and absorbance exponential functions in the black paint sample and extrapolate to infinite thickness. This correction provides a very satisfactory value.

Table 1 shows the apparent transmittance of black paint samples on aluminum plates of various thicknesses without surface reflection correction at a wave number of 2160 cm-', T1 = 1, /I
o (formula (3)) and absorbance Am-IOg (1/Tl11
), and the transmittance Tc-(1, rr'XI□ Iy'X formula (
6) % formula %(7) was compared. The measurement conditions were as follows in the apparatus shown in Figure 2.
Interference filter (center wavelength λ = 4.63 μm 1 half width △λ
= 150 ns), the incident angle was 60°, and P-polarized light was used. Note that the thickness of each sample is determined by a different measurement method.

But the correction value is! ,°~0.04331°.

Figure 7 shows the sample thickness d and absorbance A0 of this measurement result.
Shows a graph of the relationship between It can be seen that after surface reflection correction, the relationship between absorbance AC and thickness d becomes a straight line. Moreover, even if it is treated as a straight line passing through the origin, the error is small. In other words, the Habo-Beer-Lambert law holds true.

In the third correction method, as shown by broken line A in Figure 5, a baseline is drawn between point C of the saturated absorption peak at the high wavelength end and point a where the scattering spectra at the low wavelength end merge, and the necessary Read the value at the measurement wavelength point. This reading is surface reflected light!
, and so on.

The second and third correction methods described above have the advantage of not requiring correction samples.

As for the wavelength used for measurement, in general, long wavelengths are less susceptible to the influence of scattering, so it is preferable to use infrared rays on the long wavelength side and a highly transparent wavelength band of 3 to 5 μm. The signal intensity is high because of the highly transparent wavelength band, and the amount of change with respect to thickness is much larger than the amount of change with infrared characteristic absorption method. Furthermore, since the scattering spectrum changes slowly with respect to wavelength, the bandwidth of the filter can be much larger than the bandwidth used in interferometry or characteristic absorption methods. Therefore, the intensity of the measurement signal is significantly greater than that of the infrared characteristic absorption method. Moreover,
The large bandwidth has the effect of averaging local fluctuations in the spectrum and noise. For these reasons, the scope of application of the sample can be expanded.

In this way, it was possible to measure the thickness of a paint film containing pigments and a black paint film containing carbon particles over a thickness range of 50 μm or more, which cannot be measured by normal interferometry.

Note that if the applicable samples are limited to light-scattering samples or painted paints, there is no need to use the interferometry method in combination, and the measuring device can be simplified by using a dedicated device for the light-scattering method.

(Leaving space below) (c) Thickness measurement method using infrared absorption method In the light scattering method explained above, thickness was determined from data outside the infrared characteristic absorption range, but forward scattered transmitted light in the infrared characteristic absorption range The thickness can also be determined from the data using the infrared absorbance method. This method provides high accuracy in thickness measurements of relatively transparent film and coating samples.

Quantification of thickness by infrared absorption method is based on the Beer-Lambert law in principle. In this infrared absorption method, the influence of interference phenomena must be removed. In order to quantify characteristic infrared absorption, the ratio method or baseline method using a reference wavelength is used, but the reference wavelength or baseline wavelength point is usually selected at a wavelength that has no characteristic absorption. Easily exhibits interference phenomena. And it obstructs precise measurement. These wavelength bands are the same as in the case of a transparent optical thin film, and when P-polarized light is incident at a polarization angle, surface reflection components are reduced or not generated. That is, since interference phenomena are reduced or not generated, corrections can be made accurately using the baseline method or the like using data in a wide wavelength range outside the infrared characteristic absorption range, and therefore the infrared absorption of the sample can be measured with high precision.

FIG. 8 shows an infrared absorption type film thickness meter used in this example. In this film thickness meter, the white light emitted from the light source +1 is transmitted through the rotating interference filter 1 provided on the rotatable disk 12.
3 through the wavelength λ1. λ7. The light becomes monochromatic light of λ, and further passes through the polarizer 14 and enters the sample 15. The reflected light from the sample 15 is detected by a detector 16, and the output signal of the detector 16 is processed by a signal processing section 17, sent to the outside, and recorded and displayed. This thickness gage is essentially the same as the thickness gage of FIG. 2, except that the angle of incidence is fixed.

<How to take the baseline of the transmittance spectrum> FIG. 9 shows the transmittance spectrum of a brick-colored paint sample on a steel plate. Moreover, FIG. 10 shows an absorbance spectrum obtained by converting the transmittance spectrum shown in FIG. 9.

When the sample is a light-scattering opaque thin film, the incident light is attenuated by scattering by the scattering particles. This is a physical phenomenon that is essentially different from attenuation due to absorption of infrared characteristics. Since the light scattering intensity is a function of wavelength, particle density, etc., the scattered and transmitted light shifts depending on the thickness of the sample, as shown in FIGS. 5 and 9. For this reason, the ratio method, which assumes that the thickness is proportional to the absorption rate at a certain wavelength, cannot provide practical quantitative accuracy. Therefore, the baseline method is generally used. The background of the infrared transmission spectrum is this forward scattering/transmission spectrum, and this background becomes important and problematic because it becomes a reference when quantifying the infrared characteristic absorption.

Conventionally, the baseline method for quantifying characteristic absorption has been to fit a straight line or smooth the absorbance spectrum, not the transmittance spectrum, assuming a true-like background in a narrow range sandwiching the characteristic absorption band. Draw a curved line. However, as can be seen from the absorbance spectrum of the paint shown in Figure 1θ, drawing the baseline is not simple. For example, 2 shown by the dashed line
It is not possible to determine which of the book baselines CI and C2 is closer to the true baseline. It can be seen that straight line fitting has limited applicability.

The background is the scattering spectrum itself, the intensity of which is a function of the wave number (the reciprocal of the wavelength).

Therefore, rather than making judgments in a narrow range that spans the characteristic absorption band, a functional model of the power of the wave number (+/λ) is fitted over a wide wavelength band in the transmittance spectrum, as shown by the broken line in Figure 9. This has good accuracy and reproducibility. We define this as a hypothetical new transmittance within the characteristic absorption band (
100% standard). Then, the base value of a necessary characteristic absorption wavelength point, for example, 3440c+a-' is read.

Correction of sample surface reflected light> As explained above, the P-polarized light/polarization angle incidence method is used to reduce the surface reflected component Ir, but even so, the surface reflected light cannot be completely removed. Correction is performed in the same way as in the light scattering method described. That is, the first
In the correction method, a sample thick enough to make the transmitted light intensity Ij = 0 is prepared and photometered, and the detected signal in the characteristic absorption band is taken as the surface reflection component (2). In the second correction method,
Extrapolate the data of transmitted light intensity 1j to infinite thickness to find the surface reflection component! , and so on. In the third correction method, as shown in FIG. 9, a baseline D is drawn by connecting the saturated absorption peak f on the high wavelength side and a point e where the scattering spectra on the low wavelength side merge. Then, using this line as a new transmittance reference of 0%, the base value of a necessary characteristic absorption wavelength point, for example, 3440 cm-' is read.

Quantification of characteristic absorption In particular, in paint samples, the main reason why the linear relationship between the absorbance of the characteristic absorption band and the thickness, that is, the Beer-Lambert law, breaks down is that the sample surface reflected light component is not corrected. And how to take the baseline.

Furthermore, there is a flaw in the attempt to correct the absorbance spectrum rather than the transmittance spectrum.

Therefore, in this example, first, in the normal transmittance spectrum shown in FIG. 9, a curve obtained by fitting a mathematical model that takes into account the light scattering effect of pigment particles is used as the standard for transmittance of 100%. Next, the spectrum of only the light component reflected from the surface of the sample is set as a reference for transmittance of 0%. The characteristic absorption profile lies in this range. characteristic absorption wavelength,
For example, the value at 3440cI11-' is continued and the corrected transmittance T0 is calculated based on the new transmittance 0%-100% standard. Then, from this value, absorbance AC (
-log(1/To)).

When viewed from this new 0%-100% transmittance standard, it can be seen that in samples with large scattering and attenuation, the characteristic absorption is significantly compressed and distorted. This is the reason why the Beer-Lambert law breaks down.

FIG. 11 shows the relationship between absorbance and thickness of a paint sample on a steel plate. Curve (G) is based on the conventional method without the above correction. The amount of change in absorbance with respect to thickness is small and decreases as the thickness increases. In other words, the Beer-Lambert law is broken. On the other hand, it can be seen that when the above correction is made, the linear relationship is restored and the line passes through the origin, as shown in line (F). In other words, veil
Lambert's law almost holds true.

<Switching to light scattering method> However, when the sample thickness becomes thicker or the sample has stronger scattering properties, most of the incident light is scattered,
Attenuation occurs, and forward scattered/transmitted light becomes smaller and smaller. Moreover, in the characteristic absorption band, attenuation due to absorption is added, so the photometry of transmitted light is related to the S/N of the photometry system.
It gets extremely bad. Moreover, as a quantitative procedure, characteristic absorption cannot be calculated unless the standard of transmittance is 100% - 0% is drawn, so it is difficult to quantify characteristic absorption accurately.
Or it becomes impossible.

On the other hand, as already explained in Section (b), the forward scattered/transmitted light intensity in the wavelength band outside the infrared characteristic absorption band has a quantitative relationship with the sample thickness, as shown in Figure 7. In other words, it follows the Beer-Lambert law. The optical system of the infrared absorption type film thickness meter shown in FIG. 8 is designed so that several interference filters 13 can be set on the rotating disk I2. If this is set in advance, the scope of application can be expanded by simply switching the quantitative method from the infrared characteristic absorption method to the infrared scattering/transmission method.

Quantification is simple because there is no need to draw a so-called baseline and measurement is performed at ■wavelength.

Regarding the wavelength used for measurement, in general, long wavelengths are less affected by scattering, so it is best to use a wavelength band where the long wavelength side is infrared and is highly transparent, for example, the 3 to 5 μm band. good. The signal strength is high due to the highly transparent wavelength band. Moreover, the amount of change with respect to thickness is much larger than the amount of change in the characteristic absorption band. Also,
Since the scattering spectrum changes very slowly with respect to wavelength, the bandwidth of the filter 13 can be much larger than the bandwidth when measuring characteristic absorption.

For example, a wide band filter of several 1100n can be used, further increasing the signal strength. Furthermore, stability is improved because local fluctuations and noise in the spectrum are averaged out.For these reasons, sensitivity and stability are
This is significantly higher than the conventional infrared characteristic absorption method, which measures only characteristic absorption bands. Therefore, the scope of application of the sample can be expanded.

(Effect of the invention) It has become possible to quantify the thickness of an opaque thin film over a fairly wide range of film thicknesses.

[Brief explanation of drawings]

FIG. 1 is a diagram showing a specular reflection optical system for a parallel plane sample. FIG. 2 is a cross-sectional view of a thickness gauge using an incident angle scanning interference method. FIG. 3 is a graph showing interference fringes of a red paint sample on an aluminum plate using an incident angle scanning interference type thickness gauge. FIG. 4 is a diagram for explaining the Mie scattering effect due to particles. FIG. 5 is a diagram of the transmittance spectrum of a black paint sample on an aluminum plate. FIG. 6 is a diagram of the absorbance spectrum of a black paint scepter on an aluminum plate. FIG. 7 is a graph of the relationship between black paint sample thickness and absorbance. FIG. 8 is a diagram showing an infrared absorption type film thickness meter. FIG. 9 is a diagram of the transmittance spectrum of a paint sample painted on a steel plate. FIG. 1θ is an absorbance spectrum of a paint sample painted on a steel plate. FIG. 11 is a graph of the relationship between the absorbance of the paint film on the steel plate and the sample thickness. l... Opaque thin film, 3... Light source, 8... Detector. Patent applicant/Kurashiki Boseki Co., Ltd. agent Patent attorney Blue and white Blue and white! Figure 11 Figure 2 S Figure 3 Figure 4 Incident angle θ (hawk)

Claims (2)

[Claims] (1) Light (intensity I_o) is incident on the opaque thin film, and the reflected light (intensity I_m), which includes the reflected light transmitted through the opaque thin film and the surface reflected light, is measured outside the characteristic absorption band, and the surface reflection included in the reflected light is measured. The light intensity I_r is determined, the transmittance T is calculated as (I_m-I_r)/(I_o-I_r), and the thickness is determined from the transmittance data at a wavelength outside the characteristic absorption band using the Beer-Lambert law. An optical method for measuring the thickness of opaque thin films. (2) Inject light (intensity I_o) into the opaque thin film, measure the reflected light that has passed through the opaque thin film and the reflected light (intensity I_m), which includes the surface reflected light, and calculate the intensity I_r of the surface reflected light included in the reflected light. , it was separated from the apparent transmittance data outside the characteristic absorption band as a thickness-independent part, and it was assumed that there was no characteristic absorption in the characteristic absorption band by interpolation from the transmittance data in a wide wavenumber range including the characteristic absorption band. At a certain wavelength within the characteristic absorption band, find the standard of 0% transmittance obtained from the surface reflected light intensity and the standard of 100% transmittance calculated from the interpolation data. Regarding the transmittance data at the above-mentioned wavelength within the absorption band, the transmittance is calculated based on the above-mentioned transmittance 0% standard and 100% standard, and the thickness is calculated from this transmittance using the Beer-Lambert law. Optical method for measuring the thickness of opaque thin films.