EP2539693A1

EP2539693A1 - Method for determining carbon in cast iron

Info

Publication number: EP2539693A1
Application number: EP11708185A
Authority: EP
Inventors: Roland Van Driel; Bruno Van Stuijvenberg
Original assignee: Spectro Analytical Instruments GmbH and Co KG
Current assignee: Spectro Analytical Instruments GmbH and Co KG
Priority date: 2010-02-22
Filing date: 2011-02-17
Publication date: 2013-01-02
Also published as: US8976350B2; US20120300204A1; WO2011101143A1; DE102010008839A1; DE102010008839B4

Abstract

The invention relates to a method for determining the carbon content of an iron alloy, comprising the following steps: a) starting the measurement (13) of a sample in a spark spectrometer, b) forming a plasma in a pre-spark phase, c) detecting (13) and recording an intensity signal of the carbon, d) calculating and masking (15) an unstable plasma phase, e) calculating an excess increase in the carbon signal (16, 17) and calculating the content of dissolved and undissolved carbon (18).

Description

Method for the determination of carbon in cast iron

The present invention relates to a method for the determination of carbon in cast iron, in particular spheroidal graphite cast iron with the aid of sparking spectrometry on a solid sample.

Spark spectrometry is a method for the chemical analysis of metals. An electrical discharge is generated which vaporizes a portion of the sample and generates a plasma. In this plasma, the sample atoms are excited and produce emission lines that are characteristic of the elements contained in the sample. Such analyzes are routinely used in the steel industry to control alloys. This measuring method achieves a very high measuring accuracy. However, it has long been known that spark spectrometry in alloys can accurately measure the carbon content only when carbon is present in completely dissolved form in the alloy. Elementary precipitated carbon, such as is present in ductile iron (GGG), regularly leads to a falsification of the measurement. The carbon content of such alloys is measured too low.

In spark spectrometry, a metallic sample is first exposed to sparks under a protective atmosphere. In this

CONFIRMATION COPY In the pre-radio phase, the sample surface is homogenized with sparks of high energy at a frequency of 200 to 800 Hz. Each spark melts the area of the sample surface around its point of incidence in the radius of a few tens of microns. After the pre-radio phase, this homogenized sample surface is then exposed to the actual measuring sparks which generate the signal to be analyzed. When precipitated elemental carbon or other precipitated elements or compounds such as Al ₂ O ₃ are present, the precursors preferentially attack the grain boundaries of these precipitates. In the case of carbon, this results in the elemental carbon being sublimated and removed from the sample. The homogenized by Vorfunken parts of the sample surface thus contain less carbon than the original alloy.

This problem is avoided in practice by cooling the withdrawn liquid sample as quickly as possible. Thus, the existing carbon is not eliminated elementary. This method is not well reproducible in practice. It leads to samples with different cooling rates and thus a different content of elemental carbon. Therefore, with higher requirements for the accuracy of analysis, it is unavoidable to use other analysis methods for carbon determination. A common alternative method of analysis is to chop a sample and burn it in a controlled manner. The resulting carbon dioxide is measured and used to determine the total carbon content of the sample. This process is very complex, because it requires time and additional equipment.

It is therefore an object of the present invention to provide a method by means of which the carbon content of alloys can be measured precisely by means of spark spectrometry even when carbon is present in elemental form.

This object is achieved by a method having the features of claim 1. Because the intensity signal of the carbon is already recorded in the pre-radio phase, the carbon content removed by sublimation on the alloy can be taken into account. If, after the pre-radio phase, the concentration of carbon is measured in a conventional manner and the previously determined amount of sublimed carbon is taken into account, the measured value is corrected by the sublimed carbon content and the correct result for the carbon content of the sample determined.

It is advantageous if in the first measurement phase, the carbon signal is measured at a wavelength of 148.176 nm. The pre-radio phase is preferably carried out over a period of 8 to 15 seconds and in particular of 12 seconds. A particularly good control of this method is possible if during the pre-radio phase, the iron signal is detected. With the aid of the iron signal, it can be determined when a stable signal, which is also meaningful for the sublimating carbon, will be present in the pre-radio phase. In this case, the iron line is preferably measured at 149 nm.

Hereinafter, the present invention will be described in detail with reference to the drawings. FIG. 1 shows the intensity distribution in the pre-radio phase for samples with completely dissolved carbon content;

FIG. 2 shows the intensity distribution of the precursor phase for the iron content of the sample from FIG. 1;

FIG. 3 shows the intensity distribution of a sample with elemental carbon during the pre-radio phase;

Figure 4: the iron signal of the sample of Figure 3;

FIG. 5 shows an intensity distribution of the signal according to FIG. 3 as a bar graph;

FIG. 6 shows a flow chart for the various method steps the measuring method according to the invention; such as

FIG. 7: a microscopic representation of samples with globally precipitated elementary carbon.

FIG. 1 shows the intensity profile over time, which is measured in the pre-radio phase on the carbon line 148.176 nm. Specifically, the number of the measurement interval is specified on the X axis, beginning at 0. Each measurement interval corresponds to approximately 0.025 sec. The sampling frequency is corresponding to 40 Hz. The diagram depicts a pre-radio phase of approximately 12 sec. In the Y-axis, the intensity of the radiation is approximately proportional to the number of measured photons in arbitrary units. In the example Fig. 1, the intensity varies from about 50,000 per measurement interval to about 200,000 per measurement interval. The first measurement intervals deliver a signal of only about 50,000 units. During this so-called Einfunkphase the plasma is formed. Their length varies greatly and is e.g. dependent on contamination of the sample surface. Then the signal rises to about 200,000 units. The entire representation of FIG. 1 relates only to the usual pre-radio phase, in which the sample is first homogenized in a spark spectrometer. This pre-radio phase is not evaluated in the known measuring methods.

In the present method, the data of the pre-radio phase according to FIG. 1 are processed. For this purpose, the signals lying below a threshold value 1 are discarded. In concrete terms, this means for FIG. 1 that the first approximately 60 measurement intervals are discarded until, in a rise approximately at the point 2, the limit value 1 is exceeded. The mean value of intensity from point 2 is used to calculate the carbon content from the pre-radio phase.

FIG. 2 shows the corresponding iron signal at the line 149.653 nm. It can also be seen in FIG. 2 that in the first 60 measuring intervals the intensity is around 5,000 units per measuring interval. After that, the intensity increases very steeply and exceeds at one point 3 is a lower limit 4. From point 3, the average value of the pulses per measurement interval is about 75,000 units. Also for the iron signal, the plasma is unstable up to the point 3 of the pre-radio phase. These measurements are discarded. From point 3, the signal of the pre-spark phase is evaluated. Here, an upper limit value 5 is additionally defined, wherein in the evaluation individual measuring intervals are dropped, which fall below the lower limit value 4 or exceed the upper limit value 5. From the intensities between the limits, the signal for iron can be calculated. It should be noted that FIGS. 1 and 2 represent only the preliminary radio phase which is also provided in conventional methods and which is intended to homogenize the sample. This phase is followed both in the prior art and in a preferred embodiment of the invention, the actual measurement phase. However, the signals shown in FIGS. 1 and 2 are already so stable from points 2 and 3 that the measurement range for the concentration calculation of the element carbon can be obtained therefrom.

The sample shown in FIGS. 1 and 2 is a sample without precipitated elemental carbon. Here, the readings are stable over time. It may be expected that in such an ideal sample the carbon signal will be measured correctly during the actual measurement phase.

A non-ideal sample with globally precipitated carbon has been measured in FIGS. 3 and 4. In detail, Figure 3 again shows the time course of the intensity of the carbon line at 148.2 nm as in Figure 1. The other measurement parameters are the same. The length of the measuring range shown is also about 12 sec. The intensities are initially about 50,000 units, then rise rapidly thereafter. A lower limit of about 150,000 units is labeled 6. The lower limit value 6 is exceeded at the measuring interval number 25 in point 7, for example. The measuring points before the point 7 represent the unstable plasma Point 7, the measured values are evaluated.

It can be seen in FIG. 3 that the intensities initially rise to about 300,000 units per measuring interval and then drop approximately exponentially to a nearly constant intensity of 200,000 units per measuring interval. Approximately up to the measuring interval 200, which is denoted by 8, takes place over the later continuous signal, an increase in the intensities. This elevation is attributed to the subliming carbon that exits the sample and is first measured in the sublimation phase, but then lost to the measurement. The elevation between points 7 and 8 in FIG. 3 thus represents the lost carbon content due to the homogeneity of the sample.

Finally, FIG. 4 shows the iron signal at the line 149.7 nm recorded on the sample of FIG. The iron signal is stable from about point 9. Before the point 9 is here the Einfunkphase. From the point 9, that is approximately from the measuring interval 45, the signal can be recorded and evaluated.

FIG. 5 schematically shows the frequency distribution of the individual intensities as a bar chart. In the X-axis, the intensities are shown, which have been recorded in Figure 3 for each individual measuring point. The Y-axis shows the number of measurement intervals in which the corresponding intensity was measured. The shape of this representation corresponds approximately to a Gaussian function with the maximum at 200,000 units per measurement interval, as is expected from FIG. However, from an intensity of about 230,000, which is designated by the point 10, a tail is observed towards higher intensities. This spur extends from about 240,000 to over 300,000 units per measurement interval. The upper limit is marked with the number 11. The high intensity values between the points 10 and 11 correspond to the elevation of the measured values from FIG. 3 between the points 7 and 8. These measurements should be used to determine the elemental carbon.

The measuring method itself is described in a preferred embodiment in FIG. The flowchart of FIG. 6 initially foresees the start of the measurement at 12. At 13, carbon is measured at 148.2 nm and iron at 149.7 nm in the first measurement phase. In total, about 500 measurement intervals are recorded. This is illustrated by method step 14. In the process step 15, the unstable plasma phase for carbon and iron is calculated, which lies in front of the point 2 in Figure 1, in Figure 2 before the point 3, in Figure 3 before the point 7 and in Figure 4 before the point 9. Die previous measured values are discarded. In method step 16, parameters of the overshoot of the carbon signal are determined, ie in FIGS. 3 the period between points 7 and 8 is evaluated or in FIG. 5 the intensity distribution between points 10 and 11.

In a method step 17, it is determined which measurement intervals are used for obtaining a carbon and an iron sum intensity. In method step 18, the carbon concentration of the sample is then calculated from the sum intensities, including the undissolved graphitic fraction, which is determined on the basis of FIG

Measured value overshoot can be calculated. By method step 18, the pre-radio phase of the measurement is completed. This is followed by the actual measurement, which is carried out as in the prior art. In a process step, denoted overall by 19, the further elements are measured, for example silicon, chromium, nickel, magnesium and also carbon. In step 20, for the elements other than carbon, the concentration in the sample is calculated. In a method step 21, the carbon concentration is then calculated in a conventional manner from the stable signal of the measurement, wherein the undissolved graphitic carbon content is not included in the calculation. in the

Step 22 then calculates the undissolved graphitic carbon fraction by calculating the difference between the total carbon content present after step 18 and the dissolved carbon fraction determined after step 21. substance content is formed. In step 23, finally, the measurement result is output. In the output both the total carbon content can be specified as well as a separate measured value for the dissolved and the undissolved carbon content are output. The evaluation of the elevation of the carbon signal in the pre-radio phase between points 7 and 8 or 10 and 11 thus makes it possible to take account of the undissolved carbon in the measurement.

To illustrate the structure of the sample used in Figures 3 and 4, Figure 7 is attached as an example. FIG. 7 shows micrographs at 100 × magnification, which show the spherical graphite part in nodular cast iron. This spherical graphite part can be detected by the method according to the invention.

Table 1 below compares the result of a sample as shown in FIG. 7 with conventional spark spectrometry, the new method and the combustion analysis with CO 2 determination. The percentages are by weight carbon. It can be seen that at carbon contents of about 3.6% to 3.7%, depending on the nature of the sample, the deviation from the conventional spark spectrometry method to the combustion analysis is between 0.08% and 0.50% absolute, whereas for the same samples with the new ones Method a deviation between 0.004% and 0.14% is achieved absolutely. The deviation of the conventional method from the combustion analysis systematically leads to lower carbon contents, whereas the deviations of the new method compared to the combustion analysis statistically result in partly higher and partly lower measured values. A systematic deviation is not recognizable.

This shows that the elementary or undissolved carbon could not be systematically detected by the conventional method, while the new method also takes into account this carbon content. LIST OF REFERENCE NUMBERS

1st threshold

2nd point

3rd point

4. lower limit

5th upper limit

6. lower limit

7th point

8. Measuring interval

9th point

10th point

11th upper limit

12. Start

13th first measurement phase

14th step

15. Process step

16. Process step

17. Process step

18th method step

19. Process step

20. Process step

21. Process step

22nd step

23. Process step

Claims

Patent claims

1. A method for determining the carbon content of an iron alloy, comprising the steps of:

a) starting the measurement (13) of a sample in a spark spectrometer,

b) training a plasma in a pre-radio phase,

c) detecting (13) and recording (14) an intensity signal of the carbon,

d) calculating and hiding (15) an unstable plasma phase, e) calculating an excess of the carbon signal (16, 17) and calculating the content of dissolved and undissolved carbon (18).

2. The method of claim 1, wherein after one of steps c), d) or e), the concentration of carbon in a step f) is measured in a conventional manner (19).

3. Method according to one of the preceding claims, characterized in that after one of the steps c), d) or e) further elements are measured in a conventional manner.

4. The method according to any one of the preceding claims, characterized in that the result of step f) than the concentration of dissolved carbon is issued.

5. Method according to one of the preceding claims, characterized in that in the first measuring phase (13) the carbon signal is measured at a wavelength of 148.176 nm.

6. Method according to one of the preceding claims, characterized in that the pre-radio phase is carried out over a period of 8 seconds to 15 seconds.

7. Method according to one of the preceding claims, characterized in that the iron signal is recorded and recorded during the pre-radio phase.

8. A method according to any one of the preceding claims, wherein a ferric emission line is measured at 149.653 nm.