Using Handheld LIBS for Carbon Analysis in Low-Alloy and Carbon Steels

This article illustrates how carbon content in low alloy and carbon steels can be analyzed utilizing the technique of handheld laser induced breakdown spectroscopy (HH LIBS). This method recommends SciAps Z-200, the world’s only handheld analyzer that can analyze carbon content in alloys. A pulsed, 1064 nm laser, working at 5.5 mJ/pulse and 50 Hz repetition rate is used by the Z-200 analyzer. The on-board spectrometer spans 190 nm – 620 nm, with resolution less than 0.13 nm in the 193 nm range of the carbon line utilized. An on-board, user-replaceable argon purge gas is also used by the analyzer. Almost 600 tests are provided by the argon canister located in the handle, before it requires replacement.

The Carbon App Overview

Given below is a list of features included with the Carbon App:

  • Full calibration in carbo and low-alloy steels for other elements including Cr, V, Ti, Al, Si, Mn, (Fe by difference), Pb, Mo, Nb, Cu, Ni, Co
  • ProfileBuilder desktop/tablet software for user-generated carbon calibrations on a variety of ranges or bases
  • Carbon cal check and drift correction standards (3)
  • Carbon calibration from 0-1% in low-alloy steels and carbon steels

It is possible to add the Carbon App to any existing Z-300 or Z-200 analyzer

Performance Summary

Carbon data has been acquired from several analyzers, on carbon steels and low alloy steels (LAS) ranging from a pure iron (<0.005%) up to 1.2%. The test times for all analysis range between 15-22 seconds for properly ground materials, based on the level of shot rejection carried out by the method algorithm (more later). Users separating steels that vary by 0.2% carbon or more can normally complete tests in 10-15 seconds. Table 1 summarizes the performance results.

Table 1. Summary performance parameters Z-200 for carbon

Parameter Value (% absolute) Comment
Limit of Detection 0.12 3-sigma detection level for C in carbon and LA steels
Precision (absolute) 0.035
Accuracy 0.1
Test time, Properly ground materials 9-12 sec. for 0.2% carbon delta
15-22 sec. for 0.1% carbon delta
Includes pre-burn and purging time

On applying the conventional 3-sigma rule for detection, the detection limit of the current analyzer was found to be 0.12% C. Therefore, this method should not be used to separate H and L grade stainless from standard grade stainless. At any single carbon level, the precision (repeatability) was generally about ± 0.035% absolute, i.e. a 1030 steel with 0.30% C will produce a standard deviation of 0.035% C. The accuracy (bias) of any measurement can be as much as ± 0.1% for an accurately calibrated Z-200.

Steels that vary by 0.1% can be separated, even though the accuracy is specified as ± 0.1%. The accuracy cited in Table 1 is for a global calibration spanning carbon steels, a range of Cr-Mo steels and low-alloy steels. An increasingly expansive calibration range results in less accuracy for any given type of alloy.

Multiple calibrations for particular carbon steel types are supported by the Z’s software. The analysis uncertainty is dominated more by precision than calibration-curve bias, as a result of reducing the range of steels in the calibration. Generally, users limit the calibration to a range of matrix-similar alloys in order to establish where precision dominates the measurement uncertainty, and to separate alloys that vary by 0.1% carbon. For instance, a user can effortlessly produce a calibration from the global factory calibration that uses only carbon steels 1050, 1117, 1030, 1020, 1010, to attain separation of 1020 from 1030, etc. It is also possible for the user to produce a second type curve using low alloy steels such as 41XX’s, 4820, 4620 and 4340 to examine 4130 from 4140. The analyzer software allows easy switching between curves and enables the use of multiple calibration curves.

Calibration and Precision Data

Figure 1 shows the global calibration curve. As with spark OES, the calibration ratios the 193.1 nm carbon intensity to the intensity of a close Fe line. The calibration is then a fit of known C assays to the intensity ratio of carbon to iron. A variety of carbon steels 10XX, and 1117, plus many LAS including 46XX, 41XX, 43XX, and 86XX are used by the global curve. When equipped with the Carbon App, the Z-200 analyzer will be factory-calibrated with the global carbon curve referenced above. Use of the global curve is generally acceptable for basic separation of carbon steels that vary by 0.2% C or more.

Global carbon calibration spanning carbon steel and various alloy steels.

Figure 1. Global carbon calibration spanning carbon steel and various alloy steels.

Calibration to Carbon Steel Sub-Types, When to Use it

It is recommended that the calibration curve be limited to a family of alloys that cover the steels of interest in order to achieve more precise sorting of carbon steels – those that vary by as little as 0.1% C. For instance, in order to separate a series of carbon steels such as 1030, 1020 and 1010, the global calibration curve can be modified by turning off the low-alloy steels, and then maintaining only carbon steels in this concentration range of interest. Figure 2a, for example, displays a particular carbon steel calibration, starting with the global factory carbon calibration, and then limiting it to carbon steels from 0 to 0.5%. As illustrated, with this more type-specific curve, the Z-200 analyzer will then be able to yield reliable separation of these carbon steels, such as separating an A106 from 1018 or 1040 or separating a 1020 from a 1030.

Same calibration with several carbon steels in limited range.

Figure 2a. Same calibration with several carbon steels in limited range.

Two Important Notes:

  • SciAps always recommends using a linear fit and at least 4 calibration points (iron blank can be one). This prevents artifacts from incomplete sample prep from biasing the calibration. If an incorrectly prepped calibration sample is incorporated, it will not lie on a straight line fit.
  • SciAps does not recommend attempting to separate carbon steels that differ by less than 0.1% carbon.

Case Study

Separate 4130 from 4140 Low Alloy Steels

As a second example, let us consider the separation of 4130 from 4140 low alloy steels. Beginning with the global curve in the calibration software (ProfileBuilder), the user can produce a new calibration curve by enabling only the 8620, 4620, 4140, 4130, together with a few other low alloy steels with a carbon range from blank to 0.5%. Figure 2b displays the calibration curve. Tables 2 and 3 show the repeatability data for both 4130 and 4140, respectively, and clearly demonstrate the reliable separation of these two common low alloy steels by their difference in carbon content.

Carbon calibration with low alloy steels 41XX, 43XX, 4620, 4820

Figure 2b. Carbon calibration with low alloy steels 41XX, 43XX, 4620, 4820

Handheld XRF operators may try to separate 4130 from 4140 low alloy steels depending on the Mn content. However, this can be risky due to the presence of a large Fe interference with XRF. Even minimal drift or surface contamination can affect the Mn result in either direction. The Z provides a means to measure both the C and Mn content, for a more reliable analysis of these alloys.

4140 Low-alloy Steel
Result Carbon %
1 0.431
2 0.401
3 0.39
4 0.4386
5 0.3828
6 0.3489
7 0.4002
8 0.3974
9 0.3727
10 0.4292
Avg. 0.399
Std. Dev 0.028
RSD (%) 7.0%

4130 Low-alloy Steel
Result Carbon %
1 0.3454
2 0.2995
3 0.3152
4 0.3511
5 0.3127
6 0.2911
7 0.3081
8 0.3492
9 0.3532
10 0.3403
Avg. 0.327
Std. Dev 0.024
RSD (%) 7.2%

Repeatability of results for 4130 and 4140 low-alloy steels is shown in Tables 2 and 3, respectively. Assayed carbon content is 0.29% and 0.41% C for 4130 and 4140, respectively.

Carbon Steel Example

Tables 4 and 5 show the precision data for carbon steels A108 (0.15% C) and 1030 (0.331% C). Precision and relative standard deviation is similar to the low alloy steels. Separation of carbon content in alloys differing by approximately 0.1% or more is achieved through using a more limited curve again.

Single-Standard Type Calibration

Single alloy type standardization is not currently supported by the Carbon App. A surface prep artifact (i.e. improper grinding) can be added into the type standard when testing it. If other pieces of the same alloy material are ground differently, they may not be properly identified, in spite of being the same alloy. SciAps’ studies leading to this method validate that a minimum 4-point calibration with a linear fit are sufficient to expose improper sample preparation.

Analysis of Real-World Materials

Materials from several real-world applications were tested as part of the field testing of this new carbon method. This article presents results from some refinery piping provided by a top refining company. These components were earlier in-service components and are common A108 and 1010 carbon steel alloys. They were prepared based on the grinding method described later, and examined with the same testing procedure. The same carbon steel calibration displayed in Figure 2a was utilized for these materials. The real-world carbon steel measurements functioned similarly to the reference material tests.

A108 Carbon Pipe Steel
Result Carbon %
1 0.1895
2 0.1452
3 0.1517
4 0.2315
5 0.1748
6 0.1762
7 0.187
8 0.165
9 0.1735
10 0.2369
Avg. 0.183
Std. Dev 0.030
RSD (%) 16.5%

1030 Carbon Steel
Result Carbon %
1 0.3761
2 0.3102
3 0.3445
4 0.3291
5 0.2873
6 0.3193
7 0.3612
8 0.3546
9 0.3963
10 0.3785
Avg. 0.346
Std. Dev 0.034
RSD (%) 9.9%

There are four steels that contain carbon concentrations of 0.23%, 0.18%, 0.12%, and 0.073%. The results of the Z-200 LIBS data versus the assayed data are shown in Figure 3. Repeatability data for the 0.12% and 0.18% carbon contents is shown in Tables 6 and 7. Repeatability on the other two samples was similar and was omitted for brevity. These results on actual refinery piping are similar to the data obtained from reference materials. Carbon steels that differ by 0.1% carbon or more are effortlessly separated, for example 0.073 and 0.18 or 0.12 and 0.23% carbon. The results for the 0.12% and 0.18% steels are also shown to demonstrate that even with a good level of precision (± 0.02% carbon, good for a handheld device), when considering the precision band, these two alloys cannot be separated.

Results for refinery piping materials.

Figure 3. Results for refinery piping materials.

Piping Steel C 0.12%
Result Carbon %
1 0.202
2 0.138
3 0.152
4 0.153
5 0.171
6 0.174
7 0.115
8 0.169
9 0.130
10 0.160
Avg. 0.157
Std. Dev 0.025
RSD (%) 15.8%

A108 Pipe C 0.18%
Result Carbon %
1 0.2222
2 0.1984
3 INC
4 0.1983
5 0.1877
6 0.1888
7 0.2336
8 0.2329
9 0.2033
10 0.2005
Avg. 0.207
Std. Dev 0.018
RSD (%) 8.6%

Method

Sample preparation with a handheld grinder is required by the analysis method, followed by testing with the Z-200. The R980 quick change discs and the Milwaukee M12 grinder were used for the data collected during the development of this app.

Milwaukee M12 grinder and quick change discs

Figure 4. Milwaukee M12 grinder and quick change discs

Definitions

A test refers to a single test of the material with the Z LIBS analyzer. A result is treated to be a final answer that comprises of five (5) valid LIBS tests which are automatically averaged by the analyzer software.

A result is typically 15 seconds, due to the 5-test average, as each test takes 3 seconds. A result may need up to 22 seconds if the software rejects one or two of the tests based on built-in data quality rejection criteria. These times are required to attain carbon separations that vary by 0.1% carbon.

A test is defined as a single analysis on the material, comprising of spectral and pre-burn data from six different raster locations. Figure 5 displays a test showing the six laser burns in the material.

A result is defined as an average of 5 valid tests, and shows the measured percentage of carbon and the measurement uncertainty.

The testing works as follows. The operator carries out five tests on the ground material, which is automatically averaged by the analyzer software to obtain a final result. One or more tests could be rejected by the built-in rejection algorithm, thus requiring a total of 6 or 7 tests to obtain an average of 5 good tests. Rejection of more than 1 or 2 tests highlights the fact that the material was not properly ground and should therefore be re-ground.

The rejection algorithm exploits the 2D rastering of the laser carried out by the Z-series analyzer. The laser is rastered to six discrete positions on the material, spaced about 200 µm apart, for every single test. The sample is pre-burned at every single location for 0.2 seconds and spectral intensity data collected for 0.3 seconds for each position (for a total of 0.5 seconds per location). The software rejects any test where the standard deviation is more than a factory preset threshold. The user is prompted by the software for additional tests until the required 5 good tests are attained and averaged for a result.

Improper sample grinding is indicated by a high carbon uncertainty from a test. In these cases, the laser has probably struck a region with high carbon surface contamination that was not removed by grinding. The overall result will be biased high if the resulting test is not rejected. The sample can be considered to be properly ground if zero or perhaps one test is rejected during a carbon measurement. A reliable quality indicator, establishing the fact that the sample material was properly prepared, is provided by the test rejection criteria built into the SciAps Carbon App.

Conclusion

The SciAps Z-200 or Z-300 handheld LIBS analyzers currently offer carbon concentration measurements in low-alloy steels and carbon steels. Sample grinding using a specified handheld grinder is required by the method, followed by a (typical) 15 second test with the Z. Pre-burn and purging time is included in the testing time. If operators follow the described procedures, the Z will enable reliable separation of carbon steel grades that differ by 0.1% C or more. The Z uses argon purge to achieve the required precision, and a data rejection algorithm to guarantee proper grinding. Consistently good argon purge and sample preparation are important for carbon analysis with HH LIBS.

This information has been sourced, reviewed and adapted from materials provided by SciAps, Inc.

For more information on this source, please visit SciAps, Inc.

Citations

Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    SciAps, Inc.. (2019, May 29). Using Handheld LIBS for Carbon Analysis in Low-Alloy and Carbon Steels. AZoM. Retrieved on February 29, 2020 from https://www.azom.com/article.aspx?ArticleID=13562.

  • MLA

    SciAps, Inc.. "Using Handheld LIBS for Carbon Analysis in Low-Alloy and Carbon Steels". AZoM. 29 February 2020. <https://www.azom.com/article.aspx?ArticleID=13562>.

  • Chicago

    SciAps, Inc.. "Using Handheld LIBS for Carbon Analysis in Low-Alloy and Carbon Steels". AZoM. https://www.azom.com/article.aspx?ArticleID=13562. (accessed February 29, 2020).

  • Harvard

    SciAps, Inc.. 2019. Using Handheld LIBS for Carbon Analysis in Low-Alloy and Carbon Steels. AZoM, viewed 29 February 2020, https://www.azom.com/article.aspx?ArticleID=13562.

Ask A Question

Do you have a question you'd like to ask regarding this article?

Leave your feedback
Submit