Titania aerogels with tailored nano and microstructure: comparison of lyophilization and supercritical drying

Introduction

The wet-chemical approach is a very important one among the different methods to prepare new nanostructured materials. A lot of different wet chemical approaches, so-called bottom-up fabrication methods, of nanostructured solids of different sizes, shapes and material compositions are known, both in polar and non-polar solvents [1]. It has been published recently [2, 3] that a network of thin platelets building up the gel network can be prepared by freezing aqueous colloids in situ and subsequent freeze drying (lyophilization) or by critical point drying. The advantage of this strategy is its suitability for various nanoparticle systems, since there is no chemical selectivity. At the best case, the volume and shape of the resulting aerogel is the same as the volume and shape of the initial hydrogel.

Hydrogels are usually considered as aqueous jelly-like material, in properties ranging from soft and weak to hard and tough. The continuity of the liquid permeating jellies was demonstrated by diffusion, syneresis, and ultrafiltration, and the fact that the liquid may be replaced by other liquids of very diverse character indicates clearly that the gel structure may be independent of the liquid in which it is bathed. The attempt to remove the liquid by evaporation usually results in extensive shrinkage and possible structural damage of the material [4, 5].

To avoid this, the sample can be dried via two possible alternate paths from the liquid phase to the gas phase without crossing the liquid–gas boundary on the phase diagram. In freeze drying, solutions are frozen in a cold bath and then the frozen solvents are removed via sublimation under vacuum, leading to formation of porous structures. Pore size, pore volume and pore morphology are dependent on variables such as freeze temperature, solution concentration, nature of solvent and solute, and the control of the freezing direction [6]. Supercritical drying, on the other hand, proceeds at high temperature and high pressure above the critical point. This route from liquid to gas does not cross any phase boundary, instead passing through the supercritical region, where the distinction between gas and liquid ceases to apply.

Nanostructured titania both in the crystalline form of anatase and rutile represents important material widely used as catalysts, pigments, photocatalysts, components of batteries, and other technically important materials. However, except of classical sol-gel processes based on hydrolysis of titanium alkoxides, purely aqueous processes are rather rare in synthesis of such material despite the fact that use of water as solvent is clearly the most advantageous fromboth economic and ecological point of view. We successfully applied lyophilization for synthesis of highly photoactive 2D titanium dioxide nanostructures prepared from aqueous colloids of peroxo-polytitanic acid. Reaction of hydrogen peroxide with suspension of thoroughly washed precipitates obtained by neutralization of aqueous solution of titanyl sulphate with aqueous ammonia leads to transparent yellow colloid suspension. In this way prepared colloid suspension is a two phase system in which, beyond the water phase, also a liquid hydrated peroxo-polytitanic acid component is present [3]. During time the viscosity of the originally fluid liquid increases and the material transforms into viscous hydrogel as a result of polymerization processes of planar titanium peroxo-complexes [7, 8].

In this work, yellow colloid suspensions of peroxo-polytitanic acid were dehydrated using lyophilization of fast frozen materials obtained by immersion in liquid nitrogen as well as by drying process using supercritical CO₂. Structural and morphological investigations have shown some common features as well as some differences between products of drying resulting from different mechanisms of arrangement of nanoparticles during the drying processes.

Materials and methods

Characterization methods

Results and discussion

Samples of titania aerogels prepared by two different methods of drying, lyophilization and critical point drying, were prepared. Their morphology, microstructure, and phase composition were studied.

SEM micrographs of LYO and CPD titania aerogels are displayed in Fig. 1. While the lyophilized material consists of well-separated very thin foils, the CPD aerogel is rather bulk material, although also exhibits signs of lamellar structure (Fig. 1b2). This fact is probably related to planar structure of titanium peroxide complex described in literature earlier [3, 7, 16, 17]. Morphological difference between LYO and CPD samples is best seen from the micrographs taken at the highest magnification. While the lyophilized sample (Fig. 1a3) shows very thin foils with rather smooth surface, the CPD sample (Fig. 1b3) shows highly porous material consisting of randomly oriented very small, probably planar, units.

Fig. 1:

SEM micrographs of lyophilized (a1–a3) and critical point dried (b1–b3) titania aerogels.

Fig. 2:

TEM micrographs of lyophilized (a1–a3) and critical point dried (b1–b3) titania aerogels.

Fig. 3:

In situ HT XRD analysis of (a) lyophilized (left) and critical point dried (right) and (b) mass fraction of crystalline phases upon heating of titania aerogels.

Morphology and phase composition at nanoscale was studied by TEM (Fig. 2). The LYO sample shows the typical foil morphology. The foils have lateral size of several tens of micrometers and thickness of only about 10 to 20 nm, and are composed of tightly packed grains. The SAED pattern displays weak broad circles corresponding to peroxo polytitanic acid (Pdf 47-0124) with a very small grain size of about 2 nm bordering on amorphous character. On the contrary, the CPD sample displays 3D nanoporous microstructure composed of flat nanoparticles about 5 to 10 nm in size and 2 nm thick. Most of the particles are composed of crystalline anatase, whereas others seem to be amorphous.

Such differences can be easily understood on the basis of recently published concept of formation of self-supported aerogels of nanoparticle building blocks by freezing and subsequent freeze-drying of colloidal nanoparticle suspensions [2]. Authors suppose, that during fast freezing of nanoparticle containing aqueous colloids, nanoparticles originally evenly distributed within the aqueous medium aggregate on the surface of ice grains. After full freezing and removal of ice due to the lyophilization process, the remaining nanoparticles form assembled aggregates shape and size of which depends on conditions of freezing as well as on properties of the initial nanoparticle units. It is clear that by fast freezing aqueous colloidal nanoparticle suspensions, which were as droplets injected directly into liquid nitrogen, and by subsequent freeze drying the frozen objects, highly porous and voluminous aerogels consisting solely of the respective material can be obtained.

The CPD dried material, on the other hand, is not affected by the effect of growing ice grains and the structure of aerogel is probably closer to the structure of the original aqueous colloid. Figure 1b3 shows highly porous surface of the CPD material consisting of small randomly oriented nano-sized units. The planar shape of these nanostructures was confirmed also by HRTEM. Figure 2a3 and b3 display clearly planar nanoparticles in both lyophilized and CPD dried samples, which are more densely packed in the lyophilized sample. This is in agreement with the above mentioned mechanism of lyophilization of the fast frozen colloids.

The drying procedure also affects the atomic structure of the titania aerogel and its behavior upon heating as is documented by the results of in situ HT XRD analysis shown in Fig. 3. The samples display significant differences. The lyophilized non-annealed material is amorphous up to 300°C, when it suddenly crystallizes to anatase with sharp diffraction lines corresponding to crystallite size ranging from 25 nm at 400°C to 40 nm at 900°C, when the transition into rutile begins. The CPD sample, on the other hand, shows diffraction lines belonging to anatase even in the non-annealed sample. The lines are very broad corresponding to very fine crystallite size of 2 nm. With increasing temperature, the lines are getting sharper as the crystallite size grows to 25 nm at 800°C, when the transformation to rutile is initialized [18, 19].

It is well known that anatase → rutile phase transformation depends significantly on the preparation method, it was reported that the transformation begins above ~600°C, mostly close to this temperature [20]. Nevertheless, this transformation in both samples begins at much higher temperature, and 50% of the transformation proceeds between 850 and 1100°C (Fig. 3b). It is well known that bulk rutile is more thermodynamically stable than anatase at all temperatures and pressures owing to its lower free energy. However, the lower surface energy of the anatase planes relative to those of rutile cause the former to be more stable for crystallites of extremely small sizes and correspondingly high surface areas. Further, the size above which rutile becomes more stable depends on stress [20]. Thus we can well understand the high anatase stability in both samples owing to very small average crystallite size of anatase lying in the region of tens of nanometers. The transformation temperature to rutile is significantly lower for the CPD sample by about 100°C.

The above described differences are likely to be associated with different morphology of the samples. Whereas in the LYO sample with a typical laminar morphology the mass transport is limited and does not allow growth of anatase crystals, the bulk morphology of the CPD sample creates more favorable conditions for mass transfer and thus the growth of crystals.

The FTIR spectra of both prepared titania aerogels are very similar (Fig. 4a). The analysis shows that the samples evidently contain OH⁻ groups (characteristic bands at 3400 cm⁻¹ and at 1600 cm⁻¹), NH₄⁺ (at 3200 cm⁻¹ and at 1400 cm⁻¹) and peroxogroups (920 cm⁻¹). None of the bands characteristic to sulfates can be observed at 1200 cm⁻¹, which means that the initial TiOSO₄·nH₂O was decomposed and all sulfur was removed during the synthesis.

Fig. 4:

Comparison of infrared (a) and Raman (b) spectra of lyophilized and critical point dried titania aerogels.

The analysis of the Raman spectra (Fig. 4b) reveals that both LYO and CPD samples contain spectral features which can be attributed to Ti-peroxygel, anatase or metatitanic acid. Both LYO and CPD clearly differ from anatase or rutile. The spectra of both LYO (especially) and CPD are significantly similar to the spectra of Ti-peroxy gel formed at surface of Ti metal by action of hydrogen peroxide described by Muyco et al. [21]. Both samples, LYO and CPD, contain weak or middle intensity bands with peaks at positions: 197, (198), 299 (shoulder), (286) and 387, (385) cm⁻¹, which correspond to 191, 287.5 and 387 cm⁻¹ attributed to Ti-peroxy gel, and bands of LYO (CPD) at 281 (272) and 448 (442) cm⁻¹ can be related to 269 and 449 cm⁻¹ attributed to titania. The strongest band in LYO (or CPD) at 705 (700), and bands at 823 (821) and 916 (920) cm⁻¹ can be considered as analogues to the bands ascribed tentatively to metatitanic acid (H₂TiO₃). Furthermore, in the CPD spectrum, a very strong band at 150 cm⁻¹ and less intense bands at 635, 514, and 385 cm⁻¹ can be ascribed to anatase Raman active modes. These bands can be considered as a manifestation of nascent anatase crystals formed in the original Ti-peroxy gel matrix under high pressure conditions. The strong band at 150 cm⁻¹ (belonging to anatase) is missing in LYO sample since the sample is amorphous up to 300°C. This is in good agreement with data obtained by TEM and XRD.

The results of the specific surface area and porosity measurements for both, LYO and CPD sample are given in Table 1. According to the pore size LYO sample is mesoporous and in CPD sample pores are predominantly mesoporous with a small portion of micropores and macropores. CPD sample has more than three times higher specific surface area in comparison with LYO sample. Also the total pore volume is much higher in the CPD sample. The pore diameter between 3 and 4 nm ranks the substances among typical nanoporous aerogels. The measurements presented in Table 1 are in good agreement with the results of SEM and TEM observations (Figs. 1 and 2). Compact foils with flat surface of LYO samples exhibit lower surface area and total pore volume than the bulk material composed of porous 3D network of anatase nanocrystals forming the CPD sample.

In order to compare the thermal decomposition of prepared materials, the thermoanalytical measurements were performed (Fig. 5). The analysis of thermal decomposition of prepared titania aerogels showed that four reaction steps can be distinguished at the LYO material and two reaction steps can be distinguished at the CPD material, by combining information from TG and DTA measurements. The total mass loss is approximately 30% for LYO and 20% for CPD material, the main difference is likely the different amount of water in the sample. At the DTA curve of LYO, crystallization of amorphous phase into anatase can be observed which is in good agreement with XRD data. Since the CPD material is crystalline from the beginning, no crystallization can be observed on the DTA curve.

Fig. 5:

Comparison of TG (a), DTG (b) and DTA (c) of lyophilized and critical point dried titania aerogels (heating rate 5°C min⁻¹).

Conclusions

Yellow aqueous colloids of peroxopolytitanic acid can be transferred by lyophilization as well as by critical point drying into yellow highly porous nanostructured aerogels. Products of both types of drying exhibit lamellar morphology with significant differences. Lyophilization results in products consisting of very thin leaves with flat surface, while CPD provides rather bulk highly porous samples with less pronounced foil morphology. In accordance with this finding are significantly higher values of surface area and total pore volume found for the CPD sample. While the lyophilized sample is predominantly amorphous, the CPD dried sample contains significant content of nanocrystalline anatase phase even at room temperature. Upon heating, the CPD sample crystallizes continuously from the room temperature, while the lyophilized sample transforms to anatase suddenly at ~350°C. However, the anatase crystallites in the lyophilized sample do not grow significantly with rising temperature until the transformation to rutile, whereas in the CPD sample the size of anatase crystallites increases continuously. Such behavior could be explained by the different morphology of the samples. In the lyophilized sample with a typical laminar morphology the mass transport is limited and does not allow growth of anatase crystals. The bulk morphology of the CPD sample, on the other hand, creates more favorable conditions for mass transfer and thus the growth of crystals.