Evaluation of Inclusion Characteristics in Low-Alloyed Steels by Mainly Using PDA/OES Method

Abstract

The characteristics of non-metallic inclusions (such as number, size and volume fraction) in liquid steel samples taken during ladle treatment and casting of industrial heats of two low-alloyed Ca-treated steel grades were evaluated by using the Pulse Distribution Analysis with Optical Emission Spectroscopy (PDA/OES) method. These results were compared to data obtained by Scanning Electron Microscope observations of inclusions after electrolytic extraction from steel samples (the EE method). It was found that the PDA/OES method can be used for a relative estimation of the homogeneity of the distribution of non-metallic inclusions in steel samples. Bottom and middle parts of the steel samples showed more homogeneous results with respect to the characteristics of the investigated Al₂O₃, CaO–Al₂O₃ and CaO–Al₂O₃–CaS inclusions. The numbers of inclusions in the size ranges 2.0–5.7 µm and 1.4–5.7 µm in samples taken before and after a Ca addition, respectively, showed a relatively good agreement between both methods. Furthermore, the calculated volume fractions for inclusions in the size range 2–13 µm obtained by the PDA/OES method agreed satisfactorily well with those obtained from the EE method. Finally, the minimum sizes of inclusions in steel samples, which can reliably be detected by the PDA/OES method, were estimated for steels with different concentrations of Al in steel and Al₂O₃ in inclusions.

1. Introduction

Due to tightened requirements of the steel quality (particularly for high quality steel grades) in recent years, it has become necessary to obtain fast estimations of inclusion characteristics in liquid steel in order to control and correct the steelmaking processes during ladle treatment and casting. Pulse Distribution Analysis combined with Optical Emission Spectroscopy (the PDA/OES method) is an analytical technology for fast inclusion characterization, which has been developed during the recent decades.^1,2,3,4) Some details regarding the analysis equipment, data acquisition and data processing can be found in earlier publications.^5,6) Today the PDA/OES analysis of inclusion characteristics in steel samples may take 3–5 minutes in an industrial practice. Currently, some steelmaking companies of the European Union try to use the PDA/OES method for fast determination of the inclusion characteristics in steel samples taken from different stages of the productions of various steel grades.^{7,8,9,10,11,12)} In most cases, the number and size of analyzed non-metallic inclusions are usually characterized by the PDA index (the amount of outliers on the intensity chart, which are considered as non-metallic inclusions)^12,13) or the B-factor (defined as the total summed weight of elements in inclusions),⁹⁾ which correspond to the total relative quantitative characteristics of inclusions in steel.

Pande et al.¹¹⁾ applied the PDA/OES method for determination of the number of non-metallic inclusions and the average inclusion size in metal samples taken during ladle treatment and casting of ultra-low carbon steels (≤30 ppm C). The average size of inclusions (mostly Al based) was determined by using three different methods: (A) determination from the average weight of inclusions, which was evaluated by dividing the weight of the insoluble Al content by the number of Al based inclusions detected by the PDA/OES method; (B) determination from the inclusion size distribution calculated from the individual Al peak intensity PDA/OES data, and (C) determination from the inclusion size distribution obtained by using chemical extraction (dissolution of 0.2 g of steel in HCl acid at 90–100°C) with the following SEM investigation of non-metallic inclusions. It was reported that the average size of inclusions in steel samples determined by the both PDA/OES methods was marginally greater (on ~15–36% for method A and on ~4–18% for method B) than that obtained by the chemical extraction (method C). However, an influence of the concentrations of soluble Al and the number of small size particles in steel samples on the minimum inclusion size, which can be detected by the PDA/OES method, is not determined quantitatively in this article. Moreover, the sizes of inclusions, which were measured after chemical extraction by using acids (method C), can significantly be underestimated due to the partial or complete dissolution of some phases in inclusions in the HCl acid.^14,15)

This study is focused on the evaluation of reliable inclusion characteristics (such as number, size and volume fraction) in industrial samples using the PDA/OES method. The PDA/OES results were compared to inclusion data obtained by electrolytic extraction of steel samples followed by observations using a Scanning Electron Microscope (SEM). Moreover, some limitations of the PDA/OES method, which depend on the concentration of the main elements and on the number of small size inclusions present in the steels, are considered in this study.

2. Experimental

2.1. Industrial Scale Sampling

Six industrial heats of two low-alloyed Ca-treated steel grades were carried out. Typical compositions of the steel grades are presented in Table 1. Samples of the type “dual thickness lollipop” (12/4 mm) were taken from the liquid steel from the ladle (samples L1, L2 and L3) during the CAS-OB (Composition Adjustment by Sealed argon bubbling – Oxygen Blowing) treatment and from the tundish (samples T4 and T5) during continuous casting of the steel melt, as shown in Fig. 1. All samples were taken manually by using argon-protected samplers.

Table 1. Typical composition of investigated industrial low-alloyed steel grades [in mass%].

Steel grade	Heat No	C	Si	Mn	P	S	Al	Nb
A	1, 2, 3	0.14	0.2	1.5	0.01	0.002	0.04	0.015
B	1, 2, 3	max 0.09	max 0.031	max 1.3	max 0.025	max 0.004	max 0.015	max 0.05

Fig. 1.

Schematic illustration of the main steps of steel production and the sampling moments during the trials.

2.2. PDA/OES Measurements

The principle of the PDA/OES technique is shown in Fig. 2.^16,17,18) A high-energetic discharge of electric sparks (3000–4000 sparks) with a frequency of 100–800 Hz ablates the material of the sample surface and excites sample atoms and ions in the plasma to elevated electronic energy levels. Thereby, light of different wavelengths specific for each element is emitted. These wavelengths are separated by a diffraction grating in the spectrometer. The light intensity for each wavelength is measured with photomultipliers. Based on a specific calibration function, the total mass fraction of each element in the ablated material is determined.

Fig. 2.

Schematic illustration of a PDA/OES measurement.

Examples of single spark intensities (“pulsograms”) for Al and Ca in a steel sample obtained from one PDA/OES measurement are shown in Fig. 3. The high light intensity peaks (outliers), which are related to the given elements present in non-metallic inclusions (NMI), can be evaluated by the Pulse Distribution Analysis. These outliers are identified as intensities exceeding the median intensity of the metallic background (bulk intensity), with a specified number of standard deviations (σ). The σ value is determined by means of an iterative calculation, including a removal of large outliers. The identification of inclusion types is based on the coincidences of outliers detected in single sparks. If high intensity outliers e.g. Al and Ca are present in the same spark, a complex inclusion containing both elements has most likely been ablated. The obtained information is quantified to provide the size, number and composition of inclusions in steel samples.

Fig. 3.

Typical intensity distributions obtained from PDA/OES measurements of the (a) Al and (b) Ca contents in the steel samples.

The PDA/OES measurements were performed using parameters which ablate on average about 5·10⁻⁸ g of material per spark.⁹⁾ The total ablated weight after 4000 sparks was about 2·10⁻⁴ g per measurement (spot). This corresponds to an analyzed volume of about 2.56·10⁻² mm³ per measurement. Five PDA/OES measurements were done on each sample after milling the surface of the thicker part of the lollipop sample, as is shown in Fig. 4. The total analyzed area of five spots on each metal sample was about 100 mm². Based on the industrial practice at the plant where the PDA/OES software was used, all outliers outside of +2.5σ, +1σ and +2.7σ interval from the median intensity of Al, Ca and S respectively in the metal matrix were considered as inclusions in the steel samples. The details of the calculation algorithm can be found elsewhere.¹⁾

Fig. 4.

Schematic illustration of a steel sample, including the zones used for PDA/OES determinations and electrolytic extraction.

The used PDA/OES software determines the number of outliers of intensity peaks for each type of non-metallic inclusion and recalculates these values to the number of inclusions per unit volume of a steel sample, N_V(PDA/OES). The size of inclusions is calculated from the intensity of each single outlier. More specifically, the weight of ablated inclusions is determined from the outlier intensity using the calibration function for the respective element. Thereafter, an equivalent diameter (assuming a spherical shape of inclusions) is calculated for each outlier using a “global” density of 0.0039 g/mm³.¹⁾ The obtained results were presented as the Particle Size Distributions (PSD) in which fixed steps of the size ranges 0–1.4, 1.4–2.0, 2.0–2.8, 2.8–4.0, 4.0–5.7, 5.7–8.0 and 8.0–13.0 μm were used.

2.3. Electrolytic Extraction and Investigation of Non-metallic Inclusions

During the last years three-dimensional (3D) investigations of inclusions and clusters on a film filter after electrolytic extraction (EE) of steel samples have successfully been used together with conventional two-dimensional (2D) investigations on a polished surface of a steel sample.^19,20,21) By application of an EE process, the metal matrix of the steel sample can be dissolved in an electrolyte. However, the NMI and clusters do not dissolve during the electrolytic extraction. They can be collected on a surface of a film filter by filtration of the electrolyte after an EE process.

In this study, 18 electrolytic extractions followed by 3D investigations of NMI by using SEM were performed in the steel samples before (L2) and after (L3 and T4) a Ca addition. After the PDA/OES measurements, a specimen for EE was cut from the steel samples, as is shown in Fig. 4. These specimens from the steel samples taken before and after a Ca addition were dissolved using 10% AA (10 v/v% acetylacetone - 1 w/v% tetramethylammonium chloride - methanol) and 2% TEA (2 v/v% triethanolamin-1 w/v% tetramethylammonium chloride-methanol) electrolytes, respectively. The 2% TEA electrolyte is preferably used for the samples taken after a Ca addition, due to the risk of dissolution of CaO inclusions when using the 10% AA electrolyte.¹⁴⁾ A charge of 500 C and a voltage of 150 mV were used for all extraction experiments. The total weight of the dissolved metal during the electrolytic extractions varied in the range of 0.12 to 0.14 g. A polycarbonate (PC) film filter with an open pore size of 0.05 μm was used during the filtration of the electrolyte to separate the undissolved NMI after the EE process.

The observations of inclusions on the film filters were done using a SEM. The observed area varied from 0.17 to 1.21 mm². On average about 400 inclusions were investigated per sample using magnifications of ×1000 and ×3000 to observe large and small size particles, respectively. This corresponds to an analyzed volume of about 2.3·10⁻³ to 9.9·10⁻³ mm³. The size was determined by measuring the diameter of inclusions or in case of big irregular inclusions, the average of the length and the perpendicular width of inclusions.

The total number of inclusions per unit volume, N_V(EE), and number for each i-th size range class in each particle size distribution, N_Vi, were calculated as follows:   

$$N_{V\left(EE\right)}= {{\displaystyle \sum }}^{}{N}_{Vi}$$(1)

  

$$N_{Vi}=n_i\cdot \frac{A_f}{A_{obs}}\cdot \frac{{\rho }_{Me}}{W_{dis}}$$(2)

where n_i is the number of inclusions in the i-th size range. A_f and A_obs are the total area of the film filter with inclusions (~1260 mm²) and the total observed area, respectively. ρ_Me is the density of the steel (=0.0078 g/mm³ in this study) and W_dis is the dissolved weight of the metal sample during extraction.

The compositions of the different inclusions on the film filter were analyzed using SEM-EDS.

3. Results and Discussion

3.1. Distribution of Non-metallic Inclusions in Different Sample Zones

In this study, the deviation from the average values of Al₂O₃, CaO and CaS concentrations, ∆C, obtained from the PDA/OES measurements in different zones of each steel sample is used as a parameter for the evaluation of the inclusion distributions in the steel samples. The value of ∆C for the i-th zone was calculated as follows:   

$$\mathrm{\Delta }{C}_i=100\%\cdot \left|\overline{C}-C_i\right|/\overline{C}$$(3)

where C_i and C are the concentration of a given component in the i-th zone and the average concentration for all five zones in each sample, respectively.

The results obtained in different sample zones (Zones 1–5) of 30 samples from 6 heats are shown in Fig. 5. The concentrations of CaO and CaS in steel samples taken before Ca-addition are very small and scattered. Therefore, they are not shown in this figure. It can be seen that the deviations of the PDA/OES results in Zone 5 are larger for most of the samples (particularly for the Al₂O₃ content in inclusions in steel samples L1 and L2 before Ca-addition) in comparison to those found in the other zones. The larger deviations of Zone 5 data can be explained by the larger heterogeneity of the inclusion distribution in the upper region of the thicker part of the steel samples. The heterogeneity may have occurred during the filling and solidification of the steel samples. Moreover, the deviations of the PDA/OES results for the Al₂O₃ and CaS contents in inclusions tend to significantly decrease during ladle treatment and casting. As a result, the average ∆C_i values for Al₂O₃, CaO and CaS contents in inclusions in steel samples taken during casting (samples T4 and T5) are smaller than 20% in most cases.

Fig. 5.

Average deviations of Al₂O₃, CaO and CaS concentrations, ∆C_i, in non-metallic inclusions obtained by the PDA/OES measurements in different sample zones.

Based on the results obtained from the PDA/OES measurements, specimens including Zones 1–3, in which the inclusion characteristics on the surface of a steel sample are most homogeneous, were selected for further 3-D investigations of non-metallic inclusions after electrolytic extraction (see Fig. 4).

3.2. Particle Size Distribution and Volume Fraction of Inclusions Obtained from the PDA/OES and EE Methods

Before a Ca addition, the samples of both steel grades contain mostly regular pure Al₂O₃ inclusions. After a Ca addition, the samples contain mostly globular Ca aluminates with varying contents of CaS (35–65% CaO, 8–38% Al₂O₃ and 0–50% CaS). Typical inclusions observed in all steel samples before and after a Ca addition are shown in Fig. 6.

Fig. 6.

Typical (a) Al₂O₃ and (b) CaO–Al₂O₃ inclusions observed after electrolytic extraction of steel samples taken before and after a Ca addition.

Typical particle size distributions obtained from the PDA/OES and EE methods are shown in Fig. 7 for steel samples taken before (samples L2) and after (samples T4) a Ca addition. It can be seen that the number of inclusions per unit volume, N_V, obtained from the PDA/OES and EE methods agrees satisfactorily well for particles larger than 2.0 μm in the L2 samples and larger than 1.4 μm in the T4 samples. However, the N_V values for small size inclusions determined using the PDA/OES method are significantly underestimated compared to the data obtained from the EE method. Similar results were obtained for all investigated samples for both steel grades.

Fig. 7.

Typical particle size distributions obtained by the EE and PDA/OES methods in steel samples from the A1 trial taken (a) before and (b) after a Ca-addition.

In order to compare the PSD obtained from the PDA/OES and EE methods, a ratio, r_n, between the N_V(PDA/OES) and N_V(EE) values was calculated. The r_n values for all samples taken before (L2 samples) and after (L3 and T4 samples) the Ca addition are presented in Fig. 8. It can be seen that the majority of the data can be separated into three regions: Region I – extremely low r_n values (r_n ~ 0.002–0.03), Region II – r_n ~ 0.5–4, and Region III – somewhat lower r_n values (r_n ~ 0.1–0.3). The best agreement between PDA/OES and EE data (r_n ~ 0.5–2.0) is obtained in the size range 2.0–5.7 μm for the L2 samples and in the size range 1.4–5.7 μm for the L3 and T4 samples. It should also be pointed out that the r_n values calculated for most samples from Steel A are considerably larger than those from Steel B. The significant disagreements between the N_V(PDA/OES) and N_V(EE) values in the Regions I and III can be explained by a consequence of an ablation of more than one small inclusion (< 2 μm) and an ablation of only some parts of a large inclusion (> 5.7 μm) by one spark.

Fig. 8.

Ratio of the numbers of inclusions per unit volume, r_n, in steel samples from the A and B trials taken (a) before and (b) after a Ca addition.

The results for the volume fractions of inclusions, f_V, for different size ranges estimated from the PDA/OES and EE methods (Fig. 9) show similar tendencies as for the number of inclusions: f_V(PDA/OES) << f_V(EE) in Region I, f_V(PDA/OES) ~ f_V(EE) in Region II and f_V(PDA/OES) < f_V(EE) in Region III. The very low f_V(PDA/OES) values in Region I may be explained by the effect of the Al concentrations in steel samples on the detection limit of small size inclusions. Furthermore, a high density of small inclusions in steel samples leads to that most of them are being measured by OES as an elevated bulk level rather than as inclusions (outliers).¹⁾

Fig. 9.

Volume fraction of inclusions, f_V, in steel samples from the A and B trials taken (a) before and (b) after a Ca addition.

The total numbers of particles per unit volume, N_V, volume fractions, f_V, and average size of inclusions, d_V, obtained from the EE and PDA/OES methods are shown for all inclusions and for inclusions larger than 2 μm in different steel samples in Figs. 10, 11 and 12, respectively. It can be seen in Figs. 10(a) and 11(a) that the total N_V(PDA/OES) and f_V(PDA/OES) values for all inclusions obtained from the PDA/OES method are very low in comparison to those from the EE method. However, it should be pointed out that the agreements between the PDA/OES and EE data are significantly better in steel samples taken after a Ca addition for both steel grades. The N_V and f_V values for inclusions larger than 2 μm show a satisfactorily good agreement between both methods for most steel samples taken before a Ca addition and for all samples taken after a Ca addition, as is shown in Figs. 10(b) and 11(b).

Fig. 10.

Total number of particles per unit volume, N_V, determined for (a) all inclusions and for (b) inclusions larger than 2 μm in different steel samples using the EE and PDA/OES methods.

Fig. 11.

Volume fraction, f_V, determined for (a) all inclusions and for (b) inclusions larger than 2 μm in different steel samples using the EE and PDA/OES methods.

Fig. 12.

Average size, d_V, determined for (a) all inclusions and for (b) inclusions larger than 2 μm in different steel samples using the EE and PDA/OES methods.

Although Pande et al.¹¹⁾ reported that the average size of all inclusions determined from the PDA/OES method was only marginally greater than that from the chemical extraction method, the average values of d_V(PDA/OES) and d_V(EE) in most steel samples investigated in this study are significantly different, as shown in Fig. 12(a). The ratio between the average values of d_V(PDA/OES) and d_V(EE) varies from 1.6 to 2.0 for most samples. However, as follows from Fig. 12(b), the average sizes of inclusions in the size range 2–13 μm obtained from the PDA/OES and EE methods agree satisfactorily well for most investigated steel samples. A possible explanation of the significant disagreement between the results obtained by Pande et al.¹¹⁾ and in this study is the possible partial or complete dissolution of some inclusions (especially small size inclusions) in the HCl acid, which was used for the chemical extraction by Pande et al.

3.3. Limitations of the PDA/OES Method in the Evaluation of the Particle Size Distribution

3.3.1. Limitations in the Evaluation of the Inclusion Number

In order to ablate only one inclusion per spark and to ablate the bulk matrix without NMI by the following spark, the maximum theoretical number of inclusions per unit volume, N_V,max, for the given parameters of the PDA/OES measurements should not exceed a value of ~7.8·10⁴ mm⁻³ (= 4000 sparks/(2·2.56·10⁻² mm³)). However, in practice this number is much lower (~10⁴ mm⁻³) in order for the PDA/OES method to give reliable results.¹⁾ If the inclusion density is higher, the PDA/OES method will detect several small inclusions as one outlier peak of a larger inclusion. In addition, this can be seen as an elevated bulk level of the steel matrix. A correlation between the r_n and N_V(EE) values is shown in Fig. 13. The dashed vertical line represents the critical N_V,max value for the PDA/OES measurements. It can be seen that the r_n value tends to significantly decrease with an increased number of inclusions in each size range. Moreover, the values of r_n for inclusions in the size range 0–1.4 μm are very low in all samples because the N_V(EE) values in this size range (~3.5·10⁶–1.2·10⁷ mm⁻³) are about 45 to 150 times larger than the critical N_V,max value. Although the numbers of inclusions in the size range 1.4–2.0 μm are smaller than N_V,max, the r_n values in the L2 samples (before a Ca addition) are also very low. However, the inclusion data in the L3 and T4 samples (after a Ca addition) show a satisfactorily agreement between the obtained N_V(PDA/OES) and N_V(EE) values. This big difference may be explained by the difference in the Al and Ca concentrations in the steel bulk. It is known that high contents of elements which are dissolved in the steel matrix make it difficult to detect the small size inclusions containing these elements.¹⁾ This is due to the larger variations of the intensity signals and noise levels in the measurements of the bulk concentration. The concentration of Al in the investigated steel grades is about 20 to 100 times higher than the concentration of Ca. Therefore, the detection of Al₂O₃ inclusions is more difficult compared to a detection of CaO inclusions. As a result, the N_V(PDA/OES) values for the Al₂O₃ inclusions are significantly more underestimated in that size range compared to those for CaO inclusions.

Fig. 13.

Correlation between the r_n values and number of inclusions per unit volume obtained from the EE method, N_V(EE), for different inclusion size ranges.

3.3.2. Limitations in the Evaluation of the Inclusion Size

A minimum size of detected inclusions, which depends on the concentration of a given element in the steel matrix, can be estimated theoretically. It is known that the measured intensity is proportional to the mass of an element in the ablated volume of a material. In this study, a ratio of intensities for Al, k_(Al), obtained by a spark ablation of metal with one Al₂O₃ particle, I_(Me+P), and without particles, I_(Me), was calculated by using the following equation:   

$$\begin{array}{l}{k}_{\left(\mathrm{A}\mathrm{l}\right)}=\frac{I_{\left(Me+P\right)}}{I_{\left(Me\right)}}~\frac{M{\left(\mathrm{A}\mathrm{l}\right)}_{\left(Me+P\right)}}{M{\left(\mathrm{A}\mathrm{l}\right)}_{\left(Me\right)}}\\ \hspace{1em}\hspace{1em}=1+\frac{V_{\left(P\right)}}{V_{\left(Me\right)}}\left[\frac{\left(\%\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3\right)\cdot 54/102}{{\left[\%\mathrm{A}\mathrm{l}\right]}_{\left(Me\right)}}\cdot \frac{{\rho }_{\left(\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3\right)}}{{\rho }_{\left(Me\right)}}-1\right]\end{array}$$(4)

where M(Al)_(Me+P) and M(Al)_(Me) are the mass of Al in the metal with and without an Al₂O₃ particle being ablated by one spark, respectively. V_(P) and V_(Me) are the volumes of a particle (V_(P) = d_V³·π/6) and metal matrix ablated by one spark (V_(Me) ~ 6.41·10⁻⁶ mm³), respectively. ρ_(Al2O3) and ρ_(Me) are the densities of Al₂O₃ (=0.00397 g/mm³) and the metal matrix (= 0.0078 g/mm³), respectively. (%Al₂O₃) and [%Al]_(Me) are the concentrations of Al₂O₃ in the inclusion and of Al in the metal matrix, respectively. Note, that the k_(Ca) value is evaluated in a similar way.

During practical PDA/OES measurements, some very fine Al₂O₃ inclusions cannot be distinguished as individual particles, due to the limitation of the PDA/OES method as discussed above. Therefore, the value of the bulk intensity for Al, I_(Al)bulk, determined by a spark ablation of the steel matrix can consist of the intensities of Al soluble in the steel and Al from those fine Al₂O₃ inclusions (I_(Al)bulk = I_(Al)sol + I_(Al2O3)fine). In this case, a concentration of Al in the metal matrix (bulk concentration), [%Al]_bulk, determined from the I_(Al)bulk value consists of the soluble Al content in steel, [%Al]_sol, and part of the insoluble Al from such fine Al₂O₃ inclusions, [%Al]_(Al2O3)fine. However, in the following calculations and discussion it was assumed that [%Al]_(Me) = [%Al]_bulk and [%Ca]_(Me) = [%Ca]_bulk.

Figure 14 shows the changes of the k_(Al) values for the PDA/OES measurements calculated using Eq. (4) as a function of the size of Al₂O₃ particles, d_V, as well as on the concentrations of Al in the metal matrix, [%Al]_(Me), and Al₂O₃ in inclusions, (%Al₂O₃). The experimental results showed that good agreements of the N_V values in the metal samples before a Ca addition (L2 samples) were obtained in the size range 2.0–5.7 μm (Region II). The non-metallic inclusions in these samples correspond to pure Al₂O₃ inclusions. Therefore, it was assumed that the minimum size of the Al₂O₃ particles, which can correctly be detected by the PDA/OES method in metal samples of Steel A with the given Al content (=0.040%), is 2.0 μm. As can be seen in Fig. 14(a), for a reliable detection of pure Al₂O₃ inclusions in this case, the value of k_(Al) should be larger than the critical value (k_(Al)crit = 1.44). Therefore, this value of k_(Al)crit = 1.44 is used for an evaluation of the minimum size of Al₂O₃ inclusions, which can reliably be detected by the PDA/OES measurement in steel samples with different concentrations of Al.

Fig. 14.

Relationships obtained by the PDA/OES measurements: (a) between the k_(Al) values and the size of Al₂O₃ particles, d_V, and (b) between the concentration of Al in the steel and minimum size of detected inclusions, d_V-min.

It can be seen in Fig. 14 that the minimum size of the detected Al₂O₃ inclusions, d_V-min, decreases from 2.0 to 1.46 μm with a decreased concentration of Al in steel from 0.040% (Steel A) to 0.015% (Steel B). However, the number of Al₂O₃ inclusions detected in the size range 1.4–2.0 μm in the L2 samples of Steel B is also underestimated because most of inclusions in the range 1.40–1.46 μm could not clearly detected. The minimum size of reliably detected Al outliers as inclusions increases in steel samples taken after Ca additions from 2.0 to 2.5 and 4.3 μm with a decreased concentration of Al₂O₃ in the inclusions from 100% to 50% and 10%, respectively (Fig. 14(b)).

However, the d_V-min for detection of Ca outliers as inclusions is significantly lower (<1.0 μm) in comparison to that for Al due to much lower concentrations of Ca in the steel samples, as is shown in Fig. 15. Therefore, the numbers of complex CaO–Al₂O₃ inclusions detected by the PDA/OES in the size range 1.4–2.0 μm are generally in good agreement with the data obtained from the EE method for all steel samples taken after a Ca addition. Despite that the value of k_(Ca)crit = 1.44 is used in this study (Fig. 14(b)), additional investigations are needed to clarify the k_(Ca)crit value due to significant variations of Ca contents in the observed inclusions.

Fig. 15.

Relationships obtained by the PDA/OES measurements: (a) between the k_(Ca) values and the size of CaO particles, d_V, and (b) between the concentration of Ca in the steel and minimum size of detected inclusions, d_V-min.

According to obtained results, it can be summarized that the number and volume fractions of Al₂O₃, CaO–Al₂O₃ and CaO–Al₂O₃–CaS inclusions in the size range 2.0–5.7 μm can satisfactorily be estimated by using the PDA/OES method. The agreement between results obtained from the PDA/OES and EE methods is better in steel samples taken after a Ca addition. More specifically, the CaO–Al₂O₃ and CaO–Al₂O₃–CaS inclusions can easier be detected using the PDA/OES method in comparison to the pure Al₂O₃ inclusions.

4. Conclusions

The characteristics of non-metallic inclusions (such as number, size and volume fraction) in liquid steel during ladle treatment and casting of industrial heats of two low-alloyed Ca-treated steel grades were evaluated by using the PDA/OES and EE methods. The obtained results can be summarized as follows:

(1) The PDA/OES measurements can be used for a relative estimation of the homogeneity of the distribution of non-metallic inclusions in steel samples. It was found that the characteristics of Al₂O₃, CaO–Al₂O₃ and CaO–Al₂O₃–CaS inclusions on a surface of middle and bottom parts (Zones 1–3) of the lollipop samples taken from the liquid steel are more homogeneous than the corresponding data from the upper parts of the samples.

(2) The best agreement for the numbers of inclusions per unit volume, N_V, and volume fractions, f_V, determined by the PDA/OES and EE methods was obtained for inclusions in the size range 2.0–5.7 and 1.4–5.7 μm for samples taken before (L2) and after (L3 and T4) a Ca addition, respectively.

(3) The total values of N_V, f_V and d_V for inclusions in the size range 2–13 μm obtained by the PDA/OES method agreed satisfactorily well with those from the EE method. Therefore, the PDA/OES method can be used for relative assessment of the non-metallic inclusions during ladle treatment and casting of the studied steel grades.

(4) The ratio of Al intensities obtained by a spark ablation of metal with and without Al₂O₃ particles (k_(Al) = 1.44) was estimated experimentally for the PDA/OES assessment of inclusions in the given steel grades. This k_(Al) value can be used for an estimation of the minimum size of inclusions, which can reliably be detected in steel samples for different concentrations of Al in steel and Al₂O₃ in inclusions.

Acknowledgements

The Swedish Governmental Agency for Innovation Systems (VINNOVA), Jernkontoret – The Swedish Steel Producers’ Association, technical area TO45 and the foundation Axel Hultgren are acknowledged for financial support. Arne Bengtson at Swerea KIMAB is also acknowledged for valuable comments.

References

• 1)  TGS9.04/07, Process Based Steel Cleanliness Investigations and Rapid Metallurgical Screening of Inclusions by Modern PDA Techniques (RAMSCI), Publications Office of the European Union, Luxembourg, (2012), ISBN 978-92-79-22678-6, doi:10.2777/58274, ISSN 1831-9424.
• 2)  TGS2.02/07, Optimisation of Sampling at Liquid Steel State and Correlative Inclusions Assessment of Liquid Steel for the Improvement of High Performance Steel Grades Production Process (SOPLIQS), Publications Office of the European Union, Luxembourg, (2012), ISBN 978-92-79-22144-6, DOI:10.2777/96482, ISSN 1831-9424.
• 3)   R.  Meilland and  L.  Dosdat: Rev. Metall. Cah. Inf. Tech., 99 (2002), No. 4, 373.
• 4)   A.  Bengtson,  R.  Didriksson and  M.  Sedlakova: Proc. 7th Int. Workshop on Progress in Analytical Chemistry in the Steel and Metal Industries, CETAS-2006, CETAS, Luxembourg, (2006), 213.
• 5)   F.  Reinholdsson,  A.  Lind,  R.  Nilsson and  P.  Jönsson: Proc. of 5th Int. Clean Steel Conf., Hungarian Mining and Metallurgical Society and Institute of Materials, The Institute of Materials, Minerals and Mining (IOM3), London, (1997), 96.
• 6)   F.  Reinholdsson,  A.  Lind,  R.  Nilsson,  P.  Sjödin and  P.  Jönsson: ISIJ Int., 37 (1997), No. 6, 637.
• 7)   M. W.  Egger,  A.  Pissenberger and  St.  Aigner: Proc. 8th Int. Workshop on Progress in Analytical Chemistry and Materials Characterisation in the Steel and Metal Industries, CETAS-2011, CETAS, Luxembourg, (2011), 89.
• 8)   F.  Ruby-Meyer,  S.  Evrard,  M.  Vergauwens and  P.  Balland: Proc. of 8th Int. Workshop on Progress in Analytical Chemistry and Materials Characterisation in the Steel and Metal Industries, CETAS-2011, CETAS, Luxembourg, (2011), 294.
• 9)   R.  Didriksson,  T.  Engkvist and  G.  Runnsjö: Proc. 8th Int. Workshop on Progress in Analytical Chemistry and Materials Characterisation in the Steel and Metal Industries, CETAS-2011, CETAS, Luxembourg, (2011), 126.
• 10)   M.  Goransson,  F.  Reinholdsson and  K.  Willman: Iron Steelmaker, 26 (1999), No. 5, 53.
• 11)   M. M.  Pande,  M.  Guo,  R.  Dumarey,  S.  Devisscher and  B.  Blanpain: ISIJ Int., 51 (2011), No. 11, 1778.
• 12)   J.  Ruuska,  S.  Ollila and  K.  Leiviskä: Mater. Sci. Forum, 762 (2013), 649.
• 13)   M.  Burty,  C.  Louis,  P.  Dunand,  P.  Osmont,  F.  Ruby-Meyer,  M.  Nadif,  F.  Penet,  T.  Isono,  E.  Takeuchi and  T.  Toh: Rev. Métall. Cah. Inf. Tech., 97 (2000), No. 6, 775.
• 14)   R.  Inoue,  K.  Kiyokawa,  K.  Tomoda,  S.  Ueda and  T.  Ariyama: Proc. 8th Int. Workshop on Progress in Analytical Chemistry and Materials Characterisation in the Steel and Metal Industries, CETAS-2011, CETAS, Luxembourg, (2011), 118.
• 15)   R.  Inoue,  S.  Ueda,  T.  Ariyama and  H.  Suito: ISIJ Int., 51 (2011), 2050.
• 16)   M.  Göransson and  P.  Jönsson: ISIJ Int., 41 (2001), S42.
• 17)   M-K.  Göransson: Experiences and Evaluation of the Modified OES Technique as a Metallurgical Tool for Inclusion Characterization of Bearing Steel, Licentiate thesis, KTH, Stockholm, (2001).
• 18)   H. R.  Sobel: Optical Emission Spectroscopy, Encyclopedia of Materials Science and Engineering, Vol. 5, Pergamon Press Ltd., Oxford, (1986), 3300.
• 19)   A.V.  Karasev and  H.  Suito: Metall. Mater. Trans. B, 30B (1999), No. 2, 259.
• 20)   H.  Ohta and  H.  Suito: ISIJ Int., 46 (2006), No. 1, 14.
• 21)   Y.  Kanbe,  A.  Karasev,  H.  Todoroki and  P. G.  Jönsson: ISIJ Int., 51 (2011), No. 4, 593.