Low-temperature spark plasma sintering of NiO nanoparticles

Materials Science and Engineering A 528 (2011) 2936–2940

Contents lists available at ScienceDirect

Materials Science and Engineering A journal homepage: www.elsevier.com/locate/msea

Low-temperature spark plasma sintering of NiO nanoparticles Rachman Chaim a,∗ , Ram Reshef a , Guanghua Liu b , Zhijian Shen b a b

Department of Materials Engineering, Technion – Israel Institute of Technology, Haifa 32000, Israel Department of Materials and Environmental Chemistry, Arrhenius Laboratory, Stockholm University, S-106 91 Stockholm, Sweden

a r t i c l e

i n f o

Article history: Received 28 October 2010 Accepted 16 November 2010 Available online 23 November 2010 Keywords: Spark plasma sintering Densification Nanocrystalline Particle sliding NiO

a b s t r a c t NiO nanoparticles of 20 nm in diameter were spark plasma sintered between 400 ◦ C and 600 ◦ C for 5 and 10 min durations. Application of 100 MPa pressure from room temperature resulted in densities between 75% and 92%. The final grain size was between 26 nm and 68 nm. Lower densities were recorded when 100 MPa was applied at the SPS temperature. Two shrinkage rate maxima of ∼3.4 × 10−3 s−1 and ∼2 × 10−3 s−1 were observed around 390 ± 10 ◦ C and at the SPS temperature. The two shrinkage rate maxima were related to densification by particle sliding followed by diffusional grain boundary sliding during the heating. The strong effects of the surface and interfacial processes which are active during the SPS were highlighted. © 2010 Elsevier B.V. All rights reserved.

1. Introduction Recently spark plasma sintering (SPS) was extensively used for fabrication of different nanocrystalline ceramics. The rapid heating rates in this technique, due to the application of the pulsed direct electric current, significantly shorten the overall sintering and densification processes, compared to the conventional hot-pressing or hot-isostatic pressing. In addition, the versatility of the technique enables fabrication of materials with different electrical conductivities. NiO is among the electronic ceramics whose electrical behavior strongly depends on both temperature and oxygen partial pressure. This oxide is an insulator with very low electrical conductivity ∼10−13 (ohm cm)−1 at room temperature [1,2]. However, its electrical conductivity severely increases with the temperature increase, so that it reaches 3 × 10−8 (ohm cm)−1 at 230 ◦ C and 10−2 (ohm cm)−1 at 730 ◦ C. Generally, further increase by two orders of magnitude is expected in the electrical conductivity with the decrease in the oxide particle size into the nanometer range [2]. Consequently, the contribution of the thermal and electrical field effects on sintering and densification of nanocrystalline NiO (nc-NiO) may strongly be temperature dependent. Recently, densification of nc-NiO powders by SPS between 600 ◦ C and 900 ◦ C was investigated [3]. Fully dense NiO with submicrometer grains were formed at 800 ◦ C and 100 MPa for 20 min duration. Following the strain rate and the microstructure analyses,

∗ Corresponding author. Tel.: +972 4 829 4589; fax: +972 4 829 5677. E-mail address: [email protected] (R. Chaim). 0921-5093/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2010.11.050

the densification map for the NiO nanoparticles was constructed, and dislocation glide and climb were identified as the dominating densification mechanisms at these temperatures. However, according to these calculated densification maps, densification should also proceed below 600 ◦ C, where diffusion in very limited. The present paper studied the densification of NiO nanoparticles at the lower temperature range between 400 ◦ C and 600 ◦ C, where the effect of different SPS parameters was also investigated. 2. Experimental Commercial pure (99.9%) NiO nanopowder (Inframat Adv. Materials, USA) with 20 nm average particle diameter and BET specific surface area > 50 m2 /g was used. Nanopowder samples with constant weigh of 1.60 g were used to form 12 mm diameter with 2.1–2.8 mm thick specimens, using graphite dies. Graphite foils (Grafoil) were used to isolate the specimen from the die and the plunger surfaces. The SPS experiments were performed using the SPS unit (Dr. Sinter 2050). The green powder within the graphite die was pressed to 100 MPa already from room temperature (unless otherwise indicated) to the SPS temperatures between 400 ◦ C and 600 ◦ C; the heating rate was 100 ◦ C min−1 . The specimens were densified for 5 min under 100 MPa at the SPS temperatures. Selected experiments were conducted also for 10 min durations. In addition, two specimens were densified, where 25 MPa was applied from the room temperature (instead of 100 MPa), followed by pressure increase to 100 MPa at the SPS temperature (500 ◦ C and 550 ◦ C). Finally, one specimen was sintered at 800 ◦ C with zero dwell time, where 75 MPa was applied from room temperature. The

R. Chaim et al. / Materials Science and Engineering A 528 (2011) 2936–2940

2937

Table 1 SPS parameters for the NiO nanopowder for 5 min at 100 MPa. Specimen

1 ◦

SPS temperature ( C) Duration (min) Temperature when 100 MPa was applied (◦ C) 1st Maximum Strain rate 10−3 (s−1 ) Temperature (◦ C) 2nd Maximum Strain rate 10−3 (s−1 ) Temperature (◦ C) Final density (%) * ** #

2 400

3

4

450

5

6

500

7

8

550

600

25 3.4

10 25 3.3

25 3.4

25 3.4

500 3.8

25 3.7

550 4.0

380 – – 75.1

380 – – 76.6

390 – – 82.6

380 2.8 500 84.8

400 1.6# 500 81.9

380 1.9 550 85.9

405 1.6# 550 77.5

*

9

*

10 800

25 3.3

10 25 3.3

25** 4.6

380 1.4 600 88.8

380 1.9 600 91.6

380 – – 94.8

25 MPa was applied from room temperature followed by 100 MPa at the SPS temperature. 75 MPa was applied from room temperature. The strain rate factor of 10−2 s−1 instead of 10−3 s−1 should be used.

pulse and pause durations were 40 and 6.6 ms, respectively, with a standard on–off time relation of 12:2; the vacuum was kept at ∼6 Pa. All the SPS parameters (i.e. time, temperature, pressure, ram displacement and its rate, vacuum level) were monitored during the experiments. The overall SPS conditions were summarized in Table 1. The final densities of the sintered specimens were determined by the Archimedes technique; the theoretical density of 6.67 g cm−3 was used for NiO. X-ray diffraction (Philips PW-3020 diffractometer) with Cu K␣ radiation was operated at 40 kV and 30 mA and used at the scanning rate 0.5◦ min−1 for phase identification, average grain size and accurate lattice parameter determination. The average grain size was determined from diffraction line broadening [4]. The accurate lattice parameters were determined using the cos2  extrapolation method [5]. The calculated values were corrected for the instrumental broadening and angle deviation, using a SrTiO3 standard. 3. Results The specimen characteristics and their SPS parameters were listed in Table 1. The specimens treated at the lowest SPS temperature of 400 ◦ C were porous and cracked during the release from the graphite die. The specimen’s color was dark gray to black, for all SPS temperatures, except dark green for the 800 ◦ C treatment. This color change is associated with the change in the oxide composition towards stoichiometry at high temperatures [6]. A typical temperature–pressure–time (Fig. 1a) and ram displacement–displacement rate-time dependences (Fig. 1b) under 100 MPa pressure show that continuous shrinkage started with the temperature increase up to the SPS temperature. Following the specimen weight, dimensions and displacements, and subtracting the thermal expansion effects, the typical green density of the specimens prior to heating was about 28%. Nevertheless, the common feature to the SPS experiments were two displacement rate maxima. The first maximum was common to all experiments and occurred around 390 ± 10 ◦ C (Table 1), when 100 MPa pressure was applied from the room temperature (Fig. 2a). The temperature of the first maximum did not change when lower pressures of 25 or 75 MPa were applied from the room temperature (specimens 5, 7, and 10 in Table 1). Moreover, the maximal displacement rate was almost constant, ∼9 to 10 ␮m s−1 ; the corresponding strain rates varied between 3.3 × 10−3 s−1 and 4.6 × 10−3 s−1 (Table 1). This onset temperature and shrinkage behavior may then be considered as an intrinsic property of the present NiO nanopowder, as will be discussed below. The second displacement rate maximum was observed only for the specimens sintered above 500 ◦ C. The second maximum displacement rate appeared immediately when the SPS temperature was reached; the corresponding strain rates varied between 1.4 × 10−3 s−1 and 2.8 × 10−3 s−1 (Table 1). How-

ever, when the main pressure was applied at the SPS temperature (Fig. 2b, and specimens 5 and 7 in Table 1), rather than at room temperature (Fig. 2a, and specimens 4 and 6 in Table 1), the second maximum strain rate increased by one order of magnitude to 1.6 × 10−2 s−1 . As may be expected, the final density increased with the increase in the SPS temperature, while other SPS parameters were unchanged (Fig. 3a). On the other hand, the temperature at which the main SPS pressure was applied, significantly affected the final density. This can be noticed by comparing the two specimens sintered for 5 min at 550 ◦ C (Fig. 3a and Table 1): 85.9% density was reached when the 100 MPa pressure was applied at room temperature (circle in Fig. 3a), compared to 77.5% density, when the same pressure was applied only at the SPS temperature (25 MPa preloaded at room temperature, triangle in Fig. 3a). These relative densities correspond to the second stage of sintering. In this respect, densification during the second stage sintering may proceed by particle sliding and/or plastic deformation of the nanoparticles. The yield stress of NiO single crystals in the 400–600 ◦ C range decreased from 50 to 42 MPa, with the temperature increase [7]. Assuming the NiO nanoparticles to be dislocation-free [3], and follow the experimental yield stress of their single crystals [7], the above observed change in the density can be related to the lack of plastic deformation, when the applied pressure (i.e. 25 MPa preload) is below the yield stress at the corresponding SPS temperature. The density difference between the counterpart specimens at 500 ◦ C was ∼3% (84.8% versus 81.9%, Fig. 3a) smaller than the ∼9% measured for the 550 ◦ C treatments. The larger difference at the higher SPS temperature may be associated with enhanced particle neck formation, when using low preload, as will be discussed below. Particle sliding in non closed packed compact under load is expected to be more favorable than the plastic deformation of the same particles. Consequently, the first and the second maxima in the displacement rate should be related to the particle sliding and plastic deformation, respectively. The brittle nature of the specimens sintered at the lowest SPS temperature (400 ◦ C) may explain the lack of plastic deformation, as well as the limited diffusion in these sintered specimens. The average grain size calculated from the XRD line broadening is shown in Fig. 3b. The grain size increased with the increase in both SPS temperature and duration. Comparing the grain size in the sintered specimens to the particle size in the as-received nanopowder indicated that some particle coarsening took place even at the lowest temperature of 400 ◦ C. Previous microstructure study of this system by SPS above 600 ◦ C clearly showed grain growth by grain coalescence while sliding or rotating, rather than normal grain growth by grain boundary migration [3]. The lower grain sizes measured for the lower preload (triangles in Fig. 3b) are in agreement with the lesser propensity to form dense regions where grains coalesce.

2938

R. Chaim et al. / Materials Science and Engineering A 528 (2011) 2936–2940

Fig. 1. (a) Temperature–pressure and (b) displacement–displacement rate versus SPS duration of nc-NiO at 400 ◦ C and 100 MPa. Maximum displacement rate occurred at ∼380 ◦ C.

Fig. 2. Two displacement rate maxima appeared during SPS above 500 ◦ C. The 1st maximum at the lower temperature is screened by the 2nd maximum when the main pressure was applied at the SPS temperature (b).

Analysis of the lattice parameters derived from XRD (Fig. 4) revealed a monotonous increase with the SPS temperature. These changes were in agreement with previous experimental observations, where changes in the lattice parameter and specimen color were related to the defect type and concentration in NiO [6,9]. Black

non-stoichiometric NiO (with excess Ni2+ ) tends to stoichiometry and color change to dark green, when treated at high temperatures; the lattice parameter is also grain size dependent [2,10]. However, the most important aspect of the lattice parameter change here is to express the low temperature diffusion activity, albeit low,

Fig. 3. (a) Relative density and (b) average grain size of nc-NiO versus temperature at 100 MPa for 5 min (circles) and 10 min (cubes) durations. The specimens pre-loaded at 25 MPa were less dense and revealed smaller grain size (triangles).

R. Chaim et al. / Materials Science and Engineering A 528 (2011) 2936–2940

Fig. 4. Lattice parameter of the cubic NiO versus the SPS temperature, showing lattice expansion compared to the as-received nano-powder.

and its characteristic distance, of the order of the particle/grain size. 4. Discussion Different aspects and effects of the nano-powder and the SPS process can be highlighted by a comparison between the specimens sintered at different conditions. Increase in the SPS duration from 5 to 10 min improved the final density only marginally, both at the lowest (i.e. by 1.5% at 400 ◦ C, Fig. 3a) and at the highest (i.e. by 3.0% at 600 ◦ C, Fig. 3a) sintering temperatures. These results indicate that the main densification processes occur during the heating up to the SPS temperature. Apparently, the shear component of the external applied stress, overcomes the frictional forces at the surface of the mutually sliding particles, even at low applied pressures. Contribution of the plastic deformation to densification becomes significant at high applied pressures and temperatures, especially when the relative density reaches those of random (∼64%) or ordered closed packing (∼74%). Further densification of the plastically deformed nano-powder compact may take place by diffusional processes at constant SPS temperature, especially at the final stage sintering [8]; the latter is ineffective at short durations. This limited final densification is in agreement with saturation in the ram displacement observed at all SPS temperatures. In order to evaluate the active densification mechanisms during low-temperature SPS of nanocrystalline NiO, its deformation mechanism map was constructed [3,11,12]. Although this map holds true for the deformation of dense polycrystalline ceramic, it can also be used for the evaluation of the deformation mechanisms close to the final stage sintering. The effective pressure during hot-pressing of porous compacts is often set to higher values [3]; this was neglected here due to the particle sliding mechanism. As long as particle sliding proceeds, no stress intensification is expected; moreover, some pressure decrease may be observed (i.e. at 75 s in Fig. 1a). As was noted above, the two maxima in the strain rates may be related to the maximum sliding rate and maximum plastic deformation rate of the nanoparticles, respectively. However, the physical causes which lead to such maxima should be considered. The maximum in the particle sliding rate can be explained by two processes which have two opposing trends with the temperature increase. On the one hand, the frictional forces between the particles are expected to decrease with the temperature increase, due to decreased surface viscosity; hence, an increase in the densification rate occurs. On the other hand, too high temperatures may cause sticking/bonding between the sliding surfaces, due to enhanced diffusion, hence to a decrease in the densification rate via particle sliding. The larger

2939

Fig. 5. Calculated strain rates for different deformation mechanisms in 20 nm NiO grain polycrystal at 100 MPa pressure. The measured strain rates (filled triangles) coincide with those for grain boundary sliding (GBS) and Coble creep at the lower boundary of the working range.

difference in the density observed between the specimens under different preloads (i.e. 25 MPa versus 100 MPa) at 550 ◦ C than at 500 ◦ C (Fig. 3a) can be a manifestation of this latter sticking/bonding process which resists particle sliding. The two opposing processes result in the first maxima with temperature. Since the particle surface viscosity (elasticity) and diffusivity are the major factors to affect the particle sliding rate, the temperature of the first strain rate maximum can be considered as an intrinsic property of the nano-powder. Lower shear stresses are expected for particle sliding (cold or warm friction) than for particle plastic deformation, as long as a free volume for particle rearrangement and compaction exists. Therefore, the second maximum strain rate, which coincided with the SPS temperature, can be related to the plastic deformation of the relatively dense aggregates, of the order 64–74%, as was mentioned above. The second maximum was screened by the first maximum, and even disappeared when the SPS temperature was 800 ◦ C. The observed densification behavior may be explained in terms of the plastic deformation as following. Application of relatively high pressure (i.e. 100 MPa) from the room temperature is expected to enhance particle sliding, hence form larger and denser regions at lower temperatures. However, as was noted, particle sliding process screened the second displacement rate maximum, which is associated with plastic deformation. This screening may be associated with increased elastic stiffening of the compact [13], and higher effective stress needed for plastic deformation, resulting in the less effective densification. Finally, the expected strain rates at the SPS working window (i.e. temperature and pressure) were compared to the expected strain rates from three different deformation mechanisms (Fig. 5) using the published equations and data for NiO [3]. These mechanisms included grain boundary sliding (GBS) controlled by dislocation glide near the grain boundary and its recovery by a gb diffusion, Coble creep and Nabarro–Herring (N–H) creep. The strain rates calculated from the SPS data are between 3.3 × 10−3 s−1 and 4.6 × 10−3 s−1 for the first maximum (particle sliding) and 1.4 × 10−3 to 1.6 × 10−2 s−1 for the second maximum (plastic deformation) (filled triangles in Fig. 5). The corresponding Nabarro–Herring (N–H) creep rates varied between 10−3 s−1 and 10−6 s−1 at the upper boundary of the working range (Fig. 5), and cannot be responsible for the observed strain rates at the lower temperatures. The strain rates resulting from GBS and Coble creep are in good agreement with the measured values. These later processes are associated with diffusion along the grain boundaries.

2940

R. Chaim et al. / Materials Science and Engineering A 528 (2011) 2936–2940

This confirms that plastic deformation during the low-temperature densification of the NiO nanoparticles proceeded by diffusional processes at the particle surfaces and near the grain interfaces. Acknowledgments The financial support by the Israel Ministry of Science under Contract #3-3429 and Swedish Research Council are gratefully acknowledged. References [1] D. Adler, J. Feinleib, Phys. Rev. B 2 (1970) 3112–3134. [2] S.A. Makhlouf, M.A. Kassem, M.A. Abdel-Rahim, J. Mater. Sci. 44 (2009) 3438–3444.

[3] R. Chaim, O. Reinharz Bar-Hama, Mater. Sci. Eng. A 527 (2010) 462– 468. [4] C. Suryanarayana, M. Grant Norton, X-Ray Diffraction: A Practical Approach, Plenum Press, New York, 1998, pp. 207–215. [5] B.D. Cullity, Elements of X-Ray Diffraction, Addison-Wesley Pub., Reading, 1978, p. 355. [6] N.L. Paterson, Mater. Sci. Forum 1 (1984) 85–107. [7] A. Dominguez-Rodriguez, J. Casting, J. Philibert, Mater. Sci. Eng. 27 (1977) 217–223. [8] R. Chaim, M. Margulis, Mater. Sci. Eng. A 407 (2005) 180–187. [9] S.J. Bhatt, H.D. Merchant, J. Am. Ceram. Soc. 52 (1969) 452–453. [10] F. Fievet, P. Germi, F. de Bergevin, M. Figlarz, J. Appl. Crystallogr. 12 (1979) 387–394. [11] M.R. Notis, J. Am. Ceram. Soc 57 (1974) 271. [12] M. Jimenez-Melendo, A. Dominguez-Rodriguez, J. Casting, R. Marquez, Scripta Metall. 20 (1986) 739–742. [13] A. Piccolroaz, D. Bigoni, A. Gajo, Eur. J. Mech. A/Solids 25 (2006) 334– 357.