Preparation and characterisation of alumina-based composite support layers

Introduction

Porous ceramics have recently attracted considerable attention ¹ as researchers have endeavoured to exploit their unique properties, such as low density, ² low thermal conductivity^3,4 and a low dielectric constant. ⁵ Many types of porous ceramics have been developed over many years, going back centuries, including γ-alumina,^6,7 α-alumina,^8,9 titania,^10,11 zirconia, ¹² silica ¹³ and composites of these materials.^14,15

In particular, ceramic membranes^6,8, ¹⁶ are among the most useful applications of porous ceramics. For these, it is important to maintain precise control of pore size, mechanical strength and permeability while minimising processing costs.

One of the most significant properties of porous ceramic membranes is their permeability. Higher permeability is generally obtained by increasing porosity, pore size or pore interconnectivity. Permeability can be enhanced by several methods, such as incomplete sintering or the use of a sacrificial polymer template, for example, alumina, ¹⁷ titania, ¹⁸ lanthanum strontium manganite ¹⁹ and clay. ²⁰ However, these methods inevitably reduce the mechanical strength of the membranes. ¹ Because of this trade-off between mechanical strength and permeability with current methods, ²¹ much recent research has focused on the optimisation of both mechanical strength and permeability. ²²

In this study, to enhance simultaneously the air permeability and flexural strength of a porous ceramic support layer, we investigate the feasibility of adding diatomite as a pore former and bonding phase.

First, it is expected that diatomite can act as a pore former owing to the inter-particle voids induced by the irregular shapes of the diatomite particles and the inherent pores inside these particles. Additionally, within the microstructure, microcracks, which form an escape path for the gas phase that is generated during pyrolysis of the sacrificial polymer beads, can be avoided. In addition, the sintering time required for complete pyrolysis can be reduced. ¹ Furthermore, elimination of the expensive sacrificial polymer template would reduce the high cost that acted as a barrier to wider utilisation of porous ceramic membranes.

Second, diatomite might also act as a bonding phase, increasing the mechanical strength, because the sintering temperature of an alumina support layer (1400°C) is higher than the temperature (1300°C) required for initiation of coalescence and collapse of porous diatom frustules and grains in the diatomite matrix.

Diatomite is a sedimentary rock formed from the siliceous fossilised skeletons of diatoms, which are composed of rigid cell walls called frustules.^23-26 We have previously reported on possible approaches for the utilisation of diatomite that allow control over the pore characteristics of a porous ceramic membrane. ^{27 35}

As a possible diatomite substitute, the effects of pyrophyllite addition to an alumina support layer have also been investigated. Both diatomite and pyrophyllite are composed mainly of silica^36-38 and their average particle size is similar. However, pyrophyllite particles are plate-like in shape, which presents a disadvantage in terms of pore formation compared with diatomite. Also, the impurity level in pyrophyllite is far lower than that in the diatomite, and therefore there may not be as effective a reduction in the eutectic temperature with silica (as a bonding phase). Therefore, we decided that a comparison of the effects of pyrophyllite addition and those of diatomite addition warranted further investigation.

The present study considered several important aspects. First, it investigated whether the flexural strength and air permeability of an alumina–diatomite composite support layer could be enhanced simultaneously beyond their trade-off relationship. Second, it was investigated whether the pore characteristics of the alumina–diatomite composite support layers could be controlled by tailoring the alumina matrix through the addition of the diatomite. Third, in a comparative investigation, alumina–pyrophyllite composite support layers were also prepared and characterised.

Materials and methods

α-Alumina (AM-210, Sumitomo Chemical Co., Ltd, Japan), diatomite (Celite 499, Celite Korea Co., Ltd, Korea) and pyrophyllite (Korea Powder Co., Ltd, Korea) were used for the preparation of the porous ceramic support layers. The average particle sizes of the starting powders were determined using a particle size analyser (LSTM 13 320 MW, Beckman Coulter, USA). The average particle sizes of the as-received pyrophyllite, the as-received diatomite and the as-received alumina were 6.95, 7.43 and 4.80 µm, respectively.

Distilled water was used as a solvent, and the slurry was ball-milled for 24 h with an alumina ball-to-powder volume ratio of 2:1. After ball-milling, the slurry was dried at room temperature for 24 h. To incorporate diatomite or pyrophyllite into the alumina matrix, alumina and diatomite/pyrophyllite were mixed for 3 h by ball-milling with a ball-to-powder volume ratio of 0.5:1. With the use of polyethylene glycol as a binder, the alumina-based mixture was dry-pressed at 18.7 MPa and then sintered between 1400 and 1600°C for 1 h.

For comparison, alumina–pyrophyllite support layers were prepared by following the same procedures. The pore characteristics of the diatomite were investigated using scanning electron microscope (SEM; JSM-5800, JEOL, Japan) and mercury porosimetry (Autopore IV 9510, Micromeritics, USA). The air permeability was measured using capillary flow porosimetry (CFP-1200-AEL, Porous Materials, Inc., USA). A sintered specimen (of diameter 4 cm and thickness 0.4 cm) was inserted between the O-rings at the bottom of the chamber, and the bottom of the chamber was inserted into the capillary flow porosimeter. The flux was subsequently measured automatically by sensors, as the diameter of the motorised valve and the pressure of the regulator were increased incrementally.

Results and discussion

Figure 1 shows the flexural strengths and air permeabilities of typical porous ceramic membranes. First, the flexural strength of an alumina support layer sintered at 1400°C was significantly increased by increasing the sintering temperature to 1600°C. However, the air permeability was slightly decreased by this increase in sintering temperature. Second, the flexural strength of a diatomite support layer sintered at 1200°C could be significantly increased by incorporating 15 wt-% kaolin into the diatomite matrix. However, the air permeability of the diatomite–kaolin composite support layer decreased markedly. ³⁵ Third, the flexural strength of a pyrophyllite support layer sintered at 1200°C was significantly increased by increasing the sintering temperature to 1400°C. However, as with the alumina support layer, the air permeability of the pyrophyllite support layer decreased when the sintering temperature was increased. Fourth, the air permeability of a pyrophyllite support layer sintered at 1200°C was enhanced by incorporating diatomite into the pyrophyllite matrix, although at the cost of a decrease in flexural strength. ²⁷ Thus, there is a trade-off between mechanical strength and the air permeability for all of these porous ceramic membranes.

OPEN IN VIEWER

1

Flexural strengths and air permeabilities of an alumina support layer, a diatomite support layer, a diatomite–kaolin composite support layer, a pyrophyllite support layer and a pyrophyllite–diatomite composite support layer. The typical trade-off between flexural strength (left vertical axis, bar graph) and air permeability (right vertical axis, line graph) of porous ceramic membranes can be seen

To identify the pore structure in the alumina support layers, typical SEM images of the alumina support layers sintered at 1400 and 1600°C for 1 h are shown in Fig. 2a and b, respectively. In Fig. 2a, the typical porous alumina matrix in the microstructure can easily be identified. Although the sintering temperature of the alumina support layer was increased from 1400 to 1600°C, the average pore size of the alumina support layer was not significantly changed, as shown in Fig. 2c.

OPEN IN VIEWER

2

Typical SEM images of alumina support layers sintered at a 1400°C for 1 h; b 1600°C for 1 h; c pore size distributions of the alumina support layers sintered at 1400°C for 1 h and 1600°C for 1 h; d particle size distributions of the as-received pyrophyllite and as-received diatomite

For dense ceramics, the average pore size generally decreases as the sintering temperature increases. However, experimental results on porous ceramics such as diatomite, ²⁶ silicon carbide,^39,40 alumina, ⁹ corundum–mullite ⁴¹ and zirconia ⁴² in different temperature ranges show that the average pore size of these materials increases or does not change significantly with increasing sintering temperature. Although this is generally explicable as a result of pore coarsening, ⁹ or densification, ⁴⁰ and processing routes for the production of porous ceramics have been extensively documented in the literature, the relationship between average pore size and sintering temperature has yet to be established.

In the above results, we determined that the air permeability and average pore size of an alumina support layer cannot be effectively tailored by changing just the sintering temperature, as is the case for diatomite ³⁴ and pyrophyllite support layer. ²⁷

In this study, diatomite and pyrophyllite were introduced as pore formers. Figure 2d shows the particle size distributions of the as-received pyrophyllite and as-received diatomite, for which the average particle sizes were 6.95 and 7.43 µm, respectively. Therefore, considering that the average particle size of the as-received alumina was 4.80 µm, we could avoid the problem of clogging of pore channels in the alumina matrix by the addition of diatomite or pyrophyllite.

X-ray diffraction (XRD) patterns of alumina–diatomite composite support layers with varying amounts of diatomite sintered at 1400°C for 1 h are shown in Fig. 3a. When an alumina support layer with 10 wt-% diatomite was sintered at 1400°C for 1 h, the major phase was found to be alumina, whereas for an alumina support layer with 30 wt-% of diatomite sintered under the same conditions, a SiO₂ (cristobalite) phase was observed in addition to the alumina phase. As the amount of diatomite added was increased to 40 wt-%, the peak intensity of the SiO₂ (cristobalite) phase gradually increased; however, neither a mullite phase (3Al₂O₃·2SiO₂) nor an unwanted secondary phase was observed.

OPEN IN VIEWER

3

a X-ray diffraction patterns and b pore size distributions of alumina–diatomite composite support layers sintered at 1400°C for 1 h with various amounts of added diatomite

It has been suggested that melting of diatomite is induced by impurities such as Na₂O, K₂O, Al₂O₃, CaO and MgO. ²⁵ These impurities in diatomite do not form low-temperature eutectics with Al₂O₃·SiO₂, and thereby a mullite phase; rather, they form eutectics with SiO₂, and thereby a glassy bonding phase at 1400°C. This is because the important eutectic temperatures with Al₂O₃, such as those for the CaO–Al₂O₃, Al₂O₃–SiO₂ and CaO–Al₂O₃–SiO₂ systems, are usually higher than 1400°C.^43,44

Figure 3b shows that the average pore sizes of the alumina support layers increased significantly from 0.70 to 1.53 µm or 1.80 µm on the addition of diatomite. However, the increased average pore size of the alumina–diatomite composite support layers does not necessarily lead to an increased air permeability, as in the case of the pyrophyllite support layer ²⁷ ; because of pore interconnectivity, we measured the air permeability of the alumina–diatomite composite support layers, and this will be discussed below.

To determine the microstructure of the alumina–diatomite composite support layers, typical SEM images of these layers for various amounts of diatomite addition, sintered at 1400°C for 1 h, are shown in Fig. 4a–d.

OPEN IN VIEWER

4

Typical SEM images of alumina–diatomite composite support layers with the addition of a 10 wt-%, b 20 wt-%, c 30 wt-% and d 40 wt-% diatomite, sintered at 1400°C for 1 h; e flexural strength and air permeability of alumina–diatomite composite support layers with various amounts of added diatomite, sintered at 1400°C for 1 h

With the addition of 10 or 20 wt-% diatomite, the microstructure of the alumina–diatomite composite support layer exhibited microscopic features intermediate between those of the irregular porous diatomite matrix and the alumina matrix as shown in Fig. 4a and b, respectively. Notably, Fig. 4c and d shows that with the addition of 30 wt-% diatomite or more, the pore channels of the alumina–diatomite composite support layers became significantly clogged by the silica bonding phase.

Figure 4e shows the air permeabilities and flexural strengths of alumina–diatomite composite support layers with the addition of 10–40 wt-% diatomite, sintered at 1400°C for 1 h. Interestingly, the air permeabilities and flexural strengths of these support layers increased simultaneously until the amount of added diatomite reached 20 wt-%. The flexural strengths of the layers increased significantly (approximately fourfold). The abrupt decrease in the air permeability of the support layer with 30 wt-% diatomite could be explained by clogging of the pore channels, as already indicated microscopically in Fig. 4c and d.

The flexural strengths of the alumina–diatomite composite support layers began to decrease with further addition of diatomite. This behaviour can be explained by the inherently weak strength of silica. The flexural strength of an alumina-based membrane ⁴⁵ can be greater than 50 MPa, but the flexural strengths of silica-based membranes such as kaolin ⁴⁶ and diatomite ⁴⁷ are usually below 30 MPa. Therefore, as the amount of silica reached a certain point, the strength enhancement by the silica bonding phase was cancelled out by the increased proportion of silica matrix.

To determine whether the enhanced air permeability and flexural strength induced by the addition of diatomite could be obtained only in the presence of a source of silica, alumina–pyrophyllite composite support layers were also prepared.

Figure 5a shows XRD patterns of alumina–pyrophyllite composite support layers with varying amounts of pyrophyllite sintered at 1400°C for 1 h. For an alumina support layer with 10 wt-% pyrophyllite, the major phase was found to be alumina. For 20 wt-% of pyrophyllite, both SiO₂ (cristobalite) and alumina phases were observed. As the amount of the added pyrophyllite was increased to 40 wt-%, the peak intensity of the SiO₂ (cristobalite) phase gradually increased; however, a mullite phase (3Al₂O₃·2SiO₂) did not appear, as was also the case with the alumina–diatomite composite support layer.

OPEN IN VIEWER

5

a X-ray diffraction patterns and b pore size distributions of alumina–pyrophyllite composite support layers sintered at 1400°C for 1 h for various amounts of added pyrophyllite

Figure 6a and b shows SEM images of alumina–pyrophyllite composite support layers sintered at 1400°C for 1 h with the addition of 20 and 40 wt-% pyrophyllite, respectively. Unlike the addition of 30 wt-% of diatomite to an alumina support layer, the addition of 40 wt-% of pyrophyllite did not induce any significant microscopic change in the overall alumina matrix. Even if the amount of pyrophyllite was sufficient to affect the overall microstructure of an alumina–pyrophyllite composite support layer, pore channels were neither induced nor clogged significantly by the pyrophyllite addition. Note that the air permeabilities and flexural strengths of the alumina–pyrophyllite composite support layers increased slightly up to 20 wt-% of added pyrophyllite, after which they decreased slightly, as shown in Fig. 6b. Thus, although the overall trends of changes in air permeability and flexural strength of the alumina–pyrophyllite composite support layers were similar to those of the alumina–diatomite composite support layers, the pore characteristics of the alumina–diatomite composite support layers showed more significant changes.

OPEN IN VIEWER

6

Typical SEM images of alumina–pyrophyllite composite support layers with addition of a 20 wt-%, b 40 wt-%, sintered at 1400°C for 1 h; c flexural strength and air permeability of alumina–pyrophyllite composite support layers with various amounts of added pyrophyllite, sintered at 1400°C for 1 h

This result can be explained as follows. First, diatomite is composed of SiO₂ and approximately 20 wt-% impurities such as CaO, Na₂O and MgO, ⁴⁸ but pyrophyllite is primarily composed of SiO₂ and Al₂O₃ (in this study, 74.97 wt-% SiO₂, 19.26 wt-% Al₂O₃, 0.19 wt-% Fe₂O₃, 0.48 wt-% TiO₂, 0.10 wt-% K₂O, 0.05 wt-% S and ignition loss). Therefore, it is thought that the levels of impurities in the pyrophyllite are too low to form both eutectics with Al₂O₃·SiO₂ and thereby a mullite phase and eutectics with SiO₂ and thereby a glassy bonding phase at 1400°C. Second, although the average particle sizes of the pyrophyllite and diatomite are similar, the irregular and inherently porous particle shape of the diatomite is more suitable than the plate-like particle shape of the pyrophyllite to serve as a pore former in the alumina matrix.

Figure 7 provides a summary of the air permeabilities and flexural strengths of various types of porous ceramic support layer. The air permeability and flexural strength of the diatomite support layers, ³⁵ the pyrophyllite layers and the pyrophyllite–diatomite composite support layers ²⁷ can be enhanced at the expense of just one of two properties by controlling the sintering temperature or the mixing ratio of the composite. Notably, the air permeability and flexural strength of an alumina support layer can be enhanced simultaneously by the addition of diatomite, as both a pore former and a bonding phase. Although, the addition of pyrophyllite to an alumina support layer could affect both air permeability and flexural strength, its effectiveness as a pore former and bonding phase is limited both by its plate-like particle shape and by its composition.

OPEN IN VIEWER

7

Flexural strength as a function of air permeability of various types of porous ceramic support layers with control of sintering temperature or changes in composite mixing ratio

Conclusion

The flexural strength and air permeability of alumina support layers cannot be enhanced simultaneously by just increasing the sintering temperature to 1600°C. The average pore sizes and air permeabilities of alumina–diatomite composite support layers sintered at 1400°C can be controlled in the approximate ranges of 0.70–1.80 µm and 1.38–2.36 L min⁻¹ cm⁻², respectively, by changing the amount of added diatomite (utilised as a pore former). Concurrently, the flexural strength of the alumina–diatomite composite support layers can be enhanced from 29.85 to 78.32 MPa by changing the amount of diatomite (utilised as a bonding phase).

Note that, through the addition of diatomite, the flexural strength and air permeability of alumina support layers can be tailored while still keeping the overall processing cost down. The addition of diatomite provides an effective means of tailoring the pore characteristics of the alumina–diatomite composite matrix. The effectiveness of diatomite both as a pore former and as a bonding phase has been investigated here through a comparison of alumina–diatomite composite and alumina–pyrophyllite composite support layers.