JPH10267915A - Sampler for steel slag and its use - Google Patents

Abstract

PROBLEM TO BE SOLVED: To quickly obtain a highly accurate analysis result of steel slag, adjust the sample in a short time and obtain a sample with high representativeness. SOLUTION: A steel columnar member 1 having a length of 150 mm or more.
And a holding rod 3, where V / A is 10 (mm) or more, and V / A is 10 mm or more, and the diameter of the pillar 1 is 10 mm or more. The flat surface 2 has a surface roughness of 30 μm or less. This sampler is immersed in the molten slag layer for a short time of 5 seconds or less, and the slag adhered to the flat surface 2 is subjected to analysis. [Effect] Since the sample is adjusted in a very short time and the analysis surface of the slag sample is quenched, the precipitated particles are fine, the uniformity is high, and the analysis value with good reproducibility can be obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sampling method for quickly and accurately analyzing the components of slag floating on steel melted in a steel refining process.

[0002]

2. Description of the Related Art With the improvement of quality of steel products, it is necessary to know the components of slag in the refining process to grasp the refining status and to carry out strict operation management. Therefore, an analysis technique for quickly and accurately measuring a slag component on the machine side has been demanded.

[0003] The components of the slag are X-ray fluorescence analysis and laser IC.
It is measured by P analysis and the like. In fluorescent X-ray analysis,
The surface of the analysis sample is irradiated with X-rays, and characteristic X-rays generated by the elements contained in the sample are analyzed to determine the component amount. Laser IC
In the P analysis, a pulsed laser beam is condensed and irradiated on the sample surface, the surface layer is vaporized, and the sample is cooled and turned into fine particles.
The sample is transported to a CP analyzer for emission spectroscopy. In these analyses, it is known that the quality of the analytical sample affects the analytical accuracy. In particular, it is important to measure and disperse the measurement components uniformly in the sample. It is required that the surface be smooth and flat. The area of the analysis surface is preferably large, but an analysis surface having a diameter of about 10 mm is used.

Conventionally, a glass bead sample has been used as a highly reliable sample, and its preparation is specified in JIS M 8205 (hereinafter, referred to as a glass bead method). The sample is uniformly melted in a melting agent to prepare a disk-shaped glass bead, using platinum or platinum rhodium for the melting vessel,
Sodium tetraborate or lithium tetraborate is used as the melting agent. The slag component is uniformly melted by heating to a constant temperature in the range of 1000 ° C. to 1150 ° C. for a predetermined time with stirring for a predetermined time to remove bubbles sufficiently to form a smooth and flat free surface, and then allowed to cool as it is To prepare a glass bead of uniform texture.

[0005] In the case of a slag sample, the molten slag is pumped up, cooled, crushed, stirred sufficiently, a part of the slag is collected, and a glass bead is prepared by the above method. And
The surface of the prepared bead is irradiated with X-rays to perform X-ray fluorescence analysis. Although the analysis results can be obtained with sufficient accuracy by the glass bead method, it takes a long time to prepare the sample, and it takes about one hour to obtain the analysis results. For this reason, a result was not obtained within the refining time of ten to several minutes to several tens of minutes, and could not be used for operation management analysis.

[0006] A briquette method is intended to shorten the sample preparation time. In the briquette method, the slag that has been pumped up from the molten slag layer, collected and cooled is crushed into fine powder, mixed thoroughly, and then filled into a cylindrical pressure vessel.
It is pressurized and solidified into a column to prepare a briquette sample. However, this briquetting method also requires a time of about 20 minutes, and furthermore, the analysis accuracy is inferior, so that the operation management analysis is insufficient in both speed and accuracy.

[0007] Among these problems, there is a steel plate immersion method which has improved the speed. In this method, a steel plate is used as a sampler, and a strip-shaped steel plate is immersed in a slag layer, the slag attached when the steel plate is pulled up is peeled off, and the slag is directly subjected to X-ray fluorescence analysis. For example, Japanese Patent Application Laid-Open No. 54-153694
Japanese Patent Application Laid-Open Publication No. H11-15064 describes a method in which a blast furnace molten slag is adhered to a stainless steel plate heated to 600 ° C. to 700 ° C. having a surface roughness of 25 μm, a surface area of 120 cm ² and a thickness of 10 mm, and peeled off.

[0008]

In the steel plate immersion method, the sample preparation time is reduced to almost half that of the briquette method, but it takes time to adjust the temperature of the sampler, and the analysis accuracy is the same as that of the briquette method. It was necessary to repeat the analysis several times in order to obtain it, and it had not yet come to practical use as an analytical method used for operation management.

[0009] In more detail regarding the problem of analytical accuracy, in the briquette method, when two briquette samples prepared from the same charge mixed well are analyzed, there is a large difference between the two analytical values. Also, when two analysis samples were taken from the slag attached to the same sampler and analyzed by the steel plate immersion method, the difference between the two analysis values was large as in the case of the briquette sample.

As described above, in the briquette method and the steel plate immersion method, the reproducibility of the analysis is inferior. Therefore, in order to obtain the true value, the analysis is performed on a large number of samples, and the analysis results are obtained by averaging appropriate analysis values. In practice, the measurement time becomes several times even if the sample preparation time is shortened, and the effect of shortening the sample preparation time is reduced.

The present invention has been made in order to solve the above-mentioned problems, and a method for analyzing a steel slag which can shorten the time for preparing a sample, secure the reproducibility of the analysis, and can be used for the operation management of steel refining. It is intended to provide.

[0012]

Means for achieving this object are the following inventions.

A first aspect of the present invention is a sampler for steel slag having the following structure. (A) It is composed of a columnar body having a length of 150 mm or more and a holding rod for holding the columnar body. (B) The columnar body is made of steel and has a volume of V, a side surface area of A, and V / A. 10 (m
m) and at least one smooth flat surface with a diameter of 10 mm or more on the side surface, and (c) the surface roughness of the flat surface is 30 μm or less.

The sampler of the present invention aims at quenching and solidifying the slag when the columnar body is immersed in the slag to be melted. Figure 1 shows the structure of this sampler.
Shown in (A) In the drawing, 1 is a columnar body, 2 is a flat surface, and 3 is a holding rod. The length of the columnar body 1 is 150 mm or more, and its material is iron or steel. The columnar body only needs to be substantially columnar, and may be slightly shaved at the corners or may be processed to attach the holding portion to the upper portion. Iron and steel are inexpensive materials, are easy to process, have high specific gravity, specific heat and thermal conductivity, and are excellent as quenching materials.
These materials are used.

Some of the surface layer of the molten slag layer is solidified in several centimeters, and the sample is obtained by quenching the molten slag directly on the sampler surface while avoiding this solid. For this reason, the columnar body 1 is made sufficiently long with respect to the thickness of the surface layer, and its length is set to 15
0 mm or more. In order to obtain a slag sample having the required area, smoothness and flatness as a fluorescent X-ray analysis sample, the columnar body 1 has at least one smooth flat surface 2 having a diameter of 10 mm or more. FIG. 2B is a cross-sectional view of the columnar body 1, and the length e of the straight portion needs to be 10 mm or more.

The reason for limiting the relationship between the surface area and the volume of the side surface of the columnar body 1 is to enhance the quenching effect. The columnar body 1 at normal temperature is immersed in the high-temperature slag layer, and thereafter the temperature is increased by endothermic heat from the surface. The effect of suppressing the rate of temperature rise is greater when the volume of the columnar body 1 is larger and the side surface area is smaller. That is, in FIG. 2B, the cross-sectional area S of the columnar body 1 is large,
The smaller the peripheral length E (including the length e of the linear portion) is more effective. When V / A is 10 (mm) or more as the volume V and the surface area A of the side surface, the effect of quenching becomes sufficiently large. Then, when the slag is rapidly cooled on the surface of the columnar body 1, a large number of solidification starting points are formed in the slag, the precipitate becomes fine, and a slag sample in which the dispersion of the components can be regarded as uniform is obtained.

The flatness and smoothness of the surface of the analysis sample improves the reproducibility of the analysis. Fluorescent X
In the line analysis, if the surface is rough, the amount of secondary X-rays absorbed varies. For this reason, the surface roughness is reduced to JIS B 0601.
The maximum height Rmax specified in the above is 30 μm or less.

The second invention is a method of using a sampler for steel slag, comprising the following steps. (A) a step of preparing the sampler according to claim 1; and (b) a step of immersing the columnar body of the sampler in a slag layer with its upper end surplus, and pulling up within 5 seconds of the immersion time. (C) a step of separating the slag layer by impact after cooling the slag layer adhered to the columnar body; and (d) a slag sample of the separated slag layer which is in contact with the flat surface of the columnar body. And subjecting the surface in contact with the flat surface to analysis.

A slag sample is collected by attaching molten slag to the column by using the sampler described above. When the sampler is immersed in the slag layer, the entire column is not immersed but the upper part is immersed. It is made to protrude above the slag layer without being immersed. This is because when the sampler is pulled up, the surface solid slag is scooped up on the upper surface and is prevented from flowing and mixing with the slag sample. There is already solidified slag on the surface layer, and this solid slag is not quenched slag, and if these are mixed with the quenched and solidified slag, the uniformity of the sample is impaired. Also, by exposing a part of the columnar body from the slag layer, an effect of cooling the sampler can be obtained.

After the sampler is immersed in the slag layer, it is preferable to lift the sampler as quickly as possible. If the sampler is immersed for more than 5 seconds, the surface temperature of the sampler rises and the rapid cooling effect is reduced. After the lifting, the surface layer of the slag attached to the columnar solidifies in a few seconds. Thereafter, when an impact such as dropping the sampler to the floor is applied, the slag pieces are separated. And what was in contact with the flat surface of the separated slag piece was used as the slag sample,
The contacted surface is used for analysis.

The surface in contact was a surface which was quenched and solidified at a temperature difference of about 1600 ° C. at the moment when the sampler was immersed in the slag layer, and the flatness and roughness were controlled by the flat surface. is there. On the opposite side, the cooling rate is lower and the smoothness and flatness are not controlled. By such a slag sample collection method, the preparation time of the analysis sample can be shortened, and the analysis reproducibility close to the glass bead method can be ensured.

[0022]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a sampler according to the present invention will be described with reference to the drawings. The shape of the columnar body is not limited to the shape in which a part of the column shown in FIG. 1 is cut flat, and may be a polygonal column. In this case, V / A is larger when the cross-sectional shape is a regular polygon than when it is a non-regular polygon. Also, since the cross-sectional area is proportional to the square of one side and the peripheral length is proportional to the length of one side, V / A increases as one side of the regular polygon becomes longer.

FIG. 2 (a) is a graph showing the S / E, ie, V / A, where S is the cross-sectional area of the columnar body 1, e is the length of one side, and E is the length of the periphery when the cross-sectional shape is an equilateral triangle. Is 10 (mm) when e is about 70 mm. (B) shows a case where the cross-sectional shape is a square, and S / E becomes 10 (mm) when e is 40 mm. (C) shows a case where the cross-sectional shape is a regular hexagon, and S / E becomes 10 (mm) when e is about 2 mm.
3 mm.

In these regular polygons, S / E is 10 (mm).
The S when made, the equilateral triangle 2078Mm ^2, the square 1600 mm ^2, the columnar member 1 as 1385Mm ² and the angular speed increases in the regular hexagon requires only a thin. In other words, if the columnar body 1 has the same thickness, as the number of corners increases, the surface area of the side surface decreases, the amount of heat input to the immersion part decreases, and the temperature rise of the columnar body is suppressed.

However, even with regular polygons having the same number of corners, S / E can be increased by rounding the corners.
FIG. 3A is a cross-sectional view of the case where the corners of the square are rounded. If the angle θ at which the circumference of the corner is desired is 30 °, e ₁ is about 3
At 8 mm, S / E is 10 (mm), and S at this time is 1336 mm ² . This is 138 for a regular hexagon.
5mm ² smaller.

FIG. 3B is a sectional view of a partial circle in which a part of the circumference is a straight line. S / E varies depending on the angle θ at which a straight line is desired, that is, the length of the straight portion and the radius r of the circumferential portion. This relationship is shown in FIG. S / E increases with an increase in the partial radius r, and increases as θ decreases. When θ is 30 °,
When r is 20.5 mm, S / E is 10 (mm). At this time, the sectional area S is 1315 mm ² , and the sectional area is further reduced. At this time, the length of the linear portion, that is, the width of the flat surface is 10.6 mm, and a flat surface having a diameter of 10 mm or more can be obtained.

In analysis such as fluorescent X-ray analysis and laser ICP analysis, characteristic X-ray (secondary X-ray) intensity is measured in fluorescent X-ray analysis, and emission intensity is measured in laser ICP analysis. Is converted to the concentration of That is, the relationship between the intensity and the concentration is obtained for the standard sample, and the measured intensity is converted into the component concentration using this as a calibration curve. The standard sample is
It is prepared by a glass bead method in which a standard substance is melted using a melting agent. For this reason, a sample having the same composition in which the standard components are uniformly dissolved and uniformly distributed and having a smooth surface can be obtained regardless of where the glass bead is taken.

The measurement sample prepared by the glass bead method is as follows:
Like the standard sample, the components are uniform and the surface is smooth and flat. On the other hand, the sample obtained by the briquette method or the steel plate immersion method is a sample in which the molten slag is solidified as it is, and the components are not uniformly dispersed as in the standard sample. During solidification, it precipitates first from the one with the lower melting point,
Precipitates having a different composition from those solidified later.

[0029] For example, the slag, CaO-SiO ₂ system, CaO-Al ₂ O ₃ composite compounds and 1200 ° C. SiO ₂ or the like having a glass transition point of the front and rear like exist with around 1500 ° C. melting point . The slag sample is a mixture of precipitates having different compositions, and the component distribution is non-uniform. In addition, the smoothness of the surface is also quite rough compared to glass beads.

First, if the surface roughness of the measurement sample is significantly different from the surface roughness of the standard sample, an error occurs when the calibration curve is applied. The relationship between the surface roughness and the characteristic X-ray intensity will be described with reference to FIG. Primary X-rays emitted from the X-ray source 11 to the sample 10 generate characteristic X-rays (secondary X-rays) from elements in the sample, and the characteristic X-rays are detected by the detector 12. Since the X-rays are absorbed according to the distance that they pass through the sample, most of the detected characteristic X-rays are generated at the surface layer (arrow a).
Thus, only a small amount is generated in the deep layer (arrow b).

When the surface is smooth like a standard sample, the intensity of the secondary X-ray draws a logarithmic function curve with respect to the distance from the surface, and its distribution shows a constant pattern. However, in the case of a measurement sample having a rough surface, even a characteristic X-ray generated near the average level of the surface may have a longer distance to pass through the sample before reaching the surface, resulting in a decrease in intensity (arrow c). On the contrary, in some cases, the passing distance is shorter and the strength is larger than that of the standard sample.

For this reason, if the surface is rough, the distribution of the secondary X-ray intensity from the surface to the deep portion does not show a fixed pattern, and differs from the pattern of the standard sample. This causes a measurement error. In the current fluorescent X-ray measurement technology, a slag sample detects secondary X-rays generated from a depth of about 100 μm, so that the influence of roughness can be ignored up to Rmax of 30 μm.

Next, even when precipitates having different compositions are mixed and the component distribution is non-uniform, an error occurs when the calibration curve is applied. The reason will be described below using an example of X-ray fluorescence analysis. First, it is known that the amount of X-ray absorption follows equation (1).

[0034]

(Equation 1)

Here, I is the X-ray intensity after transmitting the substance having the absorption coefficient μ by the distance d, and I ₀ is the intensity before transmission.

Referring to FIG. 7, two regions a and b are considered without complexity. (A) As shown in the figure, assuming that the absorption coefficient of the region a and the region b is μ _a and μ _b , and the X-ray intensity after transmission is Ia and Ib, respectively, the measured X-ray intensity is I ₁ Equation (2) is the sum of Ia and Ib.

[0037]

(Equation 2)

In the standard sample, the component distribution is uniform.
Since the region a and the region b have the same composition, and the absorption coefficient has the average value, the measured X-ray intensity I ₂ is given by the equation (3).

[0039]

(Equation 3)

[0040] Compared with the I ₁ and I _2, it becomes equal to the I ₁ and I ₂ only when the mu _a and mu _b are equal, I ₁ and I ₂ are otherwise different. The region a is occupied by the precipitate a, and the region b is occupied by the precipitate b. If the precipitate a and the precipitate b are different, the component distribution of the measurement sample is not uniform, and the measurement is performed even if the average component concentration is the same. X-ray intensity varies.

When the area shown in FIG. 7A, that is, the precipitate becomes smaller, the result becomes as shown in FIG. 7B, and the X-rays passing only through the precipitate a or only the precipitate b decrease, and many X-rays Both precipitates become permeable. When passing through both precipitates,
When the transmission distance in the precipitates a is D, the absorption coefficient mu ₁ at that time is as shown in equation (4).

[0042]

(Equation 4)

The smaller the precipitates, the more the precipitates permeate, and the number of permeating precipitates a and b approaches the same number. If evenly pass through the precipitate a and precipitates b, D is d / 2 becomes, mu ₁ is consistent with the absorption coefficient in the case (3) component is uniform in (4) below. That is, in the fluorescent X-ray analysis, the components can be regarded as uniform. The situation is the same when three or more types of precipitates are used.

In the laser ICP analysis, if the components are not uniformly dispersed, the component concentration differs for each fine particle. For this reason, the fluctuation of the analysis value is the same as in the case of the fluorescent X-ray analysis.

In the conventional briquette sample, since the slag taken out is naturally cooled, the above-mentioned precipitates become large.
Further, even in the steel plate immersion method, slag adheres to the sampler whose temperature has been increased, so that it takes time until the slag is solidified, and the precipitates become large. In the sampler made of a columnar body having a larger volume than the surface area according to the aspect of the present invention, the precipitate can be reduced by the quenching effect.

FIG. 5 shows the result of examining the difference in component concentration in each minute area when the cooling rate was changed by changing the V / A of the columnar body. EPMA is used to analyze 10 micro areas, and the difference between the analysis values is expressed as a percentage of the standard deviation with respect to the average value, that is, the coefficient of variation (Cv). The vertical axis shows the variation coefficient, and the horizontal axis shows V / A of the columnar body. The spot diameter of the micro area, ie, the electron beam, is 50 μm, and the interval between the micro areas is about 0.5 mm. □ indicates a sampler whose columnar body has a square cross section, Δ indicates a regular hexagon, and ○ indicates a sampler obtained by combining a partial circle and a straight line.

(A) shows the measurement results for CaO,
(B) shows the measurement results for SiO ₂ , and (c) shows the measurement results for Al ₂ O _3. Regarding any component, the coefficient of variation is large when V / A is small, and the coefficient of variation is large when V / A is large. It is getting smaller. This tendency is small for Al ₂ O ₃ , but is remarkable for CaO. In particular, CaO and SiO ₂ in V / A is 8 (m
m), the degree of decrease in the coefficient of variation is large, but V
When / A is about 10 (mm), the degree of reduction is small.

That is, when a quenched sample is obtained using a sampler having a columnar body having a V / A of 10 (mm) or more, the precipitates become small and fine, and the number of precipitates contained in a fine region having a diameter of about 50 μm increases. Uniformity as an analytical sample is obtained.

Since increasing the V / A suppresses the temperature rise on the surface of the columnar body, erosion due to slag is small even when the material is iron or steel, and even in the case of carbon steel having a flat surface finished to about Rmax 10 μm, If the flat surface is re-polished every 20 times with a sampler having a V / A of 10 (mm) and every 50 times with a sampler having a V / A of 15 (mm), R
It is possible to keep a smooth surface of max 30 μm or less.

[0050]

EXAMPLE A slag sample was collected from a slag layer in an arc refining furnace using a sampler having a columnar body made of carbon steel having a length of 200 mm and a square cross section, and having a diameter of 10 from eight places.
An analysis sample of mm was taken and subjected to X-ray fluorescence analysis to examine the analysis value and the time required for analysis. Lower part 150 of four sides of pillar
The slag sample was sampled by immersing it in a state in which the upper part was protruded from the molten slag layer by about 50 mm. The immersion time was about 3 seconds.

There are three types of columnar bodies, and one side of the square of the cross-sectional shape is 60 mm, 50 mm, and 20 mm, respectively, and the V / A is 15 (mm), 10 (mm), 5 (mm), and 20 mm, respectively. The case is a comparative example. Analytical samples were collected two by two from one flat surface. As a conventional example, about 200 g of slag was pumped up, spontaneously cooled, crushed, and agitated and stirred sufficiently to prepare eight glass bead samples and eight briquette samples. At the same time, width 20mm, thickness 10mm, length 200
Eight samples (four samples each from the front and back) were examined by a conventional steel plate immersion method using a rectangular columnar body of mm, and these were compared.

For the analysis values, the coefficient of variation (Cv) is determined.
The time required for the analysis should be
From the start of the sampling operation until the first analytical value is obtained
The interval was measured. The results are shown in Table 1. In addition, slag
Samples were CaO, SiO_TwoAnd Al_TwoO _ThreeAccounts for about 90 wt% of the total
Therefore, these components are contained at a ratio of about 6: 2: 1.
It was.

[0053]

[Table 1]

In the embodiment of the present invention, V / A is 15 (m
Both the column (m) (test No. 1) and the column having a V / A of 10 (mm) (test No. 2) had a small coefficient of variation and a short analysis time. That is, even if one analysis value was trusted, the error was not much different from that of the glass bead method, and it became clear that the analysis result was obtained 7 minutes after the start of slag sampling.

In contrast, the conventional briquetting method (test
In No. 4) and the steel plate immersion method (Test No. 5), the coefficient of variation was large. Therefore, in order to obtain an analysis result with the same accuracy as in the embodiment of the present invention, it is apparent that it is necessary to perform analysis twice or more and compare those analysis values to obtain an average value of the probable analysis values. It became. in this case,
The time required for analysis must allow for a few more minutes for the time required for one analysis, and eventually the required time for analysis is close to 20 minutes or 30 minutes.

In the glass bead method (Test No. 6), the coefficient of variation was sufficiently small, but the time required for one analysis was 50%.
It could not be used for operation management for a long time.

[0057]

As described above, according to the present invention, a steel column having a smooth flat surface large enough to take a sample for analysis is immersed in the molten slag layer to remove the adhered slag. The attached surface is used as the analysis surface. However, since the sampler is a sampler in which the ratio of the volume of the columnar body to the surface area of the side surface is increased to 10 (mm) or more, the attached surface of the slag attached to the flat surface is quenched and precipitates. Are fine. For this reason, the components are almost uniformly distributed, and the analysis can be performed with a reproducibility comparable to that of a glass bead sample. Then, when this sampler is immersed in slag for several seconds and subjected to impact on the column after cooling, an analysis sample can be obtained in a short time, and an analysis result can be obtained in 7 minutes from the start of immersion. The effect of the present invention, which enables the operation management analysis of steel smelting in this way, is great.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining an embodiment of the invention;
(A) is a perspective view of the sampler, and (b) is a cross-sectional view of the columnar body.

FIGS. 2A and 2B are perspective views of a columnar body for describing an embodiment of the invention, wherein FIG. 2A is a triangular prism, FIG.
(C) The figure is a hexagonal prism.

3A and 3B are cross-sectional views showing the shape of a columnar body. FIG. 3A is a square with rounded corners, and FIG. 3B is a shape obtained by combining a partial circle and a straight line.

FIG. 4 is a graph showing a radius of a partial circle and S / E (ratio of area to circumference).

FIG. 5 is a graph showing a relationship between a coefficient of variation and V / A;
(A) is for CaO, (b) is for SiO ₂ , and (c) is for Al ₂ O ₃ .

FIG. 6 is a cross-sectional view showing the relationship between the roughness of the sample surface and the characteristic X-ray intensity.

FIGS. 7A and 7B are schematic diagrams for explaining the relationship between the size of the precipitate and the characteristic X-ray intensity. FIG. 7A shows a case where the precipitate is large, and FIG. 7B shows a case where the precipitate is small.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Columnar body 2 Flat surface 3 Holding part.

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Sanji Okano 1-2-1, Marunouchi, Chiyoda-ku, Tokyo Inside Nihon Kokan Co., Ltd. (72) Inventor Yutaka Yoshioka 1-1-2, Marunouchi, Chiyoda-ku, Tokyo Nihon Kokan Co., Ltd.

Claims (2)

[Claims] 1. A sampler for steel slag having the following structure. (A) It is composed of a columnar body having a length of 150 mm or more and a holding rod for holding the columnar body. (B) The columnar body is made of steel and has a volume of V, a side surface area of A, and V / A. 10 (m
m) and at least one smooth flat surface with a diameter of 10 mm or more on the side surface, and (c) the surface roughness of the flat surface is 30 μm or less. 2. A method for using a sampler for steel slag, comprising the following steps. (A) a step of preparing the sampler according to claim 1; and (b) a step of immersing the columnar body of the sampler in a slag layer with its upper end surplus, and pulling up within 5 seconds of the immersion time. (C) a step of separating the slag layer by impact after cooling the slag layer attached to the columnar body; and (d) a slag sample of the separated slag layer that is in contact with the flat surface of the columnar body. And subjecting the surface in contact with the flat surface to analysis.