Method for measuring particle size of inclusion in metal by emission spectrum intensity of element constituting inclusion in metal, and method for forming particle size distribution of inclusion in metal, and apparatus for executing that method

Abstract

An electron probe microanalyzer determines the particle size of intermetallic inclusions in a master while scanning an area of &phgr;5 mm located at an arbitrary location on the surface of the master. A calibration curve representative of the relationship between the particle size of intermetallic inclusions and the emission spectrum intensity of an constituent element constituting intermetallic inclusions is generated based on the determined particle size of the intermetallic inclusions. Intermetallic inclusions existing on emission spots on the surface of a test sample are specified based on data of emission spectrum intensity of an element existing on the emission spots, and a particle size of the specified intermetallic inclusions is determined based on the data of emission spectrum intensity and the generated calibration curve.

Description

TECHNICAL FIELD

[0001] This invention relates to a method of determining particle sizes of non-metallic inclusions dispersed in metal materials (hereinafter referred to as an “intermetallic inclusions”) based on emission spectrum intensities of constituent elements of the intermetallic inclusions, and a method of generating particle size distributions of the intermetallic inclusions, as well as to apparatuses for performing the methods, and more particularly to a method of determining particle sizes of intermetallic inclusions based on emission spectrum intensities of constituent elements of the intermetallic inclusions by using an emission spectral analysis method, and a method of generating particle size distributions of intermetallic inclusions, as well as to apparatuses for performing the methods.

BACKGROUND ART

[0002] In general, a steel material contains various kinds of intermetallic inclusions dispersed therein. The compositions as well as the particle sizes of the compositions and the particle size distribution of these intermetallic inclusions seriously affect the quality of the steel material, particularly the pureness of roller bearings when the steel material is used as a material for the roller bearings. Therefore, to maintain the pureness of the steel material, it is important to identify the composition of each intermetallic inclusion in the steel material and determine the particle size distribution of intermetallic inclusions existing in a predetermined area of the steel material, i.e. to determine or measure the number of intermetallic inclusions existing in the steel material for each predetermined particle size. For example, when a steel material for roller bearings contains a large number of intermetallic inclusions having relatively large particle sizes of more than 3 &mgr;m, peeling or the like easily occurs at these intermetallic inclusions, which results in considerable deterioration in the pureness of products, such as roller bearings.

[0003] Therefore, the reduction of the number of intermetallic inclusions, i.e. high pureness, is demanded of steel materials, and hence it is essential to accurately and quickly determine the compositions and particle sizes of intermetallic inclusions and a particle size distribution of intermetallic inclusions in a steel material.

[0004] Conventionally, as a method of determining the compositions and particle sizes of intermetallic inclusions and a particle size distribution of the intermettalic inclusions, there are known a method using an electron probe microanalyzer (EPMA) and a method using emission spectral analysis. Particularly, the method using emission spectral analysis identifies respective compositions of intermetallic inclusions by separating an emission spectrum emitted from intermetallic inclusions contained in a steel material which was subjected to spark discharge, into emission spectra specific to respective elements, and therefore, the method is advantageous in that the composition of each intermetallic inclusion can be identified speedily, and therefore, a number of methods using emission spectral analysis have been disclosed e.g. in “Iron and Steel vol. 73 (1987) S696, S670”, “CAMP-IsIJ vol. 7 (1994) 1292, 1293”, and Japanese Laid-Open Patent Publications (Kokai) Nos. 4-238250 and 9-43150.

[0005] Of these methods, a method described in Japanese Laid-Open Patent Publication (Kokai) No. 4-238250 measures, in time sequence, ones of emission spectra obtained by spark discharge which correspond to zero to several hundreds of pulses corresponding to an initial stage of the spark discharge, and then determines the number of intermetallic inclusions and the compositions and contents of intermetallic inclusions, based on ones of the measured emission spectra falling within a predetermined intensity range, using predetermined equations.

[0006] However, the method using an EPMA requires a complicated procedure including operations of electronic probing and various arithmetic operations, and hence this method cannot speedily determine the composition and particle size distribution of each of intermetallic inclusions contained in a large amount of samples cut out from steel materials.

[0007] Further, the method described in Japanese Laid-Open Patent Publication (Kokai) No. 4-238250 separates emission spectra emitted from intermetallic inclusions contained in a steel material subjected to a spark discharge generated under discharge conditions adapted to the steel material, into emission spectra specific to respective elements, identifies the composition of the intermetallic inclusions based on the wavelength and/or intensity of the obtained emission spectra inherent to the respective elements, and then determines the concentration of each constituent element and the particle sizes of the intermetallic inclusions. However, as is distinct from the present invention which is directed to a real steel material by using a calibration relationship in a high-concentration region, the disclosed method carries out extrapolation of a calibration relationship in a trace concentration region. The extrapolation makes it difficult for the disclosed method to correctly determine the particle sizes and the particle size distribution.

[0008] Further, when it is required to perform quick and accurate determination of particle sizes and a particle size distribution of intermetallic inclusions, there has been used a calibration curve concerning concentrations e.g. of not more than 500 ppm of a specific metal element constituting the intermetallic inclusions, i.e. a calibration curve concerning concentrations of a specific metal element of the intermetallic inclusions, which is obtained from a trace amount of a master, and hence when it is necessary to carry out the identification even in a high concentration region where the concentration of the metal element of the intermetallic inclusions is high, as in the case of determining particle sizes and a particle size distribution of the intermetallic inclusions, there is no choice but to estimate the calibration relationship up to the high concentration region by extrapolation used in the low concentration region, which makes the calibration relationship inaccurate. For this reason, the use of an alloy as a master which has a high concentration of a certain metal element falling within a known range in intermetallic inclusions, e.g. the use of an Al-based alloy as the master in the case of Al being a constituent element of intermetallic inclusions, may be contemplated so as to obtain a calibration relationship in a region where the concentration of the metal element in intermetallic inclusions is high. However, when it is necessary to obtain an Al calibration curve, since the background of the Al-based alloy is not iron (Fe), it is difficult to accurately determine Al contained in the intermetallic inclusions existing in the steel. Moreover, due to existence of a high emission spectrum intensity of an element, such as Mg, whose emission wavelength is close to that of Al, it is difficult to correctly identify the concentration of Al.

[0009] Further, there is another conventional determination method in which after a steel material is cut out as test samples, the surface of the steel material is polished to form a mirror finished surface, and then the formed mirror finished surface is inspected by an optical microscope and then the inspection results are subjected to image analysis to thereby determine the average particle size of Al2O3 or other intermetallic inclusions and the particle size distribution of the intermetallic inclusions. However, since the mirror finished surface is thus formed by polishing the surface of the steel material and then the particle sizes and particle size distribution of the Al2O3 or other intermetallic inclusions are determined on the mirror finished surface, a surface of an Al2O3 or other intermetallic inclusion corresponding to its particle size (real particle size) may not be exposed on the mirror finished surface. Therefore, although the manner of preparing test samples is simple, the determination can be performed easily, and a passably accurate average particle size can be determined (see FIG. 4) so long as an appropriate averaging method and an appropriate statistical method are used, the average particle size obtained by the method may be estimated as a smaller apparent particle size than the real particle size.

[0010] From the viewpoint that the real particle size (obtained by calculating a spherical diameter from the volume of an intermetallic inclusion particle), such as the maximum particle size and the average particle size, is the very size closely related to the rolling life, the average particle size of the Al2O3 or other intermetallic inclusions, obtained through the image analysis is estimated to be smaller than the real particle size as described above. Since the correlation between the average particle size thus estimated to be smaller and the real particle size cannot be clearly determined, the relationship between the average particle size estimated to be smaller and the rolling life may not be accurately determined.

[0011] Methods for directly determining the particle size of intermetallic inclusions include a method in which intermetallic inclusions e.g. of Al2O3 are extracted from a steel material providing test samples, by an electron beam elution method or a chemical extraction method, into grains. However, it is difficult to carry out these methods conveniently, since the electron beam elution method necessitates an apparatus for performing the same method, and/or it takes time to effect the elution or extraction.

[0012] It is an object of the present invention to provide a method of determining particle sizes of intermetallic inclusions based on emission spectrum intensities of constituent elements of the intermetallic inclusions, and a method of generating a particle size distribution of the intermetallic inclusions, as well as apparatuses for performing the methods, which are capable of identifying what form is assumed by the intermetallic inclusions from the constituent elements of the intermetallic inclusions, and quickly and accurately determining the particle sizes and particle size distribution of the intermetallic inclusions.

[0013] It is another object of the invention to provide a method of determining particle sizes of intermetallic inclusions based on an emission spectrum intensities of constituent elements of the intermetallic inclusions, and a method of generating a particle size distribution of the intermetallic inclusions, as well as apparatuses for performing the methods, which are capable of quickly and accurately determining real particle sizes and a particle size distribution of the intermetallic inclusions.

DISCLOSURE OF INVENTION

[0014] To attain the first-mentioned object, a particle size-determining method as recited in claim 1 is characterized by comprising the steps of determining a particle size of the intermetallic inclusions, which is known, in a predetermined area of a reference sample, determining an intensity of emission spectra emitted from the constituent element of the intermetallic inclusions, from a relationship thereof to an intensity of emission spectra from spark discharge spots within the predetermined area, through emission spectral analysis of the reference sample, and forming an inclusion particle size-intensity calibration curve representative of the relationship between the particle size of the intermetallic inclusions and the emission spectrum intensity of the constituent element of the intermetallic inclusions.

[0015] Preferably, the particle size of the intermetallic inclusions, which is known, in the predetermined area of the reference sample is determined through surface analysis by an electron probe microanalyzer.

[0016] To attain the above other object, a particle size-determining method as recited in claim 3 is characterized by comprising the steps of determining an intensity of emission spectra emitted from a principle component having an already known concentration and contained in a reference sample having an already known concentration in spark discharge spots within a predetermined area of the reference sample, through emission spectral analysis of the reference sample, forming a principle component known concentration-intensity calibration curve representative of the relationship between the emission spectrum intensity of the principle component having an already known concentration and the known concentration of the principle component, determining an intensity of emission spectra emitted from a principle component contained in a real steel material-reference sample in spark discharge spots within a predetermined area of the real steel material-reference sample, and an intensity of emission spectra emitted from a constituent element of intermetallic inclusions contained in the real steel material-reference sample, through emission spectral analysis of the real steel material-reference sample, calculating a concentration of the principle component contained in the real steel material-reference sample from the emission spectrum intensity of the principle component of the real steel material-reference sample, based on the principle component known concentration-intensity calibration curve, calculating a concentration of the constituent element of the intermetallic inclusions contained in the real steel material-reference sample, based on the calculated concentration of the principle component of the real steel material-reference sample, forming a real steel material-contained intermetallic inclusion constituent element concentration-intensity calibration curve representative of the relationship between the concentration of the constituent element of the intermetallic inclusions and the emission spectrum intensity of the constituent element of the intermetallic inclusions, determining an intensity of emission spectra emitted from a base element of a real steel material-pure base sample in spark discharge spots within a predetermined area of the real steel material-pure base sample, and a base element evaporation amount indicative of mass of the base element having been evaporated due to spark discharge thereon, through emission spectral analysis of the real steel material-pure base sample, and forming a base element evaporation amount-intensity calibration curve representative of the relationship between the base element evaporation amount and the intensity of emission spectra emitted from the base element.

[0017] Preferably, the particle size-determining method comprising the steps of calculating a base element evaporation volume indicative of a volume of the evaporated base element, from the base element evaporation amount based on a density of the base element, and calculating, from the base element evaporation volume, a particle size of the intermetallic inclusions as a diameter thereof corresponding to the base element evaporation volume, based on a formula of calculating a spherical volume, determining a known concentration of the principle component from the emission spectrum intensity of the base element, based on the principle component known concentration-intensity calibration curve, calculating a concentration of the constituent element of the intermetallic inclusions based on the determined known concentration of the principle component, determining an intensity of emission spectra of the constituent element of the intermetallic inclusions from the calculated concentration of the constituent element of the intermetallic inclusion, based on the real steel material-contained intermetallic inclusion constituent element concentration-intensity calibration curve, and forming an intermetallic inclusion particle size-intensity calibration curve representative of the relationship between the calculated particle size of the intermetallic inclusions and the determined emission spectrum intensity of the constituent element of the intermetallic inclusions.

[0018] To attain the first-mentioned object, a particle size distribution-generating method as recited in claim 5 is characterized by comprising the steps of executing a data sorting process for counting a number of data items of emission spectra of a constituent element of the intermetallic inclusions in a test sample, and generating a particle size distribution based on the counted number of the data items and the particle size of the intermetallic inclusions in the test sample, which have been determined by any of the particle size-determining methods described above.

[0019] Preferably, in the data sorting process, the data items of emission spectra of the constituent element of the intermetallic inclusions in the test sample are rearranged in order of intensity, and then the number of the rearranged data items is counted.

[0020] More preferably, in the data sorting process, it is determined whether or not an emission spectrum intensity of the constituent element of the intermetallic inclusions in the test sample is larger than a threshold value, and then data items of intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample to be rearranged in order of intensity are extracted based on a result of the determination.

[0021] More preferably, the particle size distribution-generating method as recited in claim 8 further comprising the step of determining an intensity of emission spectra emitted from a principle component contained in the test sample, and in the data sorting process, the data items of intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample to be rearranged in order of intensity are extracted based on a result of comparison between the intensity of emission spectra emitted from the principle component contained in the test sample and the intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample.

[0022] More preferably, the particle size distribution-generating method further comprising the steps of forming an intensity correction curve representative of the relationship between a number of times of generation of spark discharge for emission spectral analysis and an amount of attenuation of an intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample after carrying out the spark discharge, and correcting the intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample according to the number of times of generation of spark discharge for emission spectral analysis, based on the generated intensity correction curve.

[0023] More preferably, a kind of the constituent element of the intermetallic inclusions contained in the test sample is identified based on a result of comparison between the intensity of emission spectra emitted from the principle component contained in the test sample and the intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample.

[0024] To attain the first-mentioned object, a particle size-determining apparatus as recited in claim 11 is characterized by comprising acquiring means for acquiring a particle size of the intermetallic inclusions, which is known, in a predetermined area of a reference sample, acquiring means for acquiring an intensity of emission spectra emitted from the constituent element of the intermetallic inclusions, from a relationship thereof to an intensity of emission spectra from spark discharge spots within the predetermined area, through emission spectral analysis of the reference sample, and forming means for forming an inclusion particle size-intensity calibration curve representative of the relationship between the particle size of the intermetallic inclusions and the emission spectrum intensity of the constituent element of the intermetallic inclusions.

[0025] Preferably, the particle size-determining apparatus comprises acquiring means for acquiring the particle size of the intermetallic inclusions, which is known, in the predetermined area of the reference sample through surface analysis by an electron probe microanalyzer.

[0026] To attain the above other object, a particle size-determining apparatus as recited in claim 13 is characterized by comprising acquiring means for acquiring an intensity of emission spectra emitted from a principle component having an already known concentration and contained in a reference sample having an already known concentration in spark discharge spots within a predetermined area of the reference sample, through emission spectral analysis of the reference sample, forming means for forming a principle component known concentration-intensity calibration curve representative of the relationship between the emission spectrum intensity of the principle component having an already known concentration and the known concentration of the principle component, acquiring means for acquiring an intensity of emission spectra emitted from a principle component contained in a real steel material-reference sample in spark discharge spots within a predetermined area of the real steel material-reference sample, and an intensity of emission spectra emitted from a constituent element of intermetallic inclusions contained in the real steel material-reference sample, through emission spectral analysis of the real steel material-reference sample, calculating means for calculating a concentration of the principle component contained in the real steel material-reference sample from the emission spectrum intensity of the principle component of the real steel material-reference sample, based on the principle component known concentration-intensity calibration curve, calculating means for calculating a concentration of the constituent element of the intermetallic inclusions contained in the real steel material-reference sample, based on the calculated concentration of the principle component of the real steel material-reference sample, forming means for forming a real steel material-contained intermetallic inclusion constituent element concentration-intensity calibration curve representative of the relationship between the concentration of the constituent element of the intermetallic inclusions and the emission spectrum intensity of the constituent element of the intermetallic inclusions, acquiring means for acquiring an intensity of emission spectra emitted from a base element of a real steel material-pure base sample in spark discharge spots within a predetermined area of the real steel material-pure base sample, and a base element evaporation amount indicative of mass of the base element having been evaporated due to spark discharge thereon, through emission spectral analysis of the real steel material-pure base sample, and forming means for forming a base element evaporation amount-intensity calibration curve representative of the relationship between the base element evaporation amount and the intensity of emission spectra emitted from the base element.

[0027] Preferably, the particle size-determining apparatus comprises calculating means for calculating a base element evaporation volume indicative of a volume of the evaporated base element, from the base element evaporation amount based on a density of the base element, and calculating, from the base element evaporation volume, a particle size of the intermetallic inclusions as a diameter thereof corresponding to the base element evaporation volume, based on a formula of calculating a spherical volume, acquiring means for acquiring a known concentration of the principle component from the emission spectrum intensity of the base element, based on the principle component known concentration-intensity calibration curve, calculating means for calculating a concentration of the constituent element of the intermetallic inclusions based on the determined known concentration of the principle component, acquiring means for acquiring an intensity of emission spectra of the constituent element of the intermetallic inclusions from the calculated concentration of the constituent element of the intermetallic inclusion, based on the real steel material-contained intermetallic inclusion constituent element concentration-intensity calibration curve, and forming means for forming an intermetallic inclusion particle size-intensity calibration curve representative of the relationship between the calculated particle size of the intermetallic inclusions and the determined emission spectrum intensity of the constituent element of the intermetallic inclusions.

[0028] To attain the first-mentioned object, a particle size distribution-generating apparatus as recited in claim 15 is characterized by comprising data sorting means for executing a data sorting process for counting a number of data items of emission spectra of a constituent element of the intermetallic inclusions in a test sample, and generating means for generating a particle size distribution based on the counted number of the data items and the particle size of the intermetallic inclusions in the test sample, which have been determined by the particle size-determining apparatuses described above.

[0029] Preferably, data sorting means rearranges the data items of emission spectra of the constituent element of the intermetallic inclusions in the test sample counts the number of the rearranged data items.

[0030] More Preferably, the data sorting means determines whether or not an emission spectrum intensity of the constituent element of the intermetallic inclusions in the test sample is larger than a threshold value, and then extracts data items of intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample to be rearranged in order of intensity, based on a result of the determination.

[0031] More preferably, the particle size distribution-generating apparatus further comprises acquiring means for acquiring an intensity of emission spectra emitted from a principle component contained in the test sample, and the data sorting means extracts the data items of intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample to be rearranged in order of intensity, based on a result of comparison between the intensity of emission spectra emitted from the principle component contained in the test sample and the intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample.

[0032] More preferably, the particle size distribution-generating apparatus further comprises forming means for forming an intensity correction curve representative of the relationship between a number of times of generation of spark discharge for emission spectral analysis and an amount of attenuation of an intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample after carrying out the spark discharge, and correcting means for correcting the intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample according to the number of times of generation of spark discharge for emission spectral analysis, based on the generated intensity correction curve.

[0033] More preferably, the particle size distribution-generating apparatus comprises identifying means for identifying a kind of the constituent element of the intermetallic inclusions contained in the test sample, based on a result of comparison between the intensity of emission spectra emitted from the principle component contained in the test sample and the intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample.

BRIEF DESCRIPTION OF DRAWINGS

[0034] FIG. 1 is a diagram schematically showing an emission spectrometer for performing a particle size-determining and particle size distribution-generating method according to a first embodiment of the present invention;

[0035] FIG. 2 is a flowchart of a particle size-determining and particle size distribution-generating process executed according to the particle size-determining and particle size distribution-generating method according to the first embodiment;

[0036] FIG. 3 is a flowchart of a calibration curve A-forming process, which is executed in a step S201 in FIG. 2, for forming a calibration curve A representative of the relationship between the Al2O3 particle size and the emission spectrum intensity of Al;

[0037] FIG. 4 is a view schematically illustrating a surface analysis result image of Al existing in Al2O3;

[0038] FIG. 5 is a diagram illustrating the calibration curve A representative of the relationship between the Al2O3 particle size and the emission spectrum intensity of Al, which is formed in a step S308 in FIG. 3;

[0039] FIG. 6 is, FIG. 6 is a flowchart of a particle size distribution-generating process which is executed in a step S202 in FIG. 2;

[0040] FIGS. 7A to 7E are diagrams illustrating distributions of data items of emission spectrum intensities of Fe, O, Al, Ca and C, which are arranged in time sequence, for comparison performed in a step S605 in FIG. 6;

[0041] FIG. 8 is a flowchart of a data sorting process which is executed in a step S606 in FIG. 6;

[0042] FIG. 9 is a continued part of the flowchart of the data sorting process which is executed in the step S606 in FIG. 6;

[0043] FIG. 10 is a still continued part of the flowchart of the data sorting process which is executed in the step S606 in FIG. 6;

[0044] FIG. 11 is a flowchart of a data sorting process which is executed according to a variation of the particle size-determining and particle size distribution-generating method according to the first embodiment;

[0045] FIG. 12 is a flowchart of a particle size-determining and particle size distribution-generating process which is executed according to a particle size-determining and particle size distribution-generating method according to a second embodiment of the present invention;

[0046] FIG. 13 is a flowchart of a calibration curves B, C-forming process which is executed in a step S1201 in FIG. 12 for forming a calibration curve B representative of the relationship between the Fe emission spectrum intensity and the Fe concentration and then forming a calibration curve C representative of the relationship between the Al emission spectrum intensity and the Al concentration based on the calibration curve B;

[0047] FIG. 14 is a diagram illustrating the calibration curve B representative of the relationship between the Fe emission spectrum intensity and the Fe concentration, which is formed in the step S1201 in FIG. 12;

[0048] FIG. 15 is a cross-sectional view taken on a section orthogonal to the axis of a steel material for actual use, for example, from which a real steel master is cut out in a step S1306 in the FIG. 13;

[0049] FIG. 16 is a flowchart of a calibration curve D-forming process for forming a calibration curve D concerning the Fe emission spectrum intensity and the Fe evaporation loss in a step S1202 in FIG. 12;

[0050] FIG. 17 is a view useful for explaining a method of determining a particle size of an intermetallic inclusion in a step S1605 in FIG. 15;

[0051] FIG. 18 is a diagram illustrating a calibration curve D representative of the relationship between the Fe evaporation amount and the Fe emission spectrum intensity, which is formed in the step S1202 in FIG. 12;

[0052] FIG. 19 is a flowchart of a calibration curve E-forming process which is executed in a step S1203 in FIG. 12 for forming a calibration curve E representative of the relationship between the Al emission spectrum intensity and the Al particle size;

[0053] FIG. 20 is a flowchart of a particle size distribution-generating process which is executed in a step S1204 in FIG. 12;

[0054] FIG. 21 is a diagram illustrating the calibration curve A representative of the relationship between the Al2O3 particle size and the Al emission spectrum intensity, which is formed in a step S1907 in FIG. 19;

[0055] FIG. 22 is a flowchart of a particle size-determining and particle size distribution-generating process according to a third embodiment of the present invention;

[0056] FIG. 23 is a flowchart of an intensity correction curve-forming process which is executed in a step S2201 in FIG. 22;

[0057] FIG. 24 is a diagram illustrating an intensity correction curve which is formed in the step S2201 in FIG. 22;

[0058] FIG. 25 is a flowchart of a particle size distribution-generating process which is executed in a step S2202 in FIG. 22;

[0059] FIGS. 26A and 26B are diagrams useful in comparison between an Al2O3 particle size distribution (a) generated by execution of the FIG. 2 process and an Al2O3 particle size distribution (b) generated by an EPMA; and

[0060] FIGS. 27A and 27B are diagrams useful in comparison between the Al2O3 particle size distribution (a) generated by execution of the FIG. 12 process and an Al2O3 particle size distribution (b) generated by the image analysis method.

BEST MODE OF CARRYING OUT THE INVENTION

[0061] A particle size-determining and particle size distribution-generating method according to a first embodiment of the present invention will now be described in detail with reference to the drawings.

[0062] The particle size-determining and particle size distribution-generating method of the first embodiment is carried out by an emission spectrometer shown in FIG. 1, described hereinbelow, in generating a particle size distribution of intermetallic inclusions contained in a test sample cut out from a steel material.

[0063] The emission spectrometer for carrying out the particle size-determining and particle size distribution-generating method of the first embodiment will be described with reference to the drawings.

[0064] FIG. 1 is a diagram schematically showing the arrangement of the emission spectrometer for performing the particle size-determining and particle size distribution-generating method according to the first embodiment.

[0065] In FIG. 1, the emission spectrometer 100 is comprised of a light-emitting section 101, a light-emitting stand 102, a spectral section 103, a photometric section 104, an interface 105, a data processing section 106, and a terminal unit 107.

[0066] The light-emitting section 101 is provided with a counter electrode, not shown, while the light emitting stand 102 contains a test sample or a master (reference sample) serving as an electrode. The spectral section 103 is comprised of a condensing lens 108, a light shielding plate 111 formed with an entrance slit 109 and a plurality of exit slits 110, a concave diffraction grating 112, a plurality of photomultipliers 113, and an oil rotary pump 114. The terminal unit 107 includes a CRT, a printer, and a keyboard.

[0067] The light emitting stand 102 is connected to the interface 105 via the light-emitting section 101, and the photomultipliers 113 are connected to the interface 105 via the photometric section 104. The interface 105 is connected to the terminal unit 107 via the data processing section 106.

[0068] The light-emitting section 101 is arranged at a location where the counter electrode can generate a spark discharge on the test sample or the master. The light emitting stand 102 is arranged on an optical axis passing through the condensing lens 108, the entrance slit 109, and the concave diffraction grating 112, and at the same time on a surface of the spectral section 103. In the spectral section 103, the exit slits 110 are arranged on optical beams rotated about the concave diffraction grating 112 through angles of diffraction specific to a plurality of elements, respectively, from the optical axis passing through the condensing lens 108, the entrance slit 109, and the concave diffraction grating 112, and at the same time on the light shielding plate 111. The photomultipliers 113 are positioned on the optical axes of the respective optical beams.

[0069] Further, at least a space between the light-emitting section 101 and the light emitting stand 102 is filled with an inert rare gas (e.g. argon gas).

[0070] The counter electrode of the light-emitting section 101 generates a spark discharge on the test sample or the master which is held in the light emitting stand 102, and the test sample or the master which is subjected to the spark discharge emits light of emission spectra (emission spectra emitted respectively by iron (Fe) or other main constituents of the test sample or the master (hereinafter referred to as “ground”)) and intermetallic inclusions) containing information on a plurality of elements. The condensing lens 108 converges the emitted light of emission spectra and irradiates the concave diffraction grating 112 with the converged light via the entrance slit 109. The concave diffraction grating 112 separates the emitted light of emission spectra into emission spectra specific to the respective elements by utilizing the respective angles of diffraction specific thereto, and the separated emission spectra enter the photomultipliers 113 via the respective exit slits 110. The separated beams of the emission spectra specific to the respective elements contain respective pieces of optical information specific to the elements.

[0071] The photomultipliers 113 each detect the incoming emission spectrum, convert the intensity of the detected emission spectrum to an electric current value and transmit the converted current value to the photometric section 104. The photometric section 104 converts the received current value to a digital value, and then transmits the converted digital value to the data processing section 106 via the interface 105. On the other hand, the light-emitting section 101 and the light emitting stand 102 transmits data of the number of times and timing of generation of spark discharge and data of positions on the surface of the test sample or the master which were subjected to spark discharge to the data processing section 106 via the interface 105.

[0072] The data processing section 106 carries out processing including identification of the composition of each intermetallic inclusion, based on the received digital values and so forth. The CRT and a print printed by the printer display and indicates the result of determination of the particle size distribution of intermetallic inclusions.

[0073] Next, a particle size-determining and particle size distribution-generating process which is executed on intermetallic inclusions by the FIG. 1 emission spectrometer 100 will be described with reference to the drawings.

[0074] FIG. 2 is a flowchart of the particle size-determining and particle size distribution-generating process executed by the particle size-determining and particle size distribution-generating method of the first embodiment.

[0075] In FIG. 2, first, a calibration curve A-forming process for forming a calibration curve A, described hereinbelow with reference to FIG. 3, is executed (step S201), and then a particle size distribution-generating process, described hereinafter with reference to FIG. 6, is executed (step S202), followed by terminating the process.

[0076] FIG. 3 is a flowchart of the calibration curve A-forming process executed in the step S201 in FIG. 2, for forming the calibration curve A representative of the relationship between the Al2O3 particle size and the emission spectrum intensity of Al.

[0077] The calibration curve A-forming process is executed by the emission spectrometer 100 at least once before the particle size distribution-generating process is repeatedly carried out on intermetallic inclusions in the test sample by the emission spectrometer 100.

[0078] In FIG. 3, first, steel materials of SUJ2 (high-carbon chromium bearing steel, Type 2) containing Al2O3, for example, having a particle size approximately equal to 4 to 18 &mgr;m, which is a normal range of particle sizes of Al2O3 (alumina) contained in the test sample, are prepared, and masters having a cylindrical shape with a diameter of &phgr;40 mm are cut out from the steel materials. After each master is held on a sample stage of an EPMA (step S301), an area e.g. of &phgr;5 mm is arbitrarily set at an arbitrary location on the surface of the master, and the location of the area is determined. Then, the area is scanned by the EPMA for surface analysis thereof, whereby, when the intermetallic inclusion concerned is Al2O3, the location of Al is determined by the surface analysis, and then the particle size of Al is determined by a method of determining an average particle size of the intermetallic inclusion from an image showing the location of Al, as described hereinbelow with reference to FIG. 4 (step S302).

[0079] In the above example, it is assumed that the intermetallic inclusion concerned is Al2O3, and Al is employed as an element for identifying the intermetallic inclusion. Similarly, if Ca is employed as an identifying element for CaO (calsia), Mg for MgO (magnesia), Si for SiO2, Ca or S for CaS, Ti or N for TiN, or Mn or S for MnS, and surface analysis using such an identifying element provides a surface analysis result image concerning the metal element of a corresponding intermetallic inclusion, thereby similarly making it possible to determine the average particle size thereof.

[0080] FIG. 4 schematically illustrates a surface analysis result image of Al in Al2O3.

[0081] A description will be given of the method which is employed in the step S302 in FIG. 3, for determining the particle size of an intermetallic inclusion, with reference to FIG. 4.

[0082] In FIG. 4, the EPMA determines a major diameter dmax of an intermetallic inclusion 400 and a minor diameter dmin of the same, and based on these sizes, an average diameter dave is calculated as the particle size of the intermetallic inclusion 400 by using the following equation (1):

dave=(dmax+dmin)/2  (1)

[0083] Referring again to FIG. 3, data of the location and particle size of the intermetallic inclusion determined by surface analysis and the identified constituent element of the intermetallic inclusion are input to the data processing section 106 (step S303), and the data processing section 106 stores the input data in a memory, not shown.

[0084] Then, after the same master that was scanned by the EPMA is held in the light emitting stand 102 (step S304), the counter electrode of the light-emitting section 101 generates spark discharge e.g. one hundred times at a surface area (set e.g. to &phgr;5 mm) of the surface of the master which is the same area that was operated by the EPMA (step S305). The master subjected to the spark discharge produces emission spectra containing information on a plurality of elements. The intermetallic inclusion in the surface of the master has a dielectric property, and hence at this time, the spark discharge is selectively guided to the intermetallic inclusion in the surface of the master by the dielectric property (electrification) thereof. Accordingly, the emission spectra generated by the master are emitted from respective discharge spots (e.g. &phgr;30 &mgr;m) thereof containing the intermetallic inclusion.

[0085] The above process will be explained by taking a case in which the intermetallic inclusion is Al2O3, as an example.

[0086] Since the size of Al2O3 is proportional to the intensity of the emission spectrum of Al, emission spectrum intensities of Al arranged in order in which the intensity of Al decreases correspond respectively to average particle sizes of Al2O3 which are obtained by the above-mentioned surface analysis by the EPMA and arranged in order in which the size of Al2O3 decreases. The data of the emission spectrum intensities and the average particle sizes obtained earlier by the EPMA are input to the memory of the data processing section 106, and then these data are processed according to the intensity of Al by the data processing section 106, whereby the calibration curve A, referred to hereinafter in a step S308 and shown in FIG. 5, can be obtained. Then, in a step S309, referred to hereinafter, the obtained calibration curve A is stored in the memory of the data processing section 106.

[0087] Referring again to FIGS. 3 and 4, an area (emission spot) within the surface (the aforementioned area of &phgr;5 mm) of the master, on which spark discharge is effected once, is a circle of &phgr;30 &mgr;m (the size is always held constant), and since this emission spot has the diameter of &phgr;30 &mgr;m which is larger than the average particle size dave of general intermetallic inclusions, the emission spot can always contain at least one intermetallic inclusion therein, so that all the respective emission spectrum intensities of elements constituting intermetallic inclusions in the master can be obtained. This makes it possible to properly associate the intermetallic inclusions from which the emission spectra were obtained with the intermetallic inclusions having the respective average particle sizes dave, which were obtained by the EPMA method.

[0088] This will be explained by referring again to FIG. 3. each of emission spectra emitted by the master is separated into emission spectra specific to the respective elements by the concave diffraction grating 112, and the separated emission spectra enter respective ones of the photomultipliers 113 via the corresponding exit slits 110 (step S306).

[0089] Each photomultiplier 113 detects the incoming emission spectrum, converts the intensity of the detected emission spectrum to an electric current value, and transmits the current value obtained by the conversion to the photometric section 104. The photometric section 104 converts each of the received current values to a digital value, and then transmits the digital values obtained by the conversion to the data processing section 106 via the interface 105 (step S307). On the other hand, the light emitting stand 102 transmits data of the locations of emission spots in the area of &phgr;5 mm on the surface of the master, i.e. the locations of intermetallic inclusions to the data processing section 106 via the interface 105. Thus, the data of the respective intensities of emission spectra of the elements constituting each of the intermetallic inclusions and the data of the locations of the intermetallic inclusions are transmitted to the data processing section 106, and stored in the memory of the same.

[0090] Then, based on the data of the locations and particle sizes of the intermetallic inclusions and the data of the constituent elements of the intermetallic inclusions and the emission spectrum intensities of the constituent elements, which are stored in the memory, the data processing section 106 generates the FIG. 5 calibration curve A, described hereinbelow, which represents the relationship between the emission spectrum intensity and particle size of one of the elements constituting the intermetallic inclusions as described hereinabove (step S308). Further, the data processing section 106 stores the generated calibration curve A in the memory (step S309), followed by terminating the process.

[0091] FIG. 5 is a diagram illustrating the calibration curve A formed in the step S308 in FIG. 3. The calibration curve A indicates the relationship between the Al2O3 particle size and the emission spectrum intensity of Al.

[0092] In FIG. 5, the ordinate represents Al2O3 particle sizes determined by the surface analysis by the EPMA, while the abscissa represents the emission spectrum intensities of Al contained in Al2O3, which correspond to the respective Al2O3 particle sizes.

[0093] In FIG. 5, the data of the particle sizes of the intermetallic inclusions determined by the surface analysis by the EPMA method and the data of the emission spectrum intensities of the constituent element existing in the intermetallic inclusions are related to each other via the data of the locations and the data on the constituent element of the intermetallic inclusions. The data of the particle sizes of intermetallic inclusions are represented by the ordinate, and the data of the emission spectrum intensities of one of the elements forming the intermetallic inclusions are represented by the abscissa. In the step S307, if the determination by surface analysis is carried out on Al by using masters containing Al2O3, it is possible to obtain a calibration curve concerning Al2O3, and if the determination by surface analysis is carried out on Ca by using masters containing CaO, it is possible to obtain a calibration curve concerning CaO. FIG. 5 shows the calibration curve concerning Al2O3, which shows that the particle size of Al2O3 is larger as the emission spectrum intensity of Al is higher.

[0094] It should be noted that the particle sizes of intermetallic inclusions determined by the EPMA method and the emission spectrum intensity of metal elements existing in the same intermetallic inclusions can be associated with each other not only by the above method using correspondence between the size-decreasing order and the intensity-decreasing order, but also by a method using position coordinates of intermetallic inclusions within the &phgr;5 mm area. The FIG. 5 calibration curve A can be formed by either of these methods.

[0095] Further, intermetallic inclusions contained in a test sample are not limited to Al2O3, but calibration curves concerning CaO, MgO, SiO, CaS, TiN, MnS, and so forth can also be formed similarly to the FIG. 5 calibration curve. Therefore, it is preferable to store the calibration curves thus formed as well in the memory of the data processing section 106.

[0096] According to the FIG. 3 process, since an area of &phgr;5 mm at an arbitrary location on the surface of a master is scanned by the EPMA (step S302), it is possible to determine particle sizes of intermetallic inclusions contained in emission spots having a diameter of 30 &mgr;m within the &phgr;5 mm area of the master containing the intermetallic inclusions having particle sizes of 4 to 18 &mgr;m, and based on the determined particle sizes of the intermetallic inclusions, the calibration curve A representative of the relationship between the particle size of the intermetallic inclusions and the emission spectrum intensity of an element constituting the intermetallic inclusions is formed (step S308), so that data of particle sizes in a range of 4 to 18 &mgr;m (i.e. a target particle size range of the test sample, described hereinafter) or in its vicinity indicated by the generated calibration curve A can be made accurate.

[0097] Consequently, when particle sizes of intermetallic inclusions whose particle sizes are within a range of 4 to 18 &mgr;m, which is substantially the same as the range of those in a test sample, are determined based on such a calibration curve A as shown in FIG. 5, it is possible to determine the particle sizes of the intermetallic inclusions and a particle size distribution thereof more accurately than in the conventional case where particle sizes of Al2O3 inclusions are determined based on the relationship between the particle size and the emission spectrum intensity thereof obtained by extrapolation based on a calibration curve determined based on a trace concentration of Al solid-solved in steel.

[0098] FIG. 6 is a flowchart of the particle size distribution-generating process which is executed in the step S202 in FIG. 2.

[0099] The present process is executed by the emission spectrometer 100 whenever a particle distribution of intermetallic inclusions in a test sample is repeatedly generated after the FIG. 3 process by the emission spectrometer 100 is executed at least once.

[0100] In FIG. 6, first, a cylindrical test sample of &phgr;40 mm cut out from a steel material of SUJ2 is held in the light emitting stand 102 (step S601), and then spark discharge is generated e.g. one thousand times at a measurement area (&phgr;5 mm) on the surface of the test sample by the counter electrode of the light-emitting section 101 (step S602). The test sample subjected to spark discharge emits light of emission spectra.

[0101] Then, the light of emission spectra emitted from the test sample is separated into emission spectra specific to respective elements by the concave diffraction grating 112, and the separated emission spectra enter respective ones of the photomultipliers 113 via the corresponding exit slits 110 (step S603).

[0102] The photomultipliers 113 each detect the incoming emission spectrum and convert the intensity of the detected emission spectrum into an electric current value and transmit the current value to the photometric section 104. The photometric section 104 converts the received current value to a digital value, and then transmits the digital value to the data processing section 106 via the interface 105 (step S604). On the other hand, the light emitting stand 102 transmits data of the location of each emission spot within the measurement area on the surface of the test sample and the number of times of generation of spark discharge to the data processing section 106 via the interface 105.

[0103] Thus, the data of the respective emission spectrum intensities of elements existing in emission spots, the locations of the emission spots, and the number of times of generation of spark discharge are transmitted to the data processing section 106, and the data processing section 106 stores the data in its memory.

[0104] Then, the data processing section 106 arranges the data of the respective emission spectrum intensities of the elements in a time sequence order to generate diagrams as shown in FIGS. 7A to 7E, described hereinbelow, each illustrating a distribution of the data of the emission spectrum intensities. Further, the data processing section 106 compares data of the respective intensities of the emission spectra of the elements emitted from the same emission spot, to thereby determine whether or not there exist intermetallic inclusions within the emission spot, and when there exist intermetallic inclusions, the data processing section 106 identifies the intermetallic inclusions (step S605).

[0105] FIGS. 7A to 7E are diagrams showing data of the respective time-sequential emission spectrum intensity distributions of Fe, O, Al, Ca and C, which are compared in the step S605 in FIG. 6.

[0106] In FIGS. 7A to 7E, a broken line in each of the diagrams related, respectively, to O, Al, Ca and C indicates a threshold value for use in determining whether or not the corresponding element is solid-solved in a material (matrix). When an emission spectrum intensity is smaller than the corresponding threshold value, it can be considered that an element which emitted the emission spectrum does not form an intermetallic inclusion, but forms a background representing the matrix.

[0107] For example, when the data of the emission spectrum intensities of O and Al in the same timing (i.e. in the same emission spot) are larger than the respective threshold values, it means that Al2O3 exists as an intermetallic inclusion in the emission spot, and when the data of the emission spectrum intensities of O and Ca in the same emission spot are larger than the respective threshold values, it means that CaO exists as an intermetallic inclusion in the emission spot. Further, when the data of the emission spectrum intensities of O, Al and Ca in the same emission spot are larger than the respective threshold values, it means that Al2O3 and CaO exist as intermetallic inclusions in the emission spot.

[0108] Referring again to FIG. 6, the data processing section 106 executes a data sorting process described hereinafter with reference to FIGS. 8 to 10, to rearrange the data of the emission spectrum intensities of an element specific to each of the intermetallic inclusions identified in the step S605 in intensity-increasing order (step S606). In executing the rearrangement of the data in the step S606, the data processing section 106 deletes data of emission spectrum intensities whose values are smaller than the corresponding threshold value.

[0109] Then, the emission spectrum intensities of Al are substituted into the data of the rearranged emission spectrum intensities and the calibration curve A representative of the relationship between the Al2O3 particle size and the emission spectrum intensity of Al, which was formed by the FIG. 3 process, whereby the Al2O3 particle sizes are calculated (step S607), and based on the calculated Al2O3 particle sizes and the number of the data, a particle distribution of the intermetallic inclusions is generated. The generated data of the particle size distribution is displayed on the CRT or the printer. According to the FIG. 6 process, intermetallic inclusions existing in an emission spot on the surface of a test sample is identified based on the data of the respective emission spectrum intensities of elements existing in the emission spot (step S605), and then the particle sizes of the identified intermetallic inclusions are calculated, based on the data of the emission spectrum intensities and the calibration curve A formed by the FIG. 3 process (step S607). This makes it possible to perform quick and accurate determination of the particle sizes and particle size distribution of the intermetallic inclusions.

[0110] FIGS. 8 to 10 are a flowchart of the data sorting process which is executed in the step S606 in FIG. 6.

[0111] In the data sorting process in FIGS. 8 and 9, the data processing section 106 rearranges data Ix(i) of a row Ix of emission spectrum intensities specific to a constituent element X as one of the constituent elements of the intermetallic inclusion identified in the step S605 e.g. in intensity-increasing order, into a row ALx while deleting data items Ix(i) whose values are smaller than the threshold value for the constituent element X.

[0112] A description will be given of the data sorting process in FIGS. 8 to 10.

[0113] In FIG. 8, first, a count i indicative of the number of data items in the row Ix, which were read from the memory, is set to 1, and a count ii indicative of the number of data items of the row Ix whose values are smaller than the threshold value is set to 0 (step S801), and thereafter one data item is fetched at random as a data item Ix(i) from the row Ix of data items stored in the memory corresponding in number to the number of times J (1000) of generation of spark discharge, and it is determined whether or not the value of the fetched data item Ix(i) is equal to or larger than the threshold value for the constituent element X (step S802).

[0114] If the value of the fetched data item Ix(i) is equal to or larger than the threshold value for the constituent element X in the step S802, a count K indicative of the number of data items Ix(i) whose values are determined to be equal to or larger than the threshold value for the constituent element X is calculated by using the following equation (2) (step S803):

K=i−ii  (2)

[0115] The fetched data item Ix(i) is substituted for a data item Ax(K) in a row Ax as a set of data items Ix(i) which are equal to or larger than the threshold value for the constituent element X (step S804). Further, the count K is regarded as the maximum value of the current count i at the present time and substituted for Kmax (step S805).

[0116] If the fetched data item Ix(i) is determined to be smaller than the threshold value for the constituent element X in the step S802, the count ii is incremented by 1 (step S806), followed by the process proceeding to a step S807.

[0117] Then, it is determined whether or not the present count i is smaller than the number of times J (1000 times) of generation of spark discharge (step S807).

[0118] If the count i is determined to be smaller than the number of times J of generation of spark discharge in the step S807, the present count i is incremented by 1 (step S808), followed by the process returning to the step S802, whereas if the present count i is determined to be equal to or larger than the number of times J of generation of spark discharge, the process proceeds to a step S901 (FIG. 9).

[0119] In FIG. 9, arithmetic operations are carried out for rearranging the set of data items Ix(i) of the row data ALx(K) obtained in FIG. 8, which are equal to or larger than the threshold value for the constituent element X in intensity-increasing order.

[0120] First, an index l is set to 1 (step S901), and the count K is set to 1 (step S902), whereby an initial minimum value in the row Ax in the case of K=1 is calculated by using the following equation (3) (step S903):

&Dgr;Ax(K)=Ax(K)−Ax(K+1)  (3)

[0121] To determine which row data Ax is the smaller on the right side of the equation (3), it is determined whether or not &Dgr;Ax(K) is equal to or smaller than 0 (step S904).

[0122] If step 9, Ax(K)≦Ax(K+1) holds, it is determined whether or not K is larger than 1 (step S905). If it is determined in the step S905 that K=1 holds (during an initial operation), the data item Ax(K) in the row Ax is substituted for a variable Bx(l) (step S906), and the variable Bx(l) is substituted for a minimum value in a row data ALx(l) in which data items Ix(i) are rearranged in value-increasing order with respect to the value l (step S907). Then, an initial value of a provisional minimum value in the row AlX is set by setting ALx(l)=Bx(l), and K is set by setting KK=K (step S908). More specifically, ALx(l)=Bx(l)=Ax(K), and KK=K=1 are set. Then, the process proceeds to a step S912.

[0123] Further, if &Dgr;Ax(K) is determined to be larger than 0 in the step S904, the data item Ax(K+1) in the row Ax is substituted for the variable Bx(l) (i.e. the value of the variable Bx(l) is updated) (step S909), and the variable Bx(l) is substituted for the minimum value in the row data ALx(l) in which data items Ix(i) are rearranged in value-increasing order with respect to 1 (step S910). Then, e.g. when K=1 holds, the initial provisional minimum value ALx(1) is set by setting ALx(1)=Bx(1)=Ax(2), and KK=K+1 is set, e.g. KK=K+1=2 is set (step S911). Then, the process proceeds to the step S912.

[0124] The value KK in the steps S908 and S911 is a count for replacing the minimum value Bx(l) (=Ax(KK)) selected from the row Ax by a value of ∞ for the next value of the index l, to prevent the original Ax(K) value from being selected again, while allowing an arithmetic operation to be carried out in a step S913, referred to hereinafter.

[0125] After completion of these arithmetic operations, it is determined in a step S912 whether or not the count K is smaller than Kmax.

[0126] If the count K is determined to be smaller than Kmax in the step S912, the count K is incremented by 1 (K=K+1) (step S913), and then it is determined whether or not a data item Ax(K) newly fetched from the row Ax is ∞ (step S914). If the data item Ax(K) is determined to be ∞ in the step S914, the process returns to the step S904, whereas if the data item Ax(K) is not determined to be ∞, the process proceeds to a step S915, wherein a minimum value in the row Ax at the present value of the count K is calculated by using the following equation (4):

&Dgr;Ax(K)=ALx(l)−Ax(K+1)  (4)

[0127] This operation is carried out in order to compare the value of the data item Ax(K+1) with the minimum value ALx(l) which has been provisionally selected from the row Ax.

[0128] Then, the process returns to the step S904, wherein a determination as to the value of &Dgr;Ax(K) is carried out again. If &Dgr;Ax(K) is determined to be equal to or smaller than 0 in the step S904, and K is determined to be larger than 1 in the step S905, the process proceeds to a step S916, wherein ALx(l)=Bx(l) is set so as to maintain the provisional minimum value obtained in the preceding operation.

[0129] If the value of &Dgr;Ax(K) is not equal to or smaller than 0 in the step S904, the process proceeds to the steps S909 et seq., wherein the count KK for causing the minimum value Bx(l) (=Ax(KK)) selected from the row Ax to be replaced by ∞ for the next value of the index l is updated to K+1. On the other hand, if K is equal to or smaller than 1, the process proceeds to the step S906, wherein the count KK to be replaced by ∞ is held at K.

[0130] An arithmetic operation of the index l is completed when it is determined in the step S912 that the count K is equal to or larger than Kmax, and hence the process proceeds to a step S917, wherein the data item Ax(KK) selected from the row Ax is finally set to ∞ (Ax(KK)=∞). As a result, this value cannot be repeatedly fetched again for the next arithmetic operation for the index l. The index l is incremented by 1 in a step S918, followed by the process proceeding to a step S919.

[0131] In the step S919, it is determined whether or not the incremented index l is smaller than Kmax. This step is executed in order to determine whether or not the arithmetic operations for all values of the index l have been completed.

[0132] If the incremented index l is determined to be smaller than Kmax in the step S919, the steps S902 to S919 are repeatedly executed with the incremented index l.

[0133] If the incremented index l is determined to be equal to or larger than Kmax in the step S919, i.e. if arithmetic operations have been completed for all the values of the index l, the process proceeds to a step S1001 (FIG. 10).

[0134] In the steps S1001 et seq. in FIG. 10, arithmetic operations are carried out for determining a frequency Cx(l) indicative of a number (m) of data items ALx(l) having the same value in the row ALx of data items ALx(l) which are all equal to or larger than the threshold value for the constituent element X, i.e. for determining the number (m) of intermetallic inclusions exist for each particle size [&mgr;m].

[0135] First, in the step S1001, the index l is incremented by 1, and then 1 is substituted for an arbitrary value n in the index l (step S1002 ). Further, the count m of a counter which calculates the number of intermetallic inclusions which have the same particle size as that of ALx(l) existing in Kmax is set to 1 (step S1003) Then, &Dgr;ALx(l) representative of the difference between the value of the data ALx(l) for the index l in the row ALx and the value of the data item ALx(n) with the arbitrary value n substituted in the index l is calculated by using the following equation (5) (step S1004):

&Dgr;ALx(l)=ALx(l)−ALx(n)  (5)

[0136] Then, it is determined whether or not the calculated &Dgr;ALx(l) is equal to 0 (step S1005).

[0137] If &Dgr;ALx(l) is not determined to be equal to 0 in the step S1005, the process proceeds to a step S1006. In the step S1006, the count m representative of the number of the data items ALx(n) having the same value is held at m, and the process proceeds to a step S1010.

[0138] If &Dgr;ALx(l) is determined to be equal to 0 in the step S1005, it is determined whether or not the arbitrary value n is equal to 1 (step S1007).

[0139] If the arbitrary value n is determined to be equal to 1 (the index l is 1, and the calculated difference &Dgr;ALx(l) is 0) in the step S1007, the data count m is set to 1 (step S1008 ), whereas if the arbitrary value n is not determined to be equal to 1, the data count m is incremented by 1 (step S1009).

[0140] In the following step S1010, the arbitrary value n is incremented by 1, and then it is determined whether or not the arbitrary value n is smaller than the maximum value Kmax of the index l (step S1011).

[0141] If the arbitrary value n is determined to be smaller than the maximum value Kmax of the index l in the step S1011, the process returns to the step S1004, wherein the steps S1004 et seq. are repeatedly executed until the arbitrary value n has covered all the values of the index l. On the other hand, if the arbitrary value n is equal to or larger than the maximum value Kmax of the index l, it means that the arbitrary values n has covered all the values of the index l, and hence the data count m at this time is set to the number Cx(1) of the data items ALx(n) (step S1012).

[0142] Then, the index l is incremented by 1 (step S1013), and this value of the index l is regarded as the present maximum value lmax of the index l, and it is determined whether or not the present maximum value lmax is smaller than Kmax (step S1014). If the present maximum value lmax is smaller than Kmax, the process returns to the step S1002, whereas if the present maximum value lmax is equal to or larger than Kmax, the present process is terminated.

[0143] According to the FIG. 10 process, if the data items ALx(l), ALx(i) having the same value, which were fetched from the row ALx of the data items ALx(l) each having a value equal to or larger than the threshold value for the constituent element X, are each converted from the emission spectrum intensity of the constituent element X (e.g. Al) to a particle size of the intermetallic inclusion by using FIG. 5, described hereinbefore, or FIG. 21, described hereinafter (step S607), this means that the frequency Cx(l) indicative of the number (m) of intermetallic inclusions existing for each particle size [&mgr;m] is obtained. Therefore, it is possible to display or output a diagram shown in FIG. 26A, described hereinafter, based on the particle sizes [&mgr;m] and the row data Cx(l) representative of frequencies, in the same manner as a diagram shown in FIG. 26B, which is obtained by the EPMA.

[0144] Next, a variation of the particle size-determining and particle size distribution-generating method according to the first embodiment of the invention will be described in detail with reference to the drawings.

[0145] The variation of the particle size-determining and particle size distribution-generating method according to the first embodiment is distinguished from the particle size-determining and particle size distribution-generating method of the first embodiment in that by utilizing the fact that in general, in an emission spot (corresponding to timing Ti in each of FIGS. 7A to 7E) where the intensity of an emission spectrum emitted e.g. by Fe as an element forming the ground of a test sample is equal to or smaller than the threshold value, there exist intermetallic inclusions other than Fe, the data sorting process of FIGS. 8 to 10 is executed with Fe as a trigger concerning the timing Ti.

[0146] In the following, points which distinguish the particle size-determining and particle size distribution-generating method according to the present variation of the method according to the first embodiment will be described with reference to the drawings.

[0147] Referring first to FIG. 4, an emission spot on the surface of a test sample has a circular shape of &phgr;30 &mgr;m. When no intermetallic inclusion exists in the surface of a measurement area of the test sample, an Fe emission spectrum, for example, emitted from Fe as a metal element which forms the ground of the test sample and has conductivity is obtained. When there exists an intermetallic inclusion in the surface of the measurement area of the test sample, an element X as a constituent element of the intermetallic inclusion emits light, which makes the intensity of the Fe emission spectrum lower than that in the case of no intermetallic inclusion existing in the surface of the measurement area of the test sample.

[0148] Further, referring to FIGS. 7A to 7E, when no intermetallic inclusion exists in the surface of the measurement area of the test sample, the emission spectrum intensity of Fe forming the ground of the test sample assumes a larger value than the corresponding threshold value. On the other hand, when there exists intermetallic inclusions in the surface of the measurement area of the test sample, for the reason described above, the emission spectrum intensity of Fe becomes lower than the threshold value, while the emission spectrum intensity of the element X as the constituent element of the intermetallic inclusions assumes a larger value than the corresponding threshold value.

[0149] In FIGS. 7A to 7E, emission spots which are identical in the number of times of generation of spark discharge generated thereat are indicated as timing Ti by broken lines.

[0150] For example, in timing T1 indicated by broken lines, it is estimated that there exists an intermetallic inclusion of Al2O3 which is a compound composed of Al and O, except the case where there exist elements other than an element or elements to be determined and the lens is deteriorated, referred to hereinafter. Similarly, in timing T2 indicated by broken lines, it is estimated that there exists intermetallic inclusions of Al2O3 and CaO each of which is a compound composed of more than one element selected from Al, O and Ca. Further, similarly, in timing T5 indicated by broken lines, where only the emission spectrum intensity of C is equal to or larger than the corresponding threshold value, it is estimated that there exists an intermetallic inclusion composed of a carbide which does not include Fe, O, Al and/or Ca as constituent elements.

[0151] FIG. 11 is a flowchart of a data sorting process which is executed by the variation of the particle size-determining and particle size distribution-generating method according to the first embodiment.

[0152] The FIG. 11 data sorting process is distinguished from the FIG. 8 data sorting process in that a row Ax is generated by deleting data of emission spots where the emission spectrum intensities of Fe are smaller than the threshold value specific to Fe (hereinafter referred to as “the threshold value Fe”) from each row Ix of data items of emission spectrum intensities specific to each of elements X as constituent elements of the intermetallic inclusions.

[0153] A description will be given of points distinguishing the FIG. 11 data sorting process from the FIG. 8 data sorting process executed by the particle size-determining and particle size distribution-generating method according to the first embodiment.

[0154] As shown in FIG. 11, first, a count i indicative of the number of data items in a row IFe in which are arranged data of the emission spectrum intensities of Fe, fetched from the memory, is set to 1, and a count ii indicative of the number of data items in the row IFe whose values are larger than the threshold value Fe is set to 0 (step S1101). Thereafter, one data item is fetched at random as a data item IFe(i) from the rows IFe stored in the memory in number corresponding to the number of times J (1000) of generation of spark discharge, and it is determined whether or not the fetched data item IFe(i) is equal to or smaller than the threshold value Fe (step S1102). If the fetched data item IFe(i) is determined to be larger than the threshold value Fe in the step S1102, the count ii is incremented by 1 (step S1103), followed by the process proceeding to a step S1110.

[0155] If the fetched data item IFe(i) is determined to be equal to or smaller than the threshold value Fe in the step S1102, a count K indicative of the number of data items IFe(i) whose values were determined to be equal to or smaller than the threshold value Fe is calculated by using the following equation (6) (step S1104):

K=i−ii  (6)

[0156] Then, in a step S1105, one of elements (e.g. O, Ca, C, Ti, Mn, S and N) existing in the surface of measurement area of the test sample is selected as a constituent element X, and then data items Ix(i) of the emission spectrum intensities of the constituent element X corresponding to emission spots of the fetched data items IFe(i) are read out from a row Ix of data of the emission spectrum intensities specific to the constituent element X, which are stored in the memory (step S1106).

[0157] Then, the fetched data items Ix(i) are substituted for data items Ax(K) in the row Ax (step S1107), and it is determined whether or not all the constituent elements existing in the surface of the measurement area have been selected. If not all the constituent elements have been selected in the step S1108, the process returns to the step S1105, whereas if all the constituent elements have been selected, the count K is regarded as a present maximum value of the count i at the time point that the data item IFe(i) was read out at random, and substituted for Kmax (step S1109).

[0158] Then, it is determined whether or not the present count i is smaller than the number of times J (1000) of generation of spark discharge (step S1110).

[0159] If the count i is smaller than the number of times J of generation of spark discharge, the count i is incremented by 1 (step S1111), followed by the process returning to the step S1102, whereas if the count i is equal to or larger than the number of times J of generation of spark discharge, a constituent element X which is one of the elements constituting the intermetallic inclusion identified in the step S605 is selected, followed by the process proceeding to the step S901 (FIG. 9).

[0160] It should be noted that when the count i is determined to be equal to or larger than the number of times J of generation of spark discharge in the step S1110, if the data row data Ax(K) of the constituent element X have to be subjected to the threshold-comparison process executed in the step S802 in FIG. 8 of the first embodiment, the process may proceed to the step S801 in FIG. 8 and be subjected to the processing in FIG. 8, and then proceed to the step S901 (FIG. 9).

[0161] However, the gist of the present variation of the particle size-determining and particle size distribution-generating method according to the first embodiment is as follows.

[0162] As the threshold value for data items IFe(i) is set to a lower value in the step S1102 in FIG. 11, the values of the data row data Ax(K) which are equal to or larger than the threshold value in the step S1107 become larger than the values of the data items Ix(i) and sufficiently exceed the threshold value in the step S802 in FIG. 8 (fully satisfy the condition of Ix(i)≧threshold value in the step S802 in FIG. 8). In other words, so long as the threshold value for data items IFe(i) in the step S1102 in FIG. 11 is properly set, the need for carrying out the threshold-comparison process executed in the step S802 in FIG. 8 can be substantially eliminated. As a result, the process can directly proceed to the step S901 (FIG. 9) without passing through the FIG. 8 process, so that it is possible to shorten the processing time required for generation of a particle size distribution and so forth.

[0163] According to the FIG. 11 process, the data items Ix(i) of emission spectrum intensities corresponding to the respective emission spots of the data items IFe(i) whose values are equal to or smaller than the threshold value Fe are taken out from each of the rows Ix (X=O, Ca, C, Ti, Mn, S, N, and so forth) of the emission spectrum intensities of all the elements existing in the measurement area on the surface of a test sample, which are stored in the memory, before execution of the data sorting process in FIGS. 9 and 10, which makes it possible to reduce the number of data items to be used in subsequent processing.

[0164] In particular, when the particle sizes and particle size distributions (frequencies) of a plurality of intermetallic inclusions are to be determined, by executing the FIG. 11 process once, it is possible to obtain rows Ax of a reduced number of data items of emission spectrum intensities of all the elements existing in the measurement area. Therefore, it is possible to eliminate the need for repeatedly carrying out the process in FIGS. 8 to 10 for executing the rearrangement in the step S607 in FIG. 6, thereby enabling quicker and more accurate determination of the particle sizes of the intermetallic inclusions and the particle size distribution of the intermetallic inclusions.

[0165] In the present process, when the data items Ix(i) of the intensities of the emission spectra of an element X as a constituent element of an intermetallic inclusion include no data items Ax(K) whose values are equal to or larger than the threshold value, it can be considered that there exist elements other than the element to be determined (e.g. a constituent element of a carbide, other than carbon) or the lens is stained or deteriorated.

[0166] A correction process related to the stain/deterioration of the lens will be described in detail as part of a third embodiment, given hereinafter. Needless to say, it is preferable that the correction process according to the third embodiment is executed together with the variation of the particle size-determining and particle size distribution-generating method according to the first embodiment.

[0167] Next, a particle size-determining and particle size distribution-generating method according to a second embodiment of the invention will be described in detail with reference to drawings.

[0168] A bearing steel actually used as a material for rolling members, such as a roller bearing, contains various kinds of intermetallic inclusions formed e.g. of Al2O3, MgO, MnS, CaO, and SiO2. Among these intermetallic inclusions, the Al2O3 inclusion, which is an oxide-based inclusion, most seriously affects the rolling life of the bearing steel. The particle sizes and particle size distribution of the Al2O3 inclusion, or the number of Al2O3 particles existing per unit area or volume, i.e. abundance thereof, are so closely related to the rolling life or the like of the bearing steel that so-called persons skilled in the art are very much interested in the particle sizes and particle size distribution and/or abundance of the Al2O3 inclusion (though the relationship between those and the rolling life of the bearing steel may be determined by another method).

[0169] The method of determining the particle sizes and particle size distribution or abundance of the Al2O3 inclusion includes two kinds of method, one of which is a three-dimensional method in which intermetallic inclusions are extracted from a sample of the bearing steel and the determination is carried out on the extracted intermetallic inclusions in a three-dimensional manner, such as an electron-beam elution method, and the other is a two-dimensional method in which the surface of a sample of the bearing steel is polished and the surface is subjected to the determination by a combination of an optical microscope and an image analyzer. However, the former method requires a large-scale apparatus for extraction of the Al2O3 inclusion, while the latter method is simple, but it is not capable of obtaining real particle sizes of the Al2O3 inclusion.

[0170] The particle size-determining and particle size distribution-generating method according to the second embodiment is also implemented by the FIG. 1 emission spectrometer 100 to determine the particle sizes and particle size distribution of intermetallic inclusions contained in a test sample cut out from a steel material.

[0171] As is distinct from the conventional EPMA method and a like method using a special master for determining the concentration of a trace amount of Al, the present method uses a real steel master cut out from a real steel material. The present method makes it a precondition that a calibration curve C, described hereinbelow, representative of the relationship between the concentration of Al and the emission spectrum intensity of Al is generated by using the FIG. 1 emission spectrometer 100.

[0172] It should be noted that inclusions contained in the bearing steel are not only the Al2O3 inclusion as described above, but since the Al2O3 inclusion is most closely related to the rolling life of the bearing steel, the following description will be given by taking the Al2O3 inclusion as an example. It goes without saying that other kinds of intermetallic inclusions can be dealt with similarly to the Al2O3 inclusion.

[0173] The particle size-determining and particle size distribution-generating method of the second embodiment is distinguished from that of the first embodiment in which the particle size of Al2O3 actually existing in the steel is determined by surface analysis by the EPMA method, in that the emission spectrometer 100 is used to determine the real particle size of the Al2O3 inclusion.

[0174] Further, the use of the calibration curve C representative of the relationship between the concentration of Al and the emission spectrum intensity of Al is effective in that it is possible to determine the particle sizes and particle size distribution of Al2O3 particles actually existing in a steel based on Al concentration (volume %) at least in a range of the order of 500 ppm to the order of percent.

[0175] A description will be given of a particle size-determining and particle size distribution-generating process executed by the FIG. 1 emission spectrometer 100 to implement the particle size-determining and particle size distribution-generating method of the second embodiment, with reference to the drawings.

[0176] FIG. 12 is a flowchart of the particle size-determining and particle size distribution-generating process executed according to the particle size-determining and particle size distribution-generating method of the second embodiment.

[0177] In FIG. 12, first, a calibration curves B, C generating process (step S1201), described hereinafter with reference to FIG. 13, and then a calibration curve D-forming process (step S1202), described hereinafter with reference to FIG. 16, a calibration curve E-generating process (step S1203), described hereinafter with reference to FIG. 19, and a particle size distribution-generating process (step S1204), described hereinafter with reference to FIG. 20, are executed, followed by terminating the process.

[0178] Next, the particle size-determining and particle size distribution-generating method of the second embodiment will be described in detail with reference to the drawings.

[0179] In the particle size-determining and particle size distribution-generating method according to the second embodiment, an EPMA is not used, but the real particle sizes of intermetallic inclusions contained in a reference sample equivalent to a test sample is determined by using the emission spectrometer 100, and a particle size distribution is generated based on the determined real particle sizes of the intermetallic inclusions.

[0180] In the following, the intermetallic inclusion particle size-determining process and the particle size distribution-generating process executed by the FIG. 1 emission spectrometer 100 will be described with reference to the drawings.

[0181] FIG. 13 is a flowchart of the calibration curves B, C generating process which is executed in the step S1201 in FIG. 12 for forming the calibration curve B representative of the relationship between the emission spectrum intensity of Fe and the concentration of Fe, and then forming the calibration curve C representative of the relationship between the emission spectrum intensity Al and the concentration of Al by the use of the calibration curve B.

[0182] The present calibration curve-forming process is carried out by the emission spectrometer 100 at least once before the method of determining the particle sizes of intermetallic inclusions and generating a particle size distribution is repeatedly executed on a test sample by the emission spectrometer 100. Seventeen types of Fe concentration masters (examples 1 to 17 shown in Table 1) are prepared whose Fe concentrations are already quantitatively determined e.g. by chemical analysis, such as an atomic absorption analysis method, and thus already known concentration values, and different from each other. These masters may be formed of an alloy steel, such as M50, 5Cr, SUH330, SUH310, or M50NiL. Alternatively, commercial masters sold by an external testing organization may be employed. 1 TABLE 1 Fe Concentration Example [Mass %] Emission Spectrum 1 46.13 7760 2 51.27 8802 3 59.63 9846 4 62.70 10098 5 65.57 10782 6 70.02 11634 7 77.69 12633 8 82.12 13570 9 85.04 13957 10 89.36 14769 11 93.28 15798 12 94.32 15854 13 95.34 15823 14 96.73 16033 15 97.39 16025 16 99.04 16246 17 99.59 16202

[0183] After each of these Fe concentration masters is held in the light emitting stand 102 of the emission spectrometer 100 of the present invention (step S1301), the counter electrode of the light-emitting section 101 generates spark discharge e.g. ten times. The Fe concentration masters subjected to spark discharge emit emission spectra (on a master-by-master basis) (step S1302).

[0184] Thereafter, the light of emission spectra emitted by each master is separated by the concave diffraction grating 112 into emission spectra specific to Fe as the ground of the Fe concentration master (step S1303), and the split Fe emission spectra are caused to enter the respective corresponding photomultipliers 113 via the exit slits 110.

[0185] Each photomultiplier 113 detects the incoming Fe emission spectrum and converts the intensity of the detected Fe emission spectrum to an electric current value and transmits the current value to the photometric section 104. The photometric section 104 converts the received current value to a digital value and then transmits the digital value obtained by the conversion to the data processing section 106 via the interface 105 (step S1304). On the other hand, the light emitting stand 102 transmits data of the positions of emission spots of &phgr;30 &mgr;m in an arbitrary measurement area on the surface of the Fe concentration master and the number of times of generation of spark discharge to the data processing section 106 via the interface 105.

[0186] Thus, data of the intensities of the emission spectra of Fe existing in the emission spots, the positions of the emission spots, and the number of times of generation of spark discharge are transmitted to the data processing section 106, and stored in the memory of the same.

[0187] Since the Fe concentration masters have already known Fe concentration values, these known concentration values are input to the data processing section 106 as data of Fe concentration. The data processing section 106 forms the calibration curve B (FIG. 14) representative of the relationship between the concentration of Fe and the emission spectrum intensity of Fe, based on the input data of Fe concentration [mass %] and the stored data of the emission spectrum intensity of Fe (step S1305).

[0188] The number of times of generation of spark discharge caused per Fe concentration master in the step S1302 is set to 10 times by way of example because it is preferable that the emission spectrum intensity is determined as an average value of emission spectrum intensities obtained ten times in this step. This makes it possible to make the average value of the emission spectrum intensities correspond to the Fe concentration of the one Fe concentration master.

[0189] Then, steps S1306 to S1313 in FIG. 13 are executed to form the calibration curve C representative of the relationship between the emission spectrum intensity of Al and the concentration of Al.

[0190] First, cylindrical real steel masters of &phgr;40 mm are cut out from a steel material for actual use, e.g. SUJ-2 in the form of a solid round rod, as shown in FIG. 15, and then held in the light emitting stand 102 (step S1306).

[0191] In the following, description will be given of properties of the steel material for actual use.

[0192] FIG. 15 is a cross-sectional view of a steel material, taken along a section orthogonal to the axis of the steel material for actual use, from which the real steel masters are cut out in the step S1306 in the FIG. 13.

[0193] In FIG. 15, the steel material 700 for actual use is produced by a method of blooming an ingot or a continuous method. When a molten material is cooled and solidified into the steel material 700 during the process of producing the same, in the core portion of the steel material 700, where the cooling rate is smaller than in the outer peripheral portion of the same, there occurs a phenomenon that intermetallic inclusions 400 are concentrated (center segregation). As shown in FIG. 15, the center segregation is conspicuous in an area (center segregated portion 500) within a range of 0.5 D with respect to the diameter D of the steel material 700, i.e. within a range of ±0.25 D from the center of the steel material 700, and more conspicuous toward the center even in this center segregated portion 500.

[0194] The center segregation is known to be correlated with the distance from the core portion, and hence by cutting out a portion of the steel material 700 including the core portion, along a plane perpendicular to the axis of the steel material 700, it is possible to obtain a real steel master containing Al2O3 inclusions varying in size with the distance from the core portion. The Al2O3 inclusions of various sizes contained in the real steel master include ones having Al concentrations at least in a range from the order of 500 ppm to the order of percent. Therefore, it is preferable to obtain a real steel master containing Al2O3 inclusions having various sizes by cutting out a portion of the steel material 700 including the core portion, along a plane perpendicular to the axis of the steel material 700, so as to obtain a calibration curve in an Al concentration region ranging from a low concentration to a high concentration thereof.

[0195] It should be noted that a plurality of real steel masters 500 containing Al2O3 inclusions of various sizes may be obtained by cutting out from the steel material 700 a plurality of pieces of the center segregated portion 500 in parallel with the axis of the steel material 700. Further, the steel material 700 may also contain intermetallic inclusions other than Al2O3.

[0196] Referring again to FIG. 13, after the real steel master is held (step S1306), the counter electrode of the light-emitting section 101 generates spark discharge e.g. one thousand times at an arbitrary location on a measurement area (&phgr;5 mm) on the surface of the real steel master such that the diameter of each spot is held to &phgr;30 &mgr;m (it is important that once arbitrarily set, the spot diameter should be held constant), the real steel master subjected to spark discharge emits emission spectra (step S1307). The intermetallic inclusions in the surface of the real steel master are dielectric, and hence at this time, the spark discharge is selectively guided to the intermetallic inclusions in the surface of the real steel master by the dielectric property thereof. Accordingly, the emission spectra generated by the real steel master are selectively emitted from the intermetallic inclusions. Therefore, when Al2O3 intermetallic inclusions exist in the surface of the real steel master, emission spectra containing information on Al and O as element information on Al2O3 can be obtained.

[0197] It should be noted that a description will be given of the following steps on the assumption that the intermetallic inclusions are Al2O3, and hence only emission spectrum intensities from respective emission spots each containing only Fe, Al and O are selected, and emission spectral information from emission spots containing Ca, Si, Mn, Mg, etc. is excluded later.

[0198] Thereafter, the emission spectra emitted from Al2O3 as the intermetallic inclusions of the real steel master are separated into emission spectra specific to the respective elements Fe, Al and O by the concave diffraction grating 112 (step S1308), and the separated emission spectra of the respective elements enter the respective photomultipliers 113 via the corresponding exit slits 110.

[0199] The photomultipliers 113 detect the incoming emission spectra of Fe, Al and O and convert the intensities of the detected emission spectra of the respective elements to electric current values and transmit the current values to the photometric section 104. The photometric section 104 converts each of the received current values to a digital value, and then transmits the digital value obtained by the conversion to the data processing section 106 via the interface 105 (step S1309). On the other hand, the light emitting stand 102 transmits data of the positions of emission spots of &phgr;30 &mgr;m on the measurement area of &phgr;5 mm on the surface of the real steel master or the distance from the center portion and the number of times of generation of spark discharge to the data processing section 106 via the interface 105.

[0200] Thus, data of the respective emission spectrum intensities of the elements Fe, Al and O existing in the emission spots of &phgr;30 &mgr;m, described above are transmitted to the data processing section 106 whenever spark discharge is carried out, and the data processing section 106 stores the received data in the memory thereof.

[0201] It should be noted that in the case where the intermetallic inclusions are Al2O3, even if information on elements other than Fe, Al and O is obtained at each time of generation of spark discharge, only the data on Fe, Al and O stored in the memory of the data processing section 106 may be used(the other data is excluded).

[0202] Then, by applying the intensities of the Fe emission spectra emitted from the real steel master to the FIG. 14 calibration curve B formed in the step S1305, Fe concentration data [mass %] concerning the Fe concentrations [mass %] of the real steel master is generated, and then the generated Fe concentration data is stored (e.g. in a manner associated with spot positions in a sequence of times of spark discharge) in the memory of the data processing section 106 such that the stored data can be read out for use when the calibration curve C representative of the relationship between the Al concentration and the Al emission spectrum intensity is formed in steps S1311 to S1312, referred to hereinbelow (step S1310).

[0203] Then, Al concentration data [mass %] of Al within Al2O3 as the intermetallic inclusions is calculated from the Fe concentration data generated in the step S1310 (step S1311). This arithmetic operation is carried out using the following equation (7) stored in the data processing section 106:

Al concentration=(100−Fe concentration)×54/102  (7)

[0204] wherein 54 represents the atomic weight (≈26.8×2) of Aluminum in Al2O3, 102 represents the formula weight (≈28.8×2+16.0×3) of Al2O3, and the Fe concentration represents a data item of Fe concentration data generated in the step S1310 which was obtained by the same spark discharge corresponding to one of particles of Al2O3 which was generated at that time.

[0205] As is apparent from the right side of the equation (7), simply by substituting one of the Fe concentration data items generated in the step S1310 into the equation (7), it is possible to calculate one of the data items of Al concentration in the Al2O3 inclusions existing in an emission spot of &phgr;30 &mgr;m subjected to the same spark discharge.

[0206] For example, when Fe concentration data items generated based on the calibration curve B concerning emission spots corresponding to respective spark discharges are 80, 45, and 20 [mass %], respectively, the concentration data items of Al in the Al2O3 inclusion existing in the respective emission spots as 10.6, 34.5 and 42.4 [mass %], respectively.

[0207] It should be noted that although the above equation (7) is for calculating Al concentration data in the case of the intermetallic inclusions being Al2O3, when the intermetallic inclusions are SiO2, CaO or MgO, the atomic weight of a corresponding one of SiO2, CaO and MgO and the formula weight of the oxide inclusions can be applied to the equation (7) so as to calculate the corresponding one of Si concentration, Ca concentration and Mg concentration.

[0208] If spark discharge is generated one thousand times, it is expected that approximately one thousand data items of the Al concentration data calculated as above are obtained. In short, a very large number of data items of Al concentration can be obtained with such a large number of times of generation of spark discharge. Then, the calibration curve C representative of the relationship between the Al concentration and the Al emission spectrum intensity is formed based on the Al concentration data items [mass %] and the emission spectrum intensities of Al (emitted from the real steel masters) at times of generation of spark discharge when the respective Al concentration data items were obtained (step S1312). Particularly in the present process, when a test sample of the center segregated portion 500 is used, a calibration curve C representative of the relationship between the Al concentration and the Al emission spectrum intensity in emission spots where the Al concentration is higher can be obtained, whereas when a test sample from a portion other than the center segregated portion 500 is used, a calibration curve C representative of the relationship between the Al concentration and the Al emission spectrum intensity in emission spots where the Al concentration is lower can be obtained.

[0209] Further, this real steel master has a ground composed of Fe, which is preferable in that Al can be more correctly identified. Thereafter, the formed calibration curve B is stored in the memory of the data processing section 106 (step S1313), followed by terminating the process.

[0210] According to the FIG. 13 process, it is possible to directly obtain a calibration curve C covering a range from a trace Al concentration in Al2O3 in the order of ppm to a considerably large concentration in the order of percent, not by extrapolation, but by using a real steel master.

[0211] Further, if the data of the Al2O3 particle sizes of the calibration curve A determined as described earlier with respect to the first embodiment, and the data of the Al emission spectrum intensities of the calibration curve C are stored in the memory of the data processing section 106, it is possible to use these data as desired when an Al2O3 particle size distribution is formed. The calibration curve C is used when the calibration curve A is applied to an embodiment which is a combination of the particle size-determining and particle size distribution-generating method according to the first embodiment and the particle size-determining (using a calibration curve E described hereinafter with reference to FIG. 21) and particle size distribution-generating method according to the second embodiment.

[0212] Moreover, when the real steel master is used, the Al concentration data are determined based on the Fe concentration data, and as a result, even when the ground of the real steel master contains Mg whose emission wavelength is close to that of Al, it is possible to distinguish Al in an emission spectrum of Al-contained inclusions from M therein, and hence form a calibration curve C more accurately than when an Al alloy containing more Mg is used as a master.

[0213] FIG. 16 is a flowchart of the calibration curve D-forming process which is executed in the step S1202 in FIG. 12 to form the calibration curve D concerning the emission spectrum intensity of Fe and the evaporation loss of Fe.

[0214] The present calibration curve forming process is executed by the emission spectrometer 100 at least once before the method of determining particle sizes and a particle size distribution of intermetallic inclusions contained in a test sample is repeatedly carried out by the emission spectrometer 100.

[0215] In FIG. 16, first, a pure master formed of a steel material containing no Al2O3 is held in the FIG. 1 emission spectrometer 100 (step S1601), and spark discharge is generated on the pure master held in Ar (step S1602), whereby spots of &phgr;30 &mgr;m (which is always held constant once set to the diameter, as described above) are generated. Then, emission spectra of Fe from the spots are separated (step S1603), and data of the emission spectrum intensities of Fe are transmitted (step S1604). It is preferred that the pure master is formed of pure iron. At the time of emission, it is possible to form spot marks having various sizes in the master by changing the discharge voltage from 10 kV by ±30%.

[0216] Then, spot marks 800 (FIG. 17) having various sizes formed at the time of the emission are observed three-dimensionally by a SEM to thereby measure the depth and diameter of each spot mark 800 (step S1605) and determine the volume thereof. Then, the mass (evaporation loss) of Fe which existed in each emission spot is calculated based on the volume of the spot mark 800 and the density of Fe (7.86 g/cm3), and the evaporation loss of Fe is input (step S1606).

[0217] The evaporation loss of Fe is proportional to the emission spectrum intensity of Fe obtained under a discharge condition which caused the Fe evaporation loss (FIG. 18). Therefore, the proportional relationship is adopted as a calibration curve D (step S1607), and the calibration curve D is stored in the memory of the data processing section 106 (step S1608). The formation of the calibration curve D is carried out at least once.

[0218] It should be noted that data obtained by averaging data items obtained when at most ten spark discharges are generated under the same spark charging-condition is sufficient to form the calibration curve D. The calibration curve D shows that the emission spectrum intensity of Fe is proportional to the evaporation loss of Fe.

[0219] The real particle size (particle size D) of an intermetallic inclusion, such as Al2O3, in an emission spot of &phgr;30 &mgr;m formed by a single spark discharge generated on a test sample (an object to be inspected) containing the intermetallic inclusion, such as Al2O3, can be calculated by using the calibration curve D. In the following, the calculation method will be explained.

[0220] FIG. 19 is a flowchart of the calibration curve E-forming process which is executed in the step S1203 in FIG. 12 to form the calibration curve E representative of the relationship between the emission spectrum intensity of Al and the Al particle size.

[0221] In the present calibration curve E-generating process, intensities of Al and Fe emission spectra emitted from the same emission spot of &phgr;30 &mgr;m whenever spark discharge is generated once at an area of &phgr;5 mm on the test sample surface of &phgr;40 mm are used to obtain the particle size D (radius) as the real particle size of an intermetallic inclusion (Al2O3) existing in the emission spot formed by the single spark discharge by utilizing the calibration curve C formed in the step S1312 in FIG. 13 and the calibration curve D formed in the step S1607 in FIG. 16. In the following, a description will be given by taking the case in which the intermetallic inclusions are Al2O3 as an example.

[0222] In FIG. 19, first, respective data items of the emission spectrum intensities of Fe, Al, (O) as information on the same emission spot transmitted in a step S604 in FIG. 20 are fetched (step S1901), and the fetched data item of Fe emission spectrum intensity is substituted into the calibration curve D stored in the memory of the data processing section 106 in the step S1202 in FIG. 12, whereby an evaporation loss [ng] of Fe is calculated (step S1902).

[0223] Then, the data item of Al emission spectrum intensity emitted from the same emission spot is substituted into the calibration curve C stored as an Al calibration curve in the memory of the data processing section 106 in the step S1201 in FIG. 12, whereby an Al concentration [mass %] is calculated (step S1903).

[0224] The mass M of Al2O3 is calculated from the evaporation loss [ng] of Fe and the Al concentration [mass %] determined as above, by using the following equation (8) (step S1904):

Al concentration [%]×10−2=M/(M+Fe evaporation loss)  (8)

[0225] Then, the volume V of Al2O3 is calculated from the mass M of Al2O3 existing in a spot mark of &phgr;30 &mgr;m corresponding to a unit volume and the density of Al2O3 (3.90 g/cm3), by using the following equation (9) (step S1905):

V=M/3.90  (9)

[0226] Next, the Al2O3 particle having the volume V is regarded as a perfect sphere, and the particle size D (diameter) of the Al2O3 particle regarded as the perfect sphere is calculated by using the following equation (10):

D/2=(3V/4&pgr;)1/3  (10)

[0227] In the FIG. 19 process, similarly to the first embodiment, the concentration of Al contained in intermetallic inclusions can be determined based on the Al emission spectrum intensity obtained by emission spectral analysis, from the calibration curve C representative of the relationship between the Al emission spectrum intensity and the Al2O3 particle size, so that it is possible to obtain the particle size D of the Al2O3 particle regarded as the perfect sphere. The Al2O3 particle size can be determined based on the corresponding Al emission spectrum intensities by forming the calibration curve E, i.e. this relationship which is a correspondence between the concentration of Al as a constituent element of the intermetallic inclusions and the Al2O3 particle size.

[0228] The mass M of Al2O3 calculated by the data processing section 106 based on the relationship expressed by the equation (8) can be also expressed by an equation (8′) given hereinbelow by rearranging the equation (8) as described below.

[0229] First, as described hereinbefore as to the step S1903, the Al concentration [%] on the left side is determined as a concentration [%] of Al existing in the spot mark of &phgr;30 &mgr;m corresponding to the unit volume, by substituting the emission spectrum intensity of Al emitted from the spot mark into the calibration curve C.

[0230] Similarly, as described hereinbefore as to the step S1902, the Fe evaporation loss [ng] on the right side is determined as an evaporation loss [ng] of Fe which existed in the spot mark of &phgr;30 &mgr;m corresponding to the unit volume, by substituting the emission spectrum intensity of Fe emitted from the spot mark into the calibration curve D obtained based on the spark discharge at emission spots having the same diameter.

[0231] Now, the concentration [%] of Al existing in the spot mark of &phgr;30 &mgr;m corresponding to the unit volume can be converted to the mass A [ng] of Al. In the calibration curve C, the concentration of Al is determined for each spot of &phgr;30 &mgr;m, and this is the same with a test sample, so that the weight of Al existing within an area of &phgr;30 &mgr;m can be considered both for the calibration curve C and for the test sample.

[0232] The mass A [ng] of Al can be obtained as a ratio of the mass A [atomic weight of 26.8, formula weight of 54] of Al2 to the mass M of Al2O3 (formula weight of 102) as follows:

A=0.53M  (11)

[0233] Using 0.53 in the above equation (11), the equation (8) can be rearranged for the mass M of Al2O3 as follows:

M=Al concentration [%]×Fe evaporation loss/(0.53−Al concentration [%])×10−4  (8′)

[0234] In this way, the mass M of Al2O3 existing in the spot mark of &phgr;30 &mgr;m corresponding to the unit volume can be determined.

[0235] As described above, the calibration curve E (FIG. 21) for use in calculating Al2O3 particle sizes based on the emission spectrum intensity of Al as a constituent element of the intermetallic inclusions can be formed (step S1907). The formed calibration curve E is stored (step S1908), followed by terminating the calibration curve E-forming process which is executed in the step S1203 in FIG. 12. By storing the calibration curve E in the memory of the data processing section 106 in the step S1203 in FIG. 12, it is possible to dispense with arithmetic operations by the equations (8) to (10).

[0236] It should be noted that as emission spectrum intensities of the test sample used in forming the calibration curve E in the FIG. 19 process, emission spectrum intensities measured in steps S601 to S604 in FIG. 20, hereinafter referred to, may be utilized. Needless to say, when the calibration curve E has already been stored, it is possible to easily calculate the particle sizes D of intermetallic inclusions contained in the test sample from the emission spectrum intensities of the test sample measured in a FIG. 20 process, described below, based on the calibration curve E.

[0237] FIG. 20 is a flowchart of the particle size distribution-generating process which is executed in the step S1204 in FIG. 12.

[0238] It should be noted that steps S601 to S605 and steps S607 to S608 are identical to those of the FIG. 6 particle size distribution-generating process of the first embodiment.

[0239] In FIG. 20, first, a test sample (an object to be inspected) is held (step S601), and spark discharge is generated e.g. one thousand times (step S602), emission spectra are separated (step S603), and data of emission spectrum intensities are transmitted (step S604). Then, similarly to the processing in the step S605 in FIG. 6, intermetallic inclusions are identified (step S605). Since this example is a case where attention is paid to the Al2O3 inclusions as intermetallic inclusions, from the emission spectrum intensities of lots of elements emitted by the spark discharge, only data containing information on Al, O, and Fe alone are extracted as follows.

[0240] Now, when a spark discharge is generated once, an area of &phgr;30 &mgr;m containing Al2O3 is selectively changed into an emission (discharge) spot due to the dielectric property of Al2O3. In the spot mark, only Fe as the ground has evaporated to disappear, whereas only Al2O3 having a higher melting point than that of Fe remains without being evaporated. It should be noted that out of the intensities of emission spectra of lots of elements emitted by the spark discharge, attention is paid to only data containing information on Al, O, and Fe alone, and hence it can be considered that only Al2O3 except for Fe exists in the emission spot of &phgr;30 &mgr;m subjected to a single spark discharge.

[0241] Thereafter, an intermetallic inclusion particle size-calculating process for calculating the particle size D of the Al2O3 from the Al emission spectrum intensity based on the calibration curve E is (step S2006). Then, the obtained particle size D of the Al2O3 is input as Ax(K) in the FIG. 8 data sorting process, described hereinbefore, and after execution of the FIG. 9 data sorting process for rearranging particle sizes in size-increasing order and the FIG. 10 data sorting process for determining frequencies (step S607), the data processing section 106 generates a diagram shown in FIG. 27A and stores the same in the memory, and at the same time displays the FIG. 27A diagram on the terminal unit(step S608).

[0242] According to the FIG. 20 process, not the apparent diameter but the particle size D as the real particle size (sphere diameter) is determined, and as a result, it is possible to determine the particle sizes of intermetallic inclusions more accurately and generate a more accurate particle size distribution of the intermetallic inclusions.

[0243] According to the second embodiment, as shown in the FIG. 27A diagram of Al2O3 particle size distribution, it is possible to extract lots of Al2O3 inclusions having small particle sizes in a particle size distribution from a steel as a material for roller bearings, and to extract lots of Al2O3 inclusions within a particle size range of 3 to 13 &mgr;m which affect the rolling life of the roller bearings, as well as to extract a large number of intermetallic inclusions as a whole. This makes it possible to determine particle sizes more accurately and generate a more accurate particle size distribution. These merits can be utilized as a correct index for predicting the rolling life of the roller bearings and for determining how to set the pureness of a steel material for a longer rolling life thereof.

[0244] In the process of the step S605 in FIG. 20, when the data of the intensities of the emission spectra of the elements X as constituent elements of intermetallic inclusions include no data items whose values are equal to or larger than the respective corresponding threshold values, it is presumed that there exist elements not to be determined (e.g. elements composing carbides, other than carbon), or that the lens is stained or deteriorated.

[0245] The correction process related to the stain/deterioration of the lens will be described in detail in the description of the third embodiment, given hereinafter. Needless to say, it is preferable that the correction process according to the third embodiment is executed in combination with the second embodiment.

[0246] Next, a particle size-determining and particle size distribution-generating method according to the third embodiment will be described in detail with reference to the drawings.

[0247] Next, a particle size-determining and particle size distribution-generating method according to the third embodiment will be described in detail with reference to the drawings.

[0248] The particle size-determining and particle size distribution-generating method of the third embodiment is also carried out by the FIG. 1 emission spectrometer 100, in generating a particle size distribution of intermetallic inclusions contained in a test sample cut out from a steel material.

[0249] The particle size-determining and particle size distribution-generating method of the third embodiment is distinguished from the particle size-determining and particle size distribution-generating method of the first and second embodiments in that compensation is made for attenuation of emission spectrum intensities due to stains on the condensing lens 108.

[0250] A description will now be given of the intermetallic inclusion particle size-determining and particle size distribution-generating process executed by the FIG. 1 emission spectrometer 100 with reference to the drawings.

[0251] FIG. 22 is a flowchart of the intermetallic inclusion particle size-determining and particle size distribution-generating process according to the third embodiment.

[0252] In FIG. 22, first, the FIG. 2 calibration curve A-forming process in the first embodiment is executed (step S201), a correction curve generating process, described hereinafter with reference to FIG. 23, is executed (step S2201), and a particle size distribution-generating process, described hereinafter with reference to FIG. 25, is executed (step S2202).

[0253] It should be noted that the process in the step S201 has been described with reference to FIG. 2, and hence description thereof is omitted. Further, as the calibration curve A, there may be used one used at least once in the FIG. 3 calibration curve A-forming process and stored in the memory of the data processing section 106.

[0254] FIG. 23 is a flowchart of the intensity correction curve-generating process which is executed in the step S2201 in FIG. 22.

[0255] The present process is also executed by the emission spectrometer 100 at least once before the particle size distribution-generating process for generating a particle size distribution of intermetallic inclusions in a test sample is repeatedly carried out by the emission spectrometer 100 (e.g. when the emission spectrometer 100 is installed in a quality inspection line).

[0256] In FIG. 23, first, after a master cut out from a steel material formed e.g. of SUJ2 containing Al2O3 is held in the light emitting stand 102 (step S2301), the surface of the master is scanned by the EPMA to thereby determine particle sizes of Al2O3 existing in the surface of the master (step S2302). Further, an Al2O3 particle whose particle size is the closest to 15 &mgr;m of all the determined particle sizes is selected (step S2303), and data of the location of the selected Al2O3 particle is transmitted to the data processing section 106 (step S2304), and the data processing section 106 stores the received data of the location in the memory thereof.

[0257] Then, spark discharge is repeatedly generated on the selected Al2O3 particle by the counter electrode of the light-emitting section 101, and whenever spark discharge is generated one thousand times, the intensity of an emission spectrum then emitted from the selected Al2O3 particle is measured (step S2305). Data of the measured emission spectrum intensity is transmitted to the data processing section 106 via the interface 105 (step S2306). On the other hand, the light emitting stand 102 transmits data of the number of times of generation of spark discharge to the data processing section 106 via the interface 105. Thereafter, the data processing section 106 stores the data of the emission spectrum intensity and the data of the number of times of generation of spark discharge which have been received in the memory thereof.

[0258] Then, the data processing section 106 forms a intensity correction curve, described hereinafter with reference to FIG. 14, which is representative of the relationship between the number of times of generation of spark discharge and the amount of attenuation of emission spectrum intensity, based on the data of the number of times of generation of spark discharge and the emission spectrum intensities which have been stored in the memory (step S2307), and stores the formed intensity correction curve in the memory (step S2308), followed by terminating the present process.

[0259] FIG. 24 is a diagram showing the intensity correction curve formed in the step S2201 in FIG. 22.

[0260] In FIG. 24, the amount of attenuation of emission spectrum intensity is calculated as the difference between an emission spectrum intensity corresponding to a 0-th spark discharge and an emission spectrum intensity corresponding to each 1000-th spark discharge, and indicated as a correction value on the ordinate. When the number of times of generation of spark discharge is represented by i, an emission spectrum intensity before correction by I(i), and an emission spectrum intensity after correction by I′(i), I′(i) is expressed by the following equation (12):

I′(i)=I(i)+0.107i  (12)

[0261] It should be noted that a test sample can contain intermetallic inclusions other than Al2O3, it is desirable that the FIG. 25 intensity correction curve should be also formed for each of the other intermetallic inclusions, such as CaO.

[0262] According to the FIG. 23 process, an Al2O3 particle whose particle size is as large as the size of an Al2O3 particle contained in a test sample and the closest to 15 &mgr;m is selected (step S2303), and the intensity of an emission spectrum emitted from the selected Al2O3 particle is measured whenever spark discharge is generated one thousand times (step S2305), whereby the intensity correction curve representative of the relationship between the number of times of generation of spark discharge and the amount of attenuation of emission spectrum intensity is formed based on the data of the number of times of generation of spark discharge and the emission spectrum intensity (step S2307). Therefore, the formed intensity correction curve is based on the Al2O3 particle having a particle size closest to that of the Al2O3 particle contained in the test sample. This enables the correction of the intensity of emission spectra condensed by the condensing lens 108 to provide a reliable emission spectrum intensity after the correction. Further, even if the condensing lens 108 is not cleaned over a predetermined time period (corresponding to approximately 3000 times of generation of spark discharge), it is possible to accurately predict an emission spectrum intensity which could be obtained if the emission spectrum were condensed by the clean condensing lens 108, based on the number of times of generation of spark discharge.

[0263] FIG. 25 is a flowchart of the particle size distribution-generating process which is executed in the step S2202 in FIG. 22.

[0264] The present process is executed by the emission spectrometer 100 whenever particle size distribution of intermetallic inclusions contained in a test sample is repeatedly generated after the FIG. 23 process by the emission spectrometer 100 is executed at least once.

[0265] In FIG. 25, steps S601 to S605 and steps S606 to S608 are identical to those of the FIG. 6 process, and hence description thereof is omitted.

[0266] First, the steps S601 to S605 are executed. Then, the data processing section 106 corrects data of the emission spectrum intensity of each element stored in the memory, based on the data of the number of times of generation of spark discharge and the intensity correction curve stored in the memory (step S2501), and stores the corrected data of the emission spectrum intensity of each element in the memory. Then, the steps S606 to S608 are executed, followed by terminating the present process.

[0267] According to the FIG. 25 process, the data processing section 106 corrects the data of the emission spectrum intensity of each element stored in the memory, based on the data of the number of times of generation of spark discharge and the intensity correction curve stored in the memory (step S2501), it is possible to correct the emission spectrum in real time.

[0268] Although in the above described third embodiment, the attenuation of emission spectrum intensity due to stains on the condensing lens 108 is corrected based on the particle size-determining and particle size distribution-generating method according to the first embodiment, it may be corrected based on the particle size-determining and particle size distribution-generating method according to the variation of the first embodiment or the second embodiment.

[0269] In the first to third embodiments of the invention, in the processes in FIGS. 3, 12 and 19, the data processing section 106 may store calibration curves for intermetallic inclusions other than the Al2O3 (e.g. CaO, SiO, MnS, etc.), or alternatively in the FIG. 23 process, the data processing section 106 may store calibration curves and intensity correction curves for the intermetallic inclusions other than Al2O3, to thereby generate particle size distributions of a plurality of kinds of intermetallic inclusions contained in the test sample at a time. Further, by providing sufficient calibration curves and intensity correction curves for intermetallic inclusions, when it is found that intermetallic inclusions contained in a test sample are composed of two or more kinds of elements, it is possible to more easily identify whether constituent elements are single substances, or form a compound or a mixture.

[0270] Further, although in the above embodiments of the present invention, in the FIG. 9 data sorting process, the row ALx in which the data items Ix are rearranged in intensity-increasing order is obtained, this is not limitative, but a row ALx in which the data items Ix are rearranged in intensity-decreasing order may be obtained.

[0271] Moreover, although the arithmetic operations, such as the data sorting, are carried out by the emission spectrometer 100 and the data processing section 106, this is not limitative, but the arithmetic operations may be carried out by an arithmetic processor, a storage medium, or any other device which is capable of storing program modules for the arithmetic operations or executing programs for the arithmetic operations, in place of the emission spectrometer 100 and the data processing section 106, or alternatively, a combination of these devices may be used.

[0272] In the following, a first example of the present invention will be described in detail.

[0273] In the first example of the present invention, the particle size-determining and particle size distribution-generating method according to the first embodiment was carried out.

[0274] A material of SUJ2 having a relatively high degree of pureness was prepared, and a cylindrical test sample having a diameter of &phgr;40 mm was cut out from the prepared SUJ2 material. Then, after the FIG. 2 process was executed to thereby generate a particle size distribution of Al2O3 existing in an area of &phgr;5 mm at an arbitrary location on the surface of the test sample, the determined area of &phgr;5 mm was scanned by the EPMA as well, whereby the particle size distribution of Al2O3 was generated and confirmed.

[0275] FIGS. 26A and 26B are diagrams for comparison between the Al2O3 particle size distribution (a) generated by execution of the FIG. 2 process and the particle size distribution (b) generated by the EPMA.

[0276] As is apparent from FIGS. 26A and 26B, it was recognized that there is a close agreement between the Al2O3 particle size distribution (a) generated by execution of the FIG. 2 process and the particle size distribution (b) generated by the EPMA, and hence it was found that the intermetallic inclusion particle size distribution-generating method of the present invention is as accurate as the conventional method utilizing the EPMA.

[0277] Further, it was found that in the method in which the Al2O3 particle size distribution is generated by execution of the FIG. 2 process, it takes only 60 seconds (one minute) or so to generate the distribution, which means that the intermetallic inclusion particle size distribution-generating method of the present invention is capable of quickly generating a particle size distribution of intermetallic inclusions.

[0278] A second example of the invention will now be described in detail.

[0279] In the second example of the present invention, the particle size-determining and particle size distribution-generating method according to the second embodiment was carried out.

[0280] Cylindrical test samples each having a diameter of &phgr;40 mm were cut out, respectively, from different steel materials of 17 types, of which the Fe concentration had been previously determined and thus known, and the emission spectrum intensity of Fe contained in each of the test samples was measured for execution of the calibration curve B-forming process in the step S1201 in FIG. 12 (Examples 1 to 17). The results of the measurements are shown in Table 1 given hereinbefore and FIG. 14. It should be noted that intermetallic inclusions contained in the 17 types of steel materials were not limited to Al2O3 and the ground thereof was Fe.

[0281] From Table 1 and FIG. 14, it was found that it is possible to obtain a calibration relationship (calibration curve B) between the Fe concentration [mass %] and the Fe emission spectrum intensity, as shown in FIG. 14, even from different types of steel materials so long as each steel material has Fe as the ground, particularly when the Fe concentration is low, i.e. when the concentration of intermetallic inclusions is high.

[0282] According to the process which is executed in the step S1201 in FIG. 12, the calibration curve B can be obtained even when Fe concentration is low, i.e. when the concentration of intermetallic inclusions is high, and it is possible to determine Al concentration with ease when there is no intermetallic inclusion other than Al2O3. Therefore, the calibration curve B and the calibration curve C can be directly obtained by the emission spectrometer 100 alone, without using the calibration curve A as shown in FIG. 5, which is formed by using the EPMA, more specifically, by using a steel material for actual use and without extrapolating a calibration curve covering a range of Al concentration in Al2O3 from a trace concentration in the order of ppm to a considerably large concentration in the order of percent.

[0283] After the execution of the step S1201 in FIG. 12, a cylindrical test sample having a diameter of &phgr;40 mm was further cut out from a steel material containing no Al2O3, and the step S1202 in FIG. 12 was executed. Then, discharge voltage to be applied to the test sample for emission spectral analysis was changed from 10 kV by ±30% to thereby determine the volumes of spot marks through the three-dimensional SEM observation, and the relationship between the amount of evaporation of Fe calculated based on the volumes of the spot marks and the density of Fe and the spot diameter of the spot marks was examined. The results of the examination are shown in Table 2 and FIG. 18. 2 TABLE 2 Spot Diameter Evaporation Loss Emission Spectrum [&mgr;m] [ng] Intensity 5 0.30 1406 10 0.50 2836 15 0.90 5121 20 1.20 7187 25 1.40 9003 30 1.65 10872 35 1.90 13569 40 2.10 16419

[0284] It was proved from Table 2 and FIG. 18 that there is a proportional relationship (calibration curve D) between the amount of Fe evaporation caused by a single spark discharge and the emission spectrum intensity of Fe.

[0285] According to the above process, it was proved that it is possible to obtain perfect-sphere radiuses (particle sizes D) of intermetallic inclusions, such as Al2O3, contained in the test sample (an object to be inspected) in which intermetallic inclusions, such as Al2O3, exist, based on the calibration curve D and the Al concentration determined from the above calibration curve C.

[0286] Further, in the process which is executed in the step S1204 in FIG. 12, a cylindrical test sample (an object to be inspected) having a diameter of &phgr;40 mm was cut out from a SUT-2 material to be actually examined, and the particle size distribution of Al2O3 existing in an area of &phgr;5 mm at an arbitrary location on the surface of the object to be inspected was generated by execution of the FIG. 12 process. Then, a particle size distribution of Al2O3 in the determined area of &phgr;5 mm was also generated by the image analysis method and confirmed.

[0287] FIGS. 27A and 27B are diagrams for comparison between the Al2O3 particle size distribution (a) generated by execution of the FIG. 12 process and the particle size distribution (b) generated by the image analysis method.

[0288] As is apparent from FIGS. 27A and 27B, the comparison between the Al2O3 particle size distribution (a) generated by execution of the FIG. 12 process and the particle size distribution (b) generated by the image analysis method showed that the Al2O3 particle size distribution (a) generated by execution of the FIG. 12 process is more accurate than the particle size distribution (b) generated by the image analysis method in that lots of Al2O3 particles having small particle sizes are extracted in the particle size distribution concerning the Al2O3 inclusion in the steel as a material for roller bearings; lots of Al2O3 particles within a range of 3 to 13 &mgr;m which affect the rolling life of the roller bearings are extracted; and a larger number of intermetallic inclusions are extracted as a whole.

[0289] This shows that the FIG. 12 process gives more accurate results than the conventional extrapolation method in which high Al concentrations in the order of percent are extrapolated into a calibration curve, since the execution of the former provides the calibration curve B indicative of the concentration of Al in intermetallic inclusions covering up to high Al concentrations in the order of percent by using the actually used SUJ-2 material and the calibration curve D which estimates the real particle sizes of intermetallic inclusions, such as Al2O3, three-dimensionally and correctly. Further, since the FIG. 12 process is executed simply by using the emission spectral analysis, without preparing special masters or using the EPMA method, it is possible to provide a simple and easy particle size-determining and particle size distribution-generating method.

[0290] Industrial Applicability

[0291] According to the invention as claimed in claims 1 and 11, an intermetallic inclusion particle size-intensity calibration curve representative of the relationship between particle size of intermetallic inclusions and emission spectrum intensity of a constituent element of the intermetallic inclusions is formed. This makes it possible to identify what form is assumed by the intermetallic inclusions from elements constituting the intermetallic inclusions and quickly and accurately determine or measure particle sizes and particle size distribution of the intermetallic inclusions.

[0292] According to the apparatuses as claimed in claims 2 and 12, the particle size of the intermetallic inclusions in the predetermined area of the reference sample is determined through surface analysis by an electron probe microanalyzer. This enables quick and accurate determination or measurement of the particle size and particle size distribution of the intermetallic inclusions.

[0293] According to the invention as claimed in claims 3 and 13, there are formed a principle component known concentration-intensity calibration curve representative of the relationship between the emission spectrum intensity of the principle component having an already known concentration and the known concentration of the principle component, a real steel material-contained intermetallic inclusion constituent element concentration-intensity calibration curve representative of the relationship between the concentration of the constituent element of the intermetallic inclusions and the emission spectrum intensity of the constituent element of the intermetallic inclusions, and a base element evaporation amount-intensity calibration curve representative of the relationship between the base element evaporation amount and the intensity of emission spectra emitted from the base element. This enables quick and more accurate determination or measurement of real particle sizes and particle size distribution of the intermetallic inclusions.

[0294] According to the invention as claimed in claims 4 and 14, there is formed an intermetallic inclusion particle size-intensity calibration curve representative of the relationship between the calculated particle size of the intermetallic inclusions and the determined emission spectrum intensity of the constituent element of the intermetallic inclusions. This enables quicker and more accurate determination of real particle sizes and particle size distribution of the intermetallic inclusions.

[0295] According to the invention as claimed in claims 5 and 15, a data sorting process for counting the number of data items is executed to thereby generate a particle size distribution of the intermetallic inclusions in the test sample. This makes it possible to identify what form is assumed by the intermetallic inclusions from elements constituting the intermetallic inclusions and quickly and accurately determine or measure particle sizes and particle size distribution of the intermetallic inclusions.

[0296] According to the invention as claimed in claims 6 and 16, the data items of emission spectra of the constituent element of intermetallic inclusions in the test sample are rearranged in order of intensity, and then the number of the rearranged data items is counted. This makes it possible to facilitate processing of the data.

[0297] According to the invention as claimed in claims 7 and 17, data items of emission spectrum intensity of a constituent element of intermetallic inclusions in the test sample to be rearranged in order of intensity are extracted by determining whether or not an emission spectrum intensity of the constituent element of the intermetallic inclusions in the test sample is larger than a threshold value. This makes it possible to reduce the number of the data.

[0298] According to, according to the invention as claimed in claims 8 and 18, the data items of emission spectrum intensity of the constituent element of the intermetallic inclusions in the test sample to be rearranged in order of intensity are extracted based on a result of comparison between the emission spectrum intensity of a principle component of the test sample and the emission spectrum intensity of the constituent element of the intermetallic inclusions in the test sample. This makes it possible to reduce the number of the data and quickly extract data items to be rearranged in order of intensity.

[0299] According to the invention as claimed in claims 9 and 19, emission spectrum intensity of a constituent element of the intermetallic inclusions in the test sample is corrected according to the number of times of generation of spark discharge. This enables quicker and more accurate determination or measurement of particle sizes and particle size distribution of the intermetallic inclusions.

[0300] According to the invention as claimed in claims 10 and 20, a kind of the constituent element of the intermetallic inclusions is identified based on a result of comparison between the intensity of emission spectra of a principle component of the test sample and the intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample. This makes it possible to positively reduce the number of data items and more quickly and more accurately determine or measure particle sizes and particle size distribution of the intermetallic inclusions.

Claims

1. A particle size-determining method for determining a particle size of intermetallic inclusions based on an emission spectrum intensity of a constituent element of the intermetallic inclusions,

the particle size-determining method comprising the steps of:

determining a particle size of the intermetallic inclusions, which is known, in a predetermined area of a reference sample;

determining an intensity of emission spectra emitted from the constituent element of the intermetallic inclusions, from a relationship thereof to an intensity of emission spectra from spark discharge spots within the predetermined area, through emission spectral analysis of the reference sample; and

forming an inclusion particle size-intensity calibration curve representative of relationship between the particle size of the intermetallic inclusions and the emission spectrum intensity of the constituent element of the intermetallic inclusions.

2. A particle size-determining method according to claim 1, wherein the particle size of the intermetallic inclusions, which is known, in the predetermined area of the reference sample is determined through surface analysis by an electron probe microanalyzer.

3. A particle size-determining method for determining a particle size of intermetallic inclusions based on an emission spectrum intensity of a constituent element of the intermetallic inclusions,

the particle size-determining method comprising the steps of:

determining an intensity of emission spectra emitted from a principle component having an already known concentration and contained in a reference sample having an already known concentration in spark discharge spots within a predetermined area of the reference sample, through emission spectral analysis of the reference sample;

forming a principle component known concentration-intensity calibration curve representative of relationship between the emission spectrum intensity of the principle component having an already known concentration and the known concentration of the principle component;

determining an intensity of emission spectra emitted from a principle component contained in a real steel material-reference sample in spark discharge spots within a predetermined area of the real steel material-reference sample, and an intensity of emission spectra emitted from a constituent element of intermetallic inclusions contained in the real steel material-reference sample, through emission spectral analysis of the real steel material-reference sample;

calculating a concentration of the principle component contained in the real steel material-reference sample from the emission spectrum intensity of the principle component of the real steel material-reference sample, based on the principle component known concentration-intensity calibration curve;

calculating a concentration of the constituent element of the intermetallic inclusions contained in the real steel material-reference sample, based on the calculated concentration of the principle component of the real steel material-reference sample;

forming a real steel material-contained intermetallic inclusion constituent element concentration-intensity calibration curve representative of relationship between the concentration of the constituent element of the intermetallic inclusions and the emission spectrum intensity of the constituent element of the intermetallic inclusions;

determining an intensity of emission spectra emitted from a base element of a real steel material-pure base sample in spark discharge spots within a predetermined area of the real steel material-pure base sample, and a base element evaporation amount indicative of mass of the base element having been evaporated due to spark discharge thereon, through emission spectral analysis of the real steel material-pure base sample; and

forming a base element evaporation amount-intensity calibration curve representative of relationship between the base element evaporation amount and the intensity of emission spectra emitted from the base element.

4. A particle size-determining method according to claim 3, comprising the steps of:

calculating a base element evaporation volume indicative of a volume of the evaporated base element, from the base element evaporation amount based on a density of the base element, and calculating, from the base element evaporation volume, a particle size of the intermetallic inclusions as a diameter thereof corresponding to the base element evaporation volume, based on a formula of calculating a spherical volume;

determining a known concentration of the principle component from the emission spectrum intensity of the base element, based on the principle component known concentration-intensity calibration curve;

calculating a concentration of the constituent element of the intermetallic inclusions based on the determined known concentration of the principle component;

determining an intensity of emission spectra of the constituent element of the intermetallic inclusions from the calculated concentration of the constituent element of the intermetallic inclusion, based on the real steel material-contained intermetallic inclusion constituent element concentration-intensity calibration curve; and

forming an intermetallic inclusion particle size-intensity calibration curve representative of relationship between the calculated particle size of the intermetallic inclusions and the determined emission spectrum intensity of the constituent element of the intermetallic inclusions.

5. A particle size distribution-generating method for generating a particle size distribution of intermetallic inclusions, comprising the steps of:

executing a data sorting process for counting a number of data items of emission spectra of a constituent element of the intermetallic inclusions in a test sample; and

generating a particle size distribution based on the counted number of the data items and the particle size of the intermetallic inclusions in the test sample, which have been determined by the particle size-determining method according to any of claims 1 to 4.

6. A particle size distribution-generating method according to claim 5, wherein in the data sorting process, the data items of emission spectra of the constituent element of the intermetallic inclusions in the test sample are rearranged in order of intensity, and then the number of the rearranged data items is counted.

7. A particle size distribution-generating method according to claim 5 or 6, wherein in the data sorting process, it is determined whether or not an emission spectrum intensity of the constituent element of the intermetallic inclusions in the test sample is larger than a threshold value, and then data items of intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample to be rearranged in order of intensity are extracted based on a result of the determination.

8. A particle size distribution-generating method according to claim 5 or 6, further comprising the step of determining an intensity of emission spectra emitted from a principle component contained in the test sample; and

wherein in the data sorting process, the data items of intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample to be rearranged in order of intensity are extracted based on a result of comparison between the intensity of emission spectra emitted from the principle component contained in the test sample and the intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample.

9. A particle size distribution-generating method according to any of claims 5 to 8, further comprising the steps of:

forming an intensity correction curve representative of relationship between a number of times of generation of spark discharge for emission spectral analysis and an amount of attenuation of an intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample after carrying out the spark discharge; and

correcting the intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample according to the number of times of generation of spark discharge for emission spectral analysis, based on the generated intensity correction curve.

10. A particle size distribution-generating method according to any of claims 5 to 9, wherein a kind of the constituent element of the intermetallic inclusions contained in the test sample is identified based on a result of comparison between the intensity of emission spectra emitted from the principle component contained in the test sample and the intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample.

11. A particle size-determining apparatus for determining a particle size of intermetallic inclusions based on an emission spectrum intensity of a constituent element of the intermetallic inclusions,

the particle size-determining apparatus comprising:

acquiring means for acquiring a particle size of the intermetallic inclusions, which is known, in a predetermined area of a reference sample;

acquiring means for acquiring an intensity of emission spectra emitted from the constituent element of the intermetallic inclusions, from a relationship thereof to an intensity of emission spectra from spark discharge spots within the predetermined area, through emission spectral analysis of the reference sample; and

forming means for forming an inclusion particle size-intensity calibration curve representative of relationship between the particle size of the intermetallic inclusions and the emission spectrum intensity of the constituent element of the intermetallic inclusions.

12. A particle size-determining apparatus according to claim 11, comprising acquiring means for acquiring the particle size of the intermetallic inclusions, which is known, in the predetermined area of the reference sample through surface analysis by an electron probe microanalyzer.

13. A particle size-determining apparatus for determining a particle size of intermetallic inclusions based on an emission spectrum intensity of a constituent element of the intermetallic inclusions,

the particle size-determining apparatus comprising:

acquiring means for acquiring an intensity of emission spectra emitted from a principle component having an already known concentration and contained in a reference sample having an already known concentration in spark discharge spots within a predetermined area of the reference sample, through emission spectral analysis of the reference sample;

forming means for forming a principle component known concentration-intensity calibration curve representative of relationship between the emission spectrum intensity of the principle component having an already known concentration and the known concentration of the principle component;

acquiring means for acquiring an intensity of emission spectra emitted from a principle component contained in a real steel material-reference sample in spark discharge spots within a predetermined area of the real steel material-reference sample, and an intensity of emission spectra emitted from a constituent element of intermetallic inclusions contained in the real steel material-reference sample, through emission spectral analysis of the real steel material-reference sample;

calculating means for calculating a concentration of the principle component contained in the real steel material-reference sample from the emission spectrum intensity of the principle component of the real steel material-reference sample, based on the principle component known concentration-intensity calibration curve;

calculating means for calculating a concentration of the constituent element of the intermetallic inclusions contained in the real steel material-reference sample, based on the calculated concentration of the principle component of the real steel material-reference sample;

forming means for forming a real steel material-contained intermetallic inclusion constituent element concentration-intensity calibration curve representative of relationship between the concentration of the constituent element of the intermetallic inclusions and the emission spectrum intensity of the constituent element of the intermetallic inclusions;

acquiring means for acquiring an intensity of emission spectra emitted from a base element of a real steel material-pure base sample in spark discharge spots within a predetermined area of the real steel material-pure base sample, and a base element evaporation amount indicative of mass of the base element having been evaporated due to spark discharge thereon, through emission spectral analysis of the real steel material-pure base sample; and

forming means for forming a base element evaporation amount-intensity calibration curve representative of relationship between the base element evaporation amount and the intensity of emission spectra emitted from the base element.

14. A particle size-determining apparatus according to claim 13, comprising:

calculating means for calculating a base element evaporation volume indicative of a volume of the evaporated base element, from the base element evaporation amount based on a density of the base element, and calculating, from the base element evaporation volume, a particle size of the intermetallic inclusions as a diameter thereof corresponding to the base element evaporation volume, based on a formula of calculating a spherical volume;

acquiring means for acquiring a known concentration of the principle component from the emission spectrum intensity of the base element, based on the principle component known concentration-intensity calibration curve;

calculating means for calculating a concentration of the constituent element of the intermetallic inclusions based on the determined known concentration of the principle component;

acquiring means for acquiring an intensity of emission spectra of the constituent element of the intermetallic inclusions from the calculated concentration of the constituent element of the intermetallic inclusion, based on the real steel material-contained intermetallic inclusion constituent element concentration-intensity calibration curve; and

forming means for forming an intermetallic inclusion particle size-intensity calibration curve representative of relationship between the calculated particle size of the intermetallic inclusions and the determined emission spectrum intensity of the constituent element of the intermetallic inclusions.

15. A particle size distribution-generating apparatus for generating a particle size distribution of intermetallic inclusions, comprising:

data sorting means for executing a data sorting process for counting a number of data items of emission spectra of a constituent element of the intermetallic inclusions in a test sample; and

generating means for generating a particle size distribution based on the counted number of the data items and the particle size of the intermetallic inclusions in the test sample, which have been determined by the particle size-determining apparatus for determining a particle size of intermetallic inclusions based on an emission spectrum intensity of a constituent element of the intermetallic inclusions, according to any of claims 11 to 14.

16. A particle size distribution-generating apparatus according to claim 15, wherein said data sorting means rearranges the data items of emission spectra of the constituent element of the intermetallic inclusions in the test sample counts the number of the rearranged data items.

17. A particle size distribution-generating apparatus according to claim 15 or 16, wherein said data sorting means determines whether or not an emission spectrum intensity of the constituent element of the intermetallic inclusions in the test sample is larger than a threshold value, and then extracts data items of intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample to be rearranged in order of intensity, based on a result of the determination.

18. A particle size distribution-generating apparatus according to claim 15 or 16, further comprising acquiring means for acquiring an intensity of emission spectra emitted from a principle component contained in the test sample; and

wherein said data sorting means extracts the data items of intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample to be rearranged in order of intensity, based on a result of comparison between the intensity of emission spectra emitted from the principle component contained in the test sample and the intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample.

19. A particle size distribution-generating apparatus according to any of claims 15 to 18, further comprising:

forming means for forming an intensity correction curve representative of relationship between a number of times of generation of spark discharge for emission spectral analysis and an amount of attenuation of an intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample after carrying out the spark discharge; and

correcting means for correcting the intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample according to the number of times of generation of spark discharge for emission spectral analysis, based on the generated intensity correction curve.

20. A particle size distribution-generating apparatus according to any of claims 15 to 19, comprising identifying means for identifying a kind of the constituent element of the intermetallic inclusions contained in the test sample, based on a result of comparison between the intensity of emission spectra emitted from the principle component contained in the test sample and the intensity of emission spectra of the constituent element of the intermetallic inclusions in the test sample.