Inclusion Characteristics in Steel with CeO₂ Nanoparticle Addition

Abstract

The application of Ce oxides in oxide metallurgy has received extensive attention, but until now, the direct adding of CeO₂ into molten steel to generate Ce oxides has not occurred. In this paper, a mixture of CeO₂ and Si nanoparticles were added into molten steel. The resultant formation of micrometer scale Ce-bearing oxides confirmed its adding validity. This behavior may be interpreted as the reactivity between CeO₂ and [Al], and the improved wettability between CeO₂ and molten steel with the assistance of Si powder. Thus, when the quantity of CeO₂ is kept constant, its added yield should increase when increasing the added quantity of Si. This was verified by the larger percentage of Ce-bearing oxides of the total oxides and the greater average content of Ce in Ce-bearing oxides after normalization. Moreover, compared with the blank sample, statistical results indicated that the oxides in CeO₂-modified samples were refined, and their dispersion homogeneity was enhanced. This comparison indicates the effectiveness of the external adding method in oxide metallurgy.

1. Introduction

Since 1990, the concept of oxide metallurgy has attracted much attention [1,2]. This technology utilizes specific kinds of inclusions to induce nucleation of intragranular ferrite (IGF) during austenite–ferrite transformation.

To generate certain kinds of non-metallic inclusions, the internal precipitation method (IPM) [3,4] and the external adding method (EAM) [5] have been developed. The former means that inclusions form either during de-oxidation or solidification processes, and needs precise process control technology. The latter means that pre-prepared particles are directly added into steel, and its key problem is how to make these particles stably exist and evenly distribute in steel.

Until now, different kinds of oxides, such as TiO₂ [6,7,8], MgTiO₃ [9], and ZrO₂ [10] have been successfully introduced into molten steel, and the resultant complex oxides can also induce the nucleation of IGF, which is similar in appearance to the inclusion originating through IPM. In our previous papers [6,7], the effective adding of TiO₂ is attributed to wettability and reactivity. First, the contact angle between TiO₂ and the pure Fe system at 1813 K is below 90°, which suggests good wettability between them. Second, TiO₂ can react with other elements in molten steel. Thus, after adding, Ti-bearing inclusions, in which the elements of Si, Mn, and Al co-existed, were formed.

Recently, the application of Ce oxides in oxide metallurgy has aroused attention. Not only Ce adding alone [11,12,13,14], but also complex treatments, such as Ce–Mg [15,16], Ce–Ca [17], and Ti–REM–Zr (REM is cerium and lanthanum) [18], have been studied. However, the direct adding of CeO₂ into molten steel is still absent to our knowledge. In our exploratory experiment, the CeO₂ nanoparticle has been added into molten steel. However, its floating has been observed, which means poor wettability between CeO₂ and molten steel. Moreover, no Ce-bearing oxides have been found in a resultant sample after analyzing more than 1000 oxides. When adding the mixture of CeO₂ nanoparticle and Fe powders, the same conclusion has also been drawn.

Thus, in this paper, CeO₂ nanoparticles were mixed with Si powders and directly added into molten steel. Its effect on inclusion characteristics has been studied.

2. Materials and Methods

2.1. Pre-Dispersion of CeO₂ Nanoparticles

The raw materials are CeO₂ and Si nanoparticles, and their average particle sizes are both 50 nm. They were mixed with the mass ratio of 4:1, and the appropriate amount of ethanol was also added. Then, the mixture was pre-dispersed through a planetary ball mill equipped with a liquid nitrogen cooling system (QM-4L, Nanjing Chishun Science and Technology Development Co., Ltd., Nanjing, China). The total rotation time was 6 h with the rotation speed of 400 rpm at a temperature lower than −10 °C, which aims to avoid the high temperature resulting from particle collisions. The final obtained mixture was put into a vacuum drying oven at 70 °C to obtain the dry powder.

2.2. Experimental Procedure

The melting experiments were carried out in a high-heat tube-type resistance furnace. The experimental temperature was 1873 K, and Ar gas (purity > 99.999%) was used to maintain the inert atmosphere. In each experiment, about 1500 g of electrical iron was used and put into an aluminum crucible.

Firstly, when melting was complete, an appropriate amount of Fe₂O₃ was added to reduce the acid-soluble aluminum (Al_s) content from 0.28% to below 0.1%, since previous studies [19] have indicated that low Al content may help to decrease the content of Al₂O₃ in oxide.

Secondly, Mn powder (99.99%, wt. %) and a pre-prepared mixture (CeO₂ 20 g; Si 5 g) were added into molten steel in sequence. Each raw material was added from the top through a gravity drop at the same time. The time interval between these two feedings was 10 min. It can be seen that the mixture dissolved into the molten steel.

Thirdly, after stabilizing for about 10 min at 1873 K, the crucible was taken out. After solidification, the sample was quenched in water, which is named as the 2# sample. It should be mentioned that the sample surface was smooth, and no laminations were observed.

For comparison, the blank sample was prepared through the addition of Fe₂O₃, Mn powder, and Si nanoparticle in the same sequence and the same amount as the 2# sample. This sample was named 1#. The steel specimens were prepared by cutting off a small cube at the bottom of each sample.

2.3. Characterization Methods

The morphology and composition of inclusions were characterized through SEM ( SEM: JSM-6510LV, JEOL, Tokyo, Japan) and EDS ( EDS: INCA Feature X-Max 20, OXFORD Instruments, Oxford, UK). The EDS data processing is carried out according to Wang Xinhua et al. [20]. Firstly, iron was excluded to avoid signals from the steel matrix, and oxygen was removed due to insufficient accuracy. Then, the content of remaining elements was normalized to 100%, and expressed as mass percentage.

Moreover, INCA Feature software (OXFORD Instruments, Oxford, UK) was applied to automatically find and analyze inclusions in the pre-selected area. This software works based on the contrast imaging technique, which can distinguish between inclusion and Fe matrix, and collect characteristic parameters of each inclusion, such as coordinates, equivalent radius, and composition. The equivalent diameter was estimated as the diameter of a spherical particle, which has the same area as the projection image of the measured inclusion.

The samples were sent to the NCS Testing Technology Co., Ltd. (China National Analysis Center for Iron and Steel, Beijing, China) for composition analysis, and the results are shown in

Table 1. Composition of samples (mass percentage/%).

No.	C	Si	Mn	P	S	Sol.Al	Ce
1#	0.045	0.21	1.39	0.010	0.0058	0.082	/
2#	0.050	0.18	1.54	0.007	0.0046	0.074	0.0003
3#	0.100	0.44	1.13	0.012	0.0066	0.018	0.0067

.

3. Results and Discussion

The typical oxides for the 1# and 2# samples are presented in Figure 1 and Figure 2, respectively. The corresponding results of the EDS point analysis (mass percentage, and elements of Fe and O are excluded) are presented in

Table 2. EDS point analysis of typical oxides shown in the figures (mass percentage).

Figure	Position	Al	Mn	S	Ce
1	Point 1	100	/	/	/
	Point 2	87.5	9.0	3.5	/
2	Point 1	16.2	/	/	83.8
	Point 2	52.4	4.2	7.0	36.4
3	Point 1	9.8	/	/	90.2
	Point 2	38.9	7.6	11.9	41.6

.

From the view of composition, it can be seen that CeO₂ addition led to the oxide evolution from Al–O to Ce–Al–O, which demonstrates its validity. To confirm this conclusion, more than 1000 oxides were analyzed for the 2# sample, and 51.9% of them (number percentage) could be classified as Ce-bearing oxides. Moreover, in these Ce-bearing oxides, the average mass percentage of Ce and Al are 34.7% and 47.8% after normalization, respectively. The remaining are Mn and S, which is attributed to MnS segregation. It is well known that during the solidification process, MnS prefers to heterogeneously nucleate on pre-existing oxides and results in accumulation of sulfur and Mn around them. This is consistent with its appearance in Figure 2.

From the point of view of size, the typical oxide in Figure 2 is larger than 2 μm. It should be pointed out that submicron inclusions were also found in the 2# sample and nanometer-scale inclusions could not be observed owing to the accuracy limits of SEM. However, many micrometer-scale inclusions still indicate the agglomeration of CeO₂ nanoparticles. In fact, similar results have been reported by Mu et al. [21]. They added TiO₂ powder, with a particle size range of 0.139–0.854 μm, into molten steel. Then, the size range of resultant Ti-bearing inclusion increased to between 0.215 and 4.802 μm. This may be due to the higher surface free energy per unit volume originated from the nano-size effect. Thus, nanoparticles tend to agglomerate to lower their Gibbs free energy. Specific to our experiment, CeO₂ nanoparticles became micrometer scale Ce-bearing oxides.

According to our previous studies [6,7], the adding validity of TiO₂ nanoparticle is due to the wettability and reactivity between TiO₂ and molten steel. For CeO₂, a similar explanation may also be effective.

For reactivity, the thermodynamic calculation of CeO₂ in molten steel is carried out based on following equations [22]:

[Ce] + 2[O] = CeO₂(s)   ΔG^θ₁ = −852720 + 249.96T  J•mol⁻¹

(1)

[Ce] + [Al] + 3[O] = CeAlO₃(s)   ΔG^θ₂ = −1366460 + 364.3T  J•mol⁻¹

(2)

Thus, Equation (3) can be derived follows:

2[Al] + 3CeO₂(s) = 2CeAlO₃(s) + [Ce]   ΔG^θ₃ = −174760−21.28T  J•mol⁻¹

(3)

$$\Delta {\mathrm{G}}_3=\Delta {\mathrm{G}}_3^{\mathsf{\theta }}+RTln\frac{a_{Ce}·a_{CeAl{O}_3}^2}{a_{Al}^2·a_{Ce{O}_2}^3}$$

(4)

where a is the activity; $\Delta {\mathrm{G}}_3$ and $\Delta {\mathrm{G}}_3^{\mathsf{\theta }}$ are the Gibbs free energy change and standard Gibbs free energy change for Equation (3), respectively. The activities of CeAlO₃(s) and CeO₂(s) are assumed as unity, and the activities of [Ce] and [Al] can be described as follows:

$$a_i=f_i·{\left[\%i\right]}_{}$$

(5)

where [%i] is the mass percentage of element i; f_i is activity coefficient of element i, which can be calculated as follows:

$$\log f_i=\sum e_i^j\left[\%j\right]$$

(6)

where $e_i^j$ is activity interaction coefficient. The activity interaction coefficients in the molten steel at 1873 K are shown in

Table 3. Activity interaction coefficients in steel at 1873 K.

${\mathit{e}}_{\mathit{i}}^{\mathit{j}}$	C	Si	Mn	P	S	Al	Ce
Al	0.091	0.0056	/	/	0.03	0.045	−0.43
Ce	−0.077	/	/	1.77	−8.36	−2.25	−0.003

[16,22].

Based on the upper formulae and the composition of sample 2#, the ΔG₃ is −267400 J•mol⁻¹, which is negative at 1873 K. Thus, the reaction between CeO₂(s) and [Al] in molten steel may occur, which leads to the formation of the Ce–Al–O inclusion.

Thus, the following question should be answered: if the upper deduction is true, why is no Ce–Al–O inclusion found in our exploratory experiments (when adding the mixture of CeO₂ nanoparticles and Fe powders, and adding the CeO₂ nanoparticles alone)? This may be explained by the wettability between molten steel and CeO₂. Although this data is still absent to our knowledge, the floating of CeO₂ in our exploratory experiments means poor wetting. Nonetheless, the mixture of CeO₂ and Si nanoparticles can be dissolved in molten steel, which hints that Si addition helps to improve the wettability. In fact, similar results have also been reported that the increasing Si content enhances the wettability between Al₂O₃ and molten steel [23]. When the Si content increases from 0% to 5%, the contact angle between Al₂O₃ and molten iron at 1773 K decreases from 140° to 124° [23]. Thus, when the admixture of CeO₂ and Si is added into molten steel, the Si content in the melt around CeO₂ significantly increases, which improves the wettability between them and makes CeO₂ steadily react with [Al].

If the upper theory is correct, more Si powder in the mixture will enhance wettability. Then, the higher yield of CeO₂ addition is expected. Therefore, another mixture of CeO₂ and Si nanoparticles was pre-prepared and added into molten steel, which was named as the 3# sample. The weight of the CeO₂ was 20 g, which was the same as that in the 2# sample. The weight of the Si powder was doubled from 5 g to 10 g. Other experimental conditions, including the weight of each raw material and operation procedures, were kept consistent.

From Table 1, it can be seen that the Ce content of the 3# sample is much higher than that of 2#, though the added weight of CeO₂ was the same for these two samples. The typical inclusions for the 3# sample are present in Figure 3. Moreover, the results of the Feature software indicated that 87.4% of the oxides (number percentage) can be classified as Ce-bearing oxides and, for them, the average content of Ce is 50% (mass percentage, and the elements of Fe and O are excluded). These values are much higher than is found in the 2# sample (51.9% and 34.7%, respectively). All these values were consistent with the experimental design and confirmed that Si powder helps to improve the efficiency of CeO₂ addition.

Based on the discussions above, the adding validity of CeO₂ can be attributed to its reactivity and wettability. The former leads to the formation of Ce–Al complex oxides due to the reaction between CeO₂ and [Al], and the latter makes this reaction feasible with the assistance of Si powder.

For clean steel manufacturing technology, the requirements for oxides are fine and disperse, which are discussed as follows.

Firstly, the size distribution of oxides for all samples is presented in Figure 4, and each average oxide size is shown in

Table 4. Characteristic parameters of oxides for samples.

No.	Average Size (μm)	Degree of Homogeneity
1#	3.4	1.03
2#	3.6	1.24
3#	1.8	1.46

. It can be seen that by keeping the adding weight of CeO₂ and other experiment conditions constant, increasing the Si quantity leads to a decrease in oxide average size. This refinement is due to two reasons. On the one hand, the dissolution of Si in the melt around CeO₂ can improve the wettability between CeO₂ and molten steel and make its addition more effective, which means higher Ce content. In fact, Song et al. has indicated that the inclusions can be obviously refined after being treated by Ce [12]. On the other hand, more Si powder means better dispersion conditions of CeO₂ nanoparticles. Thus, the nucleation sites for oxide formation are increased, which leads to refinement.

Secondly, the homogeneity of oxide dispersion is studied according to the Wang et al. method through the following two steps [24]. During Step 1, based on the statistical results of the INCA Feature software, a minimum inter-surface distance of a certain inclusion, which is defined as the inter-surface distance between this inclusion and its nearest inclusion, was calculated as follows:

$$D_{m_i}=\mathrm{MIN}\left({\mathrm{D}}_{\mathrm{i}1},{ \mathrm{D}}_{\mathrm{i}2}\cdots {\mathrm{D}}_{\mathrm{ik}}\right)$$

(7)

$${\mathrm{D}}_{\mathrm{ik}}=\sqrt{{\left({\mathrm{X}}_{\mathrm{k}}-{\mathrm{X}}_{\mathrm{i}}\right)}^2+{\left({\mathrm{Y}}_{\mathrm{k}}-{\mathrm{Y}}_{\mathrm{i}}\right)}^2}-{\mathrm{r}}_{\mathrm{k}}-{\mathrm{r}}_{\mathrm{i}}$$

(8)

where X_i and Y_i are the central coordinates of inclusions in the cross section, r_i is the equivalent radius, and D_ik is the inter-surface distance between i and k particles.

For Step 2, the degree of homogeneity in inclusion dispersion H is defined as the reciprocal of the relative standard deviation of D_mi, as follows:

$$H=\frac{\overline{D_{\mathrm{A}}}}{\sqrt{\frac{{{\displaystyle \sum }}_{i=1}^n{\left(D_{mi}-\overline{D_A}\right)}^2}{n-1}}}$$

(9)

where $\overline{D_{\mathrm{A}}}$ is the arithmetic mean value of D_mi, and n is the measured oxide number for each sample.

The H values for all samples are shown in Table 4. It can be seen that CeO₂ treatment helps to increase degrees of homogeneity, and the increased addition of Si enhances this effect. Wang et al. [24] indicated that the degree of homogeneity of inclusion distribution decreases with the increasing attractive force, i.e., the inclusions disperse uniformly in the steel when the attractive force between inclusions is small. For Al₂O₃–MgO inclusion, this attractive force decreases with the decreasing Al₂O₃ content [24]. For our samples, the average content of Al₂O₃ in oxides decreases from 1# to 3#. This may lead to a decrease in attractive force between inclusions, which can explain the increase in the H value and the decrease in average oxide size.

4. Conclusions

The mixture of CeO₂ and Si nanoparticles was added into molten steel to investigate its effect on inclusion characteristics. With the assistance of Si powder, the wettability between CeO₂ and molten steel was improved, which promotes the reactivity between CeO₂ and [Al]. This led to the formation of micrometer scale Ce-bearing oxides. Thus, when the quantity of CeO₂ is kept constant, an increased additive amount of Si helps to improve adding efficiency. This was confirmed by the higher content of Ce in steel, the higher percentage of Ce-bearing oxides in total oxide, and the higher average content of Ce in Ce-bearing oxides after normalization. Moreover, based on statistical analysis, the oxide refinement and uniform dispersion have been confirmed in CeO₂-modifed samples. These results demonstrate the feasibility and advantages of the external adding method in oxide metallurgy.