JP2009031301A - Device for measuring thin film characteristic using two-dimensional detector and its measurement method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a noncontact and non-destructive measurement device measuring the profile and refractive index distribution of multilayer thin film thickness using the principle of reflective light meter. <P>SOLUTION: The profile and refractive index distribution or the like of multilayer thin film thickness on a substrate are simultaneously measured locally by using one or more narrow band optical filter and two-dimensionally arranged photodetectors and by searching the optimum value for a principle numerical expression in which the thin film thickness and the refractive index are expressed by a nonlinear function by an iterative numerical computation method. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an apparatus for measuring the thickness shape and refractive index distribution of a multilayer thin film using the principle of a two-dimensional reflection photometer, and a measurement method therefor.

  The process of measuring the thickness and refractive index of the thin film deposited on the substrate in the coating process of products for semiconductors and various displays improves the quality by observing and monitoring the formation process of the semiconductor product, and early defective products It plays an important role in saving process costs by discovering. In particular, monitoring the thickness shape and the flatness of the thin film layer surface is an essential process for improving the quality of semiconductors.

  Recently, a thin film layer thickness and refractive index measuring apparatus widely used is based on the principle of a reflection photometer. Reflectance photometer, also called thin film layer measuring device in a broad sense, is a non-contact, non-destructive measuring device that can measure the characteristics of multi-layer thin films, and directly on the target sample without the need for special preparation or processing. Measurement is possible.

  The configuration of the reflection photometer that is generally widely used is as shown in FIGS. 1A and 1B. Light starting from the light source 100 is reflected by the light splitter 102 and enters the sample thin film 110 through the objective lens 104. The sample substrate 108 includes a substrate 106 and a sample thin film 110 formed thereon. A part of the light projected onto the sample thin film 110 is reflected at the surface 116 of the sample thin film 110, that is, one point 116 of the boundary 112 between the sample thin film 110 and the atmosphere. Is reflected by the surface 114 of the substrate 106, that is, the boundary surface 114 between the substrate 106 and the sample thin film 110, and is further projected through the objective lens 104 to the spectroscope 130 through the light detection hole 120 at the central portion of the light splitter 102 and the reflector 118. After that, the spectroscope 130 analyzes the reflected light projected, and mainly determines the intensity of the reflected light as a function of the light wavelength. As a result, the thickness, the refractive index, etc. of the sample thin film 110 are calculated through the numerical value converter 132 and the information processing unit 134 to obtain a measured value.

  In the example of FIG. 1A, a part of the light projected to one point 116 on the surface 112 (boundary surface) of the sample thin film 110 is reflected, and the remaining part is refracted (penetrated) into the sample thin film 110 through the boundary surface 112. However, a part of the penetrated light is reflected by the interface 114 between the sample thin film 110 and the substrate 106, and the rest is refracted (penetrated) inside the substrate 106. FIG. 2 illustrates the case where there are thin film layers having different thicknesses, unlike the example of FIG. 1A.

  Referring to FIG. 2, the light beam 210 projected through the objective lens 104 in FIG. 1A is partially reflected at a point 217 on the boundary surface I 207 toward the reflected light beam 222, and partly passes through the second thin film layer 202. It is transmitted and reflected (224) at a point 218 at the interface II 208, and a portion is transmitted (214) through the interface II 208. A part of the transmitted light 214 is reflected at one point 219 of the boundary surface III 209 and goes in the direction of the reflected light beam 226, and a part 216 is transmitted to the substrate 206.

  As shown in FIG. 2, the light 222, 224, and 226 that is reflected by the sample substrate 230 is an optical interface between various media such as the first thin film layer 204 and the second thin film layer 202 on the substrate 206. Since the light beams are emitted by parallel multiple reflection at I207, II208, and III209, and based on the start line 228 in the air, there is a fine optical path difference between them, and the optical Interference is caused by the path difference. Here, when this optical path difference is viewed optically, a different path difference is generated for each wavelength, thereby causing different interference phenomena such as mutual reinforcement interference or mutual cancellation interference depending on the wavelength of the light beam. If the reflectivity is expressed as a function of the light wavelength due to such an interference phenomenon, it has a typical form as shown in FIG. Here, the horizontal axis represents the light wavelength, and the vertical axis represents the reflectance obtained by dividing the reflected light by the incident light.

  Further referring to FIG. 1A, since the light beam reflected by the sample substrate 108 is a superimposed wave of various wavelength components, in order to obtain the reflectance for each wavelength from this superimposed wave, first the reflected light beam is used. Although the signal must be separated by wavelength, this reflected light separation by wavelength is performed by the spectrometer 130. Physically, a prism is the simplest form of spectroscope, but it is usually decomposed into monochromatic light components using a diffraction grating. Therefore, after detecting the intensity of reflected light for various wavelengths using a monochromator and a single detector made of a rotating diffraction grating, or a fixed diffraction grating and an array detector, a numerical converter In 132, the data is converted into numerical data, and the information processor 134 calculates the reflectance by wavelength.

  The reflected light reflectance graph as shown in FIG. 3 has a unique shape and size depending on the thickness of the thin film layer and the characteristics of the refractive index distribution of each thin film layer and the substrate. In the case of a single thin film layer, the reflectivity is theoretically given in a closed form. However, in the case of multiple thin film layers, the reflectance can be calculated numerically from the electric field-magnetic field relationship expressed by multiplication of the eigenmatrix for each thin film layer. Organizing the eigendeterminant described above, the nonlinear function, unlike the single layer, which functionally relates the three variables: refractive index, thin film layer thickness, and reflectivity. Given in. In the case of such nonlinearity, it is often a substantial method to obtain an optimal solution by an iterative trial and error method. Specifically, if the reflectance graph as shown in FIG. 3 is known, the initial value is set using the thin film layer thickness as a variable, and the reflectance calculated by the above-described nonlinear function equation using the initial value set here. And the actually measured reflectivity to obtain an error, and the thin film layer thickness is repeatedly calculated by an iterative trial and error method so that the error is minimized. If a thin film layer thickness that minimizes an error from the actually measured reflectance is obtained, the obtained thin film layer thickness becomes a desired value. Here, the intensity of incident light for calculating the reflectance is measured using a known sample substrate and light source. The refractive index may be calculated using the reflectance and the numerical information associated therewith. Such a method is one field of the so-called “model measurement method”. As described above, the principle of the reflection photometer is to obtain the thickness and refractive index of the thin film by applying the above-mentioned “optimization technique” by trial and error after measuring the reflectivity in engineering.

  A widely used reflection photometer is used to measure the thickness of a thin film layer at a specific “point” of the thin film layer, which is formed to know the uniformity of the thickness of the thin film layer. In order to measure the thickness at a specific point, a specific portion of the irradiated real image 122 of FIG. 1B through the light detection hole 120 having a diameter of about 200 μm in the central portion of the reflector 118 of FIGS. 1A and 1B. Only the reflected light is collected, that is, the thickness of the thin film layer is measured using the reflected light incident through the light detection hole 120. Depending on the spectroscope, as shown in FIG. 4, a glass fiber 424 having a diameter of about 200 μm is installed in the upper plate 423 for lighting, and the irradiated portion as shown in FIG. In some cases, a method of collecting light passing through the glass fiber 424 through the 200 μm diameter hole in the real image 122 may be used.

  On the other hand, an apparatus and a method for measuring a thickness shape over a wide area are presented in Patent Document 1 (US Pat. No. 5,333,049 (Anthony M. Ledger)). According to Leisure's invention, a thickness measuring device with a silicon wafer size of 100 mm was implemented using the principle of white light and interferometer, but the measuring method divided the wafer area into 400 small areas, After measuring the reflectivity for each small area, the thickness is read in a predetermined reflectivity vs. thickness table using a standard wafer. Here, the thickness range is made into at least 500 sections. In other words, after measuring the reflectance with the sample, the thickness value is read with a lookup table obtained with a standard wafer. This method has the advantage that the measurement speed is fast and the entire wafer can be observed, but the error that occurs when obtaining the standard wafer data is transmitted to the entire production wafer, and the CCD (Charge Coupled Devices) of commonly used cameras. ) The sensor has a disadvantage that its resolution is not sufficient for measurement over the entire area of the sample wafer, which is 100 μm or more. Here, the resolution problem occurs when a part of a circuit specified in a semiconductor process having a high degree of integration of an electric circuit is inspected. This is because in the process of processing a wafer with a high degree of circuit integration, it is necessary to observe and inspect in detail the thin film thickness and the formation state of a local portion on the wafer. Yet another disadvantage of the Leisure invention is that a new standard wafer reflectivity contrast thin film thickness database must be reconstructed in the field each time the wafer process changes. Further, in the above-mentioned leisure patent, there is a disadvantage that so-called noise included in the reflectance value measured on the sample wafer is directly reflected and transmitted to the thin film thickness value. In order to partially overcome such disadvantages, another patent of Leisure (Patent Document 2 (U.S. Pat. No. 5,365,340)) standardizes the reflectance value measured on a sample wafer and then standard wafer. The method of reading the thickness value compared with the value of the database of is presented. The reflectance value itself measured here is standardized by calculating the value of a given scale function and then obtaining the optimized value. However, all of the remaining disadvantages associated with Patent Document 1 (US Pat. No. 5,333,049) are also applicable to Patent Document 2 (US Pat. No. 5,365,340).

  The common and serious disadvantages of the two aforementioned leisure patents 1 (US Pat. No. 5,333,049) and 2 (US Pat. No. 5,365,340) are the thin film In order to determine the thickness value by comparing it with the database value previously measured and organized by the standard wafer, the thin film thickness value is influenced too much by the value of the standard wafer database. In other words, the standard wafer reflectivity contrast thin film thickness database is simply a table of reflectivity contrast thin film thicknesses determined by the overall average, so the surface conditions of films such as standard and sample wafers When most of is uniform, its accuracy is guaranteed to some extent. However, when the surface is slightly bent and the relationship between the reflectance and the thin film thickness is irregular, the accuracy of the value of the thin film thickness decreases.

  In order to overcome such disadvantages, a coefficient characteristic such as reflectance was calculated by a theoretical formula in Patent Document 3 (US Pat. No. 5,555,472 (Paul J. Clapis and Keith E. Daniell)). After creating a library of theoretical coefficient characteristics, the thickness value of the thin film is minimized by minimizing the error value between the reflectance value measured at various points on the sample wafer and the theoretical coefficient characteristic value. The process of determining the optimal method is presented. This method was applied to measure the thickness of two films, but again, a method of measuring the thickness of a film layer on the assumption that one of the two layers maintains its uniformity for the most part. presentation. The three prior art examples described above are apparatuses for measuring the entire sample wafer. Therefore, there is a limit in the resolution to measure the reflectance with the CCD camera over the entire surface of the sample wafer, and the resolution problem is serious to measure the thickness shape of the limited region in detail. Noise is generated and transmitted to the reflectance measuring device.

Compared to this, for example, Patent Document 4 (US Patent No. 4,999,014), Patent Document 5 (US Patent No. 4,999,508), and Patent Document 6 (US Patent No. 4,999,014) No. 509) presented a method for determining the value of the thin film thickness by measuring the reflectance at one “point” of the thin film on the sample wafer. These are thin film thickness measuring devices that measure the thickness of a thin film using a typical conventional spectrophotometer, and measure the thickness and refractive index of the thin film layer at a specific point. However, only information on the thin film layer thickness and refractive index measured at one point is insufficient to evaluate the characteristics and quality of the thin film layer. On the other hand, if information on the thickness shape and refractive index distribution of a wider area is obtained, a very meaningful result can be obtained compared with the conventional point measurement method for evaluating the characteristics and quality of the thin film layer. Furthermore, with existing spectrophotometers, it is impossible to measure the thin film thickness shape of the entire relatively large target region and the thin film thickness distribution at various points close to each other. Of course, if you want to find the thin film thickness distribution in a certain area, you can continue to measure while moving the sample substrate little by little to the left and right and up and down. A transfer device is required and it takes a long time to measure. Furthermore, if the thin film thickness distribution in a fine region is to be obtained, the movement of the ultra-precise sample substrate transfer device on which the sample substrate is mounted must move with a precision of 0.1 micron or less, so its control function is also There is a problem that the whole measuring instrument is functionally and structurally very complicated and expensive because it is not easy and an ultra-fine moving platform must be adopted. In this case, an expensive ultra-fine linear moving table can be used, but this is not practical in terms of economy.
US Pat. No. 5,333,049 US Pat. No. 5,365,340 US Pat. No. 5,555,472 US Pat. No. 4,999,014 US Pat. No. 4,999,508 US Pat. No. 4,999,509 US Pat. No. 5,042,949

  As described above, it is practically impossible to simultaneously measure the thickness shape and refractive index distribution of the thin film layer over a given area of the thin film on the sample substrate using a conventional reflection photometer, In addition, in order to measure the thickness shape over a given area in a stepwise manner, in order to solve the problem that the measurement device must be made with an expensive structure and the stepwise measurement method must be used at a low speed, a new measurement is performed in the present invention. An apparatus and method are presented. The main object of the present invention is to measure the thickness shape and refractive index of a multilayer thin film at a large number of points within a given fixed area on a sample thin film, and to measure the thickness shape and refraction of the thin film layer within a short time. While the rate distribution is obtained, the structure is simple and the equipment based on the principle of the spectrophotometer is presented.

  The refractive index is usually calculated from the reflectance, and the reflectance is calculated by an iterative numerical calculation method based on an optical principle in which the measured value, the thickness of the thin film layer, and the refractive index are displayed as nonlinear functions. It is obtained by searching for an optimum value so as to minimize an error from the value.

  FIG. 5 is a conceptual diagram of a thin film layer thickness shape and refractive index distribution measuring apparatus according to the present invention. Referring to FIG. 5, the multilayer thin film thickness shape and refractive index distribution measuring apparatus according to the present invention is roughly divided into an optical unit, a sample substrate transfer unit, an image capture processing unit, an image processing unit 548, an information processing unit 550, A system control unit 552 and an information display unit are included.

  The optical unit includes a light source 500 that generates measurement light, a focusing lens 502, a diaphragm 504, a projection lens 506, a light splitter 508, an objective lens 510, an auxiliary lens 530, and a filter having a large number of narrow-band-pass light filters 538. It is composed of a ring 534. As the light source 500 for measuring the thin film layer thickness, a commonly used visible light is used. The light generated from this light source is collimated through the focusing lens 502, the diaphragm 504 that adjusts the projection area of the focused light projected onto the sample substrate 514, and the parallel light lens 506, and is projected onto the light splitter 508. Then, the light is reflected and reflected by the boundary surface I 520, which is the upper surface of the single-layer thin film layer 518, and the boundary surface II 522, which is the lower drawing, above the sample substrate 514 through the objective lens 510. The light reflected by the sample substrate further passes through the objective lens 510, passes through the light splitter 508, passes through the narrow band pass optical filter 538 by the auxiliary lens 530, and is then collected on the CCD photodetector 542. The reflected light that has passed through the narrow-band light filter 538 is linked to the two-dimensional photodetector 542 with light corresponding to a specific wavelength band determined by the optical filter.

  The image capture processing unit includes a two-dimensional photodetector 542 and an image capturer 546. An optical path in which the reflected light reflected by the boundary surface I520 and the boundary surface II522 of the thin film layer 518 on the sample substrate 514 includes an objective lens 510, a light splitter 508, an auxiliary lens 530, and a narrow-band pass optical filter 538. A pixel unit through a photodetector by a charge coupling device (CCD: Charge Coupled Device) mounted in the two-dimensional photodetector 542 from the optical image projected on the two-dimensional photodetector 542 through It changes to luminosity information. The two-dimensional image is captured by the image capturer 546 and then stored in a frame memory (not shown). More specifically, CCDs are arranged in a two-dimensional type in the two-dimensional photodetector 542. The reflected light reflected by the boundary surface I520 and the boundary surface II522 of the sample thin film layer 518 passes through the narrow-band pass light filter 538, and then only the light having a wavelength belonging to the narrowband is the two-dimensional CCD. The pixels sensed by the CCD form a two-dimensional pixel group. At this time, the formed pixel group is captured through the image capturing unit 546 and then stored in a frame memory (not shown).

  The sample substrate transfer unit includes a sample substrate transfer table 524, a sample substrate transfer table drive unit 526, and a sample substrate transfer table control unit 528. The sample substrate transfer table 524 is a stage that supports the sample substrate 514, and moves in the vertical and horizontal directions and in the vertical direction by the motor of the sample substrate transfer table driving unit 526. This movement is controlled by the sample substrate transfer table controller 528, and the sample substrate transfer table controller 528 is controlled by the system controller 552. The system control unit 552 communicates information with the input / output device 558.

  The image processing unit 548 extracts the reflected light intensity from the video information captured by the image capturing unit 546 as a function of wavelength and obtains a reflectance graph as shown in FIG. 7 as a function of light wavelength. Here, the reflectance is defined as reflectance = reflected light intensity / incident light intensity, and the incident light intensity is measured using a known standard sample.

  The information processing unit 550 is a kind of variable calculation unit in terms of function, and calculates the thickness of the thin film layer, the thickness shape of the thin film, the refractive index distribution, etc. in two dimensions from the reflectance graph as shown in FIG. Perform the function to start out. Depending on the design, the image processing unit 548 and the information processing unit 550 may be controlled by a single control unit. Furthermore, the image processing unit 548, the information processing unit 550, the system control unit 552, and the sample substrate control unit 528 can be integrated and designed to be controlled by a total control unit or a computer.

  The image recognition unit 554 and the monitor 556 display the information processed and taken out by the image processing unit 548 and the information processing unit 550 based on the pixel information captured by the image capturing unit 546 through the monitor 556, thereby allowing the user of the measuring apparatus. Making the measurement device easy and convenient to use. In particular, the parametric variables including the thin film layer thickness and the refractive index distribution extracted by the image processing unit 548 and the information processing unit 550 are displayed on the screen through the monitor 556.

  The system control unit 552 serves to control and adjust the measuring apparatus including the monitor 556 according to the present invention as a whole.

  Next, the operation principle and function of the apparatus for measuring the thickness shape and refractive index distribution of the multilayer thin film according to the present invention will be described in detail with reference to FIG.

  Light emitted from the light source 500 is projected onto the sample substrate 514 through an optical path that passes through the focusing lens 502, the stop 504, the parallel light projection lens 506, the light splitter 508, and the objective lens 510. The sample thin film 518 to be measured here will be described with reference to an example in which a structure formed of two layers of thin films 202 and 204 on the substrate 516 is used as the sample substrate 514 as shown in FIG. Referring to FIG. 2, the sample substrate 230 has a structure in which a first thin film layer 204 and a second thin film layer 202 are formed on a substrate 206. In this case, these thin film layers have different optical characteristics such as refractive index, and the boundary surface I207 between the second thin film layer 202 and air, the boundary surface II208 between the second thin film layer 202 and the first thin film layer 204, There are three boundary surfaces, that is, a boundary surface III209 between the first thin film layer 204 and the substrate 206. Reflection, refraction, and transmission phenomena occur at the optical interface between thin film layers having different optical characteristics, so that the reflected light reflected by the entire sample substrate 230 as shown in FIG. → 222, 210 → 212 → 224, 210 → 212 → 214 → 226), which is composed of light 222, 224, and 226 on which the reflected light is superimposed on several boundary surfaces. In general, the thickness of the thin film layer to be measured is as thin as several microns to several tens of millimeters, so that the path difference between the reflected light reflected by the thin film is shorter than the so-called optical coherence distance. As is well known, an interference phenomenon occurs between reflected lights when the reflected lights overlap. In addition, even if the two optical paths are geometrically the same, when the wavelength of the optical signal is different, the optical path is different depending on the wavelength, thereby causing an interference phenomenon that differs depending on the wavelength. Accordingly, a destructive interference phenomenon occurs at a certain wavelength, and a reinforcing interference phenomenon occurs at another wavelength. The intensity of reflected light at this time has a value that varies depending on the light wavelength. The intensities of incident light and reflected light also have different values depending on the wavelength within the narrow band. Further, the interference phenomenon described above is caused by the geometric difference of the reflected light, that is, the optical path difference depending on the thickness of the thin film layer, and the optical path difference determined by the difference in refractive index that varies depending on the optical characteristics of the medium layer. Since the reflectance is defined as reflectance = reflected light intensity / incident light intensity, if the incident light intensity is measured using a standard sample, a reflectance graph as shown in FIG. 3 is obtained as a function of the light wavelength. It becomes like this. In the present invention, the thickness of the thin film layer and the nonlinear function relationship that relates the reflectance and the refractive index, which are the characteristics of the thin film layer, are obtained, and after the reflectance is measured by the method presented in the present invention, the thickness of the thin film layer is measured. And the refractive index distribution are calculated by a method of searching for an optimum value by repeated trial and error.

  The theoretical background for finding the optimum values of thin film layer thickness and refractive index according to the present invention will now be described with reference to "Optics" by Eugene Hecht, 4th Edition, Addison-Wedley, 2002).

  The theoretical reflectivity when knowing the thickness and refractive index of the thin film layer is given in a relatively simple closed form in the case of a single thin film layer. However, in the case of multiple thin film layers, a numerically theoretical transmissibility graph can be obtained using a linkage equation that can be understood from the boundary condition between the electric and magnetic thin films obtained by multiplication of the characteristic matrix of each thin film layer. It starts with asking.

  According to the present invention, the light emitted from the light source is projected almost perpendicularly onto the sample substrate to be measured, and in the case of a single thin film sample composed of an incident medium, that is, an air-thin film layer-substrate. First consider. Referring to FIG. 5, at this time, the reflection coefficient r is expressed in a closed form as given in the following formula [1] as a complex number, and the absolute reflectance R having an energy meaning is a reflection coefficient expressed in a complex number. It is given by equation [2] which is the square of r.

Here, r _0l and r _1S are the Fresnel reflection coefficients at the boundary surface between the incident medium (air) and the thin film layer, that is, the boundary surface I520 and the boundary surface between the thin film layer and the substrate, that is, the boundary surface II522. β is the amount of phase change that occurs when light passes through the thin film layer,

Given in. Here, η is the refractive index, d is the thickness of the thin film layer, and λ is the light wavelength. * Indicates that it is a conjugate complex number. What should be noted here is that the thickness of the thin film can be determined by measuring the phase change amount β and the refractive index η using the optical wavelength λ as a variable. Patent Document 7 (US Pat. No. 5,042,949) is an example.

  In the case of the multiple thin film layers, the intensity of the electric field and the magnetic field of the light at the upper and lower boundary surfaces of the i-th thin film layer is not expressed by the closed-form expressions such as the mathematical formula [1] and the mathematical formula [2]. Has a functional relationship as given by the following equation [3].

  Here, i = 1, 2, 3,..., P,

Is a characteristic matrix that links the relationship between the electric field and the magnetic field between the upper boundary surface and the lower boundary surface of the i-th thin film layer as a function, and each element of the matrix M _i , _{mi, 11} , _{mi, 12} , and _{mi , 21} , _{mi, 22} are functions such as the complex refractive index at the upper and lower boundary surfaces of the i-th thin film layer, the thickness of the thin film layer, and the light wavelength. The intensity of the electric field and magnetic field of light at the interface between the incident medium-the uppermost thin film and the lowermost thin film-substrate of the multiple thin film sample substrate composed of p thin films is expressed as follows from the equation [3]. A functional relationship given by the mathematical formula [4].

  here,

  Then, the following matrix equation is obtained under the boundary condition that the electric field and magnetic field at the boundary surface between the uppermost layer thin film and the lowermost layer thin film should satisfy.

  If this is rearranged, the following formula [5] is obtained.

  here,

  r is the reflection coefficient, and t is the pass coefficient. Solving the equation [5] with respect to the reflection coefficient r, the following equation [6] is obtained.

Here, γ _o and γ _S are complex refractive indexes of the incident medium (air) and the substrate, respectively, and m ₁₁ , m ₁₂ , m ₂₁ , and m ₂₂ are elements of the characteristic matrix M and the like. m ₁₁ , m ₁₂ , m _21, and m ₂₂ are obtained by multiplying all of M ₁ , M ₂ ,..., M _P as indicated by Equation [4], and each of these elements is determined by the thin film layer. It is a function of the thickness d, the reflection angle of the reflected light, and the absolute refractive index n. Then, the absolute reflectance R is obtained by Expression [2]. That is,

Given in. Here, * means a conjugate complex number.

  According to the present invention, since the theoretical reflectance equation [6] is derived as described above, if the reflectance graph is obtained by measurement, the theoretically calculated reflectance and the reflectance of the measured reflectance graph are calculated. On the contrary, the thickness and reflectivity of the thin film can be obtained. More specifically, since the reflectance r is a nonlinear function of the thin film layer thickness d in the above equation [6], in order to obtain the thickness d of the thin film layer when a measured value of the refractive index r is given. First, after setting the thickness of the thin film layer to an arbitrary value, the theoretical reflectance corresponding to this thickness is calculated from the formula [6] for each light wavelength, and the theoretical reflectance obtained here is calculated. And the measured reflectance, the error is ideally zero if the thin film layer thickness assumed above is accurate. However, in practice, there is generally an error between the measured reflectance and the reflectance calculated through Equation [6] even if the value of the thin film layer thickness is a true value. At this time, the thin film layer thickness is set as a variable, and the “optimum” thickness of the thin film layer is determined using an optimization technique that seeks the optimum value of the thickness by continuing trial and error to minimize the reflectance error. Can be determined. Here, according to the present invention, since the reflectivity varies depending on the wavelength of the reflected light, a method for minimizing the error includes non-linear minimum error that minimizes the total error with respect to the reflectivity measurement values at all wavelengths. Use the search method. For example, a nonlinear minimum error square method such as the Ravenberg-Marquardt method may be used.

  Since the refractive index can be easily obtained when the reflectance is given as described above, according to the present invention, when the refractive index of the thin film layer is not known, the distribution of the refractive index with respect to the light wavelength is expressed as a Cauchy model or Lorentz. What is necessary is just to obtain | require using models, such as an oscillator model. In this case, assuming that the refractive index of the thin film layer is expressed by the selected model, the error between the theoretical reflectance and the reflectance calculated by the assumed model is expressed by a function of the model coefficient of the assumed model. If the model coefficient is determined so that the error between the calculated refractive index and the measured refractive index is minimized, the value of the model coefficient at this time becomes an optimum value. In other words, the optimum refractive index that minimizes the error is determined. Therefore, if the coefficient of the refractive index model is set as a variable as in the case of the thin film layer thickness and the optimization technique is applied to the reflectance error, the refractive index according to the light wavelength can be obtained. The non-linear minimum error method may be applied to the optimization method such as the method described above.

  According to the present invention, as described above, the thickness and the refractive index of the thin film layer are “measured” in each pixel region of the two-dimensional array type photodetector. And the refractive index at the position on the corresponding sample area, the thickness shape and refractive index distribution on the sample area can be obtained, and these distributions can be displayed on the monitor in two or three dimensions. The thickness distribution of the thin film layer can be intuitively easily grasped from the thin film layer thickness shape visualized in three dimensions. The refractive index distribution on the sample area to be measured by the same method can be easily and intuitively grasped.

  As described above, according to the present invention, if the photometric signal sensor in the detector is used in a two-dimensional arrangement, the measurement target region of the sample is within the field of view of the detector. As long as it enters, the thickness of the thin film layer is determined at any position within the field of view of the detector. Therefore, apart from the prior art, it is sufficient to perform the primary transfer of the sample substrate only to the extent that the measurement target portion of the sample falls within the visual field range of the detector. That is, as long as the measurement target portion falls within the visual field range, it is only necessary to obtain the thin film layer thickness at that position by knowing where the measurement target portion is located in the visual field range by image recognition technology. Therefore, a secondary fine transfer device is not required separately from the prior art.

  In order to make an efficient measurement, the thickness and refractive index of the thin film layer are often measured along any specific part or specific pattern of the sample, but according to the present invention, You can easily measure the case by changing the software. That is, by mapping the thickness and refractive index of the thin film layer obtained by the two-dimensional photodetector, values relating to the thickness and refractive index corresponding to a desired specific portion on the actual sample substrate surface can be easily obtained. In other words, all the processes described above are implemented by programming with software.

  As described above, according to the present invention, in an apparatus for measuring the thickness shape and refractive index distribution of a multilayer thin film using the principle of a two-dimensional reflection photometer and its measuring method, a filter as shown in FIG. A method for separating light wavelengths using ring 534 is presented. According to the present invention, a linear variable filter structure of the form shown in FIG. 8 may be used in place of the filter ring, and a partially circular variable filter of the form shown in FIG. A structure may be used. Here, in the linear variable filter and the partially circular variable filter, the optical wavelength of the filter function may be continuously changed or may be changed stepwise as in the filter ring described above. Partially circular variable filter The structure may be made completely circular or may be made as part of a circle. Moreover, the filter function may change continuously like a linear variable filter, and may change in steps like the above-mentioned filter ring. According to the present invention, in FIG. 5, the filter ring 534 rotates about its rotation axis 536, and this rotational movement has control functions such as the system control unit 552 and the information processing unit 550. Control is performed by the block that is being operated (not shown).

  Further, according to the present invention, a tuning filter 539 (indicated by a dotted line) such as a liquid crystal tuning filter or an acousto-optic tuning filter may be used instead of the optical filter ring 534 as shown in FIG. For example, the principle and typical structure of the liquid crystal tuned filter are “Imaging Spectrometry Usable Liquid Crystal Tunable Filters”, presented by Bytom Chryn and Chris Chibi, Jet PromLaboratory, and Peter Miller. The principle and structure of the filter is “Introduction to Acoustic-Optics”, Andalso AOTF (acousto-optical tunable filter) Spectroscopy, by Brimrose Corporation of Presented by merica. In FIG. 5, only a specific wavelength that meets the conditions by the information processing unit 550 and the system control unit 552 is used in the same manner as when the optical filter ring is used for the light source incident on the tuning filter 539 used instead of the filter ring 534. The tuning filter also operates as a narrow band pass optical filter. In the case of the above-described tuned filter, since it operates by an electrical signal without a mechanical driving unit, the enabling of the operation is completed within a short time near real time in units of microseconds (μs). Therefore, the time required for measurement at the actual process site is greatly reduced.

  In the measurement apparatus according to the present invention, after the light reflected by the optical interface of the measurement sample passes through the narrow band pass optical filter, the variable filter, or the tuning filter, a real image is formed on the two-dimensional array type photodetector 542. However, according to the present invention, when the tuning filter 539 is used because the function of the optical filter is electronic, the tuning filter 539 and the two-dimensional array type photodetector are used. 542 can be integrated to make the structure more efficient.

  As described above, the refractive index is expressed as a nonlinear function of the thickness of the thin film layer. Therefore, since the refractive index and the thickness of the thin film layer have a nonlinear function relationship, it is optimal by an iterative trial and error method. It is practical to determine the value. However, for example, there may be a case where one or more minimum values of refractive index error exist within the range of the thin film layer thickness to be measured, and it is difficult to obtain the regional minimum value. In such a case, the measurement equipment user may select the minimum value based on experience and select the thickness of the thin film layer corresponding to that value. As an alternative to this method, according to the present invention, an ultraviolet or infrared light source may be used as the light source. For example, if the thin film layer thickness to be measured is reduced, the reflectance graph of FIG. 3 tends to be flat, and the reflectance is dulled. That is, when a commonly used visible light source is used, the minimum thickness of the thin film layer to be measured is limited to about 100 mm to 200 mm. On the other hand, since the reflectance graph appears densely in the ultraviolet region, it is possible to measure even when the thickness of the thin film layer is several tens of millimeters or more when an ultraviolet light source is used. On the other hand, when the thickness of the thin film layer is relatively very large, if only the visible light source is used, the reflectivity graph of FIG. 3 appears too densely in the light wavelength direction, that is, between the top and the valley. Because there are too many, there are too many local minimum points of the reflectance error value. Therefore, when an optimization method such as the nonlinear minimum error method is applied, it often does not converge well to the global minimum. In such a case, if a light source in the infrared region of a few micron wavelength band is used, the waveform of the reflectance graph as shown in FIG. 3 is more widely distributed in this region. The thickness of the thin film layer of several tens of microns or more can be easily measured by reducing the decrease in the thickness and increasing the possibility of convergence to the minimum value in the entire area.

  The above-described embodiment is merely an example for explaining the principle and contents of the present invention, and is not intended to limit the principle and basic idea of the present invention. Those skilled in the art will readily understand and devise variations such as the detailed description of the invention. Accordingly, those skilled in the art will be able to understand and devise the meaning of the principle and basic idea in a wide range through the detailed description of the present invention described above.

  An embodiment of the present invention will be described with reference to FIG. The apparatus for measuring the thickness shape and refractive index distribution of the multilayer thin film according to the present invention has the structure shown in the conceptual diagram of FIG. 5, but is roughly divided into an optical part, a sample substrate transfer part, and an image capture. A processing unit, an image processing unit 548, an information processing unit 550, and a system control unit 552 are configured.

  The optical unit includes a light source 500, a focusing lens 502, a diaphragm 504, a projection lens 506, a light splitter 508, an objective lens 510, an auxiliary lens 530, and a narrow-band pass light filter ring 534, and a sample substrate transfer unit , A sample substrate transfer table 524, a sample substrate transfer table driver 526, and a sample substrate transfer table controller 528. The image capture processing unit includes a two-dimensional photodetector 542 and an image capture unit 546, and the image display unit includes an image recognition unit 554 and a monitor 556. The measuring apparatus further includes an image processing unit 548 and an information processing unit 550. Finally, there is a system control unit 552 that controls the entire measuring apparatus, and includes a printer, various recording devices, and data transmission. An information exchange device having functions such as a device is connected.

  As described in the configuration and operation of the above-described invention, this measurement apparatus is a general-purpose measurement apparatus that measures the thickness shape and refractive index distribution of a thin film layer over a wide area. As the light source, a light source such as visible light, infrared light, and ultraviolet light can be used. In this embodiment, the case where visible light is used as the light source will be described as an example.

  The thin film that measures well in the visible light wavelength band is a photoresist (photoresist: PR) used in the manufacture of semiconductor elements for state-of-the-art technology that has been widely used in the near future, and its thickness ranges from about 0.3 microns to 3.0 microns. As a sample, a silicon substrate covered with a photoresist thin film was used, and a standard substrate was also used as a standard guide for measurement to clarify the reflectance.

  The optical part has a structure similar to that of a typical microscope, and a tungsten halogen bulb, which is generally used as a light source, is used. First, after the reflected light reflected from the standard substrate that already knows the reflectance placed on the sample substrate passes through the narrow band pass light filter ring 534, the CCD determines the intensity of the reflected light projected on the CCD 542. Measurements were made through a CCD 542 attached to a dimensional mold. Next, the silicon sample substrate 514 was placed on the sample substrate transfer table, and the intensity of the reflected light was measured through the CCD 542 in which the CCD is mounted in a two-dimensional manner, as in the case of the standard substrate. At this time, the reflected light input to the CCD 542 is filtered through the narrow-band light filter 538 attached to the narrow-band light filter ring 534, so that the intensity of the reflected light reflected at a selected point on the sample substrate. Is given as a function of the optical wavelength determined by the optical filter actually used. A schematic diagram of the narrow band pass optical filter ring 534 used in this embodiment is shown in FIG. Narrow bandpass optical filter ring 634 is equipped with 28 optical filters in this embodiment, and the pass region of the optical filter is a minimum of 400 to a maximum of 800 nanometers. The filter ring rotates about a rotation shaft 636 (536 in FIG. 5) located at the center, but the rotational movement is controlled by the system controller 552, the image processing unit 548, or the information processing unit 550. To do. The rotation of the narrow band pass filter ring is automatically controlled by software, and is rotated step by step until the measurement of the reflected light according to the light wavelength in the selected specific area on the sample thin film 520 is completed. The reflectance is given by the ratio of the reflected light intensity measured on the sample substrate and the incident light intensity measured on the standard substrate, that is, reflectance = reflected light intensity / incident light intensity, and is a function of the light wavelength. Therefore, the reflectance graph obtained from the measured reflected light intensity is as shown in FIG. According to the present invention, instead of the narrow band pass optical filter ring 534, the linear variable optical filter 800 shown in FIG. 8 or the partially circular variable optical filter 900 as shown in FIG. 9 may be used. . Of course, an appropriate driving device is also necessary here. That is, in the case of the linear variable optical filter 800, the reflected light projected on the two-dimensional CCD 542 is differentiated and measured for each light wavelength by moving the optical filter linearly. In the case of the partially circular variable optical filter 900, a partial rotational motion is caused in the same manner as in the case of the optical filter ring (534 or 634) described above.

  As described above, according to the present invention, a tuning filter 539 such as a liquid crystal tuning filter or an acousto-optic tuning filter may be used as the optical filter. Unlike the optical filter that directly handles light, the tuning filter 539 is used as an indirect electronic filter instead of the filter ring 534 in FIG. The principle is to electronically perform an operation similar to that of a narrow-band optical filter by electronically selecting only specific wavelengths that meet the conditions of the information processing unit 550 or the system control unit 552 from the incident light. In the case of a tuned filter, the filter function can be driven in real time in microseconds because it operates with an electrical signal without a mechanical drive. Therefore, since the filter function can be implemented within a time much faster than the optical filter that has to move the filter body mechanically to change the optical wavelength band, the measurement time is significantly higher than that of the optical filter. Shortened.

  According to the present invention, it is easily integrated with the CCD 542 to be a tuned filter electronic filter, and a physical location such as an optical filter or tuned filter is an optical path 532 between the light source 500 and the CCD 542 for efficient filter function. It suffices to be located where the light concentrates. For example, the filter may be positioned in front of the light source 500, before or after the stop 504, directly above the focal point 512 on the sample, and at the current position of the optical filter ring 534 as in this embodiment. Of these possible positions, the position immediately before the light source 500 is affected by the least amount of miscellaneous light, so that the filter function is more efficient. In addition, theoretically, the possible positions of the filter may be located at any point in the middle of the optical path between the light source 500 and the CCD 542. However, if the filter is substantially located at a point where light is concentrated in the optical path, it is more efficient in terms of the size, structure, and function of the filter.

  In all the cases described above, the wavelength of the light passing through each narrow-band optical filter is not a single fixed optical wavelength but an optical wavelength band applied to a narrow transmission region. Means a substantial average value with respect to the light wavelength within the narrow band pass range. In addition, a drive device that controls the movement of the linear variable optical filter and the partially circular variable optical filter is automatically configured to measure the intensity of reflected light. Once the reflected light on the standard substrate and the reflected light on the sample substrate are measured as a function of the light wavelength, the relative reflectance is the intensity of the reflected light reflected on the sample substrate. Obtained by dividing by the intensity of light. That is, reflectance = intensity of reflected light from the sample substrate / intensity of reflected light from the standard substrate. As a result, a reflectance vs. optical wavelength graph as shown in FIG. 7 is obtained.

  According to the present invention, as described above, the optical unit is generally constructed in the structure of a light microscope in principle, and the light source 500 selectively requires visible light, infrared light, and ultraviolet light. Use according to the application. As described above, in this example, the thickness and refractive index of the photoresist were measured using a visible light source. Although the magnification of the objective lens 510 can be changed together with the microscope, a 50 × objective lens is used in this embodiment. In the image captured from the sample substrate, a region 512 having a diameter of about 60 μm to 80 μm is set as a measurement target region, and reflected light reflected from the region 512 of this size passes through the objective lens 510, the light splitter 508, the auxiliary lens 530, and the narrow band. The intensity of the reflected light was measured by projecting through a light filter 538 to a sensor composed of a two-dimensional CCD mounted inside the CCD 542. The area of the CCD sensor corresponds to about 60 μm × 80 μm on the sample, and reflected light was sensed with a resolution of 640 × 480 pixels in this region. Actually, after an electrical signal generated by the CCD sensor is captured through the image capturing unit 546, it is stored in a frame memory (not shown), and the image stored in the frame memory is analyzed by the image processing unit 548. And extracted the intensity of the reflected light. More specifically, for example, a basic pixel group of (3 × 3) size is formed, and the entire pixel area of the CCD is divided into pixel groups of (32 × 32) size, and this (32 × 32) is divided. ) The average value of the intensity of the reflected light was obtained for the size pixel group. This process is repeated as many times as the number of narrow band pass optical filters in the narrow band pass optical filter ring. Then, the reflectance graph as shown in FIG. 7 was obtained by standardization. In this embodiment, (32 × 32) in order to measure over the entire image corresponding to all 640 × 480 pixels and according to the number of 28 filters mounted on the narrow band pass optical filter ring. Measurements were made {(640 × 480) / (35 × 35)} × 28 times by dividing into (35 × 35) size pixel units instead of pixel unit sizes. Then, the intensity of (640 × 480) / (35 × 35) reflected light in one image was obtained as a function of optical frequency. This process mainly proceeded in the image processing unit 548.

  Here, as described in the detailed description of the present invention, the thickness of the photoresist film, which is a thin film layer to be measured, is measured by utilizing the method for minimizing the reflectance error. Was mainly enforced by the information processing unit 550. According to the present invention, the image processing unit 548 can perform the calculation necessary at this time. Then, as described above, the measured reflected light intensity was divided by the measured standard reflected light intensity, and the measured value was displayed as reflectance, and the reflectance graph was obtained as a function of light wavelength as shown in FIG. The value of the graph thus obtained was temporarily stored in a memory and then used to calculate the thickness of the thin film layer. Here, it is also possible to directly calculate the graph value without saving it in the temporary memory. In this example, the value of the reflectance graph was temporarily stored, and then the thickness of the thin film layer was calculated by the image processing unit.

  According to the present invention, in the detailed description of the invention, the mathematical expression [2] is expressed as the square of the reflectance in the case of a single thin film layer. That is,

r is the complex reflectance given by the equation [1] and is a complex number, and this complex reflectance r is a function of the thin film layer thickness d. Here, R is an absolute reflectance. And, actually, the measured reflectance is referred to as r _m, the calculated reflectance

Speaking, error r _e of the reflectance and

Given in. Here, R, r _m , r _c , and r _e are all integers. As a result, after determining the initial value d _l of the thickness of the thin film layer, r _cl is calculated by Equation [2],

Asked. Then the thickness

After a further change, r _c2 was further calculated, and then error r _e2 was further obtained. Smaller than where if the error r _e2 is the error r _el, error while continuing to increase the value of the thin film layer thickness gradually progresses until minimized. If the error increases in the middle, the thickness value is decreased. At this time, the thickness decreasing amount is changed to be smaller than the increased value. As a result of setting and calculating the initial value d _l of the thickness, if the error r _e2 is larger than r _{el, the} error is reduced to the minimum value by calculating the error while decreasing the value of the thin film layer thickness. Repeat until. At this time, if the error increases, the thickness of the thin film layer is slightly increased as described above. This process is repeated to find the thin film layer thickness when the error is minimized. Applying the above-mentioned repetitive trial and error method, the thin film layer thickness in the measurement region on the sample selected by calculating the thickness of the thin film layer corresponding to each pixel group having the (3 × 3) size described above. The shape of was determined. The results were displayed on the monitor 556 as two-dimensional and three-dimensional images.

  The thin film layer thickness measurement process is as shown in FIG. 7 because the reflectivity varies depending on the reflected light wavelength. In this embodiment, after 28 calculations are repeated, The errors were added together and used in the calculation.

  As described above, the minimum thickness of the thin film layer that can be measured with a visible light source is limited to about 100 to 200 mm. Therefore, when the thickness of the thin film layer is, for example, 100 mm or less, the thin film layer thickness is measured using an ultraviolet light source. That is, the thinner the thin film layer, the shorter the wavelength of the light source. This result appears in a dense graph in the optical wavelength direction in the reflectance graph of FIG. In order to change the wavelength of the narrow-band-pass optical filter in the shorter direction when using the ultraviolet light source, for example, the 28 optical filter set attached to the optical filter ring in the first example of this embodiment is also light It must be changed to a narrow bandpass optical filter set with a short wavelength. Conversely, if the thickness of the thin film layer to be measured is large, for example, 5 μm to 10 μm, the measurement can be easily performed by changing the light source to an infrared light source. In this case, the reflectance graph of FIG. 7 spreads along the optical wavelength axis. Also at this time, as an example of the embodiment, the optical filter set mounted on the optical filter ring must be changed to a narrow band pass optical filter having a long wavelength. The remaining image processing steps are the same as when a visible light source is used when an ultraviolet light source is used or when an infrared light source is used.

  The sample image information received by the image processing unit 548 through the image capturing device 546 is the intensity of the light sensed at the position of the (3 × 3) size pixel group given through the CCD mounted on the two-dimensional CCD 542 and its ( 3 × 3) is the actual position of the pixel group, and information on the light wavelength incident on the pixel group is the position of the specific optical filter selected from the optical filter set attached to the optical filter ring 534. Determined by. Information on the light wavelength is directly transmitted from the optical filter ring 534 to the image processing unit 548 or the system control unit 552. At this time, information on the light incident on the pixel group of the CCD sensor 542 and the light wavelength are synchronized.

  As described above, the light intensity corresponding to the given (3 × 3) size pixel group, the position of the image represented by the pixel group, the light wavelength relative to the pixel group, and the initial value are arbitrarily assumed. Applying the value of the thin film layer thickness to Equation [2]

There are several constants and coefficients in equation [2], and there are also sine and cosine functions. Moreover, by repeatedly calculating the r _c,

Iterative operation until is minimized. According to the present invention, in order to shorten the calculation time, after partially calculating the formula [2], the value is stored in the memory in advance, and then the stored value is read and used to shorten the calculation. You can use a lookup table method that enforces time. Here, it is customary for the designer to determine how to divide the partial calculation of the formula [2], but it is also possible to design by the user selection method.

According to the present invention, for example, when the pixel group having the (3 × 3) size described above is regarded as one region, a reflectance error corresponding to each pixel is obtained with respect to a pixel group including 3 × 3 = 9 pixels. Since the average value must be minimized after calculation over all light wavelengths, the “optimization calculation process” for minimizing the average reflectance error used at this time is quite complicated. Moreover, the reflectivity error r _e at the position of a given pixel group for a non-linear function such as the wavelength of the incident light, the film thickness, must be applied to minimize the operation process of the so-called non-linear error repeatedly There is. In the present invention, as an example, the “Redenburg-Marquardt nonlinear error square minimizing method” which is one of the nonlinear error minimizing methods described above is applied. That is,

Is to minimize the value of. However, in the present invention, another error minimizing method similar to the above method may be applied.

  In this embodiment, the refractive index of the sample to be measured is known in order to measure the reflectance of a given photoresist thin film layer. However, in some cases, the refractive index of the thin film layer to be measured may not be known. At this time, a method for obtaining the refractive index distribution with respect to the optical wavelength using a model such as a Cauchy model or a Lorentz oscillator model is presented. It was done. ("Spectroscopic Ellipsometry and Reflectometry" by HG Topmpkins, WA McGahan, John Wiley, 1999). If the refractive index of the thin film layer is modeled with the above-mentioned specific model, the error between the theoretical reflectance and the measured reflectance is displayed as a function of the model coefficient of the corresponding model, and the value of the specific model with this model coefficient Is the closest to the actual refractive index value, the error is minimized. At this time, as in the case of the thin film layer, the coefficient of the refractive index model is set as a variable, and the refractive index according to the light wavelength is obtained by applying the optimization method to the reflectance error. As the optimization method, the nonlinear minimum error method is applied as described above.

  The complicated calculation as described above is performed by the image processing unit 548 and the information processing unit 550. Although the function of the image processing unit 548 also plays other roles, it performs a role of extracting a reflectance measurement value necessary for the reflectance error calculation, etc., and performs refractive index calculation, reflectance error calculation, reflectance error optimization, Refractive index model selection and calculation when the refractive index is not known are mainly processed by the information processing unit 550. However, since the function sharing between the image processing unit 548 and the information processing unit 550 depends on the designer's discretion, the image processing unit 548 and the information processing unit 550 may be combined into one in some cases. In this embodiment, for convenience, the image processing unit 548 and the information processing unit 550 are divided into two other blocks and illustrated in FIG. In order to efficiently use the look-up table constructed according to the present invention, the arithmetic system is usually configured to perform its function closely with the information processing unit 550 or the image processing unit 548.

  A signal corresponding to the pixel extracted and given by the image capturing device 546 is also transmitted to the image recognition unit 554, and the image recognition unit 554 selects a selected measurement region on the sample, that is, the narrow band pass optical filter of FIG. The image of the sample surface passing through the ring 534 is recombined and displayed on the screen through the monitor 556. The user of the measuring machine designates a desired measurement area on the monitor screen and performs calculations necessary for measurement by the image processing unit 548 and the information processing unit 550. At this time, the user can designate only a limited measurement region in order to save measurement time. For example, an image of the entire measurement region passing through the narrow-band optical filter ring 534 in FIG. It can also be displayed. The image information displayed on the screen of the monitor 556 includes the surface state of the sample substrate, the shape of the thin film layer, the refractive index distribution, the reflectance distribution, the XY coordinates of the sample, information on the narrow band pass optical filter ring, and the selected measurement. This is information such as the magnification of the area video.

  The thin film layer thickness shape and refractive index distribution measuring apparatus according to the present invention uses an automatic focusing apparatus that automatically focuses the objective lens 510 of the optical unit when measuring various portions of the sample substrate 514. . At this time, the relative positions of the objective lens 510 and the sample substrate 514 move in three directions of XYZ. At this time, the movement of the sample substrate transfer table 524 that moves together with the sample substrate 514 is controlled by the sample substrate transfer table driver 526 and the sample substrate transfer table controller 528 that adjusts the movement.

  According to the present invention, the system control unit 552 performs overall control of the thin film layer thickness shape and refractive index distribution measuring apparatus exemplified in this embodiment. The system control unit 552 is basically a controller configured mainly with a microprocessor, a microcomputer, or a microcontroller. Since the system control unit 552 includes hardware and software, the hardware is typically a microprocessor, a main memory, a hard disk, an I / O connection part, an I / O device such as a printer. The software is roughly divided into a start program, a main control program, and the like. This system control unit 552 controls various parts of the thin film layer thickness and refractive index distribution measuring apparatus. That is, an optical unit, a sample substrate transfer table control unit 528, a narrow band pass light filter ring 534, a two-dimensional CCD 542, a frame grabber (image supplement device) 546, an information processing unit 550, an image processing unit 548, an image recognition unit 554, Controls all arithmetic processes and information processing functions including complicated calculation functions as well as all mechanical movements of the monitor 556 and the like.

  The above-described embodiment is merely an example for explaining the principle and contents of the present invention, and is not intended to limit the principle and basic idea of the present invention. Those skilled in the art will be able to easily understand and conceive variations similar to this embodiment or variations thereof. Therefore, those skilled in the art will be able to understand and consider the meaning of the principle and the basic idea in a wide range through this embodiment.

  The thickness shape and refractive index distribution measuring device and method according to the present invention does not measure the thickness and refractive index at a selected point compared to existing equipment, but the thickness on a selected specific area. Measure the shape and refractive index distribution so that you can see at a glance. Therefore, it is more reliable than the measurement value at one point, and the thickness shape and refractive index distribution of the thin film layer are measured over the measurement area wider than the point, and the image of the selected measurement area on the sample surface is flat. It is much more accurate than point measurement to show in 3D. Furthermore, according to the invention, the light source and the narrow band pass light filter are thinner than when using visible light by using an ultraviolet light source or an infrared light source and using an appropriate variable light filter corresponding to the wavelength band of the light source. It is convenient and easy to measure the thickness and refractive index of a thin film layer. In addition, the measurement speed is fast by using the narrow band pass optical filter set. In addition, according to the present invention, since the thin film thickness shape, refractive index distribution, and the like are calculated for each pixel group, the resolution of the measurement value is high, and local changes can be measured. According to the present invention, in order to display the shape and refractive index distribution of the thin film layer in the selected measurement region on the monitor with a two-dimensional screen image, the user can display the thin film layer surface of the sample substrate and the refractive index state in two dimensions. Can observe and evaluate. Therefore, the user can make a strong and firm judgment with respect to the state and characteristics of the thin film layer. In other words, the two-dimensional shape is much more reliable than existing techniques that measure around a single point, and more information can be obtained.

It is a conventional reflection photometer principle diagram. FIG. 1B is a detailed view of a reflecting plate in which a fine hole used in the reflection photometer of FIG. It is a conceptual diagram explaining the multiple reflection phenomenon in the interface of a multilayer sample substrate. It is a typical reflectance graph. It is a structural diagram of an existing reflection photometer using glass fiber for daylighting. It is a figure which shows the measuring apparatus of the thin film thickness shape and refractive index distribution by this invention. It is a figure which shows the filter ring with which several optical filters were mounted | worn. It is the typical reflected light intensity graph obtained using the filter ring. It is a notional structure figure of a linear variable filter. It is a notional structure figure of a partial circular variable filter. It is a flowchart explaining the process of measuring thickness and refractive index using a filter ring.

Claims (2)

A sample substrate transfer unit including a sample substrate transfer table capable of supporting and transporting a sample substrate on which a thin film is formed;
An optical unit including a light source that provides incident light incident on the sample substrate supported by the sample substrate transfer table;
Filter means for selectively filtering incident light incident on the sample substrate or reflected light reflected from the sample substrate according to light wavelengths;
A two-dimensional photodetector having a specific light wavelength selected by the filter means and two-dimensionally detecting reflected light reflected from the sample substrate;
An image capturer for capturing video information included in the reflected light detected by the two-dimensional photodetector;
An image processing unit that extracts the intensity of the reflected light using video information included in the reflected light captured by the image capturing device, and obtains a reflectance graph from an absolute reflectance as a function of light wavelength for each pixel;
An information processing unit that calculates the thickness of the thin film, the thickness shape of the thin film, the refractive index of the thin film, and the refractive index distribution of the thin film as the characteristics of the thin film using the reflectance graph;
The thin film characteristic measuring apparatus according to claim 1, wherein the filter means can continuously change the light wavelength to be filtered. Positioning a specific region to be measured in a sample substrate on which a thin film is formed;
Projecting light onto the sample substrate;
Selectively filtering the light by light wavelength using a filter capable of continuously changing the light wavelength to be filtered;
Two-dimensionally detecting reflected light that is filtered by the specific wavelength band and reflected from the sample substrate;
Capturing video information contained in the detected reflected light;
Extracting the intensity of the reflected light from the video information and obtaining a reflectance graph from the absolute reflectance for each pixel as a function of the light wavelength;
A method for measuring thin film characteristics, comprising: calculating a thickness of the thin film, a thickness shape of the thin film, and a refractive index distribution of the thin film as the characteristics of the thin film using the reflectance graph.

Images