In fl uence of forming method and sintering process on densi fi cation and fi nal microstructure of submicrometre alumina ceramics

Conditions of green body preparation and consequently the conditions of sintering are prerequisite for the preparation of dense samples of Al2O3 with superior optical and mechanical properties. The goal of this work was to determine the optimum forming conditions for preparation of green body, and to fi nd out the optimal sintering regime facilitating the preparation of ultra-fi ne grained high purity alumina with maximal density, fi ne microstructure and small pore size. Axial pressing followed by CIP was found to yield green body with the highest density and narrowest pore size distribution. In the two-stage sintering regime the temperature T1 has to be higher or equal to 1300°C to obtain closed porosity unstable against shrinkage. The temperature T2 ≥ 1130°C was found to be too high for suppression of grain growth in the fi nal stage of sintering and sintering trajectory was identical with that from standard sintering regime.


I. Introduction
The uniform packing of particles in green (unfi red) bodies is a critical precondition for the preparation of dense, defect-free and pore-free ceramics with fi ne grained microstructure and desired optical and mechanical properties [1,2].Refi nement of the grain size of polycrystalline alumina below the wavelength of visible light is expected to result in increased optical transparency [3].In order to achieve such sub-micrometer grain sizes, various sintering, mostly pressure-assisted, techniques are usually applied, with all their disadvantages and limitations.
Chen & Wang reported on a pressureless two-stage sintering process, which was applied with success for densifi cation of a nano-meter sized yttria powder without the fi nal stage grain growth.They postulated that in certain interval of temperatures, called "kinetic window", densifi cation is already in operation, while the grain boundary motion is not yet activated [4].In this sintering regime a powder compact is sintered at a higher temperature T 1 until the residual porosi-ty becomes unstable against further shrinkage.Then the sintering temperature is lowered to the value T 2 at which densifi cation proceeds further, but no grain growth occurs.However, the application of the twostage sintering for alumina is questionable, as some works suggest that the activation energy of densifi cation in alumina is in fact higher than the activation energy of grain growth [5].Our previous work suggests that under a suitable two-stage sintering regime a refi nement of microstructure can be observed in comparison to polycrystalline alumina sintered in a standard way.Despite observed microstructure refi nement the grain growth was not suppressed entirely, and the sintered specimens still contained about 1 % of residual porosity concentrated in less dense regions.These were created in the course of green body forming by uniaxial pressing [6].
The ultimate goal of our work is to map the twostage sintering process of submicrometre alumina more thoroughly, and to identify unambiguously the temperature interval of "kinetic window", where the residual porosity in alumina ceramics prepared from a submicrometre powder can be eliminated without the fi nal stage grain growth.This paper specifi cally deals with the infl uence of pressure conditions (axial pressing and cold isostatic pressing -CIP) on microstructure of green bodies (and sintered samples) shaped from variously treated sub-micrometre-sized alumina powder (Taimicron TM DAR) and on their densifi cation.Further step was to fi nd out optimal sintering regime facilitating the preparation of ultra-fi ne grained high purity alumina with maximum density, the fi nest microstructure and the lowest pore size.

II. Experimental
An ultra-fi ne grained high purity (99.995 %) alumina powder (Taimicron TM DAR) with the primary particle size of 150 nm was used for the experiments (denoted as powder S).The powder was then either milling activated (denoted as powder SM), or granulated (denoted as powder TG) in order to evaluate the infl uence of the powder state on green body forming and sintering.Powder characteristics are shown in Table 1.
For the optimisation of compacting pressure three different pressures 50, 100 and 150 MPa were applied in axial loading, and 50 MPa axial loading followed by cold isostatic pressing, CIP, at 250 MPa. Green density of pressed samples was measured by Archimedean method in mercury.Pore size distribution of green bodies was measured by mercury intrusion porosimetry.The pressed samples were sintered for 10 minutes at 1350°C and then the density of sintered samples was measured by Archimedean method in distilled water.The relative density of specimens (ρ rel ) was related to the theoretical density of alumina 3.98 g .cm -3 .To fi nd out the relation between sintering temperature and relative densities the sample S-CIP was subjected to 5 various sintering regimes including heating at 10°C/min to a maximum temperature (1250, 1275, 1300, 1325 and 1350°C respectively) without dwell time, and subsequent cooling to room temperature.Based on the results of preliminary sintering experiments (relative density and the mean grain size determined by linear intercept method (minimum of 200 intercepts) from SEM micrographs of polished and thermally etched (1125°C for 8 hours) cross sections of the specimens three different temperatures T 1 = 1300, 1325 and 1350°C were selected, which facilitated elimination of the residual porosity to less than 15 %.These were then used for further experiments in the second stage, with special focus on 1300°C, which yielded specimens with the smallest mean grain size.Three different temperatures T 2 (1130, 1150 and 1170°C) at four different dwell times (2, 4, 8 and 24 hours) were applied and mean grain size was determined by linear intercept method from fracture surfaces.

III. Results and Discussion
With the selected powder a set of experiments was carried out, which resulted in determination of the minimum temperature, where a stage of closed porosity will be achieved, with the mean diameter of pores facilitating their closure in the second stage of sintering.For this, a preparation of suffi ciently dense green body with narrow pore size distribution is vital.The highest green density was achieved for the granulated powder pressed isostatically; only small differences were found for the as received and milling activated powders S and SM (Fig. 1).Mercury porosimetry showed signifi cant decrease of the overall pore volume, shift of the mean pore size, and more narrow distribution of pore sizes in CIP-ed specimens (Fig. 2) especially for the powders S and SM.This was caused by two complementary effects -isostatic nature of the applied pressure and higher applied pressure.The broad pore size distribution in the case of the powder TG is the consequence of two types of porosity: the pores inside individual granules (smaller), and those among them (larger).The untreat- ed S powder after CIP yielded green bodies with lowest pore radius and with suffi ciently narrow pores size distribution for sintering experiments.This was confi rmed by preliminary sintering test where the S-CIP green bodies achieved suffi ciently high green density to justify their use in two-stage sintering experiments (Fig. 3).The sintering at various temperatures without isothermal dwell indicate that the temperature as low as 1300°C is suffi cient for elimination of the residual porosity to about 15 % (Table 2).Determination of the grain size by linear intercept method moreover revealed the fi nest mean grain size of 270 nm for the specimen sintered at 1300°C without isothermal dwell (Table 2).A prerequisite for successful densifi cation during the second, low-temperature step of sintering is that the pores become subcritical and unstable against shrinkage.This stage is, according to the original paper of Chen and Wang [4], achieved in nanosized yttria when the relative density exceeds 70 %TD.They conclude that the relative density of 75 %TD after the fi rst step is suffi cient for complete densifi cation of yttria by low-temperature heat treatment in the kinetic window [4].However, our experience has shown that 75 %TD is not suffi cient in case of the used alumina powder, and the stage near to the closed porosity, which in case of alumina corresponds to about 90 %TD, must be achieved before the second, low temperature sintering step is applied [6].The sintering trajectories (relative density -grain size dependences) of the specimens after the two-stage sintering are summarised in Fig. 4 and Table 3 and compared with the specimens sintered in a standard way, i.e. by a routine pressureless sintering with maximum temperature between 1100 and 1350°C and dwell time between 0 and 60 min.The microstructures of specimens sintered at the T 1 =1130°C and afterwards at T 2 =1300°C for 2 h and those sintered at T 1 =1130°C and afterwards at T 2 =1300°C for 24 h are shown in the Fig. 5.After the second step in all cases the mean grain size of our samples, as well as the

Relative density [% TD]
Green body Sintered samples relative density increased with increasing isothermal dwell at T 2 .Interestingly, at all temperatures the sintering trajectories fell into a single line identicalwith the sintering trajectory of specimens sintered with the use of a standard one-stage regime.The only exception is the specimen, which achieved the relative density of 99 %TD at the 720 nm mean grain size after 24 h dwell at T 2 =1170°C.The mean grain size of a stand-ard way sintered specimen with comparable relative density (99.2 %TD) was 960 nm.It therefore appears that in all reported cases the T 2 (1130, 1150 and 1170°C) was too high to facilitate effi cient suppression of the grain growth.Further experiments at lower T 2 are therefore required to verify, or reject defi nitely the applicability of the two-stage sintering for densifi cation of alumina without the fi nal stage

Conclusions
Axial pressing of as-received ultra-fi ne alumina powder followed by CIP was selected as an optimum green body forming method.CIP reduces the porosity of green bodies signifi cantly narrows the pore size distributions and shifts the mean values of pore necks diameters to lower values in comparison to axially pressed pellets.This is refl ected also in better densifi cation of CIP-ed specimens.The sintering at temperatures ≥ 1300°C without dwell eliminates the residual porosity to an extent, which gives a chance for complete densifi cation without fi nal stage grain growth during the second, low temperature sintering step.Temperature T 2 ≥ 1130°C results in further densifi cation of alumina compacts, but the grain growth is not suppressed.Further experiments at lower temperatures T 2 are required to verify, or reject, the applicability of two-stage sintering for suppression of the fi nal stage grain growth in polycrystalline alumina.

Figure 2 .
Figure 2. Radius of pore necks of green bodies before and after CIP

Sintering trajectories of polycrystalline alumina sintered in a standard way and by two-stage sintering
grain growth.Moreover, further investigations involving the two-stage sintering experiments with green bodies prepared by wet forming methods (pressure fi ltration) are under way to evaluate further the infl uence of green microstructure on the sintering trajectory.