Ironmaking
Characterization of Ti(C,N) Superstructure Derived from Hot Metal
2021 Volume 61 Issue 1 Pages 138-145
Details
2021 Volume 61 Issue 1 Pages 138-145
Ti(C,N) in BF hearth plays an important role in protecting the BF lining and prolonging the life of BF. The characterization and the crystallization process of the Ti(C,N) obtained from one dead BF were clarified in the paper. It is found that the precipitated Ti(C,N) exhibits annual-ring shape features and different colors. The annual-ring shape topography is caused by the change of the C/N ratio. The precipitated Ti(C,N) is composed of one or more large grains. Ti(C,N) presents as a superstructure or mesocrystal and evolves from Ti(C,N) nanoparticles, which are self-assembled nanomaterials with highly ordered structures. Ti(C,N) deposit forms continuously with layer-by-layer grain growth because the mesocrystal has a large specific surface area. The phase interface of Ti(C,N)-Fe presents as a shape of jagged step, and the phase interface is inlaid by many random mesocrystal structures. The Ti(C,N) deposit derived from hot metal is accompanied by layer-by-layer mesocrystal structures. The results of phase transition and morphology of Ti(C,N) provide guidance for the regulation of Ti(C,N) behavior in BF hearth.
Ti(C,N), which is often referred to as relics in blast furnace (BF) hearth, is only available after the end of the BF generation, which generally takes more than 10 years.1,2,3,4) In the actual production process of the BF, Ti(C,N) plays an important role in protecting the BF lining and prolonging the life of BF,5,6,7,8) which has been the most common maintenance method at the end of the blast furnace campaign. The formation process of Ti(C,N) is mainly described as follows.9,10) At the end period of the furnace, materials containing TiO2 are added from the top of the furnace. TiO2 is reduced under high-temperature conditions and the generated Ti dissolves into the molten iron. The dissolved [Ti] reacts with [C] and [N] contained in the molten iron. Ti(C,N) is deposited at the place where is under lower temperature conditions near the sidewall. At present, the research on Ti(C,N) is mainly focused on the characterization of microscopic morphology and the properties of the TiO2-containing slag.11,12,13,14,15,16) However, the three-dimensional morphology of Ti(C,N) and further research have not been reported yet. Understanding the three-dimensional microstructure of Ti(C,N) and the crystallization behavior of Ti(C,N) are of great helpful and guidance to the formation mechanism and regulation of Ti(C,N).
In this paper, the Ti(C,N) was obtained from one dead BF on a large scale. The formation process of the Ti(C,N) mesocrystals is then studied in a systematic manner by investigating the evolution of particle morphology and compositions at various stages. In particular, the size, phase transition, and morphology of Ti(C,N) are investigated in detail, which helps explain the correlation between the structure and behavior of the Ti(C,N) and provide guidance for the regulation of Ti(C,N) behavior in BF hearth.
The effective inner volume of the dissected BF is 1050 m3. Carbon brick with high thermal conductivity is mainly used at the hearth sidewall. The sidewall of the hearth area is about 1.2 m thick. At the furnace bottom, two layers of the integrated ceramic cup were built on top of the three layers of the large carbon block. The BF was blown in October 2006 and it was shut down in October 2017 with 11 years’ lifetime. In the last time of the BF, the heat flux intensity of the hearth sidewall increased suddenly and the BF was emergency shutdown with completely preservation of burden materials, cohesive zone, deadman, and protective layer.17,18) The protective layer named BF relics which formed on the hot surface of the carbon brick in the hearth was taken in the sampling. The thickness of the protective layer is about 100 mm. The photograph of the dissection investigation area and the obtained sample was shown in Fig. 1. The sample was taken from the sidewall of the hearth and it was close to the carbon brick, showing a purple gold color, and a large particle gold phase can be seen in the sedimentary phase.
Photograph of the dissection investigation area and the obtained sample (BF relics). (Online version in color.)
The sample taken from the site was cut into 10 mm × 10 mm × 5 mm by wire cutting, and the microscopic morphology of each phase at room temperature are observed by Scanning Electron Microscope combined with Energy Dispersive Spectrometer (SEM-EDS). To determine the phase composition of the deposited layer, the treated sample is subjected to X-ray Diffraction (XRD) analysis. The elemental distribution of the sample is based on Electron Probe Micro-analyzer (EPMA). In the experiment, the equipment of the EPMA-1720 series of Shimadzu Corporation is used to carry out elemental analysis.
The characterization of the sample is also conducted by polarizing microscopy. The obtained sample is cut into 7 mm × 7 mm × 3 mm by wire cutting and then polished to prepare the metallographic sample. The microstructure of the samples is observed using a Leica 4500D oil-immersed polarized microscope, the microscope provides natural color reproduction. The sample of the deposited layer is also subjected to the ion-polished sample surface to remove the stress layer, and the sample orientation map is drawn by Electron Backscattered Diffraction (EBSD) to obtain grain size and acquire orientation information in the titanium deposit layer.
In addition, the acid-soluble method is used to separate the hot metal and extract the precipitated phase in the deposit. The secondary electron image of SEM is also used to characterize the 3D morphology of the deposit. The acid-solution method is to put the sample into a beaker with a certain amount of hydrochloric acid (1:1), and an electromagnetic stirring heating device is used to accelerate the reaction of the sample with hydrochloric acid. After the sample is partially or completely dissolved, the sample is taken out from the beaker and cleaned by an ultrasonic cleaner with alcohol as well as dried at last. The 3D topography of these precipitates is studied by field emission scanning electron microscopy (FESEM).
The microscopic appearance of the sample is shown in Fig. 2. A large amount of Ti(C,N) like a tree branch and graphite phase are distributed in the iron matrix (Fig. 2(a)). It is similar to the previously published results.19) In Fig. 2(a), the gray area is Ti(C,N),the white area is iron and the black part is carbon. The percentage of different phases is calculated by Photoshop, as shown in Fig. 2(b). The yellow area corresponds to Ti(C,N), the green area corresponds to iron, and the blue area corresponds to carbon. Ti(C,N), carbon and Fe cover the total area is 33:48:19, respectively. Figure 2(c) is the view of the partial enlarged area. It reveals that Ti(C,N) is composed of a large number of large particles and is accompanied by iron. As can be seen from Fig. 2(d), the interface between Ti(C,N) and iron is clearly visible.
Microscopic appearance of the sample. (Online version in color.)
The optical microscopic morphology of the sample is shown in Fig. 3. Ti(C,N) in Fig. 3(a) exhibits a regular annual ring-like morphology with clear lines and appears in a variety of colors, especially for the distribution of purple and yellow stripes. Each layer has a thickness of approximately 1–10 μm. Figure 3(b) shows the carbon in the sample, which indicates that the carbon in the sample is graphite. The interface between the Ti(C,N) and graphite has clear boundaries in Fig. 3(c), and the interface between the Ti(C,N) and iron has also clear boundaries.
Optical microscopic morphology of the sample: (a) OM graph of the Ti(C,N) with regular annual ring-like morphology. (b) Graphite; (c) Interface of Ti(C,N) and graphite; (d) Interface of Ti(C,N) and Fe. (Online version in color.)
EPMA has better sensitivity for the determination of the element content of the titanium precipitate in the sample. Under the condition of 750 times, a phase-boundary regular Ti(C,N) precipitate is selected from the sample, and three elements of N, Fe, and Ti were selected for map scanning analysis. Figure 4 shows the backscattered electron image of the selected area and the map distribution of each element.
EPMA mapping results of the sample. (a) SEM micrographs; (b) EDS mapping of element Ti; (c) EDS mapping of element N; (d) EDS mapping of element Fe. (Online version in color.)
The backscattered electron image shows that the region is mainly divided into three phases, Fe is mostly distributed in the left part and Ti(C,N) appeared in the right part. The black area in Fig. 4(a) is graphite. Nitrogen shows the obvious nonuniform distribution in the Ti(C,N) precipitation, and shows an annual-ring pattern parallel to the phase interface of Ti(C,N)-Fe in the range of about 30 μm from the phase boundary. The phenomenon indicates that there is a segregation of Nitrogen in the Ti(C,N) precipitation. Moreover, the segregation of elements is likely to be the real cause of the formation of Ti(C,N) under the polarizing microscope. Combined with the optical microscopic morphology of the sample in Fig. 3(a), the purple stripes are mostly TiN and yellow stripes are mostly TiC.
The results of line scanning analysis at a distance of 160 μm at the interface are shown in Fig. 5. It is found that Ti has a significant concentration gradient at the phase interface of Ti(C,N)-Fe, while the mass fractions of the C and N in the Ti(C,N) precipitation show up and down oscillation. Because the C element measured by EPMA is not accurate, the mass fraction is only characterized by the trend of change. The position marked by the dashed line is obvious. When the C element is at the peak position, the N element is located at the trough position. At these locations, the ratio of C/N is also at the peak or trough position. The EPMA results show that the C/N ratio of Ti(C,N) is not constant in a precipitated phase. Ti(C,N) exhibits a variety of colors mainly due to the changes in the C/N atomic ratio.
EPMA lining results of the sample. (Online version in color.)
Figure 6 shows the X-ray diffraction patterns of the sample, the result shows that the sample mainly consists of three phases. The main components of the deposit are iron, Ti(C,N) and graphite. Ti(C,N) phase exhibits a strong diffraction peak, wherein the ratio of C/N is 3/7. The result is different from the results of EPMA, which is probably due to the highest proportion of TiC0.3N0.7 in the deposit. In addition to TiC0.3N0.7, there is also a certain amount of graphite phase. The presence of graphite verified the results seen under optical microscopy conditions.
XRD results of the sample.
Previously, the 2D morphology of Ti(C,N) grains was characterized, and the cross-section morphology was observed. Moreover, the secondary electron image of the particles detached from the sample after acid erosion can truly characterize the microstructure of Ti(C,N) grains. In this experiment, graphite and Ti(C,N) are insoluble in hydrochloric acid, while iron is dissolved in hydrochloric acid, which can effectively retain Ti(C,N) in the sample. The microstructure of the sample after acid intrusion is shown in Fig. 7(a). Compared with the photo before erosion in Fig. 2, the Fe phase has been completely eroded by hydrochloric acid, and the sample only contains the residual Ti(C,N) phase and graphite phase. The area originally occupied by Fe becomes a gully, and the exposed area in the gully is the true phase interface morphology of Fe–Ti(C,N). The 3D image of the precipitated phase of Ti(C,N) can be obtained. Due to the insoluble nature of Ti(C,N) in hydrochloric acid, the interface of Ti(C,N)-Fe originally covered by the iron phase can be completely observed. In the image of the top view, Ti(C,N) exhibits a shape of jagged step (Fig. 7(b)). In the image of the front view, it can be seen that the phase interface is inlaid by many random island structures, and these island structures are incomplete cubes. The morphology of the Ti(C,N)-Fe phase interface suggests that the Ti(C,N) grains may not be dense inside (Fig. 7(c)), and there are a large number of pores. These island structures are completely random at the phase interface, and the intersection of island structures exhibits a distinct right angle (Fig. 7(d)). Different from the sharpness of the island structure in Fig. 7(d), there are some defects in the surface of the island structure as shown in Fig. 7(e), indicating that these island structures grow in the hot metal. More obvious features can be observed under high magnification conditions (Fig. 7(f)). The growth mode of layer-by-layer or multilayer also depends on the temperature and supersaturating of the liquid phase by [Ti], [N] and [C]. The structure shows the typical features of the superstructure or mesoscopic crystals. In recent years, the successful synthesis of a great number of mesoscopic crystals with a variety of well-defined physical parameters via various physicochemical routes has been reported.20) Among these approaches to synthesize nanomaterials (mesoscopic crystals), the solution-phase method is a very robust one to tailor physical parameters.21) There is a vast number of reports on the crystallization of inorganics from aqueous and organic solution, while mineralization from ionic liquids (ILs) is a much less studied field.22,23,24) However, the research on the superstructure and mesoscopic crystals is primarily achieved under the conditions of laboratory synthesis, the actual production of superstructure and mesoscopic crystals in the industrial production process especially in the iron making industry have not been reported yet. Therefore, Ti(C,N) mesocrystals of rather uniform hexagonal columnar morphology is first found on a large scale in BF hearth.
Microstructure of the sample after acid intrusion. (a) Microstructure of the sample after acid intrusion; (b)–(f) are the local images enlarged gradually.
The Ti(C,N) grain orientation and grain size in the titanium deposit can be characterized by electron backscatter diffraction (EBSD) statistic. Figure 8 shows the EBSD results of the sample. The phase distribution diagram (Fig. 8(a)) shows that the sample contains three phases. It is in agreement with the expected existence of XRD results. The blue part corresponds to the ferrite matrix, the red part corresponds to Ti(C,N) phase, and the black area corresponds to the graphite phase. There are many pores with different sizes in the Ti(C,N) phase region, which are filled by the iron phase. Comparing the grain boundary diagram (Fig. 8(b)) and the orientation imaging diagram (Fig. 8(c)), it can be seen that each Ti(C,N) phase has the same orientation and there is no grain boundary in the region. Therefore, it can be judged that each titanium precipitation phase region is an independent grain. The grain size of Ti(C,N) is obviously larger than the size of iron, and the grain size of Ti(C,N) is also very different, and the large one is more than 100 μm. The difference in grain size of Ti(C,N) can also be found from the low-magnification SEM results. Each Ti(C,N) region in the titanium deposit is a single grain and Ti(C,N) grain has a large grain size, which is known from the results of EBSD.
EBSD results of titanium deposit. (a) Phase distribution diagram; (b) Grain boundary diagram; (c) Orientation imaging diagram. (Online version in color.)
Ti(C,N) presented in the titanium deposit has a large grain size, so it is necessary to investigate the growth mechanism of Ti(C,N) grains in the hydrothermal fluid. In the process of iron making, iron ore with titanium-bearing material is used for BF maintenance. TiO2 is largely reduced to [Ti] and enters into molten iron in the area from the cohesive zone to the slag-iron interface. From previous understanding of the formation of Ti(C,N), the temperature has a significant influence on both morphology and phase transformation.25) There is a temperature gradient from the slag-iron interface to the sidewall lining of BF hearth, and the sidewall temperature is lower due to the cooling system. [Ti] is firstly generated in the hot metal with the form of TiC and TiN under certain temperature conditions, and the following reactions can be occurred.
| (1) |
| (2) |
The composition of hot metal has an important influence on the formation of the BF protective layer. The average composition of hot metal before the BF shut down is shown in Table 1.
| Element | Si, % | Mn, % | P, % | S, % | Ti, % | Temperature, K |
|---|---|---|---|---|---|---|
| Percentage | 0.436 | 0.341 | 0.03 | 0.104 | 0.226 | 1760 |
The critical precipitation temperatures of TiC and TiN under certain conditions of hot metal composition are calculated in our previous study.19) TiC and TiN start to precipitate with the temperature falling below the critical value. TiC and TiN are not precipitated separately after formation but precipitated in the form of Ti(C,N) solid solution. At the same time, Fe remains liquid. Ti(C,N) acts as an ideal solid solution and reacts as follows.
| (3) |
When the hot metal temperature reaches an appropriate range in the hearth, a large amount of fine Ti(C,N) particles are nucleated from the hot metal in a large amount by the above reaction. Through the observations of previous experiments have been discussed, the growth process of Ti(C,N) grains have not been reported. The precipitated Ti(C,N) exhibits single crystals or matrix-like clusters of grains that are substantially larger than the average grain size. Ti(C,N)-Fe phase boundary morphology and Ti(C,N) grains characterization of EBSD indicate that the growth of Ti(C,N) maybe by the coalescence of fine Ti(C,N) particles. Coalescence refers to the elimination of a common boundary by grain boundary migration or immediate dissolution of grain boundaries. Based on the analysis of the deposited sample, the growth process of Ti(C,N) grains can be deduced as follows, shown in Fig. 9.
Formation process of the Ti(C,N) crystals. (Online version in color.)
Firstly, [C], [Ti], and [N] dissolved in the molten iron achieve a concentration of Ti(C,N) precipitation, and Ti, C, and N atoms form short-range ordered cluster structure grains, which are nanoparticles.
Secondly, nanoparticles are coalesced for superstructure. The crystal lattices of neighboring nanoparticles must coincide. The Ti(C,N) nanoparticles have a similar lattice parameter regardless of the ratio of C/N. Therefore, it can be considered that the Ti(C,N) nanoparticles grow up by nanoparticles coalescence and accomplish by slight nanoparticles rotations or by shifting the neighboring grains to achieve the coincidence of the lattice.26,27) The hot metal continuously cools and shrinks, and there is a force transmission between the nanoparticles. This provides the necessary conditions for the rotation of the nanoparticles of the asymmetric contact. The fine Ti(C,N) nanoparticles rotate to make the contact nanoparticles orientation uniform, and the nanoparticles’ boundaries are eliminated by migration or dissolution. Since these fine Ti(C,N) nanoparticles float in hot metal, this rotation may be easy. The result of the rotation is that when the Ti(C,N) nanoparticles orientations of the contacts tend to be uniform, they are directly combined into one large grain, which is called superstructure or mesocrystal. Surface interaction between the nanoparticles plays a critical role during the formation process of a mesocrystal and may be responsible for the formation of external faces. Mesocrystals are ordered mesoscale superstructures composed of individual nanocrystals that are aligned along a common crystallographic direction, exhibiting similar to the ones of a single crystal. In a top-down view, a typical mesocrystal can be regarded as a single crystal.
Thirdly, a single crystal grows continuously, which is layer-by-layer grain growth. Because the fine Ti(C,N) crystals have larger specific surface area than the coarse Ti(C,N) nanoparticles, they are more favorable for the deposition. The coarse Ti(C,N) crystals in hot metal have many surface defects and high roughness, so they have a large surface area, which increases the mass transfer rate of the crystal surface and contributes to 2D nucleation on the interfaces or surface defects. Therefore, fine Ti(C,N) particles precipitate on the surface of coarse Ti(C,N) crystals. So the mesocrystals in the coarse Ti(C,N) crystals tend to grow up, resulting in a layered structure, which are self-assembled nanomaterials with highly ordered structures. Special emphasis has been placed on the different possible forces that may drive orientation and assembly between mesocrystals. The oriented attachment or mineral bridges existing between the crystals is extremely important, which leads to a relatively high crystallization degree and may be responsible for some unusual properties.
Finally, the deposited layer has formed. The Ti(C,N) crystals have a long growth time during the solidification of the titanium deposit. The iron matrix begins to nucleate when the temperature reaches the crystallization temperature. The Ti(C,N) crystals have a size advantage,25) and [Ti], [C], [N] in hot metal continue to diffuse around the Ti(C,N) crystals, which also promotes the continued growth of Ti(C,N) crystals. Therefore, the crystal size of Ti(C,N) in the as-cast microstructure is significantly larger than that of the iron matrix. Due to atomic mobility and thermodynamic changes, Ti(C,N) crystallizes at different periods will have a different ratio of C/N. Coincidently, the crystal size of Ti(C,N) in the titanium deposit is also large. Furthermore, the performance of the Ti(C,N) crystals is vital for BF maintenance. Therefore, the reasonable crystal size and the properties of the titanium precipitated needs to be further explored. The process of crystallization of titanium deposit will provide a certain theoretical basis for the furnace maintenance of iron ore titanium-bearing.
(1) The precipitation of Ti(C,N) exhibits annual-ring shape features with different colors. EPMA results show that the mapping of N in the Ti(C,N) grain has also characteristic of annual-ring. The annual-ring shape topography of the titanium precipitate is caused by the change of C/N ratio.
(2) The 3D morphology of Ti(C,N) precipitates and the phase interface of Ti(C,N)-Fe were obtained by acid erosion. The precipitated Ti(C,N) is composed of one or more large grains. The phase interface of Ti(C,N)-Fe appears like a shape of jagged step, and the phase interface is inlaid by many random mesocrystal structures.
(3) During the crystallization process of titanium deposit, the growth mechanism of Ti(C,N) grains is clarified. Ti(C,N) presents as superstructure or mesocrystal and evolved from Ti(C,N) nanoparticles, which are self-assembled nanomaterials with highly ordered structures. Ti(C,N) deposit forms continuously with layer-by-layer mesocrystal growth.
This work was financially supported by the Fundamental Research Funds for the Central Universities (No. FRF-TP-18-008A1) and the China Postdoctoral Science Foundation (No. 2018M631339).
