A two-step rotary chemical vapor deposition technique was developed to uniformly mix Cu nanoparticles with the γAl2O3 powders, and then the as-obtained powders were consolidated to Al2O3-Cu nanocomposites by spark plasma sintering. In the RCVD process, the metal-organic precursor of copper dipivaloylmethanate (Cu(DPM)2) reacted with O2 and then was reduced by H2 in order to erase the contamination of carbon. At 1473 K, the relative density of Al2O3-Cu increased with increasing and the maximum value was 97.7% at = 5.2 mass%. The maximum fracture toughness of Al2O3-Cu was 4.1 MPa m1/2 at = 3.8 mass%, and 1 MPa m1/2 higher than that of monolithic Al2O3, validating the beneficial effects of Cu nanoparticles.

1. Introduction

The incorporation of small amounts of metal nanoparticles, such as Cu [1], Ni [24], W [5], and Cr [6], has been proved to be beneficial for improving the densification, mechanical properties, and conductivity of Al2O3. Particularly, the incorporation of Cu nanoparticles has attracted much attention due to its high ductility and good electrical conductivity [1, 7, 8]. Kim et al. fabricated Al2O3-Cu nanopowders by ball milling for 30 h and then sintered at 1100–1400°C for 5 min in vacuum under a pressure of 50 MPa by the pulse electric current sintering (PECS) [7]. The composites sintered at 1250°C for 5 min showed a relative density above 97% and enhanced fracture toughness of 4.5 MPa m1/2. Oh and Yoon fabricated Al2O3-Cu nanocomposites by hydrogen reduction and hot pressing of Al2O3/CuO powders. The composite prepared from Al2O3-Cu nitrate mixture exhibited maximum strength of 953 MPa [8].

In general, that the wet methods such as coprecipitation usually produces γAl2O3 phase, which would transform to αAl2O3 during high temperature sintering. Thus the uniform distribution of Cu would be quite necessary for high mechanical properties. However, the agglomeration of Cu nanoparticles in sol-gel process and ball milling is usually unavoidable [9]. Compared to the ball milling and wet sol-gel, the dry CVD technique seems more promising due to its easy operation and low pollution to the environments. Fluidized bed CVD (FBCVD) is one of the most efficient CVD techniques to coat each individual particle of a powder from gaseous species [10]. However, the application of FBCVD on powder is limited depending mainly on density and particle size of powder [11, 12]. The use of a novel rotary CVD (RCVD) technique to prepare Ni nanoparticles on ceramic powders has been reported in our previous work. Ni nanoparticles 13.9–84.5 nm and 10–100 nm in diameter were precipitated on the surface of hBN [13] and cBN [14], respectively. The relative density and hardness of hBN were found to be increased by the incorporation of Ni nanoparticles. The hardness of Al2O3-30 vol% cBN/Ni was 27 GPa, about 1 GPa and 5 GPa higher than that of Al2O3-30 vol% cBN and monolithic Al2O3.

In the present study, Cu nanoparticle was precipitated on γAl2O3 for uniform mixture by RCVD using copper dipivaloylmethanate (Cu(DPM)2) as a precursor. In order to erase the carbon contamination, the powder was firstly oxidized by oxygen and then reduced by hydrogen into the metal state. Then the γAl2O3-Cu powders were consolidated by spark plasma sintering at 1373 to 1573 K for 0.6 ks. The effects of Cu nanoparticles on the sintering behavior, microstructure, and mechanical properties of Al2O3 were investigated.

2. Experimental Procedures

The detailed description about RCVD can be found elsewhere [13]. In the present study, Cu nanoparticles were precipitated on γAl2O3 powders by RCVD at 573 to 873 K for 30 min using Cu(DPM)2 as a precursor. The Cu(DPM)2 precursor in the evaporator was heated at 393 to 423 K and carried into the reaction zone by Ar at a flow rate of 1.67 × 10−6 m3·s−1. The supply rate () of Cu(DPM)2 was set at 0.56 × 10−6 kg·s−1, and O2 was also filled in to erase the carbon contamination from Cu(DPM)2. The γAl2O3 powder, <100 nm in average diameter and 2 g in weight, was fed into the reactor and preheated at 793 to 823 K. The total inner pressure of the RCVD apparatus was kept at 800 Pa with a partial pressure of 240 Pa for O2 and 560 Pa for Ar. The deposition time was fixed at 1.8 ks. After deposition, the O2 was cut off and H2 was used to reduce the CuO to Cu. The γAl2O3-Cu powder was consolidated by spark plasma sintering (SPS, model SPS-210LX, SPS Syntex Inc., Japan) at 1373 to 1673 K for 0.6 ks. The heating rate was 3.3 K·s−1 and the pressure was 100 MPa loaded in the whole sintering process. The temperature was measured by an optical pyrometer focused on a hole (ϕ 2 × 5 mm) in the graphite die.

The crystal structure and phase of the γAl2O3-Cu powder and the sintered bodies were identified by X-ray diffraction (XRD) with CuKα radiation. The microstructure was observed by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The Cu mass content (hereafter, ) in the Al2O3-Cu powder was estimated by energy-dispersive X-ray spectroscopy (EDS) and was averaged by five measurements at different areas. The specific surface area, a total surface area per unit of mass, was measured by the BET method and calculated using [15]

Here is the area occupied by a molecule of the adsorbate, is the Avogadro’s number, and is the molar volume of the analysis gas at standard temperature and pressure (STP). is the monolayer adsorbed gas quantity, which could be obtained from

Here and are the equilibrium and saturation pressure of adsorbates at the temperature of adsorption, is the adsorbed gas volume quantity, and is the BET constant.

The bulk density of the sintered body was determined by an Archimedes’ method and converted to relative density using the theoretical density of Al2O3 (3.99 × 103 kg·m−3) [16] and Cu (8.99 × 103 kg·m−3) [17]. The hardness of Al2O3-Cu nanocomposites was measured using Vicker’s indentation test under a load of 19.6 N [18] and the fracture toughness was evaluated by measuring the crack length generated by Vicker’s indentation [19]. Ten points were averaged for each value of hardness and fracture toughness.

3. Results and Discussion

Figure 1 shows the effects of precipitation temperature () on phase structure of the RCVD-treated Al2O3 powders. At  K, no Cu peaks were identified due to the low precipitation temperature. At 673 to 873 K, the diffraction peaks at 2θ = 43.3°(111) and 50.5°(200) were indexed to Cu, and no CuO and C peaks were observed. On the other hand, the intensity of Cu peaks varied with the precipitating temperature and showed the highest at 773 K. Figure 2 shows the TEM image and EDS spectrum of Cu precipitated γAl2O3 powders at 773 K. The EDS spectrum showed clearly the presence of Cu peaks. The Mo peaks were also observed due to the Mo mesh grids used to support the powder for TEM observation. However, it is difficult to distinguish the Cu nanoparticles from Al2O3 by the TEM images due to the small grain size of Al2O3 and Cu. The presence of carbon and oxygen peaks in the EDS spectra was ascribed to the thin organic film coated on the Mo mesh grid to support the powders for TEM observation. Table 1 shows the specific surface area, Cu grain size, and of Al2O3 and Al2O3-Cu powders. The specific surface area of monolithic γAl2O3 was 49.3 m2·g−1. The addition of Cu decreased the specific surface area of Al2O3 slightly from 41.8 to 45.4 m2·g−1 due to the high theoretical density of Cu (kg·m−3). The crystalline grain size of the Cu nanoparticles was obtained by Scherrer’s Formula for the strongest peak using the diffraction angle of the peak and half width of the peak area of Figure 1. The grain size of Cu increased from 18 to 28 nm with increasing the precipitating temperature from 673 to 873 K. The maximum reached 5.2 mass% at 773 K.

Figure 3 shows the effect of sintering temperature () on the phase structure of Cu precipitated γAl2O3 powders at  K. The phase transformation of γAl2O3 to αAl2O3 started at  K, where three αAl2O3 peaks appeared together with γAl2O3 and Cu peaks. At = 1373 K, the γAl2O3 transformed completely to αAl2O3.

Figure 4 shows the effect of on the relative density of Al2O3-Cu nanocomposites. At = 1373, 1473, and 1573 K, the relative density of Al2O3 was 93.6%, 98.4%, and 97.6%, respectively. The maximum value of 98.4% at = 1473 K was a little higher than Kim’s results at = 1523 K for 0.3 ks (97%) [7], indicating that RCVD is better in precipitating Cu nanoparticles for the densification of Al2O3. Hotta and Goto have consolidated αAl2O3 by spark plasma sintering from the raw powder of Al2O3 of ~200 nm in diameter [20]. The relative density at  K reached above 99%. In our present work, however, the maximum relative density was only 98.4% at = 1473 K using the γAl2O3 of 80 nm in grain size, lower than Hotta’s result using the αAl2O3 starting powder at the same sintering temperature. This could be attributed to the phase transformation of γAl2O3 to αAl2O3 in the present sintering process, which retarded the densification process slightly. At  K, the relative density of Al2O3-Cu increased with increasing and the maximum value was 97.7% at mass%. But at and 1573 K, the relative density of decreased with increasing . At  K and mass%, the relative density was only 90.6%, about 7% lower than that of monolithic Al2O3 at the same sintering temperature. The decrease of the relative density at = 1573 K was explained from the microstructural observation of Figure 5.

Figure 5 shows SEM images of the fracture surface of Al2O3 and Al2O3-Cu sintered at to 1573 K. At  K, some pores were observed in Al2O3 fracture surfaces. In contrast, almost no pores were identified in Al2O3-Cu at the same sintering temperature, indicating Cu nanoparticles improved the densification of Al2O3. Adding a small amount of nanoparticles could usually inhibit the grain growth of Al2O3. For example, Ji and Yeomans found that 5 vol% Cr apparently decreased the grain growth of Al2O3 hot pressed at 1723 K and thus increased the strength and fracture toughness of Al2O3 [6]. Zhang et al. also found that Ni could decrease the grain growth of Al2O3 [21]. In the present study, however, although the grain size of Al2O3 was not apparently inhibited, the relative density was increased due to the decrease of the pores. The reason might be that the Cu was in situ coated on the surface of γAl2O3 (80 nm in grain size) instead of ball-mill mixing. At and 1573 K, the pores were both observed in Al2O3 and Al2O3-Cu. By incorporating pores among grains, the grain size increase of Al2O3-Cu at = 1573 K induced the lower relative density as shown in Figure 4. On the other hand, at 1373 K, the nanocomposite is less porous than monolithic Al2O3, due to the filling of the Cu nanoparticles in the pores of bulk Al2O3. However, the opposite is true at = 1573 K as Cu might retard the diffusion of the densification process at a high temperature.

Figure 6 shows the effect of on the Vickers’ hardness and fracture toughness at a load of 19.6 N. The maximum hardness was  GPa at 2.1 mass% Cu, whereas the maximum fracture toughness was  MPa m1/2 at mass%. Compared to those values of 19.8 GPa and 3.1 MPa m1/2 of the monolithic Al2O3, the hardness and fracture toughness of Al2O3-Cu nanocomposites were about 4% and 32% higher at 3.8 mass%, respectively. The reason that the hardness and toughness were achieved at different compositions is that the concentration of Cu determined these two quantities differently. The Cu nanoparticle would like to fill the pores of Al2O3 and thus increased the relative density and the hardness of Al2O3. But the fracture toughness was increased mainly by crack deflecting and bridging. Thus the maximum hardness and toughness were achieved at different compositions as shown in Figure 6.

The incorporation of second particle phases, such as Cu, Cr, and Ni, would often induce the crack deflection and bridging by absorbing the energy and thus increase the fracture toughness related to the indentation crack length [6, 7, 22]. In the present study, the optimum was 3.8 mass%, with which the crack length is the lowest and then the fracture toughness was calculated to be the largest value according to the method as described in the experimental part [19]. On the other hand, the Al2O3-Cu nanocomposites in [7] has a fracture toughness of 4.5 MPa m1/2 and 0.4 MPa m1/2 higher than the value in the present study, but the hardness of the nanocomposites was not mentioned [7]. The higher value of the fracture toughness of the nanocomposite in [7] might be due to its lower hardness and the different testing techniques. When the was higher than 4 mass%, the fracture toughness and hardness of Al2O3-Cu nanocomposites both decreased, which might be due to the agglomeration of Cu nanoparticles at a higher content.

Figure 7 shows a comparison of the hardness and fracture toughness with Al2O3-based nanocomposites in literature mixed by various methods, typically ball milling. The hardness of Al2O3 body, Al2O3-Nb, and Al2O3-Nb-CNTs composites ranged from 16 to 22 GPa, with the low fracture toughness of 2.8–3.6 MPa m1/2 [6, 23]. The incorporation of soft phases, such as Ti3SiC2, would increase the fracture toughness of Al2O3 body with a decrease of hardness [24, 25]. Yao et al. incorporated Ni to increase the fracture toughness of Al2O3 to 4.3 MPa m1/2, while the hardness was 14.2 GPa [26]. It is common that the higher hardness, the lower fracture toughness. The reported values ranged in the hatched area in Figure 7. In the present study, the fracture toughness increased to 4.1 MPa m1/2, about 40% higher than that of Al2O3 body with the hardness of 20.4 GPa, higher than most of reported values for the Al2O3 body and many Al2O3-based composites in literature.

4. Conclusions

Cu nanoparticles were precipitated on γAl2O3 by rotary chemical vapor deposition at = 673 to 873 K for 30 min, and then the as-obtained composite powders were consolidated by spark plasma sintering to get bulk Al2O3-Cu nanocomposites. Although the grain size and increased with increasing precipitation temperature from = 673 to 873 K, they were still in the range of nanocomposite powders. The incorporation of Cu enhanced the densification of Al2O3 body by SPS at = 1473 K. The sintered Al2O3-Cu nanocomposites with a fine microstructure showed the maximum hardness and fracture toughness of  GPa at mass% and  MPa m1/2 at mass%, respectively. Therefore, the treatment of rotary CVD and spark plasma sintering could be a promising route to obtain high performance ceramic-based nanocomposites.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.


The authors appreciate the financial supports from the Fundamental Research Funds for the Central Universities (2013B34414), and the National 973 Plan Project (2015CB057803).