Features of Surface Structures of Alumina and Titanium Dioxide Nanoparticles Produced Using Different Synthesis Methods

1Buryat State University, 24a Smolin Str., Ulan-Ude 670000, Russia 2Khristianovich Institute of Theoretical and Applied Mechanics of Siberian Branch of Russian Academy of Science, 4/1 Institutskaya Str., Novosibirsk 630090, Russia 3Institute of Catalysis of Siberian Branch of Russian Academy of Science, 7 Lavrentiev Av., Novosibirsk 630090, Russia 4Lavrentiev Institute of Hydrodynamics of Siberian Branch of Russian Academy of Science, 15 Lavrentiev Av., Novosibirsk 630090, Russia 5Novosibirsk State Technical University, 20 Karl Marks Av., Novosibirsk 630073, Russia 6Institute of Physical Material Science of Siberian Branch of Russian Academy of Science, 6 Sahyanova Str., Ulan-Ude 670047, Russia


Introduction
Nanoparticles are a common class of inorganic materials, and although their fundamental properties have been studied, some of their characteristics are unique, despite their all having the same phase composition.These differences are determined by the synthesis conditions [1,2].The synthesis method and its conditions determine the interactions of atoms and their association when the nanoparticles are forming.Formation peculiarities may occur during nucleation, either during molecular cluster growth or during association of the nanoparticles [3].Previous studies [1,4] have shown that a change in the method used to synthesize nanosilica particles led to a change in the nanoparticle structures, from close-packed (aerosil) to strip-like (aerogel).Furthermore, increasing their size while keeping the synthesis method as is resulted in the (-OH) surface groups being replaced with (-OH) 2 .
The properties of nanoparticle surfaces, including the contents of the different functional groups and their heterogeneity, are the most important factors for determining the efficiency of their application.If the surface groups coincide, their heterogeneity, that is, the distribution of the reactivity, becomes critical.Although significant data on the surface properties of both titanium dioxide and alumina has been accumulated, interpretations of their donor-acceptor properties are sometimes contradictory.The intensity of the -OH-group absorption is proportional to the specific surface area [5], but the ratio of the bridging to terminal -OH groups [6], in terms of nanoparticles, can vary.Changes in the number of Lewis acid sites occur because of the change in the coordination number of aluminum or oxygen [7].
Obviously, the variation in the synthesis methods may play a crucial role in the catalytic activity of alumina nanoparticles and the photocatalytic activity of titanium dioxide.Thus, we carried out a comparative analysis of the surface properties of their nanoparticles, synthesized by various methods, to estimate the possible interaction between the nanoparticles and their environment.

Materials.
In the present study, Al 2 O 3 and TiO 2 nanopowders, described below, were investigated.
(A) An industrial electron-beam accelerator is used to "physically" evaporate raw solid alumina and titanium dioxide [8].The electron beam has a relativistic energy of 1.4 MeV with a power of up to 70 kW.The high-temperature vapor cools in the air flow to condense as nanoparticles.
(B) Commercial alumina and titanium oxide nanopowders are produced through the high-temperature flame hydrolysis of corresponding chlorides AlCl 3 or TiCl 4 , together with hydrogen and oxygen [9] (Evonik Industries).Nanopowders of alumina AEROXIDE Alu 130 and titanium dioxide AEROXIDE P25 were investigated.
(C) A plasma-chemistry method in which an aqueous solution of aluminum nitrate (Al(NO 3 ) 3 9H 2 O) is treated by a high-frequency plasma generator [10] (Siberian Chemical Combine).
(D) Commercial high-purity alumina nanopowder is manufactured from aluminum metal and alcohol, producing high-purity aluminum alkoxide, after which hydrated alumina is produced by the hydrolysis of aluminum alkoxide, with consequent calcination [11] (Sumitomo Chemical).The nanopowder AKP-50 was studied.
(E) To produce titanium oxide using method [12] (Nano Chemical Inc.) the modified wet sulfate process was used.This can be characterized briefly as . . .
It is necessary to emphasize that technologies (A), (B), and (C) are gas-phase methods of nanopowder production.

Specific Surface Area of Nanopowders, Phase Composition, and Shape of Nanoparticles.
The specific surface area (SSA) and average particle size (APS) of the powders were defined by a liquid nitrogen adsorption method (Brunauer-Emmett-Teller (BET) analysis, Sorbi-M) using the standard procedure.Nitrogen was used as a gas-adsorbent.The specific surface measurements were carried out in helium/nitrogen mixtures with nitrogen concentrations of 6, 9, 15, and 20%.

Fourier-Transform Infrared (FTIR) Spectroscopy.
The surface property measurements were carried out using Fourier-transform infrared (FTIR) spectroscopy with adsorbed pyridine.To obtain solid samples, the initial powders were mixed with barium fluoride, which is transparent around 1000 cm −1 and does not distort the spectrum of the -OH groups when mixed at a ratio of 1 : 2. The mixtures were pressed into thin wafers (20-30 mg/cm 2 ).Before adsorption, the samples were evacuated to 10 −3 torr at 500 ∘ C for 2 h.Pyridine was adsorbed at a temperature of 150-160 ∘ C for 15-20 min.The samples were then evacuated for 60 min to a pressure of less than 10 −2 torr at the same temperature.The spectra were recorded at a resolution of 4 cm −1 in the region of 1000-4000 cm −1 using a Fourier spectrometer (Shimadzu-8300) to acquire 200 spectra.
The concentrations of the Lewis-type centers were measured from the integrated intensity of the absorption bands at 1440-1460 cm −1 , using an integrated absorption coefficient of 3.5 [13].The strength was estimated from the location of the two bands, namely, the vibration of the pyridine ring at 1440-1460 and 1590-1630 cm −1 .

Ultraviolet-Visible Diffuse Reflectance Spectroscopy (UV-Vis DRS).
The optical properties of the materials were investigated by diffuse reflectance spectroscopy (UV-Vis DRS) using a UV-2501 PC (Shimadzu) spectrophotometer with an ISR-240A diffuse reflectance attachment.The TiO 2 powder samples were placed in a quartz cell with an optical path length of 2 mm.Spectra were recorded with respect to a reflection standard, BaSO 4 , in the range of 190-900 nm (11,000-54,000 cm −1 ) in 2-nm steps.The reflectance coefficients were converted to absorption coefficients using the Kubelka-Munk function, () = (1 − ) 2 /2, where  is the reflectance coefficient.The wavelength values were converted to wavenumbers, expressed in cm −1 .
Using UV-Vis-DRS spectra, at the edge of the fundamental absorption of the material, we calculated the width of the band gap.The dependences,  1/2 = (ℎ]), where  = ((1 − ) 2 /2) *  ℎ] , were derived.The power "1.2" was used due to the TiO 2 being an indirect semiconductor.It was allocated in an area in which the experimental points satisfied a linear law.Furthermore, the linear dependence was extrapolated to the -axis, and the intersection point corresponded to the value of the energy band gap of the semiconductor.

Structural Properties and Morphology of Nanopowders.
Table 1 lists the specific surface area and phase information.Figure 1 presents the X-ray results obtained for the alumina nanopowders.
The SSA for samples (1)-( 3) is in the range of 40-94 m 2 /g.Their main component is the -phase of an alumina boehmite line (Figure 1(a)).Sample (4), produced by high-temperature flame hydrolysis, has a higher specific surface area and, again, its main component is the -phase, as is usually mentioned by the manufacturer [9].Sample (5), produced by the plasmachemical method, has the lowest surface area of 5 m 2 /g and consists of approximately equal quantities of  and  phases (Figure 1(b)).Note that such an SSA value is not typical for either  or  alumina, which usually have specific surface areas of about 100 m 2 /g.Sample (6), produced by aluminum alkoxide hydrolysis, has a surface area of only 8 m 2 /g and consists of an -alumina phase as shown in Figure 1(c).The X-ray analysis data reveals that the alumina samples all have different phases.It is interesting that the real states of alumina samples (1)-( 5), which are produced through gasphase methods, are not strictly the -, -, and -phases, but they are all in an X-ray amorphous state.Details of X-ray studies of corresponding "semi-amorphous" states can be found in [14].
Figure 2 shows TEM images of the alumina, which clearly illustrate the differences in the shapes and sizes of the nanoparticles.The alumina particles of sample (1) have an almost ideal spherical shape (Figure 2 4) are not spherical but rather regular with rounded edges, so their APS of about 13 nm (calculated from SSA, Figure 2(b)) is a typical value [9].The particles of sample (5) are spherical but mostly hollow and exhibit a wide size distribution.Those of sample (6) are not spherical but irregular although they approximate to a rhomboid shape, which is typical for -alumina.The PSD cannot be obtained for samples ( 5) and ( 6) because the analysis software (JEM-100CX) gives very uncertain results for complex-shaped nanoparticles.Nevertheless, the PSD for samples (1) and ( 4) is narrow (Figures 3(a) and 3(b), calculated from the TEM results) while it is wide for sample (5).For sample (6), most of the particles fall into the range of 100-300 nm.
The specific surface area of the titanium dioxide powders ranges from 42 to 50 m 2 /g. Figure 4 presents the X-ray results for titanium dioxide.It is represented by two phases.Samples (7) and ( 8) are mixtures of similar amounts of anatase and rutile, while sample (9) contains only rutile (see Table 1).
Figure 2 shows that the particles of sample ( 7) have a structure that is similar to that of sample (1), that is, a spherical shape with an APS of 40 nm and a PSD similar to that of samples ( 1)-( 3).The particles of sample ( 8) are irregular in shape, with a particle size of about 30-35 nm but a sufficiently narrow size distribution.The particles of sample ( 9) are composed of comparatively large aggregates of acicular nanoparticles with a needle thickness of about 10 nm and a length of about 30-50 nm.
The bands at 120, 444, and 607 cm −1 correspond, respectively, to the  1 ,   , and  1 vibrational modes of rutile.The appearance of a band at 240 cm −1 in the rutile spectrum is associated with the multiphonon Raman process.A noticeable broadening and shift of the bands in the spectrum of sample (9), compared to the spectrum of monocrystalline rutile TiO 2 , is apparently due to the disordering of the rutile structure in this sample.
The results of Raman spectroscopy were in complete agreement with the data of the phase analysis performed by X-ray diffraction.Additionally, it was found that sample (7) is characterized by an abnormally high luminescence, considering the measurement when excited at 1064 nm.It can be concluded that the anatase in this sample has a high defect density.

Optical Properties.
The UV-Vis-DRS spectra were recorded to study the optical properties of the titanium dioxide samples (Figure 7).The results of the analysis indicate that the visible region has a weak absorption with a maximum at 12,700 cm −1 , which is likely to correspond to Ti 3+ cations with a coordination number of less than 6 [17].In the UV region, all the samples have an absorption of more than 25,000 cm −1 with various intensities, depending on the method of nanoparticle synthesis.For sample (9) (Figure 7), two absorption bands can be clearly identified with peaks at 30,000 and 35,500 cm −1 .
According to [17,18], these bands can be attributed to oxygen defects (O 2 − ), which may appear in the TiO 2 structure, depending on the method of sample synthesis.The analysis of the intensity of the absorption bands at 30,000, 35,500, and 12,700 cm −1 in the UV-Vis-DRS of all the samples showed that as the number of Ti 3+ cations with coordination numbers less than 6 increased, the number of oxygen defects (O 2 − ) in the titanium dioxide structure decreased.This points to the Ti 3+ cations with a coordination number of 6 most likely being stabilized on the surface or in the surface layer of the TiO 2 structure, while the oxygen defects (O 2 − ) exist within the TiO 2 bulk.
It is well known that TiO 2 is an indirect-gap semiconductor [18].It is characterized by a fundamental absorption edge (or band gap,   ). Figure 8 shows the absorption spectra of titanium dioxide samples with the function (ℎ]) 1/2 = (ℎ]), which determines the photon-energy values corresponding to the   value.Samples ( 7) and ( 8), based of   values, are two-phase and consist of anatase and rutile modifications, but with different ratios.This is completely consistent with the Xray diffraction (XRD) and Raman data.These   values are such that we can conclude that sample (9) has an intermediate value between the rutile   (3.0 eV) and anatase   (3.2 eV).However, the XRD and Raman methods clearly indicate that this sample has a rutile structure.This means that the band gap width for this sample is not a valid response, since sample ( 8) is significantly deficient in oxygen.
Thus, the UV-Vis-DRS data allows us to conclude that all the samples are defective to a certain degree and may exhibit different properties, depending on the nature of the defects, that is, whether they are cationic (Ti 3+ ) or anionic (O 2 − ), with their amount and localization being highly dependent on the sample synthesis method.

Surface Properties of Nanoparticles
3.4.1.Hydroxyl Cover. Figure 9(a) shows the FTIR spectra of the -OH groups in the alumina samples.For samples (1), ( 3), ( 4), and (6), the hydroxyl cover is close to that of conventional alumina.After evacuation at 500 ∘ C, each sample cleared the terminal (above 3750 cm −1 ) and the bridging (3650-3745 cm −1 ) -OH groups [13].Also, hydrogen-bonded -OH groups, represented by broad unresolved bands below 3600 cm −1 , were detected.For -alumina, we observed the same -OH groups, but the proportion of the terminal -OH groups is significantly lower.Samples (2) and ( 5) are characterized by an almost complete lack of terminal and bridging -OH groups and a very high proportion of hydrogen-bonded -OH groups.
Figure 9(b) shows the FTIR spectra of the hydroxyl groups in the titanium dioxide samples.Samples (7) and (8) show the -OH groups that are typical for anatase relating to the terminal and bridged Ti-OH groups, that is, absorption bands 3743 and 3670 cm −1 , respectively.Sample (9) has wide bands of hydrogen-bonded -OH groups.This is likely due to the lamellar crystallites of large rutile agglomerates in the sample (Figure 5).These crystallites are interconnected by hydrogen bonds, which shift the band of the -OH groups on the surface to below 3600 cm −1 and causes them to broaden.

Lewis Acid Sites.
The FTIR spectra of adsorbed pyridine are shown in Figure 10.The bands in the region below 1440 cm −1 and those between 1570 and 1580 cm −1 are attributed to the vibration of the pyridine rings (]CCN 19b and ]CCN8b, respectively [19]) of the physically adsorbed pyridine, which was not removed by evacuation at 150 ∘ C. The band at 1485-1495 cm −1 (]CCN 19a) relates to the vibration of the rings of the physically adsorbed pyridine, which is related to the Lewis acid sites.
Table 2 presents data on the concentration of the Lewis acid sites because Brønsted centers capable of protonating pyridine (region between 1530 and 1550 cm −1 (]CCN 19b)) have not been found.
Table 2 presents the concentrations of the Lewis centers in micromoles/m 2 and their strengths, characterized by the absorption-band frequency in the ranges 1440-1460 and 1590-1630 cm −1 .A previous study [5] showed that the strength of the Lewis centers, determined by the pyridine adsorption heat on the centers, can be estimated by the shift of the 1440-1460 and 1590-1630 cm −1 bands into the highfrequency region.An increase in the frequency causes the Lewis centers to become stronger.Note, however, that the strength of the Lewis centers on the surface of the titanium dioxide and alumina samples cannot be compared by the frequency shift.This is caused by the relationship between the frequency shift and heat of pyridine adsorption being different for the Zr and Al cations [5].
Note that, in certain cases, the nanoparticles displayed centers that could be comparable in terms of strength to alumina in the solid state.
The table shows that samples (1)-( 4) and ( 6), ( 7) are sufficiently homogeneous and that only a single type of center can be detected.Samples (5), (8), and (9) show the centers of two types (spectra obtained after evacuation at 400 ∘ C).An increase in the concentration of the stronger centers is obviously accompanied by a drop in the proportion of the weaker centers.
For alumina, sample (2) has the weakest acid sites, whereas those for sample (4) are the strongest.Sample (5) has acid sites with next level of strength.For titanium dioxide, sample (7) exhibits the largest concentration of acid sites at a frequency of 1453 cm −1 .The strength of the Lewis samples points to the need to involve a wider range of material analysis techniques and the importance of precisely characterizing the surface properties.
As was mentioned previously, the synthesis conditions determine the nature of the atomic interactions, the process, and the conditions of their association while organizing the nanoparticles.These differences can arise in every stage of cluster formation [3].When the cluster concentration is sufficiently high, it is possible to rebuild them, leading to a change in the surface structure.In a previous study [2], it was shown that the silica nanoparticles can exhibit a different surface distribution of silicon and oxygen atoms, which is explained by the Lewis center distribution.The samples investigated in [2] were synthesized using the same methods as those used to produce samples (1) and (4), and ( 7) and ( 8), indicating the same possibility for alumina and titanium dioxide nanoparticles.
In practice, such a variation in the surface properties results in a change in the interaction between the nanoparticles and their environment (polymers, metals, and so on).For example, it is possible to change the viscosity of a fluid by using nanoparticles which are chemically identical but which have been produced using different methods [20].

Conclusion
A comparative study of the volume and surface properties of nanoparticles was carried out by FTIR, UV-Vis DR, and Raman spectroscopy, as well as X-ray diffraction methods.It was shown that variations in the synthesis method can produce changes in the surface properties of the nanoparticles, while maintaining the phase composition of the material.The nanoparticles exhibit unique properties not normally associated with their phase.To identify the functional properties of the nanoparticles, which determine their interaction with other materials, comprehensive studies of the phase consistency and surface properties must be carried out.

Figure 9 :
Figure 9: FTIR spectra of hydroxyl groups in (a) alumina and (b) titanium dioxide samples.

Figure 10 :
Figure 10: FTIR spectra of adsorbed pyridine in (a) alumina and (b) titanium dioxide samples.

Table 1 :
Specific surface areas and phase information of samples.