Protonation and Photocatalytic Activity of the Rb 2 La 2 Ti 3 O 10 Layered Oxide in the Reaction of Hydrogen Production

The Rb2La2Ti3O10 layered oxide was synthesized by the solid-state method. Three phases with different protonation degrees and intercalated water contents were obtained from the initial compound by the treatment with distilled water and hydrochloric acid. The obtained samples were characterized by powder X-ray diffraction, SEM, X-ray microanalysis, BET, DRS, and TG. It was found that the complete ion exchange of Rb for H in the layered oxide Rb2La2Ti3O10 proceeds through the formation of two metastable intermediate phases with average protonation degrees of 0.5 and 0.75, which successively transform from one to another. Each of these phase transformations is accompanied not only by the contraction of the interlayer distance but also by the displacement of adjacent perovskite layers by 1/2 of the cell parameter which results in the change in the space group. The photocatalytic activity of obtained samples decreases with the increase in the protonation degree, which correlates with the decrease in the intercalated water content.


Introduction
Photocatalytic production of hydrogen from renewable resources such as water or biomass is regarded as a promising way to deal with the problems of exhaustion of traditional fuel feedstocks and the environment pollution caused by their use [1,2].Many inorganic materials were proposed as photocatalysts for hydrogen production under UV and visible light irradiation [3][4][5].Layered oxides represent a unique class of photocatalysts, which have the abilities to intercalate substrate molecules into the interlayer space, to exchange interlayer cations for different structural units and to undergo the exfoliation process with the formation of oxide nanosheets [6].These properties already allowed creating highly efficient photocatalytic systems and open-wide possibilities in the future [7].
The A 2 La 2 Ti 3 O 10 (A = alkaline metal) oxides belong to the triple-layered Ruddlesden-Popper perovskite-like titanates.The photocatalytic activity of K 2 La 2 Ti 3 O 10 in the reaction of water splitting under UV light was first reported by Takata et al. [8].The high activity was explained by the possibility of the water molecules intercalated in the interlayer space to take part in the photocatalytic reaction.Takata et al. [9] attempted to confirm this mechanism for the isostructural titanate Rb 2 La 2 Ti 3 O 10 , which demonstrated an activity twice as high as the potassium-containing analogue.It was found that the partial substitution of Ti with Nb results in the loss of intercalation ability and a significant decrease in the photocatalytic activity.Moreover, isostructural layered oxides Li 2 La 2 Ti 3 O 10 and Na 2 La 2 Ti 3 O 10 , which do not intercalate water at ambient conditions, show very low efficiency in the reaction of hydrogen production under UV light [10,11].Thus, the photocatalytic activity of triple-layered titanates A 2 La 2 Ti 3 O 10 is likely to be connected with their intercalation ability, which is also the case for some other layered oxides [12,13].
During the past years, the K 2 La 2 Ti 3 O 10 layered oxide has become a popular research object as a photocatalyst [7], but, surprisingly, Rb 2 La 2 Ti 3 O 10 did not, despite that it has initially demonstrated much higher photocatalytic performance.This fact may be associated with a higher cost of rubidium in comparison with potassium.Anyway, we could not find any reports about photocatalytic properties of Rb 2 La 2 Ti 3 O 10 since the original work [8], in contrast with K 2 La 2 Ti 3 O 10 , which was investigated in more than 60 papers.These works cover a wide range of research directions devoted to the enhancement of photocatalytic properties, including improved synthesis techniques, doping, and design of composite photocatalysts.Nevertheless, we believe that there is a lack in these investigations, namely, the study of reactions between the layered oxide and aqueous solution.
It is a matter of common knowledge that many layered oxides undergo protonation, that is, substitution of interlayer cations for protons in acidic solution [14].The protonation process is a widely used technique in soft chemistry of layered oxides and is a starting point for many synthetic routes [6].However, there is a nonobvious fact that layered oxides may undergo protonation not only in acidic but also in neutral and even alkaline aqueous solution.Since the protonation process changes the structure of the interlayer space and affects the intercalation ability of the layered oxide, it may also have a dramatic impact on its photocatalytic properties.Unfortunately, this fact is usually overlooked in papers concerning the photocatalytic activity of A 2 La 2 Ti 3 O 10 -type titanates.
This paper is focused on the detailed study of protonation processes, which occur under contact of the Rb 2 La 2 Ti 3 O 10 layered oxide with water and acidic solution, and their influence on its photocatalytic activity.

Materials and Methods
The layered oxide Rb 2 La 2 Ti 3 O 10 was prepared by conventional solid-state reaction (1) in air at atmospheric pressure using Rb 2 CO 3 , La 2 O 3 , and TiO 2 as reagents (99.9%).The stoichiometric amounts of oxides with 50% excess of rubidium carbonate were ground in an agate mortar together with heptane to prevent hydration.Then, the mixture was pelletized by pressure and heated in a corundum crucible at 1000 °С for 10 h.
The as-prepared sample is denoted as S0.The synthesis of protonated samples was carried out by the treatment of the initial sample S0 with distilled water and 0.01 M HCl.During this procedure, two main approaches were used.At the first stage, 5 g of the S0 sample was placed into a closed flask together with 40 ml of distilled water.The suspension was shaken for 10 min and centrifuged at 1000g.Finally, the fugate was replaced by fresh water, and the whole procedure was repeated 24 times until a single-phase product (S1) was obtained.At the second stage, the S1 sample was placed on a vacuum filter (average pore diameter 50 nm) and distilled water was added to it dropwise at an average rate of 5 ml/min by a peristaltic pump (Watson-Marlow-400) and continuously removed through the filter.After passing of 6 l of water through the filter, another single-phase product (S2) with a higher protonation degree was obtained.Finally, water was replaced by a 0.01 M HCl solution, and the S2 sample was exposed to it on the vacuum filter.After passing of 4.5 l of acid through the filter, the final product (S3) was obtained.
The phase composition of the obtained samples was determined by powder XRD (Rigaku Miniflex II, CuKα).The indexing of XRD patterns and determination of cell parameters were performed on Topas software provided by Bruker.
The degree of protonation as well as the amount of intercalated water was determined by thermogravimetric analysis (TG) (Netzsch TG 209 F1 Libra), which was performed in argon atmosphere with a heating rate of 10 K/min from room temperature to 800 °C.Typically, two steps of mass loss were observed on the TG curve.The first step (80-250 °C) corresponds to the liberation of intercalated water and the second step (300-400 °C) refers to the dehydration of the partly protonated layered oxide as follows: The values x/2 (protonation degree) and y (amount of intercalated water per formula unit) were calculated on the basis of these equations [15].
The morphology of the samples was investigated by scanning electron microscopy (Zeiss Merlin with field emission cathode, electron optics column GEMINI-II, and oil-free vacuum system).Furthermore, energy-dispersive X-ray microanalysis of the samples was carried out (Oxford Instruments INCAx-act), and the protonation degree was calculated from the relative rubidium content in each sample.
The specific surface area of the samples was determined using the Brunauer-Emmett-Teller (BET) method (Quadrasorb SI) by measuring the amount of adsorbed krypton.
The optical band gap energy of the samples was determined by diffuse reflectance spectroscopy (DRS) (Shimadzu UV-2550 with ISR-2200 integrating sphere attachment).The band gap was determined from the crosspoint of linear sections of the Kubelka-Munk function (F) plot in the coordinates F ⋅ hν 1/2 = f hν .
To prepare the suspension for the photocatalytic experiment, 60 mg of the photocatalyst sample was added to 60 ml of aqueous solution containing 0.1% (mol) isopropyl alcohol and, in some experiments, 0.1 M KOH in order to maintain the pH constant.The suspension was shaken and left for 1 hour to establish equilibrium between the photocatalyst and solution.Straight before the experiment, each suspension was sonicated for 10 min (Elmasonic S10H ultrasound bath) to disaggregate the photocatalyst particles.The pH of each suspension was measured by a Mettler Toledo S220 SevenCompact pH meter equipped with an InLab Expert Pro-ISM electrode.
The photocatalyst suspension (50 ml) was placed in an external irradiation reaction cell, equipped with a powerful magnetic stirrer and a liquid cut-off filter and connected to 2 International Journal of Photoenergy a closed gas circulation system (120 ml dead volume).A medium-pressure mercury lamp DRT-125 (125 W) was used as a radiation source.Light reaches the reaction cell only after passing through a thermostated light filter solution at 15 °C (KCl + NaBr, 6 g/l each, 2 cm optical path), which cuts off radiation with λ < 220 nm.The total irradiation time was 120 min for each experiment.During the photocatalytic reaction, hydrogen accumulates in the gas phase, whose composition was analyzed by an online gas chromatograph at certain time intervals (Shimadzu GC-2014, Rt-Msieve 5A Column, TCD, Ar carrier).At the beginning of each experiment, the system was deaerated and argon gas was introduced at atmospheric pressure.

Results and Discussion
The initial Rb 2 La 2 Ti 3 O 10 sample (S0) was synthesized as a single phase without impurities; the obtained XRD pattern (Figure 1) is consistent with that in literature data [16] and corresponds to an I4/mmm space group with cell parameters a = 3.895 Å and c = 30.46Å.When we exposed the S0 sample to liquid water, we detected two main processes: the intercalation of water molecules into the interlayer space and the exchange of interlayer rubidium cations for protons.These processes are accompanied by phase transformations, that is, we observe the disappearance of initial peaks and appearance of new peaks on the XRD patterns instead of a continuous shift of the lattice parameters.Moreover, we discovered that there exists not only the fully protonated phase H 2 La 2 Ti 3 O 10 , which is described in literature, but also partially protonated phases with different space groups.During the treatment of the initial sample with water, we managed to obtain two intermediate phases S1 and S2 with different protonation degrees and different contents of intercalated water.Furthermore, by means of treatment of the S2 sample with 0.01 M HCl, we obtained the S3 sample, which was fully protonated.
According to TG data (Figure 2 and Table 1), the sample S1 has the composition HRbLa 2 Ti 3 O 10 •0.8H 2 O, that is, half of the rubidium ions are substituted with protons.The peaks on the XRD pattern of S1 have lower intensity compared to those of the initial S0 sample, and some of them are broadened (especially in the 28 °-33 °region).This may be due to the structural disorder (i.e., stacking disorder), caused by the displacement of perovskite layers during the ion exchange and intercalation processes.The XRD pattern of S1 can be well indexed in an orthorhombic space group С222 with a = b = 3.84 Å and c = 32.52Å, which corresponds to the displacement of adjacent perovskite slabs in one direction (a or b) by 1/2 of the cell parameter compared to the I4/mmm space group of S0.
Sample S2 has a higher protonation degree and a less content of intercalated water as compared to S1.Its composition can roughly be described as H 1.5 Rb 0.5 La 2 Ti 3 O 10 •0.5H 2 O.In contrast with S1, it is well crystallized and its XRD pattern is perfectly indexed in a P4/mmm space group with a = 3.82 Å and c = 14.90 Å, corresponding to an eclipsed configuration of adjacent perovskite layers.Therefore, the S1 sample not only changes its composition during flushing with 6 l of distilled water but also transforms it into a more symmetric and ordered form-S2.
Sample S3 is almost completely protonated and contains a minor amount of water.TG data show that in this case, water desorbs at lower temperature compared to that of S1 and S2, suggesting that it may be physically adsorbed rather than intercalated.The formula of S3 can be written as H 1.9 Rb 0.1 La 2 Ti 3 O 10 •0.16H 2 O. Compared to S2, it has a lower structure symmetry and its XRD pattern can be described by 3 International Journal of Photoenergy the orthorhombic space group С222 with a = b = 3.805 Å and c = 27.44 Å, which again corresponds to the displacement of adjacent perovskite slabs in one direction (a or b) by 1/2 of the cell parameter as well as in the case of S1.We can claim that this main phase of S3 is fully protonated, because the observed residual rubidium content is apparently related to the minor impurity of the S2 phase, which can be detected on the XRD pattern at 29.55 During the protonation process, we observe the decrease in the interlayer distance, which is related to the c lattice parameter.This is caused by two factors.On the one hand, bulky rubidium cations are replaced by compact protons.On the other hand, the amount of intercalated water also decreases.DRS experiments showed that the change in the interlayer distance correlates with the change in the band gap.The measured values of the band gap lie within the range of 3.55-3.65eV (Figure 3) which is in general agreement with literature data [9].It can be seen that the band gap energy increases in the series S3 < S2 < S0 < S1, in the same order as the interlayer distance (Table 1).The valence band and conductance band of layered titanates are formed mainly by O2p and Ti3d orbitals within the perovskite layer.The contribution of the alkali cation to the band formation in such compounds is usually negligible, so its nature should not have much impact on the band gap [17].However, the band     5 International Journal of Photoenergy gap depends on the interaction between orbitals of adjacent perovskite layers, which clearly depends on the interlayer distance, explaining the observed correlation.The variation of the band bap among investigated samples is not critical, but it may have an impact on the photocatalytic activity due to the change in the electron and hole reactivity.
We have demonstrated that the interaction of Rb 2 La 2- Ti 3 O 10 with water successively yields two metastable phases S1 and S2 with different protonation degrees, different contents of intercalated water, and even different structures.However, the observed phase transformations are accompanied by even more complex effects, namely, the increase in the specific surface area due to partial delamination of the layered oxide particles.These changes in morphology can be observed on the SEM images of the samples S0-S3 (Figure 4).While the initial sample S0 consists of bulky plate-like particles, for S1, we clearly observe fissures, which divide the particles into thinner units.The particles of the S2 and S3 samples, whose morphology is the result of this delamination process, have less thickness compared to those of S0, which results in the increased surface area.The BET surface area of all protonated samples exceeds that of the initial compound more than fivefold (Table 1).Hence, we could expect from the protonated samples an increased photocatalytic activity.However, experiments showed quite the contrary.
According to Figure 5, the photocatalytic activity of the layered oxide samples gradually decreases with the increase in the protonation degree in the series S0-S1-S2-S3.The superior activity was demonstrated by the initial S0 sample in the aqueous isopropanol solution without KOH.Under these conditions, its activity is ≈22 times larger than that of TiO 2 P25 Evonic-Degussa which was used for reference.The samples S2 and S3 show a quite low photocatalytic activity.Besides other reasons, this may be caused by the strong difference of the suspension pH (Table 2), which is in this case determined only by the equilibrium between the layered oxide and the solution (hydrolysis or dissociation).
At the same time, in the presence of KOH, the difference in the photocatalytic activity is not such critical.In this case, the pH is maintained constant.But, on the other hand, the protonated forms should be not stable at such alkaline pH values as they can act as solid acids.Protons in the interlayer space may become substituted by potassium ions from the solution, and this process is promoted by the neutralization reaction in an alkaline medium.The XRD pattern of protonated phases did not show significant changes after photocatalytic experiments.But even if the reverse ion exchange proceeds slowly and only at the surface of layered oxide particles, it may significantly affect the photocatalytic properties.Maybe, this is the reason for the fact that the hydrogen evolution rate over the S3 sample slowly increases with irradiation time.Generally, we should note the correlation between photocatalytic activity and the amount of intercalated water in the interlayer space.Apparently, this is the reason why the initial sample, which is most capable of water intercalation, demonstrates the best photocatalytic performance in both cases.
To conclude, we would like to focus on the practical importance of the obtained results.When the layered oxide is used as a matrix for the creation of nanostructured composite photocatalysts, reactions of intercalation and ion exchange are usually involved in the preparation process.However, the optimal synthesis conditions as well as the quality of obtained product should strongly depend on the composition and structure of the initial layered oxide.We have shown that the Rb 2 La 2 Ti 3 O 10 layered titanate forms three different protonated phases depending on the external conditions.These phases possess a different interlayer distance and a different amount of intercalated water; thus, most likely, their reactivity in ion exchange and intercalation processes should be different.Therefore, it is important to control which exactly protonated phase is involved in the reaction, when the layered oxide is taken as a starting compound.In the current paper, we have reported for the first time the existence of two intermediate protonated forms of Rb 2 La 2 Ti 3 O 10 and characterized them.We hope that these results will be useful for the design of new materials based on layered oxides, especially highperformance photocatalysts.

Conclusions
It was found that the Rb 2 La 2 Ti 3 O 10 layered oxide is not stable against rubidium leaching in aqueous solution.Two partially protonated phases of Rb 2 La 2 Ti 3 O 10 were obtained and characterized for the first time.The intermediate phases with compositions HRbLa 2 Ti 3 O 10 •0.8H 2 O and H 1.5 Rb 0.5 La 2- Ti 3 O 10 •0.5H 2 O were obtained by washing the initial sample with distilled water, while the final fully protonated phase was obtained by the treatment with 0.01 M hydrochloric acid.With the increase in the protonation degree, we observed the decrease in intercalated water amount and the contraction of the interlayer distance.Each phase transformation between protonated forms is accompanied by the displacement of adjacent perovskite layers by 1/2 of the cell parameter.Partial delamination of the layered oxide particles occurred during the protonation of 2 La Ti 3 O 10 resulting in an increase in the BET surface area.The photocatalytic activity decreases with the increase in the protonation degree, which correlates with the decrease in the intercalated water content.The initial nonprotonated layered oxide demonstrated the highest hydrogen production rate under UV irradiation.