Nanoporous zirconia microspheres prepared by salt-assisted spray drying

Nanoporous zirconia with high surface area and crystallinity has a wide range of industrial applications, such as in inorganic exchangers for ion exchange columns, catalyst substrates, and packing material for HPLC. Spherical particles of crystalline nanoporous zirconia are highly desired in various industries due to easy handling of the materials in a fluidized bed. Here, spray drying was adopted to produce spherical nanoporous zirconia powders in both laboratory scale and pilot plant scale. Effect of salts on spray-dried ZrO2 powders and their crystallization behavior was studied. It was found that addition of salts to the zirconia precursors has a huge effect on the crystallization of nanoporous zirconia powders. These results have a great impact on the development of microspheres of nanocrystalline ZrO2 and potentially open up a new opportunity to the low-cost production of porous ceramic microspheres with the salt templating method, in general.


Introduction
In the last couple of decades, zirconia has gained increasing interest, due to its mechanical properties, ability to catalyze chemical reactions and ionic conductivity at high temperature. For this reason, zirconia and zirconia doped with varies metal oxides, such as yttria, ceria, and magnesia, have been widely studied [1]. Zirconia exhibits three crystallographic phases with increasing temperature at atmospheric pressure: monoclinic below 1175 °C, tetragonal between 1175 and 2370 °C and cubic above 2370 °C. Furthermore, a metastable tetragonal phase exists at room temperature. However, the metastable tetragonal zirconia phase transforms to the stable monoclinic phase due to a range of external factors: applied shear stresses [2], aging in humid atmosphere [3] or annealing [4]. The tetragonalto-monoclinic phase transformation is associated with a volume expansion of ~ 4 vol%. This volume expansion is the cause of the excellent mechanical properties of doped zirconia ceramics, as the stress-induced phase transformation of metastable tetragonal zirconia grains at the crack tips is the basis of transformation toughening in zirconia ceramics [5].
Porous zirconia with high surface area is of interest for various functional applications. Spherical micron sizedporous zirconia is of particular interest due to the reduced particle aggregation and free-flowing properties. Both alcohol drying [6] and spray drying [7] are known from the literature, and spray drying is furthermore widely used in various industries due to the easy operation and high reproducibility of the process. Another way to obtain small zirconia particles is by modifying of the precursor blend. Salt-assisted spray drying was tested as a possible route. It is known from the literature that salt-assisted spray drying and salt-assisted spray pyrolysis can be used to produce nano-sized ceramic particles [8].
In this study, first, different ways to prepare nano-sized zirconia particles were investigated. The tested preparation routes are azeotropic drying, spray drying of pure ZrO(OH) 2 and spray drying in combination with either CsCl or NH 4 NO 3 . An overview of the tested methods is illustrated in Fig. 1. The effect on particle size, crystal size, surface area, crystallization temperature, and crystal phase of the formed zirconia is studied.

Experimental
All chemicals were supplied by Aldrich, and they were all used as received. Dispersions of highly porous ZrO(OH) 2 powders were synthesized as previously described [9,10] by controlled hydrolysis of ZrOCl 2 and conditioning. The ZrO(OH) 2 powders are extremely porous and have specific surface area of ~ 350 m 2 /g.

Spray drying of neutral, aqueous dispersions
The pilot-scale spray dryer was used to spray drying of pure zirconia and CsCl salt containing zirconia, while the laboratory-scale spray drier was used to produce NH 4 NO 3 salt containing zirconia.
A suspension of amorphous ZrO(OH) 2 was prepared as described above. From the suspension, three different aqueous dispersions of ZrO(OH) 2 were prepared by stirring, with low (lCsZrO 2 ) and high (hCsZrO 2 ) CsCl concentration and one without addition of CsCl. The pure ZrO 2 precursor was spray dried without further preparation. The lCsZrO 2 precursor solution was prepared by stirring 4 L 5.0 wt% ZrO(OH) 2 water suspension with 100 g CsCl until the salt was dissolved. The hCsZrO 2 precursor solution was prepared by stirring 4 L 5.0 wt% ZrO(OH) 2 aqueous suspension with 200 g CsCl until the salt was dissolved. The dispersions were spray dried at pilot plant scale. Dispersions of amorphous zirconia in water were spray dried in a MOBILE MINOR™ Pilot Plant with an overall chamber diameter of 0.8 m, a cylindrical height of 0.83 m and a 60° cone. Droplets of the precursor solution were generated in the spraying chamber using a two-flow spray-nozzle with a replaceable orifice with a diameter of 1.5 mm. At drying temperatures 275/145 °C (inlet/outlet), dry-looking powders were obtained. The produced powder was calcined at various temperatures.
Originally, it was expected to be possible to wash the salts out with anhydrous methanol. However, the solubility of the salt was too low, and the salt was instead removed with water to facilitate characterization of the particles.
The lNZrO 2 precursor solution was prepared by mixing 200 g 5.0 wt% ZrO(OH) 2 water suspension, 200 ml ethanol (99.7%), 10 g acetic acid, 5 g ammonium nitrate, and 5 g of 2.0 M HCl. The resultant suspension was sonicated for 30 min with Vibracell before spray drying. The precursor solutions were spray dried at laboratory scale in a laboratory-scale spray drying reactor, which consists of a spraying chamber combined with a tube furnace and a filter [11]. Droplets of the precursor solution were generated in the spraying chamber using a two-flow spray-nozzle with a replaceable orifice with a diameter of 1.5 mm at room temperature of 25 °C. The liquid flow and gas flow were set to 1.76 ml/min and 16,000 ml/min, respectively. Outlet oven temperature was 280 °C, and dry-looking white spherical powders were collected. The produced powder was calcined at various temperatures. A schematic illustration of the custom-made laboratory-scale spray drier is shown in Fig. 2. At the end of the spray drying process,

Calcination and washing
The samples were calcined in dry air at temperatures varying from 450 to 550 °C. After calcination, the crystalline samples were cooled to RT in ambient atmosphere. The CsCl and NH 4 NO 3 containing samples were washed with water until no reaction with AgNO 3 or nitrate sticks was observed. The washed samples were dried at RT in ambient atmosphere.

XRD
A STOE theta-theta X-ray diffractometer, STOE and Cie GmbH, Damstadt, Germany, was used in reflection mode, and XRD patterns were scanned in 0.1 steps (2θ) in the 2θ range from 20° to 60°, with a fixed counting time (30 s). The XRD patterns were analyzed using WinX POW software. For the crystalline samples, the obtained values of the t-ZrO 2 and m-ZrO 2 volume fractions (v t and v m ) were compared with the values obtained from the integral intensities of the monoclinic diffraction lines (− 1 1 1) and (1 1 1) and the tetragonal diffraction line (1 0 1), following a procedure proposed by Toraya et al. [12]. The crystal size was calculated using the Scherrer equation [13].

Transmission electron microcopy (TEM)
Low-and high-resolution TEM images and selected area electron diffraction (SAED) patterns were obtained using a JEOL JEM-3010 microscope operating at 300 kV (Cs = 0.6 mm, Point resolution 0.17 nm). Images were recorded with a CCD camera (MultiScan model 794, Gatan, 1024 × 1024 μm) at a magnification of 4000-400,000 times. TEM samples were prepared by applying a drop of zirconia-isopropanol dispersion onto a carbon-coated Cu grid, and the solvent was allowed to slowly evaporate at room temperature.

Differential scanning calorimetry (DSC)
Differential scanning calorimetry was carried out using a high-sensitivity differential scanning calorimeter Seiko SII-DSC120 and a Seiko SII-DSC320. The samples (15-25 mg) were placed in aluminum pans with a mass of about 16 mg. Heating rate of 5 °C min −1 . During all the experiments, a protection gas of pure He (99.996%) was used with flow rate of 100 cm 3 /min. The DSC equipment is regularly temperature calibrated to within 0.5 °C using Zn metal.

Scanning electron microscopy (SEM)
In this study, an XL30 TMP environmental scanning electron microscope (ESEM) from Philips/FEI was used. In sample preparation, small amounts of powders were placed on an aluminum sample holder using a spatula. The ESEM investigation was carried out at room temperature (ca 25 °C) without sample coating. Accelerating voltage of 20-30 kV was used. The probe current (spot size) was varied between 4 and 6 nA, and the working distance (WD) was around 10 mm.

Results and discussion
The effects of preparation methods and the presence of either salts or isopropanol during calcination are listed in Table 1. It is observed that the different methods result in zirconia powders with very different properties. This means that the preparation route has a large effect on the properties of the produced zirconia powders; particularly, the crystal phase and crystal size are influenced. For the CsCl-containing samples, the crystallization temperature is also effected. The XRD patterns for the samples calcined at 450 °C are illustrated in Fig. 3. It is observed that the (101) tetragonal reflection at 2θ = 30.2° is more dominating for hNZrO 2 compared with lNZrO 2 and the two pure samples, which have more dominating (− 111) and (111) monoclinic reflections. The difference in peak width is a result of different crystal sizes. This difference is most pronounced when comparing the XRD patterns of the alcohol-dried sample and lNZrO 2 (Fig. 3c, e). The alcohol-dried sample has the broadest peaks because it has the smallest crystal size. Furthermore, it is found that while both the pure zirconia samples and the two samples spray dried with NH 4 NO 3 are crystalline at 450 °C, this is not the case for the two CsCl containing samples, which still are amorphous at this temperature. For this reason, these two samples were calcined at 550 °C and subjected to XRD. The results are illustrated in Fig. 4. It is observed that the tetragonal reflection is dominating for both samples. The XRD patterns of the powders prepared with saltassisted spray drying show only the existence of zirconia peaks and none of the peaks corresponding to CsCl or NH 4 NO 3 are observed. This result indicates that all the salt is removed during washing.
All the samples were analyzed with SEM and TEM. The azeotropic drying resulted in particles with a diffuse morphology (Fig. 5), whereas spray drying resulted for all samples in spherical particles with a diameter in the range of 1-2 μm for the pure sample and for the samples spray dried with CsCl (Figs. 6, 7a-d). For the two samples spray-dried with NH 4 NO 3 , the particle sizes were in the range of < 0.5 μm (Fig. 7e-h). The two NH 4 NO 3 -containing samples gave the widest particle size distributions. Except for the size differences, no changes in morphology were observed with either SEM or TEM in any of the spray dried samples.
All the tested methods resulted in crystalline nanoporous zirconia powders with particle sizes up to 2 μm. This is in contrast to the very large and hard agglomerates of zirconia obtained by just calcining the air-dried ZrO(OH) 2 dispersion (not shown).
The two pure zirconia samples have similar properties, but azeotropic drying gives the smallest crystals and micron-sized particles with a more diffuse structure. Spray drying, however, gives micron sized spherical zirconia. Both methods give pure zirconia with primarily the monoclinic phase and the samples contain only 30 vol% tetragonal crystals. From the DSC analysis, the crystallization  Table 1). The crystallization temperatures are, however, increased, and the two CsCl-containing samples crystallize at 525 °C. Calcination at 550 °C was expected to increase the sintering of the zirconia samples and hence increase the crystal size, but the crystal size does not increase significantly in any of the two samples and does not differ significantly from the spray-dried pure zirconia sample. The decrease in v m is probably a result of anionic stabilization originating from incorporation of Cl − in the tetragon crystal lattice. The stabilizing effect is among others described by Gutzov et al. [14]. Furthermore, chloride ions adsorb strongly on zirconia and are difficult to remove completely from the surface [15].
In contrast to the CsCl containing samples, the presence of NH 4 NO 3 during calcination causes no increase in the crystallization temperature, as both the samples crystallize at ~ 435 °C. For the two NH 4 NO 3 -containing samples, a significant difference is observed. hNZrO 2 has a volume fraction of monoclinic crystals (v m ) of 0.10, whereas lNZrO 2 has a v m of 0.65. The latter is close to the v m of the two pure zirconia samples (Table 1). In addition, the crystal sizes vary in these two samples and hNZrO 2 has a crystal size of 11.6 nm and lNZrO 2 a size of 13.7 nm. These crystal sizes were confirmed by TEM. Both samples show increased crystal sizes compared with all the other samples. In particular, the sample containing the lowest content of NH 4 NO 3 has increased crystal size. hNZrO 2 gives the lowest v m , but has larger crystals, than the pure zirconia samples. This indicates a stabilization of the tetragonal phase, which most likely originates from ammonium ions working as mineralizers, in the same way, as chlorine ions stabilize the samples calcined in the presence of CsCl. The stabilizing effect of ammonium ions counteracts the size effect and increases the critical size of the tetragonal crystals. Normally when the crystal size increases, the v m will increase as well. This is not the case in this study and supports the presence of ammonium stabilization. Surprisingly, lNZrO 2 is not stabilized to the same extend and is the one of all the salt-containing samples with the highest v m . lNZrO 2 and lCsZrO 2 have the same zirconia concentration, and the fact that only the CsCl is able to stabilize the tetragonal phase in the crystals strengthens hypothesis that the chloride ions have a stabilizing effect on the tetragonal phase.
In contrast to the diffuse particle structure of the samples dried with azeotropic distillation, all the spray-dried samples are spherical, but with different sizes depending on the production process. This indicates that the morphology and size are more a result of processing parameters, than a result of the properties and concentrations of the salts. Furthermore, the fact that the two NH 4 NO 3 -containing samples gave a wider particle size distribution supports this conclusion.

Conclusion
Salt-assisted spray drying method has been developed for the pilot-scale production of nanoporous zirconia microspheres.
Four different preparation routes were tested to prepare spherical zirconia powders: azeotropic drying with isopropanol, spray drying and salt-assisted spray drying with NH 4 NO 3 or CsCl (two different concentrations in both cases). From the results of this study, it is concluded that addition of high amounts of NH 4 NO 3 gives the crystals with the lowest monoclinic volume fraction and particles having a spherical shape. Spray drying of a precursor solution containing 1:1 ZrO 2 :CsCl results in almost the