Improvement of the oxidation resistance of TiC-coated carbon black by high-temperature annealing in N₂

Abstract

The annealing of TiC-coated carbon black in N₂ at 1000–1300 °C resulted in formation of a TiC_xN_y solid solution, grain growth and removal of nanocracks in coatings, which significantly increased the commencement and completion temperatures of carbon oxidation, effectively improving their oxidation resistance.

---

Titanium carbide (TiC) is considered to be an important refractory ceramic material, attributed to its high hardness, high thermal stability and superior wear resistance. Apart from the applications in cutting tools and wear-resistant metal parts,¹ TiC has been used in the preparation of coatings on carbon materials such as carbon fibers, aiming to improve their wettability and hinder excessive interfacial reaction with the liquid metal matrix.² TiC coating also provides a barrier to retard oxidation of carbon materials when they are used in high temperature applications.³

Recently, TiC coatings have been prepared on carbon powders for applications in refractory composites.⁴ It was reported that after coating with TiC by molten salt synthesis, the wettability and suspension property of carbon black had been significantly enhanced. However, the improvement in oxidation resistance of carbon powders was limited, which affected the lifetimes of carbon-containing refractory composites. There are two possible ways to increase oxidation temperature and decrease oxidation rate of TiC coated carbon black (CB), (1) increasing the thickness of TiC coating and (2) making coating denser. Dong and his colleagues⁵ revealed that the oxidation resistance of TiC coated carbon had a close relation with thickness. In their experiments, the oxidation of carbon fiber was considerably retarded when TiC thickness reached up to 200 nm. However, it is not likely to further increase TiC thickness in this case (e.g. ∼50 nm maximum), as a reasonable level of carbon content needed to remain for submicron sized TiC-coated CB (∼500 nm). Nevertheless, the densification of carbide coatings can be achieved simply by high-temperature annealing, because heat treatment will promote grain growth and reduce the nanocracks formed in the coatings.⁶

TiCN is a solid solution of TiC and TiN, and has similar crystal structure with both of them, however, it exhibits excellent combination of properties such as hardness–toughness compromise, good wear resistance and chemical oxidation.⁷ In this work, TiCN solid solution was formed by annealing of TiC-coated CB in N₂ at different temperatures. The effects of annealing on their chemical composition and oxidation resistance were investigated.

TiC coated carbon black (hereafter referred to as TiC@CB) were prepared by using molten salt synthesis technology.^4b The mixture of Ti powders (325 mesh, 99.7% pure, Sigma-Aldrich) and carbon black (N991, Cancarb, Canada) were fired in KCl at 950 °C for 4 h. The resultant powders were obtained by repeatedly washing and filtering. TiC@CB powders were then heat-treated at 1000–1300 °C in N₂ and Ar atmospheres, respectively. The annealed TiC@CB powders were fired at 500 °C for 30 min in air, in order to investigate their oxidation behavior at the beginning of oxidation.

Phase variation during annealing process was identified by a powder X-ray diffractometer (XRD) (D8Focus, Bruker, Germany). The overlapped peaks were analyzed by Jade5.0 software after slow scanning from 2theta of 35 to 45°. Microstructures and chemical composition of TiC@CB powders were characterized using a field emission gun scanning electron microscope (SEM) (Inspect FEI, USA) equipped with energy-dispersive spectroscope (EDS) (Oxford Instruments, UK). Thermogravimetry analysis (TGA) and differential scanning calorimetry (DSC) were performed in air using a SDT-Q600 thermogravimetric analyzer (TA instruments, USA) from room temperature to 1000 °C. TGA and DSC results were recorded at the heating rate of 10 °C min⁻¹.

Fig. 1a shows XRD patterns of TiC@CB particles after firing in N₂ at temperatures ranging from 1000 to 1300 °C. The peak intensity increased and the width of half peaks narrowed gradually with temperature, showing higher annealing temperature caused higher crystallinity and grain growth. On the other hand, the diffraction peaks of annealed powders deviated to high diffraction angle. This indicates elemental N might be incorporated into TiC crystal structure causing lattice reduction. Among these patterns, the peaks from heat-treatment at 1100 °C appeared unsymmetrical. As shown in Fig. 1b, a slow scan from 2theta of 35 to 45° demonstrated a clear peak profile of (111) and (200) planes of the obtained crystals. Peak analysis confirmed that two peaks overlapped. Combined with appearance of bifurcated peaks at high diffraction angles (Fig. 1a), it could be reasonably believed that there was still a small portion of untransformed TiC remaining in the samples after annealing at 1100 °C. However, only a single phase was detected in the samples treated at 1200–1300 °C, because their diffraction peaks appeared symmetrical.

Fig. 1 (a) XRD of TiC@CB particles after annealing in N₂ at 1000–1300 °C, respectively (Si was used to calibrate the position of TiC peaks, excluding peaks shift caused by instrument and sample preparation). (b) Slow scanned XRD peaks diffracted from (111) and (200) planes of TiC_xN_y cubic crystals which was prepared by annealing at 1100 °C.

By peak refinement and calculation, the lattice constant was obtained for each case. Fig. 2 shows the lattice constant of as-prepared TiC coating was 4.3247 Å, very close to that of perfect TiC crystal structure (a = 4.3274 Å, (ICCD card [38-1420])). It kept dropping to 4.3189 Å after annealing in N₂ at 1000 °C, indicating a small portion of N was solidized to TiC crystal structure. After 1100 °C treatment in N₂, the cell constant was changed to 4.2549 Å, and it continued to decrease to 4.2511 Å at 1200 °C and 4.2505 Å at 1300 °C which was very close to standard lattice constant value of TiN (a = 4.2417 Å). The change and evolution of lattice constant revealed that TiC was transformed to TiCN solid solution by annealing in N₂.

Fig. 2 Peak shift of a typical (200) plane of TiC crystalline structure and change on lattice constant as a function of annealing temperature in N₂. The corresponding data of TiC coating prepared at 950 °C was also provided for comparison.

The TiCN phase was modelled as a perfect TiC_1−x–TiN_x solid solution, 0 ≤ x ≤ 1, with a NaCl type crystal lattice. In accordance with Vegard's rule,⁸

a_(TiCN) = x_(C)a_(TiC) + x_(N)a_(TiN) (1.1)

where

x_(C) = 1 − x_(N) (1.2)

So

x_(N) = (a_(TiC) − a_(TiCN))/(a_(TiC) − a_(TiN)) (1.3)

According to eqn (1.3), the atomic percentage of carbon and nitrogen atoms could be estimated from parameter a_(TiCN). The nominal composition of TiCN phase was listed in Table 1. At 1100 °C, 83% N replaced C. It increased to 88% at 1200 °C, and 89% at 1300 °C. It indicated that the increase of N content would be very small (<1%) when keeping temperature rising.

Table 1 Cell parameter and nominal composition of as-prepared TiC@CB and samples annealed at different annealing temperatures

Annealing temperature	Modelled TiC	Synthesized 950 °C	Annealed				Modelled TiN
			1000 °C	1100 °C	1200 °C	1300 °C
Lattice constant (Å)	4.3274	4.3247	4.3189	4.2549	4.2511	4.2505	4.2417
x	1		0.8899	0.1730	0.1231	0.1144	0
y	0		0.1101	0.8270	0.8769	0.8856	1
Nominal composition (TiCxNy)	TiC	TiC	TiC0.89N0.11	TiC0.17N0.83	TiC0.12N0.88	TiC0.11N0.89	TiN

---

TiC@CB powders were also heated in Ar in order to compare the differences in microstructures and oxidation resistance of treated TiC@CB. As can be seen in Fig. 3a and b, TiC@CB particles annealed in Ar or N₂ showed similar morphologies and size to the ones before treatment. Nevertheless, larger grain size was observed in the sample annealed in N₂. After annealing in Ar, no new phases were seen (Fig. 3c), but the TiC peaks became higher and sharper, indicating an increase in its crystallinity and crystal size. No peak deviation was detected compared with the one without treatment. Chemical analysis of the sample annealed in N₂ by EDS (Fig. 3d) confirmed the incorporation of elemental N, and consequently the formation of a TiC_xN_y solid solution.

Fig. 3 (a and b) SEM images and (c) XRD of TiC@CB particles after annealing at 1300 °C for 4 h in (a) Ar and (b) N₂. (d) EDS spectrum obtained from (b) showing its chemical composition.

Fig. 4a shows TG curves of TiC@CB after heat-treatment were right shifted to higher oxidation temperatures. The weight gain for the sample treated in Ar and N₂ was retarded, reaching the highest point at ∼532 °C for the former and ∼635 °C for the latter, compared with ∼495 °C obtained from the sample before annealing. The rate of weight loss after weigh gain for these three samples was almost the same, however, TiC@CB particles after annealing in Ar lost all carbon at ∼821 °C and the one treated in N₂ terminated its weight loss at ∼838 °C, about 150 °C higher than that without heat treatment (∼688 °C). Such a significant retardation in weight loss implies annealing treatment particularly in N₂ was effective to improve oxidation resistance of TiC@CB particles.

Fig. 4 (a) TG curves and (b) DSC of as-prepared TiC@CB particles before and after annealing in Ar and N₂ at 1300 °C, respectively.

DSC testing was performed to provide more direct evidence of improvement of oxidation resistance of TiC@CB after high-temperature annealing. Fig. 4b shows the exothermic peaks of oxidation of TiC shells and carbon cores. Prior to carbon combustion, TiC were oxidized initially acting as protecting coatings. The positions of exothermic peaks of TiC and corresponding carbon were shifted simultaneously to higher oxidation temperatures for samples after annealing in Ar and N₂, clearly indicating the annealing of samples evidently slowed down the oxidation of carbon.

The samples before and after heating in Ar and N₂ were oxidized at 500 °C for 30 min and characterized by SEM and EDS in order to clearly compare their differences in oxidation resistance and durability in oxidizing condition. Fig. 5a, as an example, shows a SEM image of as-prepared TiC@CB particles after oxidizing. The spherical shape of particles remained, but high level of oxygen on particle surface was detected by EDS (Fig. 5d). However, only minor oxygen was found in the sample annealed in Ar (Fig. 5b and e) when they were oxidized. For TiC@CB after annealing in N₂, the spheres remained almost the same to the one before oxidation at 500 °C, and elemental oxygen was hardly found (Fig. 5c and f). This observation is consistent with the results of weight loss characterized by TG (Fig. 4a).

Fig. 5 SEM images of TiC@CB particles (a) before and after annealing at 1300 °C in (b) Ar and (c) N₂. (d–f) are EDS spectra corresponding to (a–c). All these samples were oxidized in air at 500 °C for 30 min.

Results of TG, DSC and chemical analysis of annealed samples after oxidation showed that the oxidation resistance of TiC@CB was improved significantly by annealing in N₂. One of the key factors is the high-temperature annealing. As shown in Fig. 3, TiC@CB before and after annealing in Ar at 1300 °C remained the same morphologies and chemical phases. Nevertheless, the oxidation of TiC@CB particles was considerably retarded (Fig. 4). This is because high-temperature treatment prompted the growth of crystalline which was clearly verified in Fig. 3c. Since nano grains of titanium carbide oxidation was very exothermic and fast,⁹ large grain size of TiC increased their thermal stability in oxidizing atmosphere to some extent. Besides, the growth of grain size at high temperature reduced the amount of grain boundary and consequently decreased the nanocracks between loose connected crystals.¹⁰

In addition, the formation of TiC_xN_y phase contributed to the remarkable improvement of oxidation resistance. XRD characterization and relevant calculation (Table 1) indicated higher annealing temperature resulted in more elemental N incorporating in as-prepared TiC crystal structure. As a result, N/C ratio in TiC_xN_y phase increased with annealing temperature, reaching 8 : 1 for sample annealed at 1300 °C (TiC_0.11N_0.89). The thermal stability of TiC and TiN was investigated in previous work.¹¹ The oxidation of TiC normally began at 483 °C and TiN started oxidizing at 737 °C. It was also reported that the incorporation of N in TiC would improve its oxidation resistance which increased with N/C ratio. In Gupta's work,¹⁰ TiCN thin film was deposited on Ti sheets by laser alloying from CH₄ and N₂. The authors investigated the TiCN on the micro level (XPS and angle-integrated photoelectron spectroscopy) and suggested that the affinity of Ti in TiC_xN_y solid solution toward oxygen had been reduced remarkably. However, in this work, we has proved that TiCN coating showed better oxidation resistance than TiC coating by using direct experimental data, for instance, oxidation rate and evolution of microstructure and chemical composition.

In summary, this work has provided a simple solution to further improve oxidation resistance of TiC-coated carbon particles. By simply heating in N₂, TiC-coated carbon could be converted to TiCN-coated carbon and became oxidation resistant. This will make TiC-coated carbon more applicable in carbon-containing high temperature composites. It also revealed the evident solid solution reaction occurred at 1100 °C, and the N/C ratio in TiCN was increased with annealing temperature. Besides, different performance of the samples annealed in Ar and N₂ indicated the improvement of their oxidation resistance was not only attributed to grain growth and reduction of grain boundary, but also to oxidation-resistant TiCN coating.

Acknowledgements

We thank the University of Sheffield and Open Foundation of the State Key Laboratory of Refractories and Metallurgy (Grant no. G201505, Wuhan University of Science and Technology) for financial support, and the University of Exeter for assistance with TG-DSC analysis.

Notes and references

1. (a) J. Barry and G. Byrne, Wear, 2001, 247, 139–151; (b) S. Zhang, W. T. Wu, M. C. Wang and H. C. Man, Surf. Coat. Technol., 2001, 138, 95–100; (c) Y. Xing, J. Deng, J. Zhao, G. Zhang and K. Zhang, Int. J. Refract. Met. Hard Mater., 2014, 43, 46–58.
2. P. Baumli, J. Sychev, I. Budai, J. T. Szabo and G. Kaptay, Composites, Part A, 2013, 44, 47–50.
3. (a) X. Li, Z. Dong, A. Westwood, A. Brown, S. Zhang, R. Brydson, N. Li and B. Rand, Carbon, 2008, 46, 305–309; (b) T. Piquero, H. Vincent, C. Vincent and J. Bouix, Carbon, 1995, 33, 455–467.
4. (a) X. Liu and S. Zhang, J. Am. Ceram. Soc., 2008, 91, 667–670; (b) J. Ye, R. Thackray, W. Lee and S. Zhang, J. Mater. Sci., 2013, 48, 6269–6275.
5. Z. J. Dong, X. K. Li, G. M. Yuan, Y. Cong, N. Li, Z. Y. Jiang and Z. J. Hu, Thin Solid Films, 2009, 517, 3248–3252.
6. (a) Y. H. Lu, J. P. Wang, S. L. Tao and Z. F. Zhou, Appl. Surf. Sci., 2011, 257, 6380–6386; (b) M. Ganjali, S. Pourhashem and M. Mozafari, Ceram. Int., 2015, 41, 13100–13104.
7. (a) P. K. Ajikumar, M. Kamruddin, S. Kalavathi, A. K. Balamurugan, S. Kataria, P. Shankar and A. K. Tyagi, Ceram. Int., 2012, 38, 2253–2259; (b) J. M. Córdoba, M. J. Sayagués, M. D. Alcalá and F. J. Gotor, J. Am. Ceram. Soc., 2005, 88, 1760–1764.
8. F. Monteverde, V. Medri and A. Bellosi, Appl. Organomet. Chem., 2001, 15, 421–429.
9. J. Sempere, R. Nomen, E. Serra, B. Sempere and D. Guglielmi, J. Therm. Anal. Calorim., 2011, 105, 719–726.
10. R. Gupta, S. Soni and D. M. Phase, Appl. Phys. A, 2015, 118, 191–196.
11. (a) S. U. A. Mitsuo, N. Nihira and M. Iwaki, Surf. Coat. Technol., 1998, 103–104, 98–103; (b) Y. H. Lu, J. P. Wang and Y. G. Shen, Appl. Surf. Sci., 2011, 257, 2769–2774.

---