A facile synthesis of well-defined titania nanocrystallites: Study on their growth, morphology and surface properties

Microporous and Mesoporous Materials 154 (2012) 187–195

Contents lists available at SciVerse ScienceDirect

Microporous and Mesoporous Materials journal homepage: www.elsevier.com/locate/micromeso

A facile synthesis of well-defined titania nanocrystallites: Study on their growth, morphology and surface properties Lenka Mateˇjová a,⇑, Zdeneˇk Mateˇj b, Olga Šolcová a a b

Institute of Chemical Process Fundamentals of the ASCR, v. v. i., Department of Catalysis and Reaction Engineering, Rozvojová 135, 165 02 Prague 6, Czech Republic Charles University in Prague, Faculty of Mathematics and Physics, Department of Condensed Matter Physics, Ke Karlovu 5, 121 16 Prague 2, Czech Republic

a r t i c l e

i n f o

Article history: Received 16 June 2011 Received in revised form 6 October 2011 Accepted 28 November 2011 Available online 6 December 2011 Keywords: TiO2 anatase Nanocrystalline structure Metal alkoxide X-ray diffraction Whole powder pattern modeling (WPPM)

a b s t r a c t The fast and simple synthesis of anatase nanocrystallites by the low-temperature hydrolysis of titanium (IV) alkoxides in the hydrogen peroxide solution followed by calcination of the amorphous titania peroxo-product at temperature in the range 300–450 °C was designed and studied. The phase composition and crystallite-size were evaluated from X-ray diffraction data by the advanced whole powder pattern modeling method together with the crystallite-size distribution and microstructure parameters such as the defect density. Besides that, the textural and surface properties, skeletal density, particle morphology and the purity of nanopowders were determined by nitrogen physisorption, helium pycnometry, mercury porosimetry, XPS, FESEM and the organic elementary analysis. The pure anatase nanocrystallites with the uniform globular structure and the narrow crystallite-size distributions were prepared at the significantly lower temperature (300 °C) in comparison with the conventional chemical methods. The size and shape of anatase crystallites were tailored by the calcination temperature and the type of titanium (IV) alkoxide. Titania nanocrystallites with the excellent uniformity were produced from titanium (IV) n-butoxide-type precursor. Correlation between the crystallite growth and the porous structure was observed up to temperature about 400 °C. Above this temperature the crystallite size monotonically increases, whereas the specific surface area is reduced steeply. Ó 2011 Elsevier Inc. All rights reserved.

1. Introduction A large part of the modern material research is focused on the synthesis of porous metal oxides (TiO2, ZrO2, SnO2, CeO2, ZrO2– CeO2, TiO2–ZrO2, Al2O3–CeO2, TiO2–CeO2, etc.) thanks to their helpful application in many research areas; ecological liquidation (oxidation caused by the UV light effect) of organic contaminants in waste water and air [1–4], dye-sensitized photo electrochemical cells [5], electro chromic devices [6–8], sensors [9], efficient supports of metal species in catalysts for VOCs [10,11] and CVOCs oxidation [12,13], etc. Applications of these porous metal oxides are based on their powdered form as well as on nanolayers deposited on various substrates like glass, silicon, metallic foils, etc. In general, various functionality and catalytic efficiency of these metal oxides are strongly affected by the differences in their textural properties (e.g. specific surface area, micropore volume, microand mesopore-size), structural properties (e.g. crystalline phase composition and/or crystallite size) [14], surface acidity, etc. Therefore, the precise and correct characterization of material properties is also highly requested.

⇑ Corresponding author. Tel.: +420 220 390 302; fax: +420 220 920 661. E-mail address: [email protected] (L. Mateˇjová). 1387-1811/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.micromeso.2011.11.054

Titania (TiO2), a member of the above mentioned group, is still under keen photo-catalytic research and many advanced preparation procedures/techniques improving its properties and catalytic performance have been developed and investigated; hydrothermal crystallization [15,16], sol–gel and templated sol–gel [17,18] in combination with super/subcritical drying [19], chemical vapor deposition [20], etc. Usually the synthesis of TiO2 crystalline phases requires the hard experimental conditions – high temperature (at least above 400 °C), high pressure (in SFE technique the required pressure is above 7.3 MPa) or some special media/ atmosphere (e.g. plasma/vacuum). Only a few preparation routes of the pure TiO2 synthesized at mild conditions have been described in literature. Owing to the higher purity of final TiO2 [21,22] the titanium (IV) alkoxides (in spite of their reactivity in moisture) are generally preferred as metal precursors instead of titanium chlorides. Wang et al. [23] studied the hydrolysis of titanium (IV) n-butoxide in H2O2 solution at near room temperature and they proved preparation of the anatase and rutile crystalline phase mixture by thermal treatment of the prepared poly-peroxotitanic complex above 150 °C in the air flow. The transformation of anatase to rutile started at significantly lower temperature (150 °C) in comparison with the other conventional methods, where the transformation usually occurs in the temperature range 500–600 °C. Moreover,

188

L. Mateˇjová et al. / Microporous and Mesoporous Materials 154 (2012) 187–195

Ti–O–O bonds were confirmed by FTIR analysis in titania samples up to 250 °C. Uekawa et al. [24] prepared pure TiO2 anatase with crystallite sizes from 9 nm to 15 nm in dependence on the boiling time (from 6 to 48 h) at 75 °C. For preparation the diluted solution made of alkoxide and hydrogen peroxide in ethanol was used. The prepared TiO2 anatase exhibited the BET surface area between 250 and 440 m2/g without subsequent calcination. Nevertheless, it possessed mainly microporous character. This study is focused on the preparation of TiO2 anatase by the hydrolysis of the titanium (IV) alkoxides in solution of hydrogen peroxide and on a thorough characterization with special attention on the complex preparation process of the pure TiO2 anatase nanocrystals with respect to the individual steps. A great attention is aimed at the precise and correct chemical and structural characterization of synthesized TiO2. The possible tailoring of the size and shape of anatase nanocrystallites by combination of the calcination temperature and the type of used titanium (IV) alkoxide is examined. Based on knowledge of peroxo bonds presented in titania samples up to 250 °C [23], the influence of calcination temperature was evaluated above this range. Besides the influence of the used metal precursor type the various calcination conditions; (i) temperature range (300–450 °C) and (ii) duration of thermal treatment (4–10 h), on textural and structural properties – mainly the crystallite growth and the phase composition – was investigated by an up-to-date whole powder XRD pattern modeling method (WPPM) [25]. This method was developed by Scardi and Leoni [26] in 2001 and it is based on a Rietveld-like fitting of the whole XRD pattern with special care of modeling width and shape of diffraction peaks. The WPPM overcomes [27] some approximations and drawbacks involved in the classical approaches as e.g. the Scherrer formula and thus it can more reliably reveal the finer details of the real sample microstructure such as the crystallite size distribution [28] or the defect density [29–31]. 2. Material and methods 2.1. Preparation Following chemicals were used for the synthesis: Titanium (IV) isopropoxide (Ti(OCH(CH3)2)4) and Titanium (IV) n-butoxide (Ti(O(CH2)3CH3)4) (Aldrich, purity > 98 + %), hydrogen peroxide (30% nonstabilized p.a., Lach-Ner) and distilled water. In a typical synthesis 20 ml of Titanium (IV) n-butoxide was added drop by drop to ice-cooled 50 ml 4.4 M H2O2 water solution at continuous agitation. The bath temperature was maintained at 1 °C, temperature of the mixture was kept constant at 1–2 °C during the whole experiment. During 45 min mixing the formation of the orange-colored TiH2O2 complex was observed together with the oxygen evolution (visible huge amount of bubbles in solution), water and alcohol. The experiment was finished after the formation of the oxygen bubbles nearly stopped. The precipitate was left on air for 24 h to release the rest of oxygen. Then the precipitate was dried in an oven at 150 °C for 5 h. During drying the orange precipitate converted into the yellow precipitate. After drying the precipitate was calcined above 300 °C for the defined time with the heating rate 1 °C/min in a muffle furnace in the air flow to guarantee the decomposition of Ti–O–O– bonds and alcohol. The sample nomenclature is explained in Table 1; footnote a. 2.2. Characterization of TiO2 powders The specific surface area, the micropore volume, the total pore volume, the mesopore surface area and the micro- and mesopore-size distribution were evaluated by the nitrogen physisorption at 196 °C performed on an automated volumetric

apparatus ASAP2020 Micromeritics (USA). The application of nitrogen and helium with high purity (99.9995%) guaranteed the precision of measured data. Before analysis samples were degassed for 24 h at 105 °C under vacuum (1 Pa). The true (helium) and apparent (mercury) density of prepared powders together with the meso- and macropore-size distribution were determined with the use of helium pycnometer (AccuPyc1330, Micromeritics, USA) and high-pressure mercury porosimeter (AutoPore III, Micromeritics, USA). Before analysis samples were dried at 105 °C for 24 h on air. XPS measurements were carried out using the ESCAProbeP (Omicron Nanotechnology) spectrometer equipped with monochromatic Al Ka X-ray source (1486.7 eV) and hemispherical electron analyzer. Sample information was determined in vacuum (1010 mbar) at measuring area with diameter 1 mm from surface layer 5–10 nm. The detection limit of analyzer was 0.1 atom.%, accuracy of binding energies ±0.2 eV. For component detection the NIST X-ray Photoelectron Spectroscopy Database was used. The field emission scanning electron microscope S-4800 Hitachi (Japan) was used for the study of the TiO2 particle morphology and the porous network topography. The purity (carbon content in wt.%) of powders was specified on the Vario EL III apparatus from Elementar. TiO2 powder (approximately 5 mg) was burned in oxygen atmosphere at temperature up to 1200 °C. Gaseous products (N2, CO2, H2O and SO2) were purified, separated to individual components and analyzed on TCD detector. All analyses were triplicated. The detection limit of the apparatus was 0.1 abs.%. XRD powder diffraction patterns were measured by the PANalytical-MPD diffractometer in the conventional focusing Bragg– Brentano geometry with variable slits. Ni-filtered characteristic CuKa radiation produced by a laboratory X-ray tube was used and diffracted intensity was registered by the PIXcel PSD detector. The patterns were collected in the diffraction angle range 2h = 8°– 140°. In order to correct the measured diffraction line broadening effects for the intrinsic instrumental broadening of the diffractometer the NIST LaB6 profile standard was measured in the same experimental setup. 2.3. XRD data evaluation procedure XRD patterns of the synthesized powders were analyzed by the WPPM method [26]. Attention was focused on a phase composition and a size of coherently scattering domains (crystallites). In WPPM the diffraction data are simulated from a microstructure model. For simplicity, crystallites were assumed to have a spherical shape. It was furthermore assumed that crystallite-size, diameter D, is distributed according to the log-normal distribution [32], which is generally found to be a suitable distribution describing the size dispersion in nanopowders of ceramic particles [28,33–35]. The log-normal distribution has two parameters [32]: the median, M, and the ‘multiplicative standard deviation’, r⁄. In literature the values of the mean crystallite size are usually presented. The arithmetic mean crystallite diameter, hDi, and the area weighted crystallite diameter, hDiA , can be easily calculated from M and r⁄ (see Appendix A). The WPPM can account the influence of the crystal structure defects as dislocations [25,31] or stacking faults. In [29] the dislocation density was evaluated in N-doped anatase nano-crystallites. However, such a proper modeling of effects induced by crystal defects is difficult and was not yet implemented for tetragonal structures in the fitting software used (MSTRUCT [36]). Hence a more simple phenomenological approach was adopted. The effect of the local lattice parameter variation (microstrain) connected with the crystal structure imperfection was simulated by a phenomenological pseudo-Voigt function, which was convoluted with other

L. Mateˇjová et al. / Microporous and Mesoporous Materials 154 (2012) 187–195

189

Table 1 Textural properties, apparent density, skeletal density and porosity of TiO2 prepared under the different calcination conditions. Samplea

SBETb [m2/g]

Smesoc [m2/g]

Vmicroc [mm3liq/g]

Vtotal,100d [mm3liq/g]

dporee [nm]

qHgf [g/cm3]

qHeg [g/cm3]

eh [%]

nBUT-TT300/4 nBUT-TT330/4 nBUT-TT380/4 nBUT-TT450/4 ISOP-TT400/4 ISOP-TT400/10

170 135 96 12.3 72 57

119 96 66 8.2 50 39

34 26 20 3 16 13

156 126 107 29 89 84

2.6 2.7 3.3 – 4.1 4.6

1.40 1.55 1.74 1.61 2.05 1.48

3.51 3.68 3.68 3.63 3.73 3.73

60 58 53 56 45 61

a

The sample nomenclature consists of ‘the used type of titanium (IV) precursor’ and experimental conditions of the thermal treatment ‘temperature(°C)/time(h)’. The specific surface area, SBET, evaluated from the nitrogen adsorption isotherm measured at 196 °C in the p/p0 range = (0.05–0.25) according to the classical BET theory [37]. c The mesopore surface area, Smeso, and micropore volume, Vmicro, evaluated from the nitrogen adsorption isotherm measured at 196 °C using the modified BET equation [38] and the t-plot method with the Lecloux–Pirard standard isotherm (C constant in the range 7–10) [39,40]. d The total pore volume, Vtotal,100, determined from the nitrogen adsorption isotherm measured at 196 °C at maximum p/p0 (0.99) = The total volume of pores with the diameter smaller than 100 nm. e The pore diameter evaluated from the adsorption branch of the nitrogen adsorption–desorption isotherm by the BJH method [41] via the Roberts algorithm [42], the empirical Lecloux–Pirard standard isotherm and the assumption of the cylindrical pore geometry. f The apparent density, qHg, evaluated from the high-pressure mercury porosimetry [43]. g The true density, qHe, evaluated from the helium pycnometry [44]. h The porosity, e, evaluated according to the equation e = 1(qHg/qHe). b

effects (crystallite-size and instrumental). The microstrain e(%) was considered as the relevant parameter characterizing the defects effects strength [36]. Parameters of the crystallite-size distribution (M and r⁄), microstrain e(%) together with the anatase crystal structure parameters (lattice parameters, isotropic temperature factors, etc.) were refined by the Rietveld-like fitting of measured XRD data using the MSTRUCT program. More details about implementation of the broadening effects employed can be found in [36].

3. Results and discussion 3.1. TiO2 textural properties Textural properties of the synthesized TiO2 powders in correlation with the calcination temperature, the length of heating time and the used type of titanium (IV) alkoxide are summarized in Table 1. Their effects on the porous structure illustrate Fig. 1(a)–(c). In Fig. 1(a) and (b) the nitrogen adsorption–desorption isotherms at 196 °C and the pore-size distributions evaluated from the nitrogen physisorption for all samples are presented. In Fig. 1(c) the pore-size distributions evaluated from the high-pressure mercury porosimetry are shown. From Table 1 the decrease of the mesopore surface area, Smeso, the micropore volume, Vmicro, as well the total pore volume, Vtotal,100, with the increasing calcination temperature and the length of heating time for all samples is evident. The increasing calcination temperature caused, in nBUT samples, a decrease of these textural parameters (DSmeso  93%, DVmicro  92%, DVtotal,100  81%) with significant impact on the mesopores. For nBUT-TT450/4 sample the higher temperature resulted in the disappearing of the small mesopores (Fig. 1(b)). These facts correspond to the difference in Fig. 1(c) in meso-macroporous region (pore radius above 103 nm) as well the significantly lower BET surface area in comparison with the other nBUT samples. The total porosity e decreased from 60% to 53% for nBUT samples. All these features could correspond to the gradual total sintration of the porous structure probably in consequence of the progressing crystallite growth. The longer heating time applied on ISOP samples caused the noticeable changes rather in larger mesopores and macropores than in the micro-mesoporous region. The increased porosity from 45% to 61% could be explained by the aggregation of the crystallites. These facts seem to be in agreement with the pore-size distribution curves shown in Fig. 1(b) and (c). Moreover, from Fig. 1(c) it can be seen that the network of larger mesopores

and macropores in nBUT samples is more irregular than in ISOP samples. However, the changes with the increasing temperature occur on the level of the whole porous structure and there are evenly progressing in the contrast to ISOP samples. The change of the porous structure in the ISOP couple with the longer heating time consists in the definite increase of the amount of macropores. The effect of the used metal precursor on the porous structure is evident from Fig. 1(a). The shape of the nitrogen adsorption– desorption isotherms of nBUT and ISOP samples differs, especially in hysteresis loops. Finally, there are only small differences in diameters of smaller mesopores. The true (helium) density of prepared TiO2 samples seems to be affected by temperature and mainly by the metal precursor used. For titanium (IV) n-butoxide the true densities vary between 3.50–3.68 g/cm3 contrary to the titanium (IV) isopropoxide for that the true densities are higher or equal to 3.73 g/cm3. The extension of heating time from 4 h to 10 h revealed no effect.

3.2. Surface properties and purity Results from XPS analysis which allow valuable insight into TiO2 surface properties are shown in Fig. 2(a) and (b). From the general XPS spectrum (Fig. 2(a)) it is evident that the surface (into 5 nm deep) of the ISOP-TT400/4 sample comprised only the expected elements: titanium, oxygen and carbon and no other contaminations on surface were detected. Titanium is present in TiOx form and according to the main peak position Ti 2p 3/2 (458.9 eV) the existence of TiO2 is obvious. The total amount of carbon on the surface (24.3 atom.%) significantly exceeds the average weight concentration of carbon (<0.1 wt.%) determined by the organic elementary analysis. Based on the values of binding energies (see Fig. 2(b)) carbon in chemically non-equivalent state was identified. The highest occurrence of C 1s (284.7 eV) belongs as usual to the removable contamination carbon; originally adsorbed CO2 from air. Other two spectral peaks (285.5 eV and 286.6 eV) could be assigned to the existence of formed hydrocarbons (C–C) or CHx. The spectral line at 288.7 eV corresponds to the carbon introduced during the preparation procedure (probably included in alkoxide molecules). The stoichiometric excess of oxygen was investigated using the Ar+ beam etching for the defined time period 600 s (i.e., up to 20 nm deep). The rapid decrease of the stoichiometric excess of oxygen from 21.5 to 3.7 atom.% in the first 300 s and the plateau in Ti/O ratio dependence during the last 300 s confirmed the

L. Mateˇjová et al. / Microporous and Mesoporous Materials 154 (2012) 187–195

190

Fig. 1. (a) Nitrogen adsorption–desorption isotherms at 196 °C, (b) pore-size distributions evaluated from adsorption branches of sorption isotherms and (c) pore-size distributions evaluated from high-pressure mercury porosimetry.

Fig. 2. (a) XPS survey spectrum measured with the ISOP-TT400/4 sample and (b) high resolution XPS spectrum of C(1s) signal.

presence of oxygen caused by the strong surface hydroxylation. Carbon was still detectable after 300 s (1.5 atom.%) and in the course of the next 300 s it decreased to 0.5 atom.%. This fact corroborates results from the organic elementary analysis in which the carbon content was below the detection limit (<0.1 wt.%). 3.3. XRD phase and profile analysis It is evident from measured XRD patterns (Fig. 3) that the anatase is a dominant crystalline phase in the all synthesised TiO2

samples. Furthermore, the anatase diffraction lines are broadened significantly due to the small crystallite size and the effect is different for various samples (see anatase 1 0 1 peak width – FWHMf values in Table 2). In the nBUT samples beside the anatase reflections a weak broad peak is visible at the diffraction angle 2h  30.8° (Fig. 4). According to Zhang and Banfield [46] the anatase nanocrystals are thermodynamically only slightly more stable than the brookite ones and the synthesis of the small anatase particles is likely accompanied by the formation of nano-brookite. Hence, the weak non-anatase peak in the nBUT samples was associated

L. Mateˇjová et al. / Microporous and Mesoporous Materials 154 (2012) 187–195

191

Fig. 3. XRD patterns of TiO2 powders prepared by hydrolysis of titanium (IV) alkoxides in solution of hydrogen peroxide under different calcination conditions.

Table 2 Structural parameters of TiO2 powders prepared under different calcination conditions.

a b c d e f

Sample

FWHMf

nBUT-TT300/4 nBUT-TT330/4 nBUT-TT380/4 nBUT-TT450/4 ISOP-TT400/4 ISOP-TT400/10

1.24 0.96 0.82 0.54 0.30 0.28

The The The The The The

a

[°]

Mb [nm]

r⁄c

hDid [nm]

hDiA e [nm]

Fractionf [wt.%]

4.6 5.9 8.5 11.5 12.2 11.9

1.46 1.46 1.39 1.46 1.82 1.84

4.9 6.3 9.0 12.4 14.6 14.3

6.6 8.5 11.2 16.5 30.0 30.2

93 93 96 97 100 100

full width in half of maximum, FWHMf, of anatase 101 reflection corrected for instrumental effects [45]. median, M, of the log-normal crystallites size distribution [26,32] as refined from XRD data fitting. multiplicative standard deviation, r⁄, of the log-normal crystallites size distribution [26,32] as refined from XRD data fitting. calculated arithmetic XRD mean crystallite size, hDi (Appendix A Eq. (2)). calculated area weighted mean XRD crystallites size, hDiA (Appendix A Eq. (2)). rough estimate of the anatase phase weight fraction as determined from XRD.

with the brookite phase (Fig. 4). From the refined weight fractions of the both TiO2 phases (Table 2) a trend of the brookite phase disappearing with the higher calcination temperature is noticeable. From the full widths in half of maximum values, FWHMf, in Table 2, it is clearly visible that for higher calcination temperatures the anatase 101 reflection is narrower which likely corresponds to the larger anatase crystallites. On the contrary it appears that the time period (4 h and 10 h) for which the ISOP samples were thermally treated had only negligible influence. These qualitative observations are confirmed by the refined values of the mean crystallite size hDi in Table 2 and by the crystallite-size distributions determined by WPPM, depicted in Fig. 5. Another observation concerns variances of the crystallite-size distributions. The multiplicative standard deviations, r⁄, in Table 2, for the nBUT samples are almost the same (r⁄  1.46), but considerably lower than that of

the ISOP samples (r⁄  1.82). This corresponds to the generally narrower crystallite-size distribution in the case of nBUT samples and it is also noticeable in the shape of crystallite-size distributions depicted in Fig. 5. In all samples microstrains e  0.2–0.3% were detected. The values indicate relatively small concentrations of the lattice defects in the synthesized anatase nanocrystals as compared with that prepared by rather ‘‘physical’’ methods, e.g. in [47], where microstrains about e  1% were observed in the magnetron sputtered titania thin films. The microstrain values showed no significant differences across the synthesized samples and no correlation with the calcination temperature. This gives no argue for a possible idea that strain fields of the some excess large structural defects, visible by XRD, can be reduced by the thermal treatment at temperatures below 450 °C.

192

L. Mateˇjová et al. / Microporous and Mesoporous Materials 154 (2012) 187–195

Fig. 4. Measured and calculated XRD patterns of the nBUT-TT300/4 sample. Detail of the anatase 1 0 1 and brookite reflections in the inset plot.

Fig. 5. Crystallite-size distributions evaluated from XRD data fitting.

From the XRD crystallite-size a theoretical total crystallites surface area per unit mass, SXRD, was calculated and compared with the specific surface area, SBET, measured by the nitrogen physisorption. SXRD is inversely proportional to the area weighted crystallites-size hDiA (see Appendix A). This relation is in literature (e.g. Weibel [34] and Audebrand [48]) rather used in an opposite way to estimate the particles size, DBET, from the measured specific surface area, SBET. Here SBET and SXRD are directly compared in Fig. 6. It is visible that with an exception of the nBUT-TT450/4 sample, treated at the highest temperature, both quantities are well comparable, which shows that majority of the crystallites surface is available to the gas thanks to the formed fine porous structure. Decrease of the specific surface area, SBET, with the inverse of the area weighted crystallites size, 1=hDiA , in the case of nBUT-TT300–380 samples, corresponds with the presumption of a total surface

Fig. 6. Correlation between the specific surface area of the porous structure, SBET, and the total surface area of crystallites, SXRD, estimated from XRD analysis. (d) nBUT samples, (N) ISOP samples, ( ) linear fit through the nBUT-300–380 samples, (  ) diagonal line.

reduction due to the crystallites growth with the increasing calcination temperature. The trend is steeply broken at the calcination temperature 450 °C, when SXRD  SBET, which likely indicates the crystallite agglomeration and the disintegration of the fine micro-mesoporous structure under these conditions. 3.4. TiO2 morphology from FESEM The field emission scanning electron microscopy was used for the morphology determination of the prepared titania powders.

L. Mateˇjová et al. / Microporous and Mesoporous Materials 154 (2012) 187–195

193

Fig. 7. FESEM microphotographs of the nBUT-TT450/4 sample, (a)–(d) different magnifications.

Fig. 8. FESEM microphotographs of the ISOP-TT400/10 sample, (a)–(d) different magnifications.

In Fig. 7 and Fig. 8 there are presented microphotographs of nBUTTT450/4 and ISOP-TT400/10 samples in various magnifications. In Fig. 7(a) and Fig. 8(a) the network of interconnected mesopores and macropores created by spaces between aggregates of nano-

crystals is clearly observable. Both, the cross-section and the shape of pores look irregularly. In the case of nBUT-TT450/4 sample the aggregates (pore walls) are composed of the highly uniform anatase nanocrystals with the pretty spherical shape from 10 to 15 nm

194

L. Mateˇjová et al. / Microporous and Mesoporous Materials 154 (2012) 187–195

in diameter (Fig. 7(b)–(d)). This observation corresponds relatively precisely with the crystallite-size and the narrow crystallite-size distribution evaluated from XRD (11.5 nmx/1.46). For ISOPTT400/10 sample (Fig. 8(b)–(d)) the shape and the size of individual titania nanocrystals do not seem to be so highly uniform – the sample appears as a mixture of the prime oval nanocrystals and the residual spherical ones. The presence of crystallites of the nonspherical shape, characteristic with a higher surface to volume ratio, could also explain an unexpectedly relatively good SBET values for the ISOP samples. In the case of ISOP-TT400/4 sample SBET is even slightly higher than SXRD (see Fig. 6). The FESEM indicates that the type of titanium (IV) alkoxide (nBUT or ISOP) used for the titania nanocrystal preparation significantly influences the morphology of the final anatase nanoparticles. 4. Conclusion Pure TiO2 anatase nanocrystals were successfully synthesized by the hydrolysis of titanium (IV) alkoxides in the dilute solution of hydrogen peroxide at temperature close to 2 °C, followed by calcination from 300 °C. Crystallite-size distributions of the prepared samples were determined by the whole X-ray powder pattern modeling method. It was shown that the type of the used metal precursor influences the size and the distribution of the TiO2 anatase crystallites. The anatase nanocrystallites prepared from the titanium (IV) n-butoxide possessed an excellent uniformity with a relatively narrow crystallite-size distribution in comparison with the crystallites prepared from the titanium (IV) isopropoxide. The size of the anatase crystallites increases with the calcination temperature, whereas the effect of the thermal treatment duration (4 h versus 10 h) on the anatase crystallite growth was not proved. These results were corroborated by FESEM, nitrogen physisorption as well the high-pressure mercury porosimetry. The influence of the type of the used titanium precursor on the TiO2 skeletal density was revealed by the helium pycnometry. XPS analysis confirmed the existence of TiO2 stoichiometry. Ar+ beam etching affirmed the strong surface hydroxylation and the high purity of the synthesized TiO2 samples. From the crystallite-size distribution determined by XRD the theoretical total surface area of crystallites for the each synthesized sample was estimated. For calcination temperatures up to 400 °C the samples show an excellent specific surface area in comparison with the theoretical values. It confirms the relatively low degree of the particles agglomeration and the formation of the fine porous structures in the final materials. Acknowledgements The financial support of the Academy of Sciences of the Czech Republic, Program Nanotechnology for Society (KAN400720701), the Grant Agency of the Czech Republic (104/09/P290) and the research program MSM0021620834 of Ministry of Education of the Czech Republic is gratefully acknowledged. The authors also thank Mr. Jirˇí Franc from J. Heyrovsky´ Institute of Physical Chemistry of the ASCR, v. v. i. for FESEM images. Appendix A The model describing broadening of XRD lines due to small crystallites size assumes the spherical particles with diameter D. The crystallite size, D, is statistically distributed according to the log-normal distribution [26,32–33]:

pffiffiffiffiffiffiffi pðDÞ ¼ 1=ð 2pDrÞ exp½ lnðD=MÞ=ð2r2 Þ:

ð1Þ

As suggested by Limpert et al. [32] with the median, M, and the ‘multiplicative standard deviation’, r⁄ = exp(r), the crystallite-size distribution can be interpreted analogously to the well-known normal distribution – as that 68.3% of crystallites have a size within the interval [M/r⁄, Mr⁄]  Mx/r⁄ and 95.5% in the interval [M/ (r⁄)2, M(r⁄)2], etc. The arithmetic mean crystallite diameter, hDi, the variance of the crystallite-size distribution, var2, and the area weighted crystallite diameter, hDiA , can be calculated from M and r [32,33] according to simple formulas:

hDi ¼ M expð1=2r2 Þ;

v ar2 ¼ hDi2 ðexpðr2 Þ  1Þ;

3

hDiA 

hD i hD2 i

¼ hDi expð2r2 Þ

ð2Þ

Within this simple model a theoretical total crystallites surface area per unit mass, SXRD, can be estimated. There are N crystallites in a gram of the sample:

N½1=g ¼

1021 ; R1 qstruct ½g=cm3   6 pðD½nmÞ3 pðDÞdD

ð3Þ

where qstruct is the structural density of the crystallites (for anatase qstruct = 3.89 g/cm3). The total surface, SXRD, of these crystallites is then:

SXRD ½m2 =g ¼ N 

Z

pD2 pðDÞdD ¼ 6  103

hD2 i

qstruct  hD3 i

3

¼

6  10

qstruct ½g=cm3   hDiA½nm

ð4Þ

and hence it is inversely proportional to the area weighted crystallite size hDiA [34,48]. References [1] D. Dong, P. Li, X. Li, Q. Zhao, Y. Zhang, C. Jia, P. Li, J. Hazard. Mater. 174 (2010) 859–863. [2] E.S. Elmolla, M. Chaudhuri, Desalination 252 (2010) 46–52. [3] T. Peng, D. Zhao, K. Dai, W. Shi, K. Hirao, J. Phys. Chem. B 109 (2005) 4947– 4952. [4] G. Waldner, M. Pourmodjib, R. Bauer, M. Neumann-Spallart, Chemosphere 50 (2003) 989–998. [5] Y.-Q. Wang, S.-G. Chen, X.-H. Tang, O. Palchik, A. Zaban, Y. Koltypin, A. Gedanken, J. Mater. Chem. 11 (2001) 521–526. [6] T. Ivanova, A. Harizanova, T. Koutzarova, N. Krins, B. Vertruyen, Mater. Sci. Eng. B 165 (2009) 212–216. [7] A. Verma, S.B. Amanta, N.C. Mehra, A.K. Bakhshi, S.A. Agnihotry, Sol. Energy Mater. Sol. Cells 86 (2005). 85–10. [8] C.O. Avellaneda, L.O.S. Bulhões, A. Pawlicka, Thin Solid Films 471 (2005) 100– 104. [9] G. Waldner, R. Gomez, M. Neumann-Spallart, Electrochim. Acta 52 (2007) 2634–2639. [10] H.L. Tidahy, S. Siffert, J.-F. Lamonier, E.A. Zhilinskaya, A. Aboukaïs, Z.-Y. Yuan, A. Vantomme, B.-L. Su, X. Canet, G. De Weireld, M. Frére, T.B. N’Guyen, J.-M. Giraudon, G. Leclercq, Appl. Catal. A 310 (2006) 61–69. [11] M. Hosseini, S. Siffert, R. Cousin, A. Aboukaïs, Z. Hadj-Sadok, B.-L. Su, Comptes Rendus Chimie 12 (2009) 654–659. [12] J.-M. Giraudon, T.B. Nguyen, G. Leclercq, S. Siffert, J.-F. Lamonier, A. Aboukaïs, A. Vantomme, B.-L. Su, Catal. Today 137 (2008) 379–384. [13] M. Kułazyn´ski, J.G. Van Ommen, J. Trawczyn´ski, J. Walendziewski, Appl. Catal. B 36 (2002) 239–247. [14] K. Kocˇí, L. Obalová, L. Mateˇjová, D. Plachá, Z. Lacny´, J. Jirkovsky´, O. Šolcová, Appl. Catal. B 89 (2009) 494–502. [15] M. Wu, G. Lin, D. Chen, G. Wang, D. He, S. Feng, R. Xu, Chem. Mater. 14 (2002) 1974–1980. [16] H. Hayashi, K. Torii, J. Mater. Chem. 12 (2002) 3671–3676. [17] H. Wang, J.J. Miao, J.M. Zhu, H.M. Ma, J.J. Zhu, H.Y. Chen, Langmuir 20 (2004) 11738–11747. [18] Y.-D. Wang, Ch.-L. Ma, X.-D. Sun, X.-D. Li, Mater. Lett. 54 (2002) 359–363. [19] L. Matejova, T. Cajthaml, Z. Matej, O. Benada, P. Kluson, O. Solcova, J. Supercrit, Fluids 52 (2010) 215–221. [20] A. Mills, N. Elliot, I.P. Parkin, S.A. O’Neil, R.J. Clark, J. Photochem. Photobiol. A 151 (2002) 171–179. [21] M. Addamo, V. Augugliaro, A. Di Paola, E. García-López, V. Loddo, G. Marcí, L. Palmisano, Colloid Surf. A 265 (2005) 23–31.

L. Mateˇjová et al. / Microporous and Mesoporous Materials 154 (2012) 187–195 [22] M. Addamo, V. Augugliaro, A. Di Paola, E. García-López, V. Loddo, G. Marcí, R. Molinari, L. Palmisano, M. Schiavello, J. Phys. Chem. B 108 (2004) 3303–3310. [23] Z. Wang, J. Chen, X. Hu, Mater. Lett. 43 (2000) 87–90. [24] N. Uekawa, J. Kajiwara, K. Kakegawa, Y. Sasaki, J. Colloid Interface Sci. 250 (2002) 285–290. [25] P. Scardi, Z. Kristallogr, Supplements 27 (2008) 101–111. [26] P. Scardi, M. Leoni, Acta Crystallogr. A 58 (2002) 190–200. [27] P. Scardi, M. Leoni, R. Delhez, J. Appl. Crystallogr. 37 (2004) 381–390. [28] M. Leoni, P. Scardi, J. Appl. Crystallogr. 37 (2004) 629–634. [29] F. Spadavecchia, G. Cappelletti, S. Ardizzone, C.L. Bianchi, S. Cappelli, C. Oliva, P. Scardi, M. Leoni, P. Fermo, Appl. Catal. B 96 (2010) 314–322. [30] M. Šlouf, R. Kuzˇel, Z. Mateˇj, Z. Kristallogr, Supplements 23 (2006) 319–324. [31] M. Leoni, J. Martinez-Garcia, P. Scardi, J. Appl. Crystallogr. 40 (2007) 719–724. [32] E. Limpert, W.A. Stahel, M. Abbt, Bioscience 51 (2001) 341–352. [33] J.I. Langford, D. Louër, P. Scardi, J. Appl. Crystallogr. 33 (2000) 964–974. [34] A. Weibel, R. Bouchet, F. Boulc’h, P. Knauth, Chem. Mater. 17 (2005) 2378– 2385. [35] S. Vives, C. Meunier, Powder Differ. 24 (2009) 205–220. [36] Z. Mateˇj, R. Kuzˇel, L. Nichtová, Powder Differ. 25 (2010) 125–131.

195

[37] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309–319. [38] P. Schneider, Appl. Catal. A 129 (1995) 157–165. [39] J.B. DeBoer, B.C. Lippens, B.G. Linsen, J.C.P. Broekhoff, A.V.D. Heuvel, Th.J. Osinga, J. Colloid Interface Sci. 21 (1966) 405–414. [40] A. Lecloux, J.P. Pirard, J. Colloid Interface Sci. 70 (1979) 265–281. [41] E.P. Barret, L.G. Joyner, P.B. Halenda, J. Am. Chem. Soc. 73 (1951) 373–380. [42] B.F. Roberts, J. Colloid Interface Sci. 23 (1967) 266–273. [43] Appendix B, C, D and H, in AutoPore III – Operator’s Manual, Micromeritics Instrument Corporation, 1998. [44] Appendix B, Analysis theory, in: Multivolume Inserts Kit for the AccuPyc™ 1330 – Operator’s Manual Addendum, Micromeritics Instrument Corporation, 1997. [45] T.H. Keijser, J.I. Langford, E.J. Mittemeijer, A.B.P. Vogels, J. Appl. Crystallogr. 15 (1982) 308–314. [46] H. Zhang, B. Chen, J.F. Banfield, G.A. Waychunas, Phys. Rev. B 78 (2008). 2141061–12. [47] R. Kuzˇel, L. Nichtová, Z. Mateˇj, J. Šícha, J. Musil, Z. Kristallogr, Supplements 27 (2008) 287–294. [48] N. Audebrand, J.-P. Auffrédic, D. Louër, Chem. Mater. 12 (2000) 1791–1799.