Iron carbide and nitride via a flexible route: synthesis, structure and magnetic properties

Abstract

Single-phase Fe₄N and Fe₃C have been controllably synthesized from a coordination precursor by adjusting the molar ratio of hexamethylenetetramine to ferric chloride hexahydrate, both presenting good paramagnetic behavior, with maximum magnetizations that can reach 150.51 and 53.03 emu g⁻¹ up to 17.5 kOe, respectively. The magnetic properties of Fe₃C can be adjusted by the molar ratio of hexamethylenetetramine. Moreover, the structure, components and the possible mechanism were investigated carefully.

---

Introduction

Apart from the potential applications in chemical-electrocatalysts, metal carbides and nitrides with excellent magnetic properties are believed to be potential candidates in rare-earth-free magnetic materials, due to their superior mechanical, saturation magnetization of iron nitrides and carbides.^1,2 The relative works concerning the fabrication of iron carbides and nitrides were in general based on direct carbonization and nitridation, such as CVD,³ plasma treatment⁴ and magnetron sputtering,⁵ and reported the successful fabrication of metal carbides and nitrides. However, harsh reaction conditions, toxic reactants and huge energy consumption introduced from the direct carbonization and nitridation have led to large limitations in physical applications, compared to ferrite. Thus, it is necessary to develop a modified method to improve the preparation difficulties. Up to now, work on the preparation of carbides and nitrides has provided some new methods, such as salt conversion,⁶ thermal injection of Fe(CO)₅,⁷ thermal decomposition of Prussian blue⁸ and some sol–gel methods,^9–11 which have tackled the limitations mentioned above to some degree. During these works, the controllable preparation of iron carbides and nitrides by adjusting the reaction parameters was highlighted, and both the carbide and nitride sources were provided in the reaction system. Constanze Schliehe and co-workers⁹ synthesized iron carbides and nitrides using ferrite and urea. However, the prepared iron nitrides were impure and coexisted with FeO and Fe, although single-phase Fe₃C was fabricated successfully. Zoe Schnepp¹⁰ et al. fabricated Fe₃N and Fe₃C at 675 °C and 750 °C, respectively, where the ratio of iron acetate to gelatin was 3 : 1 at a heating temperature rate of 10 °C min⁻¹. However the process was difficult to copy because gelatin, as a carbon and nitrogen source, is a protein compound and does not have certain components. Besides the magnetism, an important property was not investigated. One could deduce from the work mentioned above, that the organic amine should be accepted as an effective agent in the controllable preparation of iron carbides and nitrides, and simultaneously provide carbide and nitride sources by adjusting the reaction conditions. A further study exhibited that a high ratio of amine to iron source promoted the formation of carbides, while a low ratio promoted the formation of nitrides.¹² Thus, the ratio of organic amines plays a key role in controlling the crystallization of iron carbides and nitrides. However, the impurity of ferrite resulted from hydrolysis in an aqueous system, which decreased the quality of the product. To improve the purity of the iron carbides and nitrides, hexamethylenetetramine (HMTA), which has been verified to synthesize Fe₃C nanoparticles,¹³ was applied in decreasing hydrolysis for its particular structure of tertiary amine. Besides, no extra oxygen element from HMTA was brought in, which was also favorable for improving the quality of the iron carbides and nitrides. From the analysis above, one could deduce that the iron carbides and nitrides may be fabricated selectively and remained of good quality via the HMTA route with an appropriate iron source.

Herein, we demonstrated a flexible route for the controllable preparation of Fe₃C and Fe₄N by adjusting the molar ratio of HMTA to ferric chloride hexahydrate. The morphology, structure, the possible mechanisms and magnetic properties were also investigated carefully.

Experiment section

Materials synthesis

2.70 g ferric chloride hexahydrate was dissolved into 10 mL acetone. Then a certain series of HMTA was added into the prepared solution in different molar ratios. The suspension should be centrifuged and dried under the protection of a N₂ flow at 50 °C from water and oxygen. Next, the dried precursor was calcinated at 670 °C for 1 h with a continuous N₂ flow, then cooled down to room temperature before fabricating the final product.

Characterization

X-ray powder diffraction (XRD) data were recorded on a Shimadzu diffractometer X-6000 with Cu-Kα radiation (λ = 1.5406 Å). The UV-vis spectra were characterized on a PUXI TU-1901 spectrophotometer. The magnetic properties were measured on a Lake Shore vibrating sample magnetometer 7404, where 0.10 g product was measured in a nonmagnetic Teflon capsule.

Results and discussion

Fig. 1 shows the XRD patterns of the samples prepared at different molar ratios of FeCl₃ to HMTA, indicating that the molar ratio has a large impact on the product crystallization. A low molar ratio of HMTA favors the formation of iron nitrides, yet the presence of ferrite can still be observed, due to crystal water still occupying the orbits of the iron ions. Iron nitride can be obtained as the major component at a ratio of 1 : 1. The diffraction peaks at 2θ = 41.2°, 47.9° and 70.0°can be well indexed to the (1 1 1), (2 0 0) and (2 2 0) crystal planes of cubic Fe₄N, corresponding to JCPDS no. 06-0627. Then the product transfers into α-Fe (JCPDS no. 06-0696) when the ratio of ferric chloride to HMTA reaches 1 : 1.5. Fe₃C (JCPDS no. 35-0772) began to appear at a ratio of 1 : 2, as shown by the blue curve. However, the as-synthesized Fe₃C coexisted with metallic iron, which can be ascribed to the incomplete carbonization caused by insufficient HMTA added into the reaction system, implying that it is necessary to increase HMTA in order to enhance the carbonization. The single-phase Fe₃C can be obtained when the HMTA ratio is increased to 1 : 3. Nevertheless, the further increase of HMTA did not lead to a new product and the intensity of the Fe₃C characteristic peaks even decreased, showing that the excessive HMTA introduction could not realize a deeper carbonization but increased the carbon content in the reaction system during the calcination process. The XRD patterns indicate that the amount of HMTA plays a key role in controlling the species of products between Fe₄N and Fe₃C. It is obvious that a small amount of HMTA favors a nitrogen source, which also corresponds to a similar conclusion in another study.¹² Interestingly, the iron nitride synthesized by the conventional sol–gel route is ε-Fe₃N, different from the result of this study. Although, the further mechanism has not been understood, we think the crystal water has a crucial impact on the crystallization and growth of Fe₄N, contrasting the Fe₃N prepared from anhydrous ferric chloride. Briefly, a small increase of HTMA promotes α-Fe, implying that a certain amount of HMTA is indispensable to Fe₃C.

Fig. 1 XRD patterns of the samples prepared from different ratios of FeCl₃ to HMTA.

A further increase gradually brings in enough carbon to fabricate Fe₃C, inducing the single-phase Fe₃C at a proper ratio. However after the formation of the Fe₃C structure, the excessive HMTA takes no effort to change the product species, only supplying a mass amorphous carbon matrix around the Fe₃C.

Based on the XRD patterns, we proposed a possible mechanism of Fe₄N and Fe₃C. It is well-known that most methods to produce nitrides are based on thermal decomposition under nitrogen or an ammonolysis step with or without ammonia,¹⁴ which is also applicable to our study to some degree. As discussed in another study,¹⁰ gelatin releases ammonia to realize the nitridation process. This implies that the nitrogen gas not only provides protection from oxidation, but also suppresses the decomposition of ammonia at a high temperature range. Thus, nitrogen gas is necessary to realize the whole reaction. Previous studies show that the FeCl₃·NH₃ adduct is the ideal precursor of iron nitride. We believe that the HMTA will form a similar coordinate adduct with FeCl₃·H₂O, where a lone pair electrons from the nitrogen atom in HMTA can interact and coordinate with the Fe³⁺ cations. To support our proposed mechanism, UV-vis absorption spectra were ascertained to confirm the presence of a solid coordination precursor. Compared to the UV-vis absorption spectrum of ferric chloride hexahydrate, the precursors prepared from the increasing concentration of HMTA with constant FeCl₃ show different UV-vis absorptions to support the presence of a new product, based on the ligand field theory. Interestingly, the decrease, and even disappearance of the absorption intensity and a gradual negative shift at around 412 nm and 480 nm, with an increasing concentration of HMTA with constant FeCl₃, demonstrated the point that the Fe³⁺ cations are coordinated with nitrogen atoms in HMTA by UV-vis spectra, verifying our mechanism above and the different compositions of precursors (Fig. 2).

Fig. 2 UV-vis absorption spectra of the precursors prepared with an increasing concentration of HMTA and constant FeCl₃.

If the HMTA is limited, the nitridation keeps the dominant role in the preparation of cubic Fe₄N, due to the stronger interaction between the iron and nitrogen atoms. Noticeably, trace crystal water still exists, because the unsaturated Fe³⁺ orbits are occupied to keep stable, forming ferrite after calcination, when the amount of HMTA is rather small. Based on the results that the anhydrous ferric chloride will form Fe₃N via a similar route, we believe the presence of crystal water is important for the crystallization of Fe₄N, although the further mechanism is still unknown. The extra HMTA then enhances the reduction properties, resulting in the appearance of Fe. Besides, the metastability of Fe₄N promotes the process mentioned above. We believe this reaction may be instant, due to the good dispersion and instability of nano-Fe₄N. The further increasing HMTA will introduce the carbonization from nano-Fe to Fe₃C, finally realizing the fabrication of Fe₃C. As with Fe synthesis, the carbonization process will be at a high rate when the ratio of ferric chloride to HMTA is greater than 1 : 1.5. Due to the excellent chemical stability, the excessive carbon cannot reduce the Fe₃C or prepare other kinds of iron carbides by deep carbonization. Of note is that both Fe₄N and α-Fe are the intermediates and precursors to Fe₃C, analyzed from the reaction process. Thus, only by controlling the precious molar ratio of ferric chloride to HMTA, can Fe₄N and α-Fe survive in the products.

The morphology of Fe₄N and Fe₃C was investigated by HRTEM (see Fig. 3). Both Fe₄N and Fe₃C are encapsulated by a mass amorphous carbon matrix, showing irregular particles with a large range of sizes. The lattice fringe spacing of 0.22 nm can be well index to the d (lattice plane distance) value of the (111) crystal plane of Fe₄N in Fig. 3(D). It can be observed that the morphology shows more or less visible sharp edges, which can be ascribed to the particle coalescence. This sharp morphology depends on the distance between particles, ramping and the local temperature, according to a previous study.¹⁵ Noticeably, the carbon matrix surrounding Fe₃C shows a graphite shell with an average thickness of 7 nm, as shown in Fig. 3(G), which is in agreement that the Fe₃C can promote the formation of graphite. Besides, the d values of 0.34 and 0.21 nm can be indexed to the graphite and Fe₃C in Fig. 3(H), respectively, supporting the presence of graphite and Fe₃C. It is noted that the carbon grows with an increase of HMTA (Fig. 4).

Fig. 3 (A)–(D) are HRTEM images of Fe₄N at different scales; (E)–(H) are HRTEM images of Fe₃C prepared at a ratio of 1 : 3 for FeCl₃ to HMTA at different scales.

Fig. 4 TG curves of Fe₄N and Fe₃C prepared from different ratios of FeCl₃ to HMTA.

To investigate the component of the products, thermal gravimetric (TG) analysis was introduced to evaluate the carbon content in Fe₄N and Fe₃C, in which the consequently iron element transfers into Fe₂O₃. The calculated carbon contents of different ratios are 16.96%, 67.83%, 84.32% and 89.36%, corresponding to the weight loss of 111.72%, 42.89%, 20.90% and 14.19%, as shown in Fig. 4. It is noticed that the carbon contains graphitic carbon, which has a large impact on the magnetic stability and property by reducing the random flipping of the magnetic moments, which result from thermal fluctuations¹⁶ and can be measured by Raman spectra, as shown in Fig. 5. The peaks at around 1350 and 1590 cm⁻¹, can be indexed to the D- and G- bands of carbon, respectively.^17–19 The proportions of I_D/I_G are 1.17, 1.22, 1.35 and 1.39, indicating the HMTA ratio has an important impact on the graphite degree of the carbon matrix.²⁰ The intensity of I_D/I_G for the Fe₃C samples enhances with an increase in the molar ratio of HMTA, implying the decreasing degree of graphite. It can be easily understood that the catalytic activity of Fe₃C may promote the surrounding amorphous carbon to form the graphite structure, which is confirmed in the HRTEM images. When excessive HMTA is added into the reaction system, more carbon is formed after the calcination process. However, only the carbon surrounding the surface of as-synthesized Fe₃C forms graphitic carbon, leading to the increasing I_D/I_G value. Interestingly, the value of I_D/I_G for Fe₄N is much smaller than Fe₃C, which has been confirmed the catalytic property to graphite and carbon nanotube, implying a better graphite degree. However, the distinguished graphite structure is not observed in the HRTEM images. We suppose that the high graphitic degree is assigned to the synergic effect of low carbon content caused by the limited ratio of HMTA and crystallization of Fe₄N. As distinguishable graphite has not been observed in the HRTEM images, the catalytic activity of Fe₄N to graphite cannot be confirmed.

Fig. 5 Raman spectra of Fe₄N and Fe₃C prepared from different ratios of FeCl₃ to HMTA.

The magnetic hysteresis loops of Fe₄N and Fe₃C were measured on the vibrating sample magnetometer (VSM). It is obvious that Fe₄N shows good soft magnetic and paramagnetic behaviors with a maximum magnetization of 150.51 emu g⁻¹ up to 17.5 kOe (Fig. 6 and Table 1). The rather low squareness of 2.53% (maximum magnetization/retentivity) broaden the physical application of Fe₄N as a soft magnetic material. The superior maximum magnetization and trace retentivity and coercivity can be assigned to the positive crystallization degree of Fe₄N and a little carbon matrix left in the product. While, the Fe₃C also shows paramagnetism, with maximum magnetizations that can reach 53.03, 29.47 and 12.22 emu g⁻¹ up to 17.5 kOe as the HMTA ratio increases. The other magnetic parameters are listed in Table 1. It should be noted that, the as-synthesized Fe₃C presents some coercivity and retentivity, which is very different from that of Fe₄N. The unexpected coercivity and retentivity can be due to the interference of the carbon matrix, leading to magnetic exchange coupling behavior from the insulation of amorphous carbon, which is commonly reported in magnetic material synthesis. Due to the increasing content and impact of the carbon matrix from an increased amount of HMTA, both squareness and coercivity increase sharply, as shown in Table 1. The carbon matrix contributes to the decreasing value of maximum magnetization of Fe₃C (much lower than the theoretical value of 120 emu g⁻¹), due to the mass remainder from increasing HMTA carbonization. As a result, the magnetic properties of Fe₃C particles can be controlled by adjusting the amount of HMTA added. Furthermore, the carbon matrix prevents Fe₄N and Fe₃C from oxidation and corrosion.

Fig. 6 The magnetic hysteresis loops of Fe₄N and Fe₃C prepared from different ratios of FeCl₃ to HMTA. The inset is the local magnification with the applied field between −0.4 and 0.4 kOe.

Table 1 Magnetic properties of products

Ratio	Saturation magnetism (emu g^−1)	Retentivity (MR, emu g^−1)	Coercivity (HC, Oe)
1 : 1	150.51	3.81	43.36
1 : 3	53.03	5.91	274.30
1 : 5	29.47	4.56	318.39
1 : 10	16.24	3.12	324.68

---

Conclusions

In conclusion, carbon-encapsulated Fe₄N and Fe₃C can be prepared controllably by adjusting the amount of HMTA. High magnetic performance verifies that the Fe₄N will be a significant candidate to replace ferrite. Furthermore, the magnetic properties of Fe₃C can be adjusted by controlling the HMTA ratio to satisfy the different requirements.

Acknowledgements

This work is supported by the National Natural Science Foundation of China.

Notes and references

1. T. Ogawa, Y. Ogata, R. Gallage, N. Kobayashi, N. Hayashi, Y. Kusano, S. Yamamoto, K. Kohara, M. Doi, M. Takano and M. Takahashi, Appl. Phys. Express, 2013, 6, 073007.
2. S. H. Huh and A. Nakajima, J. Appl. Phys., 2006, 99, 064302.
3. M. Terada, M. Itoh, J. R. Liu and K.-i. Machida, J. Magn. Magn. Mater., 2009, 321, 1209.
4. W. Tillmann, E. Vogli and S. Momeni, Mater. Sci. Forum, 2010, 638, 812.
5. X. Wang, W. Zheng, H. Tian, S. Yu, W. Xu, S. Meng, X. He, J. Han, C. Sun and B. Tay, Appl. Surf. Sci., 2003, 220, 30.
6. J. F. Geng, D. A. Jefferson and B. F. G. Johnson, Chem. Commun., 2004, 21, 2442.
7. C. Yang, H. Zhao, Y. Hou and D. Ma, J. Am. Chem. Soc., 2012, 134, 15814.
8. C. Aparicio, L. Machala and Z. Marusak, J. Therm. Anal. Calorim., 2012, 110, 661.
9. C. Schliehe, J. Yuan, S. Glatzel, K. Siemensmeyer, K. Kiefer and C. Giordano, Chem. Mater., 2012, 24, 2716.
10. Z. Schnepp, M. Thomas, S. Glatzel, K. Schlichte, R. Palkovits and C. Giordano, J. Mater. Chem., 2011, 21, 17760.
11. X. Lao, X. Yan, J. Xie and Y.-L. Li, in High-Performance Ceramics VII, Pts 1 and 2, ed. W. Pan and J. H. Gong, 2012, vol. 512–515, p. 1429.
12. C. Giordano, C. Erpen, W. Yao, B. Milke and M. Antonietti, Chem. Mater., 2009, 21, 5136.
13. N. Fan, X. Ma, Z. Ju and J. Li, Mater. Res. Bull., 2008, 43, 1549.
14. A. L. Hector, Chem. Soc. Rev., 2007, 36, 1745.
15. W. Baaziz, G. Melinte, O. Ersen, C. Pham-Huu and I. Janowska, Phys. Chem. Chem. Phys., 2014, 16, 15988.
16. A. Wu, D. Liu, L. Tong, L. Yu and H. Yang, CrystEngComm, 2011, 13, 876.
17. Y. Du, W. Liu, R. Qiang, Y. Wang, X. Han, J. Ma and P. Xu, ACS Appl. Mater. Interfaces, 2014, 6, 12997.
18. J. Robertson, Mater. Sci. Eng., R, 2002, 37, 129.
19. A. C. Ferrari and J. Robertson, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 61, 14095.
20. C. A. Grimes, D. Qian, E. C. Dickey, J. L. Allen and P. C. Eklund, J. Appl. Phys., 2000, 87, 5642.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra00336a

---