Investigation on lanthanide-dependent z value of JEM-phase Sialon

Abstract

JEM-phase Sialon (LnSi_6−zAl_1+zO_zN_10−z) was revealed to have limited z values. To keep the structure stability, substitution of the Si–N bond by the Al–O bond is very restricted and lanthanide-dependent (La → Sm). The z values of La-, Nd- and Sm-doped JEM-phase Sialons are ∼1.0, ∼0.6 and 0.4–0.2, respectively.

---

The phase relationship in the Ln₂O₃–Si₃N₄–AlN–Al₂O₃ systems has been intensively investigated for many years and summarised by Izhevskiy et al.¹ In 1995, the crystal structure of the so-called “JEM”-phase Sialon was reported by Grins et al.² via Rietveld refinement from powder X-ray diffraction (XRD) data. Since then, JEM-phase Sialon has always been treated as an intergranular phase in Ln–Si–Al–O–N systems.^2,3 The z value of JEM-phase La–Sialon was reported to be about 1.0, while no comments were made on whether there is any difference for z values of JEM-phase Sialons doped with different lanthanides. In 2007, Takahashi et al.⁴ reported La and Ce cations co-doped JEM-phase Sialon powders, and they claimed that JEM-phase powders could be used as phosphors and suitable for solid-state lightings especially for home illumination. These JEM-phase Sialon powders, however, were impure because the researchers used the Ln₂O₃, Si₃N₄ and AlN as starting materials. In actual fact, even though they mentioned that the initial value of parameter z was set to be around 1.0, the reaction equation among them, as indicated by the eqn (1), could only get balanced with z value of 1.5 by using those raw materials. This is because the atom ratio of O/Ln in lanthanide oxide is of a fixed value of 1.5 and the lanthanide oxide cannot be nitridized to be LnN due to the thermodynamic reason.

z ≥ 1.5: 1/2Ln₂O₃ + (2 − z/3)Si₃N₄ + (2 + z/3)AlN + (z/3 − 1/2)Al₂O₃ = LnSi_6−zAl_1+zO_zN_10−z (1)

Apparently, it would not be possible to prepare pure JEM-phase Sialon materials with z values of less than 1.5 when using Ln₂O₃, Si₃N₄, AlN and Al₂O₃ as raw materials. The application of JEM-phase Sialon materials would be hampered if they were of a considerable amount of impurities. For instance, impure JEM-phase phosphors would probably affect the service performance of white light-emitting diodes (LEDs), limiting the potential application of JEM-phase Sialon as phosphors.

In our previous work, we successfully developed novel JEM-phase Nd–Sialon orthogonal microcrystal arrays⁵ and reported the crystal structure of Nd-doped JEM-phase Sialon determined from single crystal XRD data.⁶ It revealed that the Nd–Sialon single crystal could emit UV-to-blue light when excited by 325 nm UV-light. Most importantly, in these studies the average z value of JEM-phase Nd–Sialon crystals was estimated to be around 0.4⁶ to 0.6⁵, which were all smaller than the suggested value of ∼1.0. Further study on the z-value will be critical for the preparation of JEM-phase Sialon. Besides, the knowledge on the homogeneity range of JEM-phase Sialon will be of great importance to its applications (including phosphors).

To the best of our knowledge, no single-phase JEM-phase Sialon powder/ceramic has hitherto been reported in the literature. From our previous results^5,6 we have a hypothesis that, JEM-phase Sialon might be of a particular z value or a narrow range of z values. If so, then it is fair enough that no pure phase would be prepared from those starting compositions not designed to be of such a value or be within such a possible range. On the other hand, if z value of JEM-phase Sialon can vary in a relatively wide range, monophasic JEM-phase Sialon materials with various z values would be prepared. In order to verify our hypothesis, we used lanthanide nitrides (LnN) rather than lanthanide oxides (Ln₂O₃) as starting materials (where Ln = Nd and Sm) for the preparation of JEM-phase LnSi_6−zAl_1+zO_zN_10−z in this study. The nominal compositions of various z values can be designed according to either eqn (2) or (3), where z value can be various in the range of 0 ≤ z ≤ 1.5:

0 ≤ z ≤ 1.5: (1 − 2z/3)LnN + z/3Ln₂O₃ + (2 − z/3)Si₃N₄ + (1 + z)AlN = LnSi_6−zAl_1+zO_zN_10−z (2)

z ≥ 0: LnN + (2 − z/3)Si₃N₄ + (1 + z/3)AlN + z/3Al₂O₃ = LnSi_6−zAl_1+zO_zN_10−z (3)

Si₃N₄ (purity 99.7 wt%, α-Si₃N₄ 93.8 wt%), AlN (purity 99.7 wt%), Al₂O₃ (purity 99.9 wt%), LnN (Ln = Nd and Sm, purity 99.5 wt%) and Ln₂O₃ (Ln = Nd and Sm, purity 99.9 wt%) powders were used as starting materials. According to the eqn (3), Nd-doped samples with nominal z values of 0.1, 0.2, 0.4, 0.5, 0.6, 0.8, 1.0 and 1.2 were designed, and Sm-doped samples were set with nominal z values of 0.2, 0.4, 0.6, 0.8, 1.0 and 1.2, respectively. In addition, Nd-doped samples with nominal z values of 1.5, 2.0 and 3.0 were designed as per eqn (1). There are difficulties during the preparation process of those green samples involving LnN powders. It is because that lanthanide nitrides are extremely moisture-sensitive; when exposed in air, strong exothermic reaction eqn (4) occurs and LnN changes into Ln(OH)₃ with NH₃ gas which can be smelled when mixing the powders. Therefore, we adopted a surface-sacrifice method as follows to protect the inner part of samples.

LnN + 3H₂O = Ln(OH)₃ + NH₃ (4)

The mixture powders of each sample with a total weight of 3 g were dry-mixed for 5 min, and then shaped in a ϕ20 steel die and wrapped in sealed packages quickly. The green compacts were taken out from the packages and quickly placed in a graphite crucible, then immediately vacuumed in the furnace before being heated in a high-temperature furnace via gas pressure sintering (GPS) process. The samples were synthesized from ambient temperature to 1700 °C at a ramp rate of 10 K min⁻¹ and held for 3 h with a nitrogen pressure of 0.9 MPa, and then cooled down to room temperature by switching off the power after the scheduled dwell. Due to the difficulty of using lanthanide nitrides, the whole process has certain limitation that there existed some unavoidable periods (5–10 min in total) when the green samples were exposed in air, and thus caused certain degree of transformation of LnN into Ln(OH)₃ on the surface of samples.

With the sacrifice of the surface, LnN in the inside of samples were exempt from being transformed. It is fortunate that the colors of sacrificed surface and JEM-phase Ln–Sialon (inner part) are different. To be specific, the sample surfaces were loose and grey, while the inner of samples were much denser, and the color for Nd-doped samples was light blue and that for Sm-doped ones was light yellow. It indicates that before solid-state reaction of raw materials took place, LnN in the inner was protected effectively in a considerable degree with the sacrifice of the surface.

It is thus believed that JEM-phase Sialons were synthesized successfully with the sacrifice method. The inner powders were dug out carefully for phase analysis. The phase compositions were identified by X-ray powder diffraction (XRPD; D8-Advance, Bruker, Germany) with Cu K_α radiation. The quantitative phase analysis of the XRPD data (full pattern) was done using the JAVA based software namely Materials Analysis Using Diffraction (MAUD)⁷ by Rietveld method.

Although the samples were moulded by mechanical pressing, the samples heated under the above synthesis procedure were not sintered. Instead, the surface were much looser and of grey colour. This phenomenon is due to that the exothermic reaction of LnN (on the surface of green compacts) with moisture in air can give out NH₃ gas.

The light-blue powders in the inner of the Nd-doped samples were carefully dug out for phase examination. Fig. 1 shows the XRD patterns of the inner materials of Nd-doped samples. After the gas pressure sintering (GPS) process at 1700 °C for 3 h, the products of those samples with initial z values from 0.1 to 1.2, mainly consisted with JEM-phase NdSi_6−zAl_1+zO_zN_10−z, M-phase Nd₂Si_3−xAl_xO_3+xN_4−x, and NdSi₃N₅. As it can be seen that, the impurities (M-phase and NdSi₃N₅) in the samples with designed z value of 0.6 and 0.8 showed the lowest peaks among those samples, which indicates that the z value of JEM-phase Sialon should be probably within a small range around 0.6.

Fig. 1 XRD patterns of the Nd-doped samples (inner).

The quantitative phase analysis was conducted based on XRD patterns of those synthesized powders via Rietveld method. Fig. 2 presents the weight fraction of each phase against the nominal z value of JEM-phase Nd–Sialon. When z value was set in the range of 0.1–0.5, the Nd–α-Sialon was present after solid-state reaction at the synthesis temperature, apart from the major phases of JEM-phase, M-phase and minor NdSi₃N₅.⁸ The low contents of JEM-phase Sialon in these samples illustrate that the crystal structure of Nd-doped JEM-phase cannot be stable if the substitution of Si–N bond by Al–O bond is relatively low (i.e. z value is lower than 0.5). As can be seen from the figure that the weight fraction of JEM-phase reach a maximum of 90.7 wt% (92.4 vol%) when setting the nominal z value to be 0.6, higher than those with other z values. On the other hand, with z values of 0.6–1.2, JEM-phase Nd-Sialon was still the dominant phase in the powder material but decreased in the quantity from 90.7 wt% to 76.3 wt%, respectively. Higher z value (0.8–1.2) resulted in more content of M-phase Nd–Sialon (4.1–8.7 wt%) and extra AlN remained (1.5–7.8 wt%). The phase assemblages in the samples with nominal z values of 1.2 and 1.5 are similar. Additionally, M-phase rather than JEM-phase was synthesized in the samples with nominal z values of 2.0 and 3.0. Therefore, the real z value should be much closer to 0.6 than other values.

Fig. 2 The weight fraction of each phase against the nominal z value of JEM-phase Nd–Sialon.

The yellowish powders in the inner of the Sm-doped samples were dug out for XRD analysis as well. The XRD patterns are shown in Fig. 3. When the initial z value was 0.2–0.4, the diffraction peaks of JEM-phase Sm–Sialon were very intense, with only very weak peaks of secondary phases, including M-phase and an “F”-phase Sialons (crystal structure of which will be reported elsewhere). While, the relative intensity of JEM-phase decreased dramatically if the nominal z value was higher, and disappeared when z = 1.0. The higher z value (1.0–1.2) did not favor the preparation of JEM-phase but M-phase Sialon.

Fig. 3 XRD patterns of the Sm-doped samples (inner).

The results by quantitative phase analysis are shown in Fig. 4, indicating the relationship of phase content as function of nominal z value. The content of Sm-doped JEM-phase reached the maximum of 97.1 wt% (97.3 vol%) at nominal z value of 0.2, and decreased to 31.9 wt% when the nominal z value increased to 0.8, accompanying with the formation of secondary phases. In the samples with z value of 1.0 and 1.2, no JEM phase existed, and the α-phase dominated. The content of M-phase is much less than α-phase, although the peaks of M-phase were stronger than those of α-phase (Fig. 3). It can be explained by the fact that the diffraction ability of M-phase is much higher than α-phase. According the JCPDS standard cards, the typical Reference Intensity Ratio (RIR) values of M-phase and α-phase Sialons are >7.5 and <0.9, respectively.

Fig. 4 The weight fraction of each phase against the nominal z value of JEM-phase Sm–Sialon.

It is widely known that monophasic materials of β-phase Si_6−zAl_zO_zN_8−z can be prepared with a wide range of z values (0 < z < 4.2).^9–11 Unlike β-Sialon, the z value of JEM-phase Sialon is probably within a narrow range or even a constant depending on the Ln ion in the structure, as indicated from ∼1.0 for La-JEM,² ∼0.6 for Nd-JEM, and much smaller values (z = 0.2–0.4) for Sm-JEM according to this study. Fig. 5 shows a schematic diagram of phase area of JEM-phase Sialon in the Ln–Si–Al–O–N systems, which is drawn according to the eqn (3) and the z value ranges of JEM-phase Sialon with different lanthanides. It may be explained by that the substitution of Si–N by Al–O bond (lengths are 1.74 and 1.75 Å, respectively¹²) is much less in the structure of JEM-phase Sialon doped by lanthanides (La to Sm) with increasing atomic number (57 to 62). By limiting the substitution level, JEM-phase can be structurally stable, and adapt to the decreasing ionic radius of lanthanide elements. It can also explain the previous question that why the researchers^2,4–6 failed to prepare the pure JEM-phase Sialon powders when they used Ln₂O₃, Si₃N₄, AlN and Al₂O₃ as the starting materials and formulated according to the reaction eqn (1). In future study, the first-principles calculations^10,11,13,14 of thermochemistry of JEM-phase Sialon doped with various lanthanides would be worth to be conducted. Moreover, the computational structure simulation should also be a good way to evaluate the structure stability of JEM-phase Sialon.

Fig. 5 Schematic diagram of phase area of JEM-phase Sialon.

In summary, by using lanthanide nitrides (NdN and SmN) as raw materials, the z values of JEM-phase Sialons (LnSi_6−zAl_1+zO_zN_10−z, Ln = La and Sm) were investigated in this study for the first time. The following conclusions can be made:

(1) The z value of JEM-phase Sialon is not in a wide range but probably constant or in a very narrow range. Moreover, it reduces with the decrease of ion radius of lanthanide (from La to Sm), i.e. ∼1.0 for La-JEM, ∼0.6 for Nd-JEM and 0.2–0.4 for Sm-JEM. In other words, less and less Si–N bonds can be substituted by Al–O bonds in the structure of JEM-phase with dopant from La to Sm.

(2) Such a substitution is limited to adapt to the decreasing ionic radius of lanthanide elements (La to Sm) with increasing atomic number (57 to 62), so as to keep the stability of the crystal structure of JEM-phase Sialon. When the radius of the lanthanide ion is smaller than that of samarium, its crystal structure is not stable any more.

(3) Nd- and Sm- doped JEM-phase Sialon powders synthesized within these z-value ranges were dramatically purer than others. It is impossible to prepare phase-pure JEM-phase Sialon materials by using lanthanide oxides as starting materials because the z value is much smaller than 1.5 in the chemical formula.

(4) Even though pure JEM-phase can probably be prepared by optimising the z value further, the synthesis route reported here is far away from industry production of JEM-phase phosphors due to the fact that as the employed lanthanide nitrides are very active they cannot encounter air or water.

This work was partially supported by the National Natural Science Foundation of China (Grant no. 51032007) and the Fundamental Research Funds for the Central Universities (Grant no. 2011PY0173). Y. G. Liu thanks the New Star Technology Plan of Beijing (Grant no. 2007A080) and the Program for New Century Excellent Talents in University (Grant no. NCET-12-0951). S. F. Huang thanks the Scholarship Foundation for Technological Innovation of Postgraduates, China University of Geosciences (Beijing), for financial aid. S. F. Huang and X. Ouyang would like to acknowledge the China Scholarship Council (CSC) for providing a doctoral scholarship to study at the University of Auckland.

Notes and references

1. V. A. Izhevskiy, L. A. Genova, J. C. Bressiani and F. Aldinger, J. Eur. Ceram. Soc., 2000, 20, 2275–2295.
2. J. Grins, Z. J. Shen, M. Nygren and T. Ekstrom, J. Mater. Chem., 1995, 5, 2001–2006.
3. Z. Shen and M. Nygren, in Nitrides and Oxynitrides, eds. S. Hampshire and M. J. Pomeroy, 2000, vol. 325–323, pp. 191–197.
4. K. Takahashi, N. Hirosaki, R. J. Xie, M. Harada, K. I. Yoshimura and Y. Tomomura, Appl. Phys. Lett., 2007, 91, 091923.
5. S. F. Huang, Z. H. Huang, M. H. Fang, Y. G. Liu, J. T. Huang and J. Z. Yang, Cryst. Growth Des., 2010, 10, 2439–2442.
6. S. F. Huang, Z. H. Huang, Y. G. Liu and M. H. Fang, Dalton Trans., 2011, 40, 1261–1266.
7. L. Lutterotti, S. Matthies and H.-R. Wenk, IUCr: Newsletter of the CPD, 1999, 21, 14–15.
8. M. Woike and W. Jeitschko, Inorg. Chem., 1995, 34, 5105–5108.
9. J. F. Yang, Y. Beppu, G. J. Zhang, T. Ohji and S. Kanzaki, J. Am. Ceram. Soc., 2002, 85, 1879–1881.
10. K. Tatsumi, I. Tanaka, H. Adachi and M. Yoshiya, Phys. Rev. B: Condens. Matter, 2002, 66, 165210.
11. J. J. Liang, A. Navrotsky, V. J. Leppert, M. J. Paskowitz, S. H. Risbud, T. Ludwig, H. J. Seifert, F. Aldinger and M. Mitomo, J. Mater. Res., 1999, 14, 4630–4636.
12. J. J. Melendez-Martinez and A. Dominguez-Rodriguez, Prog. Mater. Sci., 2004, 49, 19–107.
13. I. Tanaka, A. Kuwabara, K. Yuge, A. Seko, F. Oba and K. Matsunaga, in Siaions and Non-Oxides, ed. K. Komeya, Y. B. Cheng, J. Tatami and M. Mitomo, 2009, vol. 403, pp. 73–76.
14. C. M. Fang and R. Metselaar, J. Phys.: Condens. Matter, 2004, 16, 2931–2939.

---