Nanoparticulate Perovskites for Photocatalytic Water Reduction

SrTiO3 and BaTiO3 nanoparticles (NPs) were activated using H2O2 or aqueous HNO3, and pristine and activated NPs were functionalized with a 2,2′-bipyridine phosphonic acid anchoring ligand (1), followed by reaction with RuCl3.3H2O and bpy, RhCl3.3H2O and bpy, or RuCl3.3H2O. The surface-bound metal complex functionalized NPs were used for the photogeneration of H2 from water, and their activity was compared to related systems using TiO2 NPs. The role of pH during surface complexation was found to be important. The NPs were characterized using Fourier transform infrared (FTIR) and solid-state absorption spectroscopies, thermogravimetric analysis mass spectrometry (TGA-MS), and matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS), and the dihydrogen generation was analyzed using gas chromatography–mass spectrometry (GC-MS). Our findings indicate that extensively functionalized SrTiO3 or BaTiO3 NPs may perform better than TiO2 NPs for water reduction.


Introduction
Increasing global economic development and a growing population have resulted in a higher demand for energy [1]. The burning of fossil fuels to provide energy releases greenhouse gases, which places immense stress on the environment. In the long term, this will cost society increasing amounts of resources [2][3][4][5]. As a consequence, research focusing on inexpensive energy solutions is increasingly focused on renewable and clean sources [6]. One solution is the use of dihydrogen derived from renewable sources as fuel or energy storage [7]. An attractive approach to a hydrogen economy is to combine energy harvesting and H 2 evolution using photocatalysts operating under solar irradiation [8]. Hence, the design and development of efficient photocatalysts are crucial. A major contribution to this development is the use of heterogeneous catalysts [9][10][11][12]. These are often easier to recover than homogeneous catalysts but have the disadvantage of inactive interior volumes, with only surface sites being catalytically active [13][14][15]. An alternative is functionalizing nanoparticle (NP) scaffolds with photocatalysts. Such immobilized photocatalysts offer greater catalyst-to-volume ratios than bulk heterogeneous catalysts, and this enhances the catalytic activity and turnover. An additional benefit of NPs is the possibility of dispersing them in liquid phases [16][17][18][19].
In previous studies, we have shown that the binding of photo-and redox-active Rh and Ru coordination compounds onto TiO 2 NP surfaces can be used successfully for H 2 production, with the catalytic activity of the NP-supported catalyst outperforming previously reported, related homogeneous catalysts [20]. Another promising class of photocatalysts are perovskite-type oxides MTiO 3 (M = Ca, Sr or Ba). These comprise cheap, earth-abundant elements [21][22][23], are water-insoluble, thermodynamically stable, and resistant to temperatureand photo-corrosion [24]. The photocatalytic properties of MTiO 3 can be improved, tuned, and modified. This has been demonstrated using alternative synthetic methods [25,26], doping with organic or inorganic compounds [24,27,28], surface nanoparticle deposition [29,30], the use of cocatalysts [31,32], or the use of composites [33,34].
In this work, we report the surface activation of pristine SrTiO 3 and BaTiO 3 NPs using either HNO 3 or H 2 O 2 . The activated NPs are abbreviated as SrTiO 3 -a, BaTiO 3 -a (HNO 3 activation) or SrTiO 3 -OH, BaTiO 3 -OH (H 2 O 2 activation), respectively. We compare the subsequent functionalization of each type of pristine and activated NP with the ligand [2,2bipyridine]-4,4 -diylbis(phosphonic acid) (1, Scheme 1) which contains phosphonic acid anchoring domains. The functionalized NPs were used for direct surface metal complex assembly by reaction with 2,2 -bipyridine (bpy) and ruthenium or rhodium trichloride to give a surface-bound complex presumed to be (but not established as) an [M(bpy) 2 (1)]species. The photocatalytic behavior of these metal-functionalized NPs was investigated. Experiments were also conducted to highlight the importance of pH control for successful metal complex assembly on the NP surface. The NPs were characterized using Fourier transform infrared (FTIR) and solid-state absorption spectroscopies, thermogravimetric analysis mass spectrometry (TGA-MS), and matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS), and the dihydrogen generation was analyzed using gas chromatography-mass spectrometry (GC-MS).
to give a composite material shown to generate H2 upon i ising feature of SrTiO3 and BaTiO3 NPs is their relatively tentials and band gaps, only ~0.2 eV larger than TiO2 [41-4 ious anchoring groups, including phosphonic acids, car groups, makes them suitable candidates to extend our ea NPs [47,48].
In this work, we report the surface activation of pristin either HNO3 or H2O2. The activated NPs are abbreviated a tivation) or SrTiO3-OH, BaTiO3-OH (H2O2 activation), res sequent functionalization of each type of pristine and act bipyridine]-4,4′-diylbis(phosphonic acid) (1, Scheme 1) w anchoring domains. The functionalized NPs were used fo assembly by reaction with 2,2′-bipyridine (bpy) and ruthe give a surface-bound complex presumed to be (but not e species. The photocatalytic behavior of these metal-functi Experiments were also conducted to highlight the importa metal complex assembly on the NP surface. The NPs we transform infrared (FTIR) and solid-state absorption spe analysis mass spectrometry (TGA-MS), and matrix-assis mass spectrometry (MALDI-MS), and the dihydrogen gen chromatography-mass spectrometry (GC-MS). Scheme 1. Anchoring ligand 1.

General
RuCl3 . 3H2O  NPs were purchased from Sigma-Aldrich Chemie GmbH (Buchs, Switzerland) and had diameter sizes of <100 and 50 nm, respectively; further characterizations are reported in the Supplementary Materials. The anchoring ligand 1 was prepared as previously described [20]. Instrumentation details are given in the Supplementary Materials. The calculated major MALDI peaks reported in the Supplementary Materials were determined using the most abundant isotopes (e.g., 102 Ru, 35 Cl). For functionalized NPs in which the surface-bound ligand 1 is coordinated to a metal, a simplified notation is introduced; for example, Ru@SrTiO 3 describes a SrTiO 3 NP modified with ligand 1 which is also coordinated to Ru. In the case of NPs functionalized with both Ru and Rh, the relative amounts of the two metals are indicated using lower case (lower concentration) or upper case (higher concentration) letters (r or R). The first letter refers to Ru and the second to Rh, e.g., rR@SrTiO 3 describes metal complex-functionalized NPs with Ru/Rh in a 1:20 molar ratio on the surface.

Experimental
2.2.1. Nanoparticle Surface Activation Using HNO 3 (SrTiO 3 -a, BaTiO 3 -a) Pristine SrTiO 3 or BaTiO 3 NPs were activated as previously reported [20]. The NPs (2.00 g) were dispersed by sonication for 15 min in dilute aqueous HNO 3 (30 mL, 3.0 M). The mixture was then stirred for 30 min. The suspension was centrifuged (10 min, 7000 rpm) and the NPs were washed once with milliQ water (40 mL). The NPs were added to milliQ water (40 mL) and dispersed by sonication for 10 min. The suspension was then stirred for 72 h. The suspension was centrifuged (10 min, 7000 rpm), and the NPs were washed with milliQ water (2 × 40 mL). The activated NPs (SrTiO 3 -a, 1.67 g and BaTiO 3 -a, 0.60 g) were stored in a sealed vial under N 2 after drying them under high vacuum. The characterization data of the pristine NPs and acid-activated NPs are given in the Supplementary Materials. The TGA and TGA-MS spectra of the pristine NPs ( Figures S1-S4)  Pristine SrTiO 3 or BaTiO 3 NPs were activated as reported in the literature [49]. The NPs (1.05 g) were dispersed by sonication for 20 min in H 2 O 2 (50 mL, 30%). The mixture was then stirred for 4 h at 110 • C under N 2 . The suspension was cooled down, centrifuged (10 min, 7000 rpm), and the NPs were washed with milliQ water (40 mL The functionalization was performed as previously reported [20]. Anchoring ligand 1 (10.0 mg, 31.6 µmol, 1.0 eq.) and milliQ water (18 mL) were added to a microwave vial and dispersed by sonication for 1 min. Acid-or H 2 O 2 -activated NPs (449.0 mg, 4.6 SrTiO 3 eq. or 7.3 BaTiO 3 ) were added. The suspension was dispersed by sonication for 10 min. The microwave vial was sealed, and the reaction mixture was heated for 3 h at 130 • C in the microwave reactor. After cooling to room temperature, the suspension was centrifuged (20 min, 7000 rpm). The NPs were separated from the solvent and washed with EtOH (2 × 15 mL). This procedure gave white f-NPs (f-NP = functionalized NP) with the following yields: 1@SrTiO 3 -a NPs (418.3 mg), 1@SrTiO 3 -OH NPs (422.6 mg), 1@BaTiO 3 -a NPs (409.9 mg), 1@BaTiO 3 -OH NPs (425.1 mg), and these were stored in a sealed vial under N 2 after drying the NPs under high vacuum. This reaction was repeated using pristine SrTiO 3 and BaTiO 3 NPs (224 mg, 4.6 SrTiO 3 eq. or 7.3 BaTiO 3 eq.) and 1 (5.0 mg, 15.8 µmol) yielding 1@SrTiO 3 NPs (218.1 mg, 221.6 mg) and 1@BaTiO 3 NPs (210.7 mg, 219.4 mg). For NMR spectroscopic measurements, NPs (5-10 mg) were dispersed in 500 µL  The metal complex was formed directly on the NP surface. 1@SrTiO 3 (70.9 mg), RuCl 3 . 3H 2 O (1.03 mg, 3.94 µmol), and bpy (1.25 mg, 8.00 µmol) were added to a vial. H 2 O (5.0 mL) and EtOH (3.0 mL) were added, and the mixture was thoroughly dispersed using sonication and stirring. The suspension was transferred into an autoclave PTFE liner with additional EtOH (2.0 mL). The autoclave was sealed and then heated in an oven at a heating rate of 320 • C/h to 160 • C. The autoclave was left at 160 • C for 1 h. After cooling, the autoclave was opened, and the suspension was centrifuged (20 min, 7000 rpm). The resulting NPs were washed with H 2 O (3 × 10 mL) and EtOH (1 × 10 mL). Ru@SrTiO 3 (65.8 mg) was obtained as a pale orange powder after drying the NPs under a high vacuum. This experiment was repeated with 1@BaTiO 3 (70. 2.2.5. Nanoparticle Surface Complexation (rR@SrTiO 3 , rR@SrTiO 3 -a, rR@SrTiO 3 -OH, rR@SrTiO 3 -OH-A, rR@BaTiO 3 , rR@BaTiO 3 -a, rR@BaTiO 3 -OH, rR@BaTiO 3 -OH-A) The notation is defined in Section 2.1. The metal complex was formed directly on the NP surface. This reaction was carried out with each isolated NP from Sections 2.2.1 and 2.2.2. Hence, 1@SrTiO 3 , 1@SrTiO 3 -a NPs, 1@SrTiO 3 -OH, 1@BaTiO 3 , 1@BaTiO 3 -a, or 1@BaTiO 3 -OH (365.6 mg) was added to a vial with RuCl 3 . 3H 2 O (0.24 mg, 0.9 µmol), RhCl 3 . 3H 2 O (5.15 mg, 19.6 µmol), and bpy (6.45 mg, 41.3 µmol). H 2 O (5.0 mL) and EtOH (3.0 mL) were added, and the mixture was thoroughly dispersed using sonication and stirring. The suspension was transferred to an autoclave with a PTFE liner with additional EtOH (2.0 mL). The autoclave was sealed and then heated in an oven at a rate of 320 • C/h to 160 • C. The autoclave was left at 160 • C for 1 h. After cooling, the autoclave was opened, and the suspension was centrifuged (20 min, 7000 rpm). The resulting NPs were washed with H 2 O (3 × 10 mL) and EtOH (1 × 10 mL). After drying the NPs under high vacuum, the reaction yielded rR@SrTiO 3 (354.2 mg) as a pale orange powder, rR@SrTiO 3 -a (358.2 mg) as a pale orange powder, rR@SrTiO 3 -OH (358.4 mg) as a grey powder, rR@BaTiO 3 (349.9 mg) as a grey powder, rR@BaTiO 3 -a (356.9 mg) as a dark brown powder or rR@BaTiO 3 -OH (357.4 mg) as a dark grey powder. This reaction was repeated with 1@SrTiO 3 -OH and 1@BaTiO 3 -OH with a pH adjusted to 1.5 using 4.0 mL H 2 O, 1.0 mL aqueous H 2 SO 4 (1 M), and the same amount of EtOH (5 mL). Other reaction conditions were kept the same.

Dihydrogen Generation
The water reduction reaction conditions were consistent with the conditions in the previous work with TiO 2 NPs [20]. Triethanolamine (TEOA) was added as a sacrificial electron donor, K 2 [PtCl 4 ] as a catalyst to facilitate H 2 formation (feasibly by the formation of Pt nanoparticles), bpy as an additive, aqueous H 2 SO 4 for altering the solution pH. TEOA (2.52 mmol, 376 mg), K 2 [PtCl 4 ] (1.7 µmol, 0.70 mg), and bpy (18.6 µmol, 2.91 mg) were added to a 5 mL microwave vial with milliQ water and aqueous H 2 SO 4 (1 M) to modify the pH. The experiments were performed at pH 7.5 and used 1 mL aqueous H 2 SO 4 (1 M) and 5 mL milliQ water. Metal complex-functionalized NPs were added (114.1 mg). The vial was flushed with N 2 and then sealed. The suspension was sonicated (10 min) and thoroughly shaken. N 2 was bubbled through the suspension for 10 min. The vial was irradiated using a sun simulator generating 1200 W m −2 (see the Supplementary Materials) for 4 h with an incident angle of light of 5 • . The suspension was stirred throughout the irradiation and periodically shaken. Headspace samples for gas chromatography were collected using a syringe and transferred to a 10 mL GC vial for analysis. The measured GC integral was converted into mL of H 2 with calibration performed by injecting several known volumes of H 2 .

Results and Discussion
3.1. NP Activation, Functionalization, Surface Complexation, and Material Characterization 3.1.1. NP Surface Activation The surface activation of NPs is a crucial step for successful and stable surface functionalization. We previously described the benefits of acid activation of TiO 2 NPs for their use in catalysis [20]. However, in the literature, other methods are also described, including using various acids (e.g., MeCO 2 H or HF), hydrogen peroxide, or plasma for surface activation [49][50][51]. These methods are used to enhance the surface reactivity and make the surface more reactive for functionalization. TiO 2 and SrTiO 3 NPs are acid-resistant, whereas BaTiO 3 NPs are sensitive to mineral acids. Hence, as described in Section 2.2.1, during acid activation with aqueous HNO 3 (3 M), the BaTiO 3 NPs partially dissolved or were lost during washing (70%). Since acid activation was unsuitable for the BaTiO 3 NPs, other activation methods were explored, and activation using the H 2 O 2 treatment [49] was chosen. This method simultaneously saturates the surface with hydroxyl groups and strips it of carbonate groups. Although the literature protocol required hydroxyl groups for a salinization reaction [49], the hydroxyl groups should also favor condensation with the phosphonic acid of the anchoring ligand.
Activation was performed with both SrTiO 3 and BaTiO 3 by boiling the dispersed NPs under N 2 in 30% H 2 O 2 . The particles were washed with water and then dried for 72 h under high vacuum. The pristine and activated NPs were characterized using FTIR and TGA-MS. The FTIR differences shown in Figure 1 were limited to the fingerprint region. For the pristine SrTiO 3 activated with H 2 O 2 , an additional absorption at 1446 cm −1 appeared. The pristine BaTiO 3 activated with acid showed the disappearance of a prominent peak at 1420 cm −1 , while the activation with H 2 O 2 caused the peak to shift and broaden. diated using a sun simulator generating 1200 W m −2 (see the Supplementary Materials) f 4 h with an incident angle of light of 5°. The suspension was stirred throughout the irr diation and periodically shaken. Headspace samples for gas chromatography were co lected using a syringe and transferred to a 10 mL GC vial for analysis. The measured G integral was converted into mL of H2 with calibration performed by injecting sever known volumes of H2.

NP Surface Activation
The surface activation of NPs is a crucial step for successful and stable surface fun tionalization. We previously described the benefits of acid activation of TiO2 NPs for the use in catalysis [20]. However, in the literature, other methods are also described, inclu ing using various acids (e.g., MeCO2H or HF), hydrogen peroxide, or plasma for surfa activation [49][50][51]. These methods are used to enhance the surface reactivity and make th surface more reactive for functionalization. TiO2 and SrTiO3 NPs are acid-resistan whereas BaTiO3 NPs are sensitive to mineral acids. Hence, as described in Section 2.2. during acid activation with aqueous HNO3 (3 M), the BaTiO3 NPs partially dissolved were lost during washing (70%). Since acid activation was unsuitable for the BaTiO3 NP other activation methods were explored, and activation using the H2O2 treatment [49] w chosen. This method simultaneously saturates the surface with hydroxyl groups an strips it of carbonate groups. Although the literature protocol required hydroxyl group for a salinization reaction [49], the hydroxyl groups should also favor condensation wi the phosphonic acid of the anchoring ligand.
Activation was performed with both SrTiO3 and BaTiO3 by boiling the dispersed NP under N2 in 30% H2O2. The particles were washed with water and then dried for 72 under high vacuum. The pristine and activated NPs were characterized using FTIR an TGA-MS. The FTIR differences shown in Figure 1 were limited to the fingerprint regio For the pristine SrTiO3 activated with H2O2, an additional absorption at 1446 cm −1 a peared. The pristine BaTiO3 activated with acid showed the disappearance of a promine peak at 1420 cm −1 , while the activation with H2O2 caused the peak to shift and broaden. The TGA-MS results are reported in Table 1. A comparison of the pristine SrTiO3 an BaTiO3 NPs with acid-activated NPs (SrTiO3 and BaTiO3) shows that there was a slig increase in the weight loss (0.1-0.4%) in the lower temperature region (<380 °C). The NP activated with H2O2 showed a greater weight loss than the pristine NPs, with an increa in the lower temperature region of 1.2% or 2.2% for SrTiO3 and BaTiO3. Using TGA-MS, peak with amu 18 (H2O) was found in each sample and assigned to physisorbed an The TGA-MS results are reported in Table 1. A comparison of the pristine SrTiO 3 and BaTiO 3 NPs with acid-activated NPs (SrTiO 3 and BaTiO 3 ) shows that there was a slight increase in the weight loss (0.1-0.4%) in the lower temperature region (<380 • C). The NPs activated with H 2 O 2 showed a greater weight loss than the pristine NPs, with an increase in the lower temperature region of 1.2% or 2.2% for SrTiO 3 and BaTiO 3 . Using TGA-MS, a peak with amu 18 (H 2 O) was found in each sample and assigned to physisorbed and chemisorbed water. For H 2 O 2 -activated NPs, this increase could also be due to the loss of the hydroxyl groups. The pristine NPs and H 2 O 2 -activated NPs showed organic decomposition products (amu 44, CO 2 ) throughout the TGA experiment, possibly due to impurities in the pristine samples; this was especially visible for BaTiO 3 . In the higher temperature region (380-900 • C), H 2 O 2 -activated SrTiO 3 and BaTiO 3 NPs also showed slightly higher weight losses (0.2%) than the pristine NPs. The acid-activated SrTiO 3 and BaTiO 3 NPs showed slightly higher (0.6%) and lower (0.1%) weight losses, respectively. In both cases, the TGA-MS experiment recorded a significant weight loss at 550 • C attributed to amu 44 (CO 2 ), and, in the case of SrTiO 3 -a, a mass loss at amu 81 was also observed. The TGA-MS experiment on SrTiO 3 -a is illustrated in Figure 2 as an example of the sharp weight loss. The origin of this weight loss is unclear.
Nanomaterials 2022, x, x FOR PEER REVIEW 6 of 16 chemisorbed water. For H2O2-activated NPs, this increase could also be due to the loss of the hydroxyl groups. The pristine NPs and H2O2-activated NPs showed organic decomposition products (amu 44, CO2) throughout the TGA experiment, possibly due to impurities in the pristine samples; this was especially visible for BaTiO3. In the higher temperature region (380-900 °C), H2O2-activated SrTiO3 and BaTiO3 NPs also showed slightly higher weight losses (0.2%) than the pristine NPs. The acid-activated SrTiO3 and BaTiO3 NPs showed slightly higher (0.6%) and lower (0.1%) weight losses, respectively. In both cases, the TGA-MS experiment recorded a significant weight loss at ~550 °C a ributed to amu 44 (CO2), and, in the case of SrTiO3-a, a mass loss at amu 81 was also observed. The TGA-MS experiment on SrTiO3-a is illustrated in Figure 2 as an example of the sharp weight loss. The origin of this weight loss is unclear. Figure 2. TGA-MS of acid-activated SrTiO3 NPs. SrTiO3-a, where black is the weight loss, red is the temperature, brown is the derivative weight against time, blue is the ion current of amu 18, green is the ion current of amu 44, and orange is the ion current of amu 81.

Nanoparticle Surface Ligand Functionalization
The bpy metal-binding domain in ligand 1 (Scheme 1) was chosen for the surface assembly of {M(bpy)3} n+ (M = Ru, n = 2; M = Rh, n = 3) moieties. Phosphonic acids bind strongly to BaTiO3 NP surfaces [47], and it is reasonable to assume a similar behavior with SrTiO3 NPs. The pristine BaTiO3 and SrTiO3 NPs had 50 and 100 nm diameters, respectively, making them larger than the p25 TiO2 NPs used in previous work [20,52,53]. This

Nanoparticle Surface Ligand Functionalization
The bpy metal-binding domain in ligand 1 (Scheme 1) was chosen for the surface assembly of {M(bpy) 3 } n+ (M = Ru, n = 2; M = Rh, n = 3) moieties. Phosphonic acids bind strongly to BaTiO 3 NP surfaces [47], and it is reasonable to assume a similar behavior with SrTiO 3 NPs. The pristine BaTiO 3 and SrTiO 3 NPs had 50 and 100 nm diameters, respectively, making them larger than the p25 TiO 2 NPs used in previous work [20,52,53]. This difference significantly changes the surface area from a volume ratio of 0.28 nm −1 for TiO 2 to 0.12 nm −1 and 0.06 nm −1 for BaTiO 3 and SrTiO 3 , respectively. This reduces the available surface area and necessitates functionalization adjustments to avoid the free ligand's physisorption. Section 2.2.3 details the functionalization methods adopted. The NPs were characterized using 1 H NMR spectroscopy, FTIR, TGA-MS, MALDI-MS, and solid-state absorption spectroscopy. The TGA-MS results are presented in Table 2 and show higher weight losses than the corresponding unfunctionalized NPs in the high-temperature region (380-900 • C), indicating successful surface functionalization. When compared to TiO 2, the weight loss due to functionalization was smaller (0.2-0.7% versus 2.6%) [20] than expected since the SrTiO 3 and BaTiO 3 NPs had considerably larger particle sizes. The NPs all showed carbon-containing impurities before functionalization, as described in Section 3.1.1. Based on the TGA-MS experiments in the low-temperature region, the impurities were lost during the functionalization. Figure 3 shows the TGA-MS measurement of 1@SrTiO 3 -a, where ligand 1 was decomposed in a single event distinct from the decompositions recorded in the starting material. The TGA-MS measurement of 1@SrTiO 3 still shows the peak observed for SrTiO 3 -a at~550 • C with an additional weight loss between 700 • C and 880 • C. The TGA-MS in this region shows amu 18 (H 2 O) and 44 (CO 2 ), corresponding to a decomposition of the anchoring ligand 1. difference significantly changes the surface area from a volume ratio of 0.28 nm −1 for TiO2 to 0.12 nm −1 and 0.06 nm −1 for BaTiO3 and SrTiO3, respectively. This reduces the available surface area and necessitates functionalization adjustments to avoid the free ligand's physisorption. Section 2.2.3 details the functionalization methods adopted. The NPs were characterized using 1 H NMR spectroscopy, FTIR, TGA-MS, MALDI-MS, and solid-state absorption spectroscopy. The TGA-MS results are presented in Table 2 and show higher weight losses than the corresponding unfunctionalized NPs in the high-temperature region (380-900 °C), indicating successful surface functionalization. When compared to TiO2, the weight loss due to functionalization was smaller (0.2-0.7% versus 2.6%) [20] than expected since the SrTiO3 and BaTiO3 NPs had considerably larger particle sizes. The NPs all showed carbon-containing impurities before functionalization, as described in Section 3.1.1. Based on the TGA-MS experiments in the low-temperature region, the impurities were lost during the functionalization. Figure 3 shows the TGA-MS measurement of 1@SrTiO3-a, where ligand 1 was decomposed in a single event distinct from the decompositions recorded in the starting material. The TGA-MS measurement of 1@SrTiO3 still shows the peak observed for SrTiO3-a at ~550 °C with an additional weight loss between 700 °C and 880 °C. The TGA-MS in this region shows amu 18 (H2O) and 44 (CO2), corresponding to a decomposition of the anchoring ligand 1.  and BaTiO 3 , respectively. Compared to the unfunctionalized NPs, 1@SrTiO 3 and 1@BaTiO 3 (independent of prior activation) showed several absorptions in the range 1900 cm −1 to 900 cm −1 , indicating a bound ligand. The possibility of traces of an absorbed and labile species being on the NP surface after the functionalization was excluded based on the 1 H NMR spectroscopy.
The FTIR spectra of the f-NPs (Figure 4) showed a broad, weak absorption aroun 3300 cm −1 , in accordance with the TGA-MS results, supporting the presence of hydrox groups. The NPs also exhibited a strong absorption at 540 cm −1 and 500 cm −1 for SrTiO and BaTiO3, respectively. Compared to the unfunctionalized NPs, 1@SrTiO3 and 1@BaTiO (independent of prior activation) showed several absorptions in the range 1900 cm −1 to 90 cm −1 , indicating a bound ligand. The possibility of traces of an absorbed and labile specie being on the NP surface after the functionalization was excluded based on the 1 H NM spectroscopy. The solid-state absorption spectra used the pristine NPs (SrTiO3 or BaTiO3) as th 100% baseline, and the results are shown in Figure 5. The f-NPs showed a broad wea absorption between 400 and 700 nm. These results are similar to the solid-state absorptio spectra of the ligand 1-functionalized TiO2 NPs [20].

Nanoparticle Surface Complexation
For all complexations, the f-NPs were dispersed in an autoclave in a mixture o H2O/EtOH together with RuCl3 . 3H2O, RhCl3 . 3H2O, and bpy (see Sections 2.2.4 and 2.2 for detailed procedures). Depending on the activation of the NPs and the pH during th complexation, differently colored NPs with different catalytic activities were isolated. Th ruthenium and rhodium metal complex-bearing NPs were tested for their ability to cata lyze dihydrogen production from water under irradiation using simulated sunlight. Th The solid-state absorption spectra used the pristine NPs (SrTiO 3 or BaTiO 3 ) as the 100% baseline, and the results are shown in Figure 5. The f-NPs showed a broad weak absorption between 400 and 700 nm. These results are similar to the solid-state absorption spectra of the ligand 1-functionalized TiO 2 NPs [20].
The FTIR spectra of the f-NPs (Figure 4) showed a broad, weak absorption aroun 3300 cm −1 , in accordance with the TGA-MS results, supporting the presence of hydrox groups. The NPs also exhibited a strong absorption at 540 cm −1 and 500 cm −1 for SrTiO and BaTiO3, respectively. Compared to the unfunctionalized NPs, 1@SrTiO3 and 1@BaTiO (independent of prior activation) showed several absorptions in the range 1900 cm −1 to 9 cm −1 , indicating a bound ligand. The possibility of traces of an absorbed and labile speci being on the NP surface after the functionalization was excluded based on the 1 H NM spectroscopy. The solid-state absorption spectra used the pristine NPs (SrTiO3 or BaTiO3) as th 100% baseline, and the results are shown in Figure 5. The f-NPs showed a broad wea absorption between 400 and 700 nm. These results are similar to the solid-state absorptio spectra of the ligand 1-functionalized TiO2 NPs [20].

Nanoparticle Surface Complexation
For all complexations, the f-NPs were dispersed in an autoclave in a mixture H2O/EtOH together with RuCl3 . 3H2O, RhCl3 . 3H2O, and bpy (see Sections 2.2.4 and 2.2 for detailed procedures). Depending on the activation of the NPs and the pH during th complexation, differently colored NPs with different catalytic activities were isolated. Th ruthenium and rhodium metal complex-bearing NPs were tested for their ability to cat lyze dihydrogen production from water under irradiation using simulated sunlight. Th method was also utilized with only RuCl3 . 3H2O to prepare Ru@SrTiO3 and Ru@BaTiO dihydrogen production from water under irradiation using simulated sunlight. The method was also utilized with only RuCl 3 . 3H 2 O to prepare Ru@SrTiO 3 and Ru@BaTiO 3 . The isolated f-NPs are shown in Figure 6 and were characterized using 1 H NMR spectroscopy, FTIR, TGA-MS, MALDI-MS, and solid-state absorption spectroscopy. omaterials 2022, x, x FOR PEER REVIEW 9 o The isolated f-NPs are shown in Figure 6 and were characterized using 1 H NMR spectr copy, FTIR, TGA-MS, MALDI-MS, and solid-state absorption spectroscopy. TGA-MS revealed an increased weight loss of the complex-bearing NPs when co paring them to ligand-functionalized NPs. The results are shown in Table 3. The grea weight loss for the metal complexed f-NPs compared to the ligand f-NPs in the high-te perature region (380-900 °C) was rather low (0.2-0.5%). This provides evidence for low degree of functionalization compared to the equivalent functionalized TiO2 NPs [2 (rR@) represents anchoring ligand 1-f-NPs bound to the surface of the NPs and complexed w RuCl3 and RhCl3 at a ratio of roughly 1:20; (-a) represents using HNO3 activated NPs during functionalization; (-OH) means H2O2 activated NPs during the functionalization; (-OH-A) refer additionally adjusting the pH to 1.5 during the complexation.
A successful complexation was expected to result in additional decomposition in TGA-MS. Hence, the TGA-MS data of SrTiO3-a ( Figure 2) and 1@SrTiO3-a (Figure 3) rR@SrTiO3-a (Figure 7) were used to identify the decomposition processes. Using the current of CO2, amu 44 (Figure 7, green), the decompositions can be differentiated fr the starting material. The decomposition of rR@SrTiO3-a occurred at a slightly lower te perature (~650 °C) than the recorded decomposition for 1@SrTiO3-a (~800 °C) and high than the impurity recorded with SrTiO3-a (~550 °C). TGA-MS revealed an increased weight loss of the complex-bearing NPs when comparing them to ligand-functionalized NPs. The results are shown in Table 3. The greater weight loss for the metal complexed f-NPs compared to the ligand f-NPs in the high-temperature region (380-900 • C) was rather low (0.2-0.5%). This provides evidence for the low degree of functionalization compared to the equivalent functionalized TiO 2 NPs [20]. (rR@) represents anchoring ligand 1-f-NPs bound to the surface of the NPs and complexed with RuCl 3 and RhCl 3 at a ratio of roughly 1:20; (-a) represents using HNO 3 activated NPs during the functionalization; (-OH) means H 2 O 2 activated NPs during the functionalization; (-OH-A) refers to additionally adjusting the pH to 1.5 during the complexation.
A successful complexation was expected to result in additional decomposition in the TGA-MS. Hence, the TGA-MS data of SrTiO 3 -a ( Figure 2) and 1@SrTiO 3 -a (Figure 3) to rR@SrTiO 3 -a ( Figure 7) were used to identify the decomposition processes. Using the ion current of CO 2 , amu 44 (Figure 7, green), the decompositions can be differentiated from the starting material. The decomposition of rR@SrTiO 3 -a occurred at a slightly lower temperature (~650 • C) than the recorded decomposition for 1@SrTiO 3 -a (~800 • C) and higher than the impurity recorded with SrTiO 3 -a (~550 • C).

Figure 7.
TGA-MS of acid-activated with ligand 1-functionalized and ruthenium-, rhodium-, and bpy-complexed SrTiO3 NPs. rR@SrTiO3-a, where black is the weight loss, red is the temperature, brown is the derivative weight against time, blue is the ion current of amu 18, and green is the ion current of amu 44.
In further experiments with 1@SrTiO3-OH and 1@BaTiO3-OH, the pH for the complexation reaction in the autoclave was adjusted from 6.9 to 1.5 with aqueous H2SO4. The resulting NPs showed large weight loss increases in the high-temperature region of 4.0% (Figure 8, bo om, labeled rR@SrTiO3-OH-A) and 6.1% (Figure 8, top, rR@BaTiO3-OH-A). However, the TGA-MS showed peaks at amu 48 (SO) and 64 (SO2) at 750 °C, suggesting the surface binding of H2SO4 [54]. The presence of amu 44 (CO2, Figure 8) suggests an organic decomposition, indicating that the decomposition of H2SO4 was not the main cause of the weight loss. The main weight loss occurred at ~700 °C, corresponding to the decomposition also observed for rR@SrTiO3-a. The earlier decomposition at ~550 °C and ~800 °C was not observed for rR@SrTiO3-OH-A. For rR@BaTiO3-OH-A, an additional decomposition at ~900 °C was observed. 1 H NMR spectroscopy was used to verify the absence of a non-bound anchoring ligand in all cases.
The solid-state absorption spectra of the complexed NPs are shown in Figure 9. In the case of SrTiO3, the spectrum of the metal complex f-NPs (Figure 9, left) shows a broad absorption between 410 nm and 480 nm, with weaker absorptions between 500 and 700 nm. These generally agree with the absorption spectra for the corresponding TiO2 NPs [20]. However, rR@SrTiO3-OH NPs (Figure 9, left, violet) exhibited more intense absorptions in the regions 540-570, 600-620, and 660-680 nm. As these NPs were grey to black, a panchromatic absorption was to be expected. Metal complex f-NPs with BaTiO3 NPs (Figure 9, right) showed similar panchromatic adsorption. For rR@BaTiO3-OH-A (Figure 9, right, red), absorptions between 600-620 and 660-680 nm were less intense than other NPs, in accordance with their pale orange color. In further experiments with 1@SrTiO 3 -OH and 1@BaTiO 3 -OH, the pH for the complexation reaction in the autoclave was adjusted from 6.9 to 1.5 with aqueous H 2 SO 4 . The resulting NPs showed large weight loss increases in the high-temperature region of 4.0% (Figure 8, bottom, labeled rR@SrTiO 3 -OH-A) and 6.1% (Figure 8, top, rR@BaTiO 3 -OH-A). However, the TGA-MS showed peaks at amu 48 (SO) and 64 (SO 2 ) at 750 • C, suggesting the surface binding of H 2 SO 4 [54]. The presence of amu 44 (CO 2 , Figure 8) suggests an organic decomposition, indicating that the decomposition of H 2 SO 4 was not the main cause of the weight loss. The main weight loss occurred at~700 • C, corresponding to the decomposition also observed for rR@SrTiO 3 -a. The earlier decomposition at~550 • C and~800 • C was not observed for rR@SrTiO 3 -OH-A. For rR@BaTiO 3 -OH-A, an additional decomposition at~900 • C was observed. 1 H NMR spectroscopy was used to verify the absence of a non-bound anchoring ligand in all cases.
The solid-state absorption spectra of the complexed NPs are shown in Figure 9. In the case of SrTiO 3 , the spectrum of the metal complex f-NPs (Figure 9, left) shows a broad absorption between 410 nm and 480 nm, with weaker absorptions between 500 and 700 nm. These generally agree with the absorption spectra for the corresponding TiO 2 NPs [20]. However, rR@SrTiO 3 -OH NPs (Figure 9, left, violet) exhibited more intense absorptions in the regions 540-570, 600-620, and 660-680 nm. As these NPs were grey to black, a panchromatic absorption was to be expected. Metal complex f-NPs with BaTiO 3 NPs (Figure 9, right) showed similar panchromatic adsorption. For rR@BaTiO 3 -OH-A (Figure 9, right, red), absorptions between 600-620 and 660-680 nm were less intense than other NPs, in accordance with their pale orange color. Figure 10 shows the FTIR spectra of rR@SrTiO 3 (left, blue) and rR@BaTiO 3 (right, blue) and their variants (-a, green; -OH, violet; -OH-A, red). Overall, the FTIR spectra of all metal complex f-NPs of SrTiO 3 and BaTiO 3 were similar (except for rR@SrTiO 3 -OH-A and rR@BaTiO 3 -OH-A). In addition, the spectra were very similar to that of the anchoring ligand f-NPs described in Section 3.1.2, with only small differences in the fingerprint region (1900 to 900 cm −1 ). In contrast, the FTIR spectra of rR@SrTiO 3 -OH-A and rR@BaTiO 3 -OH-A showed strong absorptions at 1219, 1141, and 1100 and at 1200 and 1091 cm −1 , respectively, due to the presence of H 2 SO 4 on the NP surface. This was confirmed by recording an FTIR spectrum of a mixture of pristine SrTiO 3 NPs and aqueous H 2 SO 4 (see Supplementary Materials, Figure S56). As expected, attempts to quantify the ruthenium or rhodium content of the functionalized NPs using energy-dispersive X-ray (EDX) spectroscopy were unsuccessful, as the expected values were below the 1% detection limit.

Dihydrogen Generation from Water General Procedure
The experimental details of these studies are given in Section 2.2.6, and the results are collected in Table 4. The f-NPs that produced dihydrogen showed the characteristic red color of a {Ru(bpy) 3 } 2+ chromophore, while f-NPs that were either grey or black either did not produce H 2 or did so with low efficiency. BaTiO 3 NPs and SrTiO 3 NPs activated with H 2 O 2 , functionalized with ligand 1 and then complexed with RuCl 3 . 3H 2 O and RhCl 3 . 3H 2 O were less active for water reduction. This indicates minimal or no formation of the required photocatalyst(s) on the surface. Figure 8. TGA-MS of H2O2 activated with ligand 1-functionalized and ruthenium, rhodium, and bpy under adjusted pH complexed SrTiO3 NPs (top) and BaTiO3 NPs (bo9om). rR@SrTiO3-OH-A and rR@SrTiO3-OH-A, where black is the weight loss, red is the temperature, brown is the derivative weight against time, blue is the ion current of amu 18, green is the ion current of amu 44, yellow is the ion current of amu 48, and pink is the ion current of amu 64.   Figure 10 shows the FTIR spectra of rR@SrTiO3 (left, blue) and rR@BaTiO3 (right, blue) and their variants (-a, green; -OH, violet; -OH-A, red). Overall, the FTIR spectra of all metal complex f-NPs of SrTiO3 and BaTiO3 were similar (except for rR@SrTiO3-OH-A and rR@BaTiO3-OH-A). In addition, the spectra were very similar to that of the anchoring ligand f-NPs described in Section 3.1.2, with only small differences in the fingerprint region (1900 to 900 cm −1 ). In contrast, the FTIR spectra of rR@SrTiO3-OH-A and rR@BaTiO3-OH-A showed strong absorptions at 1219, 1141, and 1100 and at 1200 and 1091 cm −1 , respectively, due to the presence of H2SO4 on the NP surface. This was confirmed by recording an FTIR spectrum of a mixture of pristine SrTiO3 NPs and aqueous H2SO4 (see Supplementary Materials, Figure S56). As expected, a empts to quantify the ruthenium or rhodium content of the functionalized NPs using energy-dispersive X-ray (EDX) spectroscopy were unsuccessful, as the expected values were below the 1% detection limit.

Dihydrogen Generation from Water General Procedure
The experimental details of these studies are given in Section 2.2.6, and the results are collected in Table 4. The f-NPs that produced dihydrogen showed the characteristic red color of a {Ru(bpy)3} 2+ chromophore, while f-NPs that were either grey or black either did not produce H2 or did so with low efficiency. BaTiO3 NPs and SrTiO3 NPs activated with H2O2, functionalized with ligand 1 and then complexed with RuCl3 . 3H2O and RhCl3 . 3H2O were less active for water reduction. This indicates minimal or no formation of the required photocatalyst(s) on the surface.   As discussed in our previous paper [20], surface functionalization is increased by the HNO 3 treatment, leading to more active catalytic sites on the surface and greater hydrogen production.

Conclusions
We explored activation methods for SrTiO 3 and BaTiO 3 NPs using H 2 O 2 and aqueous HNO 3 . Pristine and activated SrTiO 3 and BaTiO 3 NPs were functionalized with anchoring ligand 1 and subsequently elaborated using RuCl 3 . 3H 2 O and bpy, RhCl 3 . 3H 2 O and bpy, or RuCl 3 . 3H 2 O. The potentially photo-and redox-active Rh and Ru surface-bound metal complex f-NPs were used in a photochemical system for the solar generation of H 2 from water. BaTiO 3 NPs and SrTiO 3 NPs activated with H 2 O 2 , functionalized with ligand 1, and then complexed with RuCl 3 . 3H 2 O and RhCl 3 . 3H 2 O yielded f-NPs that were inactive for water reduction. rR@SrTiO 3 -a was the most efficient within our tested materials, giving 1.1 mL H 2 per hour, with rR@SrTiO 3 being roughly two-thirds as efficient. The complexation of the metal species to the H 2 O 2 -activated and ligand-functionalized NPs was modified by adjusting the pH to 1.5 with aqueous H 2 SO 4 . The resulting orange SrTiO 3 and BaTiO 3 NPs were active for water reduction and produced H 2 . rR@SrTiO 3 -OH-A were considerably less active than rR@SrTiO 3 or rR@SrTiO 3 -a, while rR@BaTiO 3 -OH-A performed better than rR@BaTiO 3 -a. Hence, pH seems to be more important during complexation than previously thought, and adjusting it can play a major role. It is unclear if the drop in efficiency for rR@SrTiO 3 -OH-A was due to residual H 2 SO 4 influencing the pH of the suspension during the water reduction or if complexation was impacted. For rR@BaTiO 3 -OH-A, the solid-state absorption spectroscopic data and the red color of the functionalized NPs further suggest the formation of surface-bound {Ru(bpy) 3 } 2+ chromophores.
We conclude that extensively functionalized SrTiO 3 or BaTiO 3 NPs may perform better than TiO 2 NPs for water reduction. However, the former NPs are more expensive than TiO 2 , and cost-benefit and scale-up limitations should be explored. Particle size might also play a significant role in surface loading, and we note that the pristine BaTiO 3 and SrTiO 3 NPs had 2 to 4 times larger radii than the TiO 2 NPs. This work included the characterization of activated f-NPs and metal complex f-NPs using FTIR spectroscopy, solid-state absorption spectroscopy, and TGA-MS, providing evidence for successful functionalization.