Application of some modern analytical techniques for investigations of non-metallic inclusions in steel samples

(1)

IN

DEGREE PROJECT

MATERIALS SCIENCE AND ENGINEERING,

SECOND CYCLE, 30 CREDITS

,

STOCKHOLM SWEDEN 2018

Application of some modern

analytical techniques for

investigations of non-metallic

inclusions in steel samples

ZHIYUE WAN

(2)

Abstract

The non-metallic inclusions (NMI) have large influence on steel properties. Therefore, evaluation of inclusion size, number, composition and morphology by using modern analytical techniques are very important for control of steel production and steel quality. Three dimensional analysis method of electrolytic extraction is applied in this work. Metals are dissolved and undissolved inclusions are collected on the film filter. Scanning Electron Microscope (SEM) with Energy Dispersive Spectrometer (EDS) is applied to observe different non-metallic inclusions.

Apart from electrolytic extraction, several other different methods can be used to analyse the inclusions. Each of them has their advantages and drawbacks. A part of this work compares different methods for investigation of inclusions.

To modify the inclusion size, number and morphology, calcium or zirconium can be added in to the steel with certain amount according to the steel grade. This work discussed how inclusion size, number and morphology changed after the modification.

In samples of Heat A, the addition of Zr leads to the appearance of ZrO2 clusters. Large size particles

disappeared while the peak value for the number of inclusions per unit volume didn’t decrease. The electrolytic extraction and fractional gas analysis (FGA) results show some differences on the inclusion compositions.

In samples of Heats B and C, the effect of calcium treatment is investigated. Calcium treatment

modified the spinels with large content of Al2O3 into spherical CaS+CaO-MgO-Al2O3 inclusions. Large

size inclusions were removed after calcium treatment.

Key words: Non-metallic Inclusions, Electrolytic Extraction, Fractional Gas Analysis, Zirconium

(3)

1. Introduction

1.1. Non-metallic inclusions (NMI)

To pursue high quality steel with less impurities, the control of non-metallic inclusions is an important aspect.

The composition, size, morphology and number distribution of the inclusions can affect the mechanical properties of the steel a lot. Three mechanisms are investigated which are related to the negative effects of steel properties: first, the notching elements which can increase the stress field near the non-metallic inclusions; second, the pressurized gas migrates into the inclusions continuously, which creates a stress field nearby; third, the non-metallic phases may produce residual stress because of the difference in thermal expansion [1]. For example, after rolling the elongated MnS inclusions will show anisotropy in mechanical property which makes the steel very brittle. There also have problems during the continuous casting if the inclusions are clogged and clusters are formed in the nozzles. For

example, the Al2O3 inclusion particles are very prone to gather together and form big size clusters

which can cause big problem during casting.

The non-metallic inclusions can be formed during the different steel making processes or from the raw materials. The main non-metallic inclusions in the steel are the sulphides, oxides and nitrides in chemical composition. In generation, removal and modification of the non-metallic inclusions could be different depending on their type and composition. The production processes including melting, refining, casting and rolling may result in different size, volume and chemical composition of non-metallic inclusions. So it is necessary to classify different non-non-metallic inclusions and find their potential influences on the steel properties. Hard, large size inclusions are supposed to be reduced because they could not be deformed with steel matrix when rolling, which may result in stress field around inclusions. Inclusions with lower melting temperature may be easier to deform because they may exist as liquid form, which shows spherical in morphology and much easier to remove from steel. As a result, to improve the fatigue resistance properties of steel, the inclusions which is of low melting point, small size and spherical in morphology is preferred.

1.2. Modification of inclusions

The inclusions in the steel can influence the mechanical properties of the steel as well as causing problems during casting process. The large size and amount of inclusions can also lead to clogging problems in ladle and tundish nozzles. As a result, the modification of the inclusions in the steel before casting is of crucial importance, which is usually achieved by the addition of Ca, rare-earth-metals and Zr. The aim of the modification of the inclusions is mainly to change their composition and decrease their size and number.

(5)

1.2.1. Ca addition on the modification of oxide inclusions

3Ca + (Al2O3) = 3(CaO) + 2Al (1)

3(CaO) + 2Al + 3S = (Al2O3) + 3(CaS) (2)

The main oxide inclusion in the liquid steel is Al2O3, which is in the solid state, and it is largely existed

in the alumina killed steel. At the temperature of liquid steel, Ca has a higher oxygen affinity than Al, thus the reaction (1) will happen. Nevertheless, the reaction products can be in other forms of calcium aluminates depends on the steel composition. According to the phase diagram, the melting point for calcium aluminates is around 1673 K. In this way, the melting point is lower and the solid form of inclusions in the steel can be changed into liquid state of calcium aluminate, which could decrease the possibility of clogging.

However, the amount of Ca reacting with oxide inclusions is difficult to be controlled accurately. First of all, Ca has a relatively low solubility in liquid steel, which results in a low efficiency of Ca addition. Moreover, Ca is supposed to distribute and react with inclusions homogeneously in liquid steel, otherwise the Ca poor or Ca rich aluminates will be formed. If the Ca aluminate of high melting point is formed, the castability of the steel would be decreased, as a result, it’s beneficial to know the transformation of aluminum oxide inclusions during Ca treatment.

In terms of modification of Al2O3 inclusions by calcium treatment, the process depends on S content

and Oxygen content for Ca in the melt, usually the amount of Oxygen in the melt is decided by

aluminum deoxidation, and reaction (2) shows how is the Al2O3 modified.

To be specific, the transformation of inclusions take place in several steps as following,

Al2O3=>CA6=>CA2=>CA=>CAx(liquid), where C stands for CaO and A stands for Al2O3. First, Ca, O

and S diffuse to the solid alumina surface, then a CA6 layer forms on the surface of alumina, and then a

CA2 layer forms on the CA6 surface, followed by a CA layer and finally a layer of molten aluminate

phase CAx is formed [2].

The activities of CaO and Al2O3 are essential for the Al2O3 modification. It is found that Al2O3 inclusion

is much easier to be modified at early treatment stage with moderate S content, in which period

temperature is high (around 1600°C), oxygen activity is high and particle size is small. The reaction

between Al2O3 and CaO will stop when the activity of Al2O3 is too low or when CaS starts to form on the

outside surface [2].

Turkdogan [3] pointed out that the size of the inclusions influences the modification process. It is much more difficult to modify a 10μm inclusion to a 1μm inclusion, and in reality the inclusion size ranges from several microns to 20μm.

(6)

1.2.2. Zr addition

Besides the addition of Ca, the addition of Zr can also have a great effect on the size distribution, number and composition of the inclusions. According to the Gibbs free energy for the formation of

different oxides (Figure 1-1), apart from Ca, Zr also has high affinity with oxygen than Al and ZrO2 is

more stable than Al2O3, which means Zr may be a better de-oxidiser than Al.

ZrO2 has low surface energy and contact angle which means that they are not easy to aggregate and

form clusters as Al2O3. On the other hand, ZrO2 and MnS have similar lattice parameter which means

that the MnS is prone to be precipitated on the pre-formed ZrO2 particles. Thus, the very elongated

and dangerous MnS inclusion can be avoided in the final product. Based on the investigation of X65 pipe steel, Du et al. [4] found that in the Zr-Ti killed steel, the oxide inclusions and sulphide inclusions are more finely dispersed while in the Al-killed steel larger inclusions and elongated sulphides are found.

Li et al. [5] has found that the size of inclusions in Zr-Ti killed steel is much smaller than that in Al-killed steel. There are approximately 65% of the oxide inclusions are smaller than 1 μm in the Zr-Ti killed steel whereas only about 47% of oxide inclusions are in the same range in the steel deoxidized using Al.

Mitja et al. [6] found that the addition of Zr changed the inclusion size distribution in the steel PK942 (X11CrNiMo12), with the majority of the inclusions (>60%) are smaller than 1 μm in diameter. The addition of Zr results in the large reduction in the total non-metallic inclusion surface area (37% reduction). Zr reduced most of the inclusions and formed new, smaller size inclusions so the size and the amount of the inclusions are largely changed.

(7)

1.3. Comparison between 2D and 3D measurement

To analysis of non-metallic inclusions, many methods are used nowadays. Each of them has their advantages and disadvantages. The ideal way to characteristic inclusions should provide accurate and reliable result with less time consuming. If the characteristic inclusions in the liquid steel can be finished before casting, certain correction can be made to ensure the steel quality. Thus, several different methods are combined to investigate the non-metallic inclusions to get the perfect and comprehensive results. However, different methods can also show different results on the same sample. For these reasons, different investigation methods are studied to see their advantages and disadvantages on the investigation of non-metallic inclusions.

1.3.1. 2D analysis method

2D analysis method is a traditional way to observe the non-metallic inclusions, which is always based on a grinded and polished cross section of steel sample in the microscope. It is a good way to determine the location of the inclusions in the metal sample. It was reported that the 2D analysis with the application of INCAFeature program provide good statistics of a large number of inclusions in tool steel [7]. However, the 2D analysis method still has a big problem that is it can not analysis the real size and shape of the inclusions. Take the elongated inclusions as an example (Figure 1-2). The elongated inclusions have to be observed on the deformation direction while they always have random orientations. It is hard to know whether the whole shape and length of an elongated inclusion is shown on the polished cross section. If the deformation direction has large intersection angle with observed polished surface, the result can be very different from the true value [8]. It can also cause some problems when measuring clusters. The cross section may not reveal the real size of the clusters and the size of them can be larger than that measured by 3D methods [9]. Moreover, this is an off-line method so it can not be used for the control of the inclusions during the steel making process.

Figure 1-2 Schematic illustration of different patterns for the cutting of elongated inclusions whose apparent length(Lobs) shows equal (a), shorter (b) and longer (c) detection compared to the

(8)

1.3.2. 3D analysis method

Three types of 3D analysis method of inclusions are studied in this project: electrolytic extraction (EE), optical emission spectrometry with pulse discrimination analysis (OES/PDA) and fractional gas analysis (FGA).

Electrolytic extraction method uses the steel sample as the anode and electrolyzes it in the organic solution. After the electrolysis process start, the steel dissolved into the solution as the form of ion and the undissolved non-metallic inclusions will be collected in the solution. Generally, 0.05g to 0.5g sample can be dissolved and around 0.1g inclusions can be collected [10]. After the cleaning of remaining steel sample, magnetic separation and filtration process, the non-metallic inclusion can be collected on the PC film filter. Then the morphology and size of the inclusions can be observed by using SEM/EDS.

Electrolytic extraction is a more reliable way to observe inclusions in the steel sample from small size to large size. The exact size and morphology can be determined. However, this method also has some drawbacks. Since the inclusions are collected from the liquid solution, it is hard to tell which part of the steel sample does the inclusion come from. Sometimes carbon can be detected from the inclusion, while the carbon is not in the inclusions but from the film filter used for filtration. This happens especially when the inclusion size is less than 6μm, which may interfere the detection of the composition of the inclusions [11]. Another possible is that some brittle inclusions may be broken during the electrolysis or filtration process, which may result in the size change of the inclusions. For the EDS detection, it is hard to detect the element with low atomic weight, which may result in the inaccurate measurement of inclusions composition. This is also a time consuming way to observe the inclusions. It may take about 12 to 24 hours from the sample preparation to get the final result. Moreover, it is an off-line analysis method which can’t be applied during the steel making process. Optical emission spectrometry with pulse discrimination analysis (OES/PDA) is especially used for bulk samples. This method has been explored a lot during the past two decades. As shown in Figure 1-3, an electrical source is used to excite the atoms and ions in the sample to emit element characteristic emission lines. These optical emission lines then pass through the spectrometer. In this step, the element-specific wavelengths are separated by grating and the intensity for each wavelength is measured. The elements in the sample can be detected according to the wavelength peak, and the peak area can show the quantity of each element.

(9)

Figure 1-3 Schematic illustration of the PDA/OES measurement [12]

By using PDA/OES, composition and size distribution of the inclusions can be obtained. The inclusions with the size range of 2 to 20-30μm can be detected [13] and the detected depth is about 10μm from the sample surface. This method has been applied during the production of low alloyed steel and stainless steel to determine the size, type and the composition of the inclusions [14 - 16]. The main advantage of the PDA/OES method is that it is very fast for the inclusions detection. It was reported that it takes only 3-5 minutes to analysis inclusions in steel samples by using PDA/OES in an industrial practice and the result is quite accurate compared with the 2D analysis method [12]. This method can be applied during the steelmaking process easily which means it is very helpful for the detection of steel quality and the optimization of the inclusion modification before the steel comes into the final product. The ability to detect different elements also makes it possible for the PDA/OES to measure a wide range of inclusions with different composition in steel samples.

There are also some drawbacks on this method. First one is that there could be some mistakes when detecting the inclusions with the size less than 1μm. This is especially for the small inclusions containing Ca, which can be detected as the metallic background and can be ignored [17]. The second one is that there are some limits for detecting Si and Mn in inclusions, because they are high soluble elements in the steel.

The fractional gas analysis (FGA) is mainly used for the detection of oxygen and nitrogen inclusions content in the steel. The main principle of this method is based on the different thermodynamic stability of different oxides and nitrides during the heating process. In the fractional gas analysis, the steel sample was put into a graphite crucible and heated at a fixed rate with a certain carrier gas flow.

With the increasing of the temperature, CO/CO2 and N2 will be released from the oxide inclusions and

nitrides inclusions. Since different oxide inclusions and nitrides inclusions have different characteristic

temperature to release CO/CO2 and N2, different compounds can be identified according to the heating

(10)

Figure 1-4 shows a typical curve of the oxygen extraction in the form of CO gas for an Alnico sample. I(t) is the intensity of CO extraction under the certain heating setup. There are two gas evolution peaks

in the figure which correspond to SiO2 (peak 1) and Al2O3 (peak 2) respectively. Different oxide

inclusions and nitride inclusions are identified in this way and the mass fractions of them can be calculated. In the research of GRIGOROVICH et al. [18], fractional gas analysis has been used to detect

the SiO2, Al2O3 and TiN inclusions in the Alnico magnet sample. It was also reported that the FGA can

be used for the steel quality control and prediction of steel properties. It has also been used to detect the oxide inclusions in wheel steels, rail steels, tire cord steels and pipe steels [19, 20].

Figure 1-4 Curve of oxygen (in the form of CO) extraction from Alnico alloys during FGA [18]

Fractional gas analysis is a reliable way to detect oxide and nitrides content in different forms of inclusions. The amount and composition of different oxide and nitride inclusions can be analyzed by this way. However, the disadvantage of this method is that it can not detect all the inclusions with different composition in the steel sample. Secondly, it can only tell the result about the composition and content of the inclusions. The information about morphology, size and number of inclusions can not be obtained from this method.

1.4. Aims of this study

()

To the first group of samples in this project (Group A), which are A1, A2, A3 and A4, the first aim is to see how the addition of Zr changed the composition, morphology, size and number of the inclusions. The second one is to figure out how the inclusion composition, size and number distribution change with time after the addition of Zr. On the other hand, the result of electrolytic extraction method and fractional gas analysis method are compared with each other to see the differences.

To the second group of samples (Heats B and C), which are B2, B4, B5, B6, C4, C5 and C6, the aim is firstly to see how the inclusion composition, morphology, size and number change in different steel-making steps (in the ladle with vacuum treatment, in the tundish during continuous casting and in the final product). The effect of calcium modification of the non-metallic inclusion can also be investigated.

(11)

1.5. Consideration on social ethic aspects

The quality of steel depends a lot on the behavior of NMI. The analysis and control of NMI can help to improve the steel making process. Thus, the energy use and the exhaust emission can be reduced which is environmental friendly.

The main aim of this study is the analysis of NMI, but it can also help to extend the service life of the steel with the consideration of environmental stainability and the economic factors.

(12)

2. Experimental Procedure

2.1. Working Material Samples

Two groups of samples are investigated in this project. The first group contains laboratory sample A1, A2, A3 and A4. As shown in Figure 2-1. 100g of Fe-10%Ni alloy were melted in an induction furnace

under Ar atmosphere at the temperature of 1600°C. Sample A1 was taken 1 minute after the addition

of 0.03wt% Ti. Then 1 minute after the addition of 0.06wt% Zr, A2 was taken. A3 was taken 2 minutes after A2 and A4 was taken 2 minutes after A3. The crucible used in these melting experiments was MgO crucible.

The second group of investigation samples is industrial pipe line steel from two industrial heats (B and C), which contains samples taken from liquid steel during ladle treatment (B2, B4 and C4), continuous casting (B5 and C5) and from final rolled product (B6 and C6). An illustration of the steelmaking process and sampling moments are shown in Figure 2-2. Sample B2 was taken before calcium addition. Sample B4 and C4 were taken from the ladle after the vacuum treatment and the addition of calcium. Sample B5 and C5 were taken from the tundish during the continuous casting in the liquid state. Sample B6 and C6 were taken from final product after hot rolling. The steel grade for the B5, B6 and C5, C6 are same, but there are some different processing technologies applied between them. According to information obtained from company, the quality (corrosion resistence) of steel from Heat C is better than that from Heat B.

Figure 2-1 Schematic illustration of the experimental procedures and sampling of A1, A2, A3 and A4 samples Tundish Final Product B6, C6 B5, C5 B4, C4 Ca addition Hot Rolling Continuous Casting Ladle Treatment

Figure 2-2 Schematic illustration of steelmaking process and sampling moments for industrial Heats B and C

B2

Ti: 0.03% Zr: 0.06%

Operations:

Samples: _A1 _A2 _A3 _A4

Time

1 min 2 min 2 min

(13)

2.2. Electrolytic Extraction

The electrolytic extraction is used to extract the non-metallic inclusions from the metal samples, for 3D investigations of the inclusions on a film filter by using SEM with EDS.

Figure 2-3 shows the apparatus used for the electrolytic extraction and filtration procedures. All the used equipment should be washed carefully with tap water, distilled water and methanol. After polishing and cleaning the metal surface in acetone and benzene, the electrolysis process can start. The electrolyte used for all the samples is 250mL 10%AA (10%acetylacetone + 1%tetramethlammonium chloride + methanol). The metal sample to be electrolyzed acts as the anode and the Pt ring used acts as the cathode. Main electrolytic parameters are listed in Table 2-1.

The choice of electric charge used for the electrolytic extraction is according to the number of the inclusion per unit volume. If the inclusion number per unit volume is relatively high, then the charge of 500 coulombs is enough to dissolve the metal and collect enough inclusions for SEM observation. If the steel is very clean with less inclusion, the higher charge should be used which can be up to 1000 coulombs.

In some cases, the current and voltage during the electrolysis process are not stable. With the electrolysis process going on, some precipitates may cover the surface of the metal which can slow down the electrolysis speed. The current and voltage will decrease with the time. If there are too many precipitates, the electrolysis process may be stopped in advance.

Potentiostat (+) (-) Electrolyte Sample Pt Ring a) Film Filter Vacuum Pump b)

(14)

Table 2-1 Experimental parameters used during the electrolytic extractions

Sample

Electrolytic extraction parameters

Charge

(Coulombs) Current(mA) Voltage(V) Electrolyte

Pore size of PC filter (μm) Extraction sides A1 500 49-71 3.6-6.9 10%AA 0.05 2 sides A2 500 61-69 3.0 10%AA 0.05 2 sides A3 500 59-61 3.0 10%AA 0.05 2 sides A4 500 57-59 3.5-3.6 10%AA 0.05 2 sides B2 800 30-62 3.0-4.5 10%AA 0.4 1 side B4 500 32-65 3.4-4.5 10%AA 0.4 1 side B5 500 40-60 3.4-3.8 10%AA 0.4 1 side B6 1000 32-58 3.5-4.0 10%AA 0.4 1 side C4 800 23-68 2.9-4.5 10%AA 0.4 1 side C5 1000 39-70 3.0-3.7 10%AA 0.4 1 side C6 1000 60-70 4.0 10%AA 0.4 1 side

After the electrolysis process is finished, certain weight of metal is dissolved and the non-metallic inclusions are collected in the solution. A film filter with certain pore size is used for the filtration of the inclusions. All the inclusions larger than the pore size can be collected on the film filter. Magnetic separation is used to remove the undissolved metal pieces from the solution. After filtration the film filter is collected in a sample box washed with methanol. The filtration process should be controlled well to ensure that the all the inclusions are located evenly on the film filter, otherwise the calculation of the inclusion number per unit volume could be inaccurate.

The hight, width and length of the sample are checked before the electrolytic extraction. The weight of the sample should be checked both before and after the electrolytic extraction, which is an important parameter used for the inclusion number calculation.

2.3. Scanning Electron Microscope (SEM) & Energy Dispersive Spectrometer (EDS)

Scanning Electron Microscope (S3700N, Hitachi) is used for the investigation of inclusions’ size, number and morphology. The chemical composition of the inclusions is checked by energy dispersive spectrometer.

(15)

To investigate the inclusions by using the SEM, a part of the film filter is cut and pasted by a conductive tape on an aluminum sample holder (Figure 2-4). The samples should always be kept in the sample box to avoid the dust in the air. Also the sample should not be tilted or flipped otherwise may cause the deviation of the inclusion number distribution.

Back-scattered electrons mode (BSE) is used during scanning and taking SEM pictures. The heavier element shows brighter color in this case. The working distance is around 10.0mm to 11.0mm. The used magnification is dependent on the inclusion size range and number. If the inclusion size range is wide, larger magnification should be used to observe smaller size inclusions and smaller magnification should be used to observe larger size inclusions. The result of them should be combined with each other to get the final particle size distribution. Continuous pictures are taken from the certain area of the film filter to make sure that systemic observation is applied.

When making EDS analysis, even larger magnification (up to 10000) is used. EDS can only detect the outer layer of the inclusions, if the inclusion size is too large. Sometimes the elements from the film filter, such as carbon and oxygen may also be analyzed.

2.4. Measurement of size and number of inclusions and clusters

After obtaining the pictures from the SEM, the inclusions’ and clusters’ size are measured and number are counted to calculate their distribution curve. The size is measure from the SEM pictures by caliper in the unit of millimeter and transformed into the real size according to the magnification.

Figure 2-5 explains how the size is measured for each type of inclusions. The dv for spherical inclusion is its real diameter. For the elongated inclusions such as b) and clusters such as c), the maximum length (l) and perpendicular width (w) are measured and the equivalent size is calculated as deq in the following equation: deq = √4 × 𝑙 × 𝑤 𝜋 (3) Sample Holder Conductive Tap

(16)

a) b) c)

Figure 2-5 Size measurement of a) a spherical inclusion, b)an elongated inclusion and c) a cluster on a film filter

According to the SEM pictures, the number of inclusion per unit volume (Nv) can be calculated by the equation below:

Nv =_{𝑊𝑑𝑖𝑠}𝑁𝑖

𝜌 ×

𝐴𝑜𝑏𝑠 𝐴𝑓𝑖𝑙

Ni is the number of inclusions from SEM pictures in each size category. Wdis is the weight decrease

after the electrolytic extraction. Afil is the area of the film filter used which is 1200 mm2. Aobs is the area

of the film filter which has been observed under the SEM, and ρ is the density of the metal which is

chosen as 0.0078g/mm3_.

Another very important parameter is the aspect ratio (AR), which is used to classify different type of inclusions:

AR = 𝑙

𝑤

Where l is the maximum length and w is the perpendicular width of the inclusions. Higher AR value corresponds to more elongated inclusion.

(4)

(17)

3. Results and discussion

3.1. Characteristics of inclusions in samples of Group A

3.1.1. Classification of inclusions and clusters

Four typical types of inclusion are found in A1, A2, A3 and A4 (Table 3-1).

Table 3-1 Classification of typical non-metallic inclusions in samples A1, A2, A3 and A4

Type Morphology Composition deq(um) AR

ІA

FeO : 29.3% ~ 100% ZrO2 : 0% ~ 38.7% TiO2 : 0% ~ 36% Al2O3 : 0% ~ 10.5% 0.81 ~ 3.36 1

ІB

FeO : 18.3% ~ 94.3% ZrO2 : 0% ~ 77.9% TiO2 : 4% ~ 32.5% Al2O3 : 0% ~ 7.2% 2.48 ~ 9.00 1

ІC

ZrO2 : 0% ~ 98.7% FeO : 0% ~ 96.5% TiO2 : 0% ~ 39.3% Al2O3 : 0.1% ~ 4.3% 2.27~16.77 1.14~2.73

ІІA

ZrO2 : 42.8% ~ 100% TiO2 : 0% ~ 35.9% FeO : 0% ~ 24.7% Al2O3 : 0% ~ 1.4% 1.76~20.14 1.13~3.81

Type ІA is the spherical inclusion with smooth surface. Type ІB also shows the spherical shape but the surface of this type of inclusion is rough compared with ІA. The diameter of type ІB inclusions are

(18)

larger than that of ІA. Type ІC looks like cluster because the particles seem to be connected with each other but each particle have clear boundary actually. The particles are not connected so well with each other so this type of inclusion is defined as the same group with inclusion ІA and ІB rather than the

cluster. The three main components of type ІA, ІB and ІC inclusions are ZrO2, TiO2 and FeO. Type ІІA

is the big size cluster with the main content of ZrO2.

To inclusion type ІB, different composition is found in the center and on the surface of the particles. In the center there is more FeO while the FeO content decreased from the center to the surface. The

possible reason is that in the liquid metal, there is a pre-formed FeO or FeO-TiOx liquid spherical

inclusion can be reduced on a surface by adding Zr. As a result, solid phases with high content of ZrO2

can form on a surface of liquid inclusion.

3.1.2. Number and particle size distribution

Table 3-2 shows some results of the inclusion size distributions in samples A1, A2, A3 and A4. The density of the inclusions distribution is higher in sample A2 and A3, so smaller area is observed. After the addition of Zr, the number of the inclusion increased a lot and then decreased with time going. Sample A2 has the largest number of particles among the four samples. 5 minutes after the addition of Zr, the particle number per unit volume decreased below the value of the sample before Zr addition (A1).

Table 3-2 Summary of inclusion size distribution analysis for sample Group A

Sample Magnification for SEM observation Observed area (mm2₎ Number of NMI observed Size range (μm)

Number of particle per unit volume (mm-3₎

A1 ×1K(11)* 125173 169 0.49-26.05 112976

A2 ×3K(4) 5202 216 0.29-4.52 2570440

A3 ×3K(8) 9826 287 0.29-7.74 1835825

A4 ×1K(11) 133601 195 0.79-13.47 94413

*( ) – Number of SEM pictures used for the measurement

Figure 3-1 shows the number percentage of each type of inclusion in the four samples. In all four samples type ІA inclusions have the highest number percentage value. The amount of type ІB inclusions is much less compared with type ІA. It is obvious that before the addition of Zr in sample A1, up to 89.69% of the inclusion is type ІA and there is no type ІІA cluster. After the addition of Zr, the number percentage of type ІІA clusters increased with the time going. Sample A4 has the highest

amount of type ІІA clusters, which means that ZrO2 tends to aggregate. More and more cluster will be

formed over time. Type ІC inclusions exist in all the four sample, which indicates that the formation of it doesn’t due to the addition of Zr.

(19)

Figure 3-1 Number percentage of different type of inclusion for sample group A

Table 3-3 lists the particle size of each type of inclusion in the four samples to discuss how the size of them is changed.

Table 3-3 Particle size of each type of inclusion and cluster in sample Group A

Sample Type Number

percentage Average diameter ±Ϭ(μm) Size range (μm) A1 ІA 89.69% 1.5±0.9 0.5 – 5.0 ІB 7.62% 3.1±0.9 1.7 – 4.5 ІC 2.69% 16.4±9.0 1.7 – 26.1 A2 ІA 92.13% 0.7±0.5 0.3 – 4.5 ІB 0.93% 2.1±.05 1.6 – 2.6 ІC 5.09% 1.8±0.9 0.9 – 4.4 ІІA 1.85% 1.7±0.8 1.1 – 3.1 A3 ІA 94.06% 0.9±0.4 0.3 – 2.7 ІB 1.75% 1.2±0.2 1.0 – 1.5 ІC 2.10% 4.2±1.9 1.4 – 7.5 ІІA 2.10% 3.4±2.3 1.6 – 7.7 A4 ІA 80.00% 1.8±0.7 0.8 – 5.2 ІB 3.08% 3.2±1.3 1.8 – 4.3 ІC 11.28% 6.7±2.8 3.9 – 13.5 ІІA 5.64% 6.3±2.6 3.0 – 12.0

To the type ІA and type ІB inclusions, the size of them decreased after the addition of Zr and they tend to grow up with the time. The particle size of type ІA and type ІB inclusions increased in the 5 minutes after the addition of Zr, and the size of them in sample A4 is larger than that in sample A1. To the type

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

A1

A2

A3

A4

Number Percentage

ІA

ІB

ІC

ІІA

(20)

ІC inclusion, the particle size of them also decreased after the addition of Zr and kept increasing with time. Different from type ІA and type ІB inclusions, the particle size in sample A4 is smaller than that in sample A1. The type ІC inclusion was modified to smaller size with the addition of Zr. To the type ІІA cluster, they appear after the addition of Zr and the size of them kept increasing from sample A2 to

A4. The main composition of type ІІA cluster is ZrO2 because ZrO2 particles tend to be aggregated

together and form large size clusters.

Figure 3-2 indicates the relation between particle size and number of NMI per unit volume. It is obvious that all the four samples have the largest number of inclusions with particle size smaller than 2μm. For particle size larger than 2μm, the four samples show almost the same tendency while sample A1 has the biggest size particles with the equivalent diameter larger than 8μm. In this sample, this part is mainly composed of the type ІC inclusions. A2 has the highest peak value for the Nv, and A1 has the lowest. After the addition of Zr, the peak value for Nv increased significantly and slightly decreased with time, but the peak value for sample A4 is still higher than that of A1. The Zr modification in this case increases the Nv at peak value. Although the large size particles in A1 are removed, the particle size is not changed significantly to lower value.

Figure 3-2 Particle size distribution with number of NMI per unit volume for sample Group A

3.1.3. Composition of inclusions and clusters

The composition of each type inclusion in the four samples is also investigated. Figure 3-3 shows how

is the composition of each type of inclusion in the four samples distributed in the FeO-ZrO2-TiO2

ternary diagram. 0.00E+00 2.00E+05 4.00E+05 6.00E+05 8.00E+05 1.00E+06 1.20E+06 1.40E+06 0 2 4 6 8 10 12

Nv (mm

-3

₎

deq(μm)

K123 K129 K128 K130 A1 A2 A3 A4

(21)

ІA

A1 A2 A3 A4

ІB

A1 A2 A3 A4

a)

b)

(22)

ІC

A1 A2 A3 A4 A1 A2 A3 A4

ІІA

c)

d)

(23)

In Figure 3-3 a), it is clear that the composition of type ІA inclusions is significantly different in the four samples. In sample A1, before the addition of Zr, the type IA spherical inclusions is mainly

composed of FeO which is larger than 94% in mass percent. After the addition of Zr, ZrO2 and TiO2

content are increased from A2 to A3 and then slightly decreased from A3 to A4. The amount of FeO

remains the highest although the ZrO2 and TiO2 content are changing.

To type IB inclusions, the main content in sample A1 is also FeO and the amount of TiO2 in type IB

inclusions is larger than that of type IA inclusions, which due to the reduction of FeO by adding Ti.

After the addition of Zr, the FeO content decreased from A2 to A4. At the same time, the ZrO2 content

increased up to 80% in mass percent while the TiO2 amount didn’t change a lot. The composition

range of sample A3 and A4 is quite similar, which indicate that the composition of type IB inclusions tends to be stable in 3 minutes after the addition of Zr.

As the composition of type IA and IB inclusions in sample A1, type IC inclusions in A1 are also mainly composed of FeO. However, the composition of type IC inclusions in sample A2 is very scattered as shown in Figure 3-3 c). Some of them are with very high FeO content (up to 63% in mass percent)

while some of them are with very high ZrO2 content (up tp 99% in mass percent). In this case, the

composition result data from the electrolytic extraction method with SEM are very scattered.

To type IIA clusters, the main composition is always ZrO2. With the addition of Zr, ZrO2 particles are

formed and they tend to aggregate to form big size clusters. With the time going, the composition of type IIA cluster tends to be more stable. Almost pure zirconium oxides clusters are observed finally. Since type IA has the high number percentage among the four type of inclusions, it is also interesting

to discuss the how the TiO2, FeO, ZrO2 and Al2O3 content in type IA inclusions change depending on

the equivalent diameter in the four samples. As shown in Figure 3-4 a), type IA inclusions in sample

A4 have the highest TiO2 amount and A1 has the lowest TiO2 amount. To the FeO, type IA inclusions in

sample A1 have the highest FeO content compared with other samples and the other data are very

scattered. The same situation happens in the content of ZrO2. All the data in sample A2, A3 and A4 are

very scattered except that there is no ZrO2 in sample A1. Most of the type IA inclusions have the Al2O3

content below 6% in mass percent and the highest content is FeO, while some of them have higher

TiO2 content and the others have higher ZrO2 content. To inclusion type IA, there is no clear trend of

how the composition changes with the increasing of equivalent diameter and the composition data from the electrolytic extraction method with SEM is quite scattered.

The FGA analysis result of these four samples is obtained from other work and it is interesting to discuss and compare it with the electrolytic extraction result.

(24)

Figure 3-4 The composition change with inclusion size increasing for type ІA inclusions: a)TiO2%,

b)FeO%, c)ZrO2% and d)Al2O3%

3.1.4. Comparison between the result of FGA and electrolytic extraction

As discussed before, the FGA results can not show the morphology or different type of inclusions, but can only tell the oxygen content for each reduction peak at different temperature ranges and the total oxygen content in oxides. Table3-4 shows the FGA results of oxygen content in inclusions of sample A1, A2, A3 and A4. There are three temperature peaks corresponding to the following temperature ranges: 1589~1768K, 1721~1932K and 1890~2121K. According to K.V.Grigorovich’s study, carbon reduction of

FeO is at the temperature of 1450℃ (1723K) and TiO2 is at 1525℃ (1798K) [21]. Thus, in the lowest

temperature range (1589K~1768K) the FeO inclusions are reduced which may also contain some

amount of TiO2. In the second temperature range (1721K~1932K), TiO2 inclusions are mainly reduced

and may also contain some amount of ZrO2. In the highest temperature range (1890K~2121K), ZrO2

inclusions are reduced. Therefore, the total FeO, TiO2 and ZrO2 contents can be calculated according to

the oxygen content in each temperature range. The total inclusion amount (volume fraction) can also be estimated according to the data obtained from FGA analysis.

0 10 20 30 40 0 2 4 6 TiO2% deq(um) K123 K129 K128 K130 A1 A2 A3 A4 0 20 40 60 80 100 0 2 4 6 FeO% deq(um) K123 K129 K128 K130 A1 A2 A3 A4 0 10 20 30 40 50 0 2 4 6 ZrO2% deq(um) K123 K129 K128 K130 A1 A2 A3 A4 0 2 4 6 8 10 12 0 2 4 6 Al2O3% deq(um) K123 K129 K128 K130 A1 A2 A3 A4 a) b) c) d)

(25)

Table 3-4 FGA result of oxygen content in samples of Group A T (K)

%O

A1 A2 A3 A4 1589~1768 0.04216 0.01228 0.01188 0.01452 1721~1932 0 0.02399 0.02989 0.01054 1890~2121 0 0.01285 0.01493 0.00568 SUM 0.04216 0.04912 0.0567 0.03074

To discuss the composition of the inclusions in A1, A2, A3 and A4, two assumptions are made to do the

calculation. First one is that only FeO, TiO2 and ZrO2 are considered, which means that the other small

content oxides such as Al2O3 are ignored. The second one is that in the lowest temperature range only

FeO is reduced, in the second temperature range only TiO2 is reduced and in the highest temperature

range only ZrO2 is reduced. Based on these two assumptions, the mass percent of FeO, TiO2 and ZrO2

in the inclusions are calculated without the consideration of morphology (Figure 3-5 a ).

To compare with the result from FGA, the total FeO, TiO2 and ZrO2 contents in non-metallic inclusions

are also calculated from the electrolytic extraction data according to the following equation:

[𝑤𝑡%]𝑡𝑜𝑡𝑎𝑙= 𝑛𝐼𝐴× [𝑤𝑡%]𝐼𝐴+ 𝑛𝐼𝐵× [𝑤𝑡%]𝐼𝐵+ 𝑛𝐼𝐶× [𝑤𝑡%]𝐼𝐶+ 𝑛𝐼𝐼𝐴× [𝑤𝑡%]𝐼𝐼𝐴 (6)

[wt%] is the average composition of each inclusion type and n is the number percentage of them. The calculation results are illustrated in Figure 3-5 b).

a) b)

Figure 3-5 Total contents of FeO, TiO2 and ZrO2 (wt%) in the inclusions in samples of Group A

obtained from a) FGA and b) Electrolytic extraction

It is obvious that there are several differences between the results obtained from these two methods. The FeO contents always show the highest in the four samples according to the electrolytic extraction

result, while FGA result shows only in sample A1 and A4 contain most FeO and TiO2 content is the

highest in sample A2 and A3. The same point is that FeO content in A4 is lower than than in A1.

However, TiO2 and ZrO2 content result shows much bigger difference between the two methods. %TiO2

0% 20% 40% 60% 80% 100% A1 A2 A3 A4

wt%

FeO TiO2 ZrO2 0% 20% 40% 60% 80% 100% A1 A2 A3 A4

wt%

FeO TiO2 ZrO2

(26)

is always higher than %ZrO2 according to the FGA results while the electrolytic extraction results show

3.2. Characteristics of industrial sample Heats B and C

3.2.1. Classification of inclusions and clusters

Table 3-5 illustrates the typical inclusion types in steel samples from Heats B and C.

Table 3-5 Classification of typical non-metallic inclusions in samples of Heats B and C

Type SEM image Composition (wt%) deq(um) AR Sample

І

CaS 97.1%~100% Al2O3 0%~2.13% FeO 0%~3.0% CaO 0~1.8% 2.5~7.4 1~3.0

B6

ІІ

CaO 9.2~69.3% CaS 9.3%~62.1% Al2O3 2.9%~54.2% MgO 0%~21.8% FeO 0%~15.3% SiO2 0.6%~6.5% 0.9~7.4 1.1~9.4

B4

B5

B6

C4

C5

C6

ІІІ

TiNx 1.2~53.1% CaS 0%~50.0% Al2O3 3.9%~45.1% CaO 3.5~44.9% MgO 0%~18.9% FeO 0%~11.9% SiO2 0.9%~4.7% 1.1~5.0 1~5.0

B4

B5

B6

C5

C6

ІV

TiNx 76.7~100% FeO 0%~16.7% CaS 0%~12.8% Al2O3 0%~4.9% MgO 0%~4.7% CaO 0%~4.6% 0.7~2.9 1.1~1.8

B4

B5

B6

C4

C5

C6

V

Al2O3 74.6% ~ 84.1% MgO 1.9% ~ 18.4% CaO 0% ~ 16.73% CaS 0% ~ 5.5% 1.8~4.9 1~2.4

B2

Five types of typical inclusions are found. Type I inclusions were found only in sample B6, which is

(28)

the excessive addition of calcium during the calcium treatment, and CaS is prone to be formed during the process of steel solidification. Type II inclusions show the spherical shape and contain CaS with

Al2O3, FeO, MgO and CaO. Type IV is titanium nitrides with cubic shape. Type III inclusions are the

combination of Type II and Type IV. Some titanium nitrides were precipitated on surface of the Type III incluisons. Type V inclusions were observed only in sample B2 before the calcium treatment (with

the main composition of Al2O3-MgO). Calcium treatment changed the morphology and composition of

inclusions from Type V spinels to Type II and Type III spherical inclusions.

3.2.2. Number and particle size distribution

Table 3-6 shows the number and size ranges of NMI observed in steel samples of Heats B and C. To compare sample B2 with B4, the inclusion size decreased and the inclusion number also decreased a little after the calcium treatment. From the ladle treatment(B4) to continuous casting(B5), the inclusion size range increased and number decreased. The inclusion number in final product (B6) is higher than that in B5 and the inclusion size is also increasing. In sample C4, C5 and C6, the inclusion number in C6 is the highest but the inclusion size range is smallest.

Table 3-6 Number and size ranges of NMI observed in steel samples of Heats B and C

Sample Magnification for SEM observation Observed area (mm2₎ Number of observed NMI Size range (μm)

Number of particle per unit volume (mm-3₎ B2 ×2K(7)* ×1K(5) ×500(4) 257304 77 0.56 – 7.05 73644 B4 ×3K(214) 270577 432 0.43 – 2.86 73347 B5 ×2K(247) 687744 138 0.09 – 4.34 18782 B6 ×3K(26) ×2K(22) 94783 114 0.56 – 6.62 50099 C4 ×2K(106) 254032 95 0.47 – 3.57 19308 C5 ×2K(106) 295145 143 0.55 – 6.12 18126 C6 ×3K(30) ×2K(18) 88521 146 0.29 – 2.78 64729

(29)

It was reported that the corrosion resistance of steel C is better than that of steel B. This fact can be explained by the differences of inclusion characteristics in steel samples B6 and C6 obtained from the final product (hot rolled plates). It can be seen that the inclusion number in sample C6 is higher than that in B6 and the inclusion size range in C6 is significantly smaller. Percentages of different types of NMI in steel samples of steels B and C are shown in Figure 3-6.

a) b)

Figure 3-6 Number percentage of different type of inclusion in a) B4, B5, B6 and b) C4, C5, C6 samples

In Figure 3-6 a), it is obvious that the number percentage of Type II inclusions decreased from sample B4 (98%) to B6 (36%). The number percentage of Type III inclusions increased from 1% to 28% and Type IV inclusions increased from 2% to 31%. The increase of Type III inclusions may due to that titanium nitrides and sulfides are prone to nucleate on oxides during the steel solidification [22]. The same tendency appears in sample C4 to C6. The difference is that only in sample B6 there is some amount of Type I inclusion (4%) and no Type III inclusion is found in sample C4. The number of Type IV inclusions in C4, C5 and C6 is significantly higher than that in B4, B5 and B6 respectively. Type II and Type III inclusions count for lower in C4, C5 and C6. This may be due to the nitrogen and sulphur contents are different in steels B and C.

The main NMI in these samples are Type II and Type IV inclusions. Each of them has different influence on the steel properties.

CaS inclusions can be harmful to the steel castability because of its high melting point and hardness, which may lead to the erosion of ladle slide and nozzles during casting. To Type II inclusion, the

possible form is that there may be a CaO-Al2O3 oxide core in the center which is wrapped by CaS or

oxide-sulfide duplex shell. For excessive calcium treatment, redundant calcium will react with sulfur 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% B4 B5 B6

Number Percentage

ІV ІІІ ІІ I 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% C4 C5 C6

Number Percentage

ІV ІІІ ІІ I

(30)

and form hard CaS inclusions. The dissolved aluminum and sulfur may also react with the CaO in excessive calcium modification to form oxide-sulfur duplex inclusion. Therefore, a large number of CaS-bearing inclusions will form after excessive calcium treatment. The deformation ability of this type of inclusion depends on the composition and depth of its layer. A pure and thick CaS layer has some protective and cushioning effect to the calcium aluminates. An oxide-sulfide shell with less CaS

content can not prevent the plastic deformation of the whole inclusion. Thus, the modification of Al2O3

inclusion should be controlled in the proper range and the S content should also be controlled to reduce the negative effect of CaS-bearing inclusions [23].

It is reported [24] that primary titanium nitrides inclusions can act as nucleants during solidification which can promote the formation of intragranular ferrite in low-alloy steels. And the formation of intragranular ferrite can improve the toughness of the coarse grained heat affected zone. The addition of titanium nitrides in certain steel can improve the yield strength and decrease the fracture appearance transition temperature. However, C.Bernhard reported that titanium nitrides and other

titanium oxides should be removed by using CaO-Al2O3-SiO2-MgO during the making of stainless steel

[25]. Mirjam Bajt Leban and Robert Tisu’s experiment results show that the titanium nitrides didn’t show much influence on the stress corrosion cracking property of AISI 321 stainless steel [26]. Nevertheless, Meng et al. found that stress corrosion cracking can initiate at titanium nitrides inclusion stringers underneath the surface for alloy 690TT [27]. Tan et al. found that the fatigue crack was observed to initiate at titanium nitrides inclusions on the surface and the titanium nitrides inclusions also enhance fatigue crack propagation in alloy 690 [28].

Figure 3-7 illustrates the total particle size distribution for all types of inclusions. In Figure 3-7 a), it is obvious that B2 sample has more larger size inclusions compared to the B4, B5 and B6 samples. The inclusion equivalent diameter is up to 7.05μm in B2. Then after the calcium treatment in sample B4, the inclusion size decreased but the peak value for Nv increased a lot (from less than 15000 per cubic millimeter to more than 45000 per cubic millimeter). Large size spinels were modified into smaller

size spherical CaS+(CaO-Al2O3-MgO) inclusions and the number increased. Sample B5 has the lowest

inclusion number but the inclusion size range is larger than B4. In the final product (B6 sample), the inclusion number increased and the inclusion size kept increasing from B4 to B6. Calcium treatment modified the large size spinels to Type II and Type III inclusions but the spherical Type II and Type III inclusions tend to grow up in the liquid steel during the continuous casting process. Also, large size Type I inclusions may precipitate during the steel solidification. Although the inclusion size was larger in final product (B6) compared with B4 and B5, it didn’t excess the inclusion size in sample B2 before calcium treatment.

In sample C4 to C6, the tendency seems a little bit different. Same as B4 and B5, C5 has more large size inclusions but the peak value for C4 is larger. But the peak value for Nv increased a lot from C5 to C6, which is different from B5 to B6. This may because that more titanium nitrides in C6 sample are precipitated during the steel solidification process.

(31)

Figure 3-7 Particle size distributions in steel samples a) B2, B4, B5, B6 and b) C4, C5, C6 0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000 0 2 4 6 8 Nv(mm-3₎ deq(um) B2 B4 B5 B6 0 5000 10000 15000 20000 25000 30000 35000 0 2 4 6 8 Nv(mm-3₎ deq(um) C4 C5 C6 a) b)

(32)

3.2.3. Composition of inclusions

As discussed before, sample B2 has 100% of Type V inclusions and the others are mainly composed with Type II and Type IV inclusions. Since Type IV inclusions are mainly titanium nitrides, the composition change of Type IV inclusion will not discussed more. The composition of the inclusions is recalculated for 100% and ternary composition diagrams are plotted.

Figure 3-8 a) and b) illustrate the composition change of spinel in sample B2 before the calcium

treatment to the CaS+(CaO-Al2O3-MgO) Type II inclusion in samples B4, B5 and B6 after the calcium

treatment. Type III inclusions in sample B6 is also shown in Figure 3-8 a) and b) since the number percentage of them is up to 30% (Figure 3-6 a)) .

Figure 3-8 Ternary composition diagrams a) CaS-CaO-Al2O3 for B2 to B6, b) MgO-CaO-Al2O3 for B2 to B6, c) CaS-CaO-Al2O3 for C4 to C6 and d) MgO-CaO-Al2O3 for C4 to C5

Before calcium treatment, the Type V inclusions are aluminum oxides (more than 80%), which are very prone to aggregate and form clusters. Calcium modification changed the composition of them to lower aluminum oxides content spherical inclusions. With the change of composition, the melting

C4 II C5 II C6 II C4 II C5 II C6 II B5 II B2 V B4 II B6 II B6 III B5 II B2 V B4 II B6 II B6 III a) b) c) d)

(33)

steel-making process because they can be easily formed in the liquid steel and flow up into the slag. From Figure 3-8 a) and b), it is obvious that aluminum oxides content decreased dramatically after the calcium treatment. The main contents in B4 Type II, B5 Type II and B6 Type II inclusions are CaO and

CaS. The MgO content is the lowest. From ladle treatment to the final product, CaS and Al2O3 content

in Type II inclusions increased.

The S content in the liquid steel in this case may influence the composition change of inclusion Type II.

Figure 3-9 shows the unreacted core model of evolution CaO-Al2O3 system into CaO-Al2O3-CaS system.

Figure 3-9 Unreacted core model of evolution CaO-Al2O3 system into CaO-Al2O3-CaS system [29]

The outer layer of Type II inclusion is CaO-Al2O3-CaS, which is stable and can not react with [Al] and

[S] in liquid steel. The following reaction continues happening in the reacting layer until the CaO in this layer will not react with [Al] and [S] in the liquid steel any more.

3𝐶𝑎𝑂 + 2[𝐴𝑙] + 3[𝑆] = 3𝐶𝑎𝑆 + 𝐴𝑙2𝑂3

The similar tendency can be seen in samples C4 to C6. The CaS content kept increasing from C4 Type II to C6 Type II inclusions. To compare sample Heat B with sample Heat C, it is obvious that the CaS content in Heat B is higher than that in C. The CaS contents in inclusions of Type II from samples of

Heat C are mostly under 40% in CaS-CaO-Al2O3 system and in Heat B it can be up to 60%. This may

due to the sulphur content is different in this two steel.

(34)

4. Conclusion

To investigate the inclusion morphology, number, size and composition, electrolytic extraction was applied. The investigation of samples of Group A aims at finding out the effect of zirconium modification. The result of electrolytic extraction for samples of Group A is compared with FGA result to see the difference. Electrolytic extraction is also applied for samples of Heats B and C to investigate the effect of calcium treatment. Based on this work, main conclusions can be summarized as follow:

1) Three types of inclusions (Type IA, Type IB and Type IC) are found in samples of Group A, and

ZrO2 clusters (Type IIA) appeared in samples A2, A3 and A4 after the addition of Zr. The

number percentage of ZrO2 clusters increases from sample A2 (1.85%) to A4 (5.64%). The size

range of Type IIA clusters also increases from 1.1-3.1μm in A2 to 3.0-12.0μm in A4.

2) The total number of non-metallic inclusions per unit volume of steel sample (Nv) increased 13 times after the addition of Zr in comparison to the sample A2, and Nv decreased continuously from sample A2 to A4. The peak value for Nv in sample A4 is still higher than that in A1. Large size inclusions disappeared after the Zr modification.

3) The main components of inclusions of Type IA, Type IB and Type IC in samples of Group A are

FeO, ZrO2 and TiO2. Before the addition of Zr in sample A1, FeO content can be up to 100%.

After the addition of Zr, FeO content decreased significantly. ZrO2 and TiO2 content increased.

The composition of Type IC inclusion is very scattered.

4) The composition results of EE and FGA method show some difference. FeO content in samples

A2-A4 is higher from electrolytic extraction results. From another side, TiO2 and ZrO2

contents are higher from FGA results.

5) FGA method can evaluate the volume fractions of oxide inclusions and total content of each oxide component without the information on morphology, number and size. Nevertheless, the electrolytic extraction method can be used for evaluation of the morphology, number, size and composition for different types of inclusions.

6) Spinels (Type V) were found in sample B2 before the calcium treatment. After calcium

treatment, Type V inclusions were modified into spherical CaS+(CaO-Al2O3-MgO) inclusions

(Type II). Composition and morphology of the inclusions are changed during the calcium treatment. Some titanium nitrides (Type IV) were also found in samples of Heats B and C. 7) Type II and Type IV inclusions count for the most in samples B4 to B6 and C4 to C6. The

number percentage of Type II inclusion decreased from B4 (97%) to B6 (36%) and from C4 (88%) to C6 (20%). C6 has more titanium nitrides compared with B6.

(35)

9) After calcium treatment, Al2O3 content in the inclusions decreased. CaO and CaS contents

(36)

5. Further work

To improve the analysis result, following work is suggested.

1. To compare 2D analysis result with 3D analysis result, 2D analysis method could be obtained to see how each type of inclusion is located in each sample. It will be interesting to see how the inclusions distributed in different part of the sample.

2. Methodology analysis and error analysis of electrolytic extraction method could be made to improve the accuracy of this method.

3. The reason for the formation of Type IC inclusions in sample A2, A3 and A4 is not discussed in detail in this work. Not many similar inclusions were found in previous work. Therefore, the mechanism of the formation of this type of inclusion and how will it influence the steel quality may be investigated further deeper.

4. In this work, only electrolytic extraction result of sample Heat B and Heat C is investigated. To compare the result of electrolytic extraction method with FGA method, the FGA result of sample Heat B and Heat C should also be obtained and discussed.

5. To sample Heat C, the analysis result of the inclusion morphology, number distribution and composition before the calcium treatment could be made to compare with sample C4, C5 and C6 to see the effect of calcium treatment.

(37)

6. Acknowledgements

I would like to say thanks to my supervisor Docent Andrey V. Karasev at KTH, who teaches me a lot about the experimental methods and academic knowledge. Without his help l can not finish my experimental work on electrolytic extraction and SEM. His kind and patient guide inspire my interests on the research work. I would also want to say special thanks to Wenli Long, who teaches me on the SEM operation. Finally, I want to thank my parents for always supporting me to do the work I like and my boyfriend for always keeping company with me.

(38)

7. Reference

[1] C. Mapelli, "Non-metallic inclusions and clean steel," La Metallurgia Italiana, pp. 43-52, 2008. [2] P. YE, "Thermodynamics and Kinetics of the Modification of Al2O3 Inclusions," ISIJ

International, pp. 105-108, 1996.

[3] E.T.Turkdogan, "Int. Calcium Treatment Symp," Univ. of Strathclyde, Glasgow, 1998.

[4] K. M. Wu, L. Cheng, Y. Li, O. Isayev, O. Hress and Y. Du, "Effect of Zr–Ti deoxidisation on the hydrogen-induced cracking of X65 pipeline steels," Materials Science and Technology, vol. 32, pp. 728-735, 2016.

[5] X. Wan, W. Lu, A. Shirzadi, O. Isayev, O. Hress, K. Wu and Y. Li, "Effect of Zr-Ti combined deoxidation on the microstructure and mechanical properties of high-strength low-alloy steels,"

Materials Science & Engineering A, pp. 179-187, 2016.

[6] Mitja Koleznik, "DE-OXIDATION OF PK942 STEEL WITH Ti AND Zr," MATERIALI IN

TEHNOLOGIJE/MATERIALS AND TECHNOLOGY, pp. 1031-1036, 2017.

[7] Andrey. K, P.G. Jösson and H. Doostmohammadi, "A Comparison of a Two-Dimensional and a Three-Dimensional Method for Inclusion Determinations in Tool Steel," steel research, pp. 398-406, 2010.

[8] Yuichi Kanbe, Andrey. K, Hidekazu T. and Pär G. Jösson, "Analysis of Largest Sulfide Inclusions in Low Carbon Steel by Using Statistics of Extreme Values," steel research, pp. 313-322, 2011. [9] Yuichi KANBE, Andrey. K, Hidekazu T. and Pär G. Jösson, "Application of Extreme Value

Analysis for Two- and Three-Dimensional Determinations of the Largest Inclusion in Metal Samples," ISIJ International, vol. 51, pp. 593-602, 2011.

[10] Diana Janis, Ryo Inoue, Andrey Karasev and Pär G. Jösson, "Application of Different Extraction Methods for Investigation of Nonmetallic Inclusions and Clusters in Steels and Alloys," Hindawi Publishing Corporation, 2014.

[11] Ryo INOUE, Shigeru UEDA, Tatsuro ARIYAMA and Hideaki SUITO, "Extraction of Nonmetallic Inclusion Particles Containing MgO from Steel," ISIJ International, pp. 2050-2055, 2011. [12] Diana JANIS, Andrey KARASEV and Pär Göran JÖSSON, "Evaluation of Inclusion

Characteristics in Low-Alloyed Steels by Mainly Using PDA/OES Method," ISIJ International, vol. 10, pp. 2173-2181, 2015.

[13] L. Dosdat, "Rapid characterization of inclusionary cleanliness in steels by PDA-OES mapping,"

Metallurgical Research and Technology, vol. 99, pp. 373-382, 2002.

[14] TORTOERTO Giampietro, "Routine qualitative analysis of inclusions in steel," Metallurgy

Analysis, vol. 27, pp. 1-6, 2007.

[15] D. Janis, P.G.Jösson, A.APPELL and J.Janis, "Application of pulse distribution analysis with optical emission spectroscopy (PDA/OES) method during production of duplex stainless steel,"

Ironmaking & Steelmaking, vol. 43, pp. 121-129, 2016.

[16] Mia GÖRANSSON and Pär G. JÖNSSON, "ldeas for Process Control of Inclusion Characteristics During Steelmaking," ISIJ International, vol. 41, pp. 42-46, 2001.

[17] "Process based steel cleanliness investigations and rapid metallurgical screening of inclusions by modern PDA techniques (RAMSCI)," Publications Office of the European Union , 2012.

[18] K. GRIGOROVICH, G.BURKHANOV, S.SHIBAEV and A.DALMATOV, "FRACTIONAL GAS ANALYSIS PROCEDURE FOR DETERMINING THE CONTENTS OF OXIDE AND NITRIDE PHASES IN ALNICO ALLOYS," METALURGIJA, vol. 48, pp. 187-191, 2009.

[19] K. Grigorovich and S. S. Shibaev, "OPTIMIZATION OF THE CLEAN STEELS LADLE

TREATMENT AND NON-METALLIC INCLUSION CONTROL," METAL, vol. 5, pp. 1-8, 2007. [20] K. V. Grigorovich, T. V. Shibaeva and A. M. Arsenkin, "Effect of a PipeSteel Killing Technology on

the Composition and Number of Nonmetallic Inclusions," Russian Metallurgy (Metally), pp. 927-933, 2011.

(39)

[22] S. M. VINCENT DESCOTES, "TEM Characterization of a Titanium Nitride (TiN) Inclusion in a Fe-Ni-Co Maraging Steel," The Minerals, Metals & Materials Society and ASM International, vol. 46A, pp. 2793-2795, 2015.

[23] ZHOUHUA JIANG, YANG LI and GUANG XU, "Formation Mechanism of CaS-Bearing Inclusions and the Rolling Deformation in Al-Killed, Low-Alloy Steel with Ca Treatment,"

METALLURGICAL AND MATERIALS TRANSACTIONS B, vol. 47B, pp. 2411-2420, 2016.

[24] W. Mu, "Microstructure and Inclusion Characteristics in Steels with Ti-oxide and TiN Additions," Stockholm , 2015.

[25] C. Bernhard, "Experimental Study on the Behavior of TiN and Ti2O3 Inclusions in Contact with CaO‐Al2O3‐SiO2‐MgO Slags," Scanning, 2017.

[26] Mirjam Bajt Leban and Robert Tisu, "The effect of TiN inclusions and deformation-induced martensite on the corrosion properties of AISI 321 stainless steel," Engineering Failure Analysis, vol. 33, pp. 430-438, 2013.

[27] Fanjiang Meng, Jianqiu Wang, En-Hou Han and Wei Ke"The role of TiN inclusions in stress corrosion crack initiation for Alloy 690TT in high-temperature and high-pressure water,"

Corrosion Science, vol. 52, pp. 927-932, 2010.

[28] J. Tan, "Role of TiN inclusion on corrosion fatigue behavior of Alloy 690 steam generator tubes in borated and lithiated high temperature water," Corrosion Science, pp. 349-359, 2014.

[29] G. XU, "Formation Mechanism of CaS-Bearing Inclusions and the Rolling Deformation in Al-Killed, Low-Alloy Steel with Ca Treatment," METALLURGICAL AND MATERIALS

Application of some modern analytical techniques for investigations of non-metallic inclusions in steel samples

IN

DEGREE PROJECT

MATERIALS SCIENCE AND ENGINEERING,

SECOND CYCLE, 30 CREDITS

,

STOCKHOLM SWEDEN 2018

Application of some modern

analytical techniques for

investigations of non-metallic

inclusions in steel samples

ZHIYUE WAN

Abstract

Contents

1. Introduction

1.1. Non-metallic inclusions (NMI)

1.2. Modification of inclusions

1.3. Comparison between 2D and 3D measurement

1.4. Aims of this study

1.5. Consideration on social ethic aspects

2. Experimental Procedure

2.1. Working Material Samples

2.2. Electrolytic Extraction

2.3. Scanning Electron Microscope (SEM) & Energy Dispersive Spectrometer (EDS)

2.4. Measurement of size and number of inclusions and clusters

3. Results and discussion

3.1. Characteristics of inclusions in samples of Group A

ІA

ІB

ІC

ІІA

A1

A2

A3

A4

Number Percentage

ІA

ІB

ІC

ІІA

Nv (mm

)

deq(μm)

ІA

ІB

a)

b)

ІC

ІІA

c)

d)

%O

wt%

wt%

3.2. Characteristics of industrial sample Heats B and C

І

B6

ІІ

B4

B5

B6

C4

C5

C6

ІІІ

B4

B5

B6

C5

C6

ІV

B4

B5

B6

C4

C5

C6

V

B2

Number Percentage

₎