Effect of pretreated microstructure on subsequent sintering performance of MgAl2O4 ceramics

Effect of pretreated microstructure on subsequent sintering performance of MgAl2O4 ceramics

Ceramics International 45 (2019) 7544–7551 Contents lists available at ScienceDirect Ceramics International journal homepage: www.elsevier.com/locat...

2MB Sizes 0 Downloads 1 Views

Ceramics International 45 (2019) 7544–7551

Contents lists available at ScienceDirect

Ceramics International journal homepage: www.elsevier.com/locate/ceramint

Effect of pretreated microstructure on subsequent sintering performance of MgAl2O4 ceramics ⁎

T



Shengqiang Guoa, Hao Wanga, , Pengyu Xua, Bin Wanga, Yan Xiongb, , Bingtian Tua, Weimin Wanga, Zhengyi Fua a

State Key Lab of Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology, 122 Luoshi Road, Wuhan, Hubei 430070, China School of Materials Science & Engineering, Hubei Province Key Laboratory of Green Materials for Light Industry, Hubei University of Technology, 28 Nanli, Wuhan, Hubei 430068, China

b

A R T I C LE I N FO

A B S T R A C T

Keywords: MgAl2O4 transparent ceramic Microstructure Densification Pore-boundary separation Sintering performance

Pore-grain boundary separation at the final stage of pressureless sintering was successfully suppressed via limiting grain growth in MgAl2O4 ceramics. A pretreatment process by spark plasma sintering (SPS) was utilized to obtain a homogeneous porous body with relative density of 88%, small pores (~28 nm) and uniform grains (~230 nm). The subsequent pressureless sintering results shows that the SPS-ed sample exhibited superior densification with limited grain growth, in contrast to the sample pretreated by pressureless sintering. Coble's model for intermediate stage of sintering was applied to correlate the densification behavior with the instantaneous grain size. It was demonstrated that the densification process could be promoted significantly by maintaining small grain size. In the final stage of sintering, the grain growth mechanism is pore displacement controlled by lattice diffusion. After final densification by hot isostatic pressing, transparent ceramics possessing fine-grained microstructure and acceptable transmittance had been fabricated. Average grain size of the transparent ceramics was 1.9 µm, which enhanced its Vickers hardness (13.9 ± 0.2 GPa).

1. Introduction Magnesium aluminate spinel (MgAl2O4) is inherently difficult to be sintered [1,2]. Since the grain growth and densification processes during pressureless sintering have the comparable activation energy value, there are difficult to be decoupled with each other [3,4]. With the microstructure coarsening in the sintering of ceramics, the key to achieve nearly full density is to prevent the pore-grain boundary separation [5]. Once the pore-grain boundary separation occurs, the pores are entrapped ultimately within grains and cannot be eliminated subsequently by pressureless sintering (PS) or hot isostatic pressing (HIP) sintering [6,7]. Since the pore detachment is strongly depended on both pore size and grain size, this phenomenon is prevailing in MgAl2O4, in particular, when coarse starting powder is used in sintering experiments [8]. However, as a defect-intolerable material, the residual porosity in transparent ceramics needs to be less than 0.01% [9]. Furthermore, the fine-grained MgAl2O4 transparent ceramics with enhanced mechanical properties and exceptional optical transparency is beneficial to expand its applications (e.g. in transparent armors and missile domes) [10]. Generally, the microstructural evolution of ceramic could be ⁎

described as the sintering trajectory or sintering path, which commonly shows a positive curvature (i.e. a flat trajectory at low density and a steeper slope at high density). According to Brook's model, a feasible solution for avoiding the pore-grain boundary separation is to keep the sintering trajectory away from the separation zone during the whole sintering stage [7]. The limited grain size with flattened sintering trajectory is beneficial to alleviate the pore separation, which could be realized by shifting densification conditions to quite low temperature or short time [11]. The prerequisite for excellent sintering performance of a green body is the microstructural homogeneity. Agglomeration forms large interagglomerate pores, thereby deteriorating the homogeneity of green body which may lead to coarse grains and intragranular pores at elevated sintering temperature [12]. Since the large pore is more likely to lag behind grain boundary, thus, the homogeneous microstructure could also alleviate the pore separation phenomenon. Various methods are available to eliminate these pores including deaggregation of the raw powders [9,13,14], Ostwald ripening process for powder compact [15], warm-pressing in the sintering body [16], and advanced shaping process using the agglomeration-free powders [14,17]. Besides, a green body with high packing density and fine particles

Corresponding authors. E-mail addresses: [email protected] (H. Wang), [email protected] (Y. Xiong).

https://doi.org/10.1016/j.ceramint.2019.01.048 Received 2 December 2018; Received in revised form 4 January 2019; Accepted 7 January 2019 Available online 08 January 2019 0272-8842/ © 2019 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

Ceramics International 45 (2019) 7544–7551

S. Guo et al.

The microstructure and phase compositions were characterized by SEM (Model S-3400, Japan) and X-ray diffraction (XRD, χ’Pert PRO of Panalytical, Netherlands). The particle size distribution of the synthesized powders was determined by a dynamic light scatterometer (NanoZS ZEN3600 Zetasizer, British). The density was evaluated using the Archimedes method and relative density was derived by assuming the theoretical density of 3.579 g/cm3 for MgAl2O4. The pore size distribution was measured by a mercury intrusion porosimeter (AutoPore IV95/0, USA), and the average pore size and full width at half maximum (FWHM) were obtained by fitting the pore size distribution based on the Gaussian function [13]. Average grain size was determined using linear intercept method (with correction factor 1.56) on SEM micrographs [27]. The in-line transmittance spectrum were recorded using spectrometers (Lambda 750 S, USA for 0.2–1 µm; Model Nexus, WI for 1–7 µm). The Vickers hardness was measured through a hardness tester (Model 430 SVD, China) at a load of 9.8 N for 15 s.

could promote the sintering performance. Both high packing density and fine particles increase the number of particle-particle contact points per volume [18]. In addition, the fine particles could also promote sintering by shortening the distance of mass transport. The packing density is equivalent to the average pore size when evaluating the forming result of the homogenous green body [19]. Krell et al. [17] found that average pore size (defined by pore diameter, D) decreased from 1/3–1/5 of average grain size (G) in green bodies when gelcasting was substituted by slip-casting using the same alumina powder. During the following pressureless sintering, sintering temperature of slip-cast samples was reduced by 40–50 °C compared with the gelcast counterparts. The similar correlation between average pore size and sintering results is also valid for MgAl2O4 [13], TiO2 [20]and ZrO2 [21]. In addition, cold isostatic pressure (CIP) was used to maintain the average pore size at 1/3 G when two MgAl2O4 powders with the particle size of 120 nm and 53 nm were used. For the green body with 53 nm raw powder, the pre-sintering temperature was also decreased from ~1530 °C to ~1250 °C [9]. The above results clearly indicate that the pressureless sintering process could be significantly promoted by either reducing D individually or reducing D and G proportionately. However, the green density by the shaping process under ambient conditions could not exceed the theoretical packing density (74%), for the close-packing of uniform-size spheres [22], which corresponds to a lower bound of D (0.234 G) [19]. In order to surpass this bound, particles must be faceted [23]. As one of the most widely techniques for fast sintering, spark plasma sintering (SPS) is effective for obtaining fine-grained MgAl2O4 ceramic [24,25]. Unfortunately, residual porosity has been reported to deteriorate its transmittance [26]. In spite of these problems, SPS still shows great potential for sustainably reducing the pore size of porous body on the premise of maintaining a homogenous microstructure and restricting grain growth. In this study, we regulated the microstructure of MgAl2O4 porous body by pretreatment process, and investigated its effect on the sintering performance. For this purpose, two porous bodies associated with different microstructures were obtained by pretreatments of SPS and PS, respectively, from coarse raw powder (~180 nm). The sintering performance, microstructural evolution, densification kinetics and grain growth kinetics during subsequent pressureless sintering were investigated in detail. Finally, in-line transmittance and hardness test were conducted on transparent ceramics prepared by post-HIP sintering.

3. Results and discussions 3.1. Characteristic of MgAl2O4 powder The XRD pattern of the synthesized powder is displayed in Fig. 1, which is well with the spinel structure (PDF#77-1193). Fig. 2 shows the particle size distribution of the synthesized powder. The average particle size was determined to be 180 nm, which was also confirmed by the SEM observation (Fig. 2, inset). 3.2. Comparison of microstructure between porous bodies after different pretreatments The fracture and etched surfaces of SPS-88 and PS-88 samples are shown in Fig. 3. The SPS-88 sample shows a fine-grained (230 nm) and homogeneous microstructure, as show in Fig. 3a and b. Comparatively, PS-88 sample shows a larger grain size (730 nm) and a sparse particle space compared with the SPS-88 sample (Fig. 3c and d). It is demonstrated that the pretreatment by SPS is very powerful for suppressing grain growth, as expected. Due to the simultaneous occurrence of densification and grain growth at 1575 °C in pressureless sintering of MgAl2O4 [3], the grain size of PS-88 reached 730 nm, which consumed part of the surface free energy of particles by distinct grain growth. However, the grain growth occurred slightly in SPS-88 when the relative density also reached 88%. The densification may be related to intensive particle rearrangement and the enhancement of diffusion by applied pressure or electric field [28,29]. The pore size distribution of SPS-88 and PS-88 samples is shown in Fig. 4. The SPS-88 sample possessed a smaller average pore size and a

2. Experimental procedure The spinel powder was synthesized by solid-state reaction with 50 mol% MgO (> 99.9%) and 50 mol% α-Al2O3 (> 99.99%). For simplicity, porous bodies were labelled with two parts, the first indicated the pretreatment method, the second demonstrated the relative density (RD). As-synthesized powders was sintered into porous body by SPS apparatus (FCT-HPD60, Germany) at 1425 °C for 1 min under a pressure of 20 MPa (named as SPS-88). Another sample was compacted by CIP (LDJ320/700–400, China) under 200 MPa for 5 min (designated as CIP-49). The sample with 88% RD (named as PS-88) was obtained from CIP-49 after soaking at 1575 °C in air for 1.5 h at a heating rate of 10 °C/ min. For comparison, SPS-88 and PS-88 samples were sintered at 1575 °C in air at a heating rate of 10 °C/min. Moreover, SPS-88 and PS88 were pre-sintered at 1575 °C for 1 h and 4 h, respectively, to achieve closed porosity, and then were densified by HIP at 1500 °C for 5 h under an argon gas pressure of 180 MPa. Similarly, according to their pretreatment methods, the transparent ceramics obtained after HIP (Avure Technologies, USA) were labelled with HIP-SPS-88 and HIP-PS-88, respectively. After grinding and polishing, the final thickness of transparent ceramics for in-line transmittance measurement was 1.0 mm. To observe the grains clearly, the sintered specimens were either fractured or polished and chemically etched with phosphoric acid at 160–210 °C for 10 min.

Fig. 1. XRD patterns of synthesized powder and porous body of SPS-88. 7545

Ceramics International 45 (2019) 7544–7551

S. Guo et al.

Fig. 2. Particle size distribution and SEM image (inset) of the synthesized powder.

Fig. 4. Pore size distribution of SPS-88, CIP-49 and PS-88 samples.

RDT ≈ 1 −

narrower pore size distribution than that of the PS-88 sample. The average pore size, FWHM of pore size distribution, cumulative pore volume and the average grain size are summarized in Table 1. Compared with that of CIP-49, the average pore size of PS-88 increased from 73 nm to 85 nm. Moreover, the FWHM expanded from 16 nm to 36 nm. The average pore size and FWHM of SPS-88 were 28 nm and 10 nm, respectively. Both RDV and RDT were relative density obtained by different methods: RDV was calculated from the cumulative pore volume, while RDT was calculated from the average pore size and the average grain size of sample by assuming that each grain is an equal-sized tetrakaidecahedron with cylindrical pores along its edges [30–32]. The calculation formula of RDV is

RDV =

1 ⎛ 1 /⎜ + ΣVi ⎞⎟ ρth ⎝ ρth ⎠

3π r 2 ⎛ ⎞ 2 2 ⎝l⎠

(2)

where r is the radius of cylindrical pore channel, l the tetrakaidekahedron edge length [31]. When r and l are replaced by pore diameter and grain size [30,32], respectively, Eq. (2) could be rewritten as

RDT ≈ 1 −

3π D 2 ⎛ ⎞ 2 ⎝G ⎠

(3)

where D and G is average pore size and the average grain size, respectively. For SPS-88 and PS-88 samples, the RDT and RDV are consistent with the relative density measured by Archimedes method, which implies that their microstructures were close to the geometric model for the intermediate stage of sintering [33]. For a certain relative density, D is linearly increasing with G in a homogeneous porous body [6,34], which has been confirmed in our experiments (D/G data, as shown in Table 1). The smaller pore size and a narrower pore size distribution of SPS-88 were obtained with the smaller grain size. Furthermore, SPS-88 and PS88 were indeed two geometrically similar sintering systems with

(1)

where ρth is the theoretical density of MgAl2O4, Vi the incremental pore volume with pore diameter of Di . Neglecting variations caused by the intersection of the cylindrical pore channel at the vertices, the solid volume fraction RDT is

Fig. 3. SEM images of fractured and etched surfaces of porous bodies: (a) and (b) SPS-88 sample, (c) and (d) PS-88 sample. 7546

Ceramics International 45 (2019) 7544–7551

S. Guo et al.

Table 1 Pore size distributions and its analysis results. Sample

Grain size (nm)

Average pore size (nm)

FWHM (nm)

Cumulative pore volume (mL/g)

RDV (%)

RDT (%)

Pore size/Grain size

CIP-49 SPS-88 PS-88

180 230 730

73 28 85

16 10 36

0.2476 0.0370 0.0283

53.01 88.30 90.80

– 90.12 90.96

0.405 0.121 0.116

the interconnected pore channels exert a stronger pinning effect on the grain boundaries, which typically results in a retarded grain growth [36]. For SPS-88 sample with a fine and homogenous microstructure (see Table 1), most of the continuous pore network collapsed at higher critical density, which in turn prolonged the intermediate stage of sintering and the pore-pinning effect [15,37]. In the final sintering stage, the pores halt shrinking while the grain growth result in a fractional decrease in the number of pores and thus in a proportional increase in the pore size [6]. The increase of pore size leads to the decrease in its mobility, and when the pore velocity becomes less than the boundary velocity, a pore-grain boundary separation occurred. As shown in Fig. 5, the sintering path ran into the separation zone. As shown in Fig. 6f and h, several grains with trapped pores inside could be occasionally observed in the PS-88 sample after pressureless sintering at 1575 °C for 4 h and 8 h, which is well consistent with Maca's results [38]. With the same grain size or density, SPS-88 has a smaller pore size than PS-88, these pores always attached to the grain boundary and moved with the grain boundary migration. The advantage of SPS88 at the final stage of sintering should be attributed to the reduction of pore size by SPS pretreatment prior to the start of this stage. Furthermore, with the division of the sintering stage by the critical relative density, the densification and grain growth processes can be discussed separately. The densification kinetics of SPS-88 sample is presented in Fig. 7, where the results of PS-88 sample is included for comparison. The prolonged soaking time resulted in the decrease of the densification rate, which was obvious in the PS-88 sample. For SPS-88 sample, the relative density increases rapidly to 98.0% within 1 h and levelled off without significant densification afterward. Moreover, SPS-88 sample had a higher densification rate in the range 88–97% RD. Coble's intermediate stage models have been applied to correlate the densification rate with the grain size during pressureless sintering. Up to now, the oxygen vacancies by lattice diffusion is commonly accepted as the mechanism for controlling the densification of MgAl2O4 [39–41]. The densification by lattice diffusion during intermediate sintering stage is given by [42]

numerical factor slightly larger than 3, which could be substantiated by the ratio of all corresponding features in linear dimensions (G, D and FWHM, as listed in Table 1) [35]. According to the Herring scaling law, the time required for densification will be shortened with the decrease of pore size or grain size, regardless of the diffusion process [35]. The sintering performance of SPS-88 sample should be superior to PS-88 sample, which needs to be evaluated by subsequent pressureless sintering experiments.

3.3. Impact of pretreatment on subsequent pressureless sintering Based on the identical isothermal sintering cycle, the effects of pretreated microstructure on pressureless sintering was investigated. The sintering trajectory of SPS-88 and PS-88 samples was shown in Fig. 5. SPS-88 and PS-88 samples followed two separate sintering trajectories, which mainly depended on the initial grain size. For PS-88 sample, the grain growth is moderated when the RD increase to 95.4%. However, a fast grain growth was triggered when the density exceeded 97.4%. For SPS-88 sample, the flat part of the sintering trajectory extended to 97.2% and the corresponding relative density for rapid grain growth was delayed to 98.0%. For PS-88 and SPS-88, the turning point of the sintering path corresponded to relative densities of 97.4% and 98.0% with grain size of 2.1 µm and 1.2 µm, respectively. In the works of Benameur and co-workers [39], the sintering path with similar trend has been reported in the use of 55 nm MgAl2O4 powder. When the density reached 88% and 98%, the grain size was increased from 55 nm to about 250 nm and 1.0 µm (grain size can be increased at least by a factor of 18), respectively. The grain size of 1.2 µm in our study was close to Benameur's result, when the relative density of 98% was obtained. However, the coarse powder with particle size of 180 nm was used in this study, which indicates a pronounced effect of pretreated microstructure on sintering performance. At the critical relative density 97.4% and 98.0%, the open porosity cannot be detected in PS-88 and SPS-88 by the Archimedes test, respectively. In the meantime, the continuous pore channels were completely pinched off to form closed pores which were located at grain corners. Compared with the isolated pore at the final sintering stage,

Aγa03 Dl dP 1 =− × 3 dt kT G

(4)

where

P = 1−ρ

(5)

a03

where A is a numerical constant, γ is the surface energy, is the volume of the diffusing species, Dl is the diffusion coefficient, k is the Boltzmann's constant, T is the absolute temperature, G is the grain size, P is the porosity, and ρ is the relative density. If T is fixed, the first term in the right of Eq. (4) is constant and the densification rate is converted to be grain size-dependent. It seems that the densification kinetics of SPS-88 and PS-88 samples were derived from the difference in initial grain size. Actually, it is always the cube of the instantaneous grain size, which is inversely proportional to the instantaneous densification rate (Eq. (4)). The decrease of the densification rate of samples comes from unwanted grain growth during the high temperature soaking (Fig. 6a for SPS-88, Fig. 6b and d for PS-88). Similarly, the higher densification rate of SPS-88 sample throughout the intermediate stage of sintering could be primarily attributed to a lower level of instantaneous grain size in SPS-88.

Fig. 5. Sintering trajectories of porous bodies SPS-88 and PS-88 during pressureless sintering at 1575 °C. (The corresponding separation zone is schematically illustrated on the basis of SEM observations.). 7547

Ceramics International 45 (2019) 7544–7551

S. Guo et al.

Fig. 6. SEM images of etched surfaces of sample SPS-88 and PS-88 pressureless sintering at 1575 °C: (a, c, e, g) SPS-88 sample for 10 min, 1 h, 4 h, 8 h, (b, d, f, h) PS88 sample for 10 min, 1 h, 4 h, 8 h. (The entrapped pores in the sintered sample was highlighted by open circles.).

is the absolute temperature [43,44]. Nevertheless, the used grain growth model assumed a normal grain growth in a nearly full dense and polycrystalline material. In this study, as-sintered sample with closed porosity was selected to ensure that the use of Eq. (6) is valid. The grain growth exponent m was adjusted to provide the best fit to the experiment data [45]. Those data yielded an excellent fit to a m value of 3 (R2 = 0.9849 and 0.9959 for SPS-88 and PS-88 respectively, see Fig. 8). For the cubic grain growth kinetics, the following two possible grain growth mechanisms could occur in pure system [45,46]:

To identify the mechanism for grain growth, it is common to use the follow relation

G m − G0m = k (T )(t − t0)

(6)

where

Qg ⎞ k (T ) = k 0 exp ⎜⎛− ⎟ RT ⎠ ⎝

(7)

where G is the reference grain size at time t, G0 is the average grain size at time t0, m is an exponent characteristic of the grain growth mechanism, k0 is the pre-exponential constant of the diffusion coefficient, Qg is the activation energy for grain growth, R is the gas constant, and T

(i) Pore displacement controlled by lattice diffusion, (ii) Pore displacement controlled by vapor transport (m = 2 or 3). 7548

Ceramics International 45 (2019) 7544–7551

S. Guo et al.

Fig. 7. Densification kinetics of porous bodies SPS-88 and PS-88 during pressureless sintering at 1575 °C.

Fig. 10. In-line transmittance spectrum of transparent ceramics HIP-SPS-88 and HIP-PS-88 and the inset shows their apparent photos.

growth of SPS-88 sample led to small pore growth. In addition to the small initial pores of SPS-88, the pore-grain boundary separation becomes more difficult. The microstructure of porous body and kinetic analysis of SPS-88 sample are affirmative factors for inhibiting the pore-grain boundary separation. Firstly, SPS-88 sample has smaller grain size, which ensured a low-level starting point in the sintering trajectory. Secondly, the shorted sintering time and the delayed pore closure limited the grain growth during intermediate stage of sintering. Finally, the smaller grain size, as well as sluggish grain coarsening, maintained the fine pores in the final stage of sintering. These pores with high mobility could move together with the grain boundary, thereby the pore-grain boundary separation was suppressed. 3.4. Transparent ceramics after hot isostatic pressing Fig. 8. Grain growth kinetics of porous bodies SPS-88 and PS-88 during pressureless sintering at 1575 °C.

A complementary HIP step was performed on pre-sintered samples (SPS-88 pressureless sintered at 1575 °C for 1 h and PS-88 pressureless sintered at 1575 °C for 4 h). Probably, the pore-boundary separation occurs during pressureless sintering by prolonging the sintering time. However, the high pressure forced the pores to shrink with limited grain size during HIP, thereby reducing the chance of separation [44]. The microstructure of transparent ceramics HIP-SPS-88 and HIP-PS-88 is shown in Fig. 9. In the case of HIP-SPS-88, a monodispersed and homogeneous microstructure was maintained (Fig. 9a). The HIP-PS-88 exhibited rather broad grain size distribution with the biggest grains exceeding 35 µm, which indicates that the abnormal grain growth occurred (Fig. 9b). The average grain size was measured to be 1.9 µm, which showed a relatively higher hardness (13.9 ± 0.2 GPa) than that of HIP-PS-88 (13.2 ± 0.4 GPa). The hardness value of both HIP-SPS-88 and HIP-PS-88 compares favorably with the results in literatures [51,52].

The volatilization of MgO occurs at high temperature, although MgAl2O4 with high-melting (> 2100 °C) [47]. It is extremely severe over the surface of the sample (0.15 mm) and under vacuum [48]. Considering that our sintering process was conducted in the air and the bulk was embedded with protective MgAl2O4 powder, the vapor transport for mass transport should be more difficult to be implemented, whereas the lattice diffusion could be favored. Using k0 reported by Bratton and the fitted k [3,49], we calculated the Qg of SPS88 and PS-88 to be 454 and 434 kJ/mol, respectively, according to Eq. (7). The activation energy for grain growth is close to the reported values of Bratton (462 kJ/mol) and Chiang (422 kJ/mol) [3,50]. Controlled by the identical solid-state diffusion mechanism, slow grain

Fig. 9. SEM images of etched surfaces of MgAl2O4 spinel transparent ceramics: (a) HIP-SPS-88, (b) HIP-PS-88. 7549

Ceramics International 45 (2019) 7544–7551

S. Guo et al.

Table 2 The experiment parameters and properties of MgAl2O4 transparent ceramics prepared by pressureless sintering and HIP in the literatures. Powder

Primary particle size

Pre-sintering

Post-HIP

Grain size

Transmittance (thickness, wavelength)

Reference

FSP

20–50 nm 53 nm

1280 °C/2 h 1260 °C 1400 °C/80 h

1320 °C/200 MPa/3 h 1260 °C/200 MPa/15 h 1500 °C/200 MPa/3 h

G¯ = 0.45 µm G¯ = 0.34 µm G¯ = 2.2 µm

80% (2 mm, 635 nm) 84% (3.7 mm, 640 nm) 63% (2 mm, 635 nm)

Goldstein [55] Krell [9] Goldstein [56]

S30CR

55 nm

1500 °C/2 h 1550 °C/5 min

1450 °C/180 MPa/5 h 1500 °C/200 MPa/1 h

79.3% (1 mm, 550 nm) 60.2% (1.1 mm, 632.8 nm)

Kim [13] Maca [38]

80% (4 mm, 550 nm)

Ramavath [57]

FSP Our research

120 nm 180 nm

35% (3.7 mm, 640 nm) 63% (1 mm, 550 nm)

Krell [9] This study

1650 °C/1 h

1800 °C/ 195 MPa/5 h

1540 °C 1575 °C/1 h

1550 °C/200 MPa/15 h 1500 °C/ 180 MPa/5 h

GM = 8 µm G¯ = 1.4 µm G¯ = 1 µm GM = 5 µm G¯ = 40–50 µm G¯ = 4.3 µm G¯ = 1.9 µm

Note: G¯ is average grain size. GM is maximum grain size.

Acknowledgements

Fig. 10 shows the in-line transmittance spectrum and photo of transparent ceramics. The in-line transmittance of HIP-SPS-88 reached 62% at 550 nm wavelength, which showed a higher transmittance than that of HIP-PS-88 (32%). The transmittance of HIP-SPS-88 reached to 83% range over 1.0–5.0 µm. In polycrystalline ceramics with cubic crystal structure, the residual porosity is a dominant factor for degrading the transmittance [53]. Especially, pores with a size near 1/10 of the incident light's wavelength severely deteriorate the transmittance according to Mie scattering theory [54]. Therefore, the reason for the higher transmittance of HIP-SPS-88, especially, in the visible light region, should be light scattering by residual pores. Table 2 summarizes the experiment parameters and properties of MgAl2O4 transparent ceramics, which were prepared from pure phase powder by pressureless sintering and HIP. The most outstanding result was reported by Krell and Goldstein using powder FSP, where the sintering temperature was reduced to < 1300 °C, therefore, the grain size less than 0.5 µm was obtained [9,55]. In addition, independent research by Kim [15], Maca [41], and Ramavath [59] showed that lowering the sintering temperature could effectively reduce grain size. However, the reduction of the sintering temperature in the studies above should attributed to the improvement of the sintering performance of the fine powders. Interestingly, both Goldstein and Maca found abnormal grain growth, which also appeared in our sample HIP-PS-88 [38,56]. This phenomenon may be related to inferior microstructure of the green bodies because all of them was formed by cold isostatic pressing. In addition, we use the coarsest powder (180 nm) as tabulated in Table 2, but the final grain size of the product is similar with that obtained by Kim and Maca using 55 nm powder. By contrast, the innovative pretreatment by SPS in this study led to a trade-off result as a grain size of 1.9 µm and a comparable visible transmittance.

The authors would like to acknowledge the supports from National Key R&D Program of China (No. 2017YFB0310500) and National Natural Science Foundation of China (Nos. 51472195 and 51502219). References [1] A. Goldstein, A. Krell, Transparent ceramics at 50: progress made and further prospects, J. Am. Ceram. Soc. 99 (2016) 3173–3197. [2] R.D.M. Marc, I.E. Reimanis, C. Smith, H.J. Kleebe, M.M. Müller, Effect of impurities and LiF additive in hot-pressed transparent magnesium aluminate spinel, Int. J. Appl. Ceram. Technol. 10 (2012) 1–16. [3] R.J. Bratton, Sintering and grain-growth kinetics of MgAl2O4, J. Am. Ceram. Soc. 54 (1970) 141–143. [4] S. Kochawattana, A. Stevenson, S.H. Lee, M. Ramirez, V. Gopalan, J. Dumm, V.K. Castillo, G.J. Quarles, G.L. Messing, Sintering and grain growth in SiO2 doped Nd:YAG, J. Eur. Ceram. Soc. 28 (2008) 1527–1534. [5] S.J.L. Kang, Grain growth and densification in porous materials, 2005. [6] J.L. Shi, Relation between coarsening and densification in solid-state sintering of ceramics: experimental test on superfine zirconia powder compacts, J. Mater. Res. 14 (1999) 1389–1397. [7] R.J. Brook, Pore-grain boundary interactions and grain growth, J. Am. Ceram. Soc. 52 (1969) 56–57. [8] T. Sone, H. Akagi, H. Watarai, Effect of pore-grain boundary interactions on discontinuous grain growth, J. Am. Ceram. Soc. 74 (2010) 3151–3153. [9] A. Krell, T. Hutzler, J. Klimke, A. Potthoff, Fine-grained transparent spinel windows by the processing of different nanopowders, J. Am. Ceram. Soc. 93 (2010) 2656–2666. [10] A. Krell, T. Hutzler, J. Klimke, Transmission physics and consequences for materials selection, manufacturing, and applications, J. Eur. Ceram. Soc. 29 (2009) 207–221. [11] S.J. Bennison, M.P. Harmer, Effect of magnesia solute on surface diffusion in sapphire and the role of magnesia in the sintering of alumina, J. Am. Ceram. Soc. 73 (2010) 833–837. [12] J.L. Shi, J.H. Gao, Z.X. Lin, T.S. Yen, Sintering behavior of fully agglomerated zirconia compacts, J. Am. Ceram. Soc. 74 (2010) 994–997. [13] J.M. Kim, H.N. Kim, Y.J. Park, J.W. Ko, J.W. Lee, H.D. Kim, Fabrication of transparent MgAl2O4 spinel through homogenous green compaction by microfluidization and slip casting, Ceram. Int. 41 (2015) 13354–13360. [14] M.J. Mayo, Processing of nanocrystalline ceramics from ultrafine particles, Int. Mater. Rev. 41 (1996) 85–115. [15] F.J.T. Lin, L.C.D. Jonghe, M.N. Rahaman, Microstructure refinement of sintered alumina by a two-step sintering technique, J. Am. Ceram. Soc. 80 (1997) 2269–2277. [16] J.W. Son, D.Y. Kim, P. Boch, Enhanced densification of In2O3 ceramics by presintering with low pressure (5 MPa), J. Am. Ceram. Soc. 81 (2010) 2489–2492. [17] A. Krell, J. Klimke, Effects of the homogeneity of particle coordination on solid-state sintering of transparent alumina, J. Am. Ceram. Soc. 89 (2006) 1985–1992. [18] J.R. Lee, T.J. Chung, S.H. Yang, G.S. Hong, K.S. Oh, Prevention of tapering in the tube-shaped sputtering target via initial heat treatment under external pressure, Ceram. Int. 41 (2015) 3677–3683. [19] S.M. Sweeney, M.J. Mayo, Relation of pore size to green density: the Kozeny equation, J. Am. Ceram. Soc. 82 (2010) 1931–1933. [20] E.A. Barringer, H.K. Bowen, Effects of particle packing on the sintered microstructure, Appl. Phys. A 45 (1988) 271–275. [21] M. Trunec, K. Maca, Compaction and pressureless sintering of zirconia nanoparticles, J. Am. Ceram. Soc. 90 (2010) 2735–2740. [22] H.E. White, S.F. Walton, Particle packing and particle shape, J. Am. Ceram. Soc. 20 (1937) 155–166. [23] J. Liu, D.P.D. Lo, Particle rearrangement during powder compaction, Metall. Trans. A 32 (2001) 3117–3124. [24] K. Morita, B.N. Kim, H. Yoshida, H. Zhang, K. Hiraga, Y. Sakka, Effect of loading

4. Conclusions The pretreatment by SPS is an effective way to manipulate the microstructure of porous body, especially in terms of significantly reducing pore size (28 nm) and limiting grain growth (230 nm). An enhanced densification and limited grain growth were observed during the subsequent pressureless sintering. The net result is to displace the sintering path to a lower grain size for a given density, thereby, the pore-grain boundary separation is avoided. After the final densification by HIP, a MgAl2O4 transparent ceramic with a combination of high transmittance and fine grain size was fabricated. The in-line transmittance was remained at 83% throughout 1.0–5.0 µm, and the hardness (13.9 ± 0.2 GPa) was enhanced by decreasing the grain size to 1.9 µm. The innovative approach in this study could also be an alternative for sintering other ceramics, especially, where the enhancement of the sintering performance of the available powder is desired.

7550

Ceramics International 45 (2019) 7544–7551

S. Guo et al.

[25] [26]

[27] [28]

[29] [30] [31]

[32]

[33] [34] [35] [36] [37]

[38] [39] [40] [41]

[42] R.L. Coble, Intermediate-stage sintering: modification and correction of a latticediffusion model, J. Appl. Phys. 36 (1965) (2327–2327). [43] J. Wang, Z. Fang, C. Feng, Z. Jian, H. Zhang, R. Tian, Z. Wang, J. Liu, Z. Zhao, C. Shi, Effect of Y2O3 and La2O3 on the sinterability of γ-AlON transparent ceramics, J. Eur. Ceram. Soc. 35 (2015) 23–28. [44] L. Chrétien, L. Bonnet, R. Boulesteix, A. Maître, C. Sallé, A. Brenier, Influence of hot isostatic pressing on sintering trajectory and optical properties of transparent Nd: YAG ceramics, J. Eur. Ceram. Soc. 36 (2016) 2035–2042. [45] F.A. Nichols, Theory of grain growth in porous compacts, J. Appl. Phys. 37 (1966) 4599–4602. [46] R.J. Brook, Pores and grain growth kinetics, J. Am. Ceram. Soc. 52 (1969) 339–340. [47] E.F. Osborn, Subsolidus reactions in oxide systems in the presence of water at high pressures, J. Am. Ceram. Soc. 36 (2010) 147–151. [48] C.J. Ting, H.Y. Lu, Deterioration in the final-stage sintering of magnesium aluminate spinel, J. Am. Ceram. Soc. 83 (2010) 1592–1598. [49] Y. Mordekovitz, L. Shelly, M. Halabi, S. Kalabukhov, S. Hayun, The effect of lithium doping on the sintering and grain growth of SPS-processed, non-stoichiometric magnesium aluminate spinel, Materials 9 (2016) 481. [50] Y.M. Chiang, W.D. Kingery, Grain-boundary migration in nonstoichiometric solid solutions of magnesium aluminate spinel: II, effects of grain-boundary nonstoichiometry, J. Am. Ceram. Soc. 73 (2010) 1153–1158. [51] A. Krell, A. Bales, Grain size-dependent hardness of transparent magnesium aluminate spinel, Int. J. Appl. Ceram. Technol. 8 (2011) 1108–1114. [52] M. Sokol, M. Halabi, S. Kalabukhov, N. Frage, Nano-structured MgAl2O4 spinel consolidated by high pressure spark plasma sintering (HPSPS), J. Eur. Ceram. Soc. 37 (2017) 755–762. [53] A. Ikesue, K. Yoshida, Influence of pore volume on laser performance of Nd: YAG ceramics, J. Mater. Sci. 34 (1999) 1189–1195. [54] R. Apetz, M.P.B. Van Bruggen, Transparent alumina: a light-scattering model, J. Am. Ceram. Soc. 86 (2010) 480–486. [55] A. Goldstein, A. Goldenberg, M. Hefetz, Transparent polycrystalline MgAl2O4 spinel with submicron grains, by low temperature sintering, J. Ceram. Soc. Jpn. 117 (2009) 1281–1283. [56] A. Goldstein, A. Goldenberg, Y. Yeshurun, M. Hefetz, Transparent MgAl2O4 spinel from a powder prepared by flame spray pyrolysis, J. Am. Ceram. Soc. 91 (2008) 4141–4144. [57] P. Ramavath, P. Biswas, K. Rajeswari, M.B. Suresh, R. Johnson, G. Padmanabham, C.S. Kumbhar, T.K. Chongdar, N.M. Gokhale, Optical and mechanical properties of compaction and slip cast processed transparent polycrystalline spinel ceramics, Ceram. Int. 40 (2014) 5575–5581.

schedule on densification of MgAl2O4 spinel during spark plasma sintering (SPS) processing, J. Eur. Ceram. Soc. 32 (2012) 2303–2309. C. Wang, Z. Zhao, Transparent MgAl2O4 ceramic produced by spark plasma sintering, Scr. Mater. 61 (2009) 193–196. K. Morita, B.N. Kim, H. Yoshida, K. Hiraga, Y. Sakka, Influence of pre- and postannealing on discoloration of MgAl2O4 spinel fabricated by spark-plasma-sintering (SPS), J. Eur. Ceram. Soc. 36 (2016) 2961–2968. M.I. Mendelson, Average grain size in polycrystalline ceramics, J. Am. Ceram. Soc. 52 (1969) 443–446. Z.A. Munir, U. Anselmi-Tamburini, M. Ohyanagi, 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. 41 (2006) 763–777. C. Liu, M. Xiang, Z. Fu, Z. Shen, Y. Xiong, Microstructural refinement in spark plasma sintering 3Y-TZP nanoceramics, J. Eur. Ceram. Soc. 36 (2016) 2565–2571. J. Zhao, M.P. Harmer, Effect of pore distribution on microstructure development: I, matrix pores, J. Am. Ceram. Soc. 71 (1988) 113–120. W.S. Slaughter, I. Nettleship, M.D. Lehigh, P.O. Tong, A quantitative analysis of the effect of geometric assumptions in sintering models, Acta Mater. 45 (1997) 5077–5086. R.M. German, Coarsening in sintering: grain shape distribution, grain size distribution, and grain growth kinetics in solid-pore systems, Crit. Rev. Solid State Mater. Sci. 35 (2010) 263–305. R.L. Coble, Sintering crystalline solids: I, intermediate and final state diffusion models, J. Appl. Phys. 32 (1961) 787–792. P.L. Chen, I.W. Chen, Sintering of fine oxide powders: I, microstructural evolution, J. Am. Ceram. Soc. 79 (1996) 3129–3141. C. Herring, Effect of change of scale on sintering phenomena, J. Appl. Phys. 21 (1950) 301–303. R. Chaim, M. Kalina, J.Z. Shen, Transparent yttrium aluminum garnet (YAG) ceramics by spark plasma sintering, J. Eur. Ceram. Soc. 27 (2007) 3331–3337. F.S. Shiau, T.T. Fang, T.H. Leu, Effect of particle-size distribution on the microstructural evolution in the intermediate stage of sintering, J. Am. Ceram. Soc. 80 (2010) 286–290. K. Maca, M. Trunec, R. Chmelik, Processing and properties of fine-grained transparent MgAl2O4 ceramics, Ceram. Silik. 51 (2007) 94–97. C.J. Ting, H.Y. Lu, Defect reactions and the controlling mechanism in the sintering of magnesium aluminate spinel, J. Am. Ceram. Soc. 82 (1999) 841–848. I. Reimanis, H.J. Kleebe, A review on the sintering and microstructure development of transparent spinel (MgAl2O4), J. Am. Ceram. Soc. 92 (2009) 1472–1480. Y. Oishi, K. Ando, Self-diffusion of oxygen in polycrystalline MgAl2O4, J. Chem. Phys. 63 (1975) 376–378.

7551