The phase change behavior and thermal conductivity of Aluminum nitride (AlN) ceramics with Y(NO₃)₃·6H₂O and Y₂O₃ as an additive were studied. Sintering was performed in the temperature of 1,900 ^oC for up to 3 h under a N₂ atmosphere to optimize the sintering conditions for each composition. The microstructure and assemblage of the secondary phase have a significant effect on the final thermal conductivity of the sintered AlN. The mechanical property and thermal conductivity of the AlN composition of using the Y₂O₃ additive were improved by adding Y(NO₃)₃·6H₂O, which decreased the porosity. At 13.56 wt% Y(NO₃)₃·6H₂O, the AlN ceramic exhibited the highest strength of 375 MPa, the highest hardness of 10.60 GPa, and the highest thermal conductivity of 200.2 W/m·K.

Keywords: AlN, Y(NO₃)₃·6H₂O, Phase change behavior, Thermal conductivity, Mechanical property

Aluminum nitride (AlN) may be highly suitable as substrates and packages for IC/LSI because of its high thermal conductivity (theoretical value of 320 W/m·K), low dielectric constant (8.0 at 1 MHz), and thermal expansion coefficient (4.8 × 10⁻⁶ K⁻¹ at 20-500 ℃) that is close to that of silicon [1-3]. However, because of its high covalent bonding, it is difficult to sintering. Several sintering additives, such as CaO, CaC₂, C, CaO-Al₂O₃, and Y₂O₃, were reported to be useful for the fabrication of high densification and thermal conductivity AlN ceramics [4-6]. Y₂O₃ is widely used as a sintering additive because it forms binary eutectics at temperatures around of ~1,800 ℃ with native Al₂O₃, which is present on the surface of AlN particles, resulting in a material with high density and high thermal conductivity [7, 8]. A typical feature in ceramic processing is the addition of a limited fraction of sintering additives, typically oxides, to promote densification. However, an associated disadvantage is that non-negligible alterations may occur of the bulk property for which the ceramic phase has been originally selected. The thermal conductivity of poly phase ceramics is strongly affected by internal phase geometry and other microstructural details [9-11]. High thermal conductivity constitutes an attractive property of AlN ceramics for high-power semiconductor devices [12, 13]. However, the usual preparation of dense polycrystalline AlN bodies involves the additive Y₂O₃ which, by virtue of its good wet-ability of the AlN grain surface, enables ready densification by pressureless sintering [4, 14, 15]. As substrates and packages, the thermal conductivity of AlN ceramics is very important. Most studies look at the influence of different sintering additives on the thermal conductivity of AlN ceramics.
In this work, we studied to compare the sintering behavior and crystalline phases change related with the densification of AlN with Y₂O₃ and Y(NO₃)₃·6H₂O additives.

Commercially available AlN powders (Grade H, Tokuyama Soda, Japan) was used as a starting material. First, the Y(NO₃)₃·6H₂O (99.99%, High Purity Chemicals, Japan) or Y₂O₃ (99.9%, High Purity Chemicals, Japan) was dissolved in ethanol, and the AlN powders were dispersed in the solution using ultrasonic technology. The Y(NO₃)₃·6H₂O and Y₂O₃ were added into a AlN suspension and then sonicated for 24 h to obtain a metal ion doped AlN suspension via surface adsorption. After dried and calcinated (500 ℃), the mixed powders were densified by hot pressing at 1,900 ℃ for 3 h under a pressure of 20 MPa in flowing N₂. The sintering additive of Y(NO₃)₃·6H₂O was added in the following amounts: 10.1198, 11.8684, 13.5638, and 15.2594 wt% (equivalent to 6, 7, 8 and 9 wt% Y₂O₃ respectively), referred to here after as A-6YP, A-7YP, A-8YP and A-9YP for the sintered samples. For comparison, monolithic AlN was also prepared in the same process. The composition of powder mixture was indicated in Table 1.
Crystalline and 2^nd phases were identified with X-ray diffraction pattern (D/MAX-2500V, RIGAKU) with Cu Kα radiation (λ = 1.5406 Å). All the samples for mechanical test were prepared by diamond saw and then grinded by diamond disc. Mechanical properties of specimen were evaluated by four-point flexural strength (5882, Instron, USA) with a cross head speed of 0.5 mm/min and micro vickers hardness tester (HM124, Akashi, Japan) at a load 9.8 N for 5 sec. The dimension of mechanical test specimen was 3 × 4 × 40 mm (width × length × height) and the inner span and outer span were 10 and 30 mm, respectively. Fracture surfaces of the pellets were observed by Field emission scanning electron microscopy (FE-SEM, S-4700, HITACHI, Japan). Thermal diffusivity measurement (LFA 457, NETZSCH, Germany) at room temperature is measured by the laser flash method. Thermal conductivity (K) is calculated using a heat capacity (C_p) of AlN at room temperature where λ and d are the thermal diffusivity and the density, respectively [16].

The densities and mechanical properties of the sintered AlN ceramics are summarized in Table 2. The density increases with the additive content. Compared with AlN when the additive is Y₂O₃, better mechanical properties, such as Vickers’ hardness, flexural strength, are obtained for sintered samples when the additive is ≤8 wt% Y(NO₃)₃·6H₂O, A-8YP exhibits the highest strength of 375 MPa, hardness of 10.60 GPa. This high density is due to the formation of an aluminum oxy nitride (Al–O–N) liquid by reaction between surface alumina and AlN. The addition of Y₂O₃ promotes more extensive liquid formation due to reaction to form a quaternary liquid (Y–Al–O–N) at lower temperature.

Fig. 1 gives the X-ray powder diffraction patterns for the all samples after sintering at 1,900 ℃ for three hours. Each composition shows strong peaks associated with AlN in addition to minor peaks associated with the secondary grain boundary phase.
Fig. 1(a) showed peaks for AlN, YAlO₃ (YAP) on the 2 wt% of Y₂O₃. The YAP phases disappeared with increasing Y₂O₃ content and then Y₃Al₅O₁₂ (YAG) phases (3 wt% of Y₂O₃) Y₄Al₂O₉ (YAM) phases (4 wt% of Y₂O₃) appeared. But Fig. 1(b) showed peaks for AlN, YAP and YAG phases with increasing Y(NO₃)₃·6H₂O content. Because compared with AlN when the additive is Y₂O₃, the improved dispersion properties are attributed to the increase in reactivity between Al₂O₃ and Y(NO₃)₃·6H₂O.
Al₂O₃ peaks were not detected because the amount of residual Al₂O₃ is very small and below the detection limit of the equipment. Moreover, Al₂O₃ occurs as a thin surface layer on AlN grains. Because of this surface layer and residual spinel, thermal conductivity is very low for this sample, as can be seen in Fig. 3. This means that the amount of added Y₂O₃ was insufficient to fully react with residual Al₂O₃ and to purify the AlN and to complete the oxygen removal from both the lattice and the surface.
From previous reports [17, 18], the Y₂O₃ decomposed from Y(NO₃)₃·6H₂O reacts with Al₂O₃, which is also present on the surface of AlN particles, to form yttrium aluminates:



The formation of liquid phase (Y₄Al₂O₉ and/or Y₃Al₅O₁₂) during sintering process will play an important role in the densification of the AlN ceramics.
The FE-SEM microstructure of composition containing Y₂O₃ and Y(NO₃)₃·6H₂O is shown in Fig. 2. The secondary phase (white) shows up at the grain junctions only and not along grain edges. Compared with AlN when the additive is Y₂O₃, the addition of Y(NO₃)₃·6H₂O appears to have a further effect of changing the wetting behavior of the liquid phase with respect to the AlN grains. Indeed, the dihedral angle of the secondary phase at the grain junction is high. This has a significant effect on the thermal conductivity (k).
Jackson et al. [19] suggest that thermal conductivity rises to a maximum and then decreases due to the increase in volume fraction of aluminates with further Y₂O₃ addition.

Fig. 3 shows the trend in thermal conductivity of the compositions. The thermal conductivity of the aluminate phase is low (<10 W/m·K), hence as the volume fraction increases, a decrease in the overall thermal conductivity is expected. The thermal conductivity of AlN sintered without sintering additives has a significantly low value of 92.2 W/m·K. Therefore, the effect of the additive has a major impact on the conductivity in two ways: (a) it removes oxygen from particle surface and (b) a micro structural change occurs as the aluminates de-wets the grain boundaries and segregates to the grain junctions leading to AlN–AlN grain boundary contact. Also, this figure shows, that compositions of using the Y(NO₃)₃·6H₂O additive have higher thermal conductivity than composition of using the Y₂O₃ additive. Because of the oxide layer was consumed to produce YAP, YAG as secondary phases. Furthermore, it was shown by the experimental work of Medraj et al. [20] that YAP wets the surface of AlN more than YAG or YAM if all the experimental conditions are considered. Also, the presence of YAP phase will prevent AlN–AlN surface contact. But higher YAP content is associated with lower thermal conductivity.

Aluminum nitride (AlN) ceramics were prepared by hot-pressing with Y(NO₃)₃·6H₂O and Y₂O₃ as sintering additive. When increasing Y₂O₃ content, the YAlO₃ (YAP) phase disappeared and then Y₃Al₅O₁₂ (YAG) phase (3 wt% of Y₂O₃) Y₄Al₂O₉(YAM) phase (4 wt% of Y₂O₃) appeared. But increasing Y(NO₃)₃·6H₂O content showed peaks for AlN, YAP and YAG phase. Because compared with AlN when the additive is Y₂O₃, the improved dispersion properties are attributed to the increase in reactivity between Al₂O₃ and Y(NO₃)₃·6H₂O. The mechanical properties and thermal conductivity are obtained for sintered samples when the additive is ≤13.5638 wt% Y(NO₃)₃·6H₂O, A-8YP specimen exhibits the strength of 375 MPa, hardness of 10.60 GPa and the highest thermal conductivity of 200.2 W/m·K. The addition of Y(NO₃)₃·6H₂O appears to have a further effect of changing the wetting behavior of the liquid phase with respect to the AlN grains. It was confirmed that AlN using Y(NO₃)₃·6H₂O showed relatively higher thermal conductivity and mechanical properties than the Y₂O₃.

This work was financially supported by Ministry of Science and ICT(MSIT) in Korean government and Korea Industrial Technology Association (KOITA) as “A study on the programs to support collaborative research among industry, academia and research institutes”