JP5871242B2 - Film thickness measuring apparatus and film thickness measuring method - Google Patents

Description

  The present invention relates to a film thickness measuring device and a film thickness measuring method.

Patent Document 1 discloses a measurement method for measuring the thickness of a SiO ₂ film in a non-contact manner using optical interference. In the measurement method of Patent Document 1, measurement is performed by switching light of a plurality of wavelengths using a confocal optical system. Specifically, the measurement data of the reflectance for the first wavelength and the second wavelength are respectively obtained, and the calculated data indicating the relationship between the wavelength and the reflectance for each film thickness is obtained from the measurement data. It is calculated by approximating the thickness.

JP 2009-204313 A

By the way, in the manufacturing process of a power device, a bipolar device, a CMOS (Complementary Metal Oxide Semiconductor) device, a sensor, and the like, there is a step of forming a narrow groove (trench) in order to completely separate various devices. In some cases, SiO ₂ is formed at the bottom of the trench. Furthermore, an SiO ₂ film may be formed on the bottom of the through hole in the process of manufacturing the through electrode provided in the stacked three-dimensional device. It is desired to measure the thickness and residue of the SiO ₂ film.

  As the aspect ratio (trench depth d / width W) of a trench formed by an etching technique increases, a technique for measuring whether the trench structure is formed correctly by measuring the residue and oxide film thickness at the bottom of the trench. desired.

  In general, the bottom of the sample is evaluated by observing the cut surface of the sample with a scanning electron microscope (SEM). However, in SEM, since it is necessary to cut the sample, non-destructive measurement is desired.

  In addition, it is difficult to apply to a trench by a conventional optical film thickness measurement method (white spectroscopy, reflection spectroscopy, ellipsometry, etc.). In particular, when measuring a trench having a large aspect ratio, it becomes difficult to measure accurately.

  For example, in the method using reflection spectroscopy, a spectroscopic spectrum is measured by illuminating a sample with white light and dispersing the reflected light. The spectral spectrum is theoretically calculated in the form of wavelength dependence of absolute reflectance according to the conditions determined by the wavelength and thickness of light, the refractive index (n) and attenuation coefficient (k) of the film and substrate. be able to. The film thickness is determined by fitting the measured data to the theoretical value using the film thickness as a parameter.

In the case of the normal of the plane sample, I'll be using the reflection reference sample, it is possible to convert the reflectance of the sample to an absolute reflectivity. However, light reflected trench bottom is affected by the shape of the trench, want intends is difficult to measure the absolute reflectance. In general, it is difficult to measure the spatial distribution of the film thickness with high resolution, and the average value of the area illuminated by light is obtained. Since it is difficult to separate the reflected light from the bottom of the trench and the reflected light from the upper surface of the trench, it is difficult to accurately measure the absolute reflectance.

  In the method using ellipsometry (ellipsometry), the film thickness can be obtained by calculating the reflected light. However, since it is necessary to make the illumination light obliquely incident on the sample, it is difficult to illuminate the bottom of the trench and receive the reflected light. In general, it is difficult to measure the spatial distribution of the film thickness with high resolution, and the average value of the area illuminated by light is obtained.

In the method using white interference fringe scanning, the sample or objective lens is Z-scanned. The height can be obtained by calculating the maximum intensity position of the interference fringes of white light when the Z scan is performed. In order to measure the film thickness, it is necessary to separately measure the interference intensity signals from the base substrate and the film surface . However, when the film thickness is 1 μm or less, it is difficult to separate these intensity signals. Furthermore, when a transparent film having a low reflectance exists on a substrate having a high reflectance, the interference intensity from the transparent film cannot be accurately measured, and the film thickness measurement becomes more difficult.

In the process according to the confocal microscope, the distance between the objective lens and the sample is changed by the Z scanning, by detecting the in-focus point position, allowing the height measurement and the film thickness measurement. When the film thickness is several μm or less, it is difficult to accurately measure the film thickness because the reflected signals on the base substrate and the film surface cannot be separated . Furthermore , in order to accurately measure the in-focus position, it is necessary to increase the numerical aperture of the objective lens. However, the larger the numerical aperture, the more illumination light components tilted with respect to the optical axis of the objective lens. Therefore, the amount of light reaching the bottom of the trench is reduced, making it difficult to detect the in-focus position.

  The present invention has been made against the background of such circumstances, and provides a film thickness measuring apparatus and a film thickness measuring method capable of accurately measuring the film thickness of a thin film provided at the bottom of a recess. It is intended.

  A film thickness measurement apparatus according to an aspect of the present invention is a film thickness measurement apparatus that measures the film thickness of a thin film formed on the bottom of a recess provided on a sample, and includes at least a first wavelength illumination light A light source unit capable of switching between illumination light of the second wavelength, a photodetector for detecting reflected light reflected by the sample and acquiring an image in a predetermined region of the sample, and illumination light from the light source unit And a confocal optical system for guiding the reflected light from the sample to the photodetector, and a first image by the illumination light of the first wavelength to calculate the film thickness of the thin film, And a processing unit for obtaining reflectance measurement data for the first wavelength and the second wavelength based on the second image by the illumination light of the second wavelength, and the processing unit includes the wavelength and the reflectance. For each thin film thickness. Referring to calculate data, said calculated by approximating the thickness of the thin film from the measured data, in the approximation of the thickness, the correction coefficient corresponding to the shape of the recess is one that has been introduced. Thereby, the film thickness of the thin film provided in the bottom part of the recessed part can be measured correctly.

  In the above-described film thickness measuring apparatus, the correction coefficient may be a calculated value of a ratio between the light reception intensity from the plane and the light reception intensity from the bottom of the recess. Thereby, the correction coefficient can be calculated.

  In the above-described film thickness measurement apparatus, the light reception intensity from the plane and the light reception intensity from the bottom of the recess may be calculated using an approximate expression function or a reference table indicating the incident angle dependence of the reflectance. . Thereby, the correction coefficient can be easily calculated.

  In the film thickness measurement apparatus, the correction coefficient may be set according to the wavelength of illumination light. Thereby, since correction can be performed more appropriately, accurate film thickness measurement can be performed.

  In the film thickness measurement apparatus, the processing unit may calculate the film thickness of the thin film from the measurement data and the calculation data by a least square method. Thereby, it is possible to calculate by approximating the film thickness easily.

  A film thickness measuring method for measuring a film thickness of a thin film provided on a sample, wherein a first wavelength illumination light and a second wavelength illumination light are switched to the sample via a confocal optical system. The reflected light that is irradiated and reflected by the sample is detected through the confocal optical system, and a first image by the light of the first wavelength and a second image by the light of the second wavelength are acquired. In order to calculate the film thickness of the thin film, based on the first image and the second image, the measurement data of the reflectance for the first wavelength and the second wavelength are respectively obtained, and the wavelength and the reflectance are calculated. Referring to the calculation data in which the relationship is shown for each film thickness of the thin film, the film thickness of the thin film is approximated from the measurement data and calculated according to the shape of the recess. A correction coefficient is introduced.

  In the film thickness measurement method, the correction coefficient may be a calculated value of a ratio between a light reception intensity from a plane and a light reception intensity from the bottom of the recess. Thereby, the correction coefficient can be calculated.

In the above-described film thickness measurement method, the light reception intensity from the plane and the light reception intensity from the bottom of the recess may be calculated using an approximate expression function or a reference table indicating the incident angle dependency of the reflectance. Good. Thereby, the correction coefficient can be easily calculated.

  In the film thickness measurement method, the correction coefficient may be set according to the wavelength of illumination light. Thereby, since correction can be performed more appropriately, accurate film thickness measurement can be performed.

  In the film thickness measurement method, the film thickness of the thin film may be calculated from the measurement data and the calculation data by a least square method. Thereby, it is possible to calculate by approximating the film thickness easily.

  ADVANTAGE OF THE INVENTION According to this invention, the film thickness measuring apparatus and film thickness measuring method which can measure the film thickness of the film | membrane provided in the bottom part of the recessed part correctly can be provided.

It is sectional drawing which shows the structure of the sample which has a trench. It is sectional drawing which shows the structure of a trench with a film | membrane. Is a diagram illustrating an optical model of the air / _SiO 2 / Si. It is a figure explaining the optical path difference by the refractive index of reflected light. It is a graph which shows the film thickness dependence of a reflection spectrum. It is a figure which shows the measuring apparatus which concerns on this embodiment. It is a figure which shows the reflectance fitting by a film thickness parameter. It is a graph which shows the relationship between a film thickness parameter and a residual change. It is a figure which shows the relationship between NA of an objective lens, and a film thickness analysis result. It is a figure explaining the restriction | limiting of the reflected light by a trench structure. It is a figure which shows the relationship between the reflection intensity of a trench, and an absolute reflectance. It is a figure explaining the three-dimensional coordinate of incident light and reflected light. It is a figure explaining the angle range in the case of a plane. It is a figure explaining the angle range in the case of a trench. It is a figure explaining the angle range in the case of VIA. It is a figure explaining the incident angle dependence of a reflectance. It is a graph which shows the incident angle dependence of Si reflectance at the time of illumination wavelength 546nm. It is a graph which shows the incident angle dependence of Si average reflectance. It is a table | surface which shows the calculation result of a trench coefficient. It is a graph which shows the relationship between an aspect-ratio and a trench coefficient when changing an illumination wavelength. It is a graph which shows the relationship between an aspect-ratio and a trench coefficient when NA is changed. It is a graph which shows the film thickness analysis result without trench correction. It is a graph which shows the film thickness analysis result without trench correction. It is a graph which shows the film thickness analysis result with trench correction. It is a graph which shows the film thickness analysis result with trench correction. It is a figure which shows notionally the calculation model of the film thickness analysis in the case of correct | amending with a trench coefficient. It is a figure which shows notionally the calculation model which calculates a trench coefficient.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. The following description shows preferred embodiments of the present invention, and the scope of the present invention is not limited to the following embodiments. In the following description, the same reference numerals indicate substantially the same contents. In this embodiment, as a basic principle of measurement, a reflection spectroscopy type film thickness measurement method is used. The measurement apparatus according to this embodiment includes a light source, an optical system, a detector, and the like, and optically measures the film thickness. For example, the optical system of the measuring apparatus has a configuration similar to that of Patent Document 1.

[Theoretical calculation of absolute reflectance]
First, the structure of the sample will be described with reference to FIGS. FIG. 1 is a cross-sectional view showing the structure of a sample having a trench. FIG. 2 is an enlarged cross-sectional view showing a trench portion. As shown in FIG. 1, the sample 30 includes a Si substrate 31 and a SiO ₂ film 32.

The Si substrate 31 is, for example, a semiconductor substrate such as silicon. A plurality of trenches 34 are formed in the Si substrate 31. For example, the trench 34 is formed by etching a silicon wafer that is the Si substrate 31. The trench 34 is formed perpendicular to the surface of the Si substrate 31. A transparent SiO ₂ film 32 is formed at the bottom of the trench 34.

Here, the width (opening width) of the trench 34 is W and the depth is d. Further, the thickness of the SiO ₂ film 32 is assumed to be t. Therefore, the structure of the bottom of the trench 34 is air (Air) / SiO ₂ / Si in order from the top.

Here, an optical model of the air / SiO ₂ / Si structure is shown in FIG. The complex refractive indexes of air / SiO ₂ / Si are N ₀ , N ₁ , and N ₂ , respectively. Similarly, the refractive indexes of air / SiO ₂ / Si are n ₀ , n ₁ , and n ₂ , and the extinction coefficients are k ₀ , k ₁ , and k ₂ . The above complex refractive index, refractive index, and extinction coefficient are values with respect to the wavelength λ of the illumination light. For air, n ₀ = 1 and k ₀ = 0. The complex refractive indexes N ₀ , N ₁ and N ₂ are expressed by the following formulas (1) to (3), respectively.

Next, the thin film interference strength due to the SiO ₂ film 32 on the Si substrate 31 will be considered. It is assumed that light having a wavelength λ is incident on the SiO ₂ film 32 perpendicularly. FIG. 3 shows a state in which illumination light is incident obliquely in consideration of the NA (numerical aperture) of the objective lens. In FIG. 3, the interface between air and SiO ₂ is the interface A, and the interface between the SiO ₂ film 32 and the Si substrate 31 is the interface B.

Part of the light incident on the sample 30 is reflected at the interface A. A part of the light incident on the sample 30 is transmitted through the SiO ₂ film 32 and reflected at the interface B. Part of the reflected light reflected at the interface B passes through the SiO ₂ film 32 and passes through the interface A. Further, part of the reflected light reflected at the interface B is reflected at the interface B and is reflected again at the interface A toward the Si substrate 31. In general, multiple reflection occurs as shown by the dotted line in FIG.

Here, the amplitude reflectivity _{r 0} and respectively formula amplitude reflectance _{r 1} of the interface B of the interface A (4), (regarded as normal incidence) indicated by (5).

At this time, the phase change δ and the amplitude change γ of the light transmitted once through the SiO ₂ film 32 having the thickness t are expressed by equations (6) and (7), respectively.

The amplitude reflectance r of the entire film structure taking into account multiple reflection is expressed by Equation (8).

Since the reflectance R is the square of the absolute value of the amplitude reflectance r, Equation (9) is obtained.

Therefore, if the complex refractive indexes N ₁ and N ₂ are known, the normal incidence reflectance R can be calculated from the wavelength λ and the film thickness t. Here, correction of the oblique incidence effect by the NA of the objective lens is considered. As shown in FIG. 4, incident light incident at an angle θ ₀ with respect to the optical axis will be described. A part of the incident light is reflected at the interface A (reflected light E ₀ ). Further, a part of the incident light is refracted by the angle θ ₁ at the interface A according to Snell's law expressed by the equation (10).

Further, the incident light refracted by the angle θ1 at the interface A is reflected by the interface B and is refracted again by the interface A (reflected light E ₁ ). Therefore, interference due to multiple reflection occurs between the reflected light E ₀ reflected at the interface A and the reflected light E ₁ reflected at the interface B.

Unlike the case of normal incidence, the optical path difference between the reflected light E ₀ and the reflected light E ₁ is as shown in Expression (11).

As the incident angle θ0 of incident light increases, the apparent film thickness t decreases. The illumination light includes light having various incident angles ranging from an angle θ _{NA to} 0 degrees depending on the NA of the objective lens. Therefore, strictly speaking, it is necessary to calculate the interference of reflected light at all angles. However, if all the angle dependences of the amplitude reflectivity r and the optical path difference are handled, the calculation load increases. Therefore, approximation is performed using the amplitude reflectance r of normal incidence and the average optical path difference.

Since the incident light is non-polarized light, the angle dependence of the optical path difference is considered to be dominant rather than the angle dependence of the reflectance R. Then, the relationship between NA and the refraction angle is as shown in Equation (12).

Therefore, the average optical path difference can be calculated by obtaining the average of the angular distribution in the range of 0 to + θ _max for the phase change δ in the equation (11). The average optical path difference δ can be calculated by the following equation (13).

Therefore, the reflectance R for an arbitrary NA can be easily calculated by replacing the optical path difference δ from the equations (6) to (13) and calculating the reflectance R of the equation (9). FIG. 5 shows an example in which the film thickness dependence of the reflectance R in the visible light region (400 to 700 nm) is calculated for NA = 0.3. In FIG. 5, the horizontal axis indicates the wavelength λ of the illumination light, and the vertical axis indicates the reflectance R. Further, FIG. 5 shows the wavelength dependence of the reflectance R when the thickness t of the SiO ₂ film 32 is changed to 0 nm, 100 nm, 300 nm, 500 nm, and 1000 nm.

[Verification of NA film thickness analysis on plane]
First, the NA dependency of the film thickness analysis of a plane other than the trench 34 will be described. Here, a case where 436 nm, 486 nm, 514 nm, 546 nm, 578 nm, and 633 nm are selected as the wavelength λ of the illumination light will be described. Of course, the wavelength λ to be selected is not limited to the above value.

Reflectance R can be the luminance value Isample or RaMotomu capture images Mel. And the film thickness t is calculated | required from the measured value of the reflectance R of each wavelength. Specifically, the film thickness t can be obtained by fitting the reflectance R in the equation (9) using the measured value at each wavelength λ and using the least square method using the film thickness t as a parameter.

When measuring the reflectance R of the sample 30, the reflectance of a reference sample having a known reflectance can be measured. By using the reference sample, it is possible to correct the characteristics of the optical system and the light source used for the measurement. For example, Si or quartz glass can be used as the reference sample. When Si is used as a reference sample, a reflection image of Si at each wavelength is picked up by a measuring device. Here, imaging conditions such as color balance and gain control are assumed to be constant. The reflectance R of the sample 30 is calculated as a relative value with respect to the luminance value I ^Si of ^Si . Furthermore, the wavelength-dependent data R ^Si (λ) of ^Si reflectance is already known. Then, the reflectance is corrected with the wavelength-dependent data R ^Si (λ). Therefore, the reflectance R ^sample (λ) of the sample 30 can be obtained by Expression (14). Note that I ⁰ is a dark luminance value received when the shutter of illumination light is closed.

  Here, the film thickness measuring apparatus according to this embodiment will be described with reference to FIG. The measuring apparatus 100 is a confocal microscope having a confocal optical system 101. The measuring apparatus 100 includes a light source unit 10 capable of selecting a wavelength. The light source unit 10 includes a light source 11 and an interference filter 12. As the light source 11, a white light source including a plurality of bright lines in a continuous spectrum such as a mercury xenon lamp is used. For example, a xenon lamp having a wide continuous spectrum from the ultraviolet to the infrared region (185 nm to 2000 nm) may be used. Of course, the light source 11 is not limited to a xenon lamp, and a white diode, a white laser, or the like may be used. As will be described later, any light source may be used as long as the wavelength can be selected.

  An optical system for observing the sample 30 with light from the light source 11 will be described. The light emitted from the light source 11 passes through the interference filter 12 and is converted into light having a specific wavelength. As the interference filter 12, for example, a plurality of bandpass filters that selectively transmit light of a specific wavelength can be used. Thereby, a plurality of single-wavelength illumination lights are selectively transmitted. For example, 436 nm, 488 nm, 515 nm, 546 nm, 578 nm, and 633 nm can be selected as the wavelength of the illumination light. When a mercury xenon lamp is used, light having a wavelength corresponding to the emission line can be selected. It is also possible to select light having a wavelength other than the wavelength corresponding to the bright line with a filter. Since light other than the wavelength of the emission line has low intensity, it is possible to balance by widening the half-value width of the interference filter. The wavelength switching may be continuous or intermittent. For example, a wavelength of 5 to 7 may be selected between 400 nm and 650 nm.

  Note that a laser light source that emits single-wavelength laser light may be used as the light source 11 and a wavelength conversion element may be provided. For example, wavelength conversion of single wavelength light incident on the wavelength conversion element can be performed by second harmonic generation. Further, a variable wavelength laser can be used as the light source 11. Furthermore, a plurality of laser light sources that emit laser beams having different wavelengths may be provided, and light having a desired wavelength among the plurality of laser light sources may be selected.

  The single-wavelength illumination light that has passed through the interference filter 12 passes through the lens 13 a and enters the slit 14. The illumination light is shaped into a line in the X direction through the slit 14. The line-shaped illumination light enters the beam splitter 15. The beam splitter 15 branches the light so that the amount of reflected light and transmitted light is approximately 1: 1 regardless of the polarization state. Accordingly, approximately half of the illumination light is transmitted through the beam splitter 15.

  Thereafter, the light traveling in the right direction in FIG. 1 enters the vibrating mirror 16. The vibrating mirror 16 scans the sample 30 in the Y direction with X-direction linear illumination light. Thereby, the surface of the sample 30 can be scanned XY. As the vibration mirror 16, for example, a galvanometer mirror, a polygon mirror, or the like can be used.

  The illumination light reflected downward by the vibration mirror 16 enters the objective lens 17 through the lens 13b. The objective lens 17 collects the illumination light and irradiates the sample 30. The sample 30 is placed on the stage 18. Then, the reflected light from the sample 30 passes through the objective lens 17 and the lens 13b again, is reflected again by the vibrating mirror 16, and enters the beam splitter 15. Thereafter, approximately half of the incident light is reflected by the beam splitter 15 and enters the lens 13c. The lens 13 c forms an image of the combined light on the light receiving surface of the photodetector 19. The light transmitted through the lens 13 c is received by the photodetector 19.

  In the present embodiment, the photodetector 19 is a CCD line sensor that captures a confocal image of the sample 30. Illumination light transmitted from the light source 11 through the slit 14 is reflected by the sample 30 and detected by the CCD line sensor. A slit confocal image is captured by scanning the sample 30 with the vibrating mirror 16. As long as the method of the confocal optical system 101 is used, the scanning method and the like may be different, and slits and photodetectors that are suitable for the method can be used as appropriate. For example, a vibrating mirror for scanning in the X direction and the Y direction may be used, and an AOD that is an acousto-optic element may be used in the X direction.

  The stage 18 has a Z-axis drive motor (not shown) and can move the sample 30 in the vertical direction of FIG. The stage 18 is controlled so that the sample surface comes to the focal position by moving in the Z-axis direction (optical axis direction). Note that the focus position can be adjusted by moving the objective lens 17 instead of moving the stage 18 in the Z direction. That is, Z scanning is performed so that the distance between the sample 30 and the objective lens 17 changes. Of course, when measuring a reference sample, a reference sample such as Si is placed on the stage 18 instead of the sample 30.

  In the confocal optical system 101, it is conceivable that the focal position changes when the observation wavelength is changed, and a change in luminance due to this is expected. This is because the deviation of the focal position of each wavelength is measured in advance and stored in the PC, and when the wavelength is switched, the Z position of the sample 30 or the objective lens 17 is automatically finely adjusted by the deviation. You can cancel with Alternatively, an omnifocal image may be created by Z scanning at each wavelength. The wavelength dependence of the observation optical system itself can be corrected by calculation by measuring in advance a sample having a known reflection spectrum, such as silicon or quartz glass.

The processing device 20 includes a reflectance measurement unit 21 and a film thickness calculation unit 22. The reflectance measuring unit 21 measures the reflectance from the images obtained by the respective photodetectors 19 when the illumination light having a plurality of different wavelengths is irradiated. The film thickness calculator 22 calculates the film thickness of the SiO ₂ film 32 formed on the bottom of the trench 34 using the reflectance measurement data. That is, in order to calculate the film thickness of the SiO ₂ film 32, the processing apparatus 20 measures the reflectance for each wavelength based on an image with illumination light of a certain wavelength and an image with illumination light of a different wavelength. Ask for data. Then, the processing unit 20 refers to the calculation data the relationship between wavelength and reflectance are shown respectively for each film thickness of the SiO ₂ film 32, and approximates the thickness of the SiO ₂ film 32 from the measurement data calculate. This film thickness measurement method will be described in detail later.

In the present invention, the reflectance is measured using the confocal optical system 101 while switching the illumination wavelength. Thereby, scattered light from the edge of the trench 34 and reflected light from the upper surface of the Si substrate 31 can be removed, and the film thickness t of the SiO ₂ film 32 provided on the bottom of the trench 34 can be measured. In this way, the film thickness of the SiO ₂ film 32 can be measured in a short time without contact and without destruction. It is possible to measure the film thickness at the bottom of the trench 34 having a high aspect ratio. Since the measurement is based on the relative change in the reflection intensity depending on the wavelength, it is not affected by the pseudo focus which has been a problem in the film thickness measurement by detecting the reflected light focus position. Moreover, it can respond to the film thickness measurement of the surface with a level | step difference larger than the focal depth of an optical microscope. By measuring the absolute reflectance of the bottom of the trench 34 without the SiO ₂ film 32, it is also possible to determine the surface roughness of the trench bottom from the attenuation of the reflectance.

With the gain adjustment of the photodetector 19 and the like fixed, images of the same field of view are acquired for the reference sample Si and the sample. Here, the measuring apparatus 100 captures an image by changing the wavelength as described above. That is, a reference sample and an image of the sample are acquired for each of the illumination wavelengths of 436 nm, 486 nm, 514 nm, 546 nm, 578 nm, and 633 nm. Then, an average value of luminance values for area specified on the image ^{I Si,} and ^{I sample.} Then, the absolute reflectance R ^sample of the sample 30 is obtained by substituting the average luminance values I ^Si and I ^sample into the equation (14). The absolute reflectance R ^sample is obtained for each illumination wavelength.

For verification of the NA correction calculation method, measurement is performed using the SiO ₂ film 32 on a normal silicon wafer instead of the trench. That is, the substrate 31 having the SiO ₂ film 32 formed on the entire flat surface is used. The film thickness of the SiO ₂ film 32 measured with a spectroscopic ellipsometer is about 505 nm. The optical model is the same as in FIG. The NA of the objective lens 17 is 0.3.

  The reflectance obtained from the equation (14) is defined as a measured reflectance Rm. Moreover, the reflectance calculated | required from Formula (9) is set to calculated reflectance Rc. FIG. 7 shows a graph plotting the measured reflectance Rm and the calculated reflectance Rc. In FIG. 7, the horizontal axis represents the wavelength λ of the illumination light, and the vertical axis represents the reflectance R. In FIG. 7, the measured reflectance Rm is plotted as a black square. Further, in FIG. 7, the calculated reflectances Rc when the film thickness t is 490 nm, 504 nm, and 515 nm are indicated by a dotted line, a solid line, and a one-dot chain line, respectively.

A residual Σ between the measured reflectance Rm and the calculated reflectance Rc is defined as in Expression (15).

The residual Σ calculated according to the film thickness t is shown in FIG. In FIG. 8, the horizontal axis represents the film thickness, and the vertical axis represents the residual Σ. From the graph shown in FIG. 8, since the residual Σ is minimized at t = 504 nm, the film thickness t can be determined to be 504 nm. Thus, the film thickness that minimizes the residual Σ can be estimated as the film thickness (analytical film thickness) of the SiO ₂ film 32. When illuminated with 6 wavelengths, it is estimated that the film thickness can be measured with an error of about ± 2 nm. In this way, the film thickness can be accurately measured by correcting the angular distribution of the illumination light by the NA of the objective lens 17 using the equation (13).

  FIG. 9 shows the result of film thickness analysis when the objective lens 17 having a different NA is used. In the graph of FIG. 9, the horizontal axis represents the NA of the objective lens 17, and the vertical axis represents the analysis film thickness. In the graph of FIG. 9, the data before NA correction is plotted in white, and the data after NA correction is plotted in black. Regardless of the magnification of the lens, the analysis film thickness before correction decreases as the NA increases. That is, as the NA increases, the influence of the angular distribution of illumination light increases, and the analysis film thickness decreases from the actual film thickness (505 nm). On the other hand, the analysis film thickness after the NA correction is constant in the vicinity of about 505 nm. Therefore, it can be confirmed that the NA correction works effectively.

  Since the film thickness can be obtained for each pixel of the image by the same method as described above, the film thickness distribution can be displayed.

[Measurement of absolute reflectance of trench shape]
Next, measurement of the absolute reflectance of the SiO ₂ film 32 provided at the bottom of the trench 34 will be described. When measuring the absolute reflectance of the SiO ₂ film 32 at the bottom of the trench, the amount of incident light is limited by the aspect ratio of the trench 34. Therefore, in the formula (14) in the case of a flat sample, the luminance value of the reflected image cannot be converted into the reflectance.

  The limitation of reflected light by the trench structure will be described with reference to FIG. FIG. 10 is a diagram showing the incident light incident on the sample and the reflected light reflected on the sample in a simplified manner. In FIG. 10, the reflection on the planar structure is shown on the left side, and the reflection on the trench structure is shown on the right side. In the case of a planar structure, all of the incident light is received by the photodetector 19 as reflected light.

  On the other hand, in the case of the trench structure, only a part of the incident light is received by the photodetector 19 as reflected light. That is, since the incident light does not reach the bottom of the trench, the luminance value at the photodetector 19 is lowered. Therefore, even if the surface of the trench structure has the same reflectance as that of the planar structure, the reflection intensity (luminance value) becomes small. Therefore, if the equation (14) is used as it is, the absolute reflectance becomes small. That is, it is estimated lower than the actual reflectance.

  Therefore, a coefficient for converting the reflection intensity at the trench structure into an absolute reflectance is introduced. This coefficient is defined as a trench coefficient Tr. The trench coefficient Tr is a correction coefficient that takes a value corresponding to the shape of the trench. For example, the trench coefficient Tr mainly depends on the aspect ratio β of the trench 34 and the NA of the objective lens 17. The trench coefficient Tr and the aspect ratio β can be expressed by the following expressions (16) and (17), respectively.

As shown in FIG. 11A, the absolute reflectance of the flat reference sample 40 is R ^ref and the reflection intensity is I ^ref . As shown in FIG. 11B, the absolute reflectance at the bottom of the trench 34 without the SiO ₂ film 32 is R ^{Tr — 0} and the reflection intensity is I ^{Tr — 0} . As shown in FIG. 11C, the absolute reflectance of the trench 34 provided with the SiO ₂ film 32 is R ^Tr_t and the reflection intensity is I ^Tr_t .

Absolute reflectance ^{R Tr_0} of the SiO ₂ film 32 is not trench 34 on the bottom, the absolute reflectance ^{R Tr_T} trench 34 there is SiO ₂ film 32, respectively, formula (18), when written as equation (19) To do.

  If the reflection intensity of the trench 34 without a film can be measured, the trench coefficient Tr can be calculated backward from the equation (18) using the theoretical value of the absolute reflectance. However, the filmless trench 34 is not always prepared.

Equation (18) is also the reflection when the bottom of the trench 34 is Si. Therefore, the trench coefficient Tr can be obtained by calculating the ratio between the reflection intensity of the trench 34 and the reflection intensity of the plane from the model as shown in Expression (20).

[Calculation of trench coefficient]
In FIG. 10, the relationship between the incidence and reflection of the cross section of the trench 34 is shown. However, in order to quantitatively calculate the relationship between the incidence and the reflection, it is necessary to deal with it three-dimensionally. Therefore, reflection with respect to the origin is considered with the coordinate system of incidence and reflection as spherical coordinates as shown in FIG.

It is assumed that the incident light has a uniform angular distribution, and the reflection intensity with respect to the minute solid angle dΩ is integrated within the solid angle range. That is, the received light intensity (received light amount) P _tot can be obtained using the following equation (21).

(For flat surface)
The reflection when the sample 30 is a plane, that is, the reflection at the reference sample 40 will be described with reference to FIG. FIG. 13 shows an XY plan view, an XZ sectional view, and a YZ sectional view for explaining the case of a planar sample (reference sample 40). The Z direction is the optical axis direction, and the X direction and the Y direction are directions perpendicular to the Z direction. As shown in FIG. 13, the angular distribution is calculated in the range of a cone having a radius L and a height d. Here, as shown in Expression (22), the maximum incident angle given by the NA of the objective lens 17 is θm.

Then, reflection can be received within a range of ± θm. At this time, Φ is in the range of 0 to 2π. The received light intensity P ^Flat _tot can be expressed by Expression (23).

(For trench)
The case of the trench 34 having a width W and a depth d will be described with reference to FIG. FIG. 14 shows an XY plan view, an XZ sectional view, and a YZ sectional view for explaining the case of a trench. In FIG. 14, the Y direction is the longitudinal direction of the trench. The size of the trench 34 in the Y direction is assumed to be sufficiently longer than the radius L of the cone. The size of the trench in the X direction is the width W.

Of the cone distribution with radius L and height d, consider the angular range that can be reflected at the bottom of the trench 34. As in the XY plan view of FIG. 14, the incident angle θm changes depending on the azimuth angle Φ. Focusing on the range of 0 <Φ <90 °, the limit at which the cone contacts the edge of the trench 34 is Φc. Therefore, up to 0 <Φ <Φc °, the incident angle is restricted by the trench opening, and θm is expressed by Equation (24). On the other hand, when Φc <Φ <90 °, there is no restriction on the aperture of the incident angle, so θm is as shown in Equation (25). However, Φc is represented by the following equations (26a) and (26b) depending on the size of the cone and the trench 34.

The light receiving intensity P ^Tr _tot in the trench 34 can be expressed by Expression (27).

(In the case of VIA)
Next, the case where the concave portion formed in the sample 30 is not the trench 34 but a circular VIA 35 will be described with reference to FIG. The incident angle is determined by the opening of the VIA 35. Here, it is assumed that the opening of the VIA 35 is circular with a radius W and has a depth d. When the incident angle θm is defined as in Expression (28), the received light intensity P ^VIA _tot in the ^VIA 35 can be expressed in Expression (29).

[Estimation of angle dependence of reflectivity R (θ)]
The incident angle dependence of the reflectance of Si at the bottom of the trench 34 will be described. The incidence angle dependency of the reflectance R can be calculated from the formula of the amplitude reflectance r in the case of oblique incidence. In the case of oblique incidence, as shown in FIG. 16, the calculation is divided into s-polarized light in the vibration direction parallel to the paper surface and p-polarized light in the vertical vibration direction. As an unpolarized reflectance, an average of the reflectances of s-polarized light and p-polarized light is obtained. That is, if the reflectance of s-polarized light is Rs and the reflectance of p-polarized light is Rp, the reflectance Rav = (Rs + Rp) / 2 in the case of non-polarized light.

  As an example, FIG. 17 shows the calculation result of the incident angle dependence of the reflectance of Si when the wavelength is 546 nm. In FIG. 17, the horizontal axis indicates the incident angle, and the vertical axis indicates the reflectance. FIG. 17 shows Rs, Rp, and Rav, respectively. The reflectance Rav, which is an average value of Rs and Rp, changes gently.

  The result of calculating the average value Rav of the reflectance for each wavelength is shown in FIG. FIG. 18 shows the reflectance Rav calculated for wavelengths 436 nm, 546 nm, and 633 nm. In FIG. 18, the horizontal axis represents the incident angle, and the vertical axis represents the reflectance Rav. Thus, the reflectance Rav changes according to the wavelength. Accordingly, data obtained by calculating the incident angle dependence of the reflectance R with respect to the measurement wavelength is used as a reference table. That is, for each of 436 nm, 486 nm, 514 nm, 546 nm, 578 nm, and 633 nm, a reference table showing the relationship between the incident angle and the reflectance Rav is prepared.

Thus, the incident angle dependence of the reflectance Rav in FIG. 18 is difficult to handle analytically. Therefore, the reference table can be created by numerically calculating Equation (30). Alternatively, the incident angle dependence of the reflectance Rav may be expressed by a polynomial approximation function.

[Calculation of trench coefficient]
Using the numerical calculation result of Expression (30), P ^Flat _tot of Expression (23) and P ^Tr _tot of Expression (27) are calculated. Then, the received light intensity ratio P ^Tr _tot / P ^Flat _tot is obtained as in equation (31). The ratio of the received light intensity P ^Tr _tot / P ^Flat _tot becomes the trench coefficient Tr.

  A calculation example in which numerical calculation is performed with the angular increment set to 1 ° in Expression (31) is shown in the table of FIG. Here, the width W of the trench 34 is 5 μm, and the NA of the objective lens 17 is 0.8. Furthermore, the trench coefficient Tr when the depth d is changed to 1 μm, 2 μm, 5 μm, 10 μm, 25 μm, 50 μm, 75 μm, and 100 μm is shown. The trench coefficient Tr is calculated for each illumination wavelength.

If the NA of the objective lens 17 is 0.8, the trench coefficient Tr cannot be ignored in the trench 34 having an aspect ratio of 1 or more. That is, in the reflection of Si, when the aspect ratio is 1 or more, the trench coefficient Tr is considerably larger than 1. Therefore, the light reception intensity P ^Tr _tot in the trench structure is greatly reduced as compared with the light reception intensity P ^Flat _tot in the plane reflection. Therefore, it is understood that correction with the trench coefficient Tr is necessary.

  In order to confirm the calculation result of Equation (31) in more detail, the relationship between the aspect ratio and the trench coefficient Tr is shown in FIG. In FIG. 20, the horizontal axis represents the aspect ratio of the trench 34, and the vertical axis represents the trench coefficient Tr. Here, calculation is performed with NA = 0.8. The aspect ratio and the trench coefficient are calculated for each of the illumination wavelengths 436 nm, 486 nm, 514 nm, 546 nm, 578 nm, and 633 nm. However, since the difference due to the illumination wavelength is small, the graphs overlap in FIG.

  In the case of Si, although the difference depending on the wavelength is small, the trench coefficient Tr increases rapidly from an aspect ratio of about 1 at any illumination wavelength and cannot be ignored. The fact that the trench coefficient Tr is larger than 1 means that even if the reflection intensity detected by the light detector 19 is converted into the absolute reflectance as it is, it apparently decreases. Therefore, correction with the trench coefficient Tr is required.

  Next, FIG. 21 shows changes in the trench coefficient Tr when the NA of the objective lens 17 is changed for the illumination wavelength 546 nm. In FIG. 21, the horizontal axis represents the aspect ratio of the trench 34, and the vertical axis represents the trench coefficient Tr. FIG. 21 shows the calculation results when NA is 0.13, 0.3, 0.46, 0.55, 0.8, and 0.95. Thus, the greater the NA of the objective lens 17, the greater the trench coefficient Tr. Further, it can be seen that as the NA increases, the correction by the trench coefficient Tr cannot be ignored even when the aspect ratio is small.

An analysis example of the film thickness when the SiO ₂ film 32 having a thickness of 500 nm is provided at the bottom of the trench 34 having an aspect ratio of 5 will be described. First, FIG. 22 and FIG. 23 show the film thickness measurement results of the absolute reflectance calculated using the equation (14) from the luminance value of the designated area of the image acquired by the measuring apparatus 100. 22 and 23 are diagrams showing analysis results when the reflectance R is converted as planar reflection without performing correction with the trench coefficient Tr. Here, it is a calculation result when NA = 0.3 of the objective lens 17.

  In FIG. 22, the horizontal axis indicates the illumination wavelength, and the vertical axis indicates the reflectance. In FIG. 22, the measured values are plotted with black squares, and the calculated values are shown with solid lines. In FIG. 23, the horizontal axis indicates the film thickness t, and the vertical axis indicates the residual Σ. As shown in FIG. 23, the film thickness t when the residual Σ is minimized is 87 nm. FIG. 22 shows the calculated values when the film thickness t = 87 nm.

Next, FIG. 24 and FIG. 25 show the film thickness analysis results when the correction is performed with the trench coefficient Tr under the same conditions (the aspect ratio is 5, the thickness of the SiO ₂ film 32 is 500 nm, and NA = 0.3). . FIG. 24 shows a graph in which the measured value is corrected by the trench coefficient Tr (= 2.53). That is, FIG. 24 is a graph showing the measured value corrected by the trench coefficient Tr. Note that the measured value corrected by the trench coefficient Tr can be obtained by multiplying the actual received light intensity detected by the photodetector 19 by the trench coefficient Tr. By introducing the trench coefficient Tr in this way, it is possible to correct the received light intensity at which the photodetector 19 cannot receive light by the trench 34. Therefore, the reflectance R can be accurately estimated from the luminance value of the photodetector 19.

As shown in FIG. 25, the film thickness that minimizes the residual Σ is 502 nm. Since the actual film thickness of the SiO ₂ film 32 is 500 nm, the measurement error of the film thickness is about 2 nm. Further, the minimum value of the residual Σ is almost 0, and fitting errors can be suppressed. Thus, even with the trench 34 having a large aspect ratio, the thickness of the bottom SiO ₂ film 32 can be accurately measured.

As described above, the fitting using the trench coefficient Tr improves the fitting fit by orders of magnitude. That is, the residual Σ is reduced by using the trench coefficient Tr. Therefore, the assumed film thickness can be accurately calculated as an analysis result. In the above description, the thickness of the SiO ₂ film 32 at the bottom of the trench 34 is measured. However, as shown in FIG. 15, the thickness of the SiO ₂ film 32 at the bottom of the VIA 35 can also be measured. In this case, a coefficient corresponding to the VIA shape is used instead of the trench coefficient Tr described above. In this case, the coefficient is calculated using P ^VIA _tot in equation (28) instead of P ^Tr _tot in equation (31). As described above, the film thickness can be accurately measured by performing correction using the correction coefficient corresponding to the concave portion such as the trench 34 or the VIA 35. The correction coefficient can be set based on the shape of a recess such as the trench 34 or the VIA 35.
Furthermore, a correction coefficient can be set for a gap between pillars on a sample provided with a plurality of pillars. That is, the pillars are convex portions and the gaps between the pillars are concave portions. In this way, the correction coefficient can be set using a geometrical arrangement according to the height, interval, and shape of the pillars, with the region surrounded by the pillars as recesses. Thereby, the film thickness of the thin film provided in the area | region enclosed by the pillar can be measured.

  FIG. 26 conceptually shows a calculation model for film thickness analysis when correction is performed using the trench coefficient Tr. When correction is performed using the trench coefficient Tr, a measured value of the absolute reflectance and a theoretical value of the absolute reflectance are used.

Measurements of the absolute reflectance reflection intensity I ^Si of the reference sample ^can be determined from the reflection intensity I ⁰ of the reflected intensity I ^sample, and the background of the trench bottom (oxide film). That is, as shown in Expression (14), based on the luminance value I ^Si at the reference sample 40, the luminance value I ^Sample at the bottom of the trench, and the luminance value I ⁰ at the background, the reflectance measuring unit 21 calculates the absolute reflectance. R ^Sample is calculated. Further, the reflectance measuring unit 21 refers to the wavelength dependency R ^Si (λ) of the reflectance of the reference sample 40 to calculate the absolute reflectance R ^Sample . As the reference sample 40, a sample whose absolute reflectance has a known wavelength dependency is used. By measuring the luminance value while changing these illumination wavelengths, the absolute reflectance R ^Sample is calculated for each illumination wavelength.

Further, the reflectance measurement unit 21 corrects the measured value of the absolute reflectance R ^Sample with the trench coefficient Tr. The trench coefficient Tr can be stored in advance in the reflectance measurement unit 21 according to the shape of the trench 34. Correcting with the trench coefficient Tr makes it possible to obtain an absolute reflectance independent of the shape of the trench. In other words, the absolute reflectance at the bottom of the trench is replaced with the absolute reflectance at the plane by the trench coefficient Tr. In this way, the measured value Rm of the absolute reflectance R ^Sample is obtained.

Further, the theoretical value Rc of the absolute reflectance can be obtained by determining an optical model on a plane, a known optical constant, and an oxide film thickness. Here, as shown in FIG. 3, an optical model in which air / SiO ₂ / Si is formed is used as the optical model. The optical constant in the material of the optical model is used. As the optical constant, a complex refractive index, a refractive index, and an extinction coefficient are used. Therefore, if the film thickness t of the SiO ₂ film 32 is determined, the theoretical value Rc of the absolute reflectance R ^Sample is determined.

Then, using the film thickness t as a parameter, parameter fitting between the measured value of the absolute reflectance and the theoretical value of the absolute reflectance is performed. That is, the film thickness t is changed so that the residual Σ converges. For example, the film thickness t at which the residual Σ between the absolute reflectance measured value Rm and the theoretical value Rc is minimized is obtained by the least square method. The film thickness measurement unit 22 determines the film thickness t that minimizes the residual Σ as the analysis film thickness of the SiO ₂ film 32. Thus, the film thickness can be estimated from the measured value and the theoretical value of the absolute reflectance. Of course, regression analysis other than the method of least squares may be performed. In this way, curve fitting is performed so that the residual Σ is minimized.

  Next, a calculation model of the trench coefficient Tr will be described with reference to FIG. The trench coefficient Tr can be obtained by setting a theoretical value and an integration range of incident angles. For example, the incident angle dependency R (θ) of the reflectance is obtained with the incident intensity being constant. Here, the optical constant of Si is used. For example, the incident angle dependency R (θ) of the reflectance may be a reference table based on the equation (30) or a polynomial approximation equation. Then, the angular distribution of the reflection intensity is calculated. By integrating the angular distribution of the reflection intensity within the solid angle range, an integrated value of the reflection intensity can be obtained (see formula (21)).

Further, by setting the integration range of the incident angle, the light reception intensity P ^Flat _tot of the plane and the light reception intensity P ^Tr _tot of the trench are obtained. The integration range is set by Expression (23) to Expression (27). The trench coefficient Tr can be determined from the ratio of the light receiving intensity between the plane and the trench (P ^Tr _tot / P ^Flat _tot ).

  When calculating the trench Tr, the aspect ratio of the trench 34 needs to be known. If the aspect ratio is known in advance, the numerical value can be used. However, when the aspect ratio is unknown, three-dimensional measurement is performed with a confocal microscope using the confocal optical system 101. Thereby, the width W and the depth d of the trench 34 can be obtained.

  If the width W and depth d of the trench 34 cannot be measured accurately, the trench coefficient Tr can be taken in as a parameter for film thickness analysis. That is, in the parameter fitting of FIG. 26, not only the film thickness t but also the trench coefficient Tr is used as a parameter. By doing so, the trench coefficient Tr can be optimized. If the width W of the trench 34 is measured, the trench depth d can be determined.

In the case of the VIA 35 as shown in FIG. 15, the same calculation process can be performed to obtain the VIA coefficient. Since it is only necessary to calculate the angular distribution of the conical range with the VIA opening as the radius, it is possible to calculate by simply replacing the NA in the calculation for the plane with the apparent NA of the VIA 35. Of course, not only the circular VIA 34 and the trench 34, but also the thickness of the SiO ₂ film 32 provided on the bottom of the concave portions of various shapes can be calculated. The film thickness t can be obtained even when the opening is rectangular or has a honeycomb shape. Of course, the film thickness can be calculated not only for the Si substrate 31 and the SiO ₂ film 32 but also for other material structures. For example, in a configuration in which a transparent thin film is formed at the bottom of the trench 34, the thickness of the transparent thin film can be measured.

  The correction coefficient is a calculated value of the ratio between the received light intensity from the plane and the received light intensity from the bottom of the recess. For this reason, it is possible to convert the absolute reflectance measurement data of the bottom of the concave portion into the flat absolute reflectance measurement data. It is preferable to calculate the correction coefficient using an approximate expression function or a look-up table indicating the incidence angle dependency of the reflectance. In this way, the correction coefficient can be easily obtained. Furthermore, by calculating the correction coefficient for each illumination wavelength, the film thickness can be measured more accurately. In this case, an approximate expression function or a reference table indicating the dependency of the reflectance on the incident angle is prepared for each illumination wavelength.

  As described above, by using the confocal optical system 101, scattered light from the edge of the opening of the trench 34 and reflected light from the upper surface of the Si substrate 31 can be removed. Thereby, the net reflected light at the bottom of the trench 34 can be accurately measured. Furthermore, the film thickness of the thin film at the bottom of the trench can be measured in a short time without contact and without destruction. Further, since the measurement is based on the relative change in the reflection intensity depending on the wavelength, the pseudo focus, which has been a problem in the film thickness measurement by detecting the reflected light focus position, is not affected.

Relationship between the wavelength and reflectance with reference to the computation data shown respectively for each film thickness of the SiO ₂ film 32, is calculated from the measured data by approximating the thickness of the SiO ₂ film 32. Then, in the approximation for obtaining the thickness of the SiO ₂ film 32, a correction coefficient corresponding to the shape of the trench 34 is introduced. Therefore, the film thickness can be accurately measured.

  Further, in this embodiment, instead of dispersing the reflected light as in the spectral film thickness meter, the reflection intensity of the sample is measured using the illumination light as monochromatic light. Each time the illumination wavelength is changed, the reflection intensity is measured with respect to the same visual field. Therefore, the wavelength dependence of the reflectance, that is, the reflection spectrum can be obtained. Furthermore, even in an optical system in which the focal position is shifted depending on the illumination wavelength, measurement at the in-focus position can be performed at any illumination wavelength. Therefore, the film thickness can be accurately measured.

  By using the confocal optical system 101, the shape of a recess such as a trench can be measured three-dimensionally. If the thickness of the thin film at the bottom is sufficiently small with respect to the depth of the recess, an approximate aspect ratio can be estimated. From the NA of the objective lens 17 used for the reflectance measurement and the aspect ratio of the trench 34, a coefficient to be converted into the absolute reflectance can be calculated for each wavelength.

  A spectral spectrum can be obtained for each pixel of the observation image. The luminance value detected by the photodetector 19 is converted into an absolute reflectance. Using the measured value Rm of the absolute reflectance based on the detected luminance value and the calculated value Rc of the absolute reflectance, fitting using the film thickness as a parameter is performed. By doing so, a film thickness of 1 μm or less can be obtained with high accuracy. It is also possible to perform fitting using the aspect ratio as a parameter.

  Note that the switching of the illumination wavelength may not be continuous. For example, it may be several wavelengths between 400 and 700 nm. One wavelength can be measured if it is determined whether or not there is an etching residue of 100 nm. In this case, for example, it may be determined whether or not the minimum value of the residual Σ in parameter fitting is larger than a threshold value.

  In the confocal optical system 101, when the illumination wavelength is switched, it is conceivable that the in-focus position changes, so that a change in luminance value is expected. The shift amount of the in-focus position due to each wavelength is stored in the processing device 20 in advance. When the wavelength is switched, the z position of the sample 30 or the objective lens 17 is automatically adjusted by the shift amount. By doing so, the shift of the in-focus position can be canceled. Alternatively, an omnifocal image may be formed by z scanning at each illumination wavelength.

  As the reference sample 40 of absolute reflectance, silicon, quartz glass, or the like can be used. That is, by measuring a sample having a known reflection spectrum with the measuring apparatus 100 in advance, the wavelength dependence of the optical system itself of the measuring apparatus 100 is corrected by calculation.

  The confocal optical system 101 is used. Therefore, the surface shape can be independently measured by selecting an illumination wavelength that reflects on the surface of the sample 30 and has a low transmittance to the inside. By measuring the film thickness distribution and the surface shape distribution independently, the substrate surface shape (interface between the substrate and the film) can be obtained.

  In addition, when the film thickness is several μm or more, the difference in focus position can be detected with a confocal microscope, so that the film thickness can be measured as usual by reading the focus position. That is, according to the present invention, when the film thickness is 1 μm or less, the reflection intensity can be measured while switching the wavelength of the illumination light as described above. Can be switched. Furthermore, even when the surface has undulations exceeding the depth of focus, the reflectance of the entire surface can be measured by creating an omnifocal image with a confocal microscope.

  In the above example, the reflectance is intermittently measured for a plurality of wavelengths, but the wavelengths may be continuously switched. By using a continuous wavelength light source, a large number of wavelengths to be measured can be selected. In this case, if a continuous curve can be obtained without using an approximation that perfectly matches the calculated reflection data with the measured data, the film thickness can be measured by a general method that approximates the position of the maximum or minimum reflectance. Can be calculated. In other words, when using the continuous wavelength method, the film thickness can be calculated using only the peak wavelength, and therefore it is not necessary to accurately correct the reference sample with a known reflectance.

10 light source unit 11 light source 12 interference filter 13a, 13b, 13c lens 14 slit 15 beam splitter 16 oscillating mirror 17 objective lens 18 stage 19 photodetector 20 processing unit 30 sample 31 substrate 32 SiO ₂ film 34 trench

Claims (10)

A film thickness measuring device for measuring the film thickness of a thin film formed on the bottom of a recess provided on a sample,
A light source unit capable of switching at least illumination light of the first wavelength and illumination light of the second wavelength;
A photodetector that detects reflected light reflected by the sample and acquires an image in a predetermined region of the sample;
A confocal optical system that guides illumination light from the light source unit to the sample and guides the reflected light from the sample to the photodetector;
In order to calculate the film thickness of the thin film, reflection on the first wavelength and the second wavelength based on the first image by the illumination light of the first wavelength and the second image by the illumination light of the second wavelength. A processing unit for obtaining rate measurement data,
The confocal optical system has an objective lens that collects the illumination light and irradiates the sample, and passes the reflected light;
The processing unit refers to calculation data in which the relationship between the wavelength and the reflectance is shown for each film thickness of the thin film, and calculates the film thickness of the thin film from the measurement data,
In the approximation of the film thickness, a film thickness measurement apparatus in which a correction coefficient corresponding to the shape of the recess and the NA of the objective lens is introduced.   The film thickness measuring apparatus according to claim 1, wherein the correction coefficient is a calculated value of a ratio between a received light intensity from a plane and a received light intensity from the bottom of the recess.   The film thickness measuring device according to claim 2, wherein the light reception intensity from the plane and the light reception intensity from the bottom of the concave portion are calculated using an approximate expression function or a reference table indicating an incidence angle dependency of the reflectance. .   The film thickness measuring apparatus according to claim 1, wherein the correction coefficient is set according to the wavelength of illumination light.   The film thickness measuring apparatus according to any one of claims 1 to 4, wherein the processing unit calculates the film thickness of the thin film from the measurement data and the calculation data by a least square method. A film thickness measuring method for measuring the film thickness of a thin film formed on the bottom of a recess provided on a sample,
Via a confocal optical system including an objective lens, the sample is irradiated with switched illumination light of the first wavelength and illumination light of the second wavelength,
The reflected light reflected by the sample is detected through the confocal optical system to obtain a first image by the light of the first wavelength and a second image by the light of the second wavelength,
In order to calculate the film thickness of the thin film, based on the first image and the second image, the measurement data of the reflectance with respect to the first wavelength and the second wavelength, respectively,
Referring to the calculation data in which the relationship between wavelength and reflectance is shown for each film thickness of the thin film, the film thickness of the thin film is approximated and calculated from the measurement data,
In the approximation of the film thickness,
Refer to the film thickness analysis result when using the objective lens with different NA,
A film thickness measuring method in which a correction coefficient corresponding to the shape of the concave portion and the NA of the objective lens is introduced.   The film thickness measurement method according to claim 6, wherein the correction coefficient is a calculated value of a ratio between a light reception intensity from a plane and a light reception intensity from the bottom of the recess.   The film thickness measurement method according to claim 7, wherein the received light intensity from the plane and the received light intensity from the bottom of the recess are calculated using an approximate expression function or a reference table that indicates the incident angle dependence of the reflectance. .   The film thickness measurement method according to claim 6, wherein the correction coefficient is set according to the wavelength of illumination light.   The film thickness measuring method according to claim 6, wherein the film thickness of the thin film is calculated from the measurement data and the calculation data by a least square method.

Images