Fabrication, microstructure and mechanical properties of Al 2O3/Ni nanocomposites by a chemical method

Article abstract of DOI:10.1016/S0025-5408(03)00202-2

Al₂O₃/Ni nanocomposites were prepared by a chemical method and hot-pressing technique using Al(NO)₃·9H ₂O and NiO as the starting materials. At 1450 °C the composites can achieve >98% theoretical density after hot-pressing. Microstructural results revealed that Ni grains were uniformly dispersed in the Al ₂O₃ matrix, some fine Ni grains of ～100 nm were entrapped within Al₂O₃ grains, some Ni grains of 100-300 nm were dispersed at the grain boundary, forming intra/inter-type nanocomposite. The three-point bending strength and the fracture toughness of the dense Al₂O₃ composite increased by 26 and 79%, respectively, with the addition Ni content. The strengthening and toughening mechanisms of the composite with respect to the microstructure are analyzed.

Full text of DOI:10.1016/S0025-5408(03)00202-2

Materials Research Bulletin 38 (2003) 1591–1600
Fabrication, microstructure and mechanical properties of
Al O /Ni nanocomposites by a chemical method
2
3
a,b,*
Guo-Jun Li , Xiao-Xian Huang , Jing-Kun Guo
b
b
a
National Key Lab of Advanced Composites, Beijing Institute of Aeronautic Materials, Beijing 100095, PR China
b
State Key Lab of High Performance Ceramics and Superfine Microstructure, Shanghai Institute of Ceramics,
Chinese Academy of Sciences, Shanghai 200050, PR China
Received 9 September 2002; received in revised form 10 March 2003; accepted 21 July 2003
Abstract
Al O /Ni nanocomposites were prepared by a chemical method and hot-pressing technique using Al(NO) ꢀ9H O
2
3
3
2
and NiO as the starting materials. At 1450 8C the composites can achieve >98% theoretical density after hot-
pressing. Microstructural results revealed that Ni grains were uniformly dispersed in the Al O matrix, some fine
Ni grains of ꢁ100 nm were entrapped within Al O grains, some Ni grains of 100–300 nm were dispersed at the
2
3
2
3
grain boundary, forming intra/inter-type nanocomposite. The three-point bending strength and the fracture
toughness of the dense Al O composite increased by 26 and 79%, respectively, with the addition Ni content. The
2
3
strengthening and toughening mechanisms of the composite with respect to the microstructure are analyzed.
2003 Elsevier Ltd. All rights reserved.
#
Keywords: A. Nanostructures; A. Structural materials; B. Chemical synthesis; C. Electron microscopy; C. X-ray diffraction;
D. Mechanical properties
1
. Introduction
Al O ceramic has good properties in its hardness, its chemical stability and its refractory character.
2
3
However, the application of Al O ceramic as engineering parts is handicapped by its brittleness.
2
3
Researchers prove the brittleness of Al O ceramic can be toughened by the incorporation of ductile
2
3
metal [1–4]. The advantages of the addition of second phase were reported in several aspects, namely,
the reduction of grain size of the matrix grain or the improvement of fracture behavior [5,6]. Recently,
more and more evidence highlights that the mechanical properties of ceramics can be significantly
enhanced by adding nano-sized metallic inclusions [7–9]. A composite containing nano-sized
*
Corresponding author. Present address: Department of Materials Science and Engineering, Dalian Railway University, 794
Huanghe Road, Dalian 116028, PR China. Tel.: þ86-411-4106828.
E-mail addresses: guojunlee@163.com, guo_junlee@yahoo.com.cn (G.-J. Li).
0025-5408/$ – see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0025-5408(03)00202-2

1
592
G.-J. Li et al. / Materials Research Bulletin 38 (2003) 1591–1600
inclusions was coined by Roy et al. as a nanocomposite [10]. Reinforcement models show uniformity
and fine grain are two major microstructural requirements for the preparation of ceramic composites
with superior properties [11]. However, in general, Al O /metal nanocomposite was very difficult to
2
3
achieve the uniform microstructure by traditional techniques involving mechanical mixing of Al O
2
3
and nano-sized metallic inclusions followed by hot-pressing. In order to achieve ultimate performance,
Al O -based composites should start from the stage of powder synthesis. Although a small-scale
2
3
homogeneity can be obtained using the sol–gel route [12–14], the relatively high cost of some reactants
and the difficulty to control the gel-drying step are drawbacks of the method. The results of some
studies [15,16] have demonstrated that the heterogeneous precipitation method is able to overcome
difficulties in the second phase dispersed in the matrix. Especially, Wang et al. [17] prepared a g-Al O /
2
3
SiC mixture powder by using this chemical method, and then obtained intra-type Al O /SiC
nanocomposite (SiC is the second phase) with uniform and special microstructure and high mechanical
2
3
properties by hot-pressing. The special microstructure resulted from g-Al O phase transformation and
2
3
grain growth. The literature on Al O /Ni nanocomposite prepared by the method are few, so this
research aims at producing Al O /Ni nanocomposite from g-Al O–NiO mixture powders prepared by a
2
3
2
3
2
simple chemical method, by hydrogen reduction, followed by sintering. The microstructure and
mechanical properties of nanocomposites are analyzed and investigated.
2
. Experimental procedures
Starting nano-sized NiO material was prepared in the lab using a given process as described in the
literature [18]. As observed from its microstructure (Fig. 1), the average size of NiO particles is about
1
0 nm. The obtained NiO and Al(NO ) ꢀ9H O (Analytically Pure) powders were taken in the desired
3
3
2
proportion and the distilled water and 2 wt.% poly-ethylene glycol (PEG, molecular weight 1500) as
dispersant were added to make a slurry of NiO in Al(NO ) solution by sonicating. To this slurry, the
3
3
exceeding amount of NH ꢀH O was added to hydrolyze Al(NO ) completely into Al(OH) . The
3
2
3 3
3
resulting precipitates were filtered, thoroughly washed with distilled water and dried at 70 8C for 24 h.
The solid mixture was calcined at the desired temperature for 2 h at a heating rate of 5 8C/min to
decompose the Al(OH) to Al O . Then, composite powders of oxides of Ni and Al were wet ball-milled
3
2 3
in highly pure ethanol for 24 h with a-Al O ball, reduced at 700 8C for 4 h in hydrogen atmosphere at a
2
3
heating rate of 5 8C/min [19] and finally hot pressed at the different temperatures for 1 h with an applied
an axial pressure of 20 MPa in Ar atmosphere. The heating and cooling rates were 30 8C/min.
The hot-pressed bodies were cut, ground with a diamond wheel, and polished using diamond pastes.
The final size of specimens was 5 mm  2:5 mm  30 mm. Bulk density was measured by the
Archimedes method. The bending strength was evaluated using a three-point bending test with a span
of 20 mm and a cross-head speed of 0.5 mm/min at room temperature by an Instron-1195 Universal
Test machine. Fracture toughness, K , was determined using an indentation technique with a Vickers
IC
indenter (AKASHI) using 98 N load. The formula used for calculating K was expressed as
IC
1=2
2=3
KIC ¼ AðE=HÞ ðp=c Þ [20]. A minimum of six specimens was tested to obtain a single datum.
X-ray diffraction (XRD) patterns were obtained at a scanning rate of 48/min with 2y range from 10 to
708 using a fully automated diffraction (Rigaku RAX-10, Japan) with CuKa (0.15406 nm) radiation. A
transmission electron microscope (TEM) (Model JEM-200CX, JEOL, Tokyo, Japan) was applied to
investigate the particle size, shape of the mixed powder and composite. Fracture surfaces were

G.-J. Li et al. / Materials Research Bulletin 38 (2003) 1591–1600
1593
Fig. 1. TEM micrograph of the obtained nano-sized NiO powder.
examined by scanning electron microscopy (SEM) (EPMA-8705QHz) to investigate fracture model,
grain size and crack propagation.
3
. Results and discussion
3
.1. The composite powders
Fig. 2 is an XRD patterns of the composite powders treated at the different condition. The XRD
pattern before calcinations (Fig. 2a) contains the characteristic peaks for NiO and boehmite. After
Fig. 2. XRD patterns of the composite powders treated at the different condition: (a) uncalcination; (b) calcination at 900 8C;
(
c) calcination at 1100 8C; (d) reduction at 700 8C.

1
594
G.-J. Li et al. / Materials Research Bulletin 38 (2003) 1591–1600
2 3
Fig. 3. TEM image of Al O /5 vol.% Ni (AN5) composite powder.
calcination at 900 and 1100 8C the Al(OH) peaks are no longer observed and g-Al O peaks are found
3
2 3
except for NiO peaks (Fig. 2b). When calcinations at 1100 8C, a-Al O and NiAlO peaks appear and
2
3
2
NiO peaks disappear, which shows NiO has reacted with Al O at this temperature. In this experiment,
9
2
3
00 8C is selected as the calcination temperature because nano-sized g-Al O is of highly sintering
2
3
activity [21] and no chemical reaction is complete. After as-calcined composite powders reduced at
00 8C in hydrogen, NiO is converted to Ni by XRD analysis (Fig. 2d). Before reduction, the ball-
milling process is necessary. Two advantages: (1) introducing a-Al O seed and reducing phase
7
2
3
transformation temperature [22]; (2) breaking agglomeration.
Fig. 3 is a TEM image of as-calcined and reduced composite powders. As observed from the figure,
the spherical Ni particles with the size of ꢁ25 nm are uniformly dispersed in g-Al O . In contrast to
2
3
Fig. 1, it shows that NiO particles are agglomerated and grown during heat treatment. The result of
the literature [23] proved that g-Al O was a porous structure and its porous walls consist of a lot of
2
3
g-Al O particles. NiO particles are surrounded by g-Al O , so the close nano-sized NiO particles can
2
3
2 3
be agglomerated and grown through porous walls during calcinations.
3
.2. Densification of materials
Fig. 4 is the variation of relative densities for Al O /Ni composites containing the different Ni
2
3
content with hot-pressing temperature. The relative density of the samples enhances with increased
sintering temperatures. When Ni content is relatively lower (for example Al O /3 vol.% Ni (AN3)
2
3
samples), the samples attain the highest relative densities of 98.5% at 1450 8C for 1 h. In contrast to
monolithic Al O ceramic with the same density, the sintering temperature decreases by about 100 8C.
2
3
When Ni content is relatively higher (for example Al O /10 vol.% Ni (AN10) samples), the sintering
temperature is at 1400 8C (a possibly reason is that when Ni content is relatively higher, low melting
point nano-Ni can decrease sintering temperature), which is lower than that of AN3 samples, but only
2
3
97.8% theoretical density can be achieved. At the same temperature, the relative density of the samples
containing high Ni content is lower than that of the samples containing low Ni content. Al O is not
2
3
soaked by solid or liquid Ni and Ni grains coarsen [24] so the sintered samples produce more pores at
the interface of Al O /Ni (Fig. 5). As a result, the relative density of the samples containing high Ni
content is relatively lower.
2
3

G.-J. Li et al. / Materials Research Bulletin 38 (2003) 1591–1600
1595
1
300
1350
1400
1450
1500
1550
1600
1
00
100
95
90
85
80
75
70
9
9
8
8
7
7
5
0
5
0
5
0
2
Al O3
AN3
AN10
1
300
1350
1400
1450
1500
1550
1600
Temperature (˚C)
Fig. 4. Variation of relative densities for Al O /Ni composites containing the different Ni content with hot-pressing
temperature.
2
3
3
.3. Microstructure
The TEM micrograph of AN5 composites is shown in Fig. 5. The grain structure of the matrix phase
and inclusion are noted. Because the atomic number of Ni is far higher than Al O , the scattering
contrast results in a black feature for Ni grains and gray for Al O grains. Ni grains are located both at
2
3
2
3
the intergranular sites and within the Al O grains. But most of them seat at sites, i.e. the triple
2
3
junctions or the Al O /Al O grain boundaries. As known, the thermal expansion coefficient of Ni is
3
2
2 3
larger than that of Al O , when the composites are cool from a sintering temperature; the shrinkage
2
3
of Ni grain is more than that of Al O . Therefore, wherever Ni grains are located, there are pores at
2
3
the interface of Al O /Ni. The thermal expansion mismatch certainly exerts a radial tensile stress at the
2
3
interface of Al O /Ni, so the bonding strength of the interface is very weak. Therefore, in the
2
3
preparation process of the thin TEM sample, some Ni grains depart from the matrix and produce empty
places (Fig. 5).
Fig. 5. A typical TEM micrograph of AN5 composites.

1
596
G.-J. Li et al. / Materials Research Bulletin 38 (2003) 1591–1600
Fig. 6. Dislocation produced in the composite. Arrows indicate dislocation.
On the whole, Ni grains are homogeneously dispersed in Al O matrix. The size of intergranular Ni
2
3
grains is between 100 and 300 nm, and the average size of the intragranular Ni grains is about 100 nm.
The average grain size of the Al O matrix is less than 0.8 mm. Thus, according to Niihara’s
2
3
nanocomposite concept [25], the prepared Al O /Ni composite is a kind of inter/intra-type
3
2
nanocomposite. Nano-Ni grains are entrapped in Al O grains two ways [26,27]. First, the vermicular
2
3
growth of a-Al O during Al O phase transformation would have small Ni grains entrapped in Al O
2
3
2
3
2 3
grain; second, because the mobile velocity of the boundary is faster than that of Ni grain in the process
of Al O growth, Ni grains are separated from the grain boundary and entrapped in the Al O grain, at
2
3
2 3
the same time, the dislocation observed in Fig. 6 (arrows) is produced. When the Ni grains grown
during sintering enter into Al O grain, a higher energy was required. Thus, they have difficulty in
2
3
entrapping in the Al O grain although Al O phase transformation and growth have to seat at the triple
2
3
2 3
junctions of the low energy.
0
2
4
6
8
10
5
5
5
4
4
4
4
4
40
6.5
6.0
5.5
5.0
4.5
4.0
3.5
3.0
20
00
80
60
40
20
00
Bending strength
Fracture toughness
0
2
4
6
8
10
Ni content (vol%)
Fig. 7. Variations of bending strength and fracture toughness of the composites as a function of Ni content.

G.-J. Li et al. / Materials Research Bulletin 38 (2003) 1591–1600
1597
3
.4. Mechanical properties
The strength and the toughness of the composites are shown as a function of Ni content in Fig. 7.
Each point in the figures represents the average value of four specimens. The error bars indicate one
standard deviation. Compared with the monolithic Al O , Al O /Ni nanocomposites show a marked
2
3
2 3
increment in both strength and toughness. With 5 vol.% Ni addition, the strength increases by 26%.
With 3 vol.% Ni addition, the toughness is enhanced by 79%.
Fig. 8, typical SEM micrographs of the fracture surface of Al O /Ni composite. Evidently, the
2
3
composites have a fine matrix grain size when compared with the monolithic Al O , which shows the
2
3
introduction of Ni particles can restrain the growth of Al O grains. The fracture mode of the composite
2
3
changed from intergranular fracture to transgranular fracture with increasing Ni content.
Fig. 8. Typical SEM micrographs of fracture surface of Al
d) 10 vol.% Ni.
2 3
O /Ni composite: (a) 0 vol.% Ni; (b) 3 vol.% Ni; (c) 5 vol.% Ni;
(

1
598
G.-J. Li et al. / Materials Research Bulletin 38 (2003) 1591–1600
0
00
.7
0.8
0.9
1
2
6
600
-
0.35
¦
Ò=486d
f
5
00
500
4
00
0
400
.7
0.8
0.9
1
2
Grain size (µm)
Fig. 9. Fracture strength vs. grain size for Al
2 3
O /Ni composites.
According to the Griffith theory, the fracture strength (s ) of brittle material is expressed as the
f
following equation:
ꢀ
1ꢁ
KIC
1
c2
sf ¼
(1)
Y
where K , c and Y are the fracture toughness, one-half the width of the initial flaw and the geometrical
IC
parameter of a flaw, respectively. In general, c is proportional to the grain size (d) in dense
polycrystalline materials. Thus, the fracture strength increases when the grain size is small (Eq. (1)).
Therefore, the higher strength of the Al O /Ni composite is attributed to the decrease in grain size of
2
3
the Al O matrix. When Ni content is higher, the decrease in the strength of the composites results from
2
3
the decrease of the densification of the composites [28].
The strengths of all compositions (0 ꢂ Ni vol:% ꢂ 10) studied depended only on the grain size of the
Al O and were not affected by the second phase Ni, which is in agreement with Rankin’s result [29].
2
3
Fig. 10. SEM micrograph showing the crack propagation from the polished surface of AN3 composite.

G.-J. Li et al. / Materials Research Bulletin 38 (2003) 1591–1600
1599
ꢃx
A log-log plot of the strength values (s ) versus grain size (d) (Fig. 9) is fit to the relation s ¼ kd .
f
f
A value of 0.35 was calculated for x in agreement with other investigations [28,30]; k was 486. Thus,
ꢃ0:35
the relationship of the fracture strength and grain size can be expressed as: s ¼ 486d
.
f
Fig. 10 is an SEM micrograph showing the crack propagation from the polished surface of the
composite. Except for the crack deflection toughening, the crack branching and crack bridging (arrow)
contributed to the increase of the fracture toughness.
4
. Conclusions
The sintered density of nanocomposite prepared by a chemical method is more than 98% of the
theoretical density, which was obtained by hot-pressing at 1450 8C for 1 h at 20 MPa in Ar atmosphere.
The intragranular and intergranular Ni grains are in ꢁ100 nm and a few hundred nanometers scales,
respectively. Ni grain addition can inhibit the grain growth of the Al O matrix, reduces the sintering
2
3
temperature. The Al O grain size can be controlled to less than 1.0 mm.
2
3
The thermal stress analysis shows that the fracture mode changes from intergranular to intragranular
with increasing Ni content. No microcracks are observed by TEM.
The fracture strength and the fracture toughness of the Al O /Ni nanocomposite increased in contrast
2
3
to the monolithic Al O . The strengthening mechanism is mainly the refinement of the matrix grain.
2
3
ꢃ0:35
The relationship of the fracture strength (s ) versus grain size (d) can be expressed as: s ¼ 486d
.
f
f
The improvement of toughness is related by means of crack deflection, crack branching, crack bridging
and the pullout of metallic grains.
In brief, compared to the sol–gel method [12,13], this method is relatively simple and inexpensive,
and in contrast to traditional mechanical mixing method [1,2], it overcomes the problem on the
inhomogeneity of the microstructure of the composites. Additionally, the samples obtained by this
method have higher fracture toughness in contrast to other methods [2,13,31,32] when metal phase
content is less than 10 vol.%. Thus, the work has practical value for improving toughness and extending
the application of the Al O matrix cermet.
2
3
References
[
[
[
[
[
[
[
[
[
1] W.H. Tuan, R.J. Brook, J. Euro. Ceram. Soc. 6 (1990) 31–37.
2] T. Sekino, K. Niihara, J. Mater. Sci. 32 (1997) 3943–3949.
3] W.H. Tuan, H.H. Wu, T.J. Yang, J. Mater. Sci. 30 (1995) 855–859.
4] S.T. Oh, T. Sekino, K. Niihara, J. Euro. Ceram. Soc. 18 (1998) 31–37.
5] C.O. McHugh, T.J. Whalen, M. Humenik Jr., J. Am. Ceram. Soc. 49 (1966) 486–491.
6] L.A. Simpson, A. Wasylyshyn, J. Am. Ceram. Soc. 54 (1971) 56–57.
7] T. Sekino, K. Niihara, Nanostruct. Mater. 6 (1995) 663–668.
8] T. Sekino, T. Nakajima, K. Niihara, Mater. Lett. 29 (1996) 165–169.
9] M. Nawa, T. Sekino, K. Niihara, J. Mater. Sci. 29 (1994) 3185–3192.
[
[
[
[
[
10] R. Roy, R.A. Roy, D.M. Roy, Mater. Lett. 4 (1988) 323–328.
11] A.G. Evans, Mat. Sci. Eng. A 105/106 (1988) 65–75.
12] E. Breval, G. Dodds, C.G. Pantano, Mat. Res. Bull. 20 (1985) 1191–1205.
13] E. Breval, Z. Deng, S. Chiou, C.G. Pantano, J. Mater. Sci. 27 (1992) 1464–1468.
14] E.D. Rodeghiero, O.K. Tse, J. Chisak, E.P. Giannelis, Mater. Sci. Eng. A195 (1995) 151–161.

1
600
G.-J. Li et al. / Materials Research Bulletin 38 (2003) 1591–1600
[
[
[
[
[
[
[
[
[
[
[
15] S.W. Wang, X.X. Huang, J.K. Guo, Mater. Lett. 28 (1996) 43–46.
16] M.D. Sacks, N. Bozkurt, G.W. Scheiff, J. Am. Ceram. Soc. 74 (1991) 2428–2437.
17] H.Z. Wang, L. Gao, L.H. Gui, J.K. Guo, J. Inorg. Mater. 12 (1998) 671–674 (in Chinese).
18] G.J. Li, X.X. Huang, Y. Shi, J.K. Guo, Mater. Lett. 51 (2001) 325–330.
19] T. Sekino, T. Nakajima, S. Ueda, K. Niihara, J. Am. Ceram. Soc. 80 (1997) 1139–1148.
20] G.R. Anstis, P. Chantikul, B.R. Lawn, D.B. Marshall, J. Am. Ceram. Soc. 64 (1981) 533–538.
21] T.-S. Yeh, J. Am. Ceram. Soc. 71 (1988) 841–844.
22] M. Kumagai, G.L. Messing, J. Am. Ceram. Soc. 68 (1985) 500–505.
23] F.W. Dynys, J.W. Halloran, J. Am. Ceram. Soc. 65 (1982) 442–448.
24] W.H. Tuan, J.M. Wu, J. Mater. Sci. 28 (1993) 1409–1414.
25] K. Niihara, A. Nakahira, T. Tanaka, in: S. Komarneni, J.C. Parker, G.J. Thomas (Eds.), Nanophase and Nanocomposite
Materials, Mater. Res. Soc. Symp. Proc. 286 (1992) 405–412.
[
[
[
[
[
[
[
26] A. Piciacchio, S.H. Lee, G.L. Messing, J. Am. Ceram. Soc. 77 (1994) 2157–2164.
27] L.C. Stearns, M.P. Harmer, J. Am. Ceram. Soc. 79 (1996) 3013–3019.
28] S.C. Wang, W.C.J. Wei, Nanostruct. Mater. 10 (1998) 983–1000.
29] D.T. Rakin, J.J. Stiglich, D.R. Petrak et al., J. Am. Ceram. Soc. 54 (1971) 277–281.
30] R.M. Spriggs, T. Vasilos, J. Am. Ceram. Soc. 46 (1963) 224–228.
31] W.H. Tuan, M.C. Lin, H.H. Wu, Ceram. Inter. 21 (1995) 221–225.
32] T. Sekino, T. Nakajima, S. Ueda et al., J. Am. Ceram. Soc. 80 (1997) 1139–1148.