- Email: [email protected]
Effect of carbon on the nitridation behavior of aluminum powder
Journal of Alloys and Compounds 689 (2016) 218e224
Contents lists available at ScienceDirect
Journal of Alloys and Compounds journal homepage: http://www.elsevier.com/locate/jalcom
Effect of carbon on the nitridation behavior of aluminum powder Kon-Bae Lee a, Yong Hwan Kim b, Hyun Joo Choi a, *, Jae-Pyoung Ahn c, ** a
School of Advanced Materials Engineering, Kookmin University, 77 Jeongneung-ro, Seongbuk-gu, Seoul 136-702, Republic of Korea Korea Institute of Industrial Technology (KITECH), 156 Gaetbeol-ro, Yeonsu, Incheon, 406-840, Republic of Korea c Advanced Analysis Center, Korea Institute of Science and Technology (KIST), 39-1 Hawolgok-dong, Wolsong-gil 5, Seongbuk-gu, Seoul 130-650, Republic of Korea b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 24 February 2016 Received in revised form 14 June 2016 Accepted 11 July 2016 Available online 14 July 2016
The effect of carbon on Al powder nitridation was investigated from an atomic-scale chemical and crystallographic viewpoint. Carbon delays the initiation of nitridation and melting of Al powder, possibly because of its insulation effect. Therefore, carbon provides a pathway through which a constant supply of nitrogen gas is provided before molten Al powder is fully consolidated. Carbon also alters the nitridation mechanism by acting as a reaction agent of aluminum oxides and providing the path for nitrogen gas for constant nitridation, thereby stimulating the formation of the AlN with a typical triplet band structure. As a result, when carbon covers the entire surface of the powder, about 85% nitridation can be achieved. © 2016 Elsevier B.V. All rights reserved.
Keywords: Nitridation Aluminum Powder Carbon EELS TEM
1. Introduction Aluminum nitride has high thermal conductivity (320 W/m K, ten times higher than that of Al2O3), high electrical insulation (9 1013 U cm), low thermal expansion coefficient (4 106/ C, close to that of silicon), and excellent mechanical and chemical stability [1e12]. As such, aluminum nitride is widely used as a thermal barrier material for semiconductors and compound semiconductor substrates and as a reinforcing phase in composite materials. The application scope of this material has recently been broadened to heat-dissipating materials in light-emitting devices (LEDs) owing to its high thermal conductivity and low thermal expansion coefficient [1e12]. Two methods, the (i) carbothermal reaction and nitridation of alumina with a reaction agent (carbon) and (ii) direct nitridation of Al powder in a nitrogen atmosphere, have been extensively used for the commercial fabrication of aluminum nitride (AlN) powder [13e18]. The typical precursors of carbothermal reaction are aluminum oxide (Al2O3) and a source of carbon serving as a reaction
* Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (J.-P. Ahn). http://dx.doi.org/10.1016/j.jallcom.2016.07.109 0925-8388/© 2016 Elsevier B.V. All rights reserved.
(H.J.
Choi),
[email protected]
agent. However, this process is limited by the surface areas of the precursors and requires an excess amount of carbon (~15%e30% additional carbon) to complete the nitridation. Hence, it is costly, energy intensive, and emits a significant amount of carbon-based byproducts (CO and CO2) leading to a severe environmental impact. Preparing a homogeneous mixture of precursors is also essential in accomplishing the successful synthesis of AlN [15e18]. In comparison, the direct nitridation method uses cheaper starting materials and a lower manufacturing temperature as well as involves a simpler process than that associated with the carbothermal reaction. The direct nitridation reaction of aluminum powder proceeds in accordance with Equation (1) and is thermodynamically possible at both room and elevated temperatures [19e22]. Al (l) þ 1/2N2 (g) ¼ AlN (s)
(1)
This reaction is highly exothermic and generates a considerable amount of heat. As a result, unreacted aluminum (melted by this reaction heat) consolidates, which then impedes further nitridation by blocking the diffusion pathways that supply nitrogen gas. To prevent this blockage, commercial direct nitridation methods employ prolonged heating at elevated temperatures, of 1000 C2000 C, leading to the complete nitridation of the consolidated aluminum. To increase the overall yield, the resulting aluminum nitride must also be subjected to repeated nitridation and
K.-B. Lee et al. / Journal of Alloys and Compounds 689 (2016) 218e224
pulverization. Otherwise, further processing steps, such as the addition of aluminum tri-fluoride (AlF3) or AlN, are required to facilitate the reaction completion [13]. These processing steps may, however, have negative side effects. For instance, during the pulverization process, the fraction of impurities (e.g., oxygen) may increase and have a negative impact on the thermal conductivity. These additional steps may also increase the overall cost of manufacturing. Recently, the addition of AlN powder or other additives (e.g., carbon) have been reported to prevent the coalescence of
219
unreacted aluminum, thereby greatly enhancing the degree of nitridation at lower temperatures during the process [3e6,23,24]. However, the effects of these additives on the nitridation behavior have yet to be clearly examined. As such, in the current study, we used relatively cheap lamp carbon (LC) as an additive to suppress the coalescence of unreacted aluminum. The new facile process proposed in this study facilitates a higher degree of nitridation at a relatively low temperature, which has yet to be achieved by the aforementioned processes, although it may generate reaction
Fig. 1. SEM images of (a) aluminum and (b) Al and (c) LC powder at (b) low and (c) high magnification.
Fig. 2. SEM images of LC-free and Al-3 wt% powder beds after heat treatment. (a), (c), and (e) show LC-free powder beds after heating for 20, 30, and 60 min, respectively; (b), (d), and (f) show Al-3wt% LC powder beds after heating for 20, 30, and 60 min, respectively; insets show EDS spectra obtained from the corresponding area.
220
K.-B. Lee et al. / Journal of Alloys and Compounds 689 (2016) 218e224
byproducts (e.g., Al4C3) that can degrade the purity of AlN. Furthermore, this study examined the role of carbon in inducing the thermal and structural changes of Al during nitridation. For the systematic investigation, variations of the microstructure, phase, and thermal behavior at each reaction step were thoroughly studied from an atomic-scale chemical/crystallographic viewpoint. The investigation was conducted by using in situ thermal analysis, in situ X-ray diffraction (XRD), and transmission electron microscopy (TEM) combined with electron energy loss spectroscopy (EELS). 2. Material and methods Commercial atomized Al powder (average particle diameter: 9.4 mm; Metal Player Co. Ltd., Korea) was used as the starting material (Fig. 1). To examine the effect of LC addition, 1e15 wt% of LC powder (average particle diameter: 55 mm, aggregate sizes; Fischer Co. Ltd., Korea) was mixed with the Al powder using a Turbula mixer for 40 min. The surface areas of the particles were 0.64 and 197.3 m2/g, respectively, as determined via BrunauereEmmetteTeller (BET) analysis. Mixed powder (40 g) was placed in a graphite crucible (3 3 3 cm3). The initial thickness of Al2O3 was measured to about 10 nm by TEM and AES analyses [25]. The assembly was heated to 700 C in a retort furnace and held at this temperature for 10 mine60 min under flowing nitrogen (5000 cc/ min). To prevent additional nitridation during cooling, the weight gain was measured after air-cooling the crucible outside the furnace (i.e., subsequent to each heat treatment). The efficiency of nitridation was macroscopically measured based on the weight change of the crucible during heat-treatment. Efficiency was also microscopically examined based on the weight gain obtained through the thermogravimetry and differential scanning calorimetry (TG/DSC) measurement. The TG/DSC measurement was performed under flowing nitrogen (100 cm3/min) at a heating rate of 4 C/min and under temperatures of 700 C1000 C for up to 1 h, by using a thermal analyzer. Next, the heatflow and the weight change of a powder bed containing 0e15 wt% of LC were examined. Furthermore, the phase-change behavior of the powder was investigated via room-temperature X-ray diffraction (XRD) analysis and in situ high-temperature XRD analysis, which were performed while the sample was heated to 800 C. The reaction products and the corresponding microstructures were examined by means of scanning electron microscopy (SEM) and TEM. The microstructures of the LC-free and 3 wt% LC-containing powder beds after undergoing heating for 20, 30, and 60 min at 700 C were examined by SEM, while the microstructures of those that underwent heating for 60 min at 700 C were analyzed by TEM. Cross-sectioned specimens for TEM examination were prepared via focused ion beam milling (FIB), which was performed using a Gaþion source. The TEM examination was performed at an operating voltage of 300 kV. Thin-window energy-dispersive X-ray spectroscopy (EDS) and EELS were also performed during this examination.
consisted of a higher number of reaction products than the top portion, which remained in powder form. However, the bed was fully consolidated after 30 min of heat treatment, which also resulted in an increase in the number of reaction products. The polished surfaces of the powder beds heat-treated for 60 min are shown in Fig. 2 (e) and (f). EDS analysis in the insets in Fig. 2 revealed that the surfaces of the Al powder particles of both the carbon-containing and carbon-free powder beds were covered by a dark-gray (i.e., AlN) phase. The AlN content of the carbon-containing bed increased significantly, in contrast to that of its carbon-free counterpart. The complete transformation of Al to AlN resulted in a theoretical weight gain of ~52%. The weight gains of the LC-free and LC-containing powder beds increased from 0.95%e4.25% and 0.52%e12.26%, respectively, with increasing heat treatment time of 20e60 min. Next, the heat flow and the weight change of a powder bed with different LC contents were examined to quantitatively determine the effect of carbon content on the nitridation of Al powder. The bed was heat-treated at 700 C-1000 C, for 1 h in a nitrogen atmosphere. The corresponding results are summarized in Table 1 and Fig. 3. As shown in Table 1, “Weight gain at” indicates the weight gain at which the temperature reached the target temperature, and “Weight gain after” indicates the weight gain after heating of 1 h at the target temperature. As can be seen from Table 1, the addition of carbon had a significant effect on the nitridation behavior of the Al powder. For
3. Results and discussion The SEM images of the LC-free [Fig. 2 (a), (c), and (e)) and LCcontaining powder beds (Fig. 2 (b), (d), and (f)] are shown in Fig. 2. The powder beds were heat-treated for 20 [Fig. 2 (a) and (b)], 30 [Fig. 2 (c) and (d)], and 60 min [Fig. 2 (e) and (f)] at 700 C. As previously mentioned, additional nitridation during cooling was suppressed by air-cooling the powder bed outside the furnace after the respective heat treatments. During the initial stage of reaction (after 20 min of heat treatment), island-type reaction products formed on the surface of the Al powder of both powder beds; meanwhile, significant amounts of un-reacted Al were retained. Furthermore, the bottom portion of the powder bed became consolidated and
Fig. 3. Weight gain as a function of LC content. Weight gain was measured (a) during and (b) after heating at different temperatures. Data are also listed in Table 1.
K.-B. Lee et al. / Journal of Alloys and Compounds 689 (2016) 218e224
221
Table 1 Reaction temperature and weight change of powder beds containing 0e15 wt% of LC. LC content (%)
Set temperature ( C)
Endothermic reaction temperature ( C)
Reaction start temperature ( C)
Exothermic reaction temperature ( C)
Weight gain at (%)
Weight gain after (%)
0
700 800 1000 800 1000 700 800 1000 800 1000 800 1000 800 1000 800 1000
667.5 669.8 670.1 677.6 675 669.5 671.5 673.1 672.1 673.8 672.2 672.5 672.6 672.7 671.1 671.8
642 650 639 690 683 680 680 675 717 740 710 705 718 710 720 722
677.7 690 682 778.3 784.8 703 766.2 756.7 772.2 817.2 770.4 768 773.3 779.3 784.1 787.2
1.92 2.4 4.95 10.92 15.58 0.91 45.64 46.44 42.48 38.95 41.78 44.08 42.56 42.23 35.41 32.94
2.22 2.57 7.84 12.14 24.15 44.65 45.65 46.74 42.49 40.59 41.97 44.62 42.58 42.95 35.73 34.59
1 3
5 7 10 15
Note: All samples had the same size, and the results were obtained from a single experimental run.
example, the temperature at which nitridation and exothermic reactions began increased with increasing LC content. Carbon effectively delayed the nitridation and melting of the powder, possibly owing to its insulation effect. The powder was initially covered by LC, which had a low thermal conductivity (<1 W/m C at 1000 C [26]). Therefore, LC may act as a barrier preventing heat transfer from the atmosphere to the powder. Despite the aforementioned delay, however, the degree of nitridation increased continuously with increasing LC content. Furthermore, the carbonfree Al powder experienced insignificant weight gains of only 1.9% and 2.2%, when the temperature reached 700 C and after 1 h of heat-treatment at this temperature, respectively. The powder containing 3 wt% of LC experienced weight gain of 0.9% upon reaching 700 C and weight gains of 43.4% and 44.7% after 10 min and 1 h of heat-treatment at this temperature, respectively. Extensive reaction, which seemed to occur at the beginning of the heat treatment, was almost complete after 10 min, resulting in negligible weight gain (1.3%) during the final 50 min. Therefore, intensive nitridation was completed within a short time frame. The total weight gain at 700 C increased ~20-fold (from 2.27% to 44.7%) with the addition of only 3% of LC, indicating that ~85% of aluminum was transformed to aluminum nitrides. Furthermore, for a given amount of LC, the degree of nitridation increased with increasing temperature, although the temperature itself had a negligible influence on this increase. Meanwhile, the weight gain decreased with increasing LC content of >3 wt%, suggesting that the carbon content must be optimized in order to achieve maximum nitridation. Extensive physical contact between the LC and Al powders is essential for maximizing the effect of carbon on the nitridation of Al. However, the LC used in this study tended to agglomerate; hence, it only partly covered the surface of the powder. The BET analysis yielded surface areas of 0.64 and 192.27 m2/g for the Al and LC, respectively, corresponding to the respective total surface areas of 25.34 and 76.90 m2 in 40 g of a powder containing 1 wt% of LC. Given that LC tends to agglomerate, an amount that is significantly greater than the theoretical value is required to achieve the complete surface coverage of the Al powder. In fact, excessive carbon has been added to Al to form AlN via the commercial carbothermal method [13]. Therefore, carbon may only partially cover the surface of the Al powder containing 1 wt% of LC. At the same time, incomplete nitridation may also occur when a temperature of 1000 C is reached. In the current study, this phenomenon resulted in additional weight gain (from 15.6% to 24.2%) after heat treatment for 1 h at this temperature. However, the surface of Al powders containing >3 wt% of LC was already
completely covered by carbon; hence, the degree of nitridation varied only slightly with the LC content or heat-treatment time. These results are analogous to the observation that nitridation is significantly enhanced with the addition of carbon (Fig. 2). The representative TG and DSC curves of both samples heated at 700 C for 1 h in a nitrogen atmosphere are shown in Fig. 4. These curves reveal the difference in the nitridation behaviors (i.e., weight gain) of the samples. In the case of pure Al, nitridation in the solid state began at temperatures lower than the melting point, whereas
Fig. 4. TG and DSC curves of (a) pure Al and (b) Al-3 wt% LC powder beds when the setting temperature was 700 C.
222
K.-B. Lee et al. / Journal of Alloys and Compounds 689 (2016) 218e224
Fig. 5. In situ XRD patterns of (a) pure Al and (b) Al-3 wt% LC powder bed. Samples were heated to 800 C under nitrogen atmosphere, after which the spectra were obtained in the temperature ranging from room temperature to 800 C.
nitridation occurred over the melting point of Al in the case of the LCcontaining powder. In the case of the sample fabricated from the pure Al powder, the reaction rate increased slowly and then rapidly at temperatures near the melting point of Al. The rate decreased significantly at temperatures higher than the melting point. The corresponding rate of conversion of Al to AlN was ~4.2%. In the case of the LC-containing powder, the weight increased slowly with increasing temperature. However, the weight increased exponentially when the Al melted, and the reaction stopped abruptly once the nitridation process was complete. Furthermore, the corresponding rate of conversion (~85.9%) of Al to AlN was 20 times higher than that of the pure Al powder. The addition of carbon particles was, therefore, highly effective in improving the nitridation efficiency of the Al particles. This was confirmed by the in situ XRD analysis of the sample heated to 800 C under a nitrogen atmosphere.
The results in Fig. 5 showed that most of the peaks observed at lower temperatures were no longer observed when the pure Al powder was heated to 700 C. In fact, only a single peak occurred, i.e., an extremely high-intensity peak corresponding to the amorphous phase resulting from the melting of Al. Distinct AlN-related peaks appeared at 800 C. However, the Al-3 wt% LC powder mixture exhibited sharp AlN-related peaks, even at 700 C, indicating that the addition of carbon had an effect on the nitridation of the Al particles. The formation of the AlN phase for each sample was also confirmed via XRD analysis, as shown in Fig. 6. In accordance with Equation (1), Al powder reacted readily with nitrogen gas to form AlN. Thermodynamic calculations performed by using data from the NIST-JANAF tables [20], revealed that reaction (1) is highly exothermic And can occur at 700 C (Fig. 7). The particles of most metallic powders, including those of Al, become coated with a
Fig. 6. XRD patterns of pure Al and Al-3 wt% LC powder beds after heat treatment at 700 C for various times.
Fig. 7. Thermodynamical changes in the Gibbs free energy (DG) as a function of temperature. Thermodynamic calculations were performed by using data from the NIST-JANAF tables [20].
K.-B. Lee et al. / Journal of Alloys and Compounds 689 (2016) 218e224
surface oxide layer during exposure to air. Therefore, during exposure to nitrogen gas, the coating of the Al particles with a layer of their native oxide can be expected [27,28]. Accordingly, reaction (2) may be favored over reaction (1). 2Al2O3 (s) þ 2N2 (g) ¼ 4AlN (s) þ 3O2 (g)
(2)
Thermodynamic calculations (Fig. 7) revealed that reaction (2) did not occur at 700 C; hence, the removal of the oxide layer proved to be essential in the successful nitridation of the Al powder. In a previous study, we employed a new mechanism that yielded AlN through the direct nitridation of pure Al particles, along with a layer of AlON formed on the surfaces of the particles; this layer was able to provide catalytic sites that can supply nitrogen atoms [25].
In the current study, however, the unreacted liquid Al (formed during the nitridation) coalesced within the powder bed and the nitrogen diffusion paths were consequently blocked; hence, the nitridation process was retarded. This resulted in a very low nitridation ratio, as shown in Table 1. The thermodynamic calculations (Fig. 7) revealed that nitridation can occur at temperatures lower than 1700 C; therefore, reaction (3) may be expected, in the case of the carbon-containing Al powder bed. During nitridation, the temperature of the powder bed was measured with a thermocouple inserted directly into the bed. As shown in Fig. 8, the temperature increased to a maximum of 1732 C, which was significantly higher than the temperature difference (1032 C) between the furnace chamber and the powder bed. This temperature (1732 C) was also significantly higher than the highest temperature (1290 C) measured in the carbon-free Al powder bed. The degree of nitridation of the carbon-containing Al powder bed was, therefore, higher than that of the carbon-free bed. Al2O3 (s) þ 3C (s) þ N2 (g) ¼ 2AlN (s) þ 3CO (g)
Fig. 8. Temperature difference between the furnace chamber and powder beds.
223
(3)
In addition, the carbon powder may act as a barrier that prevents the coalescence of the liquid Al (as reported in previous studies on AlN particles [23,24]). This possibly leads to a constant nitrogen supply and an enhanced nitridation process. Therefore, the carbon added to the Al powder acted as both a reaction agent for the native oxide and a dispersing agent that prevented the blockage of the nitrogen diffusion path. A few studies have focused on the synthesis of AlN from different carbon sources [3e6]. However, the lowtemperature (i.e., temperatures lower than 1000 C) nitridation behavior of low-carbon Al has barely been investigated. TEM images and EELS spectra of the pure Al and Al-3 wt% LC powder beds after nitridation at 700 C for 60 min are shown in Fig. 9. The reaction product, AlN (not shown here), was identified via EDS and fast Fourier transform (FFT) analyses. The samples
Fig. 9. (a) Bright-field TEM image and (c) EELS spectra of pure Al, and (b) STEM image and (d) EELS spectra of Al-3 wt% LC powder beds.
224
K.-B. Lee et al. / Journal of Alloys and Compounds 689 (2016) 218e224
exhibited considerably different characteristic N K-edges of AlN, as revealed by the high-loss EELS spectra shown in Fig. 9 (c) and (d). For example, in the case of the carbon-free Al bed [Fig. 9 (c)], the first of the three N K-edge peaks corresponded to the outermost shell; this peak had the highest intensity, consistent with the formation of AlON. The N K-edge corresponding to the region just beneath the AlON layer can be attributed to the typical triplet band structure of AlN. The EELS data corresponding to the AlN and AlON layers concur with those shown in previous studies [29e34]. In contrast to the reaction product associated with the carbon-free bed, the product associated with the Al-3 wt% LC bed exhibited only the aforementioned typical triplet band structure over the entire region, thereby confirming that the added carbon acted as a reaction agent for the native oxide of Al. 4. Conclusions The effect of carbon on the nitridation of Al powder was investigated from a microstructural and thermodynamical viewpoint. With the addition of carbon, nitridation occurred at a relatively low temperature of 700 C, which led to an exothermic reaction and a temperature increase by up to 1700 C. This facilitated the removal of the oxide layer and promoted the nitridation of the powder. In other words, carbon acted as a reaction agent of the native oxides on the surface of the Al particles. Furthermore, carbon successfully inhibited the consolidation of molten Al, thereby providing a pathway for a constant supply of nitrogen gas and an increase in the degree of nitridation. Carbon also acted as a dispersing agent (diluent) that prevented the consolidation of the Al powder. Therefore, carbon plays a dual role (i.e., of reaction and dispersing agent) when added in an optimal amount that uniformly covers the entire surface of the Al powder. The TEM and EELS analyses revealed that, depending on the carbon content, two types of AlN (with differing band structures) were formed on the surface of the Al particles, i.e., AlN with a typical triplet band structure was formed when the bed contained carbon, whereas the carbon-free bed gave rise to the formation of AlN covered by an AlON layer.
[7]
[8] [9]
[10]
[11]
[12] [13] [14]
[15]
[16]
[17]
[18]
[19] [20] [21] [22]
[23] [24] [25]
[26]
Acknowledgments [27]
This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (Grant Nos. 2014R1A1A2057846 and 2009-0093814). References [1] Y. Zhang, J. Liu, R. He, Q. Zhang, X. Zhang, J. Zhu, Synthesis of aluminum nitride nanowires from carbon nanotubes, Chem. Mater 13 (2001) 3899e3905. [2] B.S.S. Daniel, V.S.R. Murthy, Directed melt oxidation and nitridation of aluminum alloys: a comparison, Mater. Des. 76 (1995) 155e161. [3] R.-C. Juang, C.-J. Lee, C.-C. Chen, Combustion synthesis of hexagonal aluminum nitride powders under low nitrogen pressure, Mater. Sci. Eng. A Struct. 357 (2003) 219e227. [4] H. Wang, J. Han, Z. Li, S. Du, Effect of additives on self-propagating hightemperature synthesis of AlN, J. Eur. Ceram. Soc. 21 (2001) 2193e2198. [5] W. Zhang, Z. Li, D. Zhang, Synthesizing AlN powder by mechanochemical reaction between aluminum and melamine, J. Mater. Res. 25 (2010) 464e470. [6] M. Qin, X. Du, J. Wang, I.S. Humail, X. Qu, Influence of carbon on the synthesis
[28]
[29] [30]
[31] [32] [33] [34]
of AlN powder from combustion synthesis precursors, J. Eur. Ceram. Soc. 29 (2009) 795e799. M. Balog, P. Yu, M. Qian, M. Behulova, P. Svec Sr, R. Cicka, Nanoscaled AleAlN composites consolidated by equal channel angular pressing (ECAP) of partially in situ nitride Al powder, Mater. Sci. Eng. A Struct. 562 (2013) 190e195. P. Yu, M. Balog, M. Yan, G.B. Schaffer, M. Qian, Powder injection moulding of an AleAlN metal matrix composite, J. Mater. Res. 26 (2011) 1719e1725. Q. Zhang, G. Chen, G. Wu, Z. Xiu, B. Luan, Property characteristics of a AlNp/Al composite fabricated by squeeze casting technology, Mater. Lett. 57 (2003) 1453e1458. Y.Q. Liu, H.T. Cong, W. Wang, C.H. Sun, H.M. Cheng, AlN nanoparticlereinforced nanocrystalline Al matrix composites: fabrication and mechanical properties, Mater. Sci. Eng. A Struct. 505 (2009) 151e156. K. Vijayalakshmi, V.V. Pillay, Influence of deposition time on the microstructure and dielectric properties of AlN/Si thin films for enhanced hydrogen sensing application, Microelectron. Eng. 113 (2014) 29e34. C.R. Rambo, D. Hotza, Microchanneled biomorphic AlN-coated Al2O3 by pressureless infiltrationenitridation, Ceram. Int. 40 (2014) 12567e12571. G. Selvaduray, L. Sheet, Aluminium nitride: review of synthesis methods, Mater. Sci. Technol. 9 (1993) 463e473. J. Haibo, K. Chen, Z. Heping, S. Agathopoulos, O. Fabrichnaya, J.M.F. Ferreira, Direct nitridation of molten Al(Mg,Si) alloy to AlN, J. Cryst. Growth 281 (2005) 639e645. E. Kroke, L. Loeffler, F.F. Lange, R. Riedel, Aluminum nitride prepared by nitridation of aluminum oxide precursors, J. Am. Ceram. Soc. 85 (2002) 3117e3119. T. Suehiro, N. Hirosaki, R. Terao, J. Tatami, T. Meguro, K. Komeya, Synthesis of aluminum nitride nanopowder by gas-reduction-nitridation method, J. Am. Ceram. Soc. (2003) 1046e1048. T. Yamakawa, J. Tatami, K. Komeya, T. Meguro, Synthesis of AlN powder from Al(OH)3 by reduction-nitridation in a mixture of NH3-C3H8, J. Eur. Ceram. Soc. 26 (2006) 2413e2418. A. Chu, M. Qin, Rafi-ud-din, B. Jia, H. Lua, X. Qu, Effect of urea on the size and morphology of AlN nanoparticles synthesized from combustion synthesis precursors, J. Alloy. Comp. 530 (2012) 144e151. T. Okada, M. Toriyama, S. Kanzaki, Direct nitridation of aluminum compacts at low temperature, J. Mater. Sci. 35 (2000) 3105e3111. http://kinetics.nist.gov/janaf/. (accessed 31.10.14).. Q. Zheng, B. Wu, R.G. Reddy, In-situ processing of Al alloy composites, Adv. Eng. Mater 5 (2003) 167e172. A.V. Korshunov, Influence of dispersion aluminum powders on the regularities of their interaction with nitrogen, Russ. J. Phys. Chem. A 85 (2011) 1202e1210. S. D. Dunmead, W. G. Moore, K. E. Howard, K. C. Morse, T. A. Guiton, US patent 5,710,382, 1998. H. Miura, H. Matunaga, US patent 7,022,301 B2. S.H. Kim, J.H. Noh, J.P. Ahn, J.C. Lee, H. Kwon, J.G. Lee, K.B. Lee, Effects of surface oxide on the nitridation behavior of aluminum particles, Metall. Mater. Trans. A 46 (2015) 496e504. P.E. Khizhnyak, A.V. Chechetkin, A.P. Glybin, Thermal conductivity of carbon black, J. Eng. Phys. 37 (1979) 475e478. G.B. Schaffer, B.J. Hall, The influence of the atmosphere on the sintering of aluminum, Metall. Mater. Trans. A 33A (2002) 3279e3284. R.N. Lumley, T.B. Sercombe, G.B. Schaffer, Surface oxide and the role of magnesium during the sintering of aluminum, Metall. Mater. Trans. A 30A (1999) 457e463. M. MacKenzie, A.J. Craven, Quantifying the oxidation of AlN using electron energy loss spectroscopy, J. Phys. D. Appl. Phys. 33 (2000) 1647e1655. V. Serin, C. Colliex, R. Brydson, S. Matar, F. Boucher, EELS investigation of the electron conduction-band states in wurtzite AlN and oxygen-doped AlN(O), Phys. Rev. B 58 (1998) 5106e5115. P.B. Guillemaud, G. Radtke, M. Sennour, Electron spectroscopy imaging to study ELNES at a nanoscale, J. Microsc. 210 (2003) 66e73. S.P. Gao, A. Zhang, J. Zhu, J. Yuan, Anisotropic spectroscopy of nitrogen K-edge in group-III nitrides, Appl. Phys. Lett. 84 (2004) 2784e2786. G. Radtke, T. Epicier, P.B. Guillemaud, J.C.L. Bosse, N-K ELNES study of anisotropy effects in hexagonal AlN, J. Microsc. 210 (2003) 60e65. K.L. Jablonska, T. Suski, I. Gorczyca, N.E. Christensen, R.C.C. Perera, E.M. Gullikson, J.H. Underwood, D.L. Ederer, Z.L. Weber, Electronic states in valence and conduction bands of group-III nitrides: experiment and theory, Phys. Rev. B 61 (2000) 16623e16632.
Contents lists available at ScienceDirect
Journal of Alloys and Compounds journal homepage: http://www.elsevier.com/locate/jalcom
Effect of carbon on the nitridation behavior of aluminum powder Kon-Bae Lee a, Yong Hwan Kim b, Hyun Joo Choi a, *, Jae-Pyoung Ahn c, ** a
School of Advanced Materials Engineering, Kookmin University, 77 Jeongneung-ro, Seongbuk-gu, Seoul 136-702, Republic of Korea Korea Institute of Industrial Technology (KITECH), 156 Gaetbeol-ro, Yeonsu, Incheon, 406-840, Republic of Korea c Advanced Analysis Center, Korea Institute of Science and Technology (KIST), 39-1 Hawolgok-dong, Wolsong-gil 5, Seongbuk-gu, Seoul 130-650, Republic of Korea b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 24 February 2016 Received in revised form 14 June 2016 Accepted 11 July 2016 Available online 14 July 2016
The effect of carbon on Al powder nitridation was investigated from an atomic-scale chemical and crystallographic viewpoint. Carbon delays the initiation of nitridation and melting of Al powder, possibly because of its insulation effect. Therefore, carbon provides a pathway through which a constant supply of nitrogen gas is provided before molten Al powder is fully consolidated. Carbon also alters the nitridation mechanism by acting as a reaction agent of aluminum oxides and providing the path for nitrogen gas for constant nitridation, thereby stimulating the formation of the AlN with a typical triplet band structure. As a result, when carbon covers the entire surface of the powder, about 85% nitridation can be achieved. © 2016 Elsevier B.V. All rights reserved.
Keywords: Nitridation Aluminum Powder Carbon EELS TEM
1. Introduction Aluminum nitride has high thermal conductivity (320 W/m K, ten times higher than that of Al2O3), high electrical insulation (9 1013 U cm), low thermal expansion coefficient (4 106/ C, close to that of silicon), and excellent mechanical and chemical stability [1e12]. As such, aluminum nitride is widely used as a thermal barrier material for semiconductors and compound semiconductor substrates and as a reinforcing phase in composite materials. The application scope of this material has recently been broadened to heat-dissipating materials in light-emitting devices (LEDs) owing to its high thermal conductivity and low thermal expansion coefficient [1e12]. Two methods, the (i) carbothermal reaction and nitridation of alumina with a reaction agent (carbon) and (ii) direct nitridation of Al powder in a nitrogen atmosphere, have been extensively used for the commercial fabrication of aluminum nitride (AlN) powder [13e18]. The typical precursors of carbothermal reaction are aluminum oxide (Al2O3) and a source of carbon serving as a reaction
* Corresponding author. ** Corresponding author. E-mail addresses: [email protected] (J.-P. Ahn). http://dx.doi.org/10.1016/j.jallcom.2016.07.109 0925-8388/© 2016 Elsevier B.V. All rights reserved.
(H.J.
Choi),
[email protected]
agent. However, this process is limited by the surface areas of the precursors and requires an excess amount of carbon (~15%e30% additional carbon) to complete the nitridation. Hence, it is costly, energy intensive, and emits a significant amount of carbon-based byproducts (CO and CO2) leading to a severe environmental impact. Preparing a homogeneous mixture of precursors is also essential in accomplishing the successful synthesis of AlN [15e18]. In comparison, the direct nitridation method uses cheaper starting materials and a lower manufacturing temperature as well as involves a simpler process than that associated with the carbothermal reaction. The direct nitridation reaction of aluminum powder proceeds in accordance with Equation (1) and is thermodynamically possible at both room and elevated temperatures [19e22]. Al (l) þ 1/2N2 (g) ¼ AlN (s)
(1)
This reaction is highly exothermic and generates a considerable amount of heat. As a result, unreacted aluminum (melted by this reaction heat) consolidates, which then impedes further nitridation by blocking the diffusion pathways that supply nitrogen gas. To prevent this blockage, commercial direct nitridation methods employ prolonged heating at elevated temperatures, of 1000 C2000 C, leading to the complete nitridation of the consolidated aluminum. To increase the overall yield, the resulting aluminum nitride must also be subjected to repeated nitridation and
K.-B. Lee et al. / Journal of Alloys and Compounds 689 (2016) 218e224
pulverization. Otherwise, further processing steps, such as the addition of aluminum tri-fluoride (AlF3) or AlN, are required to facilitate the reaction completion [13]. These processing steps may, however, have negative side effects. For instance, during the pulverization process, the fraction of impurities (e.g., oxygen) may increase and have a negative impact on the thermal conductivity. These additional steps may also increase the overall cost of manufacturing. Recently, the addition of AlN powder or other additives (e.g., carbon) have been reported to prevent the coalescence of
219
unreacted aluminum, thereby greatly enhancing the degree of nitridation at lower temperatures during the process [3e6,23,24]. However, the effects of these additives on the nitridation behavior have yet to be clearly examined. As such, in the current study, we used relatively cheap lamp carbon (LC) as an additive to suppress the coalescence of unreacted aluminum. The new facile process proposed in this study facilitates a higher degree of nitridation at a relatively low temperature, which has yet to be achieved by the aforementioned processes, although it may generate reaction
Fig. 1. SEM images of (a) aluminum and (b) Al and (c) LC powder at (b) low and (c) high magnification.
Fig. 2. SEM images of LC-free and Al-3 wt% powder beds after heat treatment. (a), (c), and (e) show LC-free powder beds after heating for 20, 30, and 60 min, respectively; (b), (d), and (f) show Al-3wt% LC powder beds after heating for 20, 30, and 60 min, respectively; insets show EDS spectra obtained from the corresponding area.
220
K.-B. Lee et al. / Journal of Alloys and Compounds 689 (2016) 218e224
byproducts (e.g., Al4C3) that can degrade the purity of AlN. Furthermore, this study examined the role of carbon in inducing the thermal and structural changes of Al during nitridation. For the systematic investigation, variations of the microstructure, phase, and thermal behavior at each reaction step were thoroughly studied from an atomic-scale chemical/crystallographic viewpoint. The investigation was conducted by using in situ thermal analysis, in situ X-ray diffraction (XRD), and transmission electron microscopy (TEM) combined with electron energy loss spectroscopy (EELS). 2. Material and methods Commercial atomized Al powder (average particle diameter: 9.4 mm; Metal Player Co. Ltd., Korea) was used as the starting material (Fig. 1). To examine the effect of LC addition, 1e15 wt% of LC powder (average particle diameter: 55 mm, aggregate sizes; Fischer Co. Ltd., Korea) was mixed with the Al powder using a Turbula mixer for 40 min. The surface areas of the particles were 0.64 and 197.3 m2/g, respectively, as determined via BrunauereEmmetteTeller (BET) analysis. Mixed powder (40 g) was placed in a graphite crucible (3 3 3 cm3). The initial thickness of Al2O3 was measured to about 10 nm by TEM and AES analyses [25]. The assembly was heated to 700 C in a retort furnace and held at this temperature for 10 mine60 min under flowing nitrogen (5000 cc/ min). To prevent additional nitridation during cooling, the weight gain was measured after air-cooling the crucible outside the furnace (i.e., subsequent to each heat treatment). The efficiency of nitridation was macroscopically measured based on the weight change of the crucible during heat-treatment. Efficiency was also microscopically examined based on the weight gain obtained through the thermogravimetry and differential scanning calorimetry (TG/DSC) measurement. The TG/DSC measurement was performed under flowing nitrogen (100 cm3/min) at a heating rate of 4 C/min and under temperatures of 700 C1000 C for up to 1 h, by using a thermal analyzer. Next, the heatflow and the weight change of a powder bed containing 0e15 wt% of LC were examined. Furthermore, the phase-change behavior of the powder was investigated via room-temperature X-ray diffraction (XRD) analysis and in situ high-temperature XRD analysis, which were performed while the sample was heated to 800 C. The reaction products and the corresponding microstructures were examined by means of scanning electron microscopy (SEM) and TEM. The microstructures of the LC-free and 3 wt% LC-containing powder beds after undergoing heating for 20, 30, and 60 min at 700 C were examined by SEM, while the microstructures of those that underwent heating for 60 min at 700 C were analyzed by TEM. Cross-sectioned specimens for TEM examination were prepared via focused ion beam milling (FIB), which was performed using a Gaþion source. The TEM examination was performed at an operating voltage of 300 kV. Thin-window energy-dispersive X-ray spectroscopy (EDS) and EELS were also performed during this examination.
consisted of a higher number of reaction products than the top portion, which remained in powder form. However, the bed was fully consolidated after 30 min of heat treatment, which also resulted in an increase in the number of reaction products. The polished surfaces of the powder beds heat-treated for 60 min are shown in Fig. 2 (e) and (f). EDS analysis in the insets in Fig. 2 revealed that the surfaces of the Al powder particles of both the carbon-containing and carbon-free powder beds were covered by a dark-gray (i.e., AlN) phase. The AlN content of the carbon-containing bed increased significantly, in contrast to that of its carbon-free counterpart. The complete transformation of Al to AlN resulted in a theoretical weight gain of ~52%. The weight gains of the LC-free and LC-containing powder beds increased from 0.95%e4.25% and 0.52%e12.26%, respectively, with increasing heat treatment time of 20e60 min. Next, the heat flow and the weight change of a powder bed with different LC contents were examined to quantitatively determine the effect of carbon content on the nitridation of Al powder. The bed was heat-treated at 700 C-1000 C, for 1 h in a nitrogen atmosphere. The corresponding results are summarized in Table 1 and Fig. 3. As shown in Table 1, “Weight gain at” indicates the weight gain at which the temperature reached the target temperature, and “Weight gain after” indicates the weight gain after heating of 1 h at the target temperature. As can be seen from Table 1, the addition of carbon had a significant effect on the nitridation behavior of the Al powder. For
3. Results and discussion The SEM images of the LC-free [Fig. 2 (a), (c), and (e)) and LCcontaining powder beds (Fig. 2 (b), (d), and (f)] are shown in Fig. 2. The powder beds were heat-treated for 20 [Fig. 2 (a) and (b)], 30 [Fig. 2 (c) and (d)], and 60 min [Fig. 2 (e) and (f)] at 700 C. As previously mentioned, additional nitridation during cooling was suppressed by air-cooling the powder bed outside the furnace after the respective heat treatments. During the initial stage of reaction (after 20 min of heat treatment), island-type reaction products formed on the surface of the Al powder of both powder beds; meanwhile, significant amounts of un-reacted Al were retained. Furthermore, the bottom portion of the powder bed became consolidated and
Fig. 3. Weight gain as a function of LC content. Weight gain was measured (a) during and (b) after heating at different temperatures. Data are also listed in Table 1.
K.-B. Lee et al. / Journal of Alloys and Compounds 689 (2016) 218e224
221
Table 1 Reaction temperature and weight change of powder beds containing 0e15 wt% of LC. LC content (%)
Set temperature ( C)
Endothermic reaction temperature ( C)
Reaction start temperature ( C)
Exothermic reaction temperature ( C)
Weight gain at (%)
Weight gain after (%)
0
700 800 1000 800 1000 700 800 1000 800 1000 800 1000 800 1000 800 1000
667.5 669.8 670.1 677.6 675 669.5 671.5 673.1 672.1 673.8 672.2 672.5 672.6 672.7 671.1 671.8
642 650 639 690 683 680 680 675 717 740 710 705 718 710 720 722
677.7 690 682 778.3 784.8 703 766.2 756.7 772.2 817.2 770.4 768 773.3 779.3 784.1 787.2
1.92 2.4 4.95 10.92 15.58 0.91 45.64 46.44 42.48 38.95 41.78 44.08 42.56 42.23 35.41 32.94
2.22 2.57 7.84 12.14 24.15 44.65 45.65 46.74 42.49 40.59 41.97 44.62 42.58 42.95 35.73 34.59
1 3
5 7 10 15
Note: All samples had the same size, and the results were obtained from a single experimental run.
example, the temperature at which nitridation and exothermic reactions began increased with increasing LC content. Carbon effectively delayed the nitridation and melting of the powder, possibly owing to its insulation effect. The powder was initially covered by LC, which had a low thermal conductivity (<1 W/m C at 1000 C [26]). Therefore, LC may act as a barrier preventing heat transfer from the atmosphere to the powder. Despite the aforementioned delay, however, the degree of nitridation increased continuously with increasing LC content. Furthermore, the carbonfree Al powder experienced insignificant weight gains of only 1.9% and 2.2%, when the temperature reached 700 C and after 1 h of heat-treatment at this temperature, respectively. The powder containing 3 wt% of LC experienced weight gain of 0.9% upon reaching 700 C and weight gains of 43.4% and 44.7% after 10 min and 1 h of heat-treatment at this temperature, respectively. Extensive reaction, which seemed to occur at the beginning of the heat treatment, was almost complete after 10 min, resulting in negligible weight gain (1.3%) during the final 50 min. Therefore, intensive nitridation was completed within a short time frame. The total weight gain at 700 C increased ~20-fold (from 2.27% to 44.7%) with the addition of only 3% of LC, indicating that ~85% of aluminum was transformed to aluminum nitrides. Furthermore, for a given amount of LC, the degree of nitridation increased with increasing temperature, although the temperature itself had a negligible influence on this increase. Meanwhile, the weight gain decreased with increasing LC content of >3 wt%, suggesting that the carbon content must be optimized in order to achieve maximum nitridation. Extensive physical contact between the LC and Al powders is essential for maximizing the effect of carbon on the nitridation of Al. However, the LC used in this study tended to agglomerate; hence, it only partly covered the surface of the powder. The BET analysis yielded surface areas of 0.64 and 192.27 m2/g for the Al and LC, respectively, corresponding to the respective total surface areas of 25.34 and 76.90 m2 in 40 g of a powder containing 1 wt% of LC. Given that LC tends to agglomerate, an amount that is significantly greater than the theoretical value is required to achieve the complete surface coverage of the Al powder. In fact, excessive carbon has been added to Al to form AlN via the commercial carbothermal method [13]. Therefore, carbon may only partially cover the surface of the Al powder containing 1 wt% of LC. At the same time, incomplete nitridation may also occur when a temperature of 1000 C is reached. In the current study, this phenomenon resulted in additional weight gain (from 15.6% to 24.2%) after heat treatment for 1 h at this temperature. However, the surface of Al powders containing >3 wt% of LC was already
completely covered by carbon; hence, the degree of nitridation varied only slightly with the LC content or heat-treatment time. These results are analogous to the observation that nitridation is significantly enhanced with the addition of carbon (Fig. 2). The representative TG and DSC curves of both samples heated at 700 C for 1 h in a nitrogen atmosphere are shown in Fig. 4. These curves reveal the difference in the nitridation behaviors (i.e., weight gain) of the samples. In the case of pure Al, nitridation in the solid state began at temperatures lower than the melting point, whereas
Fig. 4. TG and DSC curves of (a) pure Al and (b) Al-3 wt% LC powder beds when the setting temperature was 700 C.
222
K.-B. Lee et al. / Journal of Alloys and Compounds 689 (2016) 218e224
Fig. 5. In situ XRD patterns of (a) pure Al and (b) Al-3 wt% LC powder bed. Samples were heated to 800 C under nitrogen atmosphere, after which the spectra were obtained in the temperature ranging from room temperature to 800 C.
nitridation occurred over the melting point of Al in the case of the LCcontaining powder. In the case of the sample fabricated from the pure Al powder, the reaction rate increased slowly and then rapidly at temperatures near the melting point of Al. The rate decreased significantly at temperatures higher than the melting point. The corresponding rate of conversion of Al to AlN was ~4.2%. In the case of the LC-containing powder, the weight increased slowly with increasing temperature. However, the weight increased exponentially when the Al melted, and the reaction stopped abruptly once the nitridation process was complete. Furthermore, the corresponding rate of conversion (~85.9%) of Al to AlN was 20 times higher than that of the pure Al powder. The addition of carbon particles was, therefore, highly effective in improving the nitridation efficiency of the Al particles. This was confirmed by the in situ XRD analysis of the sample heated to 800 C under a nitrogen atmosphere.
The results in Fig. 5 showed that most of the peaks observed at lower temperatures were no longer observed when the pure Al powder was heated to 700 C. In fact, only a single peak occurred, i.e., an extremely high-intensity peak corresponding to the amorphous phase resulting from the melting of Al. Distinct AlN-related peaks appeared at 800 C. However, the Al-3 wt% LC powder mixture exhibited sharp AlN-related peaks, even at 700 C, indicating that the addition of carbon had an effect on the nitridation of the Al particles. The formation of the AlN phase for each sample was also confirmed via XRD analysis, as shown in Fig. 6. In accordance with Equation (1), Al powder reacted readily with nitrogen gas to form AlN. Thermodynamic calculations performed by using data from the NIST-JANAF tables [20], revealed that reaction (1) is highly exothermic And can occur at 700 C (Fig. 7). The particles of most metallic powders, including those of Al, become coated with a
Fig. 6. XRD patterns of pure Al and Al-3 wt% LC powder beds after heat treatment at 700 C for various times.
Fig. 7. Thermodynamical changes in the Gibbs free energy (DG) as a function of temperature. Thermodynamic calculations were performed by using data from the NIST-JANAF tables [20].
K.-B. Lee et al. / Journal of Alloys and Compounds 689 (2016) 218e224
surface oxide layer during exposure to air. Therefore, during exposure to nitrogen gas, the coating of the Al particles with a layer of their native oxide can be expected [27,28]. Accordingly, reaction (2) may be favored over reaction (1). 2Al2O3 (s) þ 2N2 (g) ¼ 4AlN (s) þ 3O2 (g)
(2)
Thermodynamic calculations (Fig. 7) revealed that reaction (2) did not occur at 700 C; hence, the removal of the oxide layer proved to be essential in the successful nitridation of the Al powder. In a previous study, we employed a new mechanism that yielded AlN through the direct nitridation of pure Al particles, along with a layer of AlON formed on the surfaces of the particles; this layer was able to provide catalytic sites that can supply nitrogen atoms [25].
In the current study, however, the unreacted liquid Al (formed during the nitridation) coalesced within the powder bed and the nitrogen diffusion paths were consequently blocked; hence, the nitridation process was retarded. This resulted in a very low nitridation ratio, as shown in Table 1. The thermodynamic calculations (Fig. 7) revealed that nitridation can occur at temperatures lower than 1700 C; therefore, reaction (3) may be expected, in the case of the carbon-containing Al powder bed. During nitridation, the temperature of the powder bed was measured with a thermocouple inserted directly into the bed. As shown in Fig. 8, the temperature increased to a maximum of 1732 C, which was significantly higher than the temperature difference (1032 C) between the furnace chamber and the powder bed. This temperature (1732 C) was also significantly higher than the highest temperature (1290 C) measured in the carbon-free Al powder bed. The degree of nitridation of the carbon-containing Al powder bed was, therefore, higher than that of the carbon-free bed. Al2O3 (s) þ 3C (s) þ N2 (g) ¼ 2AlN (s) þ 3CO (g)
Fig. 8. Temperature difference between the furnace chamber and powder beds.
223
(3)
In addition, the carbon powder may act as a barrier that prevents the coalescence of the liquid Al (as reported in previous studies on AlN particles [23,24]). This possibly leads to a constant nitrogen supply and an enhanced nitridation process. Therefore, the carbon added to the Al powder acted as both a reaction agent for the native oxide and a dispersing agent that prevented the blockage of the nitrogen diffusion path. A few studies have focused on the synthesis of AlN from different carbon sources [3e6]. However, the lowtemperature (i.e., temperatures lower than 1000 C) nitridation behavior of low-carbon Al has barely been investigated. TEM images and EELS spectra of the pure Al and Al-3 wt% LC powder beds after nitridation at 700 C for 60 min are shown in Fig. 9. The reaction product, AlN (not shown here), was identified via EDS and fast Fourier transform (FFT) analyses. The samples
Fig. 9. (a) Bright-field TEM image and (c) EELS spectra of pure Al, and (b) STEM image and (d) EELS spectra of Al-3 wt% LC powder beds.
224
K.-B. Lee et al. / Journal of Alloys and Compounds 689 (2016) 218e224
exhibited considerably different characteristic N K-edges of AlN, as revealed by the high-loss EELS spectra shown in Fig. 9 (c) and (d). For example, in the case of the carbon-free Al bed [Fig. 9 (c)], the first of the three N K-edge peaks corresponded to the outermost shell; this peak had the highest intensity, consistent with the formation of AlON. The N K-edge corresponding to the region just beneath the AlON layer can be attributed to the typical triplet band structure of AlN. The EELS data corresponding to the AlN and AlON layers concur with those shown in previous studies [29e34]. In contrast to the reaction product associated with the carbon-free bed, the product associated with the Al-3 wt% LC bed exhibited only the aforementioned typical triplet band structure over the entire region, thereby confirming that the added carbon acted as a reaction agent for the native oxide of Al. 4. Conclusions The effect of carbon on the nitridation of Al powder was investigated from a microstructural and thermodynamical viewpoint. With the addition of carbon, nitridation occurred at a relatively low temperature of 700 C, which led to an exothermic reaction and a temperature increase by up to 1700 C. This facilitated the removal of the oxide layer and promoted the nitridation of the powder. In other words, carbon acted as a reaction agent of the native oxides on the surface of the Al particles. Furthermore, carbon successfully inhibited the consolidation of molten Al, thereby providing a pathway for a constant supply of nitrogen gas and an increase in the degree of nitridation. Carbon also acted as a dispersing agent (diluent) that prevented the consolidation of the Al powder. Therefore, carbon plays a dual role (i.e., of reaction and dispersing agent) when added in an optimal amount that uniformly covers the entire surface of the Al powder. The TEM and EELS analyses revealed that, depending on the carbon content, two types of AlN (with differing band structures) were formed on the surface of the Al particles, i.e., AlN with a typical triplet band structure was formed when the bed contained carbon, whereas the carbon-free bed gave rise to the formation of AlN covered by an AlON layer.
[7]
[8] [9]
[10]
[11]
[12] [13] [14]
[15]
[16]
[17]
[18]
[19] [20] [21] [22]
[23] [24] [25]
[26]
Acknowledgments [27]
This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (Grant Nos. 2014R1A1A2057846 and 2009-0093814). References [1] Y. Zhang, J. Liu, R. He, Q. Zhang, X. Zhang, J. Zhu, Synthesis of aluminum nitride nanowires from carbon nanotubes, Chem. Mater 13 (2001) 3899e3905. [2] B.S.S. Daniel, V.S.R. Murthy, Directed melt oxidation and nitridation of aluminum alloys: a comparison, Mater. Des. 76 (1995) 155e161. [3] R.-C. Juang, C.-J. Lee, C.-C. Chen, Combustion synthesis of hexagonal aluminum nitride powders under low nitrogen pressure, Mater. Sci. Eng. A Struct. 357 (2003) 219e227. [4] H. Wang, J. Han, Z. Li, S. Du, Effect of additives on self-propagating hightemperature synthesis of AlN, J. Eur. Ceram. Soc. 21 (2001) 2193e2198. [5] W. Zhang, Z. Li, D. Zhang, Synthesizing AlN powder by mechanochemical reaction between aluminum and melamine, J. Mater. Res. 25 (2010) 464e470. [6] M. Qin, X. Du, J. Wang, I.S. Humail, X. Qu, Influence of carbon on the synthesis
[28]
[29] [30]
[31] [32] [33] [34]
of AlN powder from combustion synthesis precursors, J. Eur. Ceram. Soc. 29 (2009) 795e799. M. Balog, P. Yu, M. Qian, M. Behulova, P. Svec Sr, R. Cicka, Nanoscaled AleAlN composites consolidated by equal channel angular pressing (ECAP) of partially in situ nitride Al powder, Mater. Sci. Eng. A Struct. 562 (2013) 190e195. P. Yu, M. Balog, M. Yan, G.B. Schaffer, M. Qian, Powder injection moulding of an AleAlN metal matrix composite, J. Mater. Res. 26 (2011) 1719e1725. Q. Zhang, G. Chen, G. Wu, Z. Xiu, B. Luan, Property characteristics of a AlNp/Al composite fabricated by squeeze casting technology, Mater. Lett. 57 (2003) 1453e1458. Y.Q. Liu, H.T. Cong, W. Wang, C.H. Sun, H.M. Cheng, AlN nanoparticlereinforced nanocrystalline Al matrix composites: fabrication and mechanical properties, Mater. Sci. Eng. A Struct. 505 (2009) 151e156. K. Vijayalakshmi, V.V. Pillay, Influence of deposition time on the microstructure and dielectric properties of AlN/Si thin films for enhanced hydrogen sensing application, Microelectron. Eng. 113 (2014) 29e34. C.R. Rambo, D. Hotza, Microchanneled biomorphic AlN-coated Al2O3 by pressureless infiltrationenitridation, Ceram. Int. 40 (2014) 12567e12571. G. Selvaduray, L. Sheet, Aluminium nitride: review of synthesis methods, Mater. Sci. Technol. 9 (1993) 463e473. J. Haibo, K. Chen, Z. Heping, S. Agathopoulos, O. Fabrichnaya, J.M.F. Ferreira, Direct nitridation of molten Al(Mg,Si) alloy to AlN, J. Cryst. Growth 281 (2005) 639e645. E. Kroke, L. Loeffler, F.F. Lange, R. Riedel, Aluminum nitride prepared by nitridation of aluminum oxide precursors, J. Am. Ceram. Soc. 85 (2002) 3117e3119. T. Suehiro, N. Hirosaki, R. Terao, J. Tatami, T. Meguro, K. Komeya, Synthesis of aluminum nitride nanopowder by gas-reduction-nitridation method, J. Am. Ceram. Soc. (2003) 1046e1048. T. Yamakawa, J. Tatami, K. Komeya, T. Meguro, Synthesis of AlN powder from Al(OH)3 by reduction-nitridation in a mixture of NH3-C3H8, J. Eur. Ceram. Soc. 26 (2006) 2413e2418. A. Chu, M. Qin, Rafi-ud-din, B. Jia, H. Lua, X. Qu, Effect of urea on the size and morphology of AlN nanoparticles synthesized from combustion synthesis precursors, J. Alloy. Comp. 530 (2012) 144e151. T. Okada, M. Toriyama, S. Kanzaki, Direct nitridation of aluminum compacts at low temperature, J. Mater. Sci. 35 (2000) 3105e3111. http://kinetics.nist.gov/janaf/. (accessed 31.10.14).. Q. Zheng, B. Wu, R.G. Reddy, In-situ processing of Al alloy composites, Adv. Eng. Mater 5 (2003) 167e172. A.V. Korshunov, Influence of dispersion aluminum powders on the regularities of their interaction with nitrogen, Russ. J. Phys. Chem. A 85 (2011) 1202e1210. S. D. Dunmead, W. G. Moore, K. E. Howard, K. C. Morse, T. A. Guiton, US patent 5,710,382, 1998. H. Miura, H. Matunaga, US patent 7,022,301 B2. S.H. Kim, J.H. Noh, J.P. Ahn, J.C. Lee, H. Kwon, J.G. Lee, K.B. Lee, Effects of surface oxide on the nitridation behavior of aluminum particles, Metall. Mater. Trans. A 46 (2015) 496e504. P.E. Khizhnyak, A.V. Chechetkin, A.P. Glybin, Thermal conductivity of carbon black, J. Eng. Phys. 37 (1979) 475e478. G.B. Schaffer, B.J. Hall, The influence of the atmosphere on the sintering of aluminum, Metall. Mater. Trans. A 33A (2002) 3279e3284. R.N. Lumley, T.B. Sercombe, G.B. Schaffer, Surface oxide and the role of magnesium during the sintering of aluminum, Metall. Mater. Trans. A 30A (1999) 457e463. M. MacKenzie, A.J. Craven, Quantifying the oxidation of AlN using electron energy loss spectroscopy, J. Phys. D. Appl. Phys. 33 (2000) 1647e1655. V. Serin, C. Colliex, R. Brydson, S. Matar, F. Boucher, EELS investigation of the electron conduction-band states in wurtzite AlN and oxygen-doped AlN(O), Phys. Rev. B 58 (1998) 5106e5115. P.B. Guillemaud, G. Radtke, M. Sennour, Electron spectroscopy imaging to study ELNES at a nanoscale, J. Microsc. 210 (2003) 66e73. S.P. Gao, A. Zhang, J. Zhu, J. Yuan, Anisotropic spectroscopy of nitrogen K-edge in group-III nitrides, Appl. Phys. Lett. 84 (2004) 2784e2786. G. Radtke, T. Epicier, P.B. Guillemaud, J.C.L. Bosse, N-K ELNES study of anisotropy effects in hexagonal AlN, J. Microsc. 210 (2003) 60e65. K.L. Jablonska, T. Suski, I. Gorczyca, N.E. Christensen, R.C.C. Perera, E.M. Gullikson, J.H. Underwood, D.L. Ederer, Z.L. Weber, Electronic states in valence and conduction bands of group-III nitrides: experiment and theory, Phys. Rev. B 61 (2000) 16623e16632.