FABRICATION of TITANIUM CARBIDE REINFORCED IRON… Chem

Available on line at Association of the Chemical Engineers of Serbia AChE

Chemical Industry & Chemical Engineering Quarterly www.ache.org.rs/CICEQ Chem. Ind. Chem. Eng. Q. 25 (1) 21−29 (2019) CI&CEQ

KE-HAN WU1 FABRICATION OF TITANIUM CARBIDE HAI-PENG GOU2 REINFORCED IRON MATRIX CERMET BY GUO-HUA ZHANG1 1,2 VACUUM CARBOTHERMAL REDUCTION KUO-CHIH CHOU OF ILMENITE 1State Key Laboratory of Advanced Metallurgy, University of Science Article Highlights and Technology Beijing, Beijing, • Vacuum carbothermic reduction was used to remove impurity elements Mg, Ca and Mn China • After reduction, most of Si and part of Al were dissolved in the iron matrix 2 Collaborative Innovation Center of • High reduction temperature helped increase hardness and bending strength of Fe-TiC Steel Technology, University of cermet Science and Technology Beijing, • Both excessive and insufficient carbon were detrimental to performances of the cermet Beijing, China Abstract SCIENTIFIC PAPER Iron matrix cermet reinforced with TiC has been produced by vacuum carbothermal reduction of ilmenite followed by sintering processes. The influences of UDC 669.018.9:669.1:544:66 reduction temperature and carbon mass ratio were discussed in detail. X-Ray diffraction (XRD), electron probe micro-analyzer (EPMA) and scanning electron microscope (SEM) with energy dispersive spectrometer (EDS) were emp- loyed to characterize the phase composition and microstructures. After carbothermic reduction, most of Mg, Mn, Ca evaporated from the sample; Si and part of Al was dissolved in the iron matrix. The obtained powders were used as the raw materials to produce TiC-Fe cermet by vacuum sintering. Density, hardness and bending strength of the samples were examined. The optimal cermet products after heat treatment had a density of 5.38 g·cm-1, a hardness of 1125.5 HV and a bending strength of 667 MPa, which was obtained at the carbon/ilmenite mass ratio of 0.378:1 at 1773 K under the pressure of 10 Pa. Keywords: titanium carbide reinforced iron, ilmenite, vacuum, carbothermal reduction.

Ceramic-metal composites (cermet) have become Fe/TiC cermet is a prospective cermet owing to its the focus of present studies due to their excellent excellent wettability among most of metal/TiC cermets wear resistance and hardness, and admirable specific [6-11]. modulus and strength. Steel bonded carbide first For decades, various methods have been emp- appeared in the early 1960s as a creative cermet loyed to synthesize titanium carbide-reinforced Fe- integrating the characteristics of both steel and based composites, such as carbothermal reduction carbide [1]. In various carbides, titanium carbide is a [12,13], powder metallurgy [14], in situ synthesis [15- promising reinforcing material in cermet due to its –18], and electrochemical synthesis [19]. In addition, high melting point, high chemical and thermal stabil- several densification routes have been developed, ity, excellent wear resistance and hardness [2-5]. including vacuum sintering, hot-pressing, microwave sintering, and spark plasma sintering [20-23]. Among these methods, the carbothermal reduction of ilmenite Correspondence: G.-H. Zhang, State Key Laboratory of Adv- is one of the most promising routes due to the low anced Metallurgy, University of Science and Technology Beijing, Beijing 100083, China. cost of raw materials. After carbothermal reduction, E-mail: [email protected] vacuum sintering is an effective densification tech- Paper received: 24 November, 2017 nique because of its simplicity [24]. Welham and Wills Paper revised: 16 May, 2018 Paper accepted: 29 May, 2018 put forward the production of TiN/TiC-Fe composites directly from ilmenite, and the phase evolution during https://doi.org/10.2298/CICEQ171114015W

21 K.-H. WU et al.: FABRICATION OF TITANIUM CARBIDE REINFORCED IRON… Chem. Ind. Chem. Eng. Q. 25 (1) 21−29 (2019) carbothermic reduction process of ilmenite in Ar Ilmenite and activated carbon were blended in a atmosphere was studied [25]. Gupta et al. found that blender for 5 min with a rotating speed of 11000 rpm. the addition of FeCl3·6H2O provided nuclei of iron and The mixed powders, with the addition of polyvinyl increased the reaction rate during the reduction pro- alcohol solution (PVA, 3 wt.%), were compressed into cess [26]. cylindrical pellets of ∅ 18 mm×18 mm by uniaxial In China, more than 90 wt.% of ilmenite is loc- pressing in a stainless-steel die under 230 MPa. Then ated in Panzhihua, Sichuan Province [27]. However, the pellets, together with alumina crucible, were put besides FeO, Fe2O3 and TiO2, there are many other into an electric resistance furnace with a vacuum components in ilmenite, such as MgO, SiO2, Al2O3, pressure control system. The sample was heated to MnO and CaO. Especially, the content of MgO is as the desired temperature at a heating rate of 5 K∙min−1 high as 6 wt.% [28]. Gou et al. found that it was hard and held for 4 h, and then was cooled down to to separate the main impurity element Mg by the con- ambient temperature at a cooling rate of 5 K∙min−1. ventional process [29]. MgAl2O4 and Mg2SiO4 existed The detailed experimental parameters are shown in in the products after carbothermic reduction of ilme- Table 2 where the No. 6L-8L was the product No. 6-8 nite in Ar atmosphere and were hard to remove [30]. reduced at low temperature (1673 K). The atmo- Even if many investigations have been done on sphere of the system was maintained at 10 Pa by a the carbothermic reduction of ilmenite, most of them vacuum gauge (ZF-VPM-1, Huachang eternal Beijing have focused on preparation of composite powders Vacuum Technology Co., Ltd.) for all of the reduction without extending their investigation on densification process. In previous work [27], it was found that air of the powders and mechanical performance testing. atmosphere (or without special control) and argon In this paper, vacuum carbothermal reduction served atmosphere had no difference in carbothermal reduct- to remove the impurity elements (such as Mg, Mn) ion when the pressure was maintained at 10 Pa. from Panzhihua ilmenite, and vacuum sintering served Therefore, by considering the cost, the air atmo- to densify the products. The hardness, density, and sphere was selected for the reduction process. The the bending strength of the cermet were measured. obtained products were examined by X-ray diffraction (XRD, Rigaku Ultima IV), scanning electron micro- MATERIALS AND METHODS scope (SEM, MLA-250, voltage 200 V-30 kV), and electron probe micro-analyzer (EPMA, JXA-8230, Activated carbon and ilmenite were used in the voltage 0.2 V-30 KV). Contents of main impurity ele- experiment. The activated carbon (analytical reagent) ments are shown in Table 3, where content of ele- was supplied by Sinopharm Chemical Reagent Bei- ment Si was detected by sulfuric acid dehydration jing Co., Ltd. The ilmenite was supplied by Panzhihua gravimetric method and contents of elements Al, Ca, Iron and Steel (Group) Co., Ltd. Its compositions were Mg and Mn were detected by ICP-AES. detected by inductively coupled plasma atomic emis- sion spectrometry (ICP-AES) method and are shown Table 2. Experimental conditions of carbothermic reduction in in Table 1 [31-35]. To determine the contents of Al2O3 vacuum; pressure: 10 Pa; atmosphere: air; holding time: 4 h and SiO , 0.2 g sample mixed with 3 g anhydrous 2 Group Mass ratio (carbon:ilmenite) Temperature, K sodium hydroxide was melted at 923 K in a nickel No. 1 0.358:1 1773 crucible. Then the sample was leached by hot deion- No. 2 0.365:1 1773 ized water, and after that leached by concentrated hydrochloric acid. After the salt was completely dis- No. 3 0.371:1 1773 solved, the above solution was diluted with deionized No. 4 0.378:1 1773 water to 200 ml for ICP-AES measurement. To det- No. 5 0.385:1 1773 ermine the contents of iron oxide and titanium oxide, No. 6 0.391:1 1773 the sample was leached in polytetrafluoroethylene cru- No. 7 0.424:1 1773 cible by aqua regia, sulfuric acid and hydrofluoric acid, No. 8 0.456:1 1773 with heating. Then the solution was measured by No. 6L 0.391:1 1673 ICP-AES. The titration method with potassium dichro- No. 7L 0.424:1 1673 mate was used to measure Fe3+ and Fe2+ contents. No. 8L 0.456:1 1673

Table 1. Chemical compositions of Panzhihua ilmenite (wt.%)

Component FeO Fe2O3 TiO2 MgO SiO2 Al2O3 CaO MnO Total Content 36.85 5.59 45.34 5.76 3.46 1.35 0.96 0.69 100

22 K.-H. WU et al.: FABRICATION OF TITANIUM CARBIDE REINFORCED IRON… Chem. Ind. Chem. Eng. Q. 25 (1) 21−29 (2019)

Table 3. Contents of the main impurity elements in the products ports, and loading rate were 16 mm×2 mm×2 mm, after vacuum reduction 13.1 mm, 0.5 mm/min, respectively. Each group of Contents of the main impurity samples was tested three times and then the average Temperature, Group elements, wt.% value was taken. The density was investigated by the K Si Al Ca Mg Mn Archimedes’ principle. The detailed experimental No. 0 – 2.19 0.97 0.93 4.70 0.73 procedure flowchart is given in Figure 1. No. 1 1773 2.58 1.70 0.33 0.023 <0.0005 Table 4. Effect of carbothermal reduction temperature and ball No. 2 1773 2.55 1.78 030 0.037 <0.0005 milling on the particle size of iron and TiC (μm) No. 3 1773 2.60 1.55 0.31 0.044 <0.0005 No. 4 1773 2.54 1.23 0.25 0.037 <0.0005 Temperature, K Process Iron TiC No. 5 1773 2.55 1.32 0.23 0.038 <0.0005 1673 Before ball milling 19.9 3.3 No. 6 1773 2.58 1.30 0.37 0.0047 <0.0005 1673 After ball milling 7.8 3.3 No. 7 1773 2.64 1.14 0.23 0.015 <0.0005 1773 Before ball milling 55.6 7.5 No. 8 1773 2.52 1.65 0.29 0.019 <0.0005 1773 After ball milling 7.8 7.5 No. 6L 1673 2.62 1.12 0.58 0.27 <0.0005 No. 7L 1673 2.50 1.06 0.63 0.62 <0.0005 No. 8L 1673 2.50 1.01 0.65 0.78 <0.0005

In order to synthesize cermet, the powder products after vacuum carbothermic reduction were milled with alcohol in air atmosphere at normal atmo- spheric temperature by a planetary ball mill (QM- -3SP2) with a rotating speed of 580 rpm for 8 h. The products after ball milling were measured by particle size analyzer (SA-CP3) and the results are shown in Table 4. It was found that both the particle sizes of iron and TiC reduced at 1773 K were larger than those obtained at 1673 K, owing to the faster rate of grain growth at higher temperatures. The particle size of iron was substantially decreased after ball milling. However, the particle size of TiC had no change after ball milling. The mixture of 5 g sample and 0.18 g PVA were compressed into a cylindrical pellet with Figure 1. Experimental procedure flowchart. the size of ∅ 18 mm×4 mm by uniaxial pressing in a stainless-steel die under 308 MPa. Then the pellet RESULTS AND DISCUSSION was placed into alumina crucibles which were put into an electric resistance furnace connected with a turbo Vacuum carbothermal reduction molecular pump (TMP, JTFB 300). The samples were The theoretical contents of main impurity ele- heated to 600 K at a heating rate of 2 K∙min−1 and ments in the products are shown as No. 0 in Table 3. held for 2 h to remove PVA. Then the samples were It was calculated by assuming that final products were heated to 1673 K at a heating rate of 5 K∙min−1 and Fe, TiC, [Si], [Al], [Ca], [Mg] and [Mn], where [Si], [Al], held for 6 h. The samples were ultimately cooled [Ca], [Mg] and [Mn] were elements Si, Al, Ca, Mg and down to ambient temperature at a cooling rate of 5 Mn dissolved in liquid iron, respectively. By compar- K∙min−1. The vacuum pressure was maintained at ing the actually measured contents of various ele- 0.002 Pa during the whole sintering process. The ments with the theoretical contents shown in Table 3, obtained products were characterized by EPMA and the products of different elements can be approx- SEM with energy dispersive spectrometer (EDS) after imately deduced. If the measured content of one ele- polishing. The Vickers hardness (HV) of the cermet ment was lower than the corresponding theoretical was investigated by a micro-hardness tester (MH-6, value, it was conjectured that the element was separ- 100X). The bending strength was examined by the ated by gas state. Therefore, it could be seen that three-point bending method (CDW-5, 5KN). The size element Mn was eliminated; contents of elements Mg, of the bending test sample, distance between sup- Ca substantially decreased and a high temperature

23 K.-H. WU et al.: FABRICATION OF TITANIUM CARBIDE REINFORCED IRON… Chem. Ind. Chem. Eng. Q. 25 (1) 21−29 (2019) was beneficial for the removals of them; elements Al, Si almost remained in the products, revealing that gas products Al2O and SiO were not generated during the reduction process. The volume fraction of TiC in this composite was about 65 vol.%, which was approx- imately calculated by Eqs. (1) and (2): WW 100 TiC 100 TiC ρρ vol. % ==TiC TiC (1) TiC − WWWTiC++ Fe TiC 1 WTiC ρρρ ρ TiC Fe TiC Fe

==MMTiC WTi TiC WWTiC Ti WSi (2) MWMTi Si Ti where WTiC is the mass fraction of TiC in the composite; WFe is the mass fraction of Fe in the composite; WTi is the mass fraction of element Ti in the composite; WSi is the mass fraction of Si in the composite (about 2.56% in Table 3); ρTiC is the density of -3 TiC (4.93 g·cm ) [36]; ρFe is the density of Fe (7.8 -3 g·cm ) [36]; MTiC is the relative molecular mass of TiC

(60); MTi is the relative molecular mass of Ti (48); wTi/wSi is the mass ratio of element Ti to element Si in Panzhihua ilmenite (about 21:1 in Table 1). XRD patterns of the products after vacuum reduction are given in Figure 2. It was found that the products were composed of a titanium carbide phase and α-Fe phase after reduction. However, there were graphite phases formed in products as shown in Figure 2a. The graphite phase would reduce the purity of Fe/TiC composites and the strength of the composites materials. Therefore, the mass ratio of carbon to ilmenite should be less than 0.424:1. It was also found that there was a Ti2O3 phase formed in products No. 1 and No. 2 which indicated the initial Figure 2. XRD patterns of product obtained after reduction: carbon content was not enough to reduce ilmenite a) No. 1-5; b) No. 6L-8L and No. 6-8. completely, while the mass ratio of carbon to ilmenite Table 5. Oxygen content in the products and extent of reduct- was lower than 0.365:1. ion after vacuum reduction The oxygen contents in Panzhihua ilmenite and products were measured by oxygen, nitrogen and Group Oxygen content, wt.% Extent of reduction hydrogen analyzer (EMGA-830 OK, HORIBA) and are Ilmenite 33.2 0% shown in Table 5. The extent of reduction was cal- No. 1 2.3 93.1% culated by Eq. (3): No. 2 1.4 95.8% W No. 3 0.7 97.9% α =−100(1 O ) (3) ° No. 4 0.3 99.1% WO No. 5 0.3 99.1% where α is the extent of reduction; WO is the mass No. 6 0.3 99.1% o fraction of oxygen in products; WO is the mass No. 7 0.3 99.1% fraction of oxygen in ilmenite. It was found that when No. 8 0.3 99.1% the mass ratio of carbon to ilmenite was larger than No. 6L 1.0 97.0% 0.378:1, the initial carbon content was enough to No. 7L 1.0 97.0% make the degree of reduction be over 99%. Whereas No. 8L 0.9 97.3% when the ratio was lower than this value, reduction was incomplete.

24 K.-H. WU et al.: FABRICATION OF TITANIUM CARBIDE REINFORCED IRON… Chem. Ind. Chem. Eng. Q. 25 (1) 21−29 (2019)

SEM images of the products after vacuum red- Table 6. Contents of the main elements in the iron area after uction with different molar ratios of ilmenite to carbon reduction and sintering were almost the same and Figure 3 shows the typical Contents of the main elements, pictures of products No. 6L and No. 6, a light white Group T / K Process wt.% area and grey area were iron and titanium carbide by Fe Si Al Ti C EDS analysis, respectively. It was found that particle No. 1 1773 After reduction 92.4 5.35 1.10 0.75 0.38 sizes of TiC obtained at 1773 K were about 7 μm, After sintering 92.5 5.44 0.22 1.47 0.36 larger than those obtained at 1673K (3 μm), owing to No. 2 1773 After reduction 91.4 6.49 1.04 0.64 0.42 fast crystal growth rate at high temperatures. The After sintering 91.2 6.84 0.23 1.36 0.41 contents of the main elements in the iron area were No. 3 1773 After reduction 91.6 6.42 0.83 0.70 0.41 investigated by EPMA and are given in Table 6. It was After sintering 91.2 6.82 0.21 1.41 0.40 found that, in the iron area, contents of elements Fe, Si, Al and C were about 91.9, 6.0, 0.9 and 0.5 wt.%, No. 4 1773 After reduction 91.6 6.31 0.90 0.83 0.38 respectively. The contents of these elements in the After sintering 91.2 6.51 0.30 1.65 0.37 iron area were not affected by temperature and car- No. 5 1773 After reduction 92.6 5.38 0.81 0.80 0.42 bon ratio essentially. It was worth noting that 6.0 wt.% After sintering 91.8 5.55 0.24 2.01 0.43 of Si content in liquid iron was not enough to form the No. 6 1773 After reduction 89.6 6.95 1.55 1.55 0.37 SiC phase. The critical silicon content for SiC and C After sintering 86.6 10.94 0.18 1.97 0.35 saturated Fe-Si-C melts was about 22.4 wt.% by No. 7 1773 After reduction 92.6 5.47 1.19 0.16 0.56 Chipman et al. at 1763 K [37]. Only when the content After sintering 91.5 5.53 0.05 1.99 0.95 of Si in liquid iron was higher than this value, SiC No. 8 1773 After reduction 91.2 6.64 0.83 0.72 0.61 could be formed. Based on the above analyses, SiC After sintering 90.7 7.00 0.21 1.48 0.60 cannot be formed during the vacuum reduction pro- No. 6L 1673 After reduction 92.3 5.64 1.03 0.60 0.40 cess. Therefore, nearly all of the element Si should be After sintering 89.2 8.00 0.13 2.15 0.47 dissolved in the iron matrix. Besides, the element No. 7L 1673 After reduction 93.1 5.27 0.04 0.84 0.70 compositions were detected by EPMA, and it was After sintering 92.1 5.57 0.01 1.79 0.50 found that only Ti and C existed in the titanium car- No. 8L 1673 After reduction 92.2 5.99 0.64 0.73 0.42 bide phase. After sintering 91.7 6.24 0.20 1.44 0.38

According to the above analyses, it was specul- ated that Eqs. (4)-(11) may happen during the reduction process. The changes in Gibbs energy of different reactions under 10 Pa are given in Figure 4. It was found that at 1673 or 1773 K, the thermodynamic

reduction sequence of the different oxides was Fe2O3

> FeO > MnO > TiO2 > MgO > SiO2 > CaO > Al2O3. Most of Mg, Mn, Ca evaporated from the sample,

while Si and part of Al were dissolved in the iron Figure 3. SEM images of product obtained after reduction: matrix: a) No. 6L; b) No. 6. += + FeO(solid)( Csolid)(g COas)( Fe liquid) (4) Data of Al and Si shown in Tables 3 and 6 rep- Δ=−+ GT((CO)10p = Pa) 232.18 159580 resent the contents of Al and Si in the products and the iron matrix after the reduction process, respect- 11 Fe O+= C CO + Fe (5) ively. If all the Al and Si existed in the iron matrix, the 332 3(solid) (solid) (gas) (liguid) mass ratio of Al:Si in Tables 3 and 6 should be the Δ=GT−+248.25 162012 same. However, it was found that the mass ratio of ((CO)10p = Pa) Al:Si shown in Table 6 was about 1:5~8, while that 13 1 shown in Table 3 was about 1:2, which indicated that TiO+=+ C CO TiC (6) 222(solid) (solid) (gas) 2(solid) not all of the element Al existed in the iron matrix. A Δ=−+ certain amount of element Al may still exist in the GT(p (CO)= 10 Pa) 242.32 260711 form of oxides. += + MgO(solid)( Csolid)(g COas)(g Mg as) (7)

25 K.-H. WU et al.: FABRICATION OF TITANIUM CARBIDE REINFORCED IRON… Chem. Ind. Chem. Eng. Q. 25 (1) 21−29 (2019)

Δ=−+ GT((CO)(Mg)10pp== Pa) 445.09 617233 brightness were observed in Figure 5, including a continuous white area, grey angularity, small black += + MnO(solid)( Csolid)(g COas)(g Mn as) (8) sphere, thin longish black strip and the large black pore. Map scanning analyses by SEM and EDS were Δ=GT−+430.37 531589 ((CO)(Mn)10pp== Pa) used to determine these phases as shown in Figure 126. It was found that element Si was dissolved in the AlO+= C CO + [Al] (9) 332 3(solid) (solid) (gas) iron matrix; element Ti was combined with carbon as TiC; element Al was combined with oxygen as Al2O3. Δ=−+ GT(p (CO)= 10 Pa) 270.45 445930 Accordingly, the continuous light white area, grey angularity, small black sphere and thin longish black 11+= + SiO2(solid) C (solid) CO (gas) [Si] (10) strip in Figure 5 were iron, TiC, Al2O3 and graphite, 22respectively. The graphite should be precipitated from Δ=−+ the iron phase during cooling. The existence of Al O GT(p (CO)= 10 Pa) 261.01 357924 2 3 could decrease the purity of Fe/TiC cermet. However, += + CaO(solid)( Csolid)(g COas)(g Ca as) (11) the particle size of Al2O3 was ultrafine (<1 μm) which contributed to the pinning effect on the grain bound- Δ=−+ GT(pp (CO)== (Ca) 10 Pa) 435.38 678070 ary [38]. Accordingly, the existence of Al2O3 could reinforce the strength of Fe/TiC cermet [39]. where ΔG is the change of Gibbs energy of reaction [36]; T is the reaction temperature; p(CO), p(Mg), p(Ca) and p(Mn) are the partial pressures of CO, Mg, Ca and Mn, respectively.

Figure 4. Changes of Gibbs energy of different reactions.

Vacuum sintering SEM images of products No. 6L-8L and No. 6-8 after vacuum sintering are given in Figure 5 and the volume fraction of TiC in Figure 5 calculated by Image-Pro Plus 6.0 Software was given in Table 7. This value was roughly in consistency with the value calculated by chemical compositions in Eqs. (1) and (2). As shown in Figure 5, the size of TiC particles at Figure 5. SEM images of as-prepared cermets: a) No. 6L, 1673 1773 K was larger than that at 1673 K. Furthermore, K, AC/Ilmenite 0.391:1; b) No. 6, 1773 K, AC/Ilmenite 0.391:1; the pores of products reduced at 1773 K in Figure 5 c) No. 7L, 1673 K, AC/Ilmenite 0.424:1; d) No. L, 1773 K, was also less than that at 1673 K, which indicated AC/Ilmenite 0.424:1; e) No. 8L, 1673 K, AC/Ilmenite 0.456:1; that high reduction temperature was beneficial for the f) No. 8, 1773 K, AC/Ilmenite 0.456:1. densification of products during sintering. It was worth noting that several phases of different shapes and

26 K.-H. WU et al.: FABRICATION OF TITANIUM CARBIDE REINFORCED IRON… Chem. Ind. Chem. Eng. Q. 25 (1) 21−29 (2019)

Table 7. Volume fraction of TiC in SEM patterns of products Properties No.6L-8L and No. 6-8 Hardness, density, and bending strength of the Group No. 6L No. 6 No. 7L No. 7 No. 8L No. 8 obtained cermet are given in Table 8. The bending

vol.(TiC) % 67.2 65.4 64.7 63.5 68.3 61.9 strength of No. 8L was too low to test. Considering the data shown in Table 8, it was found that both the reduction temperatures and initial carbon ratio had an effect on the properties of the final products. By contrasting the density data in Table 8, it was found that a high reduction temperature was beneficial for dec- reasing the porosity of the product, which was also revealed by comparing Figure 5a, c and e with Figure 5b, d and f. The dense structure improved the adhe- sion strength of the interface between titanium carbide and iron matrix and decreased dislocation move- ments within the iron matrix. Therefore, high temperatures could enhance the hardness and bending strength by the aforementioned effects.

Table 8. Properties of the obtained cermet

Density Micro Vickers Bending strength Group T / K Figure 6. SEM images with map scanning analyses. g·cm-3 hardness MPa No. 1 1773 4.77 648 291 Contents of the main elements after sintering in No. 2 1773 5.10 733 326 the iron area investigated by EPMA are given in Table No. 3 1773 5.40 1001 603 6. By contrasting the contents of various elements No. 4 1773 5.38 1126 667 before and after sintering shown in Table 6, it was No. 5 1773 5.33 953 597 found that, in the iron area, the contents of elements No. 6 1773 5.30 950 509 Fe and Al decreased during the sintering process, No. 7 1773 5.30 950 509 while the contents of elements Si and Ti increased. No. 8 1773 5.12 705 412 Content of C was almost unchanged. Iron existed in No. 6L 1673 5.00 648 138 liquid state at high temperature. Under a high vacuum No. 7L 1673 4.50 644 149 degree during the sintering process, Fe could be vol- atilized in the gas phase. Therefore, the loss of Fe No. 8L 1673 3.84 641 N/A could be attributed to the evaporation under high vacuum conditions [40]. This evaporation of iron also Figure 7 shows the effect of carbon ratio on the resulted in the slight increase of Si and Ti. properties of the products. Each point was tested three times and then the average value was taken.

Figure 7. The relationship between the carbon ratio and properties of the products No. 1-8.

27 K.-H. WU et al.: FABRICATION OF TITANIUM CARBIDE REINFORCED IRON… Chem. Ind. Chem. Eng. Q. 25 (1) 21−29 (2019)

While the mass ratio of carbon to ilmenite was beyond CONCLUSION 0.378:1, excess carbon was essentially a weakened phase with a low density and low hardness in the cer- In this paper, iron matrix cermet reinforced with met. The existence of residual carbon could decrease TiC (65 vol.%) has been produced by carbothermal both the density and the mechanical properties. While vacuum reduction of ilmenite followed by the sintering the mass ratio was lower than 0.371:1, carbon was process. The main conclusions were drawn as insufficient to reduce ilmenite completely. The lack of follows: 1. During the reduction process, the element Mn carbon resulted in the existence of Ti2O3, which was another brittle phase with low density, low hardness, was eliminated; contents of elements Mg, Ca and poor wettability with the iron matrix [41]. There- substantially decreased and the high temperature fore, deficient carbon could also decrease both the was beneficial for their removal; element Si and density and the mechanical properties, too. Conse- part of element Al were dissolved in the iron mat- quently, a high reduction temperature could increase rix. density, hardness, and bending strength, while both 2. The high reduction temperature helped increase excessive and insufficient carbon decreased them. the density, hardness and bending strength, while The highest hardness was 1125.5 HV while the bend- both excessive and insufficient carbon decreased ing strength and density were 667 MPa and 5.38 them. -1 g·cm-3, respectively. This hardness data was close to 3. The optimal cermet with a density of 5.38 g·cm , the others’ works, as shown in Table 9 [14,20,42,43]. a hardness of 1125.5 HV, and bending strength of 667 MPa was obtained at the carbon/ilmenite Table 9. Hardness of TiC/Fe cermet from reference [14,20,42,43] mass ratio of 0.378:1 at 1773 K under the pressure of 10 Pa. Sintering TiC Hardness Substrate material temperature, K (vol.%) HV Acknowledgements 1703 61 35CrMo steel 1249 This work was supported by the Natural Science 1693 62 High manganese 1004 Foundation of China (No. 51734002 and U1460201). steel 1623 61 465 stainless steel 920 REFERENCES 1623 70 465 stainless steel 1140 Self-propagating 52 High manganese 600 [1] Y. Wu, X. Wang, F. Long, Y.F. Shen, Z.G. Zou, Mater. synthesis steel Sci. Forum (2009) 687-691 [2] N.R. Oh, S.K. Lee, K.C. Hwang, H.U. Hong, Scripta Figure 8 shows the cross-section perpendicular Mater. 112 (2015) 123-127 to the fractured surfaces in product No. 4 after bend- [3] V.K. Rai, R. Srivastava, S.K. Nath, S. Ray, Wear 231 ing testing. Transgranular cracking through TiC was (1999) 265-271. found which revealed that TiC was the brittle phase [4] C.C. Degnan, P.H. Shipway, J.V. Wood, Wear 250 (2001) and TiC had excellent bonding with the iron matrix. 1444-1451 Each TiC particle failed essentially by cleavage frac- [5] B. Yang, F. Wang, J. Zhang, Acta Mater. 51 (2003) 4977- -4989 ture with well-defined facets [2]. [6] M. Kiviö, L. Holappa, S. Louhenkilpi, M. Nakamoto, T. Tanaka, Metall. Mater. Trans., B 47 (2016) 2114-2122 [7] A. Contreras, E. Bedolla, R. Perez, Acta Mater. 52 (2004) 985-994 [8] P. Xiao, B. Derby, Acta Mater. 44 (1996) 307-314 [9] N. Frage, Metall. Mater. Trans., B 30 (1999) 857-863 [10] G.L. Borofsky, Metall. Mater. Trans., A 23 (1992) 709-727 [11] P. Zwigl, D.C. Dunand, Metall. Mater. Trans., A 29 (1998) 565-575 [12] M.R. Hasniyati, H. Zuhailawati, S. Ramakrishnan, S.A.R.S.A. Hamid, J. Alloy. Compd. 587 (2014) 442-450 [13] M.A.R. Dewan, G. Zhang, O. Ostrovski, Metall. Mater. Trans., B 41 (2010) 182-192 [14] Z. Wang, T. Lin, X. He, H. Shao, B. Tang, X. Qu, Int. J. Figure 8. SEM images showing the cross-sections perpen- Refract. Met. H. 58 (2016) 14-21 dicular to the fractured surfaces after the bending test.

28 K.-H. WU et al.: FABRICATION OF TITANIUM CARBIDE REINFORCED IRON… Chem. Ind. Chem. Eng. Q. 25 (1) 21−29 (2019)

[15] M. Razavi, M.S. Yaghmaee, M.R. Rahimipour, S. Sal- [29] H.P. Gou, G.H. Zhang, K.C. Chou, Chem. Ind. Chem. man, R. Tousi, Int. J. Miner. Process. 94 (2010) 97-100 Eng. Q. 23 (2017) 67-72 [16] B. AlMangour, D. Grzesiak, J.-M. Yang, J. Alloy. Compd. [30] H.P. Gou, G.H. Zhang, K.C. Chou, ISIJ Int. 55 (2015) 706 (2017) 409-418 928-933 [17] L. Zhong, Y. Xu, M. Hojamberdiev, J. Wang, J. Wang, [31] C. Shen, Chem. Ind. Times 30 (2016) 30-32 Mater. Design 32 (2011) 3790-3795 [32] X. Yang, Shandong Land Res. 30 (2014) 74-79 [18] Z. Ni, Y. Sun, F. Xue, J. Bai, Y. Lu, Mater. Design 32 [33] J.W. Li, W.H. Yu, S. Zhang, S.T. Hu, H. Yang, Wuhan (2011) 1462-1467 Iron Steel Cor. Tech. 55 (2017) 35-39 [19] S. Lin, Q. Song, X. Qian, Z. Ning, X. Lu, F. Derek, New J. [34] J.R. Qin, Y. Fan, Hunan Nonfer. Metals 31 (2015) 71-72 Chem. 39 (2015) 4391-4397 [35] I. Barin, F. Sauert, E. Schultze-Rhonhof, S.S. Wang, [20] Z. Wang, T. Lin, X. He, H. Shao, J. Zheng, X. Qu, J. Thermochemical data of pure substances, Parts I and II. Alloys Compd. 650 (2015) 918-924 VCH Verlags-gesellschaft, Weinheim (1989) 1-1829 [21] J.-J. Wang, J.-J. Hao, Z.-M. Guo, S. Wang, Int. J. Min. [36] J. Chipman, J. Fulton, N. Gokcen, G. Caskey, Acta Met. Mater. 22 (2015) 1328-1333 Metall. 2 (1954) 439-450 [22] W. Zhang, X. Zhang, J. Wang, C. Hong, Mat. Sci. Eng., A [37] J.F. Guo, J. Liu, C.N. Sun, S. Maleksaeedi, G. Bi, M.J. 381 (2004) 92-97 Tan, J. Wei, Mater. Sci. Eng., A 602 (2014) 143-149 [23] M.R. Rahimipour, M. Sobhani, Metall. Mater. Trans., B 44 [38] Z.P. Zhao, A.T. Tang, S.M. Liu, M. Chen, Mater. Rev. 27 (2013) 1120-1123 (2013) 54-59 [24] P. Ananthapadmanabhan, P.R. Taylor, J. Alloy. Compd. [39] P. Saha, D. Tarafdar, S.K. Pal, P. Saha, A.K. Srivastava, 287 (1999) 121-125 K. Das, Int. J. Adv. Manuf. Tech. 43 (2009) 107-116 [25] N.J. Welham, P.E. Willis, Metall. Mater. Trans., B 29 [40] C. Xuan, H. Shibata, S. Sukenaga, P.G. Jönsson, K. (1998) 1077-1083 Nakajima, ISIJ Int. 55 (2015) 1882-1890 [26] S.K. Gupta, V. Rajakumar, P. Grieveson, Metall. Mater. [41] F. Akhtar, S.J. Askari, J.A. Shah, S. Guo, J. Alloy. Trans., B 18 (1987) 713-718 Compd. 439 (2007) 287-293 [27] G.H. Zhang, H.P. Gou, K.H. Wu, K.C. Chou, Vacuum 143 [42] S.W. Hu, Y.G. Zhao, Z. Wang, Y.G. Li, Q.C. Jiang, Mater. (2017) 199-208 Design 44 (2013) 340-345. [28] R. Huang, P. Liu, X. Qian, J. Zhang, Vacuum 134 (2016) 20-24

KE-HAN WU1 IZRADA KERMETA TITAN-KARBIDOM OJAČANE HAI-PENG GOU2 MATRICE GVOŽĐA VAKUUM KARBOTERMIČKOM 1 GUO-HUA ZHANG REDUKCIJOM ILMENITA KUO-CHIH CHOU1,2

1State Key Laboratory of Advanced Kermet titan-karbidom ojačane matrice gvožđa je proizveden vakuum karbotermičkom Metallurgy, University of Science and redukcijom ilmenita i procesima sinterovanja. Detaljno su razmatrani uticaji temperature Technology Beijing, Beijing, China redukcije i masenog udela ugljenika. Za karakterizaciju sastava faze i mikrostruktura 2 Collaborative Innovation Center of korišćeni su rentgenska difrakcija (XRD), elektronska sonda mikro-analizatora (EPMA) i Steel Technology, University of skenirajući elektronski mikroskop (SEM) sa energetskim disperzijskim spektrometrom Science and Technology Beijing, (EDS). Nakon karbotermičke redukcije, veći deo Mg, Mn i Ca je ispario iz uzorka; Si i Beijing, China deo Al je rastvoren u gvožđevoj matrici. Dobijeni praškovi su korišćeni kao sirovina za proizvodnju TiC-Fe kermeta vakuumskim sinterovanjem. Ispitane su gustine, tvrdoća i NAUČNI RAD čvrstoća savijanja uzoraka. Optimalni proizvodi kermeta nakon toplotne obrade imaju gustinu od 5,38 g/cm, tvrdoću 1125,5 HV i čvrstoću savijanja od 667 MPa, koja je dobi- jena pri masenom odnosu ugljenik/ilmenit 0,378:1 na 1773 K i pritisku 10 Pa.

Ključne reči: gvožđe ojačano titan-karbidom; lmenit; vacuum; karbotermička redukcija.