TW201007117A - Film thickness measuring device and film thickness measurement method - Google Patents

Abstract

The present invention relates to a film thickness measuring device and a film thickness measurement method. A model part (721A) based on parameter update instructions of the fitting part (722) is used to sequentially update the first one layer film thickness d1, and then based on the updated first layer film thickness d1, function of illustrating theoretical reflectivity is updated. Further, according to the updated function the model part (721A) calculates the theoretical reflectivity of various wavelengths (spectrum). According to this method, the first layer film thickness d1 is determined by using a fitting method. When the fitting method can not converge in a preset number of iteration, then Fourier transform is used to determine the first layer film thickness d1.

Description

201007117 VI. [Technical Field] The present invention relates to a film thickness measuring device and a film thickness measuring method, and more particularly to a film thickness measuring structure of a plurality of layers of a test object formed on a substrate method. [Prior Art]

In recent years, in order to achieve low power consumption and high speed in a complementary metal oxide semiconductor (CMOS) circuit, the insulating layer is covered with silicium (silic〇n 〇n insulat〇r, s〇i). The substrate structure is attracting attention. The SOI substrate is provided with an insulating layer (tantalum layer) such as SiO2 (Si〇2) between two 矽(sU丨(3)^, si) substrates, so that the (four) bonding surface formed in one of the 矽 layers The parasitic diode and the electrostatic capacitance generated between the other layer (substrate) and the like can be reduced. In the above, the method for manufacturing the SOI substrate is to form an acidified film on the surface of the silicon wafer and paste the other germanium wafer to make the acidified film between $AI^, in between, further The silicon wafer forming the circuit component is ground to a predetermined thickness. According to this grinding process, in order to control the thickness of the silicon wafer, the mold thickness needs to be continuously monitored. The method for measuring the film thickness of the polishing process is disclosed in Japanese Patent Application Laid-Open No. Hei. No. Hei. No. Hei. No. Hei. No. Hei. No. Hei No. 05-308096, which is a method for measuring the film thickness of the polishing process, that is, using Fourier's, the infrared-spectrum infrared spectrophotometer (Fourier transform infrared spectrophotometer) , FTIR) method. In addition, Japanese Laid-Open Patent Publication No. Hei. No. 0 discloses a method of measuring a reflection spectrum "pea·) by using a distributed multi-channel spectroscope. In the above, a method for measuring an infrared ray from an infrared ray source is irradiated onto a polishing object through a polishing object to detect the reflected light. Further, in Japanese Patent Laid-Open Publication No. 2-221-2842, it is disclosed. The method of measuring the film thickness of the ruthenium film is based on the interference of the infrared ray of 9 Um (micrometer) or more on the surface of the ruthenium film, and then the reflection of the light reflected from the surface of the shi shi film and the reflected light inside the shi shi film. Further, in Japan Patent Publication No. 20034 141〇7 discloses a method for measuring the interference of (X) thickness by using infrared rays as the light of the shirt. However, the Japanese Patent Laid-Open Publication No. (10)-3G691G and the Japanese Patent Publication No. 3-6_6 In the exposed film thickness method, the film thickness relative value of the sample based on yttrium cannot be determined, and thus the absolute thickness of the film thickness cannot be determined. In the item disclosed in GG5-1, 9920, it is assumed that the refractive index does not depend on the wavelength and is based on the self-regressive model (m〇de refractive index and ancient, "&

Half, with wavelength dependence, the error of the I. In addition, the method of measuring the wavelength is the same as the method of measuring the damage of the patent: 2°〇3]141°7. (1) In the measurement method according to the Japanese Patent Laid-Open Publication No. 2GG2-22842G, it is necessary to form a through portion on the non-broken sample exposed by the 敎 object, and the film is continuously measured in a non-destructive manner by the I method. thick. SUMMARY OF THE INVENTION In order to solve this problem, an object of the present invention is to provide a film thickness measuring device and a film thickness measuring method for measuring a film thickness having higher accuracy. The film thickness measuring device of the present invention includes a light source, a spectrometry portion, a first determining unit, a converting unit, an analyzing unit, and a second determining unit. The light source irradiates the measurement light having a predetermined wavelength range to the object to be tested which forms a plurality of layers on the substrate. The object to be tested includes a first layer closest to the light source and a second layer adjacent to the first layer. The spectrometry portion is used to obtain the wavelength distribution characteristic of the reflectance or the transmittance according to the light reflected by the object to be tested or the light penetrating the object to be tested. The first-determining means performs the adaptation of the wavelength distribution characteristic by the model formula to determine at least the film thickness of the first layer, and the model has the film thickness of each layer included in the object to be tested. The conversion unit converts the correspondence between the values of the reflectance or the transmittance of each wavelength and wavelength of the wavelength distribution characteristic into a correspondence relationship between the wave number of each wavelength and the converted value calculated according to the predetermined relational expression, and is used to generate The law of arrogance and publicity LA μ inverse knife cloth characteristics. The analyzing unit is configured to obtain an amplitude value of each wavenumber component included in the wave number distribution characteristic. The second decision unit 用以 is used to determine at least the film thickness of the first drawer layer according to the large wave number component of the amplitude value included in the wave number distribution characteristic. Thereafter, the first decision unit and the second decision unit are selectively made valid. Preferably, the film thickness measuring device further comprises a third determining unit, wherein the value of the media thickness of the first layer determined by the second determining unit is used to set a model, and the long-distribution characteristic is adapted to be used for The film thickness of the second layer is determined, and the model has a film thickness of each layer included in the to-be-contained material. Preferably, the second decision unit is valid when the first - 洫 s _ red < 疋 疋 疋 兀 兀 未 未 未 未 未 未 未 。 第二 第二 第二 第二 第二 第二201007117 Better' model includes a wavelength correlation function used to represent the refractive index. More preferably, the predetermined wavelength range includes the wavelength of the infrared light domain. More preferably, the parsing unit comprises a Fourier transform unit for performing discrete Fourier transform on the wavenumber distribution characteristics. Preferably, the analysing unit utilizes an optimization method for obtaining an amplitude value of the wavenumber component included in the wave number distribution characteristic/the other aspect, the film thickness measuring method of the present invention, including the illuminating step, reference 2 The measuring light having the wavelength range of the gamma is irradiated to the object to be tested formed on the base plate, and the object to be tested includes the illusion object including the first layer closest to the light source and the second layer adjacent to the layer. The method for measuring the film thickness of the step-step includes: a wavelength distribution characteristic obtaining step of obtaining a wavelength distribution characteristic of the reflectance or the transmittance according to the light reflected by the object to be tested or the light penetrating the object to be tested; The determining step 'utilizes the model to perform the adaptation of the wavelength distribution characteristic to determine at least the film thickness of the first layer, the model has the thickness of each layer included in the object to be tested; and the generating step 'each of the wavelength distribution characteristics Correspondence between the values of the reflectance or the transmittance of the wavelength and the wavelength 'converted to the correspondence between the wave number of each wavelength and the converted value calculated according to the predetermined relationship, for generating the wave number distribution characteristic; the amplitude value obtaining step For obtaining the wave number distribution characteristic: the amplitude value of each wavenumber component; the second determining step is for determining at least the first film according to the large wave number component of the amplitude value included in the wave number-characteristic Thick; and an activation step for selectively making the first decision unit and the second decision unit valid. According to the present invention, it is possible to have a film thickness of a test object having higher accuracy. The above described objects, features and advantages of the present invention will become more apparent from the following description. [Embodiment]

The symbol indicates 'and the repeated description is omitted. Device Architecture

Fig. 1 is a view showing a film thickness measuring device 100 according to an embodiment of the present invention, which is generally used for measuring the film thickness of each layer for a single layer or a laminated structure. The film thickness measuring apparatus 100 of this embodiment is particularly suitable for film thickness measurement of a test object having a relatively large thickness (usually 2 ΙΟΟΟμπι). Specifically, the thickness measuring device 1 is a micro-spectroscopic measuring device that irradiates light onto the object to be tested, and reflects light reflected by the object to be tested. Wavelength distribution characteristics (hereinafter also referred to as "spectrum" ))), which can be used to determine the film thickness of each layer constituting the object to be tested. Further, it is not limited to the measurement of the film thickness, and can be used for the measurement of the reflectance of each layer (absolute and relative) and the analysis of the layer structure. Further, the spectrum of the light that penetrates the object to be tested (the spectrum of the transmitted light) may be used instead of the spectrum of the reflected light. In the specification, a case is described in which a substrate or a substrate as a substrate to be tested is formed with one or more layers. In one embodiment, the object to be tested is a substrate having a laminate structure such as a thick substrate substrate such as a Si (Si) substrate, a glass substrate, and a sapphire substrate, and an SOI substrate 7 201007117. In particular, the film thickness measuring apparatus 100 of the present embodiment is suitable for measuring the thickness of the germanium substrate after the cutting and polishing, the thickness of the germanium layer (active layer) of the S01 substrate, and the chemical mechanical polishing (CMP). The thickness of the substrate is the thickness of the substrate. Referring to Fig. 1, the film thickness measuring device 100 includes a measuring light source 10, a condensing lens (col 1 i mating lens) 12, a filter (cutting fi Iter) 14, an imaging lens 16 and 36, and an aperture 18, Beam splitter 20 and 30, observation light source 22, optical fiber 24, injection portion 26, pinhole mirror 32, axis conversion mirror 34, observation camera 38, display Part 39, and data processing part 7〇. In order to obtain the reflectance spectrum of the test object, the light source 1 is used as a light source for generating light having a predetermined wavelength range, and particularly has a wavelength component in the infrared light field (for example, 900 nm (nano) to i6) A light source of 〇〇nm, or 1470 nm to 160 〇 nm). A hologram lamp (Hal〇gen lamp) is usually used as a light source for measurement. The focus lens 12, the filter 14, the imaging lens 16, and the diaphragm 18 are disposed on the optical axis for connecting the measurement light source 1 and the beam splitter 3 to optically adjust the light source from the measurement unit. The measured light that is emitted. Specifically, the focus lens 12 is an optical light at which the measurement light of the measurement light source is initially incident, and the measurement light that propagates as the diffusion light is refracted into parallel light. The measurement light passing through the focus lens 丨 2 is incident on the filter i 4 . The filter 14 is for blocking unwanted wavelength components in the measurement light. The filter 14 is usually formed of a multilayer film deposited on a glass substrate or the like. In order to adjust the beam diameter of the measuring light, the imaging lens 丨6 will pass through the filter 1* 201007117 Μ = light is converted from flat light to concentrated light. The light that has passed through the imaging lens 16 is incident on the aperture 18. The aperture 18 adjusts the amount of light of the measurement light to a predetermined amount for emission to the beam splitter 3〇. More preferably, the aperture 18 is configured in accordance with the imaging lens 16 (four) for the position of the illuminating image. Furthermore, the apricot „ ^ $ ' of the aperture 18 is appropriately set to correspond to the measurement glazing incident on the object to be tested and the necessary light intensity, etc. The surface light source 22 is a light source for generating observation light for The object is focused and the measurement position is confirmed. And the observation light source 22 is selected to include a wavelength at which the object to be detected may be reflected. It is observed that 99, 2^ is connected to the emitting portion 26 by the optical fiber 24, so that the observation light of the observation light propagates through the optical fiber 24 as the optical waveguide, and is emitted from the emitting portion 26 toward the spectroscope 2. 22 Μ 26 includes a part of the heart of the mask to cover the light source for observation. Partially observing the light 'projects the predetermined reference image onto the nose. For the object to be tested (such as a glass substrate, etc.) formed without any pattern on the surface, the observation reference image is also easy to be filled, and any shape of the reticle can be used, like 疋(4) Use concentric circles and cross-shaped patterns. In the middle, the direct (four) force observation light beam 22 produces the observation beam of the light beam back '26 reverse t (light quantity) is approximately the same 'and part of the observation light is masked; the light two (shadow) ^, will make observation The light forms a region with a first intensity of about zero on the beam profile (projected with the observation reference image in the region of the object to be tested). The shaded area is used as the stage of the sample stage on which the object to be tested is placed, and its placement surface is flat 9 201007117. For example, the movable mechanism 5 in which the stage 5 is mechanically coupled can be freely driven in three directions (X direction, γ direction, and z direction). The movable mechanism 51 is usually composed of a three-axis servo motor and a servo driver for driving each of the feeding motors. Further, the movable mechanism 51 is driven by a user or a control device (not shown) or the like to respond to the position of the stage. According to the driving of the stage 50, the positional relationship between the object to be tested and the objective lens 40 to be described later is changed. The objective lens 40, the beam splitter 3'', and the pinhole mirror 32 are disposed on the optical axis 延伸ι extending in the direction perpendicular to the flat surface of the stage 50. The reference light generated by the reference light source 30 reflects the measurement light for converting the propagation direction to the lower side of the paper surface toward the optical axis AX1. Further, the beam splitter 30 allows the reflected light of the object to be detected which propagates over the plane of the optical axis AX1 to penetrate through the light.

On the other hand, the dichroic mirror 20 reflects the observation light generated by the observation light source 22, and converts the propagation direction to the right of the paper surface toward the optical axis Αχ2. That is, the beam splitter 30 has a function of a light injection portion for injecting observation light into a predetermined position from the optical source 10 for measurement to the optical path of the objective lens 4 of the light collecting optical system. The measurement light and the observation light synthesized by the spectroscope 2 are reflected by the spectroscope 30 and then incident on the objective lens. In particular, since the measurement light has a wavelength component of the infrared light region, and the observation light has a wavelength component of the visible light region. Therefore, from the visible light to the infrared light, the beamsplitters 20 and 30 are capable of maintaining the target value of the transmission/reflection characteristics. The objective lens 40 is a collecting optical system for collecting the measurement light and the observation light propagating below the plane of the paper 10 201007117 of the optical axis aX1. That is, the objective lens collects the light and the observation light for imaging on the object to be tested or its adjacent position. Further, the objective lens 40 is a magnifying lens having a predetermined magnification (for example, 1 〇, 2 〇, 倍, 4 〇, etc.). Therefore, this magnifying lens enables the measurement area of the optical characteristics of the measuring light to be miniaturized as compared with the beam section of the human lens that is incident on the objective lens 40. In addition, the measurement light incident on the object to be tested by the objective lens 40 and

The observation light is transmitted through the reflection of the object to be tested, and is transmitted above the optical axis. The reflected light passes through the objective lens 4 and then passes through the beam splitter 3 to reach the pinhole mirror 32. The pinhole mirror 32 has a function of a light separating portion, and separates the measured reflected light and the observed reflected light from the reflected light generated by the object to be tested. Specifically, the pinhole mirror 32 includes a reflecting surface for reflecting the reflected light from the object to be tested and propagating above the plane of the optical axis AX1, and forming a hole at the center of the intersection of the reflecting surface and the optical axis AX1. Part (pinhole) 32a. The measured reflected light generated by the measurement light of the measuring light source 10 is reflected by the object to be tested, and the diameter of the pinhole 32a formed at the position of the pinhole mirror 32 becomes smaller. Smaller. Further, the pinhole mirror is arranged to match the measurement light and the observation light to the imaging positions of the measured reflected light and the observed reflected light which are generated by the reflection of the object to be detected. In this configuration, the reflected light generated by the object to be tested is incident on the spectrometry portion 60 through the pinhole 32a. On the other hand, the direction of propagation of the residual reflected light is converted to be incident on the axis conversion mirror 34. The spectrometry portion 60 is for measuring the reflection seat #4^, 3曰 of the reflected light by the pinhole mirror 32, 201007117, and outputs the measurement result to the data processing portion 70. More detailed (.ία, for the light measurement part 6〇 including diffraction grating: er (grating) 62, detection part of the field 66 and shutter (sh coffee 68 〇 filter 66, shutter RQ η, 8, And the diffraction grating 62 is disposed on the optical axis αχ 1. The filter 66 is an optical, optically-preserving mirror for detecting the 佘G„ 疋 reflected light incident on the spectroscopic measuring portion 60 through the pinhole 32a. The wavelength is formed by limiting the wavelength outside the measurement range contained therein. In particular, it is used to block the wavelength component outside the measurement range. 快门 When the detection portion 64 is reset (reset), the shutter 68 is used to block the incident. The light is detected by a mechanical shutter that is driven by an electromagnetic force. The diffraction grating 62 splits the measured reflected light of the human beam, and then transmits the optical waveguides to the detecting portion 64. Specifically, the diffraction grating Μ is a reflective diffraction grating' that reflects each of the diffraction waves to a corresponding direction in the diffraction grating 62 having the structure at a predetermined wavelength interval, when the reflected wave is incident. Reflecting the wavelength components contained in the corresponding wavelengths, and then incident on them Further, the wavelength interval corresponds to the wavelength resolution of the spectrometry portion 60. The diffraction grating 62 is usually composed of a flat focus type spherical grating. In the measured reflected light of the diffraction grating 62, an electronic signal corresponding to the light intensity of each wavelength component is outputted by the detecting portion 64 for measuring the reflectance spectrum of the object to be tested. The detecting portion 64 is made of infrared rays. The indium gamium arsenide (InGaAs) array is composed of an optical field sensitivity. 12 201007117 The data processing material 70 obtains the characteristic procedure related to the present invention for the detection portion β4, and the film thickness of the luminosity spectrum. The data processing part is the reflectivity and layer structure of each layer of the phase. In addition: Guan:: to analyze the underlying layers of the object to be tested. After that, 7 ()% ί^ material details such as optical characteristics. 7G wheel ^ On the other hand, the film thickness of the object to be tested is first, and the reflected light reflected by the pinhole mirror 32 propagates along the optical axis AX1, and then enters the cylinder wire 4 & r_

Silly to the axis conversion mirror 34. Observe that the reflected light propagates from the optical axis AX3 to the station AY/1, to the first axis AX4. In this way, the reflected light is reflected by the optical axis AX4, and the latter is shot to the observation camera. The observation camera 38 is used as an image capturing portion for obtaining a reflection shirt image by observing the reflected light. Usually, the charged_c〇Upled device (CCD) and the complementary gold-oxygen half (four) are C〇mpiementary .Μ oxide SemiC0nduct0r' CM〇s) sensor (sens〇r). Further, the observation camera 38 generally has a visible light range sensitivity, and in many cases, the sensitivity characteristics are different from the detection portion having a sensitivity of a predetermined measurement range of 64 °. Then, the observation camera 38 self-observes the reflected light. After the reflected image is taken, the corresponding video signal is output to the display portion 39. The display portion 39 displays the reflected image on the screen based on the video signal of the observation camera 38. When the user sees the reflected image displayed on the display portion 39, the focus of the object to be tested and the position of the measurement are confirmed. The display portion 39 is usually composed of a liquid crystal display ([i. Furthermore, a viewfinder (f inder) can be provided so that the user can directly see the reflected image' instead of the observation camera 38 and the display portion 39. 13 201007117 "Analytical examination of reflected light" First, when the measurement light is irradiated onto the object to be tested, the observed and reflected light is examined mathematically and physically. Fig. 2 is a cross-sectional view showing the object to be tested 〇Bj which is a measurement target of the film thickness measuring apparatus 1() of the embodiment of the present invention. Referring to Fig. 2, an SOI substrate is taken as a representative example of the object to be tested 〇BJ. That is to say, the object to be tested 〇Bj is configured with a three-layer structure: a bismuth (si) layer 1, a base (base) layer 3 (substrate layer), and a cerium oxide (Si 〇 2) layer 2 in between ( BOX layer). Further, the film thickness measuring device 1 is irradiated with light from the upper side of the paper to the object to be tested 〇BJ. In other words, the measurement light is initially incident on the Si layer 1. For the sake of easy understanding, the reflected light generated by the reflection of the interface between the Si layer 1 and the Si 〇 2 layer 2 is considered next to the measurement light incident on the object OBJ. In the following description, each layer is represented by ί. That is, "0" is used to indicate the atmosphere such as air and vacuum, and the object to be tested 丨BJ is represented by r丨

Layer 1 and "2" indicate its Si〇2 layer 2. In addition to this, the refractive index of each layer is expressed by i, that is, the refractive index η·. Since the reflection of light is generated at the interface of each layer having different refractive indices (1), the amplitude reflectance coefficient of the P-polarized component and the S-polarized component can be selected in each interface between the ith layer and the i+1th layer having different refractive indices. And, %+1 means as follows:

Frensnel 14 201007117 r(P) — ”/+1 gas C_“丄nf+\ cos^. -\-n. cos^.+jr(S) _ cos^.-«/+1 cos^ i,i+l n. cos^. +«z-+| cos^ ill /+1 Here, it represents the incident angle of the i-th layer. According to the Snell rule described below, the incident angle can be calculated from the incident angle of the uppermost atmosphere (the second layer):

A^〇 sin = Nj sin φι In the film thickness layer with light interference, the light reflected by the reflectance shown by the above formula will travel back and forth multiple times in the layer. For this reason, the light directly reflected by the interface of the adjacent layer and the light after multiple reflections in the layer are different in phase from each other due to the different optical path lengths between the two layers, and light interference occurs on the surface of the 81-layer i. In order to show the effect of this light interference in each layer, the optical phase angle in the i-th layer can be expressed as follows:

Let f) rij cos^· Here, 4 denotes the film thickness of the i-th layer, and Λ denotes the wavelength of the incident light. In order to be more simplistic, when the light is irradiated perpendicularly to the object to be tested 〇BJ, that is, when the incident angle is =0, there is no difference between the p-polarized light and the s-polarized light, and the amplitude reflectance and film phase of the interface between the layers are not different. The angle burn is as follows: n〇+n\

15 201007117 β\=2π η\ , ν ^ y Further in the three-layer system shown in Fig. 2, the reflectivity of the object to be tested OBJ is expressed as follows:

? 4 ^ + 2yDF12 CQS2A 1 + r0 fl 2+2 于 In the above equation, for the phase angle A frequency conversion (Fourier transform), phase = (Phase factor) 〇〇 82 two pairs of reflectivity and non- Linear. ❹ Next, the phase factor C0S2 is converted into a linear function. In one embodiment, the reflectance and the following equation are converted and defined as independent variables "wavenumber converted reflectivity" and,: R': { R T^R2 2 r01+r12 Ό1

Ra +Rfj cos2 12y 2m\ Ding

V 2mni^ΐϊ1-rl22 cos 2^i

d\

Install dj 2^?7J ^dry' A ^~ indicates the light (electromagnetic wave) in the material, that is, the wave number AT (propagation number) in the layer. The wave number conversion reflectance i?' is a linear expression of the phase factor cos2, so it is linear. Here, in the equation, the wave number conversion reflectance 疋 is shown as a slice, and the wave number conversion reflectance /?, the inclination angle is expressed. In other words, for 16 201007117 phase factor C0S2 work related to frequency conversion, the wave number conversion reflectance π is a function for linearizing the reflectance and the value of each wavelength. Furthermore, the function 1/(1-, can also be used as a function of the linearization of the phase factor. Therefore, the wave number in the Si layer 1 can be defined as follows: κλΑ is here, when the Si layer 1 is Wavelength; when the L light velocity is S, and the wavelength in the vacuum and the light velocity are c, the refractive index is expressed as 叼=:e/χ.

Using the wave scale 1, the angular frequency ω, and the phase 5, the electromagnetic wave generated by the light traveling along the X direction in the Si layer i can be expressed as five M=E〇exp[攸-印+小, in other words, The propagation characteristics of electromagnetic waves in the Si layer 1 depend on the wave number. Through these relationships, it is known that there is a wavelength in the vacuum; the light of I, because of its light velocity falling within the layer, the wavelength will also become longer; «丨. Considering this wavelength dispersion phenomenon, the wavenumber conversion reflectivity can be defined as follows: • and ' ([1) = 4 + /3⁄4 cos 2 [the river by this relationship, when the wave number conversion reflectance W When the frequency conversion (Fourier transform) related to the wave scale is performed, the peak of the period component corresponding to the film thickness* is used to specify the position of the peak, and the film thickness ή can be calculated. After the correspondence between the reflectance spectrum measured by the object to be tested OBJ and the reflectance of each of the long wavelengths is converted into a correspondence relationship between the wave number calculated by each wavelength and the wave number conversion reflectance γ calculated by the above relational expression, The wave function of the wavenumber conversion reflectance containing the wave scale is performed A few feet related to the 17 201007117 frequency conversion 'after 'based on the peak appearing in the characteristics after the frequency conversion' can calculate the film thickness of the Si layer 1 constituting the object OBJ. This is because the wave number distribution characteristics are included. The amplitude value of each wavenumber component is then used to calculate the film thickness of the layer 1 based on the large wavenumber component of the amplitude value. Further, as described below, discrete Fourier transform such as fast Fourier transform (FFT) can be used. The method of conversion, or the optimization method using the maximum entropy method (hereinafter also referred to as "MEM"), is used to analyze the large wavenumber component of the amplitude value from the wave number distribution characteristic. φ in wave number conversion In the definition of the reflectance brother, /^ and the operation are the values of the interference phenomenon in the unrelated layer, but it depends on the amplitude reflectance of the refractive index „丨 of the Si layer 1 in the interface between the layers. For this reason, when the refractive index 波长 has a wavelength dispersion, the value is a function value depending on the wavelength (that is, the wave number is particularly), and therefore, it is related to the wave number X and cannot be a constant value. Therefore, in order to represent the Fourier transform, the ruler, the ~, the office and the coffee 2 Youhe, after Fourier transform with a wavenumber, 'will be used as a function of the power spectrum (Qing strike spectrum) are set to p, Pa, 匕 and F, then the following formula holds: 罾 which represents the convolution. In the formula, depends on? The composition of the 8 film thickness is relatively small, and the power spectrum F has independent peaks, so it does not affect the power spectrum? Make an impact. On the other hand, since Pb in the equation is decomposed with the power spectrum F, the film thickness component of Pb is modulated into the film thickness component of the power spectrum F. However, Pb has nothing to do with the intra-layer interference phenomenon, because it is only affected by the wavelength dependence of the refractive index of 18 201007117 in the adjacent two layers. For one point, compared with the power spectrum, the film thickness is eight K" ~ film thickness into the (four) male The film thickness component of ρ曰' can be neglected to a lesser extent. ... 6 is used as a periodic function of the film thickness q, and the Pb is converted into a film thickness component of the power spectrum f by the finger product. Therefore, the peak displayed by the spectrum will be “h”_“d+Q”, and since the q value is very small, the influence on the peak position d is small. Step, when performing the Fourier transform, as described below, consider the measurement. The maximum film thickness of the image layer is sampled according to the Nyquist sampling theorem for the wave number conversion reflectance &, with appropriate sampling interval and number of samples. The wave number converted reflectance # sampled based on this method is calculated relative to the calculated The thickness spectrum of the power spectrum is r, and the film thickness of Pb is smaller (Or), so it can be said that the film thickness is hardly affected by the film thickness. In this way, the calculated reflectance spectrum is ' Conversion to a wave number related letter 'It takes into account the wavelength dispersion of the film, and then performs Fourier transform to further calculate the film thickness of the film. In addition, in the above description, the reflectance spectrum is used, but the penetration can also be utilized. Rate spectrum. In this case, the measured penetration rate '卩Γ indicates 'wavenumber conversion transmittance', and the relationship is expressed as follows: Γ:

T

Ta+TbQ〇s2Kd\ Even with the penetration spectrum, the transmittance Γ is also nonlinear for the phase factor cos2A. For this reason, for the above reasons, it is related to the phase factor cos2, and it is a linear wave number conversion transmittance. According to the above formula, the wave number is changed to 19 201007117. The transmissivity T' is the phase factor cos2. In the same manner as described above, the film thickness of the film can be accurately calculated. That is, for the phase factor cos2^ associated with the frequency conversion, the wavenumber conversion transmittance Γ is a function for linearizing the transmittance values of the respective wavelengths. Referring again to Fig. 2, the reflected light generated by the interface between the Si 〇 2 layer 2 and the substrate 矽 layer 3 is considered. When π represents the refractive index of Si layer 1, 4 represents the film thickness, "2 represents the refractive index of Si 〇 2 layer 2, and the sentence represents the film thickness, the wave number conversion reflectance and ' can be expressed as follows: R' = Ra + RfyC〇s2K\di +Rccos2K2d2 + R^ cos2(K\d\ +^2^2)+^ cos2(K\d\ -^2^2) where Κλ _ 2πη\ 尤1=丁

Kr, _ 1πη2 Κ2= Here, the wavenumber conversion reflectance and l(Xl) and 2 (fire 2) converted by the wave number, especially 1 and 2, respectively, are used to separate and calculate the Si layer. 1 φ film thickness 4 and Si 〇 2 layer 2 film thickness. Specifically, it is as follows: R\(Ki) = Ra +Rjj cos2K\d\ +RC cos2K\d2 + Rd cos 2K\ {^d\ +d2^j + Re cos 2K\ R2(K2)- Rq + ^bc〇s2i^2^1 +^c cos + cos2K2^1 +d2^j + Re cos_j 20 201007117 Its order, 4 = ^Ldl «2 d2=^~d2 in these formulas _, though, d tube and ' It is not the correct film thickness, but the wavenumber conversion reflectance is in the power spectrum corresponding to the second term of (6); the original film thickness can be obtained from the peak, and the wave number conversion reflectance

And the power spectrum towel corresponding to the third item of (4), the original film thickness d2 can be obtained from the peak. Furthermore, the refractive indices of 'actually' Si|丨 and the coffee layer 2 are similar, and in most cases, the reflectance at the interface between the two is relatively small compared to the reflectance at other interfaces. As a result, the value contained in the function of the wavenumber conversion reflectance is small. Therefore, in many cases, it is difficult to confirm the peak corresponding to the three values of the wave number conversion reflectance from the power spectrum. In this case, the first item of the wavenumber conversion reflectance anvil (6) is calculated first, and the peak position of the power spectrum is +^, and the second item of the wave number^reflectance &&(&) After the peak position |/ij' of the power spectrum is obtained, the difference between the two is obtained, and the measurement result of the film thickness "with respect to the wavelength range and the wavelength resolution" to measure the SOI substrate can be calculated. Furthermore, in the case of the measurement example of the factory, in the case of the third (a) diagram, the wavelength range of the light is measured: 9 〇〇 nm to 丽 (10)', and in the case of the third (8) diagram, the wavelength range is 1340nm ~ (10) ―. The step-step corresponds to the measurement wavelength, and the diffraction (4) 62' having the appropriate characteristics is selected so that the reflected light is incident on the detection portion ^, and the detection of the portion 6 4 is detected in 21 201007117; γ ·... ," (point) number (number of detected channels) such as (1) channel). In other words, the smaller the wavelength interval of the wavelength range (that is, the wavelength resolution), the smaller the mother detection is based on the aforementioned analytical test ' The wavelength of the reflectance should be periodically changed. In the measurement results shown in Figure 3(4), although the reflectance can be seen to have a periodic variation in wavelength, it is impossible to obtain an accurate film thickness—in this case In the measurement results shown in the figure, it is clear that the peak of the reflectance and the valley (vaUey) are not shown, and the period of change of the reflectance can also be measured. Figure 3(c), The measured result (reflectance frequency 4) shown in Fig. 3(b) is converted to the function of the above-mentioned wavenumber conversion reflectance and is used to display the frequency conversion result particularly related to the wave number. As shown in Fig. 3. 'The system can use the value corresponding to the main peak as the layer thickness. Further' Fig. 4 and Fig. 5 show the results of the s〇I substrate. Fig. 4 is a view showing the film thickness measuring apparatus 1 according to the embodiment of the present invention. ❹ The measurement result of the SOI substrate is measured. In the measurement example of Fig. 4, the film thickness of Si layer 1 is 1〇·〇(10) (design value), and the thickness of layer 2 of layer 2 is 0.3 (design value). Further, in figure 4(a) The measurement shows that the measurement light having a wavelength component of the visible light region (330 nm to n00 nm) is used, and in the fourth (b) diagram, the measurement is performed using a wavelength component having an infrared light region (9 〇〇 nm 160 Onm). Further, as described above, the number of detection points (the number of detection channels) of the detection portion injury 6 4 (Fig. 1) is the same. 22 201007117 As shown in Fig. 4(a), when the wavelength of the visible light region is utilized In the case of measuring light of a component, a periodic variation of the reflectance is exhibited in a wavelength region of more than 86 〇 nm, but in a short visible light region, it is known that a significant periodic change does not occur. As shown in Fig. 4(b), when the measurement light having the wavelength component of the infrared light region is used, it is known that there is a significant inverse Fig. 5 is a view showing another measurement result of S (H substrate) by using the film thickness measuring device ι of the embodiment of the present invention. The film thickness of 'Si! in the measurement example of Fig. 5 8 〇〇 μιη (design value), the film thickness of the scratch layer 2 is 〇._ (design value). Further 'in the 5th (a) figure, the display shows the use of infrared light field (900nm~16〇〇nm) The measurement light of the wavelength component is shown in the figure 5(b), and the measurement light having a wavelength component having a narrower infrared light field (147 〇 nm to 160 〇 nffl) is used. The number of detection points (number of detection channels) of the 64 (Fig. 1) are the same. As shown in Fig. 5(a), even when the measurement light having the wavelength component of the infrared light region is used, the measured reflectance does not show a significant periodic change. As shown in Fig. 5(8), when the measurement light having a narrower infrared wavelength region is used, it is known that a significant change in the periodicity of the reflectance occurs. According to the above measurement examples, in order to measure the film thickness of the thick layer with high accuracy, it is necessary to appropriately set the wavelength range and wavelength resolution of the measurement light. Since this is a method of utilizing the interference of the light in the layer, and the wavelength resolution of the reflected portion of the detecting portion 64 is limited, a more appropriate wavelength of the measuring light can be set according to the method described below. In the following inspection, the lower limit of the wavelength detection of the detecting portion 64 is 23 201007117 ^in, and the detection portion ^ j = the upper limit of the wavelength detection of the object 64. The wavelength range of the measurement light to be irradiated by the light source 1 彳 (the first solid, _ Α diagram) can be any range if the oil of the detection portion 64 is included and the detection range of the heart length is recognized. Further, the number of detection points (the number of detection channels) of the detection portion 4 rπ and (Fig. 1) is 6/7. Fig. 6 is a view showing the measurement range of the thickness and the detection wavelength range of the detecting portion 64 according to the embodiment of the present invention, and the relationship with the number of detection points. ❹ () Qi thick is the relationship between the lower limit of the range and the detection wavelength range. According to the film thickness measuring method described above, since it is necessary to find the wavelength of the light interference generated in the object to be tested as the object, the detecting portion 64 is required to have a wavelength range capable of generating light interference. That is, as shown in Fig. 6(a), in the detection wavelength range of the detecting portion 64, the reflectance waveform to be measured of the object to be tested needs to have a change of more than one cycle. This is because the detection wavelength range of the detecting portion 64 is changed from the lower limit value @ to the upper limit value, which means that the change in the optical distance generated is sufficient to carry out the film thickness of the object to be tested. Therefore, the relationship between the lower limit c/min of the film thickness measurement range and the wavelength range of the measurement light needs to satisfy the following conditional expression (1): >__j:min. and max_ ...(1) 2(Anax ·%ιίη - Ληίη .«max ) where 'wmin denotes the refractive index of the wavelength, and 'max denotes the refractive index of the wavelength & nax. 24 201007117 ) The relationship between the upper limit of the film thickness measurement range and the number of detection points.

The arrangement interval of each array element is larger. The coordinates. At this time, for the wavelength, each array element such as InGaAs is waited for

Interval configuration, such as jl Bu - A, ^ I So, the smaller the S wave number, the φ corresponding to the wave number. Therefore, the reflectance waveform corresponding to the wave number and changing with a predetermined period can be correctly sampled. The arrangement interval (wavelength resolution Δ) of each array element needs to satisfy the Nyquist sampling theorem, and the sampling upper theorem is satisfied to determine the upper limit of the film thickness measurement range. The wavelength resolution ΔΑ of the detection portion 64 can be expressed as ΔΑ by the number of detection points (the number of detected channels) Sp. Since the longer the wavelength of the measurement light is, the shorter the period of the reflectance waveform is. Therefore, in the reflectance waveform, when the extreme value ❶ (Peak or valley) of the upper limit value of the light is measured, The wavelength generated by the extreme value adjacent to the extreme value is expressed as: ^, and the following condition is satisfied between the upper limit value dmax of the film thickness measurement range: 2 (Anax . "1 ~ A · Here, when measuring When the film thickness of the target layer is large, η·3χ and ηι can be assumed, and therefore, the above condition is expressed as the following conditional expression (2):

-λ\) 25 ...(2) 201007117 At this time, the wavelength resolution ΔΑ needs to satisfy the following conditions: ΔΛ = ^111^—&lt; - λ\ SP '~2 Substituting the relationship of the upper limit value dmax into the above In the relational expression of wavelength resolution, after elimination, ll can be expressed as the following conditional expression (3): Αλ = ——m^n &lt; 乂^iax _ SP 2(乂max + 2 . wmax . dmax ) (3) The above is the result of the inspection. If the film thickness measurement range (lower limit i/min to upper limit value) required for the test object is to be determined in advance, in order to satisfy the above conditional expression (1) and (2) It is necessary to determine the wavelength range of the measurement light (lower limit value Ληώ~upper limit value Ληβχ) and the number of detection points 5^. <<Calculation Example>> When the film thickness of the Si layer i is measured in the SOI substrate shown in Fig. 2, the calculation of the necessary conditions will be described below. In this calculation example, the upper limit of the Si layer 1 of the SOI substrate is _ 1 〇〇 μ Π 1 ' and the refractive index is constant (n = 3. 5) without depending on the wavelength. Furthermore, in this calculation example, the lower limit c?min of the Si layer 1 of the S01 substrate is not considered. By substituting the above-mentioned set values into the above conditional expressions (2) and (3), the upper limit value ^ax = 1424. Onm can be calculated, and the wavelength resolution ΔΑ_1· 445375 nm can be calculated. Therefore, when the detection portion 64 having 512 channels is used to measure the film thickness of the object to be tested having a maximum film thickness of 1 ΟΟμπι, if the measurement light used covers a wavelength range of about 684 to 1424 (10), then 26 201007117 The reflected light of the range is detected by the detecting portion 64 (wavelength analysis sound △ = 1.4453125 nm). &amp; wherein the wavelength resolution Δ; ι calculated according to the above conditional expression is used to explain the theoretical minimum specification 'when actually measuring, the phase is preferably at the calculated wavelength resolution Δ乂' Improve its accuracy. Furthermore, it is preferably several times (for example, 2 to 4 times). Furthermore, increasing the accuracy of the household means setting the value of the wavelength resolution ΔΑ to be smaller. X, that is, in the actual film thickness measuring device, the influence of the incident angle of the light to be measured by the object to be measured, and the influence of the opening angle when using the lens collecting system, etc., cause the spectral accuracy to be lowered. In this case, the peak height on the power buccal spectrum becomes small, and it becomes difficult to calculate the film thickness. Further, when the discrete frequency conversion is performed by using the FFT or the like with the sampled value, the influence of the aliasing is also affected, and the conversion error generated during the wave number conversion is also increased. Further, the refractive index dispersion of the test object will also be due to the wavelength range of the measured light, @烈变&amp;, or 彳 can not meet some of the conditions. 0, Figure 7 shows the use of a film having a wavelength resolution close to the theoretical value. A schematic diagram of the simulation results of the thickness measurement device. Fig. 8 shows a schematic diagram of the simulation results of the measurement results without using a film thickness measuring device having a wavelength resolution higher than the theoretical value. Further, the film thickness of the object to be tested as the object is 1 ΟΟ μιη. More specifically, in the 7th (a) diagram, the reflectance spectrum (wavelength resolution ΔΛ=2.734375ηιη) in the range of 9〇〇nm to 16〇〇nm is detected by the detection portion 64 having 512 channels. The result of the measurement, and in the eighth...Fig. 2Ί 201007117, the power spectrum after frequency conversion of the reflectance spectrum shown in Fig. 7(a) (here, FFT conversion) is shown. As shown in Fig. 7(b), in this case, although there is a peak near 1 ΟΟμιη, the level of the noise is smaller than that of the film side (ghost). On the other hand, in the 8th (a) diagram, the measurement result of the wavelength range is determined, and the accuracy of the wavelength resolution of the detection portion 64 becomes twice the theoretical value, and the result is 8th (b). In the figure, the power spectrum after frequency conversion of the reflectance spectrum shown in Fig. 8U) (here, FFT conversion) is shown. In this embodiment, the number of detection points and the wavelength range are determined so that the wavelength resolution ΔΑ of the detecting portion 64 becomes 1 367 875 nm. As shown in Fig. 8(8), in this case, a very strong peak appears in the vicinity of the original film thickness, and the film thickness of the test object can be accurately measured. "Overview of the film thickness difference procedure" ^, according to the periodicity of the reflectance _, the film thickness to be measured can be calculated. In other words, the detected reflectance is changed to obtain the power bump, the early turn ❹, the rich frequency 4, and the peak appearing in the power spectrum can be used to calculate the film thickness. In fact, this power spectrum is a diffusion Fourier transform method, which uses "m equals", "Tian Shi..." and there is also a situation in which the FFT can not be used to take the power spectrum of 77. For this reason, in the present embodiment, in addition to discrete Fourier transform such as FFT, a method of converting surface optimization or a conversion method can be performed. That is, $, as the calculation of the power spectrum, the film thickness of the ❹ ❹ ❹ ❹ , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , Details of the inauguration process, details of the 2010 201007117 "Microcomputer / PC utilization technology for scientific measurement of waveform data processing and measurement system" edited by Minami Shigeo '10th edition' August 1992 Issued on the 1st, CQ Press, is hereby provided as a reference.

Further, in the film thickness measuring apparatus 1 of the present embodiment, in addition to the film thickness calculated analytically from the detected reflectance spectrum, the physical model calculated by the measurement target may be theoretically After calculating the two-spot rate frequency, the reflectance spectrum actually detected is compared, and based on the deviation value between the two, the optical characteristic value of the measurement target is exploratoryly calculated, that is, the fitting method is performed. In addition, as shown in Fig. 2, the film thickness of the material 'f Μ(10)i of the second layer of the Si〇2 layer 2 is two or more digits, and the object is measured by the fiiing method. It is impossible to calculate the film thickness of each layer with the degree of accuracy. _ 9th ® shows the measurement results of the reflectance spectrum associated with the SGI substrate. In the measurement example of Fig. 9, the film thickness of the i-th layer & layer i 1 〇〇 Mm 'the second layer Si 〇 2 layer 2 is 0.48 叩 〇 〇: Γ =: interval change "Fig. 9 Even if the film thickness of the second layer-layer 2 changes, the change in the reflectance spectrum of the measured layer does not cause a large change in the measured reflectance spectrum. In other words, the square is too much to say the reflectance measured from the object to be tested. Frequently, due to the influence of the i-th layer ςι•思, 千鸿”曰, 1 layer of Si layer, it means that even if the parameters of the second layer 2 change, it cannot be adequately fitted. Next, the film thickness 豕 substrate thickness measuring device 100 of the present embodiment performs one of the above-described Fourier transform 29 201007117 conversion, optimization method, and fitting method, or the like. They are appropriately combined to separate the film thicknesses of each layer and can be correctly analyzed. Hereinafter, details of the film thickness calculation program of the film thickness measuring device 丨00 in the present embodiment will be described. Further, the film thickness calculation program is executed by the data processing portion 70 (Fig. 1). The structure of the data processing section shows the outline hardware architecture of the data processing section according to the embodiment of the present invention. Referring to FIG. 10, a data processing portion 70 is generally implemented by a computer, and includes a central processing unit (CPU) 200 for executing various programs including an operating system (OS); The portion 212 is configured to temporarily store necessary data when the CPU 200 executes the program, and a hard disk drive (HDD) 210 for non-volatilely storing the program executed by the CPU 200. Furthermore, in the hard disk portion 210, the program for implementing the program described later is stored in advance. The program is floppy disk drive (FDD) 216.

A or a CD-ROM drive 214 is read from a floppy disk 216a or a compact disk-read only memory (CD-ROM) 214a. The CPU 200 receives an instruction from the user or the like through an input portion 208 composed of a keyboard and a mouse, and outputs the measured measurement result or the like to the display portion 2 by execution of the program. 04. The parts are connected to each other through the bus bar 200. "Calculation program structure" 30 201007117 The data processing part 7 of this embodiment corresponds to the type and quantity of unknown values in the "film thickness" film thickness range, refractive index, decay coefficient, etc. of each layer of the material to be tested. 'And the resolution accuracy, etc. can be selected from the following (4) patterns 1 to 6 to select them. Furthermore, in the following

In the description, as shown in Fig. 2, the SOI substrate is not calculated by the method of independently calculating the film thickness of each of the two layers. However, the same method can be used to calculate the film thickness of each of the layers. . (1) Program pattern i The program 31 is a film thickness calculation program that can be performed when the refractive index and the attenuation coefficient of the i-th layer and the second layer are known. In the procedure ι, the film thickness of each layer is determined by the fitting method. Furthermore, in one embodiment, a least squares method can generally be utilized as the adaptation method. Figure U is a diagram showing the execution of a program in accordance with an embodiment of the present invention! A block diagram of the control structure of the associated film thickness calculation program. In the block diagram of Fig. 5, the CPU 200 reads out the program previously stored in the hard disk portion, etc., to the memory portion 212, and then executes it. Referring to Fig. 11, the data processing section 70 (Fig. 1) includes a buffer Wbuff... portion 71, a model (m〇del) portion 721, and a fitting portion 722. The buffer portion 71 is configured to temporarily store the measured reflectance spectrum RU output from the spectrometry portion 60. More specifically, the 'splitting portion 6' outputs the reflectance according to the wavelength resolution of each of the pupils. Therefore, the buffer portion 71 stores the wavelength and the reflectance of the wavelength correspondingly. The model part 72 receives the parameter related to the object to be tested, and according to the parameter received by 31 201007117, determines the model (function) used to represent the theoretical reflectance of the object to be tested, and then uses the determined function to calculate Theoretical reflectance (spectrum) for each wavelength. The calculated theoretical reflectance of each wavelength is output to the appropriate portion 722. More specifically, the modeled portion 721 receives the refractive index m and the attenuation coefficient k1' of the i-th layer and the refractive index (1) and the attenuation coefficient k2' of the second layer, and simultaneously receives the initial value of the film thickness dl of the first layer and The initial value of the film thickness of the second layer. Furthermore, the parameters may be input by the user, or the parameters of the standard material may be stored in advance as files, and then read if necessary. Furthermore, if necessary, the refractive index n of the atmosphere can also be input. And attenuation system &amp; ^.模型 The model used to display the theoretical reflectivity is the same as the reflectivity of the test object of the above three-layer system, which is a function of at least the film thickness value of each layer. Furthermore, the modeled portion 721 updates the function of displaying the theoretical reflectance according to the parameter update command of the matching portion 722 described below, and then calculates the theoretical reflectance of each wavelength according to the updated function ( Spectrum). More specifically, the (four) portion 721 sequentially updates the film thickness d! of the second layer of the material parameter and the film thickness d2 of the second layer. The adaptive portion 722 reads the reflectance frequency error _ value of the buffer portion 71, and then uses the reflectance spectrum theoretical value outputted by the modeled portion 721 to sequentially calculate the square deviation of each wavelength between the two. value. Next, the fitting portion 722 calculates the residual | from the deviation value of each wavelength, and then determines whether the residual is equal to or lower than a predetermined critical value. &gt; Material, θ phase. In other words, the matching part 7 2 2 determines whether the current parameter is convergence. When the residual is not below the predetermined threshold, the adaptation portion 722 transmits the parameter update command to the modeled (four) 721 and then waits until the new reflection 32 201007117 rate spectrum is output. On the other hand, when Tian Xigegui is below the predetermined critical value, the fitting portion 722 takes the first film thickness dl and the film thickness d2 of the second layer as analytical values and rotates. The ^2 diagram shows a flow chart of a method for the film thickness deviation procedure associated with the program type i in accordance with an embodiment of the present invention. = Test Figure 12 The user places the object to be tested (sample) on the stage (Fig. S100). After that, when the user releases the measurement preparation command, the observation light is started by the household umbrella, the language/## observation source (Fig. 1). The user refers to the reflected image obtained by the observation camera 38 and (4) the display portion 39, and the stage position command is issued to the movable mechanism 5 for adjusting the measurement range and focusing (step sl2). After the adjustment of the measurement range and the focus are completed, the user releases the measurement start I, and the measurement light source is used! . (1K) starts to generate measurement light. The spectroscopic portion receives the reflected light of the object to be tested, and outputs a reflectance spectrum based on the reflected light to the data processing portion 7G (step siG4). Next, the CPU (10) of the processing section 70 temporarily stores the inverse spectrum detected by the spectrometry section 6() in the spectrometry unit ❾60 or the like (step coffee). The data processing section 7〇〇2〇〇 performs the following film thickness calculation procedure. The CPU 20 0 displays the input screen on the display part (Fig.) or the like, and asks the user to input the parameters (step siQ8). The user inputs the refractive index (1) of the first layer of the object to be tested and the coefficient of cancellation k1 according to the input screen displayed, and the refractive index (1) and the attenuation coefficient 第 of the second layer of the object to be tested, and simultaneously enters the first The film thickness of the layer &amp; and the film thickness of the second layer (1) initial value (step 33 201007117 step S110). Further, the CPU 200 calculates a theoretical value of the reflectance spectrum based on the parameters that the user has entered (step S112). Next, for the measured value of the reflectance spectrum and the theoretical value of the reflectance spectrum stored in the memory portion 212 and the like, the CPU 200 sequentially calculates the squared deviation value of each wavelength between the two to calculate the residual between the two. Poor (step S114). Further, cpu 200 determines whether or not the calculated residual is below the predetermined threshold (step SU6). When the calculated residual is not equal to or less than the predetermined threshold (in the case of NO in step S116), the CPU 200 changes the film thickness d1 of the i-th layer and the current value of the film thickness ch of the second reference layer (step S118). The thickness of the film and the film thickness (how much change is made in which direction h depends on the degree of occurrence of the residual. Then, the process returns to step S112 of the program. In contrast, when the calculated residual is in the predetermined When the critical value is equal to or less than the critical value (in the case of YES in step S116), the CPU 2 作为 sets the film thickness d1 of the second layer and the film thickness of the second layer to the film thickness (analytical value) of each layer of the object to be tested. And outputting (step S120). Thereafter, the program is terminated. Further, in the block diagram shown in Fig. 11, although the refractive indices ηι, n2 and the attenuation coefficients k!, k2 are input at a fixed value, they may be considered. The refractive index and the decay coefficient of the wavelength dispersion. For example, the following Cauchy model can be used as the refractive index and the attenuation coefficient considering the wavelength dispersion: "(1) = containing + containing 'a, 6, c, i/, e, / represents the correlation coefficient in each layer. 34 201007117 When using this formula, 'In the long-term &amp; + Λ 各 各 各 亦可 亦可 亦可 事先 事先 事先 事先 事先 事先 事先 & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & λ), + λ2 -h where /, g, " Sellmeier coefficient, and (2) program type 2 program pattern 2, when the refractive index and the coefficient of nuisance of the first layer and the second layer are known The film thickness calculation program can be executed. In the program pattern 2, the film thickness of each layer is determined by a suitable method. Frequency conversion by discrete Fourier transform is used to obtain a U-thickness having a large film thickness, and the first layer is The film thickness is a fixed value', and the film thickness of the 帛2 layer is determined by a suitable method. Further, in the embodiment, the least squares method can generally be used as the fitting method. ❿ Figure 13 shows the implementation according to the present invention. For example, a block diagram of a control structure for executing a film thickness calculation program related to the program pattern 2. In the block diagram shown in the figure n, the coffee 2QG reads out a program such as a hard disk portion 2iq to The memory portion 212 is then executed. [beta] Referring to Figure 13, the data processing portion 70 (Fig. 1) includes The buffer portion 71, the wave number conversion portion 731, the buffer portion 732, the rich conversion portion 733, the peak search portion 734, the modeled portion 735, and the fitting portion 736. % buffer portion 7 It is used to temporarily store the measured reflectance frequency I# Κ(λ) outputted by the spectrometry part 6。. Furthermore, the specific description is as above, and will not be described here. 谷已详 35 201007117 Wave number conversion part 731 receiving The parameters of the first layer (refractive index ηι and the attenuation coefficient b) are subjected to wave number conversion of the reflectance spectrum R(A) temporarily stored in the buffer portion 根据 according to the received parameters. In other words, the wave number conversion portion 731 converts the correspondence between each wavelength in the reflectance spectrum R(A) and the reflectance of each wavelength into a wave number Κι (λ) associated with each wavelength and uses the above relationship Calculate the correspondence between the wavenumber conversion reflectance ^. More specifically for each wavelength stored in the buffer portion 71, the wave number conversion portion 731_ sequentially calculates the wave number Κι(λ) and the wave number conversion reflectance and (male) / (ι-diffic)) 'The output is then to the buffer portion 732.邛, 732, the wave number conversion part 731 sequentially rotates 3 wave number KK A ) and the wave number conversion reflectance 岣 ") corresponding storage, that is, the wave number Κ 1 ( λ), the wave number The wavenumber distribution characteristic of the converted reflectance, \ wavenumber conversion reflectivity (6), is stored in the buffer portion, for example. The Fourier transform portion 733' uses the converted reflectance of the buffer portion m (phantom) , , 隹 &gt; Also on 戚) Kl (four) Fourier transform, use 乂 to calculate the power spectrum p!. Then go (FFT, ^ is enough to use fast Fourier transform DCT) (dlsc-te cosine transfer, ), etc. as a Fourier transform The peak search portion 734 searches for the thickness calculated in the power frequency u, and the thickness calculated by the second conversion portion 733 is taken as the Uth and the film corresponding to the peak is obtained as the first layer. The film thickness is outputted. The modeled part 735 receives the parameter with the object to be tested, and the number of the fixed number is based on the model 4 of the theoretical reflectivity of the connected object, and then: the function determined by the method, = modulo 1 Equation (the theoretical reflectivity of each wavelength is 201007117. , the appropriate part of 736. More ^ j卬 to match _ rounded out the i-th layer = share: receiving from the peak detection * attenuation coefficient L, while =: and the second layer of the refractive index ... receive the second in time Film thickness of the layer & initial value.

'The parameters can be entered by the user, or the sex can be stored in advance by the standard, etc., and read if necessary. The model used to show the theory is the same as the above-mentioned three-layer system (4), and is a function containing at least the film thickness values of the respective layers. In addition, the modeling section 735 updates the function of the display theoretical reflectivity according to the parameters of the matching part 736, and then: according to the updated function, Theoretical reflectance (spectrum) for each wavelength. More specifically, the modeled portion 735 sequentially updates the film thickness d2 as a parameter. After the matching portion 736 reads the reflectance spectrum measured value ' of the buffer portion 71, the reflectance spectrum theory_value outputted by the modeled portion 735 is used to sequentially calculate the squared deviation value of each wavelength between the two. . Next, the fitting portion 736 calculates ❹ from the deviation value of each wavelength, and then determines whether or not the residual is below a predetermined critical value. That is, the matching portion 736 determines whether the current parameter is convergent. When the residual is not below the predetermined threshold, the adaptive portion 73 6 transmits the parameter update command to the modeled portion 735 and then waits until the new reflectance spectrum is output. On the other hand, when the residual is below a predetermined critical value, the fitting portion 736 outputs the film thickness di of the first layer and the film thickness d2 of the second layer as analytical values. 37 201007117 Figure 14 is a flow chart showing the method of calculating the film thickness associated with the program type 2 according to the embodiment of the present invention. In the steps of the flowchart shown in Fig. 14, the procedures of steps S100 to S108 and the flowcharts shown in Fig. 12 are denoted by the same reference numerals, and will not be described again. Hereinafter, the film thickness calculation program after step S132 will be described with respect to the difference from the flowchart shown in Fig. 12. In step S132, the user inputs the refractive index ηι and the attenuation coefficient k of the first layer of the object to be tested, and the refractive index n2 and the attenuation coefficient k2 of the second layer of the object to be tested according to the displayed wheeling screen. At the same time, the input layer 2 thickness of the media thickness d2 initial value.

Then, the CPU 200 performs wave number conversion on the reflectance spectrum stored in the memory portion 212 or the like based on the input refractive index (1) and the nuisance coefficient k of the i-th layer (step S134). Next, the wave number converted reflectance obtained after the wave number conversion is stored in the memory portion 212 or the like (step MM). Further, the CPU 2GG performs a Fourier transform on the wave number conversion reflectance in association with the wave number κ to calculate a power spectrum (step (1) 8). Further, the CPU 200 obtains the material appearing in the power spectrum towel and the film thickness corresponding to the material, and outputs it as the film thickness (1) of the first layer (step. Next, the film obtained by the CPU 200 according to step S21) Calculate the theoretical value of the reflectance spectrum by the thickness and the second layer parameter input by the user (step S142). Next, for the measured value of the reflectance spectrum and the theoretical value of the reflectance spectrum stored in the memory portion 212 and the like, The CPU 200 sequentially calculates the squared deviation values of the respective wavelengths between the two to calculate the residual between the two (step S144). Further, the cpu determines that the calculated residual is 38 201007117 or not below the predetermined threshold. (Step S146) When the calculated residual is not below the predetermined threshold (in the case of step S146 &quot;〇), cpu_ changes the current value of the thickness of the second layer &amp; (step S148). The extent to which the film thickness is to be changed depends on the degree of occurrence of the residual. Then, it returns to step S142. Relatively, when the calculated residual is below the predetermined threshold ( YES in step SU6), cpU2现在 The current value of the film thickness 第 of the i-th layer and the film thickness d2 of the second layer is taken as the film thickness (analytical value) of each layer of the object to be tested (step S150). Thereafter, the program is terminated. The program type! The same, it is also possible to use the refractive index and the attenuation coefficient considering the wavelength dispersion. The detailed function has been explained above, and will not be described here. (3) Program type 3 fading #^(4) 3 ' is 6 The film thickness calculation program that can be performed when the refractive index and the elimination coefficient of the first layer and the second layer are known. In the program pattern 3, compared with the above-mentioned program pattern 2, the difference is in the calculation. The first layer: when thick, 'does not perform Fourier transform, but uses the optimization method. The other program' is the same as the above-mentioned program type #2. The sample diagram is displayed according to the embodiment of the present invention, Execution and program type, the control structure block diagram of the film thickness calculation program. In Fig. 15, the CPU 200 stores the process in front of the hard disk portion 21°, etc. 2, and then perform. Refer to Figure 15, data processing part 7〇 (1st picture) including buffer 39 2010071 17 part η, optimization calculation part 74 ι, model part 743. 丨 imitation / 42 and fitting buffer part 71 ' used to 堑 left eight, measured reflectance spectrum Κ (λ) ” 'see疋 Some of the 60 rounds are as described above, and there is no stipulation. In addition, the specific structure and processing contents have been optimized for the calculation part 74. The reflection of the storage 'resolve the film thickness d of the layers” is more specifically optimized, and the self-regressive model is used to obtain the spectrum relative to the reflectance: J uses the regression (4) The self of the regression model &lt; coefficient of return. The optimization calculation section 741 is used to obtain the film thickness corresponding to the wavelength of the principal component, and to analyze the frequency, and to output it as the film thickness Ch of the first layer. Furthermore, in the implementation of the best film part 741 contact! s 10,000 law's best calculus Μ 741 receives the first! The search range, the gradation and the decay coefficient kl of the film thickness of the layer, and the refractive index η2 and the decay coefficient k2 of the second ΠΒ± 五 地 、 、 、 = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = Furthermore, the parameter of the standard material f may be stored in advance by a slot or the like, and may be read out if necessary. Parts: = part and the appropriate part 743, the film thickness 1 of the second layer transferred from the optimization optimization calculation unit and the object to be tested are determined by the fit to determine the film thickness d2 of the second layer. The procedures of Part 743 are the same as those of the above-mentioned method and the matching part 735 and the fitting part 736, which are not described here. The figure is a flowchart showing the method of calculating the film thickness calculation procedure according to the embodiment of the present invention. The method of the method for calculating the film thickness is shown in the steps of S100 to S106. The same steps are denoted by the same reference numerals, and the film thickness = the difference in the flow chart shown in the drawing, that is, the film thickness calculation program after the step (10) will be described. In step S162, the user inputs: the search range of the film thickness of the i-th layer of the object, the folding parameter 2(1) of the i-th layer of the object, and the decay coefficient kl, and the The refractive index (1) of the second layer of the object and the attenuation coefficient k2. The method parsing buffer portion ' is used to calculate the first layer. Then, the CPU 200 uses the frequency component film thickness di of the reflectance spectrum stored by the optimized side 212 (step S164). The CPU 200 calculates the theoretical value of the reflectance spectrum from the wave (step (10)). Next, the measured value of the reflectance spectrum stored in the memory portion 212 and the reflectance spectrum and the theoretical value of the reflectance spectrum 'cpu 2〇〇 are sequentially calculated to calculate the squared deviation values of the wavelengths between the two Luders to calculate the two. The residual between them (step S1 68). Further, the CPU 200 determines whether or not the calculated residual is below a predetermined threshold (step S1 70). When the calculated residual is not below the predetermined threshold (in the case of NO in step S170), the CPU 2 changes the present value of the film thickness of the second layer (step S172). Furthermore, the extent to which the film thickness &amp; in which direction is to be changed depends on the degree of occurrence of the residual. Thereafter, the process returns to step S166 of the program. On the other hand, when the calculated residual value is equal to or less than the predetermined threshold value (YES in step 41 201007117 to step S170), the CPU 200 sets the current value of the film thickness d1 of the first layer and the film thickness d2 of the second layer as The film thickness (analytical value) of each layer of the test object is output (step S174). After that, the program ends. Further, similarly to the above-described program pattern 1, it is also possible to use a refractive index and a decay coefficient which take into consideration the dispersion of the wavelength. The detailed function has been explained above and will not be described here. (4) Program type 4

The program pattern 4 is a method for improving the program pattern i to converge more reliably according to the adaptation. In other words, as for the s〇I substrate, for the test object having a large difference in film thickness between the second layer and the second layer, the initial value is important in order to match the film thickness of each layer. Here, the program pattern 4 first determines the initial thickness of each layer by the method of optimization, and (4) determines the film thicknesses of the first layer and the second layer by preparing these initial values and the appropriate values.

Fig. 17 is a block diagram showing a control structure for executing a film thickness calculation program relating to the program type 4 according to an embodiment of the present invention. In the block diagram shown in Fig. 17, the CPU 200 reads out the program previously stored in the hard disk portion (10) or the like to the memory portion 212, and then executes it. The control structure is substantially the same as the control structure of the u-th figure except that the program pattern shown in Fig. 17 is added in addition to the program pattern shown in Fig. 17. The optimization calculation part 751 extracts the frequency components of the reflection Α 1 ^ 1 a ^ ^ ® and the shards stored in the buffer portion 71 by means of the optimization method such as the resolution MUM. The film thickness 第 of the first layer and the film thickness d2 of the second layer and the 嘈 layer were calculated. In particular, the optimal calculation portion 751 analyzes the frequency of the measured rate spectrum to obtain two or more peaks obtained by the 42 201007117, and then + thick, to calculate the film thickness of the first layer 6 and 1 respectively. The film thickness corresponding to the film corresponding to the two peaks is calculated as the film thickness 第 of the first layer and the film thickness d2 of the first layer. The film thickness d2, # of the layer is used as the initial value of the fit, so it is not necessary to strictly calculate the specific frequency resolution/determination of the part 751. Furthermore, the best calculation part 741 is the same. The above-mentioned optimization model part 7 21 and the fitting part 俭79 (5) calculated the film thickness = the optimization calculation unit as the initial value, using the fit Decide that the original 2nd thickness of the mold layer I will be the thickness of the two layers. The model part 721 and the layer / layer of the thickness of the "the volume has been % Ming &amp; μ and the fitting part 722 within the program This has been explained above and will not be described here. Figure 18 shows the calculation of the % film thickness according to the present invention. The program pattern of the embodiment is related to the == process chart. In the flowchart shown in Fig. 18, the procedure of steps S111A and SniB is applied to suppress Jiezhe =: rsu °. Regarding other programs, the same symbols are used to indicate ::? steps, which are not described herein. The following description differs from the ... program. Referring to Figure 18, the step ςι no i ^ after the shoe library 8 is executed, the step S1UA is executed =: _ _ the user inputs the refraction of the first layer of the object to be tested according to the displayed input Rate - the refractive index n2 of the 2 layers and the coefficient of decay 2 = kl, and the search range of the 1st object to be tested and 2 μ of the 2nd layer: the film thickness of the 1st layer in step S111B, cpu, thick (4) range. The frequency component of the buffer rate spectrum is analyzed by the method of optimizing the anti-utilization of the next part 212 to calculate the film thickness ch of the 1st 43 201007117 layer and the film thickness d of the second layer. The film thickness d! of the t-th layer and the film thickness d2 of the second layer calculated by Sllu are used as initial values. Then, after step S111B, the same procedure is executed after step SU2 of FIG. Further, similarly to the above-described program pattern i, a refractive index in consideration of the wavelength of the blade and a > Xiaoxiao coefficient can be used. The detailed function has been described above, and will not be described herein. (5) Program pattern 5

The program pattern 5 is a modified example in which the method of applying the thickness of one layer to the case where only the thickness of the other layer is analyzed. In the following description, the media thickness of the second layer of the object to be tested is known, and the film thickness of the first layer is determined by the fitting method. Fig. 19 is a block diagram showing the control structure for executing the medium thickness calculation program relating to the program type 5 according to the embodiment of the present invention. Yutu

In the block diagram, the CPU (10) reads out the program previously stored in the hard disk portion (10) or the like to the memory portion 212, and then executes it. The control structure related to the program type 5 shown in Fig. 19 is a configuration model portion 72U for the program type shown in Fig. u! In the related control structure, the modeled portion 721 is replaced. The modeled portion 721A receives the refractive index ηι and the fading name k" of the first layer and the refractive index n2 and the fading coefficient 第 of the second layer, and at the same time, the film thickness d of the first layer, the initial value and the second layer The film thickness d2 is known (the known value is further 'can also be input by the user, or the parameters can be saved in advance by broadcasting, etc., if necessary, read again. Again, = 44 201007117 can also be input The refractive index n of the atmosphere and the decay coefficient k〇. In addition, the 'modeled part 721A is updated according to the parameters of the matching part 722. 'For the orderly update of the first drawer latitude calendar』 i then according to the update After the d&quot; is updated to show the function of the theoretical reflectivity. Further: the inverse H-wide 72U calculates the rationality _ reflectivity (spectrum) of each wavelength according to the updated function. According to this method, the ith layer It is determined by the law. The structure of the other mountains is described in detail above, and is not described here. Figure 20 is a flowchart showing the method of calculating the program according to the embodiment of the present invention. In the flow chart = the steps S11〇A, S118...(4), 12 Su 〇 s s s s 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于 关于In the meantime, the user inputs the refractive index ηι of the first layer of the object to be tested and the refractive index extension k of the second layer of the object to be tested and the refractive index n2 and the attenuation coefficient k2 of the second layer of the object to be tested, according to the displayed surface. '1 layer of heat I enters the initial value of the film thickness i of the first layer and the film thickness d2 of the second layer. In step S118A, the CPU 200 changes the film spread value of the first layer. In the case of the program type 5, only the film thickness of the second layer is suitable for the object. 1 is a match, in step S120A, when the calculated residual is at a predetermined threshold, the first will be The current value of the film thickness i of the layer is output as the medium thickness (analytical value) to be output. 45 201007117 Furthermore, similarly to the above-mentioned program type, the refractive index and the attenuation coefficient in consideration of wavelength dispersion can also be used. The function has been explained above and will not be described here. (6) Program pattern 6 Program pattern 6 ' is known as the film thickness of the layer, only the other The method applicable to the case of the film thickness is a modification of the above-mentioned program pattern 5. In the following description, the film thickness of the second layer of the object to be tested is known, and the film thickness of the first layer is determined by the matching method. Or Fourier transform decision. Fig. 21 is a block diagram showing a control structure for executing a film thickness calculation program related to the program type 6 according to an embodiment of the present invention. In the block diagram shown in Fig. 21, the CPU 200 The program stored in the hard disk portion 21, etc. is read out to the memory portion 212, and then executed. The control structure related to the program pattern 4 shown in Fig. 21 is configured with the matching portion 722A. In the control structure related to the program type 4 shown in FIG. 19, instead of the fitting portion 722, the wave number converting portion 73 is further increased by the buffer portion 732, the Fourier transform portion 733, and The crest explores part 734. m In other words, in this program type, the film thickness 1 of the 帛i layer of the object to be tested is determined by the fitting method, but when the fit cannot converge within a predetermined number of times, the Fourier transform is used to determine the first The film thickness of the layer is di. The adaptive portion 722A reads the measured value of the reflectance spectrum of the buffer portion 71, and then transmits the parameter update command to the modeled portion 72U, so that the measured value and the reflectance spectrum theory of the modeled portion 721A are rotated. The residual between the values is below the established threshold. Further, even if the matching portion 46 201007117 722A performs the calculation of the number of calls, if the residual cannot be below the predetermined threshold, the switching instruction is transmitted to the wave number conversion portion 731, and the Fourier transform is used to determine the second layer. The film thickness d]. Furthermore, with respect to the wave number conversion portion 731, the buffer portion 732, the Fourier transform portion 733, and the peak search portion 734, the program pattern 2 shown in Fig. 5 is described above, and the details are not described herein. . Fig. 22 is a flow chart showing the method of calculating the film thickness associated with the program type 6 according to the embodiment of the present invention. In the flowchart shown in Fig. 22, the step sm of the flowchart shown in Fig. 20 is added, and the steps S134 to sl4Qe of the flowchart shown in Fig. _ are added, and the other steps are denoted by the same symbols. This does not add capital. The following description differs from the first and second programs. Referring to Fig. 22, in step S117, cpu 2 determines whether or not the matching program has been repeated a predetermined number of times or more. When the fitting program is not repeated for a predetermined number of times or more (in the case of N0 in step S117), after the step su8 is executed, the program returns to step S112. On the other hand, if the appropriate program has been repeated for a predetermined number of times or more (in the case of YES in step sm), the program proceeds to step S134. In steps S134 to S140, the film thickness I of the second layer is determined by Fourier transform. The procedure for these steps has been described above and will not be described herein. <<Measurement Example>> Fig. 23 shows the measurement results of the film thickness of the substrate measured by the film thickness measuring device of the embodiment of the present invention. Furthermore, in Fig. 2, the power spectrum obtained after frequency conversion (FFT conversion) of the reflectance spectrum is shown in 47 201007117. Fig. 23(a) shows the measurement results of the SOI substrate in which the film thickness of the first Si layer is 22. 〇um, and the film thickness of the second Si 〇 2 layer is 3. Ομιη. In Fig. 23(a), the frequency conversion is performed using a component of the measured reflectance spectrum of 147 〇 nm to 16 〇〇 nm. As a result, the largest wave bee is generated at the position corresponding to 21. 8613 μηη. Fig. 23(b) shows the measurement results of the s〇i substrate in which the film thickness of the first Si layer is 32. Ομη, and the film thickness of the second Si 〇 2 layer is 2 〇. In Fig. 23(b), frequency conversion is performed using a composition of 15 〇〇 nm to 16 〇〇 nm in the measured reflectance spectrum. As a result, a maximum peak is generated at a position corresponding to 30. 6269um. Fig. 23(c) shows the measurement results of the s〇i substrate in which the film thickness of the first Si layer is 16. 〇um and the thickness of the second Si 〇2 layer is 13|jm. In Fig. 23(c), frequency conversion is performed using a composition of 14 〇〇 nm to 16 〇〇 nm in the measured reflectance spectrum. As a result, the maximum peak is generated at the position corresponding to 15.9069|Jin. From this, it can be seen that the above test results are generally good. <<Presence of Masking Material' As described above, the film thickness measuring apparatus 100 of the present embodiment mainly measures the film thickness of the object to be tested 0Bj based on the reflectance spectrum of the infrared light field, and therefore 'from the measuring light source 1 〇 (the first 1)) The path to the object to be tested 0BJ can be measured even if a masking material such as a polymer resin exists. In other words, although the light in the visible light region cannot penetrate the material of the polymer resin, the light in the infrared light field can penetrate.

Fig. 24 is a view showing the measurement of the object to be tested 〇Bj on which the opaque pad is placed by using the film thickness measuring device 1 〇 Q of the embodiment of the present invention.

Referring to Fig. 24, 'the planar object to be tested 〇BJ is placed on the stage 50 via a spacer' and the planar opaque pad 52 is placed on the object to be tested 〇BJ (measurement side of the light) ). The opaque pad 52 is an abrasive body used for polishing treatment and is mainly formed of a polymer resin. The opaque pad 52, which has a small amount of penetration, is capable of penetrating light in the infrared field (for example, '900 to I600 nm). Fig. 25 and Fig. 26 are views showing measurement results of the s〇i substrate on which the opaque pad 52 is placed by the film thickness measuring apparatus 100 of the embodiment of the present invention. Fig. 25 shows the result of using a magnifying lens having a magnification of 1 来 as the objective lens 40 (Figs. 1 and 24), and Fig. 26 showing the use of a magnifying lens having a magnification of 2, 83 times. The result of the objective lens 40 (Figs. 1 and 24). In addition, in Fig. 25 and Fig. 26, for comparison, the results in the state where the opaque pad 5 2 is not disposed are also shown. Furthermore, it should be noted that the range (absolute value) of the respective reflectance spectra is not the same. μ Fig. 27 shows the power spectrum obtained by transmitting the reflectance spectrum of the pad 52 shown in Figs. 25 and 26 without being disposed. Fig. 28 is a view showing the power spectrum obtained by transmitting the reflectance spectrum in the state in which the pad μ shown in Fig. 25 and Fig. 26 is arranged. Referring to Fig. 25, in the case where the magnifying lens having a 1G magnification is used as the 49 201007117 objective lens 40, as a result of the presence of the opaque pad 52, the noise component is compared with the result of the absence of the opaque pad 52. increase. On the other hand, referring to Fig. 26, in the case where the magnifying lens having a magnification of 2.83 times is used as the objective lens 4, the result of the presence of the opaque pad 52 is substantially the same as that when the opaque pad 52 is not present. , can also fully determine its periodicity. As shown in Figs. 27 and 28, in the case of using the magnifying lens having a magnification of 2.83 times as the objective lens 40, the same power frequency|| is substantially obtained regardless of whether or not the opaque solder 塾' is present. In the case where the magnifying lens having a magnification of 10 is used as the objective lens 4, it can be seen that a power spectrum having sufficient accuracy cannot be obtained. This is because, as the objective lens 40 changes the magnification, the number of openings also changes, and in the case of using the magnifying lens 5 having a magnification of 10, the diffused light is increased and the noise component is increased. As described above, the film thickness of the object to be tested 0BJ in which the opaque pad 52 is disposed can be measured by the film thickness measuring device 1 of the present embodiment. In the bean, the optical system for measuring the light and the optical system for receiving the reflected light need to have a design that can eliminate the influence of the diffused light. <<Modifications>> A Y-type optical fiber can also be used as an optical system for measuring the light to be measured and the reception of the reflected light. Fig. 29 is a view showing the configuration of an optical system of a film thickness measuring device 100# according to another embodiment of the present invention. Referring to Fig. 29, the film thickness measuring device 100# is an optical system, 50 201007117. The measuring light source of the measuring light source 10 (Fig. 1) is directed to the object to be tested 0BJ, and the reflected light of the object to be tested OBJ is guided to the detecting portion. 64 (figure i) and having an optical fiber 56. The light-receiving optical fiber 56 is a Y-type optical fiber, which combines two light rays into a single light, and simultaneously separates a single light into two light rays. More specifically, in one embodiment, the light-receiving optical fiber 56 is formed of a germanium (Ge) doped single-wire Y-type fiber.测定 The measurement light generated by the measurement light source 10 (Fig. 1) is incident on the object to be tested 〇BJ through the first branch fiber 56a, and the reflected light generated by the object 〇BJ is reflected, passes through the second branch fiber 56b. It is directed to the detection portion 64. In addition to this, between the light-receiving optical fiber 56 and the object to be tested 〇BJ, a pinhole optical system 54 as an aperture is disposed. According to the film thickness measuring device 丨〇〇# shown in Fig. 29, the film thickness can be measured even if the object OBJ is placed in a solution such as a polishing liquid. Fig. 30 is a view showing the film thickness of the test object 〇BJ in the solution by using the film thickness measuring device 100# of another embodiment of the present invention. Referring to Fig. 30, the table 57 is placed in a container, and then the object to be tested OBJ is placed on a table 57 via a spacer which is filled with a solution 58 of a polishing liquid or the like. Then, a portion of the light-receiving side of the light-receiving optical fiber 56 is immersed in the solution 58. With this structure, the film thickness of the analyte OBJ in the solution can be determined. Further, when the solution 58 is made of water as a solvent, when the film thickness is measured in the above-mentioned infrared light field (900 to 1 600 nm), it is preferable to use the wavelength of the wavelength of the water that has been removed 51 201007117:: long. Specifically, the water absorbs a wavelength of about 132 nm or more, and for the film thickness of the object to be tested, the spectrum of the reflected light is preferably in the range of ~132 nm. "Other Embodiments" = Ming program, used as a computer operating system (〇s) = is provided to the program module, and can also execute the relevant program after the necessary mode: : type and timing call. In this case, the above modules are executed in cooperation with the operating system. The program of this module can also be included in the program of the present invention. Further, the program of the present invention can also be programmed into: Ben:: Situation: The modules included in the above other programs are not included in the process - other programs/devices and their programs cooperate to execute related programs. The program incorporated into this program can also be included in the program of the present invention. The supplied program is installed and executed. In addition, the program saves part of the memory. The program port □ includes the program itself and the memory program. It can also be used to implement some or all of the functions of the program according to the present invention. According to the embodiment of the present invention, the spectrum of the rate of incidence (or the spectrum of the transmittance) after the measurement light is irradiated on the object to be tested is formed into a layer of the object to be tested, and the water layer is unfolded (1), etc. Discrete Fourier transform, two 1 _ and other optimization methods, to calculate the main wave number into: PCT and then determine the film thickness, (2) using the model of the fit to determine the method 'can be selectively executed . In this way, even if the number of layers constituting the object to be tested is 10,071,071, or the film thickness of each layer is large, the film thickness of each layer can be more accurately measured. In addition, according to the embodiment of the present invention, in the object to be measured as the object to be measured, the wavelength range (or the wavelength detection range) of the measurement light can be appropriately set corresponding to the thickness of each layer constituting the object to be tested. The wavelength resolution of the portion is detected, so that the film thickness of each layer can be more accurately measured. The present invention has been described in detail above, but the above description is only an example, and the present invention is not limited thereto, and the scope of the present invention is defined by the scope of the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a schematic block diagram showing a film thickness measuring apparatus according to an embodiment of the present invention. Fig. 2 is a cross-sectional view showing the object to be tested which is not a measurement target of the film thickness measuring device of the embodiment of the present invention. Figs. 3(a) to 3(c) are views showing the measurement results of the SOI substrate measured by the film thickness measuring apparatus according to the embodiment of the present invention. Figs. 4(a) and 4(b) are views showing another measurement result after measuring the SOI substrate by using the film thickness measuring apparatus of the embodiment of the present invention. Fig. 5 (a) and (b) are views showing another measurement result after measuring the SOI substrate by using the film thickness measuring apparatus of the embodiment of the present invention. Fig. 6(a) to (c) are diagrams for explaining the film thickness measurement range and the detection wavelength range of the detection portion and the relationship of the number of dots measured according to the embodiment of the present invention. Fig. 7(a) and (b) are diagrams showing the results of simulation of the measurement results using a film thickness measuring device having a wavelength resolution close to the theoretical value of 53 201007117. Fig. 8(a) and Fig. 8(b) are diagrams showing the results of simulation of the measurement results using a film thickness measuring device having a wavelength resolution and a degree of completeness higher than the theoretical value. Fig. 9 is a view showing the measurement results of the reflectance spectrum associated with the SOI substrate. Figure 10 is a schematic diagram showing an outline of a data processing portion of an embodiment of the present invention.

Figure 11 is a block diagram showing the control structure for executing the film thickness calculation program associated with the program sample 1 according to an embodiment of the present invention. The f12 diagram shows a flow chart of a method for the program type i correlation film thickness deviation procedure according to an embodiment of the present invention. Rear view: Fig. 3 is a block diagram showing a control structure for executing a film thickness calculation program relating to the circumstance 2 according to an embodiment of the present invention. The film thickness shows a flow chart of the method for calculating the program according to the program type 2 phase according to the embodiment of the present invention. ❹ μ m Read the block diagram of the control structure of the film thickness calculation program of ^^3 S according to the embodiment of the present invention. Figure 16 is a flow chart showing the method of calculating the film thickness in accordance with an embodiment of the present invention. Material Pattern 3, FIG. 17 shows a method for controlling the film thickness calculation program according to the embodiment of the present invention, according to an embodiment of the present invention, a control structure of the film thickness calculation program according to the sample 4, and a program and an 18th figure. Flow Pattern 4 Correlation 3 54 201007117 Fig. 19 is a block diagram showing the control structure of the film thickness calculation program according to the embodiment of the present invention. Lamp/, type Flag thickness (4) shows (4) Flow chart of the method for calculating the film thickness associated with the material type 5 of the application at the time of this issue. Fig. 21 is a block diagram showing the control structure of the film thickness calculation program relating to the sample 6 according to the embodiment of the present invention. The lamp, film thickness is a flow chart showing the method of the film thickness exclusion procedure according to the procedural type 6 of the embodiment of the present invention. Fig. 23 (a) to (c) show the measurement results of the film thickness of the SOI substrate measured by the apparatus of the embodiment of the present invention. , 'The measurement shows a schematic diagram of the object to be tested using the film thickness measuring device of the embodiment of the present invention to etch the opaque pad. The measurement result of the S01 substrate on which the opaque solder enamel is disposed is not measured by the film thickness measuring apparatus of the embodiment of the present invention. Fig. 26 is a graph showing the measurement results of the S〇1 substrate on which the opaque pad coat 咕n is placed by the film thickness measurement of the present invention. Figure 27 shows the power spectrum obtained by the reflectance frequency level 第 in the state where the 焊 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图 图Figure 28 shows the power spectrum obtained by the 反射 图 图 , , , , , , , 透过 透过 透过 透过 透过 透过 透过 透过 透过 透过 透过 透过 透过 透过 透过 透过 透过 透过 透过 透过 透过 透过 功率 功率 功率 功率 功率 功率 功率 功率 功率 功率 功率Figure 29 is a view showing the structure of an optical system according to the present invention. Film Thickness Measuring Apparatus of the Example Fig. 30 is a view showing the measurement of the film thickness of the test object in the solution by the film thickness measuring penetration of the present invention.坪~疋装55 201007117 [Description of main components] 100~ film thickness measuring device; 10~ measuring light source; 14, 66~ directional mirror; 16, 36~ imaging lens; 20, 30~ beam splitter; 24, 56 , 56a , 56b ~ ' 26 ~ injection part ; 32 ~ pinhole mirror ; 34 ~ axis conversion mirror ; 39 ~ display part ; 5 0 ~ stage ; 60 ~ spectrometry part ; 5 4 ~ needle Hole optical system; 5 8 solution; 6 4~ detection part; 70~ data processing part; 204~ display part; 208~ input part; 216~ floppy disk drive; 214~ CD drive device; 210~ hard disk Part; 731~wavenumber conversion section; 71, 732~buffer section 12~focusing lens; 18~ aperture; 22~ observation light source; fiber; 26a~mask part; 3 2 a ~ pinhole; 38~ observation camera; 4 0 ~ objective lens; 51~ movable mechanism; 52~烊 pad; 57~ table; 6 2 ~ diffracted light thumb; 68~ shutter; 206~ interface part; 200~CPU; 216a~soft Disc; 214 a ~ disc; 212 ~ memory part; 733 ~ Fourier transform part;

56 201007117 734~Crest exploration part; 721, 721A, 735, 742~ modeled part; 722, 722A, 736, 749, 751 ~ fitting part; 741~optimized calculation part; AX1 ' AX2 AX3, AX4~ optical axis; and OBJ~ test object.

57

Claims (1)

201007117 VII. Patent application scope: 1. A film thickness measuring device, comprising: a light source, the measuring light having a predetermined wavelength range is irradiated onto the object to be tested formed on the substrate, the object to be tested includes the light source a first layer adjacent to the first layer and a second layer adjacent to the first layer; for measuring the light, the light reflected by the object to be tested or the light penetrating the object to be tested is used to obtain reflectance or wear The wavelength distribution characteristic of the transmittance; the first determining unit performs the adaptation of the wavelength distribution characteristic by using a model formula to determine at least the film thickness of the first layer, the model having the inclusion of the object to be detected The film thickness of each layer; the conversion unit converts the correspondence between the wavelengths of the wavelength distribution characteristics and the reflectance or the transmittance of the wavelength into the wave number of each wavelength and the conversion calculated according to the predetermined relationship Corresponding relationship between values for generating a wave number distribution characteristic; an analyzing unit for obtaining an amplitude value of each wavenumber component included in the wave number distribution characteristic; and a second determining unit according to the wave number distribution Large numbers of wave component amplitude value Θ is comprised of, at least for determining the thickness of the first layers, wherein the first selectively allow determining means and said second decision means is activated. The film thickness measuring device according to claim 1, comprising a third determining unit, wherein the value of the film thickness of the first layer determined by the second determining unit is used to set the model And the wavelength distribution characteristic is used to determine the film thickness of the first layer, wherein the model has a film thickness of each layer included in the object to be tested. 3. The film thickness measuring device according to claim 1, wherein the second determining unit is enabled when the first determining unit is not converge within a predetermined number of times. 4. The film thickness measuring device according to any one of claims 1 to 3, wherein the model formula includes a wavelength correlation function for indicating a refractive index. 5. The film thickness measuring device according to any one of claims 1 to 3, wherein the predetermined wavelength range includes a wavelength of an infrared light field. 6. The film thickness measuring device according to any one of claims 1 to 3, wherein the analyzing unit comprises: a Fourier transform unit for performing discrete Fourier transform on the wave number distribution characteristic. 7. The film thickness according to any one of claims 1 to 3, wherein the analysis unit utilizes an optimization method for obtaining the wave number distribution characteristic. The amplitude value of each wavenumber component. 8. A method for measuring a film thickness, comprising: an illuminating step of illuminating a measuring light having a predetermined wavelength range onto a test object on which a layer is formed on a substrate, the object to be tested comprising a layer closest to the light source And a second layer adjacent to the first layer; a process of obtaining a characteristic of the skin knife 16 according to the light reflected by the object to be tested or penetrating the object to be tested i..t first to obtain reflectance or penetration Wavelength distribution of the rate 59 201007117 The first determining step is to apply the wavelength distribution characteristic to determine at least the film thickness of the first layer by using a model formula having a film thickness of each layer included in the object to be tested a generating step of converting the correspondence between each wavelength of the wavelength distribution characteristic and the reflectance or transmittance of the wavelength into a wave number associated with each wavelength and according to a predetermined relationship ^ > mm &amp; Corresponding relationship between the calculated conversion values is used for wave number distribution characteristics; amplitude value obtaining step, amplitude value of each wave number component; used to obtain the wave number distribution characteristic The large wave number component in the second decision step, the activation step, and the decision unit are valid. The amplitude value included in the wavenumber distribution characteristic is used to determine at least the film thickness of the first layer; and to selectively allow the first determining unit and the first 60