SCIENCE CHINA Technological Sciences • Article •
June 2014 Vol.57 No.6: 1085–1092 doi: 10.1007/s11431-014-5518-0
Microstructural evolution and mechanism of grain growth in magnesia ceramics prepared by high pressure and temperature with ultra-high heating rate LIU JiangHao1, FU ZhengYi1*, WANG WeiMin1, ZHANG JinYong1, WANG Hao1, WANG YuCheng1, LEE SooWohn2 & NIIHARA Koichi3 1
State Key Laboratory of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, Wuhan 430070, China; 2 Department of Environmental Engineering, Sun Moon University, Chungnam 336-708, Korea; 3 Extreme Energy-Density Research Institute, Nagaoka University of Technology, 1603-1 Kamitomioka, Nagaoka, Niigata 940-2188, Japan Received November 28, 2013; accepted February 10, 2014; published online March 28, 2014
The fast densification method of combustion reaction plus quick pressing was adopted to prepare nanocrystalline ceramics. The densification process of magnesia compact with a particle size of 100 nm was investigated, under the applied pressure of up to 170 MPa, and the temperature range of 1740–2080 K with ultra-high heating rate (above 1700 K/min). High-purity magnesia ceramics with a relative density of 98.8% and an average grain size of 120 nm was obtained at 1740 K, and the grain growth during the densification process was effectively restrained. The characteristic morphology of evaporation-condensation was observed in the compact prepared at 2080 K, which revealed the actual process of mass transfer by gas diffusion. Moreover, the investigation on the microstructure evolution and mechanism of grain growth was carried out, on the basis of as-preserved nanocrystalline ceramics. The result indicated that the grain growth of the nanocrystalline MgO was controlled by the mechanism of evaporation-condensation rather than surface diffusion. Furthermore, the pressure had an influence of restraining the grain growth based on solid diffusion and strengthening the effect of gas diffusion with the increasing temperature. Under the particular conditions, there existed an appropriate temperature for the densification of nanocrystalline magnesia, while the excessive temperature would exaggerate grain growth and impede densification. grain growth, densification, heating rate, evaporation-condensation, surface diffusion, nanocrystalline ceramics Citation:
Liu J H, Fu Z Y, Wang W M, et al. Microstructural evolution and mechanism of grain growth in magnesia ceramics prepared by high pressure and temperature with ultra-high heating rate. Sci China Tech Sci, 2014, 57: 10851092, doi: 10.1007/s11431-014-5518-0
1 Introduction Nanocrystalline materials have attracted considerable attentions due to their outstanding performances originated from the nanometer size effect [1], and high-density ceramics with grain size of nanoscale are widely demanded for both industrial application and scientific research. To prepare dense *Corresponding author (email:
[email protected]) © Science China Press and Springer-Verlag Berlin Heidelberg 2014
nanocrystalline ceramics, one of the common methods is to enhance densification of the compact of nano-particles at sufficiently high temperature while retaining the grain size to nanometer degree. But the densification process at high temperature is usually accompanied by grain growth, which is not only detrimental to densification, but also adverse to the preservation of nanostructure [2]. Consequently, the control of grain growth is critical in optimizing the microstructure and performance of nanocrystalline ceramics. However, it is a challenging task to prepare dense nanotech.scichina.com link.springer.com
1086
Liu J H, et al.
Sci China Tech Sci
crystalline ceramics by the conventional densification methods, including conventional (single-step) pressureless sintering and hot pressing, which is mainly due to the seemingly inevitable grain growth. Therefore, many efforts were paid to suppress the grain growth during densification of nanocrystalline ceramics. Nevertheless, on one hand, although the goal of restraining final-stage grain growth had been achieved by adopting several specific strategies, as utilizing the drag effect of grain growth inhibitor, or exploiting the difference in kinetics between grain-boundary diffusion and migration mechanism (as the so-called twostep sintering method), the negative effects such as degraded performances (e.g. optical transmittance) and reduced efficiency of production were produced, respectively [3,4]. On the other hand, the fast sintering techniques as microwave sintering and laser sintering could effectively control the grain growth during heat-up period by their high heating rates [5], but it was infeasible to apply pressure to enhance the densification process of their products. Thus the effect of final-stage grain growth was significant under the high temperatures and long soaking durations required for densification. By contrast, the pressure-assisted fast-densification method as spark plasma sintering (SPS) provided a more convenient and efficient way of manufacturing dense nanocrystalline ceramics [6,7]. In particular, the grain growth (or particle coarsening) at early stage of sintering was inhibited by the high heating rate (up to 600 K/min), and the applied pressure played a significant role in enhancing the densification of ceramics, thereby limiting the effect of final-stage grain growth via decreasing the required densification temperature and soaking duration [8]. In recent years, a developing fast-densification method based on the heating technique of combustion reaction (combustion reaction plus quick pressing, termed as CR-QP) was applied to prepare dense fine-grained ceramics [9,10]. By taking advantages of the particular conditions characterized with the ultra-high heating rate (>1600 K/min), high pressure (above 100 MPa) and short soaking duration (limited to several minutes), the grain growth during the densification process of submicron alumina ceramics was almost completely restrained [9]. However, the applicability of this method in fabrication of dense nanocrystalline ceramics has not been verified, and the impact of CR-QP conditions on the grain growth of fine-grained ceramics deserves further analysis. Moreover, it is found that the pressure of CR-QP has an obvious effect on restraining the grain growth of ceramics, which is not consistent with the prevailing viewpoint that the pressure applied on the ceramics prepared by SPS has no significant influence on the final grain size [8]. Cubic magnesia (MgO) is an appropriate model for investigating the mechanism of mass transfer at high temperature and pressure [11]. In this work, it is endeavored to densify high-density nanocrystalline ceramics by CR-QP method, and study the grain growth behavior and the re-
June (2014) Vol.57 No.6
sponsible atomic mechanisms, under the particular sintering conditions.
2 Experiment platform and procedure 2.1
Raw materials
Nanocrystalline MgO powders (99.99%, Sinopharm Co., Ltd, China) with an initial particle size of 100 nm was adopted as the starting materials. The powders were compacted by uniaxial pressing of 8 MPa and cold isostatic pressing (CIP) of 200 MPa in sequence, so as to be shaped into cylindrical compact with a diameter of 20 mm and a height of 4 mm. The relative density of as-obtained MgO green compact was about 54%. 2.2
Experimental process
In this work, the chemical reaction between chromic oxide (Cr2O3), alumina (Al) and carbon (C) was utilized as thermal source to supply the MgO compact with a heating effect: 3Cr2O3+6Al+4C→2Cr3C2+3Al2O3
(1)
The reactants of combustion reaction consisted of Cr2O3 (5 m, 99.9%), Al (29 m, 98.7%) and C (amorphous, 99.5%) powders in a molar ratio of 3:6:4, with various contents (0–30 mol%) of diluents (Al2O3, 4 µm, 99.9%) added in to adjust the temperature schedule of MgO compact. A batch of the reactants (200 g) was ball-milled in ethanol for 6 h and dried up in a vacuum oven at 293 K for 12 h. Then the reactants were compacted into a cylinder of 60 mm in diameter and 35 mm in height, with a green compact of MgO in the center. A thin layer of graphite foil was applied to coat the compact in advance, so as to prevent it from being contaminated by the reactants. The combustion reaction process was conducted in a self-designed stainless die (as illustrated in Figure 1), and the external pressure was provided by a four-column hydraulic machine. In our experiment, the cylinder of reactants was positioned in the center of die by fine sands, which would conduct the external pressure to MgO compact in a pseudo-isostatic manner and protect the die against impact damage. The combustion reaction of reactants would be ignited by an energized tungsten filament, which was in contact with the upper surface of the cylinder of reactants. As current was switched on, the filament flared up to initiate combustion reaction, and the temperature of sample (MgO compact) was monitored by a contact-type thermocouple (WRe5/26type). In CR-QP process, the external pressure was applied to sample when corresponding peak temperature was reached. The designed stress of sample was attained within one second after the application of external pressure, and it remained constant until the end of the holding duration. In this pressing mode, the external pressure displayed on the attached gage
Liu J H, et al.
Sci China Tech Sci
Figure 1 Schematic of preparing MgO ceramics by the method of combustion reaction plus quick pressing (CR-QP).
represents the pressure between the pistol and the upper surface of sands, and it has a certain percentage of loss during its conduction from the loading surface to sample. For instance, as the external pressure is 200 MPa, the actual pressure of sample is measured by a strain gauge to be 170 MPa. To be concise, the pressure data mentioned in the following context have been adjusted to the actual pressure. After the temperature of sample decreasing to the ambient, the sample is taken out from the oven of reactants and mechanically polished to remove the adhering graphite foil, so as to undergo the following tests. 2.3
Test
The microstructures of obtained MgO compacts were characterized by the field emission scanning electron microscope (FESEM, S-4800, Hitachi Co., Japan). The grain sizes were estimated by averaging the intercepts of 150 grains illustrated in the FESEM images. The relative densities of all the samples were measured by the Archimedes method, with the theoretical density of MgO as 3.58 g/cm3.
1087
June (2014) Vol.57 No.6
from the heating effect of combustion reaction [12,13], and then immediately started to decrease naturally to the ambient (Figure 2). The heating rates were so high that elevated the temperature of sample to the peaks of above 1700 K within one minute, thus their heating rates were estimated to be higher than 1700 K/min. Moreover, as listed in Table 1, the peak temperatures (1740–2080 K) decreased with increasing content of diluents while the effective soaking durations (730–810 s) gradually increased. It should be noted that, the effective soaking duration represented the length of period during which the temperature of sample was higher than the as-reported lowest sintering temperature of MgO ceramics (0.3Tm) [14,15], rather than the duration of isothermal stage in equilibrium sintering process. After treatment of the combustion reaction process (with temperature schedule A–D) without pressing, the products got various degrees of grain growth but only minor densification (Figure 3). This result indicated that the high temperatures provided by combustion reaction not only failed to promote the density of compact to a high level, but also lead to significant grain growth, even under the ultra-high heating rates and short soaking durations. Moreover, the microstructural observations (not presented) showed that the size of grain necks had increased to be comparable to that of the adjacent grains. It revealed that the pressure-independent densification process based on intrinsic diffusion (including grain boundary diffusion and lattice diffusion) were stagnated before reaching final stage of sintering, which should be attributed to both the limited capacity of diffusion path
3 Results and discussion 3.1
Results
In our experiments, the combustion reaction process with four different temperature schedules (schedule A–D, corresponding to the reactants with diluents contents of 0, 10, 20 and 30 mol%, respectively) were adopted to prepare MgO ceramics. In this process, the temperature of sample increased to the peak at the ultra-high heating rate originated Table 1
Figure 2 Curve of the temperature of MgO compact versus time of combustion reaction and CR-QP process.
Temperature parameters of the MgO prepared by combustion reaction or CR-QP process with temperature schedule A–D
Temperature schedule
Composition of reactants
Heating rate (K/min)
Peak temperature (K)
Soaking duration (s)
A
(Cr2O3, Al, C)
2080
730
B
(Cr2O3, Al, C)+10 mol% Al2O3
2000
765
C
(Cr2O3, Al, C)+20 mol% Al2O3
1980
795
D
(Cr2O3, Al, C)+30 mol% Al2O3
1740
810
>1700
1088
Liu J H, et al.
Sci China Tech Sci
Figure 3 Plot of peak temperature versus relative density and average grain size of the MgO prepared by combustion reaction process with temperature schedule A–D.
and the short effective soaking duration of combustion reaction process. Subsequently, the external pressure was applied to enhance the sintering process of the MgO. As shown in Figure 4, under the identical pressure of 170 MPa (the upper limit of our faculty) and holding duration of 60 s, the relative density of the sample prepared at the lowest peak temperature (1740 K of schedule D) reached 98.8%, while its grain size was 120 nm (Figure 5), which indicated that the grain growth during the densification process was effectively suppressed. Besides, it also revealed that the rising temperature lead to the larger grain size but lower relative density of the MgO, and the sample prepared at the highest peak temperature (2080 K of schedule A) got a grain size of 500 nm and a relative density of only 94.2%. Furthermore, compared the relative density and grain size data of the asobtained samples (denoted as sample A–D, corresponding to temperature schedule A–D, respectively) with that of the MgO prepared under the identical temperature conditions without the pressing step (Figure 3), it could be found that the external pressure exhibited the effect of increasing density
Figure 4 Plot of peak temperature versus relative density and average grain size of the MgO prepared by CR-QP process with temperature schedule A–D, under the identical pressure of 170 MPa and holding duration of 60 s.
June (2014) Vol.57 No.6
Figure 5 FESEM image on the polished and thermal-etched surface of the MgO ceramics fabricated by CR-QP at 1740 K and 170 MPa.
while decreasing grain size of the MgO. This result seems to contradict with the existing knowledge obtained from the research work SPS, which claims that the additional pressure has no significant effect on the grain growth of ceramics [8]. The microstructure on the fracture surface of sample A was presented in Figure 6. It indicated that the average grain size of was larger than that of the initial particles by several times, with a much wider size distribution. Moreover, to be of interest, it was also observed that there were plenty of spherical nano-sized particles located in the region of grooving, which were arranged orderly to the pattern of circle, just like a string of pearl necklace around the grooving. Furthermore, the tiny particles were found to be unstable at high temperature, because they would disappear after an isothermal annealing process at the temperature of 1673 K for 30 min, while the size and structure of grain neck had no observable change. The special morphology of mass transfer was rarely observed in the ceramics prepared at high temperature, and its formation process was supposed to be
Figure 6 FESEM image on the fracture surface of the MgO ceramics fabricated by CR-QP at 2080 K and 170 MPa.
Liu J H, et al.
Sci China Tech Sci
correlated with unusual atomic mechanism of grain growth. 3.2
Discussion
Compared with the MgO prepared by the combustion reaction process without pressing (Figure 3), the products of CR-QP with the identical temperatures have the higher relative densities (Figrue 4). This result manifests that the pressure promotes the densification of MgO by activating pressure-dependent mechanisms. However, under the uniform pressure conditions, the relative density of sample does not increase but decrease with the increase of temperature, because the fine particles of MgO tend to be evaporated under high temperature, and the amount of vapor would increase with the increase of temperature. As a result, for sample A–C which are prepared at the relatively higher temperatures than that of sample D, the larger amounts of vapor would be trapped by the quick pressing, and the resulting larger volumes of isolated pores are difficult to be eliminated by the inefficient diffusion process. This section just concisely explains the influence of the CR-QP conditions on the density of MgO, and a more detailed discussion about the densification process and mechanism would be reported elsewhere. In this paper, it is focused on researching the behaviors and mechanisms of grain growth. The pressure of CR-QP has an obvious negative effect on grain growth, which is deemed to be closely related with the particular conditions. In CR-QP process, the ultra-high heating rate plays an important role in suppressing time-dependent particle coarsening, thereby preserving the large specific surface area of nano-particles before the peak temperature is reached. Consequently, it is feasible for the pressure which is applied at the peak temperature of the MgO compact to produce a dual effect on the potential grain growth. On one hand, owing to the small area of grain coalescence (or grain neck), the effective stress of compact would be very high and probable to cause plastic yield of the coalescence, so as to restrain the grain growth based on solid diffusion by disrupting the channel of mass transfer. On the other hand, the high pressure exploits the as-preserved large specific surface area to enhance densification [16], which may suppress grain growth via reducing the total surface free energy. These are the major reasons for the pressure to limit grain growth while without impeding densification of the MgO, as the preparation process of sample D under the relatively lower temperature. However, the control effect of pressure on grain growth decreases with the increase of temperature, which implies that the responsible mechanisms and corresponding effects might have some changes. As a result, the investigation on the behavior and mechanism of grain growth under the high temperature is carried out, on the basis of the MgO ceramics prepared by CR-QP at 2080 K and 170 MPa (sample A). At high temperature, the grain growth of ceramics could be realized by mass transfer via the path of either gas (as
June (2014) Vol.57 No.6
1089
evaporation-condensation mechanism) or solid (as surface diffusion and volume diffusion mechanism). For the majority of ceramics, the contribution of evaporation-condensation mechanism to grain growth is not significant, which is mainly due to their low volatilities. Therefore, the grain growth during sintering process of ceramics is generally dominated by the mechanism based on solid diffusion. However, there exist a few exceptions, such as MgO which has a high volatility and a low saturated vapor pressure at the temperature range for sintering [17]. In this view, the effect of gas diffusion or evaporation-condensation on the grain growth of MgO ceramics is deemed to be important. For the compact made up of spherical particles, grain growth is realized by the neck growth of adjacent particles. To analyze the growth process of grain neck, the two-sphere model proposed by Kingery et al. [18] (as shown in Figure 7) is firstly adopted. According to this model, neck growth is realized by the mass transfer from free surface to coalescence, and the total surface area of the two particles could be calculated by [19,20]
A A0 1 1 cos / 2 π R 2 sin 3 ,
(2)
where A0 and A are the initial and transient area of total free surface, respectively. It should be noted that, this model is established on the premise of the conservation of mass, therefore it is appropriate to reflect the process of mass transfer based on surface diffusion rather than evaporationcondensation. For the MgO ceramics prepared by CR-QP, the parameter values of R and are geometrically determined from the microstructural observation (Figure 6), to be 500 nm and about 80°, respectively. In eq. (1), A0 is constant, and A reaches the maximum when dA/d=0 or increases to 29.042°. In another word, solid diffusion could not provide any further contribution to neck growth after the radius size of neck increasing to
Figure 7
The two-sphere model established by Kingery et al. [18].
1090
Liu J H, et al.
Sci China Tech Sci
Rsin29.042°≈0.5R. By contrast, the neck with a larger radius size of about 0.63R can be observed in Figure 6, which indicates that the achievement of neck growth could not be solely attributed to surface diffusion. Moreover, according to the knowledge that the efficiency of volume diffusion is lower than that of surface diffusion [21], the dominant mechanism of grain growth in the MgO is not likely to be based on solid diffusion. Consequently, the effect of evaporating-condensation is inferred to be indispensable to the grain growth of sample A, which is in consistent with the microstructure observation. In this view, the neck growth by evaporation-condensation is considered to proceed in the following way. At first, abundant vapors of MgO are evaporated from the nanocrystalline particles during heat-up period, and the amount of vapor increases with the increasing temperature. Then the pressure which is applied at the peak temperature of the compact forces the vapor to condense by elevating the vapor pressure to exceed the corresponding saturated or equilibrium vapor pressure. In this process, according to the principle of minimizing the total surface free energy, the vapor prefers to be condensed into the bottom of grooving and aggregated into spherical particles (as Figure 6 shows). Subsequently, the neck grows by assimilating the as-deposited particles, then successive particles would continue to be condensed under the holding pressure and assimilated by the neck in the same manner. This procedure represents the normal mode of mass transfer by evaporation-condensation mechanism, and the preservation of such characteristic morphology is owing to the particular conditions of CR-QP. To be specific, under the conditions typically characterized with the ultra-high heating rate and the high pressure applied at the peak temperature, the effect of evaporation-condensation is effectively enhanced. Besides, the short soaking duration and high cooling rate contributes to preserving the transient morphology, which would be eliminated during the long annealing period. To verify the correctness of the results of our theoretical deduction, the effect of surface diffusion on grain growth is evaluated, and compared with the overall achievement of grain growth. The following parameters for the specialty of MgO are consulted for calculation: diffusion coefficient (Ds) is 7.89 ×1023 m2 s1, surface energy of grain boundary (gb) is 1 J m2, atom volume () is 1.87×1029 (m3), and the thickness of grain boundary (s) is 4×1010 m [22]. Melting point (Tm) is 3125 K, critical temperature (Tc) is 5953 K, and critical pressure (Pc) is 33.48 atm. On one hand, the velocity of surface diffusion (Ks) can be calculated as [18,19,23] Ks
3D s s gb , 2πkT d 0
(3)
where k is the boltzmann constant, and d0 is the diameter of
June (2014) Vol.57 No.6
grain. It should be noted that the morphology parameters, including the diameter of grain and radius of grain neck are determined by the observation of Figure 6, and the duration of surface diffusion is approximate to the effective soaking duration of CR-QP process. 2
4
Vπ x . r
(4)
On other hand, the volume of grain neck as the achievement of overall grain growth is also measured. According to the principle of geometrical similarity [24], the volume of neck is estimated to be about 2.93×1020 m3 (by eq. (4)), which is larger than the capacity of mass transferred via surface diffusion by several orders of magnitude. This result further supports that surface diffusion is not possible to be the dominating mechanism of grain growth. However, it is infeasible to directly evaluate the effect of evaporation-condensation, due to the lack of an appropriate computation model [18,21]. For instance, Kingery et al. established a model to assess the effect of vapor-transferring in the grooving with a depth limited to several times of atomic size [18], but this model is not applicable to the grooving with a greater depth. Nevertheless, it is fortunate that Mullins [25] had formulated an acceptable empirical criterion on determining the relative influence of surface diffusion and evaporationcondensation on grain growth. As shown in the following equation, is the ratio of the surface free energy reduced by surface diffusion to that reduced by evaporation-condensation, and there exists a critical value of at 100, above which surface diffusion is certainly the dominant mechanism of grain growth [25]: 1
1 L 2πM 2 1 2 0.38 B t 2 0.38 t 2. 1 A 2Pr 1
(5)
In eq. (5), Pr is the critical or equilibrium vapor pressure in the region of condensation, which is dependent with temperature and surface curvature. is temperature-independent surface tension. Firstly, to estimate the equilibrium vapor pressure on the plane surface of MgO particle (Pvp, at K=0), the Edmier formula is applied to figure out the coefficient of acentric factor (): 3 lg P c 1, 7 1
(6)
where is equal to the ratio of boiling temperature to critical temperature (Tb/Tc), which is counted to be 0.65 for MgO. Then Pvp could be calculated by the Lee-Kesler method with the as-obtained ω: ln P vp f
0
T r f 1 T r ,
(7)
where Tr depends on the actual temperature. At the temperature of 2080 K, Pvp is figured out to be 0.27 atm.
Liu J H, et al.
Sci China Tech Sci
Moreover, the temperature-dependent surface tension (γ) can be estimated by 2
1
P c3T c3Q 1 T
11 9
r
.
(8)
In eq. (8), Q is related to both the critical pressure (Pc) and the relative boiling temperature (Tbr, equal to the ratio of Tb/Tc). The calculation result shows that the surface tension () is 68.6 mN/m at the temperature range near 2080 K. Subsequently, it is feasible to determine the following parameters in a sequence: (1) the equilibrium vapor pressure (Pg) on the concave surface of grooving (with a curvature of K0); (2) the equilibrium vapor pressure (Pr) near the convex surface of the condensed particle, which is located at the bottom of the grooving described in the previous step. Pg ln K0 . kT P vp
(9)
In the first step, the Pg is estimated by the GibbsThompson formula (eq. (9)), in which the curvature of grooving (K0) is approximate to x2/2r2 [23]. It is worthy of noting that Pg decreases as the ratio of neck size to grain size increases, until it drops to be equal to Pvp and then remains constant. In the second step, the Pr is calculated by the Kelvin equation: 2 M RT ln P r , r Pg
(10)
where r and denote the radius size and density of condensed particles, respectively. Finally, the acquired parameters are substituted into eq. (5), and the value of MgO is figured out to be 5.4, which is significantly lower than the critical value of 100 [25]. This result implies that the grain growth of MgO is not controlled by surface diffusion, under the temperature range near 2080 K and the situation of the grain neck as observed in Figure 6. Therefore, although it is still not grounded to ensure that the inverse proposition of the theorem proposed by Mullins is tenable, or the effect of evaporation-condensation is superior to that of surface diffusion, it is logical to infer that the contribution of evaporation-condensation to the grain growth of MgO is indispensable. Furthermore, to clarify the impact of pressure on evaporation-condensation, the model established by Kingery et al. [18] is adopted for analysis: 1
M 2 m P , 2πRT
(11)
where m is the rate of mass transfer by the path of gas, is regulatory coefficient, P is the difference in pressure between grooving and plan surface. Owing to the ultra-high heating rate, the area of coales-
June (2014) Vol.57 No.6
1091
cence at the instance of applying pressure is very small, thus the effective pressure in coalescence as well as P would be very high. As a result, the effect of evaporation-condensation is strengthened by the pressure under the CR-QP conditions. Nevertheless, according to the above-mentioned experimental results, the pressure of CR-QP has not positive effect but negative effect on the grain growth of the MgO. This discrepancy should be explained by differentiating the effect of pressure on gas diffusion from that on solid diffusion. For one thing, the pressure is capable of inhibiting surface diffusion by disrupting the path of mass transfer during heatup period, and the short soaking duration also helps to control the effect of time-dependent solid diffusion, therefore the grain growth based on solid diffusion is restrained under the relatively lower temperature (as sample D). For the other thing, although evaporation-condensation is also a timedependent mechanism [18], its efficiency is greatly promoted by pressure, which leads to the exaggerated grain growth based on gas diffusion [26]. But the increment in grain growth that supplied by gas diffusion is not sufficient to make up the decrement caused by solid diffusion, thus the degree of overall grain growth is reduced by the pressure of CR-QP. The special effect of the pressure of CR-QP on the grain growth of nanocrystalline ceramics is distinguished from that of the rest of methods including SPS, which is due to the combined effect of the ultra-high heating rate and the pressure conditions. The preservation of the high specific surface area of nanocrystalline compact until the peak temperature is reached is critical in restraining the grain growth at early stage of sintering, which enables the pressure applied at the peak temperature to maximize the potential of densification, and facilitates the subsequent diffusion-based densification by reducing the effective diffusion distance of nanograins. Therefore, the grain growth at final-stage of sintering is suppressed by pressure via decreasing the temperature and soaking duration required for densification of nanocrystalline ceramics. In sum, it is unusual for evaporation-condensation mechanism to play an important role in the grain growth of ceramics, which should be attributed to the particular temperature and pressure conditions of CR-QP method, as well as the intrinsic characteristic of MgO and the high surface activity of nano-particles. Moreover, to prepare dense nanocrystalline MgO ceramics by CR-QP method, an appropriate sintering temperature is required, while the excessive temperature would not only impede densification by generating isolated pores, but also exaggerate grain growth via strengthening the effect of evaporation-condensation.
4
Conclusion
The CR-QP method provides an efficient and convenient
1092
Liu J H, et al.
Sci China Tech Sci
way of preparing dense nanocrystalline ceramics. By taking advantages of the particular conditions characterized with the ultra-high heating rate (above 1700 K/min) and high pressure (up to 170 MPa), pure MgO ceramics with a relative density of 98.8% and an average grain size of 120 nm was obtained at 1740 K, and the grain growth during the densification process was effectively restrained. The rising temperature of CR-QP had an effect of decreasing density while increasing grain size of the product. Besides, the characteristic morphology of evaporation-condensation mechanism was observed in the ceramics prepared at 2080 K, which revealed the actual process of mass transfer by gas diffusion. According to the microstructural observation and the result of qualitative analysis, the grain growth of nanocrystalline MgO under the high temperature was based on the mechanism of evaporation-condensation rather than surface diffusion. Moreover, although the pressure had the influence of restraining the grain growth based on solid diffusion, it also strengthened the effect of gas diffusion under the higher temperature. Therefore, there existed an appropriate temperature range for sintering nanocrystalline ceramics by CR-QP, while the excessive temperature would lead to the strengthened grain growth and hindered densification.
June (2014) Vol.57 No.6
7
8
9
10
11 12
13
14 15 16 17
18 This work was supported by the Ministry of Science and Technology of China (Grant No. S2010GR0771) and the National Natural Science Foundation of China (Grant No. 51161140399).
19 20
1 2 3
4 5
6
Gleiter H. Nanostructured materials: Basic concepts and microstructure. Acta Mater, 2000, 48: 1–29 Lange F F. Densification of powder compacts: An unfinished story. J Eur Ceram Soc, 2008, 28: 1509–1516 Yoshida H, Kodo M, Soga K, et al. Doping effect on sinterability of polycrystalline yttria: From the viewpoint of cation diffusivity. J Eur Ceram Soc, 2012, 32: 3103–3114 Chen I W, Wang X H. Sintering dense nanocrystalline ceramics without final-stage grain growth. Nature, 2000, 404: 168–171 Oghbaei M, Mirzaee O. Microwave versus conventional sintering: A review of fundamentals, advantages and applications. J Alloys Compd, 2010, 494: 175–189 Chaim R, Levin M, Shlayer A, et al. Sintering and densification of nanocrystalline ceramic oxide powders: A review. Adv Appl Ceram, 2008, 107: 159–169
21
22 23 24 25 26
Munir Z A, Anselmi-Tamburini U, Ohyanagi M. The effect of electric field and pressure on the synthesis and consolidation of materials: A review of the spark plasma sintering method. J Mater Sci, 2006, 41: 763–777 Marder R, Chaim R, Estournès C. Grain growth stagnation in fully dense nanocrystalline Y2O3 by spark plasma sintering. Mater Sci Eng A, 2010, 527: 1577–1585 Meng F C, Fu Z Y, Zhang J Y, et al. Rapid densification of nanograined alumina by high temperature and pressure with a very high heating rate. J Am Ceram Soc, 2007, 90: 1262–1264 Fu Z Y, Huang L W, Zhang J Y, et al. Ultra-fast densification of CNTs reinforced alumina based on combustion reaction and quick pressing. Sci China Tech Sci, 2012, 55: 484–489 Chaim R. Densification mechanisms in spark plasma sintering of nanocrystalline ceramics. Mater Sci Eng A, 2007, 443: 25–32 Munir Z A, Anselmi-Tamburini U. Self-propagating exothermic reaction: The synthesis of high-temperature materials by combustion. Mater Sci Rep, 1989, 3: 277–358 Merzhanov A G. Self-propagating high-temperature synthesis: Twenty years of search and findings. In: Munir Z A, Holt J B, eds. Combustion and plasma synthesis of high-temperature materials. Lect Notes in August. New York: VCH Publishing Inc, 1990. 1–53 Ehre D, Gutmanas Y, Chaim R. Densification of nanocrystalline MgO ceramics by hot-pressing. J Eur Ceram Soc, 2005, 25: 3579–3585 Gupta T K. Sintering of MgO: Densification and grain growth. J Mater Sci, 1971, 6: 25–32 Mohamed F A. Interpretation of superplastic flow in terms of a threshold stress. J Mater Sci, 1983, 18: 582–592 Boo J H, Lee S B, Yu K S, et al. Growth of magnesium oxide thin films using single molecular precursors by metal–organic chemical vapor deposition. Thin Solid Films, 1999, 341: 63–67 Kingery W D, Berg M. Study of the initial stages of sintering of solids by viscous flow, evaporation-condensation, and self-diffusion. J Appl Phys, 1955, 26: 1205–1212 Akash A, Mayo M J. Pore growth during initial-stage sintering. J Am Ceram Soc, 1999, 82: 2948–2952 Kleiman S, Chaim R. Thermal stability of MgO nanoparticles. Mater Lett, 2007, 61: 4489–4491 Vieira J M, Brook R J. Lattice, grain-boundary, surface and gas-diffusion constants in magnesium oxide. In: Kingery W D, ed. Structure and properties of MgO and Al2O3 ceramics. Columbus: American Ceramic Society, 1984. 438–463 Arzt E, Ashby M F, Verrall R A. Interface controlled diffusional creep. Acta Metall, 1983, 31: 1977–1989 Kingery W D, Francis B. Grain growth in porous compacts. J Am Ceram Soc, 1965, 48: 546–547 Herring C. Effect of change of scale on sintering phenomena. J Appl Phys, 1950, 21: 301–303 Mullins W W. Theory of thermal grooving. J Appl Phys, 1957, 28: 333–339 Coble R L. Diffusion models for hot pressing with surface energy and pressure effects as driving force. J Appl Phys, 1970, 41: 4798–4807
