Manufacture of Fine-Pored Ceramics by the Gelcasting Method

The fine-pored materials represent a wide range of applications and searches are being continued to develop methods of their manufacturing. In the article, based onmeasurements on fine-grained powders of Al 2 O 3 , TiO 2 , and SiO 2 , it has been demonstrated that gelcasting can be relatively simple method of obtaining of nanoporous materials with high values of both specific surface area and open porosity. The powders were dispersed in silica sol, and the gelling initiator was NH 4 Cl. The usefulness of experiment design theory for developing of fine-pored materials with high porosity and specific surface area was also shown.


Introduction
The classification proposed by the International Union of Pure and Applied Chemistry (IUPAC) divides pores into 3 groups, depending on their size: micropores (up to 2 nm), mesopores (2-50 nm), and macropores (above 50 nm) [1].A commonly used term is also that of microporous materials, when the pores are in the order of micrometres (i.e., [2]).Moreover, the informal notion of nanopores conventionally includes pores having a diameter of up to 100 nm [3].Therefore, in this article, mesoporous, microporous, or nanoporous materials, regardless of the classification, will be referred to as fine-pored materials.
Such materials represent a wide range of applications.An example of fine-pored products applications include the following: biomaterials used as medical implants [4], carriers of permeating membranes or catalysts [5], molecular sieves [6], thermoinsulating materials [7], diffusion layers in fuel cells [8], composite materials matrixes [9], and porous ceramics infiltrated with metals or polymers [10].
A good method of obtaining porous materials with a set shape is gelcasting [11][12][13][14][15][16][17][18][19][20][21][22].Gelcasting is a combination of the traditional method of forming ceramic materials from casting slips with the polymerization reaction.The basic idea of this relatively new but intensively developing method involves making a ceramic powder suspension with an addition of a substance which is subject to gelation when influenced by an appropriate initiator [23][24][25].Proportions between the suspension components and the time of suspension homogenization are so selected that they enable casting the suspension into the mould before the gelation process and, at the same time, ensure that gelation does not occur too late due to sedimentation processes (if they are not desired) [26].One of the simplest gelling systems is that based on silica sol (dispersing agent) and NH 4 Cl (gelling initiator) [27].
The aim of this article is to prove that the gelcasting method can be applied in the manufacture of fine-pored materials, using relatively simple measures.In the presented examples of obtaining fine-pored ceramic materials by the gelcasting method, the suspensions were made with silica sol as a dispersing agent, whereas dispersing powders in the suspension were compounds commonly used in the ceramic industry.Investigations were conducted for the group of powders: Al 2 O 3 -TiO 2 -SiO 2 .Since the radius of pore diameter is a fraction of the diameter of grains in a particular monofraction [28], the investigations were carried out for dusty fractions.
The proportions of powders in subsequent sets of samples were selected according to the experiment design theory for investigations into the composition-properties correlation.The so-called simplex-network Scheffe test for a threecomponent system was applied [29].

Manufacture of Materials
The experiments were done on a gelling set based on silica sol, in which the gelling initiator was NH 4 Cl.The following ceramic powders from Evonic were used: SiO 2 Aerosil P90 (having a developed surface area of 90 m 2 /g), Al 2 O 3 Aeroxide Alu C (100 m 2 /g), and TiO 2 Aeroxide P90 (90 m 2 /g).Each time the sample preparation procedure was very similar: after adding a set amount of powder to the silica sol (the proportions will be given in tables in the further part of the article) the suspensions were homogenized during an appropriate time (usually half an hour) in a magnetic stirrer.Next, an appropriate amount of gelling initiator was added so that the time of sample gelling would range from several to several dozen minutes.To avoid uneven gelling (formation of lumps), the silica sol-based sets gelling initiator-NH 4 Cl-was added to the suspension in the form of a water solution.A 23% solution was prepared.As the different suspensions require the use of a solution in various concentrations, the solution prepared was most frequently diluted.Table 2 does not contain the concentrations calculated each time but the proportions of additional diluting (this manner of providing information was considered better as in reality, apart from the additive concentration, also the content of water in the solution was changed).The gelled suspensions were dried slowly (for 2-3 days the temperature was gradually increased to 195 ∘ C).After drying all the samples were fired at an appropriate temperature (600 ∘ C, 1000 ∘ C, 1400 ∘ C), with a hold time of 4 h.Temperature progression was 2 ∘ C/min.The samples fired at 600 ∘ C and 1000 ∘ C were additionally held at 250 ∘ C (60 min) and 400 ∘ C (60 min).Temperatures at which the samples were additionally held were aimed at preventing the samples' cracking due to a release of a large amount of gases-they were determined on the basis of thermogravimetric investigations (Figure 1).For the samples fired at 1400 ∘ C slow heating rate was only applied.

Measurement Techniques
The open porosity, apparent density, and water absorption of the prepared samples were determined by the method of hydrostatic weighing in water.The crystal structure of manufactured powders was investigated by X-ray Powder Diffraction (XRD), using an X'Pert PRO MPD (CuK) diffractometer produced by PANalytical.The specific surface area (BET method) was determined by means of a Gemini 2360 analyzer produced by Micromeritics using nitrogen as an adsorbed gas.Moreover, for selected samples measurements of pore distribution were taken with AutoPore IV9500 Mercury Porosimeter produced by Micromeritics.Photos on the investigating samples were taken with the use of highresolution scanning electron microscope (SEM) Mira III from Tescan.

Results and Their Analysis
As mentioned before, the experiments were based on Scheffe test designs.Seven samples of the appropriate proportions (Table 1) were prepared in a way which enabled calculating the coefficients of the third-degree polynomial (in this case) for the three-component mixture.
For the design in question one can determine the result of composition correlation from the following formula: where relevant coefficients are calculated from the following dependence: Based on initial investigations, it was found that in the case of mixture homogenized with the magnetic stirrer (for this type of devices the stirring force is relatively small) the proportion for the abovementioned oxide materials would be 9 g of oxides mixture powder per 60 g of silica sol.The first batch of samples (obtained by firing at 600 ∘ C) were based on suspensions having proportions given in Table 2.
It can be visible (Table 3) that results of apparent density, open porosity, and water absorption obtained on the basis of hydrostatic method are coherent with values of specific surface area from BET method.The highest values of porosity and specific surface area have samples with presence of Alu C powder.The selection of powders based on the experiment design theory allows observing certain general tendencies concerning the selected properties of the samples obtained, depending on the substrates' proportions.In this paper the parameter analyzed by the three-component Scheffe estimation is open porosity.The determined function of   open porosity changes (using formulas (1) and ( 2)) has the following form: Transition from the coded scale of coefficients  1 ,  2 , and  3 to the natural scale, where the role of the independent variable is played by the masses of particular powders (please compare the proportion in Tables 1 and 2) and the results calculated by substituting a series of mutual proportions of the examined oxides into formula (3) allowed drawing an isoline of the samples' open porosity depending on their initial composition.An approximate course of these lines is presented in Figure 2. The estimation indicates that, for the suspension containing 9 g of the examined oxide powders, the highest porosity values in 60 ml of silica sol should be expected to be found in the following components: 50-80% of Al 2 O 3 , 10-40% of SiO 2 , and 0-30% of TiO 2 .However, the interpretation of the diagram should be taken into consideration more qualitatively.The isolines course show, primarily, that the porosity value depends mainly on Alu C content in the mixture of powders.When the content of Alu C reduces in the range from 50% to 0% the estimated values of the samples porosity decrease rapidly.On the other hand, when the content of the alumina is between 50% and 100% the estimated values of porosity become high (close to 60%) and are only slightly sensitive to the changes of powders  stoichiometry.The estimation indicates that the highest value of porosity is obtained for the systems where content of Alu C is at least 50%.Thus it was shown that the gelcasting method allows manufacturing the materials with relatively high values of open porosity.
For samples 1, 4, and 5 pore size distribution was additionally determined and presented at Figure 3.
The results show that manufactured samples are formed as fine-pored material.The distribution of pores is similar for all three presented samples.However it should be noted that with higher content of Alu C powder the pore distribution in the sample has the most sharpness.The dominant pore size diameter is close to 10 nm.
The above results are confirmed by SEM investigations.The micrographs presented at Figure 4 show that the materials of all the samples are formed by highly packed grains with a diameter close to 100 nm.Some of the grains exceed 100 nm but there are numerous specimens with sizes of the order of tens of nanometers.It can be visible that the range of the grains diameter in sample 4 is slightly wider than that in samples 1 and 5. Moreover the particles shape in samples 1 and 5 looks more isometric than in sample 4.These observations are in correlation with the mentioned results of the pore distributions in the investigated samples.
For the compositional range with the high porosity additionally the influence of firing temperature (600 ∘ C, 1000 ∘ C, and 1400 ∘ C) was determined.The contents of powders in suspensions and the determined results have been given in Tables 4 and 5.
As can be seen, firing at 600 ∘ C and 1000 ∘ C does not cause significant differences in the parameters of the samples obtained.They are dominated by amorphous or metastable (delta Al 2 O 3 ) substrates; samples with TiO 2 also contain anatase and, to a lesser extent, rutile.The slightly higher sintering process is visible at 1000 ∘ C. On the other hand, the temperature of 1400 ∘ C clearly results in the material sintering, crystallization of silica phases, and mullitization of the material.In samples containing TiO 2 also the transformation of anatase into rutile is clearly visible.Thus it can be visible that temperature of sintering should be carefully selected for the system based on silica sol and the fine powders.
It should be also added that the described gelled suspensions were characterized by high drying shrinkage, reaching 25%, which makes it difficult to manufacture large-sized components, as they could crack in the drying process.The obtained small samples (order of cm) are characterized by sufficient operational strength, allowing their free movement and arrangement during potential subsequent processes.For example, owing to the porosity parameters and high surface area development, these materials can be applied as cores of porous matrixes infiltrated with polymers (including the ones maintained by capillary forces) or metals (macroscopically, smooth surfaces enable thin-walled casting and the finepored matrix removes the process gases).

Summary
The aim of measurements described in this work was to obtain materials characterized by high open porosity and small pore size by the gelcasting method.
The investigations were conducted using fine-grained powders of Al 2 O 3 , TiO 2 , and SiO 2 with a specific surface area reaching ca 100 m 2 /g.The powders were dispersed in silica sol, and the gelling initiator was NH 4 Cl.Using relatively simple methods, we were able to obtain ceramics characterized by high porosity, small pore diameters, and high surface area development.The relatively high drying shrinkage causes that it is hard to obtain large-sized components from the examined systems.The materials obtained are characterized by sufficient operational strength, allowing their free movement and arrangement.The usefulness of experiment design theory for developing of fine-pored materials with high porosity and specific surface area was also shown.The simplex-network Scheffe test allows observing certain general tendencies concerning the selected properties of the samples obtained, depending on the substrates' proportions.

2 AdvancesFigure 1 :
Figure 1: Thermogravimetric analysis of the investigated powders suspension in silica sol (gelled under NH 4 Cl solution).154.5 mg of sample was heated in the air with heating rate 3.5 ∘ C/min.

Figure 2 :Figure 3 :
Figure2: Open porosity isolines (circles) for the system from Table2, drawn on the basis of calculations conducted based on formula (3).The dotted lines define the specified powder portion in the mixture (100% at the indicated apex of the triangle and 0% at an opposed side).

Table 1 :
Design matrix for the third-degree polynomial for the three-component system.

Table 2 :
Compositions of suspensions for silica sol-based samples (proportion: 60 ml silica sol per 9 g of oxides mixture powder).

Table 3 :
Results of selected physical properties obtained for samples having compositions contained in Table2(obtained by firing at 600 ∘ C).

Table 4 :
Composition of suspensions based mainly on Al 2 O 3 , TiO 2 , and SiO 2 for silica sol-based samples (proportion: 60 ml silica sol per 9 g of oxides mixture powder).

Table 5 :
Results of selected physical properties for samples with compositions given in Table4.