Analysis of Steel Inclusions Using Optical Emission Spectroscopy

By AZoM

Table of Contents

Introduction

Optical emission spectrometry (OES) is a fast, simple and cost-effective analytical technique used for elemental analysis of solid iron and steel samples. It is used in several fields ranging from production to recycling and from foundries to service laboratories. The Thermo Scientific ARL 4460 Metals Analyzer (Fig. 1) is a high-performances, high-precision and high-accuracy OES spectrometer for analysis of iron and steel.

In addition, the Spark-DAT (Spark Data Acquisition and Treatment) considerably extends the capability of this spectrometer beyond the characterization of elemental concentrations by enabling as well ultra-fast inclusion analysis. In the steel industry, the ARL 4460 spectrometer with Spark-DAT option is gaining increasing popularity, offering information about the inclusions during the steel elaboration process.

Figure 1. Thermo Scientific ARL Spectrometer Series.

Benefits

The advantages of using the ARL 4460 spectrometer with Spark-DAT option are the following:

• Drastic reduction of investment costs for inclusion analysis.
• Extreme shortening of the time for inclusion analysis and related sample preparation.
• No additional cost and time for performing operations in comparison to the standard OES spectrometer.
• Capability of performing in parallel inclusion analysis and analysis of elemental concentrations at a rate of more than 30 samples/hour.
• Availability of the information on inclusions shortly after sample analysis.This offers unequalled perspectives for controlling the metal elaboration process on-line with the aid of information related to inclusions.
• Quantitative or semi-quantitative analysis of oxygen at levels of 30 ppm and lower in killed steels.
• Easier detection of rare or large inclusions.

Principles

The ARL 4460 spectrometer with Spark-DAT option has a range of acquisition and treatment principles when compared to standard OES spectrometers. Its salient features are:

• The intensity (or the amplitude) of a Spark-DAT signal depends on the composition of the sample at the position struck by the corresponding single spark.
• If the concentration of an element in the ablated sample material is considerably higher than the concentration of its soluble form in the matrix, the outcome is an intensity peak. This is typically the case when a spark hits a sample area containing an Al based inclusion (e.g. Al₂O₃), because the Al concentration is much higher than in the metal matrix due to the contribution of the inclusion. This is better explained in the following example: the concentration of Al in an Al₂O₃ particle is roughly 53%, and if an inclusion with a size of 5 mm ESD (Equivalent Spherical Diameter) is ablated together with the sample material, this gives an apparent rise in the concentration of 0.075%. Graphically, this is shown in figure 2.

Figure 2. Example of Spark-DAT acquisition for Al-related inclusions. The intensity of the flat, noisy baseline signal is proportional to the concentration of Al atoms dissolved in the matrix and the peak intensity depends on the amount of Al atoms contained in the inclusion(s) ablated by a single spark. Consequently, the number of peaks is related to the number of such inclusions and their intensity to such factors as the inclusion size and the concentration of Al in the inclusion.

Analysis Time

The Spark-DAT analysis alone takes around 7 s for a single measurement (including 2 s Ar flush). This mode is recommended only for rapid counting and confirmation of inclusion types and for obtaining raw data for off-line interpretation. The analysis time for a single measurement is typically 22 s. Hence, inclusion analysis is possible in a number of contexts, specifically during steel production, where analysis times are highly critical.

Figure 3. Step-by-step diagram of the operations performed during Spark-DAT analysis.

Sample Preparation

The standard OES sample preparation can be used for Spark-DAT inclusion analysis. However, paper grinding must be used only if the potential contamination by paper would not influence the results for the inclusions of interest. As aluminum oxide inclusions are generally critical in steel, it is recommended using Al₂O₃-free grinding papers. Milling is ideal for inclusion analysis as it guarantees a clean, uncontaminated surface.

Contexts of Use of Spark-DAT Inclusion Analysis

Spark-DAT inclusion analysis offers tremendous potentials for:

• Inclusion control for quality purposes. The most important application is inclusion analysis during steel elaboration.
• Process control through on-line monitoring of inclusions. The inclusions are “process markers or tracers” that indicate a change in the process.
• Samples screening. Hundreds of samples can be screened for inclusions on a daily basis. This may help solving rapidly critical quality problems.
• Replacement of long and costly analysis techniques. The Spark-DAT technique can replace traditional techniques of inclusion analysis. It also offers the possibility to replace techniques that determine the characteristic of steel depending on the inclusions it contains (e.g. fatigue resistance), if a correlation can be established between Spark-DAT inclusion analysis and the results of such technique.

Soluble/insoluble Analysis

The insoluble concentration and the ratio to the total concentration of some elements such as Al, Ca, Ti, B are traditional indicators of the steel making process and are still utilised extensively. The algorithm Soluble is therefore one of the most commonly used Spark-DAT application. It allows for the calculation of the soluble or insoluble part of an element. With Soluble, the ratio R_sol of the sum of the intensity signals due to the matrix soluble part to the sum of all the intensity signals is calculated as shown in the following equation:

In the case of Al, the concentrations of the soluble and insoluble parts are obtained in the following way:

where Al_tot is the concentration of total Al measured by OES.

Without the Spark-DAT option, this analysis is as well possible with an OES method based on multiple integrations: the Peak Integration Method (PIM). However, the Spark- DAT based method is rapid and needs no calibration sample. The method therefore is also applicable to any partially insoluble element, like B, Ti and Ca, even in the case where no certified reference material (CRM) exists.

The following table shows the excellent accuracy obtained on CKD low alloy steel standards having certified Al_sol concentrations:

Evaluation of Number and Type of Inclusions

The simplest Spark-DAT application includes counting intensity peaks on the channel of a given element with the algorithm Peaks. In general, a peak is defined as the intensity signal (I_peak) higher than a threshold situated at the mean intensity (m) of the element dissolved in the matrix (I_matrix), plus three times its standard deviation (SD):

      I_peak > m + 3•SD_Imatrix

Counting intensity peaks allows to determine the number of inclusions containing this element.

As shown in figure 4, clean and dirty steel samples can easily be differentiated by comparing the number of peaks counted on the channels of the inclusion elements. The algorithm Composition enables counting coincidental peaks, i.e. peaks appearing on the channels of several elements concurrently during the same single spark. The coincidence of a peak on Ca and S channels, for instance, means that these two elements are part of the same inclusion (e.g. a calcium sulfide (CaS) inclusion).

Figure 4. Examples of Spark-DAT Peaks analysis for two samples of clean (Left) and dirty (Right) steel samples.

In the previous example, one CaS coincidence was counted in the clean sample and 76 were counted in the dirty one. Coincidences of up to four channels can be counted with the algorithm Composition. This allows the chemical formulation of complex inclusions or inclusion clusters. Furthermore, the possibility to check for non-coincidences in addition to coincidences helps removing ambiguities on the inclusion type.

This is shown in the following example, where the number of Ca aluminate, Al oxide and Ca oxide inclusions were evaluated as follows (Fig. 5):

• Al oxide Al₂O₃: by counting Al peaks coincident with a peak on O channel and non-coincident with a peak of the Ca channel.
• Ca aluminate inclusions Al₂O₃-CaO: by counting Al peaks coincident with peaks of Ca and O channels.
• Ca oxide CaO: by counting Ca peaks coincident with a peak of the O channel and non-coincident with a peak of the Al channel.

Figure 5. Examples of Spark-DAT Peaks analysis for two different samples.

Qualitative Size and Size Distribution

Knowing the size of the inclusions or, better, their size distribution is important. Large inclusions are normally the most detrimental to the metal quality. The two algorithms Peaks and Composition can also be used to count signals belonging to a range of intensity classes. Because the peak intensity is related to the volume of the inclusion, these classes can qualitatively be considered as size classes. Setting the threshold 3xSD above the intensity of the element in the matrix enables counting of all visible peaks. Setting it higher, for example at 9x or 15xSD as in the following example (Fig. 6), allows counting only the inclusions of larger size. Calculating the inclusions between consecutive threshold values enables determine the number of inclusions in the size class that they delimit.

Figure 6. Examples of Spark-DAT analysis using Peaks and Composition algorithms for identification of different intensity classes.

In the example, the number of peaks and coincidences between 3x and 9xSD correspond to small size inclusions, between 9x and 15xSD to medium size inclusions and higher than 15x to large size inclusions. Such calculations allow generating qualitative inclusion size distributions.

In the following example (Fig. 7), the horizontal axis gives the intensity classes of inclusion signals of Ca, incremented in steps of 1xSD between (m + 3xSD) and (m + 20xSD) and in steps of 10xSD afterwards. The vertical axis offers the number of inclusions per cubic millimeter of steel. The resulting qualitative size distribution diagram can be used, for instance, to compare inclusion distributions in samples of different heats.

Figure 7. Diagram of Ca inclusions per cubic mm vs. intensity class.

Quantitative Spark-DAT analyses

The Spark-DAT algorithm QuIC (Quantification of Inclusion Content) enables quantitative analysis of inclusions in terms of size and size distribution. It allows calculating the average ESD (Equivalent Spherical Diameter) for size classes of various inclusions types. Other parameters like the total oxygen content can also be calculated.

Inclusions Quantifiable with Spark-DAT Option

Various types of endogenous and exogenous inclusions can be observed directly or indirectly in steel with the ARL 4460 spectrometer with Spark-DAT option, e.g. oxides (Al₂O₃, MgO, CaO, MnO, TiO₂, SiO₂, etc.), spinels (Al₂O₃-CaO, Al₂O₃-MgO, etc.), sulfides (CaS, MnS, AlS, etc.) and others. The identification of an inclusion type is limited mainly by the sensitivity of the analytical lines used, by the size of the inclusion and by the concentration level of the inclusion elements as soluble elements in the matrix: higher line sensitivity and lower soluble content allow determining smaller inclusion ESD’s. For example in a steel with 50 ppm of Al, the minimum detectable ESD is about 1 mm, while with 0.2% of Al the minimum ESD detectable is 4.5 mm.

On-line Analysis

The results of Spark-DAT analyses such as soluble concentration, number of intensity peaks and number of coincidental peaks, can be monitored simultaneously with concentration values. The Spark-DAT results can be processed, displayed, transmitted and stored like any standard OES result (Fig. 8).

Figure 8. Example of Spark-DAT results display.

Off-line Investigations

The Spark-DAT intensity data can be stored in standard text (.txt) or comma separated value (.csv) files (Fig. 9). These files can then be used off-line for investigations on inclusions or for research and development of new methods or algorithms for instance.

Figure 9. Spark-DAT intensity data shown via the Spark-DAT viewer integrated in OXSAS.

Conclusions

The Spark-DAT option of the ARL 4460 metals analyzer increases the versatility of the OES spectrometer. From routine uses to research applications, Spark-DAT methods provide quick, simple and cost-effective solutions for inclusion analysis in the steel industry. Among all the inclusions analysis methods available today for the steel industry, the Spark-DAT based methods are the fastest. The Spark-DAT standard option allows soluble/insoluble determination, ultra-fast on-line counting of inclusions, identification of their composition and qualitative size classification in a time ranging from several seconds to a couple of minutes. This makes it highly effective for controlling inclusions and steel cleanness during production.

Furthermore, the advanced option allows quantitative determinations of inclusion parameters such as size and total oxygen content.The inclusion analysis can be combined with the standard analysis of elemental concentrations. The sample and its surface preparation, as well as the instrument maintenance and consumables, are equivalent when compared to a standard OES instrument. This means extremely low operation costs compared to the other inclusions analysis techniques that require dedicated instruments. In addition, the ability to obtain elemental analysis information and inclusion contents with a single OES instrument greatly reduces investment costs.

About Thermo Fisher Scientific-Elemental Analysis

For over 75 years, Thermo Fisher Scientific has been a worldwide supplier of spectrochemical instrumentation to major industries including steel, transportation, cement, construction, food, pharmaceuticals, chemicals, academic research, petroleum and electronics. They offer unsurpassed capabilities in the areas of optical emission (OE), X-ray fluorescence (XRF), X-ray diffraction (XRD) and automation of spectrometers.

This information has been sourced, reviewed and adapted from materials provided by Thermo Fisher Scientific- Elemental Analysis.

For more information on this source, please visit Thermo Fisher Scientific- Elemental Analysis.