JPS6047943A - Controlling method and apparatus of analytical accuracy in spectrometric analysis - Google Patents

Abstract

PURPOSE:To enable continuous and rational control of an irregularity width, to simplify a program, and to enable the reduction of memory capacity necessary for the control of the irregularity width, by utilizing an experimental formula formed between irregularity and concn. CONSTITUTION:It is judged whether the actual irregularity width R<1> of n-times of analytical results is larger or smaller than N-times of a theoretical width R calculated from formula and, when R<1>>NR, it is judged that measurement is abnormal. In the formula, X is the concn. of the element to be analyzed in a specimen and, concretely, is an average value of n-times of analytical results. In this case, BEC is calculated from a calibration curve. At first, the photometric output of each element is inputted from an analytical apparatus SS and stored in a photometric value area. In the next step, the concn. of each element is calculated from calibration curve data. R<1> is calculated from each concn. value to be inputted to an operation area and R of the formula is calculated. The judgement of R<1>>2R is performed. The previously calculated X, that is, the average concn. and the above judge result are stored in a concn. value area. Thereafter, the content of said area is read to perform printing as R.

Description

DETAILED DESCRIPTION OF THE INVENTION L/f) Industrial Application Field The present invention relates to analysis reliability management in emission spectrometry using spark discharge as a light source.

(b) Conventional technology In optical emission spectrometry using spark discharge, variations in the results of multiple analyzes are unavoidable even if the material contains the same element to be analyzed. However, if the trial is determined, the magnitude of this variation is determined, so if the variation in the analysis results for a group is larger than the variation expected in advance, the analysis results for that group indicate that the sample is inappropriate. However, it is thought that there was some kind of abnormality in the analyzer, so the analysis results of that group must be excluded. However, since the dispersion of a group of analysis results varies depending on the concentration of the analyzed element, it is difficult to determine whether the dispersion of a group of analysis results is normal or not. The width had to be determined and compared with the actual variation width, which required a large amount of data and complicated judgment operations.Moreover, the variation management range was gradual and lacked rationality.

(c) Purpose The present invention does not involve subjectivity in determining the reliability of analysis results, and provides a method for managing the reliability of analysis results that is easy to automate, so that an appropriate determination is always made regardless of the concentration. This is what we are trying to provide.

(d) Constituent The variation in measured values varies depending on the number of measurements, and in the case of spark discharge emission spectrometry, it also varies depending on the concentration of the analytical element in the sample. By determining the coefficient of !111 based on a preliminary investigation, a simple relational expression can be established using the above coefficient between the width of variation, the number of measurements, and the concentration of the measured element in the sample. Heading: When the variation width of multiple analysis results is more than a certain times larger than the variation width calculated by this relational expression, that is, the theoretical variation width, it is determined that the measurement itself of that group of measurement results is abnormal. This is to prevent data from being used as analysis data.

The number of measurements is n, and the variation width of the fixed value after n measurements, that is, the difference between the maximum and minimum measured values, is R1.The standard deviation of the variation σ is σ = R/ k n...・・・・・・・・・(
1) It is approximated by the simple method. Here kn is the number of measurements n
The values vary depending on the situation, as shown in the table below.

Still defining the coefficient BEC and surface (described later), the standard deviation σ
The inventor of the present invention has discovered that σ=(X+BEC)XCV...Mountain...(2) can be confused with the empirical formula.

When calculating R by substituting σ in equation (1) for σ in equation (2), R= kn (X + BEC)
) (3) is an experimental formula that calculates the theoretical variation range from the results of n measurements. The actual variation range R1 of the 1"in analysis results of the present invention is N of the theoretical range R determined by the above equation (3).
It is determined whether p is larger or smaller than the times p, and when R') is NR, it is determined that the measurement is abnormal.

Now, X, DEC, and CV that appear in the above equation will be explained. X is the concentration of the analysis element in the sample, specifically the average value of n analysis results. BEC is determined from a calibration curve as shown in FIG. A calibration curve is a relationship curve between concentration and measured output determined by actual measurements using standard samples with known concentrations.When this curve is extended to concentration O, the point that cuts the vertical axis is the background of the measured output. show. So, on the vertical axis, let's take a point b' where ob = bb'', draw a horizontal line from b', and drop a perpendicular line from the intersection with the calibration curve.The concentration at the intersection with the horizontal axis is BEC. Concentration value (EEC) when pure measurement value and background are equal (S/IJ = 1)
is rumored to be the bank ground equivalent concentration. n is concentration Q%
When the standard deviation of the dispersion of measured values in n analyzes of a sample is σ, CV-σ/Q is measured in advance using a standard sample. ] ゛ is a
This is the average value of CV in the temperature measurement range. As stated at the beginning, the range of variation in analytical values varies depending on the concentration, and the fluctuations in the variation 'E depending on the concentration are averaged and made into a constant.

(e) Examples In the above explanation, the number of measurements n was not particularly limited.

n may be any value as long as it is greater than 2, and in practice, n = 2 is often sufficient. - As an example, we will discuss the case of n-2 in spark discharge emission spectroscopy of carbon in steel.

In this case, the wavelength used for analysis is Cl 930 XT.

The above formula (2) becomes a = (C% + 0.1) X O,008. Also, when n = = 2 in the diagram, k n = ],, 1-, so equation (3) is R = 1.1X (C%-1-0,1,) X O,008
,,, (4) Here, the carbon concentration C% is the average value of concentration values calculated from two measurements using a calibration curve. Control width is 2σ
Then, when the difference R1 between the two analysis values is greater than 2R, it is determined that the analysis is abnormal. If 0% - 0.1, then R in equation (4) will be 0.00194, and the difference between the two measured values will be 0.
If it is 00388 or more, it is an analysis abnormality. In FIG. 2, the horizontal axis is 0%, and the vertical axis is ij:R', that is, the difference between two measured values.If the value is below the diagonal line, the analysis is good, and if it is above the diagonal line, the analysis is abnormal.

FIG. 3 shows an example of a device that performs the above-mentioned management operations.

SS is an emission spectrometer, CPU is a computer, and M is a memory. The memory M is a program area that stores the above-mentioned management operation program, a data area that stores the above-mentioned DEC, CV, etc. data for each analysis element, calibration curve data for each element, etc., and a data area that stores the data directly from the analyzer SS. A photometric value area that stores the photometric values of each analytical element obtained;
It has a calculation area for storing R in the equation (3) and other data generated in the course of calculation 1 determination, and a density value area for storing the determined density value. P prints out the analysis results using a printer. R is an example of printing, in which the sample number, the concentration value of each element, and a determination as to whether the analysis is normal or abnormal (OIJT indicates abnormality) are printed. If the above-mentioned judgment is abnormal for any element for one sample, the judgment becomes OUT.

FIG. 4 is a flowchart 1 of the management operation. Input the photometric output of each element from the photometric analyzer SS and store it in the photometric value area a). Next, calculate the concentration of each element using the mass curve data (see below). The calibration curve has three coefficients entered in the data area as a two-method equation, and the concentration is calculated using the coefficients. Since the analysis is performed twice, the average of the two concentration values is calculated (c), and the value is expressed as X in equation (3) above.
Therefore, it is stored in the calculation area along with each density value (d). R1 is calculated from each density value and entered into the calculation area (e). (3) Print out R in the equation.

R': (Perform 2R judgment (G). First, store Meda X, that is, the average density and the above judgment result in the density value area (H)
. Thereafter, the contents of the same area are read out and printed as shown in FIG. 3R (step 1) These operations are performed for each element.

(f) Effect In spark discharge emission spectroscopy, the range of variation in measured values changes depending on the concentration, but the present invention found that a simple experimental formula holds between the variation and concentration, and decided to utilize that formula. It is now possible to continuously and rationally manage the variation range over the entire measured concentration range, and judgments can be made using a single formula over the entire measured concentration range, making programming easy and requiring less data. The memory capacity required for variation width management is small.

[Brief explanation of the drawing]

Figure 1 is a graph explaining BEC, Figure 2 is a graph of the relationship between carbon concentration and allowable variation width in an example where the present invention is applied to the analysis of carbon in steel, and Figure 3 is an example of the present invention. The block diagram of the device, FIG. 4, is a flowchart of the operation of the device. Agent Patent Attorney Kosuke Hayashi

Claims (2)

[Claims] (1) Calculate the theoretical variation width R using the following formula using the maximum and minimum difference R1 of concentration values in n measurements by emission spectrometry and the average X, and when the control width is N, if R')NR then A method for managing the quality of analysis in luminescence analysis, characterized in that the n measurements are determined to be abnormal. R = (X 0 BKC) Average value of the variation width divided by the relevant concentration. (2) A data area that stores data such as DEC and σ■ for each element in advance, a calculation area that stores measurement data used in judgment calculations such as R2X of each element, and the determined concentration value of each element. A control computer including a memory having a density value area for storing R-(X+BEC) Analytical accuracy control device for luminescence spectrometry combined with an analyzer.