Laser Desorption/Ionization Mass Spectrometry as a Potential Tool for Evaluation of Hydroxylation Degree of Various Types of Titanium Dioxide Materials

For many applications, TiO2 must have a unique surface structure responsible for its desirable physicochemical properties. Therefore the fast and easy methods of TiO2 surface characterization are of great interest. Heated TiO2 samples and dye-modified TiO2 samples were analyzed by laser desorption/ionization mass spectrometry. In the negative ion mode, two types of ions were detected, namely (TiO2)n− and (TiO2)nOH−. It has been established that the samples can be differentiated based on the relative ion abundances, especially with respect to the free hydroxyl group population. It indicates that laser desorption ionization mass spectrometry has the potential for the investigation of the surface properties of various TiO2 materials.


Introduction
TiO 2 is one of the world's most common material, widely used in many fields of science and branches of industry as well as in everyday life. For example, it has been applied as a pigment (e.g., in toothpaste [1]), as semiconductors (e.g., in solar cells [2]), as catalysts (e.g., in biodiesel production [3]), as gas sensors (e.g., as alcohol vapor sensor [4]), etc.
For many of its applications, TiO 2 must have a unique surface structure that provides desirable physicochemical properties of TiO 2 , necessary for its effective use. One of the most important features of the TiO 2 surface is the number and distribution of the hydroxyl groups present on it [5][6][7][8][9]. For example, it has been demonstrated and discussed in detail, that the surface hydroxylation degree is of crucial importance for selective adsorption of Cr(VI) [10], photocatalytic oxidation/degradation of methyl ethyl ketone [11], phenolphthalein [12], methyl orange, rhodamine B, p-chlorophenol [13], methylene blue [14], adsorption of CO 2 , SO 2 , NO 2 [15]. The most common method for the study of surface hydroxylation degree, or more precisely the tool which enables comparison of the TiO 2 hydroxylation degrees, is the Fourier-transform infrared spectroscopy (FTIR) [8][9][10][11][12][13][14][15], which sometimes is supported by X-ray photoelectron spectroscopy (XPS), temperature-programmed desorption or surface acid-base ion-exchange reactions method [7,10,13].
A specific TiO 2 application is in the mass spectrometry, namely as a substrate (solid matrix) in the surface-assisted laser desorption/ionization mass spectrometry [16][17][18][19]. There are also a number of papers reporting TiO 2 surface modification or generation of interesting gas-phase titanium-oxide clusters by subjectingTiO 2 surface to laser beam action [20][21][22][23][24]. The goal of this work is to check if laser desorption/ionization mass spectrometry (LDI-MS) may be used for TiO 2 surface characterization, namely to evaluate

Materials and Methods
Samples of heated TiO 2 were prepared by solvothermal hydrolysis of titanium tetraisopropoxide (Aldrich, Poznań, Poland) and heating (calcination) of the obtained material at different temperatures (100, 200, 300, 400, 500 and 600 • C) according to the procedure described elsewhere [25]. The obtained samples were characterized by FT-IR spectra recorded on an IFS-66/s spectrometer (Bruker, Billerica, MA, USA) using KBr powder as a diluent.
The LDI TOF MS (laser desorption/ionization time of flight mass spectrometry, was used to generate the gas-phase clusters. Portions of 1 µL of TiO2 suspension in methanol, were spotted onto the target (MTP 384 ground steel, Bruker Daltonics, Bremen, Germany). During MS experiments, an Ultraflex TOF/TOF spectrometer (Bruker Daltonics, Bremen, Germany) was operated in reflection mode (both positive and negative) in the range of m/z 50-1500. For m/z > 1000, the clusters were characterized by low intensity. This spectrometer is equipped with a SmartBeam II laser (λ = 355 nm). Metastable fragmentation of selected ions was induced by laser without further use of collision gas. The LIFT technology was employed [29]. The software flexControl v.3.4 was used for data acquisition and collection, whereas flexAnalysis v.3.4 was used for data manipulation, evaluation, and processing. Scheme 1. Structures of dyes used for modification of P25 TiO2. Figure 1 shows the FTIR spectra of heated TiO2 samples. The intensities of the absorption bands at about 3400 and 1630 cm −1 indicate that the number of hydroxyl groups decreases in the order TiO2100 > TiO2200 > TiO2300 >> TiO2400 >> TiO2500 ≅ TiO2600. The LDI TOF MS (laser desorption/ionization time of flight mass spectrometry, was used to generate the gas-phase clusters. Portions of 1 µL of TiO 2 suspension in methanol, were spotted onto the target (MTP 384 ground steel, Bruker Daltonics, Bremen, Germany). During MS experiments, an Ultraflex TOF/TOF spectrometer (Bruker Daltonics, Bremen, Germany) was operated in reflection mode (both positive and negative) in the range of m/z 50-1500. For m/z > 1000, the clusters were characterized by low intensity. This spectrometer is equipped with a SmartBeam II laser (λ = 355 nm). Metastable fragmentation of selected ions was induced by laser without further use of collision gas. The LIFT technology was employed [29]. The software flexControl v.3.4 was used for data acquisition and collection, whereas flexAnalysis v.3.4 was used for data manipulation, evaluation, and processing. The heating of TiO2 affects not only the hydroxylation degree, but also a number of other surface parameters, e.g., surface area, pore-volume, surface morphology, oxygen and titanium vacancies, surface roughness, surface micro-topography, etc. [30][31][32][33][34]. These parameters have a minor influence on the FTIR spectra, but may significantly affect the LDI mass spectra (which is easy to justify). Furthermore, the ions observed under LDI conditions are not only a result of the laser desorption/ionization process but also of the processes which occur in the laser plume containing energetic ions, neutrals, and electrons (in this hot cloud of particles, a number of gas-phase reactions can occur). Therefore, the key question is if moderate differences in the surface chemical composition, e.g., the hydroxylation degree, are reflected on the LDI mass spectra. Figure 2 shows the LDI mass spectrum of TiO2400, obtained in the negative ion mode, in the m/z range 50-200 and 200-1000 (for clarity), as a representative example. The other LDI mass spectra are shown in the supplementary material ( Figures S1 and S2). Thanks to the characteristic isotope signals (e.g., Figure S3), the titanium-containing ions can be easy identified and the lack of the characteristic isotope signals of titanium indicates that we deal with background/contaminant ions (e.g., that at m/z 113 most probably corresponds to [(HCOO)2Na] − , since formic acid is a common contamination of methanol, other contaminant ions may be PO3 − at m/z 79, CH3SO3 − at m/z 95, H2PO4 -and HSO4 − , both at m/z 97, [35]). Two types of titanium-containing ions were identified, namely (TiO2)n − and (TiO2)nOH − (the formers are open-shell ions). It is clearly seen that with increasing n, we deal with a decrease in the (TiO2)nOH − /(TiO2)n − ratio. It may be justified by the relative electron affinities of (TiO2)nOH and (TiO2)n [5]. The key question is if the observed relative abundances of the titanium-containing ions depend on the surface properties of TiO2 samples subjected to the LDI process. The presence of OH-containing ions suggests that it may be possible to correlate the ion abundances with the hydroxylation degree of the TiO2 surface.

Results and Discussion
The heating of the TiO2 results in an increase of particle size [25]; thus, it is expected that for TiO2 samples heated at higher temperatures, the ion abundances at a higher m/z range will be higher. However, the opposite situation was observed. It is visible that for TiO2 samples heated at higher temperatures, the ion abundances at a higher m/z range are lower ( Figure S2). Thus, it is plausible that this phenomenon may depend on the surface hydroxylation degree. Figure 3 shows the breakdown plots of the (TiO2)3OH − /(TiO2)nOH − ratio against the n number ((TiO2)3OH − is an abundant ion at m/z 257, (the ions TiO2OH − The heating of TiO 2 affects not only the hydroxylation degree, but also a number of other surface parameters, e.g., surface area, pore-volume, surface morphology, oxygen and titanium vacancies, surface roughness, surface micro-topography, etc. [30][31][32][33][34]. These parameters have a minor influence on the FTIR spectra, but may significantly affect the LDI mass spectra (which is easy to justify). Furthermore, the ions observed under LDI conditions are not only a result of the laser desorption/ionization process but also of the processes which occur in the laser plume containing energetic ions, neutrals, and electrons (in this hot cloud of particles, a number of gas-phase reactions can occur). Therefore, the key question is if moderate differences in the surface chemical composition, e.g., the hydroxylation degree, are reflected on the LDI mass spectra. Figure 2 shows the LDI mass spectrum of TiO 2 400, obtained in the negative ion mode, in the m/z range 50-200 and 200-1000 (for clarity), as a representative example. The other LDI mass spectra are shown in the Supplementary Material ( Figures S1 and S2). Thanks to the characteristic isotope signals (e.g., Figure S3), the titanium-containing ions can be easy identified and the lack of the characteristic isotope signals of titanium indicates that we deal with background/contaminant ions (e.g., that at m/z 113 most probably corresponds to [(HCOO) 2 Na] − , since formic acid is a common contamination of methanol, other contaminant ions may be PO 3 − at m/z 79, CH 3 SO 3 − at m/z 95, H 2 PO 4 − and HSO 4 − , both at m/z 97, [35]). Two types of titanium-containing ions were identified, namely (TiO 2 ) n − and (TiO 2 ) n OH − (the formers are open-shell ions). It is clearly seen that with increasing n, we deal with a decrease in the (TiO 2 ) n OH − /(TiO 2 ) n − ratio. It may be justified by the relative electron affinities of (TiO 2 ) n OH and (TiO 2 ) n [5]. The key question is if the observed relative abundances of the titanium-containing ions depend on the surface properties of TiO 2 samples subjected to the LDI process. The presence of OH-containing ions suggests that it may be possible to correlate the ion abundances with the hydroxylation degree of the TiO 2 surface. and (TiO2)2OH − have similar abundances as (TiO2)3OH − ). Except for the TiO2300 plot, the obtained plots, shown in Figure 3, reflect very well the hydroxylation degree, analogically as the FTIR spectrum shown in Figure 1.  The heating of the TiO 2 results in an increase of particle size [25]; thus, it is expected that for TiO 2 samples heated at higher temperatures, the ion abundances at a higher m/z range will be higher. However, the opposite situation was observed. It is visible that for TiO 2 samples heated at higher temperatures, the ion abundances at a higher m/z range are lower ( Figure S2). Thus, it is plausible that this phenomenon may depend on the surface hydroxylation degree. Figure 3 shows the breakdown plots of the (TiO 2 ) 3 OH − /(TiO 2 ) n OH − ratio against the n number ((TiO 2 ) 3 OH − is an abundant ion at m/z 257, (the ions TiO 2 OH − and (TiO 2 ) 2 OH − have similar abundances as (TiO 2 ) 3 OH − ). Except for the TiO 2 300 plot, the obtained plots, shown in Figure 3, reflect very well the hydroxylation degree, analogically as the FTIR spectrum shown in Figure 1.
It is worth adding that sometimes the desired specific properties of TiO 2 can be obtained by its calcination at a given temperature. For example, the highest photocatalytic activity of TiO 2 calcinated at 300 • C was attributed to the fact that at 300 • C the best trade-off between a few surface properties (including surface hydroxyl groups population) was reached [11]. In our case (Figure 3), it is clear that the surface properties of TiO 2 300 are responsible for the specific plot which deviates from the others. Materials 2021, 14, x FOR PEER REVIEW 5 of 8 It is worth adding that sometimes the desired specific properties of TiO2 can be obtained by its calcination at a given temperature. For example, the highest photocatalytic activity of TiO2 calcinated at 300 °C was attributed to the fact that at 300 °C the best tradeoff between a few surface properties (including surface hydroxyl groups population) was reached [11]. In our case (Figure 3), it is clear that the surface properties of TiO2300 are responsible for the specific plot which deviates from the others.
It can be taken for granted that the dye-modified TiO2 samples have similar free surface hydroxyl group populations and they differ only in the dye structures. Figure S4 (supplementary material) shows the LDI mass spectra of the unmodified P25 TiO2 and dye-modified samples and Figure 4 shows the breakdown plots of (TiO2)3OH − /(TiO2)nOH − ratio against the n number obtained for the unmodified P25 TiO2 and dye-modified samples.  It can be taken for granted that the dye-modified TiO 2 samples have similar free surface hydroxyl group populations and they differ only in the dye structures. Figure S4 (Supplementary Material) shows the LDI mass spectra of the unmodified P25 TiO 2 and dyemodified samples and Figure 4 shows the breakdown plots of (TiO 2 ) 3 OH − /(TiO 2 ) n OH − ratio against the n number obtained for the unmodified P25 TiO 2 and dye-modified samples. It is worth adding that sometimes the desired specific properties of TiO2 can be obtained by its calcination at a given temperature. For example, the highest photocatalytic activity of TiO2 calcinated at 300 °C was attributed to the fact that at 300 °C the best tradeoff between a few surface properties (including surface hydroxyl groups population) was reached [11]. In our case (Figure 3), it is clear that the surface properties of TiO2300 are responsible for the specific plot which deviates from the others.
It can be taken for granted that the dye-modified TiO2 samples have similar free surface hydroxyl group populations and they differ only in the dye structures. Figure S4 (supplementary material) shows the LDI mass spectra of the unmodified P25 TiO2 and dye-modified samples and Figure 4 shows the breakdown plots of (TiO2)3OH − /(TiO2)nOH − ratio against the n number obtained for the unmodified P25 TiO2 and dye-modified samples.  It is clear that dye-modification yielded similar changes in the course of the plots as heating (Figures 3 and 4). Namely, a substantial increase in the (TiO 2 ) 3 OH − /(TiO 2 ) n OH − ratio was observed for n ≥ 6. Furthermore, the dye structures have minor (or moderate at most) influences on the character of the plots. We also checked if the (TiO 2 ) n OH − /(TiO 2 ) n − ratio depends on the hydroxylation degree. The obtained plots, shown in the Supplementary Material ( Figures S5 and S6), can be used for the comparison/differentiation of the TiO 2 samples; however, the comparison of the hydroxylation degrees has to be performed with caution as briefly discussed in the Supplementary Material.
It was also found that the TiO 2 − /(TiO 2 ) 3 OH − ratio also may depend on the hydroxylation degree, as shown in the Supplementary Material, Figure S7. It may be argued that the abundances of ions at the low m/z range (TiO 2 − at m/z 80) may be affected by background/contaminant ions. However, it is plausible that the obtained TiO 2 − /(TiO 2 ) 3 OH − ratios reflect well the hydroxylation degree, especially for heated TiO 2 samples ( Figure S7). Furthermore, in contrast to the (TiO 2 ) 3 OH − /(TiO 2 ) n OH − ratio, the ratio TiO 2 − /(TiO 2 ) 3 OH − does not show any exceptional behavior of TiO 2 300. Therefore, Figure S7 can be a good supplement to the plots shown in Figures 3 and 4.
We also obtained the LDI mass spectra in the positive ion mode (Supplementary Material, Figure S8), but they were found to be useless for the purpose of this work. The ions (TiO 2 ) n + and TiO(TiO 2 ) n + were detected but in low abundance. Additionally, a number of abundant non-Ti-containing ions was detected as well (background/contamination).
There is a number of papers devoted to the gas-phase studies of neutral and charged titanium oxide clusters [23,24,[36][37][38][39][40][41][42], since they provide some insight into the properties of bulk titanium oxide at the molecular level. Therefore, the (TiO 2 ) n − and (TiO 2 ) n OH − ions were subjected to further analysis, to get the spectra of metastable ions (LIFT mass spectra). The gas-phase metastable decompositions of (TiO 2 ) n − ions were found to be trivial, namely, the loss of TiO 2 molecule occurred (Supplementary Material, Figure S9). The gas-phase metastable decompositions of (TiO 2 ) n − ions were trivial; namely, the loss of the TiO 2 molecule occurred (Supplementary Material, Figure S9). It has been reported that (TiO 2 ) n − ions are very reactive towards trace of gases inside the instrument and are not prone to lose TiO 2 molecule [23,24]. However, in tandem mass spectrometry, the gas-phase cluster ions may have different structures and dissociation behaviors [23]. The results obtained for (TiO 2 ) n OH − ions were quite surprising. In the obtained LIFT mass spectra, we detected the (TiO 2 ) n H 2 O − ions (Supplementary Material, Figure S10). Just to note, there is no doubt that in the full scan mass spectra, we deal with (TiO 2 ) n OH − ions, not with (TiO 2 ) n H 2 O − ions (Supplementary Material, Figure S3).
It is difficult rationalize how the (TiO 2 ) n H 2 O − ions can be formed from (TiO 2 ) n OH − ions. Most probably, at first (TiO 2 ) n OH − ions lose the OH • radical producing (TiO 2 ) n − ions. Such a process has already been observed for (TiO 2 ) 3 OH − ion [24]. We are aware that the occurrence of this process may be disputable since it is the formation of two odd-electron species from the even-electron ion. Furthermore, in mass spectrometry, OH • radical loss is not a favored process, even in electron ionization conditions. The produced (TiO 2 ) n − ions undergo two processes: losses of TiO 2 molecules and reactions with traces of water inside the mass spectrometer, and the processes may occur in any order. In other words, in our experiments, the (TiO 2 ) n − ions formed under LIFT conditions are very reactive towards gas-phase water impurities, in contrast to the (TiO 2 ) n − ions produced in the LDI source. It must be added that our results are not contrary to those reported in [23,34], which have been obtained in different conditions (e.g., with respect to the reaction time). On the other hand, it indicates that the behavior of the gas-phase titanium oxide cluster is very sensitive to the conditions used.