KR0159954B1 - Surface condition measurement apparatus - Google Patents

Abstract

The present invention relates to an apparatus for measuring the surface state, and since radiation light having a specific wavelength is detected from a material and a first parameter corresponding to the emissivity ratio is obtained from the detection signal, and the emissivity has a different value depending on the surface state of the material, The first parameter depends on the surface condition of the material, and there is a correlation between the property value and the first parameter indicating the state of the material surface, and the correlation is that the first parameter corresponding to the property value is used instead of the property value itself. As an example, a parameter that corresponds to a property value and has an algebraic ratio (lnε _a / lnε _b ) between emissivity corresponding to a temperature near the surface, the second parameter is obtained based on the correlation and the property value is obtained. Characterized in that it can.

Description

Surface condition measuring device

1A to 1D are timing charts showing the time variation of emissivity and emissivity ratio for explaining the principles of the present invention.

2 shows a correlation between oxide film thickness and emissivity.

3 is a diagram showing the correlation between the emissivity ratio and the emissivity ratio and the oxide film thickness.

4 shows the correlation between two emissivity squared ratios obtained when three different wavelengths are used.

5 shows a correlation between two emissivity ratios obtained from the correlation between two emissivity squared ratios.

6 shows another correlation between film thickness and emissivity.

Fig. 7 is a diagram showing a correlation between an oxide film thickness having an emissivity squared ratio and an emissivity ratio obtained from the correlation.

8 is a layout view of a first embodiment of the present invention.

9 is a layout view showing a radiation sensor in a first embodiment;

10 is a block diagram of a control unit in the first embodiment.

FIG. 11 is a graph of the material showing the correlation h ₁ with the silicon product.

FIG. 12 is a graph showing the material h _a and h _b correlated with silicon work.

FIG. 13 is a graph showing h ₂ as the material correlates with silicon work.

14 is a graph showing the change of emissivity ε _a and ε _{b in} the material of silicon.

FIG. 15 is a graph showing the change in the ratio of algebraic emissivity to Ld in the material of silicon.

FIG. 16 is a graph showing h ₁ correlating to material in time.

17 is a graph showing a correlation material ever cheolil relationship h _a and h _b turns.

FIG. 18 is a graph showing h ₂ correlating to material inevitably.

19 is a graph showing the change of emissivity ε _a and ε _{b in} the material.

FIG. 20 is a graph showing the change in the ratio of algebraic emissivity to Ld.

21 is a characteristic curve of spectral sensitivity of a detection element.

22 is an arrangement diagram of a second embodiment of the present invention.

23 is an arrangement diagram showing a radiation sensor in a second embodiment;

FIG. 24 is a graph of the material showing the correlation h ₁ with the silicon product.

25 is a graph showing the material h _a and h _b correlated with silicon work.

FIG. 26 is a graph showing h ₂ as the material correlates with silicon work. FIG.

FIG. 27 is a graph showing the change of emissivity ε _a and ε _{b in} the material of silicon.

FIG. 28 is a graph showing the change in the ratio of algebraic emissivity to Ld in the material of silicon.

29 is a graph showing the correlation between material h ₁ enemy cheolil turn.

FIG. 30 is a graph showing the correlation of materials h _a and h _b with material in stock.

FIG. 31 is a graph showing the correlation of material h ₂ with iron material.

32 is a graph showing the change of emissivity ε _a and ε _{b in} the material.

FIG. 33 is a graph showing the change in the ratio of algebraic emissivity to Ld in material.

34 is a layout view of a modified embodiment of the radiation sensor.

35 is a layout view of a modified embodiment of the radiation sensor.

36 is a layout view of a modified embodiment of the radiation sensor.

37 is a layout view of a modified embodiment of the radiation sensor.

38 is a layout view of a modified embodiment of the radiation sensor.

FIG. 39 is a layout view of the filter module of the radiation sensor of FIG.

40 is a layout view of a modified embodiment of the radiation sensor.

41 is a layout view of the filter module of the modified embodiment.

42 is a block diagram schematically showing a third embodiment according to the present invention.

43 is a block diagram schematically showing a fourth embodiment according to the present invention.

44 is a block diagram schematically showing a fifth embodiment according to the present invention.

45 is a block diagram schematically showing a sixth embodiment according to the present invention.

Fig. 46 is a block diagram showing an embodiment of a two color thermometer modified to a multi-wavelength type.

Fig. 47 is a graph showing the relationship between the wavelength and the spectral emissivity in a multi-color modified bicolor thermometer.

48 is a diagram illustrating a two-color thermometer layout transformed into a polygon.

Fig. 49 shows the relationship between the two measurement wavelengths and the inclination angle in a two-color pyrometer deformed into polygons.

50A and 50B are block diagrams showing the arrangement of the wavelength domain selection blocks.

Fig. 51 is a block diagram showing the photoelectric conversion portion of the two-color multiple emission thermometer of the seventh embodiment according to the present invention.

Fig. 52 is an enlarged plan view showing a filter disc in the seventh embodiment.

53 is a diagram showing the light detection sensitivity of the optoelectronic device.

Fig. 54 is a block diagram showing a signal processing portion of the two-color multiplex radiation thermometer of the seventh embodiment.

Fig. 55 is a block diagram showing a photoelectric conversion unit of the two-color multiple emission thermometer of the eighth embodiment.

56 is a diagram showing the light detection sensitivity of another optoelectronic device.

Fig. 57 is a block diagram showing a photoelectric conversion portion of the two-color multiple emission thermometer of the ninth embodiment.

58 is a perspective view of the outside of the photoelectric conversion portion.

Fig. 59 is a timing chart showing the change in emissivity of the tricolor triple form.

60 is a table showing output timing for selecting a signal from a wavelength domain selection block.

61 is a block diagram showing the conceptual structure of a multi-wavelength radiation thermometer of a tenth embodiment according to the present invention;

62 is a diagram of a regression function F used in the tenth embodiment shown in three-dimensional space.

FIG. 63 shows a regression function G obtained from the regression function F by applying a mapping thereto. FIG.

64A and 64B are block diagrams showing a selection block structure.

65 shows emissivity spectra of a low temperature surface without an oxide film formed thereon;

FIG. 66 shows the emissivity spectrum of an intermediate temperature surface without an oxide film formed thereon; FIG.

Fig. 67 shows the emissivity spectrum of the surface directly after formation of the oxide film thereon.

FIG. 68 shows emissivity spectra of a surface having a product oxide film thereon; FIG.

FIG. 69 shows emissivity spectra of surfaces after thick passive oxide films are formed thereon; FIG.

70 shows the emissivity spectrum of a surface for two adjacent wavelengths when no oxide film is formed.

Fig. 71 shows the emissivity spectrum of the surface for two adjacent wavelengths after the oxide film is formed.

FIG. 72 shows the correlation between spectral emissivity for two wavelengths. FIG.

73 shows a problem of the prior art.

74 is another diagram illustrating a problem of the prior art.

75 is a block diagram showing the arrangement of the prior art two-color radiation thermometer.

* Explanation of symbols for main parts of the drawings

116: thermometer block 118: home block

110: 2-color thermometer 124, 24: judgment block

128: correction block 205: photoelectric element

206,208 Conversion circuit 210,220,230 Timing area

235 detection sensor 401 selector

853,861: Linearizer

BACKGROUND OF THE INVENTION Field of the Invention [0001] The present invention relates to a surface condition measuring apparatus. The present invention relates to a surface condition measuring apparatus, which detects thermal radiation energy having two or more different wavelengths. Shape properties), surface temperature, thin film thickness, and the like.

In the process of manufacturing semiconductor materials, metal materials, etc., various reactions such as oxidation, galvannealing and deposition occur artificially or spontaneously on the surface of the material, and due to these reactions, the physical properties of the material surface are significantly increased. Change.

However, sensing inside the material is very difficult and it is practically undesirable to attach several sensors to the material in order to obtain a lot of data.

As a result, the surface temperature is selected as the most influential variable from the property value or the state value, and the surface temperature is used as the predicted value in the online sensing or theoretical calculation.

For example, in the process of forming an oxide film on a silicon wafer, the correlation data between the thermal pattern into the oxide process and the oxide film properties is measured offline to obtain a large amount of offline data in advance, and the furnace temperature is adjusted based on these data. (Even if it is desirable to perform surface temperature control, indirect control according to the furnace temperature is performed due to the difficulty of online sensing.)

For on-line material surface temperature sensing, there are contact and non-contact temperature sensing techniques:

In contact temperature sensing using thermocouples, dummysters, etc., too many problems are involved, such as the limited range of temperatures that can be measured, and where the temperature can be measured is limited to the sensing point and the material May cause contamination (impurity).

As a result, the process to which the present invention can be applied is limited.

As a non-contact temperature measuring method, there is a radiation temperature measurement.

It has practical technology such as process line of metal materials such as steel and aluminum, and radiation thermometer is marketed as a product.

For process lines of metallic materials as described above, a technique for measuring surface temperature in a non-contact manner has already been developed.

However, for the process of manufacturing semiconductors or novel materials, this measurement technique is also in the research stage.

Even though surface temperature measurement technology has been developed by various efforts, no effective technology has been developed.

For example, Watanabe et al., "Measurement of Wafer Temperature in Semiconductor Heat Treatment Apparatus Using Radiation Thermometer," Journal of Automatic Control, Vol. 25, No. 9, pp925-932 (1989), conventionally reported that the manufacturing process to which the furnace temperature control is applied enables the emission temperature to be measured smoothly using an optical fiber and a prism.

In this regard, the present situation is that no effective technology has been developed for on-line radiation temperature measurement in the field of semiconductor process.

Alternatively, radiation pyrometers according to the radiation temperature measurement technique applied to measure the surface temperature of hot materials are widely used.

There are two types of radiation thermometers: monochromatic thermometers that use a single wavelength for measurement and bicolor thermometers that use two wavelengths.

Even monochromatic thermometers and even two-color thermometers that use two wavelengths produce a large measurement error when the emissivity of the measurement target changes.

More specifically, there is no problem due to the temperature measurement accuracy of the two-color thermometer, as long as the spectral emissivity for the two wavelengths is substantially the same or a constant proportion is made therebetween.

However, as the surface conditions of the high temperature material suddenly change due to the oxidation reaction caused therein and the spectral emissivity deviates from this relationship, the measurement accuracy is extremely reduced and the error of the monochromatic radiation thermometer becomes much larger.

As a result, there is a demand for a method for calculating the temperature of a two-color thermometer according to a change in spectral emissivity, and there are studies to provide such a method, and an improved type of two-color thermometer that can be used even when the emissivity changes. Is being studied.

As such a technique, `` Tanaka and DP Dewitt: Theory of a New Radiation Thermometry Method and an Experimental Study Using Galvannealed Steel Specimens '', Vol. 25, No. 10, pp. 1031. 1037, October, 1989), and the TRACE (Thermometry Reestablished by Automatic Compensation of Emissivity) method.

Since both are basically the same calculation method, they describe the former.

The spectral emissivity of the radiant energy (light waves) emitted from the materials in the process can be obtained using Wien's approximation.

The reflectance is given by the following formulas [1] and [2] when the wavelength is λ ₁ and λ ₂ .

Formula [3] can be obtained by removing temperature T from the said formula.

(The symbols used in the above formulas will be mentioned below)

Measurement wavelength: λ ₁ [㎛]

Spectral emissivity with respect to measured wavelength: ε ₁ [㎛]

Actual temperature of high temperature material surface: T [K]

Luminance temperature of high temperature material surface at wavelength λ ₁ : Si [K]

Plank secondary mapping: C2 (1.4388 × 10 ⁴ ) [㎛ · K]

The left side of Equation [3] is the ratio of the wavelength power or the power of the spectral reflectance, and is simply called the emissivity power ratio.

The former two-color radiation thermometer measures temperature under the assumption that the ratio between spectral emissivity (ε ₁ / ε ₂ ) is 1 or constant, and there is a large measurement error because it does not respond to changes in spectral emissivity. Is later referred to as emissivity ratio)

At present, in Japanese Patent Publication No. 3-4855, the correlation function f of the spectral emissivity ratio (ε ₁ / ε ₂ ) to the emissivity squared ratio is determined by measurement in advance.

Specifically, if the radiation temperature measurement and the actual temperature measurement are simultaneously performed by using a thermocouple, for example, the spectral emissivity can be obtained and the spectral emissivity ratio and the emissivity squared ratio can be obtained by using such data.

As a result, a correlation function f can be obtained from this.

Further, since the luminance temperatures S1 and S2 can be obtained at the two-wavelength detector output, the emissivity squared ratio value can be obtained by calculating according to the right side of equation [3].

Therefore, in the temperature measurement, the luminance temperature is measured and the emissivity ratio is obtained according to the following formula [4] from the emissivity squared ratio calculated according to the above formula [3] using the correlation function f and the temperature T is expressed in the following formula [5]: Obtained by calculation.

However, when the above-described prior art is applied to surface temperature measurement in silicon semiconductor surface processing such as surface thin film formation, the function f in the formula [4] becomes very complicated, difficult to calculate, and roughly simplified if there is an error. .

In the surface process of the silicon semiconductor, the surface thin film changes.

As a result, as the surface temperature is about to be measured, the emissivity is greatly changed by the optical interference in the thin film.

This makes the actual radiation temperature measurement difficult.

In the process of manufacturing semiconductor materials, metal materials, and the like, even if the surface properties of the materials in the process are measured and the online process control is subsequently performed, such adjustment is not actually performed.

In addition, since the surface properties change over time and are closely related to the material temperature, it is essential that the surface properties and the surface temperature are simultaneously measured on an online basis.

However, it has not been possible to simultaneously measure the surface properties and the surface temperature at the same position of the material.

In addition, since the measurement method by the two-color thermometer using the above formulas [1] to [5] is applied to the temperature measurement of the high temperature material whose surface state changes with the progress of the oxidation reaction, the measurement wavelength at which the temperature measurement is used is It is made very precise by making it insensitive to surface state change.

However, when the spectral emissivity of the selected wavelength is sensitive to the change in surface state, a problem arises in that the measurement accuracy is greatly reduced.

More specifically, as a specific example in which the spectral emissivity of the selected wavelength changes sensitively with the change of the surface state of the high temperature material, the surface is oxidized and a semitransparent (at the measured wavelength) oxide film is formed on the surface.

In this case, optical interference occurs in the translucent film formed on the surface and the spectral emissivity is greatly reduced.

The sudden change in emissivity was calculated according to Heat Transfer 1986, Vol. 2, Hemisphere, (1986) pp. 577 to 582. From the experimental and optical interference theory model calculations, when surface oxidation occurs, a drop of the spectral emissivity spectrum appears in a short wavelength region, and this valley As the oxidation proceeds, it is confirmed that the characteristic change is the shift of the wavelength in the long direction.

65 to 69 are diagrams schematically showing examples of such characteristic changes in the spectral emissivity spectrum.

Referring to the drawings, the abscissa represents the spectral wavelength λ, the ordinate represents the emissivity ε, and the portion indicated by the valley is the valley in the spectral emissivity spectrum.

65 to 69 show changes in the spectral emission spectrum of the surface as the oxide film is gradually formed on the metal surface such as stainless steel.

FIG. 65 shows a low temperature spectrum in which no oxide film is formed, FIG. 66 shows an intermediate temperature state in which an oxide film has not yet been formed, and FIG. 67 shows an intermediate temperature state in which an oxide film starts to form, and FIG. 68 develops thereon at the same temperature. Fig. 69 shows the state of the material in which a thick oxide is formed on it heated to a high temperature.

The occurrence of the valley is considered to be able to contribute mainly to the optical interference caused by the oxide film and the Mikino et. al. obtains the spectral emissivity spectrum through model calculation based on the interference theory and reports that the calculated and experimental results agree well with each other.

Therefore, the change in the spectral emissivity spectrum can be regarded as occurring because the radiation energy of the spectral wavelength band is selectively trapped in the oxide film to a size below the oxide film thickness.

More specifically, it is believed that the valley shifts from shorter wavelengths to longer wavelengths because the apparent energy attenuation is made because the specifically selected wavelength bands create interference or multiple reflections in the oxide film and the unusually selected wavelength band moves as the oxide film becomes thicker. .

Since the emissivity ratio varies over time as the spectral emissivity spectrum is described above, the measurement error occurs in a spherical bicolor radiation thermometer and the improved form bicolor thermometer described above because of the difficulty of calculating the equation where the measurement error is used. Even in the same way.

The reason for this is that the improved form dichroism calculations using adjacent two wavelengths λ ₁ and λ _1x (λ ₁ λ _1x ) are based on the off-line and the two spectral emissivity ε ₁ and ε _1x as regression functions in advance from experimental data. While the correlation coefficient needs to be determined, the regression procedure is difficult.

This will be described briefly below.

_{Given that} the actual measured data of spectral emissivity ε ₁ and ε _1x are obtained while the valley shift from short wavelength side to long wavelength side occurs as in the spectral emissivity spectrum described above, the correlation of emissivity ε ₁ and ε _1x is positive. Change from correlation to injury.

It will be readily understood if the change in the values of the emissivity ε ₁ and ε _1x corresponding to two adjacent wavelengths λ ₁ and λ _1x is shown in FIGS. 65 to 69.

More specifically, the magnitude of the correlation between the spectral emissivity ε ₁ and ε _1x is reversed and the correlation between the short-wavelength portions of the valley in the diagram (the portion where the spectroscopic gradient is negative) is between wavelengths λ ₁ and λ _1x. Reverses.

It is understood that the situation is clearly presented in FIGS. 70 and 71 and that before and after the valley pass (70) and after the valley pass (71) is clearly completely inversely correlated.

Also in Tanaka et al., 3rd Paper Research No. 339 (1990) pp. 63-67, it is suggested that the graph ε ₁ vs. ε ₂ is not a single value but a loop is formed as schematically shown in FIG.

As a result, as described above, since the correlation regression graph between the emissivity ε ₁ and ε _1x cannot be simply determined, there is a problem that some measurement errors occur even in an improved dichroic thermometer.

As described above, the surface of the stainless steel sheet (SUS 304) can be measured with the improved two-color thermometer as described above. In the alloying temperature range of 600 ° C, the maximum measurement error is about 15 ° C and the standard deviation is about 5 ° C. to be.

During the surface oxidation or surface alloying and thus the change in surface properties as described above, temperature is an important process parameter of the steel manufacturing process control.

As a result, it is very important that there is an error in the measurement obtained by radiation temperature measurement, and as a result, a more accurate thermometer is required.

Temperature measurement error is ± 5 ° C (maximum error is within 5 ° C) is the preferred measurement accuracy.

Therefore, there is a need for the development of a radiation thermometer with higher measurement accuracy than the schematic two-color thermometer.

As a result, in order to solve the above-described difficulties of the improved dichroic thermometer, three or more wavelengths are used to measure and a plurality of two wavelength combinations are formed from them, and for each combination of two wavelengths to measure temperature. It is conceivable to carry out the same process as the improved two-color thermometer.

However, this method of using three or more measurement wavelengths and increasing the number of couplings of two wavelengths also has the following problems.

For simplicity, there are three measurement wavelengths: λ ₁ , λ ₂ and λ ₃ (λ ₁ λ ₂ λ ₃ ) and the two wavelength combinations are (λ ₂ , λ ₃ ) and (λ ₁ , λ ₂ ) for convenience. The case where it is made will be considered.

From luminance temperatures S1, S2 and S3 measured for wavelengths λ ₁ , λ ₂ and λ ₃ , two combinations (S2, S3) and (S1, S2) corresponding to the combination of wavelengths are made and the emissivity-ratio ε ₃ ^{λ 3} / ε ₂ ^{λ 2} and ε ₂ ^{λ 2} / ε ₁ ^{λ 1} are calculated using the equations corresponding to the above-mentioned equation [3] and the correlation functions f1 and f2 corresponding to the above-mentioned equations [4] are raised. It is applied to the ratio and emissivity ratios ε ₃ / ε ₂ and ε ₂ / ε ₁ are obtained.

However, if the function f1 is given according to the graph shown in Fig. 73 because the oxide film is formed on the surface of the object to be measured, the emissivity ratio is the emissivity-ratio ratio obtained from the equation corresponding to Equation [3] when A0 is The emissivity ratio is taken as two values A1 and A2.

Alternatively, if the function f2 is given according to the graph shown in Fig. 74, three values B1, B2 and B3 are obtained with the emissivity ratio corresponding to the calculated value of emissivity squared ratio B0.

As a result, the equation corresponding to the above-mentioned equation [5] has five points, that is, two emissivity ratios A1, A2 and three emissivity ratios B1-as long as the measurement wavelength is simply increased and applied to the above-described improved two-color thermometer. It should be used to calculate the temperature for B3 and should be investigated for temperatures consistent with each other from the temperatures obtained from A1 and A2 and from those obtained from B1 to B3, resulting in the determination of the corresponding temperature to the surface measurement temperature. do.

Therefore, there is a problem that the calculation method for the survey is complicated.

As a prior art example of radiation temperature measurement, there is one described in USP 4417822, in which a laser is additionally used and the reflectance of the laser beam from the surface of the measurement object is used to compensate for the emissivity.

This method has the disadvantage that the device is complicated and expensive because the laser is adopted.

It is also required to obtain offline data for reflectance measurements for the laser beam.

However, the data includes complex elements related to light scattering on the surface.

As a result, it is doubtful whether offline data can be used in online measurements.

In addition, since the error occurs in the on-line measurement of the reflectance, the error in the temperature measurement is large.

USP 4561786 describes radiation temperature measurement techniques using a device as shown in FIG.

In the above technique, the radiation wave is converged by the lens 813 and separated by the rotary filter 815 to obtain the two-color measurement wavelength.

The output for the two-color wavelength is detected by the detector 811 and a second calculation is made on the output to obtain the ratio and difference therebetween.

By using this calculated value, a temperature calculation is performed for practical use.

Referring to FIG. 75, reference numeral 851 denotes a division block for making a division between the outputs W1 and W2 of the sample and for supporting the circuits 845 and 847, ie its output constitutes a ratio.

Reference numeral 859 denotes a differential amplifier for performing subtraction between sample outputs and for performing circuits 855 and 857, ie their outputs are different.

Reference numerals S1 to S4 denote timing signals for synchronizing the rotation positions of the rotary filter 815 having the control system, 817 denotes a spectral filter 819 for the first wavelength, and 841 and 843 for spectral filters for the second wavelength. Denotes an amplifier.

For each output value from the split block 851 and the differential amplifier 859, the linearization networks 853 and 861 and the resistance values R1, R2, R3 and R4 are set after adjusting according to the respective measuring purposes.

The calculated temperature value is displayed on the meter 875.

In this way, the adjustment of the linearizers 853 and 861 and the resistors R1 to R4 are measured by trial and error.

Therefore, it is a calculation method that appears as a result of trial and error and is not based on theory.

In fact, no theoretical explanation is given.

According to the method, it is described that the temperature measurement of the aluminum phase with the surface change proceeding on it can be made very precisely (± 5 ° C).

However, to achieve this high accuracy, it is assumed that a long time trial and error is made (setting data is not described).

Since the device using the method described above does not have any theoretical basis, long trial and error is required to measure materials other than aluminum.

As a result, there is a problem of not having general applicability.

From the perspective of those who control the manufacturing process, such a system is desirable in that not only surface temperature is measured but also changes in other surface properties are monitored (detected), and process control is performed through feedback control to these monitored values.

Surface properties of the material are not only direct, such as surface roughness, surface reflectivity (emissivity), surface absorption rate and refractive index, but also within the material such as oxide film thickness, alloying film thickness, deposited film thickness, electrical conductivity and boundary (interlayer) refractive index. Indirectly affected by physical properties.

More specifically, the surface properties are determined by the properties of the material produced after the reaction is performed near the surface and the properties related to the interface between the material and the substrate.

Rather, it may be said that surface properties cannot even be assumed from the direct physical properties of the surface itself.

Of course it is also greatly influenced by the conditions of the materials in the process such as the surface temperature, the internal temperature and their distribution.

As a result, it is desirable that not only surface properties but also changes in physical properties and state changes in materials can be detected when the surface properties of the material in the processor are accurately obtained.

For example, in the above-described oxide film forming process on a silicon wafer, surface physical property control is indirectly performed by controlling the furnace temperature while blind control (no online sensor) is performed according to the actual conditions of the oxide film properties. It will be.

Naturally, statistical process errors occur.

Due to the difficulty of detection, such blind control is widespread through common semiconductor processes and is responsible for the low yield of semiconductors.

This situation is seen not only in the process semiconductor field, but also in the field of processing high-tech materials such as ceramics and superconducting materials.

In order to improve this blind control, which indirectly controls the properties by temperature and consequently increases the yield in the process, it is urgent to have an on-line monitor (sensing) capable of measuring the change in the surface properties of the material in the process.

The present invention has been made to solve the problems described above in the prior art.

The first object of the present invention is to obtain a physical property indicating the temperature by a simple method which does not require surface conditions and iterative calculation of the material.

The second object of the present invention is to measure the physical properties and surface temperature of the material at the same measurement position at the same time and to measure the physical properties as well as the surface temperature very accurately, even when the surface properties change and the emissivity changes significantly, and only the physical properties are measured. Makes it possible.

A third object of the present invention is to provide a practical two-color multiple-type radiation thermometer that can measure temperature very accurately even if the surface conditions of the object to be measured are changed by oxidation or the like and the emissivity is greatly changed as a result.

The fourth object of the present invention is to use three or more measurement wavelengths, and even if the surface conditions of the measurement target vary with time and the emissivity varies depending on the wavelength used, the actual surface temperature of the target can be easily calculated and output. It is to provide a wavelength radiation thermometer.

In order to solve the first problem, the present invention provides a surface condition measuring apparatus comprising: detecting means for detecting light of a specific wavelength emitted from a material and outputting a plurality of detection signals; Calculating means for calculating a first parameter corresponding to an emissivity ratio from the plurality of detection signals; And converting means for converting the first parameter into the second parameter according to the correlation between the second parameter and the first parameter indicating the surface state of the material.

The detection means may be arranged to include an optical beam having the same wavelength and forming a different angle (measurement angle) between its optical paths and a plurality of radiation sensors for detecting a condition on the surface of the material.

The detection means may be arranged to detect that the polarization angle is different from the light beam emitted from the material.

The detection means may be arranged to include polarization means for separating light beams having different polarization angles and a plurality of detection elements from light emitted from the material, each detecting each light beam having a different polarization angle.

The second parameter can be an emissivity ratio or an algebraic emissivity ratio and the calculation means can also be arranged to obtain the temperature near the material surface from the second parameter and the detection signal.

The second parameter may be an oxide film thickness formed on the material surface.

The detecting means may be arranged to detect light beams having a particular angle of polarization from the light beams emitted from the material and the light beams emitted from the material.

In the surface condition measuring apparatus of the present invention, an optical wave having a specific wavelength γ radiated from a material is detected by the detecting means and the detection signal and the emissivity ratio (ε _a / ε _b ; where ε _a and ε _b have a wavelength λ). For the first parameter (e.g. emissivity ratio, detection wavelength power of emissivity ratio (λ power of ε _a / ε _b ), corresponding to two different detection wavelengths Difference between the reciprocal of the signal values), from which an output is obtained from the detection signal by the detection means.

Since emissivity is obtained at different values depending on the surface conditions of the material, the first parameter depends on the surface conditions of the material.

There is a correlation between the first parameter and the physical properties indicating the condition of the material surface.

The correlation is used instead of the physical properties themselves (e.g. optical properties such as reflectance and absorptivity, film thickness formed on the material surface, surface roughness and alloying degree).

As an example of a parameter corresponding to the physical property value, there is an algebraic emissivity ratio (ln ε _a / ln ε _b ) corresponding to the temperature near the surface.

In the changing means, the second parameter is obtained based on the correlation and the physical properties are obtained.

When the emissivity or logarithmic emissivity ratio is to be used as the second parameter, the temperature near the material surface can be easily obtained from the second parameter and the detection signal.

According to the surface condition measuring apparatus as described above by obtaining the second parameter using the correlation between the first parameter and the second parameter corresponding to the property value, which is an indication of the material surface state, the physical properties corresponding to the second parameter are obtained. It is possible to obtain physical property values indicating the material surface state or temperature by a simple method without repeated calculation.

In the present invention, the second problem mentioned above is one or more two-color thermometers having a common field of view at any position on the material surface in the process,

Formula

Where C2 is the flank second constant

A two-color thermometer line obtained two wavelengths λ ₁ and a power bigyesan means for calculating an emissivity of a power ratio ε ₁ ^λ1 / ε ₂ ^λ2 from the brightness temperatures S1 and S2 for the λ ₂ according, to the measured and pre-built-correlation function f

The first emissivity ratio calculation means for converting the emissivity-ratio ratio ε ₁ ^λ1 / ε ₂ ^λ2 into the first spectral emissivity ratio ε ₁ / ε ₂ , with the initial value in the property value and calculating the second spectral emissivity ratio ε ₁ ^* / ε ₂ ^* A second emissivity ratio calculation means for separately calculating the spectral emissivity ε ₁ ^* / ε ₂ ^* for two wavelengths λ ₁ and λ ₂ having a theoretical or empirical formula having one or more physical properties which greatly change the change in emissivity as a variable, Irradiation calculation means for irradiating the properties having a second spectral emissivity calculated using properties substantially coincident with the first spectral emissivity ratio ε ₁ / ε ₂ and the first spectral emissivity ratio ε ₁ / ε ₂ Or by having a surface condition measuring device composed of thermometer calculation means for calculating a two-color temperature using a second spectral emissivity ratio ε ₁ ^* / ε ₂ ^* .

In the present invention, the second problem mentioned above is one or more two-color thermometers having a common field of view on the material surface in the process,

Formula

Where C2 is the flank second constant

The first power ratio calculation means for calculating the first emissivity power ratio ε ₁ ^λ1 / ε ₂ ^λ2 from the luminance temperatures S1 and S2 for two wavelengths λ ₁ and λ ₂ obtained from on-line two-color thermometer according to the second emissivity power ε ₁ ^* Spectral emissivity ε ₁ ^* and ε ₂ for two wavelengths λ ₁ and λ ₂ using theoretical or empirical formulas with one or more properties that significantly affect the change in emissivity as parameters to be calculated from the ^λ1 / ε ₂ ^{* λ2} calculation results second exponentiation calculating a ^* independently ratio calculation means, and giving an initial value to a physical data first emissivity a power ratio ε ₁ ^λ1 / ε ₂ ^λ2 and substantially matching the second emissivity a power ratio ε ₁ ^{* λ1} / ε properties with ₂ ^{* λ2} that It is also solved by having a surface condition measuring device composed of irradiation calculation means for irradiation calculation for.

The apparatus may have two or more dichroic thermometers and a thermometer selecting means for selecting one or more dichroic thermometers from them over time with the emissivity ratio in each of the two color thermometers.

In addition, when at least one of the physical properties is the oxide film thickness d, the theoretical formula may be as follows:

In the present invention, the emissivity of a power ratio ε ₁ ^λ1 / ε ₂ ^λ2 is obtained from the brightness temperatures S1 and S2 obtained by the two-color thermometer first spectral emissivity ratio ε ₁ / ε ₂ If is obtained from a power ratio the second spectral emissivity ratio ε ₁ ^* / ε ₂ ^* is obtained from the physical properties and the irradiation calculation is arranged to examine the physical properties having the second spectral emissivity ratio substantially coincident with the first spectral emissivity ratio.

As a result, it becomes possible to obtain the physical properties of the material in the process during the measurement, and also to calculate the surface temperature by substituting the first emissivity ratio or the last second spectral emissivity ratio by the dichroic temperature calculation formula corresponding to Equation [5].

As a result, it is possible to simultaneously measure both the surface temperature and the physical properties of the material in the process.

In such a case, the corresponding number of physical property measurements can be made by increasing the dichroic temperature coefficient.

Further, in the present invention, the second emissivity multiplication ratio ε ₁ ^{* λ 1} / ε ₂ ^{* λ 2} is obtained from the physical properties and the irradiation calculation has a second emissivity multiplier ratio substantially equal to the first emissivity power ratio obtained from the luminance temperatures S1 and S2. Since it is arranged to be made for irradiation, it is possible to measure the physical properties of the material in the process and to measure the surface temperature of the material.

Further, in the present invention, the time change of the spectral emissivity ratio or the emissivity squared ratio may be monitored, and the emissivity ratio obtained by using a measurement band which does not exhibit the change in emissivity with respect to the monitored time change may be adopted.

As a result, it becomes possible to measure temperature very accurately.

In the present invention, spectroscopic means for dispersing light at two different wavelengths λ ₁ and λ ₂ for a plurality of different wavelength regions, respectively,

Photoelectric conversion means for measuring luminance temperatures S1 and S2 for two wavelengths λ ₁ and λ ₂ for each wavelength region,

Formula

Where C2 is the flank second constant

To a power calculating a two-line wavelength emissivity from the measured brightness temperatures S1 and S2, a power ratio ε ₁ ^λ1 / ε ₂ ^λ2 for each wavelength range according ratio calculation means,

Measured and prebuilt built-in correlation function f

Emissivity ratio calculation means for converting the emissivity power ratio ε ₁ λ ¹ / ε ₂ ^{λ 2} into the spectral emissivity ratio ε ₁ / ε ₂ using

Thermometric means for calculating the two-color temperature using the spectral emissivity ratio ε ₁ / ε ₂ for calculating the measured temperature, and monitoring the time-dependent change in the emissivity-ratio or spectral emissivity ratio for each wavelength region. The third problem can be solved by selecting a special wavelength region, and having a two-color multiple-type radiation thermometer composed of a wavelength region selecting means for outputting the measured temperature calculated for the selected special wavelength region.

Further, in the apparatus, the wavelength region selecting means selects a wavelength region which is suitable for comparing the time change rate of the emissivity squared ratio or the spectral emissivity ratio and has a minimum time change rate.

Further, in the apparatus, the wavelength region selecting means selects a wavelength region which is suitable for comparing the absolute value of the emissivity squared ratio or the spectral emissivity ratio and whose absolute value is minimum.

Further, in the apparatus, when three or more wavelength ranges are used, the wavelength range selection means may be the same in the second wavelength, the lower wavelength direction, or the higher wavelength direction from the wavelength region where the time-varying emissivity ratio or spectral emissivity ratio is maximum. It is suitable for selecting the wavelength range underneath.

First, the principle of the present invention will be explained.

In the following description, two spectral wavelengths are denoted by λ _i and λ _ix (λ _i λ _ix ) and the luminance temperature at the wavelength is denoted by Si and Six and the corresponding emissivity is ε _i and ε _ix (i = 1 , 2, ...).

The letter i represents a different wavelength range.

In the improved two-color thermometer as described above, the emissivity-ratio is obtained from luminance temperatures S1 and S2 according to Equation [3], and the emissivity-ratio is the correlation between the emissivity-ratio and the two spectral emissivity ε ₁ and ε _1x The resultant temperature where the relationship is set by the regression function based on the experimental data. The resultant temperature T where the emissivity ratio ε ₁ / ε _1x is processed by the function [4] obtained as a function value and the emissivity ratio is replaced by the two-color temperature calculation formula [5] Is calculated.

However, as the surface oxide film reaches its limit thickness, the emissivity ratio varies greatly due to the interference of radiation waves.

When this change occurs, the regression function f becomes a function, for example, which results in problems with the calculation.

The present invention relates to a time domain in which such a change in emissivity ratio occurs on the basis of experiments and theories, which is to be used to compensate for the time domain in which this time domain thermometer result is excluded from the results and the temperature measurement results obtained in other time domains are excluded. It is assumed that the temperature can be measured very accurately even when a change in the emissivity ratio occurs.

1 is a diagram schematically showing a change in spectral emissivity based on the above experiment and theory.

FIG. 1 (a) shows the time variation of the spectral emissivity ε ₁ and ε _{1x '} and FIG. 1 (b) shows the time variation of the emissivity ratio ε ₁ / ε _1x' and FIG. 1 (c) shows the spectral emissivity ε ₂ and The time change of ε _{2x ′} is shown, and the first (d) diagram shows the time change of the emissivity ratio ε ₂ / ε _2x .

Where spectral wavelengths have the same relationship as λ ₁ λ _1x λ ₂ λ _{2x '} and λ ₁ and λ _1x' as well as λ ₂ and λ _{2x '} are adjacent wavelengths and two adjacent wavelengths are measured bands (wavelength regions) To form.

Specifically, Si and Ge may be set to be λ ₁ = 1.00 μm, and λ _1x = 1.05 μm, λ ₂ = 1.60 μm, and λ _2x = 1.65 μm, as appropriate for the radiation photoelectric device.

The spectral emissivity ε ₁ and ε _1x are for the adjacent wavelengths λ ₁ and λ _1x and as a result, they exhibit substantially the same time change.

However, when a small change in physical properties such as surface oxidation occurs, the emissivity is sensitively changed by the small change.

FIG. 1 (a) is a diagram showing an embodiment in which surface oxidation is in progress.

The interference effect of the radiation wave due to the oxide film appears earlier and disappears earlier for the smaller process λ ₁ .

Thus, epsilon ₁ and epsilon _1x result in a slightly shifted shift graph in the time axis direction as shown in the figure, resulting in a timing of epsilon ₁ = epsilon _1x as described by A and A 'in the figure.

Since the emissivity ratio takes a value of 1 at the A and A 'timings and the emissivity changes significantly before and after timing, the emissivity ratio also varies greatly.

For two wavelengths λ ₂ and λ _2x that form a separate measurement band from the measurement band formed by λ ₁ and λ _{1x '} , the slightly similar delayed from the changes in the first (a) and the first (b) degrees Changes are observed and shown in Figures 1 (c) and 1 (d).

The interval between the timing at which the sudden change in the emissivity occurs in the first (a) and the first (b) degrees and the timing at which the sudden change in the emissivity occurs in the first (c) and the first (d) degrees is 1.00 μm. It depends on the interval between the timing at which is formed and the timing at which the oxide film develops to a thickness of about 1.60 mu m.

While this time interval depends on the rate of oxidation, it is only a few seconds when the stainless steel sheet is oxidized in air.

In the improved two-color thermometer of the prior art, the emissivity ratio is shown in FIGS. 1 (b) and 1 (d) before or after the timing region Z2 or timing region Z10 or timing region Z20 in which the timing region Z1 emissivity ratio is greater than one. As shown, it varies greatly and there is a problem that the measurement error becomes large when the temperature calculation is performed using the emissivity ratio ε ₁ / ε _1x or ε ₂ / ε _2x .

In the present invention, ε ₂ / ε _2x which does not exhibit a change is used in the timing region Z1 or the timing region Z10 and ε ₁ / ε _{1x which} is already indicative of change and is now stabilized is the emissivity ratio for the temperature calculation. Suitable for use in Z20.

The equations corresponding to the formulas [3] to [5] used for the thermometer calculation using the measurement band are as follows:

In the present invention, as described in detail, it is possible to apply a two-color radiation temperature measuring method using a measuring band formed of two wavelengths in a plurality of wavelength ranges.

As a result, a wavelength region in which the valley in the spectral emissivity spectrum does not exist even when the valley moves from the short wavelength to the long wavelength side due to oxide film growth over time, for example, as shown in FIGS. 65 to 69. It is possible to select the measuring band on the phase.

As a result, since a temperature measurement with a two-color thermometer can be made using a measuring band which does not always cause any change in emissivity, very accurate temperature measurements can be made by the surface state of the object being measured, e. As the thickness increases over time, it even becomes possible over time.

Spectral means for dispersing light at three or more different wavelengths λ ₁ and λ _{n '} in the present invention, photoelectric conversion means for measuring luminance temperatures S1 to Sn at respective wavelengths λ ₁ to λ _n' ,

Formula

(Where i, j is a positive integer that does not exceed n, i and does not match j, C2 is the flank second constant)

A power-ratio calculation means for calculating the emissivity power-ratio ε _i ^λi / ε _j ^λj obtained online by each combination of two different wavelengths according to, representing the interrelationship of the emissivity power-ratio ε _i ^λi / ε _i ^λj ( n-1) Converting the (n-1) dimensional emissivity ratio regression function G and the power-divisor ratio regression function F, which express the interrelationship between the dimensional square rate regression function F and the emissivity ratio ε _i / ε _j , to the emissivity ratio regression function G Reference means incorporating mapping H previously made based on off-line measurement data or theory, and emissivity power ratio regression corresponding to (n-1) dimensional coordinates formed by the (n-1) emissivity power ratio calculated from the above equation Determination means using the luminance temperatures S1 to Sn obtained online for determining points on the function F,

Conversion means for applying a mapping H to a point determined on the power-ratio regression function F to convert the point to a point on the emissivity ratio regression function G, and calculating the (n-1) emissivity ratio ε _i / ε _j from the conversion point, and The fourth problem can be solved by having a multi-wavelength radiation thermometer composed of thermometer calculation means for calculating the measured temperature by using a portion or all of the calculated emissivity ratios.

In addition, the device is monitored such that the change in emissivity ratio or emissivity ratio is monitored and the emissivity exponential ratio or emissivity ratio showing a large change is excluded from the temperature calculation.

In addition, in the device the time change of the emissivity-ratio or absolute ratio of emissivity ratios is monitored and the emissivity-ratio or emissivity ratio indicating a large change is excluded from the temperature calculation.

First, the basic principle of the present invention will be explained.

2 shows the correlation between the thickness d of the oxide film formed on the silicon wafer and the emissivity ε ₁ with respect to the measurement wavelength λ _i (i is an integer) from the following equation [6] or [7].

Equations [6] and [7] are well suited for thin oxide films formed on silicon surfaces, for example by Watanabe et al. 25, No. 9, pp925-931 (1989).

From here,

ρ _a : reflectance at the boundary between the oxide film and air (= 0.034)

ρ _b : reflectance at the boundary between the oxide film and the non-oxidized part (= 0.186)

θ: Angle between measuring device and measuring surface state (= 0)

n: refractive index of oxide film (= 1.45)

2 shows a graph of three emissivity ε ₁ , ε ₂ and ε ₃ corresponding to wavelengths λ ₁ , λ ₂ and λ ₃ used to calculate.

Emissivity ε ₁ is shown by a solid line, the broken line ε _2, ε ₃ are indicated by the chain line.

The wavelengths used are λ ₁ = 1.0 μm, λ ₂ = 2.0 μm and λ ₃ = 4.0 μm with an integer ratio therebetween.

Currently, emissivity rises for two combinations of two wavelengths and emissivity ratios ε ₂ / ε ₁ and ε ₃ / ε ₁ , using the relationship between the oxide film thickness d and the emissivity ε ₁ to ε ₃ for each wavelength. The ratio ε ₂ ^{λ 2} / ε ₁ ^{λ 1} and ε ₃ ^{λ 3} / ε ₁ ^{λ 1} are obtained.

If the ratio obtained is plotted in relation to the film thickness d, we can obtain a third degree.

In FIG. 3, the relationship for the coupling of the wavelengths λ ₁ and λ ₂ is drawn in solid lines and the relationship for the coupling of the wavelengths λ ₁ and λ ₃ is drawn in broken lines.

If the correlation between the emissivity ratios themselves and the correlation between the emissivity ratios themselves are indicated in the two-dimensional coordinate plane, we obtain FIGS. 4 and 5, respectively.

Functions F and G are one-to-one correspondence, as the functions represented by the FIG. 4 graph are denoted by F and the functions represented by the FIG. 5 graph are denoted by G. FIG.

By preparing the mapping H for converting the function F into the function G in advance, applying the mapping H to a specific point on the function F makes it easier to calculate the point corresponding to the function G.

As a result, by having the functions F and G and the mapping H prepared in advance, the two emissivity multiplication ratios obtained by substituting the luminance temperatures S1, S2 and S3 obtained as actual measurements with the equations corresponding to equation [3] are identified. If the value is obtained as point P (X, Y) with a specific value as the coordinate on function F, then point Q (X, Y) on function G of FIG. 4 obtained by applying mapping H to point P Can be obtained with the corresponding specific point.

As a result, it is possible to obtain two emissivity ratios ε ₂ / ε ₁ and ε ₃ / ε ₁ from the values of the coordinates (X, Y) of the point Q, and then the two emissivity ratios according to the equations corresponding to equation [5]. By using a specific value of, it is possible to obtain two temperatures T1 and T2 on the same measurement point.

Since two sets of temperature information are obtained by one measurement, a very accurate measurement of the surface temperature can be made by determining the average of the two temperatures T1 and T2 to be the actual temperature T.

As the number of measured wavelengths simply increases, the correlation between the emissivity ratio and the emissivity ratio is complicated, and as shown in Figs. 72 and 73, the number of emissivity ratios corresponding to one emissivity ratio is shown.

As a result, it is essential to investigate the actual temperature.

According to the invention, however, the corresponding emissivity ratio can be reliably obtained only by applying the mapping without making complicated irradiation calculations.

As a result, the actual temperature can be easily and quickly obtained.

In addition, since the information obtainable is larger than that obtained in the prior art improved bicolor thermometer as described above, the surface temperature can be measured more accurately.

Although the principles of the present invention are described above on the basis of the ideal case where the correlation between the emissivity ε and the film thickness d is obtained from equations [6] and [7], the present invention is based on theory or quasi-theory. It is not limited.

When theoretical processing is not applicable, a function F expressing the correlation between the emissivity ratios, a function G expressing the correlation between the emissivity ratios and a mapping H that converts the function F into a function G are for example measured. The object (silicon wafer, aluminum, stainless steel sheet, etc.) is offline in a pre-regressive manner.

Currently, examples of how to determine the functions F and G as regression functions will be conveniently described using FIGS. 4 and 5, respectively.

The luminance temperatures S1, S2 and S3 for the three wavelengths λ ₁ , λ ₂ and λ ₃ are actually measured for the number of objects (samples) having different oxide film thicknesses d.

When using the actually measured luminance temperatures S1, S2 and S3, an emissivity squared ratio based on the actual measurement is obtained for each sample according to the equation corresponding to equation [3].

The ratios are plotted on the XY coordinate plane where ε ₂ ^{λ 2} / ε ₁ ^{λ 1} is taken as the X axis and ε ₃ ^{λ 3} / ε ₁ ^{λ 1} is taken as the Y axis.

The regression function F is obtained as the mean value of these calibration points, as the emissivity squared ratio actually measured on several samples is obtained with the black points shown in FIG.

Although not indicated, the emissivity ratios corresponding to the two emissivity squared ratios are obtained in advance by the actual measurement offline and the values are taken with ε ₂ / ε ₁ taken on the X axis and ε ₃ ε ₁ taken on the Y axis. The regression function G, plotted on the XY coordinate plane and corresponding to the regression function F, is obtained equally as the mean value of the plot shown in FIG.

In addition, a mapping H for converting the regression function F into a regression function G is prepared.

As a result, the functions and the mapping are preempted.

As the regression function F is obtained as shown in Fig. 4, the value of the emissivity squared ratio actually obtained at the time of actual temperature measurement is generally slightly different from the regression function F.

As a result, the emissivity-ratio according to the actual measurement is obtained at point P 'and the point P on the function F at the shortest distance from the point P' is determined as a function of the temperature calculation.

As a result, it is determined that the point P (X, Y) on the regression function F represents the actual emissivity power ratio, while the emissivity power ratio value according to the actual measurement will be obtained as the point P '(X', Y '). By applying the mapping H to the point P, the point Q (X, Y) is obtained on the regression function G of the emissivity ratio, and also corresponds to the emissivity ratio ε ₂ / ε ₁ and the coordinate Y corresponding to the coordinate X of the point Q. The emissivity ratio ε ₃ / ε ₁ can be easily obtained.

Thereafter, two surface temperatures T1 and T2 can be obtained by substituting two emissivity ratios in a manner corresponding to equation [5].

As long as two surface temperatures T1 and T2 are obtained, their average value can be determined as the actual temperature T as described above.

If not, the temperature obtained from the emissivity-raising ratio or emissivity ratio unaffected by an unexpected change in the emissivity can be taken as the actual temperature T, while the two emissivity-ratios or the time variation of the emissivity ratios are arranged to be monitored. .

The ideal case where the emissivity ratio can be obtained from equations [6] and [7] and additionally the three wavelengths used for measurement have an integer ratio between them is described as above.

When the measurement wavelength has an integer ratio therebetween, it exhibits a simple periodicity among the three emissivity as shown in FIG. 2 and as a result a simple period also appears in the emissivity power-ratio and emissivity ratio as shown in FIG.

As a result, the advantage of easier handling can be obtained.

However, it is not necessary that the wavelength used for the measurement necessarily have an integer ratio therebetween.

For example, at λ ₁ = 1.5 μm, λ ₂ = 2.0 μm and λ ₃ = 4.0 μm, no simple periodicity is observed among the three emissivity as indicated in FIG. 6 (corresponding to FIG. 2). As a result, no periodicity is observed in the emissivity squared ratio and the emissivity ratio as shown in FIG. 7 (corresponding to FIG. 3).

Even if no constant ratio is present in the measurement wavelength as described above, by tracking the correlation curve of the two emissivity ratios and the two emissivity ratios and having the results embedded in the computer, Of course, this may be the case.

Although the measurement wavelengths are described above with respect to the case where the measurement wavelengths are three different wavelengths and both the emissivity power ratio and the emissivity ratio can be displayed in the two-dimensional coordinate plane, the present invention uses four or more wavelengths and the emissivity power ratio and the emissivity ratio are three-dimensional. Even if displayed in the above coordinate space can be applied.

The computational process for such multi-wavelength arrays can be facilitated by a computer.

Preferred embodiments of the present invention will be described below with reference to the accompanying drawings.

Corresponding elements relating to the prior art described above will be briefly described or omitted.

8 shows a material surface state measuring apparatus as a first embodiment of the present invention.

The apparatus has radiation sensors 16a and 16b having the same measurement wavelength λ and the measuring angles θ _a and θ _b of the radiation sensors 16a and 16b (on the light path of the detection light of the radiation sensors 16a and 16b and on the material surface). The angle formed between the normals is adapted to be appropriately changed according to different materials 21 (objects to be measured) or different processes in the process.

Each part of the device will first be described below.

The radiation sensors 16a, 16b are disposed in this relative position in the material 21 in the process, such that they have different measurement angles θ _a and θ _b as shown in FIG. 9.

The radiation sensor detects an output luminance temperature at wavelength λ as radiation energy of the radiation at wavelength λ and detection signals Sa and Sb from material 21 in the process.

The radiation sensor is a detection element such as Si (detection wavelength λ ₁ = 1 μm), Ge (detection wavelength λ ₂ = 1.6 μm), PbS (detection wavelength λ ₃ = 2 μm), PbSe (detection wavelength λ ₄ = 4 μm), etc. Use

The control unit 22 calculates the surface property value or temperature near the surface of the material 21 in the process from the luminance temperatures Sa and S and exchanges data D _in and D _out containing the measurement results with the external circuit.

The control unit 22 has a parameter calculation block 31, a correlation block 22 and a thermometer calculation block 26 as shown in FIG.

The parameter calculation block 31 is the wavelength power of the emissivity ratio from the luminance temperatures Sa and Sb (λ power of ε _a / ε _b ; where ε _a and ε _b represent the measurement angles θ _a and θ _b )

Here, the wavelength power of the emissivity ratio is used as the parameter Ku, and is obtained by calculating Equation [8].

The correlation block 22 provides the correlation H between the parameter Ku and the surface property of the material in the process and provides the surface conditions of the material 21 in the process 21 on the material surface from the parameter Ku, ie optical properties such as reflectance and absorbance, film. Quantities relating to thickness, surface roughness, alloying process diagrams, or parameters related to the calculations are obtained.

The correlation H is given by a conversion table or by algebraic calculation according to experimental or theoretical calculations.

Examples given by theoretical calculations will be presented.

FIG. 11 shows the correlation h ₁ between the parameter Ku and the logarithmic emissivity ratio (ln ε _a / ln ε _b ) Ld and FIG. 12 shows the correlation h _a and h _b between the parameter Ku and emissivity ε _a and ε _{b '} . 13 shows the correlation h ₂ between the parameter Ku and the oxide film thickness d.

The data relate to the case where it is used as the material 21 for processing silicon, and the measurement wavelength λ is set to 1.6 mu m and the measurement angles θ _a and θ _b are set to θ degrees and 45 degrees, respectively.

For reference, a change in the emissivity ε _a and ε _b (FIG. 14) and a logarithmic emissivity ratio Ld (FIG. 15) are shown in relation to the oxide film thickness d.

Ever be used as material 21 of iron is processed correlation between (the other conditions are same as before;), a parameter correlated between Ku and logarithmic emissivity ratio Ld h _1, parameters Ku and emissivity ε _a and ε _b h _a and h The correlation h ₂ between _b and the parameter Ku and the oxide film thickness d is shown in FIGS. 16, 17 and 18, respectively.

The changes in the emissivity ε _a and ε _{b and} the change in the logarithmic emissivity ratio Ld in relation to the oxide film thickness d are shown in FIGS. 19 and 20.

The thermometer block 26 obtains the temperature T near the surface of the material 21 in the process from the detection wavelength λ and luminance temperature Sa (or luminance temperature Sb) using the emissivity ε _a or ε _b or the logarithmic emissivity ratio Ld.

When the emissivity ε _a or ε _b is used, the temperature T is obtained by calculation of equation [9a] or [9b] and the temperature T is obtained by calculation of equation [10] when the logarithmic emissivity ratio Ld is used.

The operation of the apparatus according to the first embodiment is explained next.

As described above, the spectral emissivity ε of the material 21 in the process depends on the surface conditions.

For example, as the oxide film is formed on the surface, the light beam from the material 21 in the process interferes with each other due to reflection between the object to be measured and the oxide film, reflection between the oxide film and air, and the spectral emissivity ε It depends on the reflectance of the material, the oxide film thickness d, the refractive index of the complex oxide film, the wavelength, the emission angle of the light (that is, the measurement angle), the polarization angle, and the like.

This is the reason for the error that occurred within the monochromatic and bicolor radiation thermometers as described above.

Emissivity ratios also vary.

The spectral emissivity ε can be expressed as a function of the wavelength λ and the measurement angle or polarization angle θ as indicated by the following formula [11].

Assuming that the material 21 in the process is silicon and the measured wavelength λ is 1.6 μm, the changes in the emissivity ε _a and ε _b detected by the radiation sensors 16a, 16b are presented for example in FIG. .

As the detection signals from the radiation sensors 16a and 16b, the luminance temperatures Sa and Sb are outputs to the control unit 22 and the parameter Ku is calculated from the luminance temperatures Sa and Sb by the parameter calculation block 31.

The parameter Ku represents the wavelength power of the emissivity ratio.

By substituting the detection wavelength [lambda] into the formulas [1] and [2] described above, equations [12a] and [12b] are obtained.

By removing the temperature T from the above, equation [13] is obtained.

If ε _a = ε _b , the value of equation [13] is 1.

However, since the measurement angles θ _a and θ _b are different from 0 degrees and 45 degrees, and the emissivity ε _a and ε _b are different values, the parameter Ku is not 1 but changes in the surface conditions of the material 21 in the process. Depends.

In the correlation block 22, the surface property value is obtained from the parameter Ku, that is, the oxide film thickness d is obtained using the correlation h ₂ (Fig. 13), and the logarithmic emissivity ratio Ld is obtained from the correlation h ₁ (11th). Ε _a and ε _b are obtained using the correlations h _a and h _b (FIG. 12).

Another surface property can be obtained using the correlation H for the property.

In the temperature calculation block 26, the temperature T is calculated from equation [9a] or [9b] or equation [10] made by using the emissivity ε _a or ε _b or the logarithmic emissivity ratio Ld obtained in the correlation block 22. Is obtained through.

The above formulas are obtained by modifying formulas [12a] and [12b].

When iron is to be used as the material 21 to be processed, those indicated in FIGS. 16-18 are used as correlation h ₁ , correlation h _a and h _b and correlation h ₂ .

In-process material (21) by having the same operation as in the case where silicon is used except that the change in the emissivity ε _a and ε _{b and} the change in the logarithmic emissivity ratio Ld are different as shown in FIGS. 19 and 20. The temperature T near), the oxide film thickness d and other physical properties can be obtained.

As described above, in the present invention, the physical properties or parameters related to the physical properties (such as the logarithmic ratio Ld) can be obtained by using the correlation between the parameter Ku and the material surface properties without repeating the calculation, resulting in a high speed process. Can be.

In other words, the operation is quick and the measurement can be completed in a short time.

Since the radiation sensors 16a and 16b have the same wavelength lambda, the temperature measuring range is wider.

For example, short wavelength thermometers with silicon detectors can measure temperatures up to 600 ° C and those with germanium detectors can measure temperatures up to 400 ° C.

However, two-wavelength thermometers using the two detection elements can only measure temperatures up to 400 ° C.

When the two-color thermometer has the same measurement wavelength λ, its temperature measurement range is not lowered by the use of a device having a lower temperature measurement range.

Adjacent two-wavelength thermometers use detection elements that match one another and detect by using two wavelengths near the center wavelength of the detection element (see FIG. 21).

For example, there is one that uses a germanium detection element and has 1.55 μm and 1.65 μm detection wavelengths.

With this thermometer, there is no problem as described above because this type of thermometer uses the same detection element but its sensitivity is lower and its accuracy is reduced.

The sensitivity of the short wavelength thermometer is obtained by partial differentiation of the equation (formula [14]) for obtaining the temperature.

The sensitivity of a two-wavelength thermometer can be obtained in a similar way, but is roughly given by equation [15].

Although a decrease in sensitivity occurs in a two-wavelength thermometer, there is no decrease in sensitivity in this embodiment because it is measured at the same wavelength λ.

Currently, a second embodiment of the present invention will be described.

The second embodiment can be made in view of the fact that the spectral emissivity also varies with this polarization angle.

This embodiment is arranged to measure the surface condition by detecting light beams having different polarization angles instead of different measurement angles from the previous embodiment.

Its overall arrangement is shown in FIG.

The radiation sensor 16c of the device is arranged such that its measured value θ _c is 45 degrees.

Light beams having a wavelength lambda (1.6 mu m) and different polarization angles (P polarization beams and S polarization beams) are detected and output as detection signals Ss and Sp.

23 shows its detailed structure.

Light from the in-process material 21 in the radiation sensor 16c is split by a half mirror (beam splitter) H.

One of the split light beams is passed through the S polarization filter Fs and detected by the detection element Ds output therefrom as the detection signal Ss.

The other of the split light beams is detected by the detection element Dp that passes through the P polarization filter Fp and is output therefrom as the detection signal Sp.

The control unit 22 used in the above embodiment is the same as that used in the above-described embodiment (Fig. 10), except that the detection signals Ss and Sp are used instead of the luminance temperatures Sa and Sb, so that instead of Sa and Sb, Is used in the formula where Ss and Sp are used as elements in the control unit 22.

The correlation block 22 also uses the correlation h ₁ , the correlation h _s and h _p as the material 21 to be processed is a logarithmic emissivity ratio Ld, the emissivity ε _s and ε _p and the oxide film thickness. d is obtained and correlation h ₂ is shown in FIGS. 24 to 26 instead of correlation h ₁ , correlation h _a and h _b and correlation h ₂ is used in the previous embodiment.

Correlation with different surface properties depends on different values of the parameter Ku.

27 and 28 show changes in the emissivity ε _s and ε _p and change in the logarithmic emissivity ratio Ld with respect to the oxide film thickness d.

The material 21 to be processed is ironic (other conditions are the same), the logarithmic emissivity ratio is Ld, the emissivity is ε _s and ε _p and the oxide film thickness d is the correlation h ₁ , the correlation h _s and h _p . Obtained and the correlation h ₂ is as shown in FIGS. 29 to 31.

The correlation with other surface properties is similarly different.

32 and 33 show changes in the emissivity ε _s and ε _p and change in the logarithmic emissivity ratio Ld associated with the oxide film thickness d.

In the apparatus of the present invention, the temperature near the surface of the material 21 in the process, the oxide film thickness d and other surface property values are obtained through the same operation as in the previous embodiment, and only different values of the parameter Ku are used.

The advantages obtained from the apparatus of the present invention are also the same as those obtained in the previous embodiment.

The present invention is not limited to the above described embodiment, and various modifications are possible.

For example, even if the radiation sensor 16c is arranged as shown in FIG. 23 in the second embodiment, the optical fiber OP can be used as shown in FIGS. 34-37.

In the figure, the same as in FIG. 23, H denotes a beam splitter (half mirror), Ds and Dp denote detection elements for two polarization angles, and Fs and Fp denote polarization filters.

The polarization filters Fs and Fp can be removed if the polarization plane maintenance filter is used in the optical fiber OP.

In this case, it becomes possible to change the measurement angle at the time of arranging one part of the fixed radiation sensor 16c and another movable part.

Further, by arranging interchangeable polarization filters Fs and Fp as shown in FIG. 38, two detection elements can be reduced to one detection element D. FIG.

As shown in FIG. 39, the filter module Fx is equipped with the S polarizing filter Fs and the P polarizing filter Fp, and the polarizing filters Fs and Fp are exchanged as the module Fx rotates by the motor M. As shown in FIG.

At this time, the detection signals Ss and Sp are alternately outputted and the parameter Ku is obtained from the signal.

It is also possible to combine the luminance signal So which does not depend on the polarization angle and the polarization luminance signal Ss or Sp.

In this case, referring to FIG. 40, the sensor can be made by placing one or more of the four filters Fs and Fp as they are and leaving the others as simple holes.

Since the luminance signal So includes light waves having a polarization angle other than the polarization angle of the polarization luminance signal, it becomes the same as obtained from the above-described embodiment.

Further, as shown in FIG. 41, the filter module Fx is additionally provided with the spectral filters Fλ ₁ and Fλ ₂ such that different wavelengths Sλ ₁ and Sλ ₂ are obtained.

As a result, the device has a hybrid configuration with a two-wavelength thermometer.

In general, the detection element has a spectral sensitivity as shown in FIG. 21, and the two wavelength signals Sλ ₁ and Sλ ₂ and the polarized light signals Ss and Sp are obtained by the spectral filters Fλ ₁ and Fλ ₂ .

Even if the wavelength power of the emissivity ratio is used as the parameter Ku so far, the same is performed if other mathematically equivalent parameters are used.

Table 1 shows some examples (Tables B and C) obtained by modifying both sides of Formula [8] (Table A).

For example, the difference between the inverses of the luminance signals shown in Table C, 1 / Sa-1 / Sb, may be used.

42 is a block diagram schematically showing a third embodiment according to the present invention.

The measuring device of the embodiment of the present invention can measure both the surface temperature T and the oxide film thickness d formed on the material surface in the process at the same time, and the measuring device is equipped with a two-color thermometer and the thickness d of the oxide film as a property value. Measure

The measuring device is a two-color thermometer (110) for measuring the surface temperature P of the material in the process, the emissivity power ratio ε according to the formula [3] / ε Emissivity-ratio ratio calculation block 112 using the luminance signals S1 and S2 input from the two-color thermometer 110 for calculating the equation, and a calculation block 112 for calculating the first emissivity ratio ε / ε according to Equation [4]. Thermometric calculation using the first emissivity ratio ε / ε from the first emissivity ratio calculation block 114 using the power-ratio ratio input from the calculation block 114 and the calculation block 114 for calculating the surface temperature T according to Equation [5]. Block 116.

The measuring device further has an emissivity ε with an assumption of the thickness of the assumption d. And ε Emissivity calculation block 120 for inputting two measurement wavelengths λ and λ for calculating a two-color thermometer / ε It has a division block 122 that uses the emissivity from the calculation block 120 to calculate.

The emissivity calculation block 120 is configured to calculate two emissivity using the following equations [16] and [17].

ρ _a : reflectance at the boundary between the oxide film and air,

ρ _b : reflectance at the boundary between the oxide film and the non-oxidized part,

θ: refraction angle at the boundary between the oxide film and the non-oxide portion,

n: refractive index of the oxide film, and

d: The thickness of an oxide film is shown.

In addition, the measuring device uses the first emissivity ratio ε ₁ / ε ₂ and the second emissivity ratio ε ₁ ^* / ε ₂ ^* respectively inputted from the first emissivity ratio calculation block 114 and the summing block 122. At this time, it is possible to make a decision using the following formula [18] of the evaluation function in the following least square method with the decision block 24 for determining whether or not the film thickness d is output.

More specifically, when H is substantially zero, the decision block 124 outputs the film thickness d currently used to calculate the second emissivity ratio ε ₁ ^* / ε ₂ ^* as the film thickness d of the material P in the process. If H is not 0, the correction value? D calculated in the correction value calculation block 126 and (d + Δ) made in the correction block 128 are corrected, and the hypothesized block d is set such that the corrected thickness d is set. After the calibration input, set the thickness d.

In the present embodiment, when the radiant energy from the material P in the process is measured by the two-color thermometer 110, the current luminance temperatures S1 and S2 are input to the power-ratio ratio calculation block 112 where the calculation of equation [3] is performed. The calculation result is input to the first emissivity ratio calculation block 114 in which the calculation of Equation [4] is performed, and the calculation result is input to the thermometer calculation block 116 in which the calculation of Equation [5] is performed. The surface temperature T of the material P in the processor measured by 110 is output.

On the other hand, in the film thickness measurement and the calculation part on the right side of the drawing, first, the arbitrary value d is set as the hypothetical value (initial value), and the value d having the wavelengths λ ₁ and λ ₂ together is calculated by the formulas [16] and [17]. Is received by the emissivity calculation block 120 and the result is input to the summing block 122 where division is performed, from which the second emissivity ratio ε ₁ ^* / ε ₂ ^* is output to the determination block 124.

The decision block 124 accepts the second emissivity ratio together with the first emissivity ratio from the first emissivity ratio calculation block 114 and calculates the values by substituting the above values into Equation [18] to execute the determination.

When the determination result is NO, the correction value Δd is calculated in the correction value calculation block 126, and the correction for the film thickness d using the correction value Δd is performed in the correction block 128 and assumes the film thickness d for which correction has been performed. Set at (118).

The procedure is repeated according to the above procedure and the irradiation calculation is made until the value H is substantially zero in the decision block 124.

When the value H becomes substantially zero, the film thickness d used in the last calculation is output.

According to the embodiment of the present invention, the surface temperature of the material P in the process and the oxide film thickness d formed on the material surface in the process can be measured substantially simultaneously.

43 is a block diagram showing a measuring device of a fourth embodiment according to the present invention.

The measuring apparatus according to the embodiment of the present invention has a first power-ratio ratio calculation block 112A (42 degrees power-increase-ratio calculation block 112) for receiving luminance temperatures S1 and S2 for calculating the first emissivity power-ratio ratio ε ₁ λ ¹ / ε ₂ ^{λ 2} . ) And using the emissivity square term calculated by the calculation block 130 for calculating the two emissivity square terms and the second emissivity square rate ε ₁ ^* λ ¹ / ε ₂ ^* λ 2 122A) has an emissivity squared term calculating block 130 using emissivity ε ₁ ^* and ε ₂ ^* inputs from the emissivity calculation block 120 and corresponding two wavelengths λ ₁ and λ ₂ .

The first and second power ratios are input to the determination block 124, and a determination according to the determination formula [18a] is executed in the determination block 124.

In other words, the fourth embodiment is substantially the same as the third embodiment.

In the embodiment of the present invention, the first emissivity ratio calculation block 114 and the thermometer calculation block 116 need not be necessarily provided.

According to the embodiment of the present invention, since the film thickness d is determined according to the first and second emissivity squared ratios and the value determined to be exactly the film thickness is output, the surface temperature T and the oxide film thickness d of the material P in the process are substantially It can be measured simultaneously with the third embodiment.

Further, only the oxide film thickness d can be arranged so that the first emissivity ratio calculation block 114 and the thermometer calculation block 116 are not provided and are measured.

44 is a block diagram schematically showing a measuring device of a fifth embodiment according to the present invention.

The embodiment of the present invention is provided with three dichroic thermometers 110A, 110B and 110C and is formed in a multi-wavelength type.

The basic function is substantially the same as in the third embodiment except that it has three two-color thermometers, and as a result, the points and the differences in the third embodiment will be mainly described next.

The dichroic thermometers 110A, 110B and 110C are arranged to use different measuring bands, ie two adjacent wavelengths λ ₁ and λ _1x , λ ₂ and λ _2x , λ ₃ and λ _3x and the associated luminance temperatures S1 and S1x, S2 and S2x and S3 and S3x are measured.

A power ratio calculation block (112) is the formula for the two wavelengths adjacent said three respectively to calculate the emissivity of a power ratio _{^{ε 1 λ1 / ε 1 λ1,}} ε 2 λ2 / ε 2 λ2 , and ε ₃ ^λ3 / ε ₃ ^λ3 Use [3].

A first emissivity ratio calculation block 114 is arranged to calculate the emissivity ratios ε ₁ / ε _1x , ε ₂ / ε _2x, and ε ₃ / ε _3x from the three power-ratios.

Thus, each calculation block 114 has a function f given in equation [4] above for each measurement band.

The thermometer block 116 calculates the surface temperatures T1, T2 and T3 measured by the two-color thermometers 110A, 110B and 110C from the three emissivity ratios calculated by the calculation block 114, respectively.

The surface temperatures T1 to T3 are blocked by thermometer selector block 132 in which an average of three surface temperatures is also obtained or one of the temperatures closest to the other is selected by a number of crystals and the resulting temperature is It is output as the surface temperature T of the material P in the process.

In the hypothesis block 118, hypotheses (initial values) of the oxide film thickness d and the reflectances ρ _a and ρ _b are obtained.

The three hypotheses and the three adjacent wavelengths λ _i and λ _ix (i = 1,2,3) are input to the emissivity calculation block 120 where the emissivity ε _i ^* and ε _ix ^* are represented by the equation [16]. ] And [17].

Thereafter, the emissivity ratio ε _i / ε _ix in the shunt block 122 is calculated from three series of emissivity.

In addition, the three emissivity ratios are input to the decision block 124, and a decision according to the following equation [18b] for the decision is performed in the decision block 124.

In the correction calculation block 126, the Jacobian approximation is obtained from the change in H obtained by slightly changing the values of d, ρ _a and ρ _b and in the correction block 128, d, ρ _a and ρ _b are The correction values are corrected according to the correction values for calculating the nonlinear function equations by the least square method such as the New Uton method and the Marquardt method and the correction values are presented in the hypothesis block 118.

According to an embodiment of the present invention, the reflectances ρ _a and ρ _b can be measured together with three physical properties, namely the oxide film thickness d and the surface temperature of the material P in the process.

45 is a block diagram schematically showing a measuring apparatus of a sixth embodiment according to the present invention.

The measuring device of the embodiment of the present invention is arranged such that the dichroic thermometer 110 is positioned at the inclination angles of the normals and Q ₁ , Q ₂ and Q _{3 on} the material P in the process for measuring on the material in the process.

Since the apparatus has substantially the same structure as the third embodiment, different points of the apparatus will be described below.

The emissivity-ratio ratio calculation block 112 receives luminance signals S1 and S1x together with the signals of each Q _i (i = 1, 2, 3), and calculates the radiation- _evolution ratios ε _{1θ i λ} ¹ / ε _1xθ ⁱ _λ ^1x .

The first emissivity ratio calculation block 114 calculates three emissivity ratios ε _{1θ i} / ε 1 _{× θ i} from the emissivity _squared ratio.

The three emissivity ratios are input to the decision block 124 and also to the temperature calculation block 116.

Each calculation block 114 has a function f for each tilt angle.

The temperature calculation block 116 performs a temperature calculation for each inclination angle and the thermometer selector block 132 selects different values selected by a number of crystals as the average value of the temperatures obtained for the three inclination angles or the surface temperature of the material P in the process. Output one of the temperatures close to

In the assumption block 118, the same as in the fifth embodiment, the assumption value of the oxide film thickness d and the reflectances ρ _a and ρ _b are arranged.

Within the emissivity calculation block 120, the emissivity ε _{1θ i} ^* / ε _{1xθ i} ^* is calculated using the hypothesized two adjacent wavelengths λ ₁ and λ _1x and the inclination angle θ _i according to equations [16] and [17]. .

In the summing block 122, the emissivity ratio ε _{1θ i} / ε 1 _{× θ i} is calculated from the emissivity and outputs the emissivity ratio for each tilt angle to the decision block 124.

In the determination block 124, the determination is performed in accordance with the following formula (18c), and while H becomes substantially zero, the oxide film thickness d and the reflectances ρ _a and ρ _b are output.

When H is not zero, the correction values Δd, Δρ _a and Δρ _b are calculated in the correction value calculation block 126 and the values presented in the correction block 128 are corrected as in the fifth embodiment.

According to the embodiment of the present invention, the surface temperature T, the oxide film thickness d and the reflectance of the material P in the process can be measured in the same type as in the fifth embodiment using the same type of color thermometer 110.

As shown in FIG. 46, the present invention provides a multi-wavelength arrangement for placing the dichroic thermometer 110 to perform measurements with four or more measurement bands λ _i and λ _ix (i denotes an integer of 1 to 4 or more). Can be changed to type.

In this multi-wavelength arrangement the two wavelengths of each band can be selected as shown in FIG.

This multi-wavelength arrangement allows a much larger number of properties to be measured simultaneously.

48 shows the arrangement of the dichroic thermometer as the dichroic thermometer 110 is changed to have a polygonal structure.

FIG. 49 is a diagram showing the relationship between two measurement wavelengths [lambda] ₁ and [lambda] _x and tilt angles in a polygonal arrangement as shown in FIG.

When two or more dichroic thermometers are used, one or two dichroic thermometers or more can be selected according to the spectral emissivity or the time-varying emissivity-ratio for each dichroic thermometer. And a thermometer selector block 132 as shown.

This selection method will be described below in detail for this case where the thermometer has two, first and second measuring bands.

^{Measurement of} emissivity-ratio ε ₁ ^{λ 1} / ε _1x ^λ _1x and ε ₂ ^{λ 2} / ε _2x ^λ _2x as the specific wavelength region, i.e., the measurement band, creates a smaller time variation of the emissivity-ratio The temperature T calculated for the band is chosen according to the change in emissivity power ratio.

The wavelength range selection block 132 may be in the arrangement as described in FIG. 50a or 50b.

The wavelength range selection set out in 50a a block 132 calculates block for differentiating the ε ₁ ^λ1 / ε _1x ^λ1x that for outputting a selection signal according to the time variation of the emissivity of a power ratio ^{_{(132B), ε 2 λ2 /}} ε A calculation block 132B for differentiating _2x ^lambda _2x and a comparator 132C for comparing a calculation result in blocks 132A and 132B and outputting a signal for selecting a measurement band giving a smaller change rate.

The selected wavelength regions set out in 50b a block 132 is ε ₁ calculation blocks for calculating an absolute value of ^{_{λ1 / ε 1x λ1x (132D)}} , calculation for calculating the absolute value of ε ₂ ^λ2 / ε _2x ^λ2x block ( 132E) and a calculation block 132E for outputting a signal to compare the calculation results in blocks 132D and 132E and to select a measurement band that gives the smaller of the calculation results.

The wavelength range selection block 132 function will be described below with reference to FIG.

As the emissivity is lowered due to the interference of the radiation wave, the valley is made to be ε ₁ → ε _1x → ε ₂ → ε _2x according to the relative size of the wavelength.

The decrease in emissivity is not known by the prior art two-color thermometer.

As a result, as long as the surface condition of the measurement target is related to the timing region Z1 or Z10 in the first diagram, the temperature calculation result calculated using the emissivity ε ₂ and ε _2x obtained for the second measurement band can be arranged to be selectively output. And inversely corresponding to the timing region Z2 or Z20 can be arranged to selectively output the temperature calculation results calculated using the emissivity ε ₁ and ε _1x obtained for the first measurement band.

As described in detail, the temperature calculation can be performed independently for two adjacent wavelengths λ ₁ and λ _1x of the first measurement band and two adjacent wavelengths λ ₂ and λ _2x of the second measurement band and additionally each The calculated temperature can optionally be output for the measurement band, which results in a smaller time variation of the emissivity squared ratio for the two wavelengths.

As a result, a suitable measuring band can be selected for the temperature calculation and the temperature can be output even when the oxide film is formed on the surface and the film thickness increases over time, so that the temperature measurement is always made very accurately.

The theoretical or empirical formulas representing the relationship between the surface properties and the emissivity ε _i are not limited to the above formulas [16] and [17], and any other relation can of course be used.

For example, the function of the following equation can be used, while the surface luminance is represented by R and any other physical property is represented by Xi.

As a specific embodiment to which the above formula [19] can be applied, the degree of alloying in the process of the galvanized steel sheet may be mentioned.

In the alloy zinc plating process of the alloy zinc steel sheet, the surface luminance R changes and the emissivity also changes with the alloying process.

As a result, the following formula [19] is applied when the degree of alloying is to be represented by X1.

Therefore, by using two or more two-color thermometers or two-color thermometers arranged at two or more inclination angles and applying data to the above formula [19a], the surface temperature T and the surface luminance R and the alloying degree X1, which are the surface properties T, are simultaneously obtained. It is also possible to measure.

According to an embodiment of the invention, the physical properties and surface temperature of the material in the process can be measured simultaneously as described above.

In addition, even when the surface properties change, the surface conditions of the materials in the process change, the emissivity varies greatly, the surface temperature can be measured very accurately, and the properties of the materials in the process can be measured.

Only physical properties can also be measured.

FIG. 51 is a block diagram of a portion showing a photoelectric conversion portion, which is a main part of the two-color multiple emission temperature of the seventh embodiment according to the present invention, FIG. 52 is an enlarged plan view showing an infrared transmission filter disc, and FIG. 54 is a block diagram schematically illustrating a signal process part.

The two-color multiple type radiation thermometer of the embodiment of the present invention is a two-color dual type having a photoelectric conversion portion shown in FIG. 51 and a signal process portion shown in FIG.

The photoelectric conversion portion includes a lens 202 for converging radiation waves emitted from the measurement target (thermal radiating material) 201, a first measurement band formed with wavelengths λ ₁ and λ _1x , and wavelengths λ ₂ and λ ₂ and λ _2x . Infrared transmission filter disc 203 for splitting the radiation wave converged by the lens 202 into the formed second measurement band, and photoelectric conversion of the radiation wave of the first measurement band split by the filter disc 205 A photoelectric device 203, a conversion circuit 206 for converting a signal from the photoelectric device 205 to the luminance temperature signals S1 and S1x, a photoelectric device 207 and a photoelectric device for photoelectric conversion of the radiation wave of the second measurement band. A conversion circuit 208 for converting the signal from 207 to the luminance temperature signals S2 and S2x.

The filter disc 203 has two series of lambda ₁ permeation filters 231, lambda _1x permeation filters 231x and lambda ₂ permeation filters 232 and lambda _2x permeation arranged at equal intervals in the order mentioned in FIG. A filter 232x is provided.

Each wavelength is performed twice each time the filter disc 203 is rotated by the motor 233.

Around the filter disc 203, there are protrusions 234 arranged at regular intervals so that the rotational position of the filter can be detected.

The timing signal is generated by the timing signal generating circuit 236 when the projection 234 is detected by the filter rotation position detecting sensor 235.

As a result of the timing signal being supplied to the conversion circuits 206 and 208, the radiation reception operation of the conversion circuits 206 and 208 is controlled.

The optoelectronic device 205 has two adjacent wavelengths as shown in the original Fig. 53 while the first measurement band formed by the wavelengths λ ₁ and λ _1x and the second measurement band formed by the wavelengths λ ₂ and λ _2x are sufficiently spaced apart. Photodetection element 207 has photodetection sensitivity at both λ ₁ and λ _1x and photoelectric element 207 has photodetection sensitivity at both two adjacent wavelengths λ ₂ and λ _2x as shown on the right side of the same drawing.

The signal processor shown in FIG. 54 outputs luminance temperatures S1 (corresponding to λ ₁ ) and S1x (corresponding to λ _1x ) from the conversion circuit 206 for calculating the emissivity squared ratio according to Equation [3-1]. Calculation block 301 of Equation [3-1] using the equation, an experimentally determined correlation regression equation, that is, emissivity ratio for calculating the emissivity ratio ε ₁ / ε _1x according to Equation [5-1]. Calculation block of equation [5-1] using equation, calculation block 401 and conversion of equation [4-1] using calculated emissivity ratio for calculating temperature T according to two-color temperature calculation equation [4-1]. Calculation blocks 302 of equation [3-2], namely calculation blocks that allow the circuit 208 to calculate the above similarly for the luminance temperature S2 (corresponding to λ ₂ ) and S2x (corresponding to λ _2x ) outputs. ), The calculation block 502 of equation [5-2] and the calculation block 402 of equation [4-2].

The temperature T output from both calculation blocks 401 and 402 is input to selector 410 and the temperature T calculated for either wavelength of either of the measurement bands is output therefrom.

Emissivity- ^raising ratios ε ₁ ^{λ 1} / ε _1x ^λ _1x and ε ₂ ^{λ 2} / ε _2x ^λ _2x are output from the calculation blocks 301 and 302 respectively input to the wavelength region selection block 132 and a specific wavelength region, i.e., the measurement band, It is selected over time.

The selection signal is input to the selector 401 described above and replaced with a temperature T calculated for the measurement band making a lower time variation of the emissivity power ratio and the temperature T is output.

The wavelength range selection block 132 is of the same arrangement as shown in FIG. 50A or 50B mentioned above.

In the embodiment of the present invention, when the first measurement band is used, an emissivity ratio is obtained in the calculation block 301 using the luminance temperatures S1 and Sx, and an emissivity ratio ε ₁ / ε _1x is a calculation block using the emissivity power ratio. Obtained within 301.

Similarly, when the second measurement band is used, the emissivity-ratio is obtained in the calculation block 302 using the luminance temperatures S2 and S2x and the calculation block 302 in which the emissivity ratio ε ₂ / ε _2x uses the emissivity squared ratio Is obtained within.

As inferred from FIG. 1, the timing region in which the spectral emissivity suddenly changes due to the oxide film formation can be sufficiently and reliably detected by observing the time change of the two emissivity ratios.

Specifically, it can be detected from the time derivative of the emissivity power ratio as shown in FIG. 50a or the absolute value of the emissivity power ratio as shown in FIG. 50b.

As a result, the temperature calculation result calculated by using the data obtained for the second measurement band as long as the surface condition of the measurement object is related to the timing region Z1 or Z10 in the first degree can be selected by the output selector 410. And, conversely, the temperature calculation result calculated by using the data obtained for the first measurement band, as it relates to the timing region Z2 or Z20, can be selected identically to the output.

As described in detail above, according to the two-color multiple radiation type thermometer of the present invention, independent of two adjacent wavelengths λ ₁ and λ _1x of the first measurement band and two adjacent wavelengths λ ₂ and λ _2x of the second measurement band. The temperature calculation is then performed and additionally, the calculated temperature can be selectively output for the measurement band, which results in a smaller time variation of the emissivity ratio for each of the two wavelengths.

As a result, a suitable measuring band can be selected for the temperature calculation and for example, the temperature can be output even when an oxide film is formed on the surface to be measured and produced over time to increase the film thickness.

As a result, the temperature measurement can always be made very accurately.

FIG. 55 is a block diagram corresponding to FIG. 51 and schematically shows a photoelectric conversion portion of the eighth embodiment according to the present invention, and FIG. 56 is a diagram showing the light detection sensitivity of the photoelectric element used in the eighth embodiment.

In an embodiment of the present invention, an optoelectronic device having a photodetection characteristic sensitive to four wavelengths λ ₁ , λ _1x , λ ₂ and λ _2x as indicated by reference numeral 209 of FIG. 55 and shown in FIG. 56 is provided. Instead of the conversion circuit blocks 206 and 208, a conversion circuit block 210 for outputting the luminance temperature signals S1, S1x, S2 and S2x, which is used instead of 205 and 207, and which corresponds to the photodetection signal of four wavelengths, is used.

If not, the structure is substantially the same as that of the seventh embodiment.

FIG. 57 is a block diagram corresponding to FIG. 51 and schematically showing a photoelectric conversion portion of a ninth embodiment according to the present invention.

The two-color multiple type radiation thermometer of the present invention is a two-color triple type having a photoelectric conversion unit as shown in the perspective view of FIG.

The ability to measure the temperature for the third measuring band formed by two adjacent wavelengths [lambda] ₃ and [lambda] _3x is obtained by modifying the bicolor dual emission thermometer of the seventh embodiment.

More specifically, the photoelectric conversion section (shown above in the block diagram) for photoelectrically converting the two wavelengths λ ₁ and λ _1x of the first measurement band of the photoelectric conversion section of the present invention and outputting the luminance temperature signals S1 and S1x; The photoelectric conversion section of the seventh embodiment has a similar function, that is, converts two wavelengths λ ₂ and λ _2x of the second measurement band and outputs the luminance temperature signals S ₂ and S _2x , And a photoelectric conversion section for photoelectrically converting the two wavelengths λ ₃ and λ _3x of the two measurement bands and outputting the luminance temperatures S3 and S3x by the conversion circuit block 208 '.

It is preferable that the photoelectric conversion section is arranged in a double eye type as shown in FIG. 58, but may have a filter added to the seventh or eighth embodiment.

The unmarked signal processing portion is obtained by adding the process function for the third measuring band to the signal processing portion in the seventh embodiment so that the measurement data can be selected from the data for the measuring band.

The present invention has a higher measurement accuracy than that of the two-color dual type such as the seventh embodiment.

More specifically, in the case of the two-color dual type, the timing domain detection of the wavelength domain selection block 132 cannot be made until the change in emissivity actually occurs, and as a result, a considerable amount of time that causes a large error is spent in the exchange of the wavelength domain. .

In the case of the two-color triple type of the present invention, since the detection of the timing region can be made in the wavelength region not related to the signal output, the consumption of error generation time in the exchange can be prevented.

The situation related to the above will be described in detail with reference to FIGS. 59 and 60.

59 shows three separate wavelength ranges, a first measuring band of λ ₁ and λ _1x , a second measuring band of λ ₂ and λ _2x , and a third measuring band of λ ₃ and λ _3x at each wavelength. A time chart indicating the occurrence of a change in emissivity.

Fig. 60 is a display of the selection signal output from the wavelength range selection block 132, where? Indicates a selected wavelength,? Indicates a wavelength region where a change can occur at any moment, and x indicates a wavelength region that is not selected.

As shown in the table, during the timing region 210, the third measurement bands λ ₃ and λ _{3x, which} are two regions in front of the wavelength region where a change currently occurs, are selected.

During the timing region 220, the first measurement band of λ ₁ and λ _1x that has already changed is selected and either one of the first and second measurement bands is selected during the timing region 230.

As a result, according to the present invention two-color triple type (including quadruple type or more), a more accurate temperature measurement is made than a result in which the wavelength region indicated by Δ is not selected.

Incidentally, in the case of the two-color dual type, a temperature calculation in the state of the wavelength region indicated by Δ may occur.

At present, the specific case where the present invention is actually applied will be described.

The two-color dual-type radiation thermometer takes offline data on a steel plate (manufactured by Kawasaki Seidetsu Co., Ltd.) according to Japanese Patent Publication No. 3-4855 described above, and then the first two-color measuring band and the second two-color measuring band are The structure is created by making the manufactured regression function.

1st measuring band: (Si photoelectric element)

λ ₁ : 1.00 μm

λ _1x : 1.05㎛

Second Measuring Band: (Ge Optoelectronic Device)

λ ₂ : 1.60 μm

λ _2x : 1.65 μm

In making the regression function, the method described in Japanese Patent Application No. 3-129828, which is already applied to a patent by the present applicant, is used.

More specifically, assuming that one measurement band is formed with two wavelengths λ ₁ and λ ₂ , the ratio between the spectral emissivity ε ₁ and ε ₂ with respect to the wavelengths λ ₁ and λ ₂ is the wavelength λ ₁ and λ ₂ Is calculated according to the following equation using the luminance temperatures S1 and S2 and the constant A (0 ≦ A ≦ 1).

Temperature measurements according to the invention are carried out using the regression data and similar measurements are also carried out for comparison with various types of thermometers and the results obtained (maximum temperature error) are shown in Tables 2 and 3.

The object to be measured is stainless steel SUS 304.

Table 2 shows the change in emissivity while surface oxidation is within the process and the change in emissivity is apparent, while Table 3 shows the measurement results in the temperature range where the change in Table 2 is relatively less than the change in which the measurements in Table 2 are performed.

The actual temperature is measured with a thermocouple welded to the steel sheet.

Samples are taken where ε change is apparent by qualitatively determining the surface oxidation in the process according to the trend of change of ε calculated using the actual temperature.

As is apparent from Table 2 and Table 3, in the case of the present invention, the temperature measurement error is within 5 ° C and its excellence is confirmed, depending on whether the emissivity change is apparent.

It is anticipated that more precisely if dichroism is used.

Even if the invention is specifically as described above, it is of course important that the invention is not limited to the embodiments described above.

For example, the number of measurement bands is not limited to two or three but may be four or more.

Further, as mentioned in the above described embodiment, the measurement band is formed with two adjacent wavelengths and the measurement band is separated, but other arrangements are also possible.

That is to say, the measurement bands do not have to be adjacent to one another and as a result they may even partially overlap.

For example, in the above-described embodiment, the emissivity ratio? / ε And ε / ε Is input from the calculation blocks 501 and 502, and the wavelength region selection can be arranged in accordance with the time variation of the emissivity ratio.

Further, in the above-described embodiment, although the temperature calculation is made for all measurement bands and the calculation result for the selected measurement band is output in the selection box 132, the temperature calculation can be made only for the selected measurement band. Can be arranged.

According to the embodiment of the present invention, as described above, for example, when an oxide film is formed on the surface to be measured and its thickness is increased so that surface conditions change, the surface temperature of the measurement object can be measured very accurately. .

61 is a block diagram showing the conceptual structure of the multi-wavelength radiation thermometer of the tenth embodiment according to the present invention.

The multi-wavelength radiation thermometer according to the embodiment of the present invention has a built-in spectrometer for spectroscopy to disperse light waves at four different wavelengths λ to λ and a photoelectric converter for measuring luminance temperatures S1 to S4 at each spectral wavelength λ to λ, respectively. The emissivity-ratio ratio ε is used for the photodetector 610 and the luminance temperatures S1 to S4 obtained online for the combination of the two wavelengths according to the following equation (20) corresponding to the above-mentioned equation [3]: / ε And an emissivity power ratio calculation block 612 for calculating.

i, j: positive integer less than or equal to 4, i ≠ j

C2: Planck Secondary Constant

In addition, the radiation thermometer has a three-dimensional power-ratio ratio regression function F for emissivity ratios prepared in advance from the actual measurement data obtained offline, a three-dimensional emissivity ratio regression function G and a regression function F similarly prepared for the emissivity ratios. A reference block 614 with a mapping H for converting to < RTI ID = 0.0 > H, < / RTI > corresponding to a three-dimensional coordinate value based on actual measurements formed of three emissivity squared ratios calculated according to Equation (20) using the luminance temperature obtained online. A decision block 616 for selecting points on the regression function F, the result of applying the mapping H to selected points on the regression function F, and converting them to points on the regression function G and calculating three emissivity ratios ε _i / ε _j from the conversion points. Conversion block 618, a selection block 6 for monitoring the time change of the calculated emissivity power ratio in calculation block 612 that indicates a smaller time change. 20), and a temperature calculation block 622 that obtains a measurement temperature as a result of performing the temperature calculation using some or all of the calculated emissivity ratios in the change block 618.

The reference block 614 is a three-dimensional regression function F for the emissivity squared ratio represented by the three-dimensional coordinates (X, Y, Z) in FIG. 62 that are determined offline for each measurement object. Built-in mapping H converts both the three-dimensional regression function G and the regression function F to the regression function G for the emissivity ratio represented by (X, Y, Z).

Regression functions F and G plot the emissivity power ratio or emissivity ratio based on the actual measurement as indicated by the sunspot in the figure and plot it as the mean of the group of sunspots.

In addition, [ε _i ^{λ i} / ε _j ^λ _j ] indicated in FIG. 62 as a data plot represents a series of emissivity squared ratios based on actual measurements and [ε _i / ε _j ] shown in FIG. 63 is identical to the actual measurements. Series of emissivity ratios.

The regression function F is input from the reference block 614 to the decision block 616 and the mapping H and the regression function G are input from the same to the transform block 618.

Next, the operation of the embodiment of the present invention will be described.

The luminance temperatures S1 to S4 corresponding to the four different wavelengths λ ₁ to λ ₄ obtained by the photodetector 610 are input to the emissivity-ratio ratio calculation block 612 when the light beam from the measurement target (not shown) is received. And the emissivity-ratio is calculated for each combination of (S1, S2), (S2, S3) and (S3, S4) according to equation (20).

As a result, for example, a specific corresponding point P '(X', Y ', Z') is obtained in the three-dimensional coordinate space in FIG.

Thereafter, it is determined that the point P (X, Y, Z) in the regression function F closest to the point P 'is used for the temperature calculation.

After determining the point P on the regression function F, the mapping H is applied to the point P and the point Q (X, Y, Z) on the regression function G with respect to the emissivity ratio indicated in FIG. 63 is determined.

On the other hand, the time variation of the three emissivity power ratios X ', Y', Z 'calculated in the emissivity power ratio calculation block 612 is monitored and the emissivity power ratio which represents a smaller time change is selected block 620. My X, Y and Z are selected from.

When there is a large change in emissivity ratio, the emissivity ratio is eliminated and the emissivity ratio is obtained from one or two remaining emissivity ratios, and the temperature T is the selected emissivity ratio according to equation [5] in the thermometer block 622. Is calculated and the temperature T is output.

In the temperature calculation block 622, the actual temperature T can be determined from the mean value of the emissivity power ratio, or the emissivity at which the temperature change is minimal, when there are two emissivity power ratios which have substantially no time change in the emissivity power ratio or show no time change. It can be done using a power-of-base ratio.

According to the embodiment of the present invention described in detail above, three different emissivity power ratios can be obtained which can be displayed in three-dimensional coordinates using four different wavelengths, and a two-color temperature calculation is obtained for each of three emissivity power ratios. The amount of result information that can be made can be greatly increased compared to the prior art improved dichroic thermometer in which only two wavelengths are used.

As a result, even if an unexpected sudden change in emissivity occurs in several wavelength ranges because of oxide film formation, another wavelength range which does not exhibit any time change can be used for the temperature calculation.

As a result, the surface temperature T can be measured accurately.

Also, as shown in FIG. 73 and FIG. 74, according to the present embodiment, the multi-wavelength system is simply applied to an improved two-color thermometer, and according to this embodiment, according to the operation of one mapping H, Since n-1) emissivity ratios can be obtained, the algorithm becomes simpler and it is possible to calculate the actual temperature very easily and quickly.

While the present invention is described in detail above, the present invention is not limited to the above-described embodiments, and various changes may be made without departing from the gist of the present invention.

For example, the time change of the emissivity ratio can be monitored to calculate the actual temperature in the embodiment and the time change of the emissivity ratio can be arranged to be monitored.

In addition, when it is proved by an offline experiment or the like that an unexpected change in the radiation characteristic of an object occurs at a specific emissivity, the emissivity-ratio or the emissivity ratio corresponding to the emissivity may be removed.

In general, since such an abnormal change tends to occur as the emissivity-ratio or the emissivity ratio is taken to a large value, it can be arranged to eliminate this value.

While monitoring the time change of the emissivity-ratio, the selection block 620 may be as shown in FIG. 64 or as shown in FIG. 64b while monitoring the absolute value of the emissivity-ratio.

For convenience, the structure may be as shown in FIG. 64B.

For convenience, three lambda ₁ , lambda ₂ and lambda ₃ are used as examples.

Selection block 620 displayed in the 64a also has a calculation block (620A), the calculation block (620B), and a smaller rate of change than for differentiating the ε ₃ ^λ3 / ε ₁ ^λ1 for differentiating ε ₂ ^λ2 / ε ₁ ^λ1 Comparator 620C for comparing the calculation results of the two calculation blocks (620A, 620B) for outputting a signal for selecting a corresponding measurement band.

Select block first shown in 64b is also 620 ε ₂ arithmetic block (620D), calculation for calculating the absolute value of ε ₃ ^λ3 / ε ₁ ^λ1 block (620E), and further for calculating an absolute value of ^λ2 / ε ₁ ^λ1 Comparator 620F is used to compare the calculation results made in two calculation blocks 620D and 620E for outputting a signal for selecting a measurement band corresponding to a small absolute value.

According to the embodiment of the present invention, as described above, the surface temperature can be measured accurately and easily even when the surface condition of the object to be measured changes with time and the emissivity changes with the wavelength of measurement.

Claims (16)

Detection means for detecting light having a specific wavelength emitted from the material and outputting a plurality of detection signals; Calculating means for calculating a first parameter corresponding to an emissivity ratio from the plurality of detection signals; And converting means for converting the first parameter into the second parameter according to the correlation between the second parameter and the first parameter indicating the surface state of the material. The surface condition measuring apparatus according to claim 1, wherein said detecting means includes a plurality of radiation sensors for detecting light beams forming different angles between a light path and a normal of a material surface. The surface state measuring apparatus according to claim 1, wherein the detecting means detects a light beam and a polarization angle different from each other from light emitted from the material. 4. The apparatus of claim 3, wherein the detecting means comprises polarizing means for extracting light beams having different polarization angles from light emitted from the material and a plurality of detection elements for detecting each light beam having these different polarization angles. Surface condition measuring apparatus, characterized in that. 2. A surface condition measuring apparatus according to claim 1, wherein said second parameter is an algebraic ratio of emissivity and said changing means also obtains a temperature near said material surface from said second parameter and said detection signal. An apparatus according to claim 1, wherein said second parameter is an emissivity ratio and said conversion means also obtains a temperature near said material surface from said second parameter, said wavelength, and said detection signal. An apparatus according to claim 1, wherein said second parameter is a film thickness of an oxide film formed on said material surface. The surface condition measuring apparatus according to claim 1, wherein said detecting means detects a light beam having a light beam emitted from said material and a specific polarization angle from the light beam emitted from said material. One or more dichroic thermometers having a common field of view on any surface of the process material; It said second exponentiation means for bigyesan from the two wavelengths λ ₁ and λ brightness temperature for the _second color obtained from the thermometer to the emissivity a power ratio ε ₁ ^λ1 / ε ₂ ^λ2 calculating according to the formula (C2: flank secondary constant); The following correlation function f, previously measured and built-in emissivity multiplication ratio ε ₁ ^λ1 / ε ₂ ^λ2 First emissivity ratio calculation means for converting according to a power emissivity ratio ε ₁ ^λ1 / ε ₂ ^λ2 claim 1 spectral emissivity ratio ε ₁ / ε ₂ in; The second spectral emissivity ratio ε ₁ ^* / ε ₂ ^* the two wavelengths as a variable to calculate with theoretical or empirical formula having one or more properties that affect the emissivity of the spectral emissivity of the λ ₁ and λ ₂ ε ₁ ^* and second emissivity ratio calculating means for separately calculating ε ₂ ^* ; An exploration calculation for providing an initial value to the property and searching and calculating a property value substantially equal to the second spectroscopy ratio ε ₁ ^* / ε ₂ ^* and the first spectroscopy emissivity ratio ε ₁ / ε ₂ calculated using the physical properties. Way; And a temperature calculating means for performing two-color temperature calculation using the first spectroscopy emissivity ratio ε ₁ / ε ₂ or the second spectroscopy emissivity ratio ε ₁ ^* / ε ₂ ^* after completion of the search calculation. . One or more dichroic thermometers having a common field of view on any surface of the process material; First a power ratio calculating means for calculating a first emissivity a power ratio ε ₁ ^λ1 / ε ₂ ^λ2 from the brightness temperatures according to the following formula S1 and S2 for two wavelengths λ ₁ and λ ₂ obtained from said two-color thermometer (C2: flank secondary constant); Spectroscopy for two wavelengths λ ₁ and λ ₂ using a theoretical or empirical formula with one or more properties that affect the emissivity as a parameter for calculating the second emissivity-ratio ε ₁ ^* λ ₁ / ε ₂ ^* λ ² from the calculations Second power-ratio ratio calculation means for separately calculating emissivity ε ₁ ^* and ε ₂ ^* ; And search calculation means for providing an initial value to the property value and searching and calculating a property value substantially equal to the second emissivity ratio ε ₁ ^{* λ 1} / ε ₂ ^{* λ 2} and the first emissivity ratio ε ₁ λ ¹ / ε ₂ ^{λ 2} . Surface condition measuring apparatus, characterized in that. Spectroscopic means for dispersing light at two different wavelengths λ ₁ and λ ₂ for each of a plurality of different wavelength regions; Photoelectric conversion means for measuring luminance temperatures S1 and S2 for two wavelengths for each wavelength region; A power for calculating for each of the following two of the emissivity from the brightness temperatures S1 and S2 measured over the wavelength a power ratio ε ₁ ^λ1 / ε ₂ ^λ2 to the wavelength region expression ratio calculation means (C2: flank secondary constant); The emissivity power-ratio ε ₁ λ ¹ / ε ₂ λ _{2 is} ^premeasured and built in, then the correlation function f Emissivity ratio calculation means for converting into a spectral emissivity ratio ε ₁ / ε ₂ using; Thermometer calculation means for performing a two-color temperature calculation using the spectral emissivity ratio ε ₁ / ε ₂ to calculate the measured temperature; And a wavelength range selecting means for monitoring a time variation of the emissivity squared ratio or the spectral emissivity ratio for each wavelength region, and the measured temperature calculated for the parameter wavelength region is output. 12. The two-color multiple radiation thermometer according to claim 11, wherein the wavelength region selecting means is configured to compare the time change of the emissivity ratio or the spectral emissivity ratio and to select the minimum wavelength region. The two-color multiple emission thermometer according to claim 11, wherein the wavelength region selection means compares the absolute value of the emissivity power ratio or the spectral emissivity ratio and selects the wavelength region so that the absolute value is the minimum value. The lower wavelength side or the second wavelength region or the higher wavelength direction according to claim 11, wherein the wavelength region selecting means makes the time variation of the emissivity squared ratio or the spectral emissivity ratio maximum when three or more wavelength regions are used. 2-color multiple radiation thermometer, characterized in that for selecting the wavelength range below the same. Spectroscopic means for dispersing light at three or more different wavelengths λ ₁ and λ ₂ ; Photoelectric conversion means for measuring luminance temperatures S1 and Sn at respective wavelengths λ ₁ to λ ₂ ; Two-line temperature of the obtained luminance for each combination of different wavelengths S1 and a power ratio calculation means for calculating an emissivity of a power ratio ε ₁ ^λ1 / ε ₂ ^λ2 according to the formula to from Sn i, j: Positive integer i ≠ j not exceeding n C2: flank secondary constant; Built-in (n-1) dimensional power-of-risk regression function F based on off-line measurement data or theory, and expresses the interrelationship between emissivity-ratio ε ₁ ^λi / ε ₂ ^λj , emissivity-ratio ε _i / ε _j (n-1) reference means for representing a mapping H correlation for converting the dimensional emissivity ratio regression function G and the power-ratio ratio regression function F into an emissivity regression function G; Judging means using luminance temperatures S1 to Sn obtained online to determine points on the emissivity-ratio ratio regression function F corresponding to the (n-1) -dimensional coordinates formed from the (n-1) emissivity-ratio ratios calculated from the above equation; Changing means for applying the mapping H to a pre-determined point on the power-ratio regression function F, which converts the point into a point on the emissivity ratio regression function G, and calculating (n-1) emissivity ratio ε _i / ε _j from the converted point; And thermometer calculating means for calculating the measured temperature by calculating the temperature using part or all of the calculated emissivity ratio. 16. The multi-wavelength radiation thermometer of claim 15, wherein the emissivity-ratio or emissivity ratio is monitored and the emissivity-ratio or emissivity ratio indicating a large change is excluded from the temperature calculation.