Film thickness measuring device and film thickness measuring method

Abstract

To measure the film thickness of an object with high precision. [Solution] The film thickness measuring device 1 comprises an optical measuring unit that irradiates a sample 100 with first light having a first wavelength, detects the first light reflected by the sample 100, and outputs a first signal related to the calculation of the wavelength centroid of the first light; a storage unit 33 that stores first relationship information showing the relationship between the film thickness of the sample 100 and the wavelength centroid at the first wavelength; and an analysis unit 32. The analysis unit 32 derives at least one of the following: the wavelength centroid corresponding to the film thickness candidate identified from the film thickness candidate and the first relationship information, which is the wavelength direction distance, and the difference between the calculated wavelength centroid of the first light and the film thickness candidate; and the film thickness direction distance, which is the difference between the film thickness derived from the calculated wavelength centroid of the first light and the first relationship information and the film thickness candidate. Based on the derivation results, the analysis unit 32 derives the film thickness of the sample 100.

Description

[Technical Field]

[0001] One aspect of this disclosure relates to a film thickness measuring device and a film thickness measuring method. [Background technology]

[0002] Patent Document 1 discloses a technique for determining the film thickness of an object from multiple candidates for film thickness, which are obtained from multiple wavelength centroids using multiple irradiation lights of different wavelengths and relational information showing the relationship between the wavelength centroids and film thickness values. [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] International Publication No. 2024 / 150501 [Overview of the project] [Problems that the invention aims to solve]

[0004] In this case, depending on the type of film being measured and the wavelength of the irradiated light, multiple film thickness candidates may be given from the value of the wavelength centroid. In this case, it may not be possible to uniquely estimate the film thickness, and the accuracy of the film thickness measurement may be reduced.

[0005] One aspect of this disclosure, made in view of the above circumstances, relates to a film thickness measuring device and a film thickness measuring method that can measure the film thickness of an object with high precision. [Means for solving the problem]

[0006] A film thickness measuring device according to one aspect of the present disclosure is [1] "a film thickness measuring device for measuring the film thickness of an object on which a film has been formed on a substrate, comprising: an optical measuring unit that irradiates the object with first light having a first wavelength, detects the first light reflected by the object, and outputs a first signal relating to the calculation of the wavelength parameter of the first light; a storage unit that stores first relational information indicating the relationship between the film thickness of the object and the wavelength parameter at the first wavelength; and an analysis unit that calculates the wavelength parameter of the first light based on the first signal, and derives the film thickness of the object based on the first relational information and the calculated wavelength parameter of the first light, wherein the analysis unit derives at least one of the following: a wavelength parameter corresponding to a film thickness candidate identified from a film thickness candidate and the first relational information, and the wavelength direction distance which is the difference between the calculated wavelength parameter of the first light and the film thickness candidate; and a film thickness direction distance which is the difference between the film thickness derived from the calculated wavelength parameter of the first light and the first relational information and the film thickness candidate, and derives the film thickness of the object based on the derivation result."

[0007] In a film thickness measuring device according to one aspect of this disclosure, the film thickness of an object is derived based on relational information showing the relationship between film thickness and wavelength parameters, which is stored in advance, and the wavelength parameter of a first light calculated from the detected light. Here, it is conceivable that the film thickness cannot be uniquely determined from the relational information and wavelength parameters (there are multiple candidate film thicknesses). In this regard, in a film thickness measuring device according to one aspect of this disclosure, at least one of the following is derived: the wavelength direction distance, which is the difference between the wavelength parameter corresponding to the candidate film thickness and the calculated wavelength parameter, and the film thickness direction distance, which is the difference between the film thickness derived from the calculated wavelength parameter and the relational information and the candidate film thickness. The film thickness of the object is then derived by considering the derived result. This makes it possible, for example, when a candidate film thickness cannot be uniquely determined, to consider the wavelength direction distance and the film thickness direction distance, and to determine that the candidate film thickness is closer to the actual film thickness of the object. This makes it possible to measure the film thickness of the object with high accuracy.

[0008] A film thickness measuring apparatus according to one aspect of the present disclosure may be [2] "a film thickness measuring apparatus according to [1], wherein the analysis unit derives a distance in the film thickness direction for each of a plurality of film thickness candidates, and estimates the film thickness candidate with the closest distance in the film thickness direction among the plurality of film thickness candidates as the film thickness of the object." With such a configuration, when there are multiple film thickness candidates, the film thickness candidate with the closest distance in the film thickness direction is adopted as the film thickness of the object, so that the film thickness of the object can be measured with high accuracy.

[0009] A film thickness measuring apparatus according to one aspect of the present disclosure may be [3] "a film thickness measuring apparatus according to [1] or [2], wherein the analysis unit derives a wavelength distance for each of a plurality of film thickness candidates, and estimates the film thickness candidate with the closest wavelength distance among the plurality of film thickness candidates as the film thickness of the object." With such a configuration, when there are multiple film thickness candidates, the film thickness candidate with the closest wavelength distance is adopted as the film thickness of the object, so that the film thickness of the object can be measured with high accuracy.

[0010] A film thickness measuring device according to one aspect of the present disclosure may be [4] "a film thickness measuring device according to any one of [1] to [3], comprising: a light irradiation unit that irradiates an object with first light having a first wavelength; an optical element whose transmittance and reflectance change according to wavelength in a predetermined wavelength range and separates the first light from the object by transmitting and reflecting it; and a light detection unit that detects at least one of the first light reflected by the optical element and the first light transmitted through the optical element and outputs a first signal relating to the calculation of the wavelength centroid, which is a wavelength parameter of the first light." With such a configuration, the wavelength centroid, which has a high correlation with film thickness, can be calculated as a wavelength parameter, and the film thickness of the object can be measured with high accuracy.

[0011] The film thickness measuring device according to one aspect of the present disclosure is as follows: "[5] The light irradiation unit irradiates the object with first light and second light having a second wavelength different from the first wavelength. The optical element separates the first light and the second light from the object by transmitting and reflecting them. The light detection unit detects at least one of the first light reflected by the optical element and the first light transmitted through the optical element, and outputs a first signal related to the calculation of the wavelength centroid, which is a wavelength parameter of the first light. At the same time, the light detection unit detects at least one of the second light reflected by the optical element and the second light transmitted through the optical element, and outputs a second signal related to the calculation of the wavelength centroid, which is a wavelength parameter of the second light. The storage unit stores the first relationship information and the second relationship information indicating the relationship between the film thickness of the object at the second wavelength and the wavelength parameter. The analysis unit derives at least one of the wavelength direction distance, which is the difference between the wavelength parameter corresponding to the film thickness candidate specified from the film thickness candidate and the first relationship information and the calculated wavelength parameter of the first light, and the film thickness direction distance, which is the difference between the film thickness derived from the calculated wavelength parameter of the first light and the first relationship information and the film thickness candidate. The analysis unit also derives at least one of the wavelength direction distance, which is the difference between the wavelength parameter corresponding to the film thickness candidate specified from the film thickness candidate and the second relationship information and the calculated wavelength parameter of the second light, and the film thickness direction distance, which is the difference between the film thickness derived from the calculated wavelength parameter of the second light and the second relationship information and the film thickness candidate. Based on the derivation result, the film thickness of the object is derived." It may be the film thickness measuring device according to any one of [1] to [4]. By irradiating the object with light of a plurality of wavelengths as the irradiation light, for example, when determining which film thickness candidate to adopt among a plurality of film thickness candidates, the amount of information to be considered can be increased, and the film thickness of the object can be measured with high accuracy.

[0012] The film thickness measuring device according to one aspect of the present disclosure may be the film thickness measuring device described in [6] "The analysis unit sets, as a film thickness candidate, the film thickness derived from the calculated wavelength parameter of the first light and the first relationship information, or the film thickness derived from the calculated wavelength parameter of the second light and the second relationship information." Thus, by setting the film thickness derived from the wavelength parameter and the relationship information (for example, the film thickness at the intersection of the calibration curve indicating the relationship information and the wavelength parameter) as the film thickness candidate, an appropriate film thickness candidate can be set easily and quickly.

[0013] The film thickness measuring device according to one aspect of the present disclosure may be the film thickness measuring device described in [7] "The analysis unit sets, as a film thickness candidate, the film thickness within a predetermined range, the film thickness measuring device according to any one of [1] to [6]." Usually, since the approximate film thickness of the object is known in advance, by setting a range such as ±100 nm of the assumed film thickness as the film thickness candidate, an appropriate film thickness candidate can be set easily and quickly. Although it is conceivable that a film thickness that is not the true film thickness value may have a high evaluation value based on the wavelength direction distance or the like and be selected as the film thickness of the object, by setting only the film thickness within the predetermined range as the film thickness candidate, such an incorrect selection is avoided and the film thickness can be measured with high accuracy.

[0014] The film thickness measuring device according to one aspect of the present disclosure may be the film thickness measuring device described in [8] "The analysis unit sets a plurality of film thickness candidates at a predetermined interval, the film thickness measuring device according to any one of [1] to [7]." By setting the film thickness candidates evenly at a predetermined interval, the film thickness can be measured with high accuracy.

[0015] A film thickness measuring apparatus according to one aspect of the present disclosure may be [9] "a film thickness measuring apparatus according to [2] or [3], wherein the analysis unit evaluates the slope of the calibration curve at the intersection of the calibration curve showing first relational information and the film thickness candidate for each of a plurality of film thickness candidates, and derives the film thickness of the target object such that film thickness candidates with larger slopes are more likely to be adopted as the film thickness of the target object." In areas where the slope of the calibration curve is small (near the extreme value), the change in film thickness with respect to changes in wavelength parameters becomes gradual, and there is a risk that the film thickness of the target object cannot be derived with high accuracy from relational information. In this regard, the accuracy of film thickness derivation can be ensured by making film thickness candidates with larger slopes more likely to be adopted.

[0016] A method for measuring film thickness according to one aspect of the present disclosure is

[10] "a method for measuring film thickness performed by a film thickness measuring device for measuring the film thickness of an object on which a film has been formed on a substrate, comprising: an optical measurement step of irradiating the object with first light having a first wavelength, detecting the first light reflected by the object, and outputting a first signal relating to the calculation of the wavelength parameter of the first light; and a derivation step of calculating the wavelength parameter of the first light based on the first signal, deriving the film thickness of the object based on first relational information showing the relationship between the film thickness of the object and the wavelength parameter at the first wavelength, wherein the derivation step derives at least one of the following: a wavelength parameter corresponding to a film thickness candidate identified from a film thickness candidate and the first relational information, and the wavelength distance in the direction of the calculated first light, and the film thickness distance in the direction of the calculated first light, which is the difference between the film thickness derived from the calculated wavelength parameter of the first light and the first relational information and the film thickness candidate, and deriving the film thickness of the object based on the derivation result."

[0017] A method for measuring film thickness according to one aspect of the present disclosure may be

[11] "the method for measuring film thickness according to

[10] , wherein in the derivation step, the wavelength distance is derived for each of a plurality of film thickness candidates, and the film thickness candidate with the closest wavelength distance among the plurality of film thickness candidates is estimated to be the film thickness of the object."

[0018] A method for measuring film thickness according to one aspect of the present disclosure may be

[12] "a method for measuring film thickness according to

[10] or

[11] , wherein in the derivation step, the distance in the film thickness direction is derived for each of a plurality of film thickness candidates, and the film thickness candidate with the closest distance in the film thickness direction among the plurality of film thickness candidates is estimated to be the film thickness of the object."

[0019] A method for measuring film thickness according to one aspect of the present disclosure may be

[13] "a method for measuring film thickness according to any one of

[10] to

[12] , wherein in the derivation step, the film thickness derived from the calculated first light wavelength parameter and the first relationship information is set as a candidate film thickness."

[0020] A method for measuring film thickness according to one aspect of the present disclosure may be

[14] "a method for measuring film thickness according to any one of

[10] to

[13] , wherein in the derivation step, a film thickness within a predetermined range is set as a candidate film thickness."

[0021] A method for measuring film thickness according to one aspect of the present disclosure may be

[15] "a method for measuring film thickness according to any one of

[10] to

[14] , wherein in the derivation step, a plurality of film thickness candidates are set at predetermined intervals."

[0022] A method for measuring film thickness according to one aspect of the present disclosure may be

[16] "a method for measuring film thickness according to any one of

[10] to

[15] , wherein in the derivation step, for each of a plurality of film thickness candidates, the slope of the calibration curve at the intersection of the calibration curve showing first relational information and the film thickness candidate is evaluated, and the film thickness of the object is derived such that the film thickness candidate with a larger slope is more likely to be adopted as the film thickness of the object." [Effects of the Invention]

[0023] According to one aspect of this disclosure, the film thickness of an object can be measured with high precision. [Brief explanation of the drawing]

[0024] [Figure 1] Figure 1 is a schematic diagram showing the film thickness measuring device according to this embodiment. [Figure 2]Figure 2 is a schematic diagram showing the configuration when a multicolor LED array is used as the light irradiation unit. [Figure 3] Figure 3 is a schematic diagram showing the configuration when a white light source and a filter switching mechanism are used as the light irradiation unit. [Figure 4] Figure 4 illustrates the relationship between the characteristics of a dichroic mirror and the wavelength of light emitted from a light source. [Figure 5] Figure 5 illustrates the properties of the light spectrum and the tilted dichroic mirror. [Figure 6] Figure 6 illustrates the wavelength shift corresponding to the amount of transmitted and reflected light. [Figure 7] Figure 7 shows the reflectance according to wavelength for each film thickness. [Figure 8] Figure 8 shows a waveform illustrating the relationship between the sample film thickness and the wavelength centroid. [Figure 9] Figure 9 shows waveforms illustrating the relationship between sample film thickness and wavelength centroid for each of several different wavelengths. [Figure 10] Figure 10(a) shows the system spectral characteristics according to wavelength for each combination of monochromatic light in the configuration shown in Figure 2. Figure 10(b) shows the relationship between the wavelength centroid and film thickness in the configuration shown in Figure 2. [Figure 11] Figure 11(a) shows the system spectral characteristics for each set filter according to wavelength in the configuration shown in Figure 3. Figure 11(b) shows the relationship between the wavelength centroid and film thickness in the configuration shown in Figure 3. [Figure 12] Figure 12 is a flowchart showing the specific steps involved in measuring film thickness. [Figure 13] Figures 13(a) to (c) illustrate the film thickness search block settings for three different wavelength patterns. [Figure 14] Figure 14 is a diagram illustrating the derivation of candidate film thicknesses. [Figure 15] Figure 15 illustrates the derivation of candidate film thicknesses according to the predicted film thickness range. [Figure 16] Figure 16 illustrates the derivation of candidate film thicknesses for each wavelength pattern. [Figure 17] Figure 17 shows the derived candidate film thicknesses. [Figure 18] Figure 18 illustrates the derivation of film thickness based on the selection of the optimal combination of film thickness candidates. [Figure 19] Figure 19(a) illustrates the derivation of film thickness in a sample with significant film thickness variation. Figure 19(b) shows an example of ground truth data for film thickness. Figure 19(c) shows the multi-wavelength analysis results corresponding to the ground truth data shown in Figure 19(b). [Figure 20] Figure 20(a) shows the relationship between the wavelength centroid and film thickness when light with a relatively narrow wavelength range is used. Figure 20(b) shows the relationship between the wavelength centroid and film thickness when light with a relatively wide wavelength range is used. Figure 20(c) shows the relationship between the wavelength centroid and film thickness when light with a relatively wide wavelength range and multiple peaks (two maximum values) is used. [Figure 21] Figure 21(a) shows the film thickness mapping results when red and blue wavelength light is used. Figure 21(b) shows the film thickness mapping results when green and blue wavelength light is used. Figure 21(c) shows the film thickness mapping results when red and green wavelength light is used. Figure 21(d) shows the ground truth data for the film thickness mapping results. [Figure 22] Figures 22(a) to 22(c) illustrate examples where deriving the film thickness is difficult using the standard deviation between candidate film thicknesses. [Figure 23] Figure 23 is a diagram illustrating the derivation of film thickness based on evaluation values. [Figure 24] Figure 24(a) is a diagram illustrating the overview of distance evaluation, and Figure 24(b) is a diagram illustrating the overview of calibration curve accuracy evaluation. [Figure 25] Figures 25(a) and 25(b) illustrate examples of setting candidate film thicknesses. [Figure 26] Figures 26(a) to 26(e) illustrate specific examples of distance evaluation. [Figure 27] Figure 27 illustrates an example of an evaluation element. [Figure 28] Figure 28 illustrates the evaluation of distance in the film thickness direction. [Figure 29] Figure 29 illustrates the evaluation of the wavelength shift amount of the calibration curve. [Figure 30] Figure 30 is a diagram illustrating noise evaluation. [Figure 31] Figure 31 is a diagram illustrating the evaluation of peripheral film thickness. [Figure 32] Figure 32 is a diagram illustrating the input for each evaluation value. [Figure 33] Figure 33 is a flowchart showing the process for measuring film thickness. [Figure 34] Figure 34 is a flowchart showing the process for measuring film thickness. [Figure 35] Figure 35 is a flowchart showing the process for measuring film thickness. [Figure 36] Figure 36 is a table showing the experimental results (average error). [Figure 37] Figures 37(a) to 37(f) show the difference between the estimated results and the ground truth data for a sample with a film thickness of 100 nm. [Figure 38] Figures 38(a) to 38(f) show the difference between the estimated results and the ground truth data for a sample with a film thickness of 800 nm. [Modes for carrying out the invention]

[0025] Figure 1 is a schematic diagram of the film thickness measuring device 1 according to this embodiment. The film thickness measuring device 1 is a device that irradiates light onto a sample 100 (object) in a planar manner and measures the thickness of a film formed on the sample 100 based on the reflected light from the sample 100. The sample 100 may be a light-emitting element such as an LED, mini-LED, μLED, SLD element, laser element, vertical-cavity-cell-scale (VCSEL), or OLED, or a light-emitting element whose emission wavelength is adjusted by a fluorescent material containing nanodots, etc. The sample 100 may also be an optical film, a thin film for a display panel, or a thin film for a semiconductor. Furthermore, the sample 100 may be a wafer of an integrated circuit (IC) having a PN junction such as a transistor, or a logic device, memory device, analog device, mixed-signal device that combines them, or a power semiconductor device (power device) such as a high-current / high-voltage MOS transistor, bipolar transistor, or IGBT, or a light-emitting device such as an LED or semiconductor laser. Sample 100 is an object in which a film 100b is formed on the surface of a substrate 100a. In this embodiment, it is explained that in Sample 100, only one layer of film 100b is formed on the surface of the substrate 100a. Film 100b is, for example, an oxide film or a nitride film, but it may be any other film.

[0026] As shown in Figure 1, the film thickness measuring device 1 comprises a light irradiation unit 10, a half mirror 11, a field lens 12, a camera system 20, and a control device 30 (analysis unit 32, storage unit 33). Of the above components, the light irradiation unit 10, the half mirror 11, the field lens 12, and the camera system 20 constitute the light measurement unit. The light measurement unit irradiates the sample 100 with first light having a first wavelength, detects the first light reflected by the sample 100, and outputs a first signal related to the calculation of the wavelength parameter (e.g., wavelength centroid) of the first light (details will be described later).

[0027] The light irradiation unit 10 irradiates the sample 100 with light in a planar manner. For example, the light irradiation unit 10 irradiates light in a planar manner over substantially the entire surface of the sample 100. The light irradiation unit 10 may have a light source capable of uniformly irradiating the surface of the sample 100, and may irradiate the sample 100 with diffused light.

[0028] The light irradiation unit 10 irradiates the sample 100 with a first light having a first wavelength and a second light having a second wavelength different from the first wavelength. Furthermore, the light irradiation unit 10 may irradiate the sample 100 with a third light having a third wavelength different from the first and second wavelengths. Note that the first wavelength, second wavelength, and third wavelength may each have a wavelength range (wavelength width). The term "different" for each wavelength may include cases where a portion of their respective wavelength ranges overlap. That is, "a second wavelength different from the first wavelength" includes not only wavelengths with completely different wavelength ranges from the first wavelength, but also wavelengths with partially overlapping wavelength ranges from the first wavelength, but whose wavelength ranges do not completely match. Similarly, "a third wavelength different from the first and second wavelengths" includes not only wavelengths with completely different wavelength ranges from the first and second wavelengths, but also wavelengths with partially overlapping wavelength ranges from the first or second wavelength, but whose wavelength ranges do not completely match. In the following description, the light irradiation unit 10 will be described as irradiating the sample 100 with a first light, a second light, and a third light. The light emitted from the light irradiation unit 10 passes through the half mirror 11 and the field lens 12 and is irradiated onto the sample 100 in a planar manner.

[0029] A detailed example of the configuration of the light irradiation unit 10 will be explained with reference to Figures 2 and 3. Figure 2 is a schematic diagram showing the configuration when a multicolor LED (Light Emitting Diode) array light source 10a is used as the configuration of the light irradiation unit 10. The light irradiation unit 10 has a multicolor LED array light source 10a, which is a light source capable of emitting monochromatic light of three or more wavelengths. Here, it is explained that the light irradiation unit 10 has a light source capable of emitting monochromatic light of three or more wavelengths, but the light irradiation unit 10 only needs to have a light source capable of emitting monochromatic light of at least one wavelength. Monochromatic light here means light with a narrow wavelength width. Light with a narrow wavelength width is, for example, light with a spectral full width at half maximum of less than 100 nm. Conversely, light with a wide wavelength width is, for example, light with a spectral full width at half maximum of 100 nm or more. In such a light irradiation unit 10, the sample 100 may be irradiated with a first light, which is formed by the multicolor LED array light source 10a simultaneously emitting monochromatic light of two wavelengths from among the three or more wavelengths of monochromatic light described above. In this case, the multicolor LED array light source 10a irradiates the sample 100 with a second light, which is formed by the multicolor LED array light source 10a simultaneously emitting monochromatic light of two wavelengths with a different combination than the first light. Furthermore, the multicolor LED array light source 10a irradiates the sample 100 with a third light, which is formed by the multicolor LED array light source 10a simultaneously emitting monochromatic light of two wavelengths with a different combination than the first and second light. The light irradiation unit 10 may also irradiate the sample 100 with four or more lights of different wavelengths. In addition, the number of wavelengths of light irradiated onto the sample 100 by the light irradiation unit 10 may be set according to the estimated film thickness of the sample 100, which is estimated from the design information of the sample 100.

[0030] That is, for example, if the multicolor LED array light source 10a is capable of emitting monochromatic light of red wavelength, monochromatic light of green wavelength, and monochromatic light of blue wavelength, the two monochromatic lights of the two wavelengths may be monochromatic light of red wavelength and monochromatic light of blue wavelength, or monochromatic light of red wavelength and monochromatic light of green wavelength, or monochromatic light of green wavelength and monochromatic light of blue wavelength. In this case, the first light may be light composed of monochromatic light of red wavelength and monochromatic light of blue wavelength emitted simultaneously. The second light may be light composed of monochromatic light of red wavelength and monochromatic light of green wavelength emitted simultaneously. The third light may be light composed of monochromatic light of green wavelength and monochromatic light of blue wavelength emitted simultaneously. Note that the first light, the second light, and the third light do not necessarily have to be light composed of monochromatic light of two wavelengths emitted simultaneously. Red wavelength monochromatic light is light containing wavelengths between 610 nm and 780 nm, green wavelength monochromatic light is light containing wavelengths between 500 nm and 570 nm, and blue wavelength monochromatic light is light containing wavelengths between 430 nm and 490 nm.

[0031] As described above, the multicolor LED array light source 10a can generate light with various wavelength patterns depending on the combination of LEDs. Furthermore, because the multicolor LED array light source 10a is composed of LEDs, it has a long lifespan. In addition, since the multicolor LED array light source 10a does not have a configuration in which the wavelength is set by a filter, a filter switching mechanism (movable part) for switching filters is unnecessary. This allows for quick wavelength switching with a simple configuration. Note that the light irradiation unit 10 may also use a tunable wavelength light source (not shown) with a configuration that does not have a filter switching mechanism, similar to the multicolor LED array light source 10a.

[0032] Figure 3 is a schematic diagram showing the configuration of the light irradiation unit 10 when a white light source 10b and a filter switching mechanism 10c are used. In the light irradiation unit 10 shown in Figure 3, light emitted from the white light source 10b is irradiated onto the sample 100 after passing through a filter set by the filter switching mechanism 10c. In this case, the light irradiation unit 10 irradiates the sample 100 with the first light, second light, and third light described above by switching the filter with the filter switching mechanism 10c.

[0033] The light irradiation unit 10 irradiates the sample 100 with light of a wavelength included in a predetermined wavelength range of the tilted dichroic mirror 22 (details to be described later) of the camera system 20. As will be described in detail later, the tilted dichroic mirror 22 is an optical element that separates light from the sample 100 by transmitting and reflecting it according to its wavelength. The transmittance and reflectance of the tilted dichroic mirror 22 change according to the wavelength in the predetermined wavelength range mentioned above.

[0034] Figure 4 illustrates the relationship between the characteristics of the tilted dichroic mirror 22 and the wavelength of light emitted from the light irradiation unit 10. In Figure 4, the horizontal axis represents wavelength, and the vertical axis represents the transmittance of the tilted dichroic mirror 22. As shown in the characteristics X4 of the tilted dichroic mirror 22 in Figure 4, in the tilted dichroic mirror 22, the transmittance (and reflectance) of light changes gradually in accordance with the change in wavelength in a predetermined wavelength range X10, while in wavelength ranges other than this specific wavelength range, the transmittance (and reflectance) of light remains constant regardless of the change in wavelength. As shown in Figure 4, the light X20 output from the light irradiation unit 10 includes light with wavelengths included in the predetermined wavelength range X10 described above. That is, the light irradiation unit 10 outputs light with a broad spectrum that includes the predetermined wavelength range X10. The wavelength range (interference peak wavelength) for measurement is determined by the material of the film formed on the sample 100 and the measurement film thickness range.

[0035] Returning to Figure 1, the half-mirror 11 is a mirror that reflects the light emitted from the light irradiation unit 10 towards the sample 100 (more specifically, towards the field lens 12 that guides the light to the sample 100) and transmits the light from the irradiated sample 100 (more specifically, the light that has passed from the sample 100 through the field lens 12). The field lens 12 is a lens that aligns the direction of light propagation.

[0036] The camera system 20 comprises a lens 21, an inclined dichroic mirror 22 (optical element), an area sensor 23 (light detection unit), and an area sensor 24 (light detection unit). The camera system 20 may also include a linear image sensor (detection unit) instead of the area sensors.

[0037] Lens 21 is a lens that focuses light from sample 100 that has been incident via field lens 12 and half mirror 11. Lens 21 may be positioned upstream of the inclined dichroic mirror 22, or it may be positioned in the region between the inclined dichroic mirror 22 and area sensors 23, 24. In this embodiment, it is described that lens 21 is positioned upstream of the inclined dichroic mirror 22. Lens 21 may be a finite focus lens or an infinite focus lens. If lens 21 is a finite focus lens, the distance from lens 21 to area sensors 23, 24 is set to a predetermined value. If lens 21 is an infinite focus lens, lens 21 is a collimator lens that converts light from sample 100 into parallel light, and is aberration corrected so that parallel light is obtained. The light output from lens 21 is incident on the inclined dichroic mirror 22.

[0038] The tilted dichroic mirror 22 is a mirror made using a special optical material and is an optical element that separates light from sample 100 by transmitting and reflecting it according to its wavelength. In other words, the tilted dichroic mirror 22 separates the first light, the second light, and the third light from sample 100 by transmitting and reflecting them, respectively. The tilted dichroic mirror 22 is configured such that the transmittance and reflectance of light change according to the wavelength in a predetermined wavelength range.

[0039] Figure 5 illustrates the characteristics of the light spectrum and the tilted dichroic mirror 22. In Figure 5, the horizontal axis represents wavelength, and the vertical axis represents spectral intensity (in the case of the light spectrum) and transmittance (in the case of the tilted dichroic mirror 22). As shown in the characteristic X4 of the tilted dichroic mirror 22 in Figure 5, in the tilted dichroic mirror 22, the transmittance (and reflectance) of light changes gradually in accordance with the change in wavelength in a predetermined wavelength range (wavelength range λ1 to λ2). On the other hand, in wavelength ranges other than the predetermined wavelength range (i.e., wavelengths lower than λ1 and wavelengths higher than λ2), the transmittance (and reflectance) of light may be constant regardless of the change in wavelength. In other words, in a specific wavelength range (wavelength range λ1 to λ2), the transmittance of light increases monotonically (reflectance decreases monotonically) in accordance with the change in wavelength. Transmittance and reflectance have a negative correlation, meaning that when one increases, the other decreases. Therefore, in the following, "transmittance (and reflectance)" may not be used, and only "transmittance" will be used. Note that "the transmittance of light is constant regardless of the change in wavelength" includes not only cases where it is perfectly constant, but also cases where, for example, the change in transmittance for a change of 1 nm in wavelength is 0.1% or less. At wavelengths lower than λ1, the transmittance of light may be approximately 0% regardless of the change in wavelength, and at wavelengths higher than λ2, the transmittance of light may be approximately 100% regardless of the change in wavelength. Note that "the transmittance of light is approximately 0%" includes transmittances of approximately 0% + 10%, and "the transmittance of light is approximately 100%" includes transmittances of approximately 100% - 10%. In Figure 3, waveform X1 shows the waveform of light output from the light irradiation unit 10. As shown in waveform X1 in Figure 3, the light output from the light irradiation unit 10 includes light with wavelengths included in a predetermined wavelength range (wavelength range λ1 to λ2) of the tilted dichroic mirror 22.

[0040] Returning to Figure 1, area sensors 23 and 24 image the light from sample 100. Area sensors 23 and 24 image (detect) the light separated by the tilted dichroic mirror 22. Area sensor 23 images (detects) the light transmitted through the tilted dichroic mirror 22 and outputs a signal according to the detection result. That is, area sensor 23 detects the first light transmitted through the tilted dichroic mirror 22, the second light transmitted through the tilted dichroic mirror 22, and the third light transmitted through the tilted dichroic mirror 22. Area sensor 24 images (detects) the light reflected by the tilted dichroic mirror 22 and outputs a signal according to the detection result. That is, area sensor 24 detects the first light reflected by the tilted dichroic mirror 22, the second light reflected by the tilted dichroic mirror 22, and the third light reflected by the tilted dichroic mirror 22. Regarding the detection of the first light, the signals output from area sensors 23 and 24 are the first signals related to the calculation of the wavelength centroid (wavelength parameter) of the first light. Regarding the detection of the second light, the signals output from area sensors 23 and 24 are the second signals related to the calculation of the wavelength centroid (wavelength parameter) of the second light. Regarding the detection of the third light, the signals output from area sensors 23 and 24 are the third signals related to the calculation of the wavelength centroid (wavelength parameter) of the third light. The wavelength range to which area sensors 23 and 24 are sensitive corresponds to a predetermined wavelength range in which the transmittance (and reflectance) of light changes in accordance with the change in wavelength in the tilted dichroic mirror 22. Area sensors 23 and 24 are, for example, monochrome sensors or color sensors. The imaging results (images) from area sensors 23 and 24 are output to the control device 30 by the first signal, the second signal, and the third signal described above.

[0041] A bandpass filter (not shown) may be placed upstream of the area sensors 23 and 24. Such a bandpass filter (not shown) may be, for example, a filter that removes light in wavelengths other than the predetermined wavelength range described above (the wavelength range in the tilted dichroic mirror 22 where the transmittance and reflectance of light change depending on the wavelength).

[0042] The control device 30 is a computer, and physically comprises memory such as RAM and ROM, a processor (arithmetic circuit) such as a CPU, a communication interface, and a storage unit such as a hard disk. The control device 30 functions by executing a program stored in memory using the CPU of the computer system. The control device 30 may be composed of a microcontroller or an FPGA.

[0043] The control device 30 derives the film thickness of sample 100 based on the first signal, second signal, and third signal, which are signals from area sensors 23 and 24 that image the first, second, and third lights, respectively. As processing related to the derivation of film thickness, the control device 30 performs wavelength centroid calculation processing based on the signals from area sensors 23 and 24, and film thickness derivation processing based on the wavelength centroid, etc. Furthermore, as a prerequisite for performing the film thickness derivation processing, the control device 30 stores relational information showing the relationship between the film thickness of sample 100 and the wavelength centroid (wavelength parameter). The control device 30 has an analysis unit 32 and a storage unit 33 as functional configurations that realize the above-described processing and storage. The analysis unit 32 has the function of performing wavelength centroid calculation processing and film thickness derivation processing. The storage unit 33 stores the above-described relational information. The storage unit 33 stores relational information for each of the different wavelengths. In other words, the memory unit 33 stores first relational information showing the relationship between the film thickness of sample 100 and the wavelength centroid at a first wavelength, second relational information showing the relationship between the film thickness of sample 100 and the wavelength centroid at a second wavelength, and third relational information showing the relationship between the film thickness of sample 100 and the wavelength centroid at a third wavelength. Below, the wavelength centroid calculation process will be described first, followed by the relationship information between film thickness and the wavelength centroid, and finally the film thickness derivation process.

[0044] (Wavelength centroid calculation process) The analysis unit 32 calculates the wavelength centroid for sample 100 based on the signals from the area sensors 23 and 24 that detected the light. Specifically, the analysis unit 32 calculates the first wavelength centroid based on the first signal, which is the signal from the area sensors 23 and 24 that detected the first light, and calculates the second wavelength centroid based on the second signal, which is the signal from the area sensors 23 and 24 that detected the second light. Furthermore, the analysis unit 32 calculates the third wavelength centroid based on the third signal, which is the signal from the area sensors 23 and 24 that detected the third light.

[0045] The analysis unit 32 may calculate the wavelength centroid of light for each pixel as the wavelength centroid, based on the transmitted light amount (intensity of light transmitted through the tilted dichroic mirror 22) identified based on the signal from the area sensor 23 indicating the detection result of the area sensor 23, and the reflected light amount (intensity of light reflected from the tilted dichroic mirror 22) identified based on the signal from the area sensor 24 indicating the detection result of the area sensor 24. Specifically, the analysis unit 32 calculates the wavelength centroid at each pixel based on the following equation (1). In the following equation (1), x' is the wavelength shift amount with respect to λ0, and I' T I' is the amount of transmitted light. R This indicates the amount of reflected light. x'=(I' T -I' R ) / 2(I' T +I' R ) (1)

[0046] Furthermore, the analysis unit 32 may calculate the wavelength centroid of light for each pixel by considering the center wavelength of the tilted dichroic mirror 22 (the center wavelength of a predetermined wavelength range) and the width of the tilted dichroic mirror 22. The width of the tilted dichroic mirror 22 is, for example, the wavelength range from the wavelength at which the transmittance in the tilted dichroic mirror 22 is 0% to the wavelength at which the transmittance is 100%. In this case, the analysis unit 32 may calculate the wavelength centroid of each pixel based on the following equation (2). In the following equation (2), x'' is the wavelength centroid, and I'' T I' is the amount of transmitted light. Ris the amount of reflected light, λ0 is the central wavelength of the tilted dichroic mirror 22, and A is the width of the tilted dichroic mirror 22 (that is, the difference between the upper limit value λ2 and the lower limit value λ1 of the wavelength width). x´´ = λ0 + A(I´ T - I´ R ) / 2(I´ T + I´ R ) (2)

[0047] FIG. 6 is a diagram for explaining the wavelength shift according to the amount of transmitted light and the amount of reflected light. When calculating x´´ (wavelength centroid) by the above-mentioned formula (1) or (2), as shown in FIG. 6, for a pixel where I´ T (amount of transmitted light) = I´ R (amount of reflected light), x´´ = λ0 (central wavelength of the tilted dichroic mirror 22). Also, for a pixel where I´ T < I´ R , that is, a pixel where the amount of reflected light is more than the amount of transmitted light, x´´ = λ1 (a wavelength on the shorter wavelength side than λ0). Also, for a pixel where I´ T > I´ R , that is, a pixel where the amount of transmitted light is more than the amount of reflected light, x´´ = λ2 (a wavelength on the longer wavelength side than λ0). Thus, the value of x´´ (wavelength centroid) shifts (wavelength shift) based on the amount of transmitted light and the amount of reflected light.

[0048] And since the wavelength centroid has a correlation with the film thickness, it can be used for deriving the film thickness. FIG. 7 is a diagram showing the reflectance according to the wavelength for each film thickness. In FIG. 7, the horizontal axis is the wavelength and the vertical axis is the reflectance. In the example shown in FIG. 7, the relationship between the wavelength and the reflectance is shown for each of the examples of film thicknesses of 820 nm, 830 nm, and 840 nm. As shown in FIG. 7, due to the difference in film thickness, the wavelength centroid will be different. Thus, since there is a correlation between the wavelength centroid and the film thickness, by specifying the wavelength centroid, it becomes possible to estimate the film thickness. Specific film thickness derivation processing will be described later.

[0049] (Relationship information between film thickness and wavelength centroid) Returning to Figure 1, the memory unit 33 stores information relating the film thickness of sample 100 to the wavelength centroid for each wavelength of light emitted from the light irradiation unit 10. As mentioned above, there is a correlation between film thickness and the wavelength centroid. Therefore, by having information relating the film thickness to the wavelength centroid prepared in advance, the film thickness can be derived from this relationship information and the actually measured wavelength centroid. The memory unit 33 stores this relationship information for each type of film. That is, the memory unit 33 stores information relating the film thickness of sample 100 to the wavelength centroid for each combination of the wavelength of light emitted from the light irradiation unit 10 and the type of film.

[0050] The relevant information may be derived based on the theoretical reflectance corresponding to the type of film and the spectral characteristics (spectral sensitivity) of the entire film thickness measuring device 1. The theoretical reflectance can be determined for each wavelength once the type of film (refractive index and extinction coefficient of the film) and film thickness are determined. The spectral characteristics of the entire film thickness measuring device 1 are determined for each wavelength of light emitted from the light irradiation unit 10. The spectral characteristics of the entire film thickness measuring device 1 can be determined in advance by various methods (estimated). In addition, the relevant information may be derived from measured values ​​by conducting measurements on a reference sample in advance to understand the relationship between film thickness and wavelength centroid. In the following explanation, the relevant information will be described assuming that it is derived based on the theoretical reflectance corresponding to the type of film and the spectral characteristics (spectral sensitivity) of the entire film thickness measuring device 1.

[0051] Figure 8 shows a waveform illustrating the relationship between film thickness and wavelength centroid for sample 100. Let's consider an example of estimating film thickness from the wavelength centroid using this relationship information. In Figure 8, the horizontal axis represents film thickness d, and the vertical axis represents the wavelength centroid x''. Now, suppose the wavelength centroid is the value shown by the dashed line extending horizontally in Figure 8. In this case, as shown in Figure 8, there are multiple candidate film thicknesses (10 candidate patterns in the example shown in Figure 8) that correspond to this wavelength centroid. Therefore, it is not possible to uniquely identify the film thickness solely from the relationship information shown in Figure 8.

[0052] Furthermore, as shown in Figure 8, the waveform representing the relevant information has multiple extreme values ​​(maximum and minimum values) indicated by dashed circles. At these extreme values, the slope of the waveform curve becomes gentler. In areas where the curve slope is gentler, differences in film thickness due to differences in the wavelength centroid become less apparent, and it may be difficult to accurately identify the film thickness corresponding to the wavelength centroid.

[0053] Thus, if one attempts to derive the film thickness from only the relationship information between the film thickness of sample 100 and the wavelength centroid for a single wavelength of light emitted from the light irradiation unit 10, it may not be possible to accurately derive the film thickness. For this reason, the storage unit 33 stores relationship information between the film thickness of sample 100 and the wavelength centroid for multiple types of wavelengths (wavelengths of light emitted from the light irradiation unit 10).

[0054] Figure 9 shows waveforms illustrating the relationship between the film thickness of sample 100 and the wavelength centroid for each of several wavelengths (wavelengths of light emitted from the light irradiation unit 10). In the example shown in Figure 9, waveforms illustrating the relationship between the film thickness of sample 100 and the wavelength centroid are shown for each of three wavelengths, namely the first light, second light, and third light emitted from the light irradiation unit 10. The wavelength range for the first light is 400-700 nm. The wavelength range for the second light is 610-700 nm. The wavelength range for the third light is 400-550 nm.

[0055] As shown in Figure 9, in waveforms showing relational information for each of several wavelengths, the regions of extreme values ​​(maximum and minimum values) do not coincide, so the film thickness can be appropriately derived based on any of the waveforms. In the example shown in Figure 9, for region 301, at least the first light waveform 201 is far from the extreme value and the slope of the curve is relatively large, so the film thickness can be appropriately derived based on the first light waveform 201; for region 302, at least the second light waveform 202 is far from the extreme value and the slope of the curve is relatively large, so the film thickness can be appropriately derived based on the second light waveform 202; and for region 303, at least the third light waveform 203 is far from the extreme value and the slope of the curve is relatively large, so the film thickness can be appropriately derived based on the third light waveform 203. Furthermore, when film thickness candidates are derived from waveforms showing relational information for each of several wavelengths, the consistency of the film thickness candidates for each of the multiple wavelengths is taken into consideration, allowing for a unique determination of the film thickness (details will be described later).

[0056] Examples of waveforms showing the above-mentioned relationship information stored in the memory unit 33 will be explained with reference to Figures 10 and 11. As described above, the relationship information between film thickness and wavelength centroid is derived based on the theoretical reflectance according to the type of film and the spectral characteristics (spectral sensitivity) of the entire film thickness measuring device 1. Figure 10(a) is a diagram showing the system spectral characteristics according to wavelength for each combination of monochromatic light in the configuration shown in Figure 2 (configuration using a multicolor LED array light source 10a). In Figure 10(a), the horizontal axis represents wavelength and the vertical axis represents spectral characteristics. As shown in Figure 10(a), spectral characteristics for the transmission side and reflection side are derived for each combination of monochromatic light. In detail, the derivation of the relationship information is performed by deriving the expected value of the wavelength centroid from the theoretical reflectance and the spectral characteristics of the film thickness measuring device 1, and then deriving the relationship equation by plotting the expected value of the wavelength centroid and performing curve fitting. Examples of curve fitting methods include polynomial approximation and other curve fitting methods. In addition to curve fitting, methods such as interpolation may also be used to derive a curve showing the relationship between wavelength and spectral intensity.

[0057] When the type of film is determined, the theoretical reflectance values ​​for each wavelength are determined according to the film thickness. Suppose a certain film thickness is specified and the theoretical reflectance values ​​for each wavelength are determined. In this case, the expected value of the transmitted light amount measured by the area sensor 23 can be estimated based on the theoretical reflectance for each wavelength and the spectral characteristics of the transmission side of the film thickness measuring device 1 for each wavelength. Since the area sensor 23 does not have a spectral function, the expected value of the transmitted light amount measured by the area sensor 23 is the value obtained by integrating the light intensity for each wavelength. Similarly, the expected value of the reflected light amount measured by the area sensor 24 can be estimated based on the theoretical reflectance for each wavelength and the spectral characteristics of the reflection side of the film thickness measuring device 1 for each wavelength. Then, from equation (1) or (2) above, the expected value of the wavelength centroid can be derived from the expected value of the transmitted light amount and the expected value of the reflected light amount. In this way, the expected value of the wavelength centroid at a certain film thickness can be derived when the type of film is determined. Then, the expected value of the wavelength centroid at each film thickness can be derived while changing the film thickness conditions while using the same type of film. This results in a state where the expected values ​​of the wavelength centroid for multiple film thickness conditions have been derived for a given type of film. Now, let's assume that the expected values ​​of the wavelength centroid for multiple film thickness conditions have been derived for a given type of film. Then, by performing curve fitting on the data plotted with multiple film thicknesses and expected values ​​of the wavelength centroid, a waveform (curve) showing the relationship between film thickness d and the wavelength centroid x'' can be derived. Examples of curve fitting methods include polynomial approximation and other curve fitting methods. In addition to curve fitting, a curve showing the relationship between wavelength and spectral intensity may also be derived using interpolation methods or other methods.

[0058] By deriving such curves for each combination of monochromatic light shown in Figure 10(a) (i.e., for the first to third light), waveforms 401 showing relational information for the first light, waveform 402 showing relational information for the second light, and waveform 403 showing relational information for the third light are derived, as shown in Figure 10(b). The storage unit 33 stores these waveforms 401 to 403 as waveforms showing the relationship information between the film thickness of sample 100 and the wavelength centroid for the first to third light.

[0059] Figure 11(a) shows the system spectral characteristics according to the wavelength for each set filter in the configuration shown in Figure 3 (configuration using a white light source 10b and a filter switching mechanism 10c). In Figure 11(a), the horizontal axis represents wavelength and the vertical axis represents spectral characteristics. As shown in Figure 11(a), the spectral characteristics of the transmitted and reflected sides are derived for each situation: no filter, long-pass filter set, and short-pass filter set. Now, let's assume that the light from the white light source 10b without a filter is the first light having the first wavelength. Let's also assume that the light from the white light source 10b passing through the short-pass filter is the second light having the second wavelength. And let's assume that the light from the white light source 10b passing through the long-pass filter is the third light having the third wavelength. Then, using the same method as described above, the expected value of the wavelength centroid is derived from the theoretical reflectance and the spectral characteristics of the film thickness measuring device 1, and the expected value of the wavelength centroid is plotted and a relational expression is derived by curve fitting. As a result, waveforms showing the relationship information between film thickness and wavelength centroid are derived for the first to third rays of light. That is, as shown in Figure 11(b), waveform 501 showing the relationship information for the first ray, waveform 502 showing the relationship information for the second ray, and waveform 503 showing the relationship information for the third ray are derived. The storage unit 33 stores these waveforms 501 to 503 as waveforms showing the relationship information between the film thickness and wavelength centroid of sample 100 for the first to third rays of light.

[0060] (Film thickness calculation process) The analysis unit 32 derives at least one first film thickness candidate from the relationship information between the film thickness of sample 100 and the wavelength centroid at a first wavelength (the wavelength of the first light), and from the first wavelength centroid calculated based on the first signal. Similarly, the analysis unit 32 derives at least one second film thickness candidate from the relationship information between the film thickness of sample 100 and the wavelength centroid at a second wavelength (the wavelength of the second light), and from the second wavelength centroid calculated based on the second signal. Similarly, the analysis unit 32 derives at least one third film thickness candidate from the relationship information between the film thickness of sample 100 and the wavelength centroid at a third wavelength (the wavelength of the third light), and from the third wavelength centroid calculated based on the third signal. Then, the analysis unit 32 derives the film thickness of sample 100 based on the first film thickness candidate, the second film thickness candidate, and the third film thickness candidate. Furthermore, even if the analysis unit 32 obtains signals at three wavelengths, a first wavelength, a second wavelength, and a third wavelength, when deriving the film thickness of sample 100, it may derive candidate film thicknesses for sample 100 from the relationship information between the film thickness of sample 100 and the wavelength centroid at at least one of these wavelengths, and from the wavelength centroid calculated based on the signals. Alternatively, the analysis unit 32 may select one or more wavelengths to be used to derive the film thickness of sample 100 according to the estimated film thickness of sample 100, and derive the film thickness of sample 100 based on the candidate film thicknesses at the selected one or more wavelengths.

[0061] The specific steps for determining the film thickness will be explained with reference to Figures 12 to 18. Figure 12 is a flowchart showing the specific steps for measuring (determining) the film thickness. As shown in Figure 12, the film thickness determination process begins with setting up the film thickness search block (step S1).

[0062] Figures 13(a) to 13(c) illustrate the setting of film thickness search blocks for three different wavelength patterns. The three wavelength patterns here refer to wavelength pattern 1 for the first wavelength (wavelength of the first light), wavelength pattern 2 for the second wavelength (wavelength of the second light), and wavelength pattern 3 for the third wavelength (wavelength of the third light). Figure 13(a) illustrates the setting of film thickness search blocks in the relationship information for wavelength pattern 1. Figure 13(b) illustrates the setting of film thickness search blocks in the relationship information for wavelength pattern 2. Figure 13(c) illustrates the setting of film thickness search blocks in the relationship information for wavelength pattern 3. As shown in Figures 13(a) to 13(c), in setting the film thickness search blocks, extreme values ​​are determined for each relationship information for each wavelength pattern, and the area between the extreme values ​​is set as one film thickness search block. In this case, 11 film thickness search blocks are set in the relationship information for wavelength pattern 1 shown in Figure 13(a). Also, 13 film thickness search blocks are set in the relationship information for wavelength pattern 2 shown in Figure 13(b). Furthermore, in the relationship information for wavelength pattern 3 shown in Figure 13(c), 10 film thickness search blocks are set.

[0063] As shown in Figure 12, the film thickness search blocks from which the film thickness candidates will be derived are then limited (step S2). Figure 14 is a diagram illustrating the derivation of film thickness candidates when the film thickness search blocks are not limited. As shown in Figure 14, when film thickness candidates are derived from the measured wavelength centroid, there are multiple film thickness candidates, so if film thickness candidates are derived for all film thickness search blocks, it will take a long time to derive the film thickness candidates.

[0064] Therefore, in the film thickness derivation process according to this embodiment, the film thickness search blocks that are to be used to derive film thickness candidates are limited. Figure 15 is a diagram illustrating the derivation of film thickness candidates according to the film thickness prediction range. As shown in Figure 15, the analysis unit 32 sets the film thickness prediction range in advance from the film thickness design value and derives film thickness candidates (first to third film thickness candidates) only from within the film thickness prediction range. More specifically, the analysis unit 32 designates only the film thickness search blocks within the film thickness prediction range as film thickness search blocks that are to be used to derive film thickness candidates, and derives film thickness candidates only from these film thickness search blocks.

[0065] As shown in Figure 12, a film thickness candidate is then derived for each film thickness search block that is targeted for deriving a film thickness candidate (step S3). Figure 16 is a diagram illustrating the derivation of film thickness candidates for each wavelength pattern. Figure 16 shows waveform 601 of relational information for wavelength pattern 1, waveform 602 of relational information for wavelength pattern 2, and waveform 603 of relational information for wavelength pattern 3. Currently, the film thickness search blocks are limited according to the predicted film thickness range, and the thin line portions of each waveform 601, 602, and 603 are excluded from the target for deriving film thickness candidates. For waveform 601 of wavelength pattern 1, two film thickness candidates, film thickness candidate 1 and film thickness candidate 2, are derived based on the measured value of the wavelength centroid. For waveform 602 of wavelength pattern 2, two film thickness candidates, film thickness candidate 3 and film thickness candidate 4, are derived based on the measured value of the wavelength centroid. For waveform 603 of wavelength pattern 3, one film thickness candidate 5 is derived based on the measured value of the wavelength centroid.

[0066] Figure 17 shows the derived candidate film thicknesses. In Figure 17, the horizontal axis represents the wavelength pattern number, and the vertical axis represents the film thickness. As shown in Figure 17, the aforementioned candidate film thicknesses 1 to 5 are represented by film thicknesses 701, 702, 703, 704, and 705, respectively. Specifically, film thickness 701 and film thickness 702, which is larger than film thickness 701, are shown as candidate film thicknesses for wavelength pattern 1. Film thickness 704 and film thickness 703, which is smaller than film thickness 704, are shown as candidate film thicknesses for wavelength pattern 2. Film thickness 705 is shown as a candidate film thickness for wavelength pattern 3. As shown in Figure 17, film thicknesses 701, 704, and 705 are approximately similar in size (thickness) to each other.

[0067] As shown in Figure 12, all combinations of film thickness candidates are then selected so that each wavelength pattern has one film thickness candidate (step S4). Figure 18 is a diagram illustrating the derivation of film thickness based on the selection of the optimal combination of film thickness candidates. In Figure 18, the vertical axis represents film thickness. As shown in Figure 18, the following combinations are selected so that each wavelength pattern has one film thickness candidate: combination 1 consisting of film thicknesses of 701, 703, and 705; combination 2 consisting of film thicknesses of 701, 704, and 705; combination 3 consisting of film thicknesses of 702, 703, and 705; and combination 4 consisting of film thicknesses of 702, 704, and 705.

[0068] As shown in Figure 12, finally, the optimal combination is selected from the above-mentioned combinations of film thicknesses (step S5). The optimal combination here is the combination in which the standard deviation value of each film thickness included in the combination is the smallest. As shown in Figure 18, combination 2, consisting of film thicknesses of 701, 704, and 705, has the smallest standard deviation value of film thicknesses, so combination 2 is selected as the optimal combination. Then, the values ​​of film thicknesses 701, 704, and 705 included in combination 2 (for example, their average value) are derived as the final film thickness value.

[0069] The film thickness measuring device 1 measures the film thickness of a sample 100 on which a film 100b is formed on a substrate 100a. The film thickness measuring device 1 includes a light irradiation unit 10 that irradiates the sample 100 with a first light having a first wavelength and a second light having a second wavelength different from the first wavelength, respectively; an inclined dichroic mirror 22 whose transmittance and reflectance change according to the wavelength in a predetermined wavelength range and separates the first light and the second light from the sample 100 by transmitting and reflecting them; area sensors 23 and 24 that detect the first light reflected by the inclined dichroic mirror 22 and the first light transmitted through the inclined dichroic mirror 22 and output a first signal, as well as the second light reflected by the inclined dichroic mirror 22 and the second light transmitted through the inclined dichroic mirror 22 and output a second signal; and an analysis unit 32 that calculates the first wavelength centroid based on the first signal and calculates the second wavelength centroid based on the second signal. The analysis unit 32 derives at least one first film thickness candidate from the relationship information between the film thickness of sample 100 and the wavelength centroid at a first wavelength, and from the first wavelength centroid, and derives at least one second film thickness candidate from the relationship information between the film thickness of sample 100 and the wavelength centroid at a second wavelength, and from the second wavelength centroid, and derives the film thickness of sample 100 based on the first and second film thickness candidates.

[0070] In the film thickness measuring device 1 according to this embodiment, a first light having a first wavelength is irradiated onto the sample 100, and the first light from the sample 100 reflected by the tilted dichroic mirror 22 and the first light transmitted through the sample 100 are detected to output a first signal, and a first wavelength centroid is calculated based on the first signal. Then, at least one first film thickness candidate is derived from the relationship information between the film thickness of the sample 100 at a predetermined first wavelength and the wavelength centroid, and the calculated first wavelength centroid. In addition, in the film thickness measuring device 1 according to this embodiment, a second light having a second wavelength is irradiated onto the sample 100, and the second light from the sample 100 reflected by the tilted dichroic mirror 22 and the second light transmitted through the tilted dichroic mirror 22 are detected to output a second signal, and a second wavelength centroid is calculated based on the second signal. Then, at least one second film thickness candidate is derived from the relationship information between the film thickness of sample 100 at a predetermined second wavelength and the wavelength centroid, and from the calculated second wavelength centroid. Furthermore, in the film thickness measuring device 1 according to this embodiment, the film thickness of sample 100 is derived based on the first and second film thickness candidates described above.

[0071] As described above, in the film thickness measuring device 1 according to this embodiment, one or more film thickness candidates are derived based on light of multiple wavelengths that are different from each other. For example, when attempting to derive film thickness from the relationship information between the film thickness of sample 100 and the wavelength centroid at a certain wavelength and the derived wavelength centroid, there may be multiple film thickness candidates corresponding to the calculated wavelength centroid, making it impossible to uniquely identify the film thickness. Furthermore, depending on the shape of the waveform showing the relationship information between film thickness and the wavelength centroid, the extreme values ​​of the waveform may become gradual, making it difficult to distinguish between film thickness differences due to differences in the wavelength centroid, and making it impossible to accurately derive the film thickness corresponding to the wavelength centroid. In this regard, as described above, by deriving one or more film thickness candidates based on light of multiple wavelengths that are different from each other, even when it is not possible to uniquely identify the film thickness with light of a single wavelength, the film thickness of sample 100 can be appropriately derived based on the film thickness candidates by considering the consistency of the film thickness candidates for each of the multiple wavelengths of light. Furthermore, for light of multiple wavelengths that are different from each other, the extreme values ​​of the waveform that show the relationship information between film thickness and the wavelength centroid (the range in which the film thickness cannot be accurately derived) do not coincide with each other. Therefore, the film thickness of sample 100 can be appropriately derived based on each candidate film thickness. As described above, the film thickness measuring device 1 according to this embodiment can measure the film thickness of sample 100 with high accuracy.

[0072] Figure 19(a) illustrates the derivation of film thickness in sample 100, which has large variations in film thickness. For example, as shown in Figure 19(a), even with sample 100 (sample 100 such as an oxide film) which has variations in film thickness of about 500 nm ± 100 nm, the method of the film thickness measuring device 1 according to this embodiment allows for high-range measurement, so the film thickness can be measured with high accuracy.

[0073] Figure 19(b) shows an example of ground truth data for film thickness. Figure 19(c) shows the multi-wavelength analysis results corresponding to the ground truth data shown in Figure 19(b). In Figures 19(b) and (c), the horizontal and vertical axes indicate the position of sample 100. Also, in Figures 19(b) and (c), the intensity of the color indicates the film thickness. As shown in Figures 19(b) and (c), the method of the film thickness measuring device 1 according to the embodiment described above (i.e., film thickness derivation based on multi-wavelength analysis) made it possible to perform highly accurate film thickness measurements with a high degree of agreement with the ground truth data for film thickness.

[0074] The light irradiation unit 10 further irradiates the sample 100 with a third light having a third wavelength different from the first and second wavelengths, the tilted dichroic mirror 22 further separates the sample 100 by transmitting and reflecting the third light, the area sensors 23 and 24 further detect the third light reflected by the tilted dichroic mirror 22 and the third light transmitted through the tilted dichroic mirror 22 and output a third signal, the analysis unit 32 further calculates the third wavelength centroid based on the third signal, and further derives at least one third film thickness candidate from the relationship information between the film thickness of the sample 100 at the third wavelength and the wavelength centroid, and from the third wavelength centroid, and may derive the film thickness of the sample 100 based on the first film thickness candidate, the second film thickness candidate, and the third film thickness candidate. In this way, by deriving one or more film thickness candidates based on three different wavelengths of light, it becomes even easier to uniquely identify the film thickness compared to the case where only two wavelengths of light are used. This allows for more accurate measurement of the film thickness of sample 100.

[0075] The analysis unit 32 may pre-set a film thickness prediction range from the film thickness design value and derive the first and second film thickness candidates only from within the film thickness prediction range. If the film thickness prediction range is predetermined, the processing time related to film thickness derivation can be shortened by deriving film thickness candidates only from within the film thickness prediction range.

[0076] The light irradiation unit 10 has a multicolor LED array light source 10a capable of emitting monochromatic light of three or more wavelengths, and the sample 100 may be irradiated with a first light composed of monochromatic light of two wavelengths from among the three or more wavelengths of monochromatic light. The waveform showing the relationship between film thickness and wavelength centroid has a steeper slope as the wavelength range of the light is wider, and a gentler slope as the wavelength range of the light is narrower. The steeper the slope of the waveform curve, the greater the change in film thickness in response to the change in wavelength centroid, so the film thickness can be accurately derived. For this reason, by simultaneously emitting two wavelengths of monochromatic light and widening the wavelength range of the first light, the curve in the waveform showing the relationship between film thickness and wavelength centroid can be made steeper, and the film thickness of the sample 100 can be measured with higher precision.

[0077] Figure 20(a) shows the relationship between the wavelength centroid and film thickness when light with a relatively narrow wavelength range (wavelength width) is used. Figure 20(b) shows the relationship between the wavelength centroid and film thickness when light with a relatively wide wavelength range (wavelength width) is used. Figure 20(c) shows the relationship between the wavelength centroid and film thickness when light with a relatively wide wavelength range (wavelength width) and multiple peaks (two maximum values) is used. In Figures 20(a) to 20(c), the film reflectance conditions are the same, and the film reflectance changes according to the change in film thickness. Now, as shown in Figures 20(a) and (b), when the wavelength range of the light emitted from the light source is relatively narrow, the slope of the curve in the waveform showing the relationship between film thickness and the wavelength centroid becomes smaller compared to when the wavelength range of the light is relatively wide. In this case, the change in film thickness with respect to the change in the wavelength centroid becomes small, making it difficult to accurately derive the film thickness. Conversely, by widening the wavelength range (wavelength width) of the light emitted from the light source, the accuracy of film thickness derivation can be improved. Furthermore, as shown in Figures 20(b) and 20(c), by using light with a relatively wide wavelength range (wavelength width) and multiple peaks (two maximum values), the slope of the curve in the waveform showing the relationship between film thickness and wavelength centroid becomes even greater than when light with a single peak (one maximum value) is used, thereby improving the accuracy of film thickness derivation.

[0078] The two monochromatic light wavelengths emitted from the multicolor LED array light source 10a may be red wavelength light and blue wavelength light. By using red wavelength light and blue wavelength light, which have a large difference in wavelength from each other, the wavelength range of the first light can be broadened, allowing for more accurate measurement of the film thickness of the sample 100. Red wavelength light includes some wavelengths in the wavelength range of 610 nm to 780 nm, green wavelength light includes some wavelengths in the wavelength range of 500 nm to 570 nm, and blue wavelength monochromatic light includes some wavelengths in the wavelength range of 430 nm to 490 nm.

[0079] Figure 21(a) shows the film thickness mapping results when red and blue wavelength light are used. Figure 21(b) shows the film thickness mapping results when green and blue wavelength light are used. Figure 21(c) shows the film thickness mapping results when red and green wavelength light are used. Figure 21(d) shows the ground truth data for the film thickness mapping results. Now, we will describe an example of performing film thickness mapping (mapping the film thickness of each region of sample 100) while changing the combination of two monochromatic light wavelengths emitted from the multicolor LED array light source 10a, while keeping all other conditions the same. Figure 21(d) shows the ground truth data for the film thickness mapping. As shown in Figures 21(a) to (d), the degree of agreement with the ground truth data for film thickness was highest when the combination of two monochromatic light wavelengths was red and blue wavelengths (see Figure 21(a)). In other words, by using red and blue wavelength light, which have a large difference in wavelength, the film thickness of sample 100 could be measured with high accuracy.

[0080] In the film thickness measurement method described above, that is, the method of deriving the film thickness of sample 100 based on relational information and the wavelength centroid, there are two challenges in ensuring the accuracy of the film thickness measurement. First, there is a decrease in the accuracy of film thickness derivation at the extreme values ​​of the waveform (hereinafter sometimes referred to as the calibration curve) that shows the relational information described above. As shown in Figure 8, at the extreme values ​​of the calibration curve, the slope of the waveform curve becomes gentler, and the difference in film thickness due to the difference in the wavelength centroid becomes less apparent, so it may not be possible to accurately derive the film thickness relative to the wavelength centroid. In the example described above, as shown in Figure 9, it was explained that the film thickness could be appropriately identified based on any of the calibration curves because the regions of the extreme values ​​in the calibration curves for each of the multiple wavelengths do not coincide with each other. However, when the regions of the extreme values ​​in the calibration curves for each of the multiple wavelengths are close to each other, there is a risk that the film thickness cannot be accurately identified.

[0081] Secondly, if the calculated wavelength centroid and the calibration curve showing the related information do not intersect (i.e., there are wavelengths for which it is not possible to derive a film thickness candidate in the example described above), it becomes difficult to derive the final film thickness. For example, in the example shown in Figure 22(a), for three different wavelengths of light, the calculated wavelength centroid (wavelengths shown by dashed lines) and the calibration curve intersect, so a film thickness candidate can be derived for each wavelength of light. In this case, the final film thickness can be determined based on the standard deviation between the film thickness candidates. On the other hand, in the example shown in Figure 22(b), the calculated wavelength centroid and the calibration curve do not intersect for the wavelength of light shown in the middle row. Also, in the example shown in Figure 22(c), the calculated wavelength centroid and the calibration curve do not intersect for the wavelengths of light shown in the middle row and the wavelengths shown in the bottom row. In such cases, it becomes difficult to determine the final film thickness based on the standard deviation between the film thickness candidates.

[0082] To address these challenges, the film thickness measurement method according to this embodiment employs a film thickness derivation algorithm that considers multiple factors affecting film thickness. Specifically, the film thickness measurement method according to this embodiment designs multiple factors as evaluation values ​​and derives a film thickness value based on the film thickness candidate with the highest evaluation. A high evaluation means that the evaluation value of negative factors is small or the evaluation value of positive factors is large.

[0083] Figure 23 illustrates the image of film thickness derivation based on evaluation values. In Figure 23, the following film thickness candidates are identified from the intersection of the three calculated wavelength centroids (0.662 nm, 0.714 nm, and 0.611 nm) and the calibration curve showing related information: 561 nm, 620 nm, 562 nm, 643 nm, 440 nm, and 563 nm. For each film thickness candidate, an evaluation value is calculated considering multiple factors described later, and the film thickness value with the smallest evaluation value, which means the highest evaluation, is identified as the final film thickness value: 562 nm. As mentioned above, for negative elements, a smaller evaluation value indicates a higher evaluation, and for positive elements, a larger evaluation value indicates a higher evaluation. In the graph in Figure 23, the horizontal axis represents film thickness, and the vertical axis represents the wavelength centroid. Although not shown in the diagrams, similarly in Figures 24(a)(b), 25(a), 26(a)-(e), 28(a)(b), and 29(a)(b), which will be discussed later, the horizontal axis represents film thickness and the vertical axis represents the wavelength centroid.

[0084] Specifically, the analysis unit 32 may derive evaluation values ​​for each film thickness candidate by performing distance evaluation and calibration curve accuracy evaluation, and derive the film thickness of sample 100 based on the evaluation values ​​(derivation results). As shown in Figures 24(a) and 24(b), the distance evaluation is Ψ dist , calibration curve accuracy evaluation Ψ accu In that case, the evaluation value is given by equation (3) below, for example. In equation (3), α and β are weight parameters that satisfy 0 ≤ α, 0 ≤ β, and α + β = 1. Evaluation value = αΨ dist +βΨ accu ...(3)

[0085] Figure 24(a) is a diagram illustrating the overview of distance evaluation. When performing distance evaluation, first, the film thickness at the intersection (target point) of one of the wavelength centroids and the calibration curve is set as a candidate film thickness. Then, the points of the above candidate film thickness at other wavelength centroids are set as comparison points, and the wavelength direction distance, which is the distance between these comparison points and the calibration curve (distance in the wavelength direction), is evaluated. Thus, the wavelength direction distance is the difference between the wavelength centroid corresponding to the candidate film thickness identified from the candidate film thickness and the calibration curve showing related information, and the calculated wavelength centroid. In the left example shown in Figure 24(a), both comparison points are far from the calibration curve, resulting in a low evaluation. On the other hand, in the right example shown in Figure 24(a), both comparison points are close to the calibration curve, resulting in a high evaluation.

[0086] Figures 25(a) and 25(b) illustrate examples of setting candidate film thicknesses. In the example shown in Figure 25(a), the intersections A and B are shown from top to bottom with the calibration curve SC1, which shows the first relationship information between the first wavelength centroid of light and the first relationship information; C is shown with the calibration curve SC2, which shows the second relationship information between the second wavelength centroid of light and the second relationship information; and D is shown with the calibration curve SC3, which shows the third relationship information between the third wavelength centroid of light and the third relationship information. In this case, the film thickness y corresponding to intersection A is... i (0) , film thickness y corresponding to intersection B i (1) , film thickness y corresponding to intersection C i (2) , film thickness y corresponding to intersection D i (3) Each of these is set as a candidate for film thickness. In this way, the analysis unit 32 may set the film thickness derived from the calculated first wavelength centroid and first relationship information, the film thickness derived from the calculated second wavelength centroid and second relationship information, and the film thickness derived from the calculated third wavelength centroid and third relationship information as candidate film thicknesses. In addition, as shown in Figure 25(b), the analysis unit 32 may set multiple film thickness candidates (here, film thickness y) at predetermined intervals. i (0) , film thickness i (1) , film thickness i (2) , film thickness i (3)...) may be set. Alternatively, the analysis unit 32 may set only film thicknesses within a predetermined range as film thickness candidates. The predetermined range of film thickness here may be the range of film thickness assumed to be the film thickness of sample 100.

[0087] Figures 26(a) to 26(e) illustrate specific examples of distance evaluation. Now, as shown in Figure 26(a), assume that candidate film thicknesses are set with the intersection points A and B of the first wavelength centroid of light and calibration curve SC1 showing the first relationship information, the intersection point C of the second wavelength centroid of light and calibration curve SC2 showing the second relationship information, and the intersection point D of the third wavelength centroid of light and calibration curve SC3 showing the third relationship information as target points.

[0088] As shown in Figure 26(b), when the target point is intersection A, the evaluation is low because the distance to the calibration curve (calibration curve SC2 or calibration curve SC3) is far for all comparison points. As shown in Figure 26(c), when the target point is intersection B, the distance between the second wavelength centroid of light and calibration curve SC2, which shows the second relationship information, is close, but the distance between the third wavelength centroid of light and calibration curve SC3, which shows the third relationship information, is somewhat far, resulting in a slightly higher evaluation. As shown in Figure 26(d), when the target point is intersection C, the evaluation is high because the distance to the calibration curve (calibration curve SC1 or calibration curve SC3) is close for all comparison points. As shown in Figure 26(e), when the target point is intersection D, the distance between the wavelength centroid of the second light and the calibration curve SC2 showing the second relationship information is short, but the distance between the wavelength centroid of the first light and the calibration curve SC1 showing the first relationship information is somewhat long, resulting in a slightly higher evaluation. In this way, the analysis unit 32 derives the wavelength distance for each of the multiple film thickness candidates and performs an evaluation such that the closer the wavelength distance of a film thickness candidate is, the higher the evaluation (the more likely it is to be adopted as the film thickness of sample 100). That is, the analysis unit 32 may estimate the film thickness candidate with the closest wavelength distance among the multiple film thickness candidates as the film thickness of sample 100. Then, the analysis unit 32 derives the film thickness of sample 100 based on the evaluation result (derivation result).

[0089] Figure 24(b) is a diagram illustrating the overview of calibration curve accuracy evaluation. For each of the multiple film thickness candidates, the analysis unit 32 evaluates the slope of the calibration curve at the intersection of the calibration curve, which shows the relationship information, and the film thickness candidate. The evaluation may be performed so that the film thickness candidate with a larger slope is evaluated higher (so that it is more likely to be adopted as the film thickness of sample 100). In the left figure of Figure 24(b), the intersection of the calibration curve and the film thickness candidate is near the extreme value of the calibration curve, and the slope of the calibration curve is small. In this case, the analysis unit 32 may evaluate the calibration curve accuracy lower. Also, in the right figure of Figure 24(b), the intersection of the calibration curve and the film thickness candidate is far from the extreme value of the calibration curve, and the slope of the calibration curve is large. In this case, the analysis unit 32 may evaluate the calibration curve accuracy higher. Then, as shown in equation (3) above, the analysis unit 32 calculates an evaluation value considering both the distance evaluation and the calibration curve accuracy evaluation, and derives the film thickness of sample 100 based on this evaluation value.

[0090] The derivation of the final film thickness value, taking into account distance evaluation and calibration curve accuracy evaluation, may be performed using the following equations (4) to (8).

[0091]

number

[0092] Furthermore, the elements (factors) evaluated in deriving the film thickness are not limited to the distance (wavelength distance) and calibration curve accuracy (slope of the calibration curve) mentioned above. Figure 27 is a diagram illustrating an example of evaluation elements. As shown in Figure 27, the analysis unit 32 may calculate evaluation values ​​for each film thickness candidate by performing spatial evaluations, namely noise evaluation and peripheral film thickness value evaluation, in addition to the distance evaluation and calibration curve accuracy evaluation mentioned above, and derive the film thickness of sample 100 based on the evaluation values ​​(derivation results).

[0093] For distance evaluation, in addition to the wavelength distance evaluation described above, the film thickness distance (details described later) may also be evaluated. Alternatively, instead of distance evaluation, evaluation using standard deviation as described above may be performed. For calibration curve accuracy evaluation, instead of (or in addition to) the evaluation of the slope of the calibration curve described above, the wavelength shift amount of the calibration curve (details described later) may be evaluated. Noise evaluation is an evaluation of the high-frequency components of the wavelength centroid image (details described later). Peripheral film thickness value evaluation is an evaluation of the difference with adjacent film thickness values ​​(details described later).

[0094] Figure 28 illustrates the evaluation of film thickness directional distance. When evaluating film thickness directional distance, first, the film thickness at the intersection (target point) of one of the wavelength centroids with the calibration curve is set as the film thickness candidate. Then, the intersection points with the corresponding wavelength centroids for the other calibration curves are set as comparison points, and the film thickness directional distance, which is the distance in the film thickness direction from these comparison points to the target point, is evaluated. Thus, the film thickness directional distance is the difference between the film thickness (comparison point) derived from the calculated wavelength centroid and related information and the film thickness candidate (target point). In the left example shown in Figure 28, the distance in the film thickness direction to the target point is far for both comparison points, resulting in a low evaluation. On the other hand, in the right example shown in Figure 28, the distance in the film thickness direction to the target point is close for both comparison points, resulting in a high evaluation. The analysis unit 32 may derive the film thickness directional distance for each of the multiple film thickness candidates and estimate the film thickness candidate with the closest film thickness directional distance among the multiple film thickness candidates as the film thickness of sample 100. For example, the analysis unit 32 may derive the film thickness of sample 100 in such a way that film thickness candidates with closer distances in the film thickness direction are more likely to be adopted as the film thickness of sample 100.

[0095] Figure 29 is a diagram illustrating the evaluation of the wavelength shift amount of the calibration curve. For each of the multiple film thickness candidates, the analysis unit 32 evaluates the wavelength shift amount, which is the distance in the wavelength direction from the intersection point of the calibration curve showing the relationship information and the film thickness candidate to the point where the wavelength centroid is 0. The evaluation may be performed so that the smaller the wavelength shift amount of the film thickness candidate (the film thickness candidate that is farther from the extreme value and has a large slope) the higher the evaluation (so that it is more likely to be adopted as the film thickness of sample 100). Comparing the left and right figures of Figure 29, the wavelength shift amount is smaller in the right figure of Figure 29 than in the left figure of Figure 29, so the evaluation is higher.

[0096] Figure 30 illustrates noise evaluation. In noise evaluation, for example, a low-frequency image (Figure 30(b)) is generated by removing high-frequency components from a wavelength centroid image (Figure 30(a)) using a median filter or low-pass filter. Then, a high-frequency image (Figure 30(c)) is generated from the absolute difference between the wavelength centroid image and the low-frequency image. Noise evaluation of the high-frequency components of the wavelength centroid distribution is performed on the high-frequency image. Smoothing processing may also be applied to the spatial distribution of the wavelength centroid. Noise Evaluation Ψ nois This may be done by the following equation (9).

[0097]

number

[0098] Figure 31 illustrates the evaluation of peripheral film thickness values. As shown in Figures 31(a) to 31(c), the difference between the derived film thickness of sample 100 and the film thickness of peripheral (e.g., adjacent) pixels may be evaluated as to whether the difference is large or small. In peripheral film thickness value evaluation, the smaller the difference from the peripheral film thickness, the higher the evaluation.

[0099] Figure 32 illustrates the input for each evaluation value. As mentioned above, in the derivation of film thickness, distance evaluation (specifically, evaluation of wavelength direction distance and evaluation of film thickness direction distance) is performed. For the calculation of distance evaluation, a pair of wavelength centroid image and calibration curve is required. In addition, calibration curve accuracy evaluation (specifically, evaluation of the slope and wavelength shift amount of the calibration curve) is performed. Furthermore, noise evaluation is calculated based on the high-frequency image. Furthermore, peripheral film thickness value evaluation is calculated based on the film thickness measurement results. Note that, except for distance evaluation using film thickness direction distance, it is possible to perform these evaluations from only one wavelength pattern (wavelength centroid image). At least one of these evaluations needs to be performed, and any combination of evaluation elements may be performed as appropriate.

[0100] The derivation of the final film thickness value by combining the evaluation elements described above may be performed using equations (10) to (12) below. α, β, γ, and δ are weight parameters that satisfy 0 ≤ α, 0 ≤ β, 0 ≤ γ, and 0 ≤ δ. The weight parameters may be designed so that their sum is 1. The number of evaluation elements is not limited to one or more, and the combination is not limited.

[0101]

number

[0102] Next, three examples of the film thickness measurement method performed by the film thickness measuring device according to this embodiment will be described. In the first example, the film thickness is derived based on distance evaluation (see Figure 33). In the second example, the film thickness is derived based on distance evaluation and calibration curve accuracy evaluation (see Figure 34). In the third example, the film thickness is derived based on distance evaluation, calibration curve accuracy evaluation, and spatial evaluation (noise evaluation, peripheral film thickness value evaluation) (see Figure 35).

[0103] In the film thickness measurement method shown in Figure 33, the sample 100 is first irradiated with light from the light irradiation unit 10 (step S1). The irradiated light here is, for example, a first light having a first wavelength, a second light having a second wavelength, and a third light having a third wavelength.

[0104] Next, the measurement light reflected by sample 100 is detected by area sensors 23 and 24, and signals are output (step S2). The signals output by area sensors 23 and 24 are a first signal related to the calculation of the wavelength centroid of the first light (when the irradiated light is the first light), a second signal related to the calculation of the wavelength centroid of the second light (when the irradiated light is the second light), and a third signal related to the calculation of the wavelength centroid of the third light (when the irradiated light is the third light). Steps S1 and S2 constitute the light measurement step.

[0105] Next, the analysis unit 32 calculates the wavelength centroid based on the signals from the area sensors 23 and 24 (step S3). The analysis unit 32 calculates the wavelength centroid of the first light based on the first signal, the wavelength centroid of the second light based on the second signal, and the wavelength centroid of the third light based on the third signal.

[0106] Next, in the analysis unit 32, an evaluation value for each candidate film thickness is calculated based on the relevant information and the calculated wavelength centroid (step S4). Specifically, in step S4, the processes in steps S41 to S46 are performed.

[0107] In step S41, the intersection point between the calibration curve showing the relationship information and the calculated wavelength centroid is searched. For example, as shown in Figure 26, the intersection point with the wavelength centroid is searched for each of the calibration curves for the first light, the second light, and the third light.

[0108] Next, based on the intersections mentioned above, candidate film thicknesses are determined (step S42). For example, as shown in the example in Figure 26, intersections A and B with respect to the first light, intersection C with respect to the second light, and intersection D with respect to the third light are determined as candidate film thicknesses.

[0109] Next, the wavelength centroid of the film thickness candidate is derived based on a calibration curve showing the relevant information (step S43). Then, the difference (wavelength direction distance) between the wavelength centroid calculated in step S3 and the wavelength centroid derived based on the calibration curve is calculated as an evaluation value (step S44).

[0110] Furthermore, based on the calibration curve showing the relevant information, the film thickness at the calculated wavelength centroid is derived (step S45). Then, the difference between the candidate film thickness and the derived film thickness (film thickness direction distance) is calculated as an evaluation value (step S46).

[0111] Finally, the calculated evaluation values ​​for each film thickness candidate are compared, the film thickness candidate with the highest evaluation is selected, and the film thickness of that candidate is derived as the film thickness of sample 100 (step S5). Steps S3 to S5 constitute the derivation steps.

[0112] The film thickness measurement method shown in Figure 34 is generally the same as the film thickness measurement method shown in Figure 33, and steps S101 to S103 and S105 in Figure 34 correspond to steps S1 to S3 and S5 in Figure 33. Step S104 in Figure 34 differs in part from step S4 in Figure 33; specifically, in addition to performing steps S141 to S146 similar to steps S41 to S46, steps S147 and S148 are also performed. Here, only steps S147 and S148 (calibration curve accuracy evaluation) will be explained.

[0113] In step S147, for each candidate film thickness, the slope of the calibration curve at the intersection of the calibration curve representing the relationship information and the candidate film thickness is calculated as an evaluation value. In step S148, for each candidate film thickness, the wavelength shift amount, which is the distance in the wavelength direction from the intersection of the calibration curve representing the relationship information and the candidate film thickness to the point where the wavelength centroid is 0, is calculated as an evaluation value. Then, in step S105, the evaluation values ​​calculated among the candidate film thicknesses are compared, the candidate film thickness with the highest evaluation is selected, and the film thickness of that candidate film thickness is derived as the film thickness of sample 100.

[0114] The film thickness measurement method shown in Figure 35 is generally the same as the film thickness measurement method shown in Figure 34, and steps S201 to S203, S205, and S206 in Figure 35 correspond to steps S101 to S105 in Figure 34. The film thickness measurement method shown in Figure 35 further includes steps S204 and S207. Here, only steps S204 and S207 (spatial evaluation) will be explained.

[0115] Step S204 is performed following step S203. In step S204, high-frequency components are removed from the wavelength centroid image using a median filter or low-pass filter to generate a low-frequency image. A high-frequency image is then generated from the absolute difference between the wavelength centroid image and the low-frequency image, and noise evaluation of the high-frequency components of the wavelength centroid distribution is performed in the high-frequency image. In step S207, peripheral film thickness value evaluation is performed to evaluate whether the difference between the derived film thickness of sample 100 and the film thickness of surrounding (e.g., adjacent) pixels is large or not.

[0116] Next, the operation and effects of the film thickness measuring device 1 according to this embodiment will be described.

[0117] The film thickness measuring device 1 according to this embodiment is a film thickness measuring device for measuring the film thickness of a sample 100 in which a film 100b is formed on a substrate 100a, and comprises: an optical measuring unit that irradiates the sample 100 with first light having a first wavelength, detects the first light reflected by the sample 100, and outputs a first signal related to the calculation of the wavelength centroid of the first light; a storage unit 33 that stores first relationship information showing the relationship between the film thickness of the sample 100 at the first wavelength and the wavelength centroid; and a unit that calculates the wavelength centroid of the first light based on the first signal. The system includes an analysis unit 32 that derives the film thickness of sample 100 based on first relational information and the calculated wavelength centroid of the first light. The analysis unit 32 derives at least one of the following: the wavelength centroid corresponding to the film thickness candidate identified from the film thickness candidate and the first relational information, which is the wavelength direction distance, which is the difference between the calculated wavelength centroid of the first light and the calculated wavelength centroid of the first light, and the film thickness direction distance, which is the difference between the film thickness derived from the calculated wavelength centroid of the first light and the first relational information and the film thickness candidate. Based on the derive results, the system derives the film thickness of sample 100.

[0118] In the film thickness measuring device 1 according to this embodiment, the film thickness of sample 100 is derived based on relationship information showing the relationship between film thickness and wavelength centroid, which is stored in advance, and the wavelength centroid of the first light calculated from the detected light. Here, it is conceivable that the film thickness cannot be uniquely determined from the relationship information and wavelength centroid (there are multiple candidate film thicknesses). In this regard, in the film thickness measuring device 1 according to this embodiment, at least one of the following is derived: the wavelength direction distance, which is the difference between the wavelength centroid corresponding to the candidate film thickness and the calculated wavelength centroid, and the film thickness direction distance, which is the difference between the film thickness derived from the calculated wavelength centroid and the relationship information and the candidate film thickness. The film thickness of sample 100 is then derived considering the derived results. This makes it possible, for example, when a candidate film thickness cannot be uniquely determined, to consider the wavelength direction distance and the film thickness direction distance, and to determine that the candidate film thickness is closer to the actual film thickness of sample 100. This allows for highly accurate measurement of the film thickness of sample 100.

[0119] In the film thickness measuring device 1 according to this embodiment, the analysis unit 32 may derive at least one of the wavelength distance and the film thickness distance for each of the multiple film thickness candidates, and derive the film thickness of sample 100 in such a way that film thickness candidates with closer wavelength distances are more likely to be adopted as the film thickness of sample 100. That is, the analysis unit 32 may derive the film thickness distance for each of the multiple film thickness candidates, and estimate the film thickness candidate with the closest film thickness distance among the multiple film thickness candidates to be the film thickness of sample 100. Alternatively, the analysis unit 32 may derive the wavelength distance for each of the multiple film thickness candidates, and estimate the film thickness candidate with the closest wavelength distance among the multiple film thickness candidates to be the film thickness of sample 100. With such a configuration, when there are multiple film thickness candidates, film thickness candidates with closer wavelength distances and film thickness distances are more likely to be adopted as the film thickness of sample 100, so that the film thickness of the object can be measured with high accuracy.

[0120] In the film thickness measuring device 1 according to this embodiment, the optical measuring unit may include: a light irradiation unit 10 that irradiates the sample 100 with first light having a first wavelength; an inclined dichroic mirror 22 whose transmittance and reflectance change according to wavelength in a predetermined wavelength range and separates the first light from the sample 100 by transmitting and reflecting it; an area sensor 24 that detects the first light reflected by the inclined dichroic mirror 22 and outputs a signal related to the calculation of the wavelength centroid, which is a wavelength parameter of the first light; and an area sensor 23 that detects the first light transmitted through the inclined dichroic mirror 22 and outputs a signal related to the calculation of the wavelength centroid, which is a wavelength parameter of the first light. With such a configuration, the wavelength centroid, which has a high correlation with film thickness, can be calculated as a wavelength parameter, and the film thickness of the sample 100 can be measured with high accuracy.

[0121] In the film thickness measuring device 1 according to this embodiment, the light irradiation unit 10 irradiates a first light and a second light having a second wavelength different from the first wavelength, respectively, the tilted dichroic mirror 22 separates the first light and the second light from the sample 100 by transmitting and reflecting them, the area sensors 23 and 24 detect at least one of the first light reflected by the tilted dichroic mirror 22 or transmitted through the tilted dichroic mirror 22 and output a first signal related to the calculation of the wavelength centroid, which is the wavelength parameter of the first light, and also detect at least one of the second light reflected by the tilted dichroic mirror 22 or transmitted through the tilted dichroic mirror 22 and output a second signal related to the calculation of the wavelength centroid, which is the wavelength parameter of the second light. The output is generated, and the storage unit 33 stores the first relational information and the second relational information. The analysis unit 32 derives at least one of the following: the wavelength centroid corresponding to the film thickness candidate identified from the film thickness candidate and the first relational information, which is the difference between the calculated wavelength centroid of the first light and the wavelength centroid of the first light; and the film thickness direction distance, which is the difference between the film thickness derived from the calculated wavelength centroid of the first light and the first relational information and the film thickness candidate. The analysis unit 32 also derives at least one of the following: the wavelength centroid corresponding to the film thickness candidate identified from the film thickness candidate and the second relational information, which is the difference between the calculated wavelength centroid of the second light and the wavelength centroid of the second light; and the film thickness direction distance, which is the difference between the film thickness derived from the calculated wavelength centroid of the second light and the second relational information and the film thickness candidate. Based on the derivation results, the film thickness of sample 100 may be derived. By irradiating sample 100 with light of multiple wavelengths, the amount of information to consider when deciding which of several candidate film thicknesses to adopt can be increased, enabling high-precision measurement of the film thickness of sample 100.

[0122] In the film thickness measuring device 1 according to this embodiment, the analysis unit 32 may set the film thickness derived from the calculated wavelength centroid of the first light and the first relationship information, or the film thickness derived from the calculated wavelength centroid of the second light and the second relationship information, as a candidate film thickness. In this way, by setting the film thickness derived from the wavelength centroid and the relationship information (for example, the film thickness at the intersection of the calibration curve showing the relationship information and the wavelength centroid) as a candidate film thickness, an appropriate candidate film thickness can be set easily and quickly.

[0123] In the film thickness measuring device 1 according to this embodiment, the analysis unit 32 may set film thicknesses within a predetermined range as film thickness candidates. Since the approximate film thickness of sample 100 is usually known in advance, an appropriate film thickness candidate can be easily and quickly set by setting a range of ±100 nm of the expected film thickness as the film thickness candidate. It is possible that even if the film thickness is not the true film thickness value, an evaluation value based on wavelength direction distance, etc., may be high and be selected as the film thickness of sample 100. However, by setting only film thicknesses within a predetermined range as film thickness candidates, such erroneous selections can be avoided, and the film thickness can be measured with high accuracy.

[0124] In the film thickness measuring device 1 according to this embodiment, the analysis unit 32 may set a plurality of film thickness candidates at predetermined intervals. By setting the film thickness candidates without bias at predetermined intervals, the film thickness can be measured with high accuracy.

[0125] In the film thickness measuring device 1 according to this embodiment, the analysis unit 32 may evaluate the slope of the calibration curve at the intersection of the calibration curve showing the first relational information and the film thickness candidate for each of the multiple film thickness candidates, and derive the film thickness of sample 100 in such a way that film thickness candidates with a larger slope are more likely to be adopted as the film thickness of sample 100. In areas where the slope of the calibration curve is small (near the extreme value), the change in film thickness with respect to the change in the wavelength centroid becomes gradual, and there is a risk that the film thickness of sample 100 cannot be derived with high accuracy from the relational information. In this respect, the accuracy of film thickness derivation can be ensured by making film thickness candidates with a larger slope more likely to be adopted.

[0126] [Example of experiment] Experiments on film thickness measurement were conducted under the following conditions. Substrates with film thickness values ​​of 100nm, 200nm, 300nm, 400nm, 500nm, 600nm, 700nm, and 800nm ​​were prepared. Three types of calibrations (red and blue, red and green, green and blue) were calculated from light sources with three different wavelength bands (red: 610nm-780nm, green: 500nm-570nm, blue: 430nm-490nm). Distance evaluation and calibration curve accuracy evaluation were performed, and the evaluation values ​​were calculated using α=0.8 and β=0.2 in equation (3) (i.e., the weighting of distance evaluation:calibration curve accuracy evaluation was 8:2), and the final film thickness value was derived. The measurement area was defined as within a 300-pixel radius from the center of the substrate. The estimated film thickness range was set to the entire range or ±100 nm of the correct value.

[0127] Figure 36 is a table showing the experimental results (average error). Figure 36 shows the average error for each film thickness (error of the derived result relative to the correct data) when the film thickness estimation range is set to ±100 nm of the correct value, and the average error for each film thickness (error of the derived result relative to the correct data) when the film thickness estimation range is set to the entire range. As shown in Figure 36, when the film thickness estimation range is set to ±100 nm of the correct value, the average error is within the range of 0.69 to 2.48 nm, indicating that the estimation error was kept small for all film thicknesses. Furthermore, when the film thickness estimation range is set to the entire range, excluding one outlier (7.19 nm at a film thickness of 500 nm), the average error is within the range of 1.05 to 2.48 nm, indicating that the estimation error was kept small for all film thicknesses.

[0128] Figures 37(a) to 37(f) show the difference between the estimated film thickness and the ground truth data for a sample with a film thickness of 100 nm. Figure 37(a) shows the estimation result when the film thickness estimation range is set to the ground truth value ±100 nm, Figure 37(b) shows the ground truth data, and Figure 37(c) shows the difference between Figures 37(a) and 37(b). Furthermore, Figure 37(d) shows the estimation result when the film thickness estimation range is set to the entire range, Figure 37(e) shows the ground truth data, and Figure 37(f) shows the difference between Figures 37(d) and 37(e). As shown in Figures 37(c) and 37(e), for a sample with a film thickness of 100 nm, there was a slight error when the film thickness estimation range was set to the entire range compared to when the film thickness estimation range was set to the ground truth value ±100 nm, but the film thickness could generally be estimated with high accuracy.

[0129] Figures 38(a) to 38(f) show the difference between the estimated film thickness and the ground truth data for a sample with a film thickness of 800 nm. Figure 38(a) shows the estimation result when the film thickness estimation range is set to the ground truth value ±100 nm, Figure 38(b) shows the ground truth data, and Figure 38(c) shows the difference between Figures 38(a) and 38(b). Furthermore, Figure 38(d) shows the estimation result when the film thickness estimation range is set to the entire range, Figure 38(e) shows the ground truth data, and Figure 38(f) shows the difference between Figures 38(d) and 38(e). As shown in Figures 38(c) and 38(e), for the sample with a film thickness of 800 nm, there was a slight error when the film thickness estimation range was set to the entire range compared to when the film thickness estimation range was set to the ground truth value ±100 nm, but the film thickness could generally be estimated with high accuracy. [Explanation of Symbols]

[0130] 1...Film thickness measuring device, 10...Light irradiation unit, 22...Inclined dichroic mirror (optical element), 23,24...Area sensor (light detection unit), 32...Analysis unit, 33...Storage unit, 100...Sample (object), 100a...Substrate, 100b...Film.

Claims

1. A film thickness measuring device for measuring the film thickness of an object on which a film has been formed on a substrate, A light measuring unit that irradiates the object with first light having a first wavelength, detects the first light reflected by the object, and outputs a first signal related to the calculation of the wavelength parameter of the first light, A storage unit that stores first relational information showing the relationship between the film thickness of the object at the first wavelength and the wavelength parameter, The system includes an analysis unit that calculates the wavelength parameter of the first light based on the first signal, and derives the film thickness of the object based on the first relational information and the calculated wavelength parameter of the first light, The aforementioned analysis unit, A film thickness measuring device that derives at least one of the following: a wavelength parameter corresponding to the film thickness candidate identified from the film thickness candidate and the first relationship information, which is the difference between the calculated first wavelength parameter of light and the calculated first wavelength parameter of light and the film thickness distance between the film thickness derived from the calculated first wavelength parameter of light and the first relationship information and the film thickness candidate; and derives the film thickness of the object based on the derivation results.

2. The film thickness measuring device according to claim 1, wherein the analysis unit derives the distance in the film thickness direction for each of the plurality of film thickness candidates, and estimates the film thickness candidate with the closest distance in the film thickness direction among the plurality of film thickness candidates to be the film thickness of the object.

3. The film thickness measuring device according to claim 1, wherein the analysis unit derives the wavelength distance for each of the plurality of film thickness candidates, and estimates the film thickness candidate with the closest wavelength distance among the plurality of film thickness candidates as the film thickness of the object.

4. The aforementioned optical measuring unit is A light irradiation unit that irradiates the object with first light having a first wavelength, An optical element whose transmittance and reflectance change according to wavelength in a predetermined wavelength range, and which separates the first light from the object by transmitting and reflecting it, The film thickness measuring apparatus according to claim 1, comprising: a photodetection unit that detects at least one of the first light reflected by the optical element and the first light transmitted through the optical element, and outputs the first signal relating to the calculation of the wavelength centroid, which is a wavelength parameter of the first light.

5. The light irradiation unit irradiates the first light and the second light having a second wavelength different from the first wavelength, The optical element separates the first light and the second light from the object by transmitting and reflecting them. The light detection unit detects at least one of the first light reflected by the optical element and the first light transmitted through the optical element, and outputs the first signal related to the calculation of the wavelength centroid, which is a wavelength parameter of the first light, and also detects at least one of the second light reflected by the optical element and the second light transmitted through the optical element, and outputs the second signal related to the calculation of the wavelength centroid, which is a wavelength parameter of the second light. The storage unit stores the first relational information and also stores second relational information showing the relationship between the film thickness of the object and the wavelength parameter at the second wavelength. The aforementioned analysis unit, The wavelength distance is the difference between the wavelength parameter corresponding to the film thickness candidate identified from the film thickness candidate and the first relationship information, and the calculated first wavelength parameter of light, and at least one of the film thickness distance is the difference between the film thickness derived from the calculated first wavelength parameter of light and the first relationship information, and the film thickness candidate. The wavelength distance is the difference between the wavelength parameter corresponding to the film thickness candidate identified from the film thickness candidate and the second relationship information, and the calculated second wavelength parameter of light, and the film thickness distance is the difference between the film thickness derived from the calculated second wavelength parameter of light and the second relationship information, and the film thickness candidate. The film thickness measuring device according to claim 4, which derives the film thickness of the object based on the derivation result.

6. The film thickness measuring device according to claim 5, wherein the analysis unit sets a film thickness derived from the calculated first wavelength parameter of light and the first relationship information, or a film thickness derived from the calculated second wavelength parameter of light and the second relationship information, as the candidate film thickness.

7. The film thickness measuring device according to claim 1, wherein the analysis unit sets a film thickness within a predetermined range as the candidate film thickness.

8. The film thickness measuring apparatus according to claim 1, wherein the analysis unit sets a plurality of film thickness candidates at predetermined intervals.

9. The film thickness measuring device according to claim 2 or 3, wherein the analysis unit evaluates the slope of the calibration curve at the intersection of the calibration curve showing the first relational information and the film thickness candidate for each of the plurality of film thickness candidates, and derives the film thickness of the object such that the film thickness candidate with a larger slope is more likely to be adopted as the film thickness of the object.

10. A method for measuring film thickness performed by a film thickness measuring device that measures the film thickness of an object on which a film has been formed on a substrate, A photometric step of irradiating the object with first light having a first wavelength, detecting the first light reflected by the object, and outputting a first signal related to the calculation of the wavelength parameter of the first light, The system includes a first step of calculating a first wavelength parameter of light based on the first signal, first relational information showing the relationship between the film thickness of the object at the first wavelength and the wavelength parameter, and a derivation step of deriving the film thickness of the object based on the calculated first wavelength parameter of light, In the derivation step, A method for measuring film thickness, comprising deriving at least one of the following: a wavelength parameter corresponding to the film thickness candidate identified from the film thickness candidate and the first relationship information, which is the difference between the calculated first wavelength parameter of light and the calculated first wavelength parameter of light, and a film thickness distance which is the difference between the film thickness derived from the calculated first wavelength parameter of light and the first relationship information and the film thickness candidate; and deriving the film thickness of the object based on the derivation results.

11. The method for measuring film thickness according to claim 10, wherein in the derivation step, the wavelength distance is derived for each of the plurality of film thickness candidates, and the film thickness candidate with the closest wavelength distance among the plurality of film thickness candidates is estimated to be the film thickness of the object.

12. The method for measuring film thickness according to claim 10, wherein in the derivation step, the distance in the film thickness direction is derived for each of the plurality of film thickness candidates, and the film thickness candidate with the closest distance in the film thickness direction among the plurality of film thickness candidates is estimated to be the film thickness of the object.

13. The method for measuring film thickness according to claim 10, wherein in the derivation step, the film thickness derived from the calculated first light wavelength parameter and the first relationship information is set as the candidate film thickness.

14. The method for measuring film thickness according to claim 10, wherein in the derivation step, a film thickness within a predetermined range is set as the candidate film thickness.

15. The method for measuring film thickness according to claim 10, wherein in the derivation step, a plurality of candidate film thicknesses are set at predetermined intervals.

16. The method for measuring film thickness according to any one of claims 10 to 15, wherein in the derivation step, for each of the multiple film thickness candidates, the slope of the calibration curve at the intersection of the calibration curve showing the first relational information and the film thickness candidate is evaluated, and the film thickness of the object is derived such that the film thickness candidate with a larger slope is more likely to be adopted as the film thickness of the object.

Images