Optimized loading of Pt single atoms from platinic acid solution and the sweet spot for activity

Abstract

In recent years, platinum single atoms (SAs) have emerged as exceptionally efficient co-catalysts for photocatalytic hydrogen (H₂) evolution. In the present work, we systematically investigate the solution parameters that govern the deposition of Pt SAs from dilute H₂PtCl₆ precursors onto well-defined sputtered anatase TiO₂ thin films and evaluate the effect on photocatalytic H₂ evolution. We show that both precursor concentration and solution composition critically determine the oxidation state, dispersion, and reactivity of surface-bound Pt species. Ultra-dilute, additive-free solutions (0.001–0.005 mM) enable strong electrostatic adsorption (SEA) and aquation-assisted anchoring of isolated Pt²⁺ atoms, leading to atomically dispersed Pt–O–Ti surface motifs with maximal H₂ evolution per Pt. In contrast, increased ionic strength or unfavorable speciation caused by higher concentration or pH adjustment induces non-selective uptake, formation of Pt(IV)-rich 2D rafts, and diminished catalytic performance. Overall, we show that the placement and photocatalytic activity of Pt SAs are dictated by the pre-reaction deposition chemistry. From a practical view, this study defines a narrow yet actionable parameter window for scalable, efficient SA co-catalyst decoration and establishes the mechanistic basis for performance optimization via precursor-chemistry design.

---

1. Introduction

In recent years, supported single atoms (SAs) have attracted significant attention in various research domains, including heterogeneous catalysis,^1–4 electrocatalysis,^5–8 and most importantly, in photocatalytic processes such as water splitting,⁹ waste degradation,¹⁰ and CO₂ reduction.¹¹ In typical photocatalytic systems, upon illumination with photons exceeding a semiconductor's band gap energy, photogenerated charge carriers (electrons and holes) are generated that can migrate to the surface and drive redox reactions in the surrounding environment, such as the generation of H₂ from water or aqueous solutions. One of the most crucial steps in this process is the efficient transfer of photoexcited electrons from the conduction band of the semiconductor to H⁺ or water molecules.^12,13 However, this electron transfer often faces kinetic limitations and thus requires the assistance of co-catalysts, specifically noble metals such as Pt, Pd, or Rh, which are well established to enhance the H₂ production.^12,14,15 Conventionally, these co-catalysts are attached onto semiconductor surfaces in the form of nanoparticles (NPs), typically a few nanometers in size.^16,17 In efforts to reduce the consumption of costly noble metals while maintaining high catalytic efficiency, much research has focused on optimizing both the size and loading of the co-catalyst.^18–20 The downscaling from nanoparticles to SAs has emerged as a significant advancement in this context, offering a dramatically increased surface-to-volume ratio, providing more active sites, and the potential for enhanced reactivity.²¹ This is evident from several studies on photocatalytic performances of SAs co-catalysts supported on oxide semiconductors, particularly, on Pt SAs over TiO₂ supports,^14,22,23 which is the most extensively exploited photocatalyst, over the past few decades. A most promising approach for Pt SA decoration on TiO₂ is reactive deposition based on reductive anchoring of SAs from dilute hexachloroplatinic (IV) acid (H₂PtCl₆) solution, providing SA co-catalyst loading without any thermal post-deposition treatment.^24,25 Despite a number of studies using this approach for Pt SA deposition and elaborating consequent co-catalysts effects, a comprehensive and systematic investigation of the Pt precursor parameters on the resulting loading and the impact on the photocatalytic activity remains largely open. Therefore, the present work aims at identifying critical factors when using reactive deposition from H₂PtCl₆ solutions to decorate Pt SAs co-catalysts on a well-defined titania substrate – that is, DC sputter-deposited anatase thin films.

In many previous studies, single atoms (SAs) have been deposited onto TiO₂, specifically on various nanostructures such as nanotubes,^26–28 nanoflakes,^29,30 and powders,^23,31 primarily to take advantage of the high surface area, which can enhance overall catalytic performance. In contrast to these approaches, our work utilizes TEM-transparent TiO₂ sputter-deposited substrates. These flat, compact films offer a well-defined geometry that simplifies illumination and characterization processes (XPS, TEM), allowing for precise evaluation of SA loading, spatial distribution, as well as their catalytic activity for photocatalytic H₂ generation under controlled illumination conditions.

These thin film anatase layers are then decorated using different concentrations of H₂PtCl₆ precursor in an aqueous medium (and under different pH-adjustment conditions) to systematically investigate the effect of precursor concentration and deposition conditions on Pt SA co-catalyst loading, and finally, to evaluate the effect of different Pt loading (conditions) on photocatalytic H₂ generation.

The results show that the deposition of Pt SAs from H₂PtCl₆ solutions is strongly dependent on precursor concentration, pH and conductivity of the solution. At low precursor concentrations, Pt is stabilized as isolated SAs, while under specific conditions, partial aggregation into less reactive nm-sized two-dimensional (2D) rafts can occur due to unfavorable solution speciation.

2. Experimental section

2.1 Fabrication of TiO₂ film on FTO substrate

TiO₂ films were fabricated following a procedure reported previously.³² Briefly, TiO₂ thin films were deposited on cleaned FTO substrates by reactive DC magnetron sputtering. Before deposition, the Ti target was pre-conditioned in an argon atmosphere to remove surface oxides. A thin Ti interlayer was first applied to the FTO, followed by TiO₂ deposition under controlled Ar/O₂ flow and pressure conditions, with the substrate holder rotated to ensure uniform growth. Films were deposited at room temperature using 500 W sputtering power, with ∼200 nm thickness adjusted by deposition time, and subsequently annealed in air at 450 °C for 1 h – these films were used for photocatalytic testing and surface analyses. In addition, ultrathin TiO₂ layers (∼20 nm) were prepared using the same DC sputter conditions on TEM grids – these were used for HAADF-STEM and TEM-EDX analyses. These 20 nm films provide electron transparency for high-resolution HAADF-STEM and EDX analyses, allowing direct imaging and compositional mapping of atomically dispersed Pt species. These films were fabricated on a commercial Si/SiO₂ TEM grid with photolithographically defined 50 × 50 µm windows.

2.2 Pt SAs deposition on TiO₂ layers

In a standard SAs deposition process, aqueous solutions of H₂PtCl₆·6H₂O (Metakem, Germany) were prepared at different concentrations (0.0002, 0.001, 0.005, 0.05, 0.01, 1, 2, 5, or 10 mM), using ultrapure deionized water in a quartz cell. These prepared solutions were purged with argon gas for 15 min to remove dissolved oxygen. Subsequently, TiO₂ films deposited on FTO substrates were submerged in the precursor solutions. The reaction cells were properly sealed and covered to avoid any interaction with light, and kept for 1 h. After 1 h, the samples were thoroughly rinsed with deionized water and finally dried under a nitrogen flow.

2.3 Characterization

The surface and cross-section morphologies of Pt-TiO₂ samples were investigated by field-emission scanning electron microscopy (FE-SEM, S-4800, Hitachi). High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and Energy Dispersive X-Ray Spectroscopy (EDS) mapping were performed using a high-resolution transmission electron microscope (HRTEM, FEI Titan G2 60–300). The chemical states and composition of samples were analyzed by X-ray photoelectron spectroscopy (XPS, PHI 5600). The XPS spectral deconvolution was carried out using MultiPak software, and all XPS spectra were aligned by a binding energy of Ti 2p (458.5 eV).

2.4 Photocatalytic H₂ evolution

The photocatalytic H₂ evolution activity of TiO₂ layers loaded with Pt SAs was assessed under ultraviolet (UV) light irradiation. The experiments were conducted using a quartz reactor containing 10 mL of an aqueous solution with 50% methanol as a hole scavenger. Prior to illumination, the system was purged with argon gas for 15 min to remove dissolved oxygen. The reactor was then sealed and exposed to 365 nm LED light with a power density of 65 mW cm⁻² for a predetermined period. Hydrogen production was monitored at set time intervals using a gas chromatograph instrument (GCMS-QO2010SE, SHIMADZU) equipped with a thermal conductivity detector (TCD).

3. Results and discussion

In a first set of experiments, Pt SAs were decorated on the anatase TiO₂ layers using immersion in a 0.005 mM H₂PtCl₆ precursor solution, as described in the Experimental section. High-resolution top-view and cross-sectional SEM images of the TiO₂ surface after Pt deposition are shown in Fig. 1(a) and (b), respectively. As expected, no deposition of particles or agglomeration is visible. These layers are ∼200 nm thick and were later used further for photocatalytic experiments. On the 20 nm thick “sister” sample, to detect Pt SAs on the TiO₂ layers, high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) was utilized. Fig. 1(c) shows an SEM image of the TEM grid, and Fig. 1(d) schematically illustrates the samples used for TEM. Fig. 1(e) presents a HAADF-STEM image of the TiO₂ layer after Pt SA deposition from the 0.005 mM H₂PtCl₆ solution. The image and the corresponding EDX mapping (Fig. S1) distinctly reveal that Pt SAs are well-dispersed and uniformly distributed across a TiO₂ substrate. The lattice spacing of 0.35 nm corresponds to the (101) planes of anatase. The surface density of Pt SAs was evaluated from several STEM images at a level of ∼7.9 × 10⁴ μm⁻². Further, neither from STEM, SEM nor the TEM-EDX, any Pt SA agglomeration is visible on the surface. The XPS Pt 4f spectrum from the surface of the Pt SAs loaded TiO₂ layer from 0.005 mM H₂PtCl₆ solution is presented in Fig. 1(f). The spectrum evidences the presence of Pt with an intense doublet of Pt 4f_7/2 and Pt 4f_5/2 at the respective positions of 72.5 eV and 75.9 eV, which corresponds to the typical speciation of Pt SAs coordinated with an oxide in a (+2) state. Quantitative analysis of the spectrum indicates a total Pt concentration of 0.8 at%, of which 0.65 at% corresponds to surface-anchored Pt²⁺ species. It is important to note that there is a significant difference in the position of these peaks from the metallic Pt⁰ (in nanoparticles), which are usually located for the Pt 4f doublet at ∼70.9 eV and ∼74.1 eV, respectively (an example is given in Fig. S2(a)). Note that in Fig. 1(f), small contributions are also observed at 74.4 eV and 77.8 eV, which correspond to Pt⁴⁺ – likely traces of adsorbed H₂PtCl₆ precursor. Notably, the XPS spectrum (Fig. S3(b)) reveals a negligible amount of Cl.

Fig. 1 (a) Top-view and (b) cross-sectional SEM images of Pt SAs decorated TiO₂ layers from H₂PtCl₆ (0.005 mM) solutions on FTO substrate, (c) SEM image of Pt SAs loaded 20 nm TiO₂ deposited onto the Si/SiO₂ TEM window, (d) schematic drawing of the layer on the TEM grid, (e) HAADF-STEM image, (f) XPS spectra of Pt 4f, and (g) H₂ evolution vs. duration from the Pt SAs decorated TiO₂ layer deposited using 0.005 mM H₂PtCl₆.

These layers, as in Fig. 1(a) and (b), were then tested for photocatalytic H₂ evolution using a 365 nm LED (65 mW cm⁻²) as a light source. The samples were illuminated in 10 mL of an aqueous solution containing 50% methanol as a hole scavenger, and evolving H₂ was detected using gas chromatography (GC), as described in the Experimental section. Fig. 1(g) shows the produced H₂ volume over illumination time. From the linear slope of the data, an H₂ evolution rate of ∼117.9 μL h⁻¹ was determined.

From this starting point, we then screened critical factors in the reactive deposition process and their correlation with Pt deposition amount, speciation and the resulting photocatalytic performance. In a first assessment, we used different concentrations of the H₂PtCl₆ precursor in the range from 0.0002 mM to 10 mM, where the precursor was simply dissolved in DI water and no further solution adjustments (pH, ionic strength) were performed. All samples were then characterized by SEM and XPS. Fig. 2(a) gives the Pt loading on the TiO₂ surface obtained for the different concentrations of H₂PtCl₆. The result shows that initially, the Pt loading increases with increasing H₂PtCl₆ concentration in the precursor solution and reaches a highest loading amount of 5.4 at% from the 0.1 mM Pt precursor solution. However, for higher concentrations of H₂PtCl₆ beyond 0.1 mM, the loading amount of Pt gradually decreases. To investigate the activity of the Pt deposited on the TiO₂ surfaces, the photocatalytic H₂ evolution was evaluated for all samples (Fig. 2(b)), and Fig. 2(c) plots the photocatalytically generated H₂ rate (determined from Fig. 2(b) using the volume of H₂ produced over 3 h normalized with time) for the different concentrations of Pt precursor. In terms of precursor concentration, a maximum Pt loading is achieved for a Pt loading of ∼5.4 at% on the TiO₂ surface from the 0.1 mM H₂PtCl₆ solution. Beyond this point, further increase in the H₂PtCl₆ concentration in the solution leads to lower Pt loading, and the H₂ evolution rate deteriorates. Overall, the activity shows an even maximum distribution (“Bell-type”) curve except for two distinct data points at 0.1 and 0.05 mM. These two samples show, in XPS, SEM, and STEM, unusual findings. SEM images reveal a similar microscopic observation for all samples in the entire concentration range from 0.0002 mM to 10 mM, that is, no detectable deposition of particles or clusters, except for the samples prepared at 0.1 mM and 0.05 mM, respectively, where clear particle deposition is observed (Fig. 2(d) and (e)). This is confirmed in the HAADF-STEM, where, except for Pt SAs, also 2D rafts can be seen (Fig. S4). The HAADF-STEM images clearly show that the majority of Pt is in the SA form with a surface density of ∼1.7 × 10⁵ μm⁻². However, clearly 2D rafts (∼5.7 × 10⁴ μm⁻²) are also observed. For all samples, the XPS spectra (Fig. S3(a)) show peak positions corresponding to Pt²⁺ oxidation states, reflecting the presence of Pt in SA form. However, for the layers Pt deposited in the range from 0.05 mM to 1 mM H₂PtCl₆, a strong doublet peak is also observed at ∼74.8 eV (Pt 4f_7/2) and ∼78.1 eV (Pt 4f_5/2), indicating the presence of significant amounts of Pt⁴⁺. Since there is hardly any Cl⁻ species detectable (Fig. S3(b)) nor any metallic species, and an additional contribution from adsorbed oxygen species is observed in ∼531–532 eV range in the O 1s spectra (Fig. S3(c)), it can be concluded that the H₂PtCl₆ in this case reacted to a fully hydrolysed Pt⁴⁺ species, i.e., a hydroxy- or aquo-species that agglomerated on the TiO₂ surface. Table 1 presents the atomic concentration of Pt²⁺ and Pt⁴⁺ oxidation states for the Pt SAs loaded TiO₂ layers from different concentrations of the precursor solution. From the HAADF-STEM, even the larger assemblies still consist of 2D rafts rather than 3D particles. All in all, this suggests that coalescence of complexes (likely non-charged Pt(IV) species in solution) has taken place for the two critical concentrations. If the XPS Pt loading data are used to plot the photocatalytic H₂ evolution activity with respect to the Pt loading amount (Fig. 2(g)), overall, an increase in the Pt loading amount enhances the H₂ generation activity, but only up to a certain threshold. For Pt surface concentrations up to ∼0.2–0.5 at%, the Pt loading strongly affects the H₂ evolution activity; for higher loadings, the effect saturates. Therefore, the results suggest that an optimal concentration of deposited Pt for achieving maximum Pt activity is ∼0.2–0.5 at%, beyond which the Pt loading does not increase the activity, i.e., leading to Pt material waste. Remarkably, this loading can be achieved already with very dilute precursor solutions (0.001 mM–0.005 mM), indicating that a maximum beneficial effect is achieved in this very low concentration range in terms of deposition efficiency as well as reactivity of the deposited species.

Fig. 2 (a) Pt loading vs. H₂PtCl₆ concentration, (b) H₂ evolution vs. duration using different concentrations of H₂PtCl₆, (c) H₂ evolution rate vs. concentration of H₂PtCl₆ in solution from the Pt SAs decorated TiO₂ layer, SEM images of Pt SAs decorated TiO₂ layers from (d) 0.1 mM and (e) 0.05 mM H₂PtCl₆ solutions, (f) XPS spectra of Pt 4f from the Pt SAs decorated TiO₂ layer deposited using 0.1 mM H₂PtCl₆, and (g) H₂ evolution vs. Pt loading amount from different concentrations of H₂PtCl₆. A few examples of the SA assembly/2D rafts are marked with yellow circles in (d) and (e).

Table 1 Atomic concentration of Pt²⁺ and Pt⁴⁺ for different concentrations of H₂PtCl₆

Concentration of H2PtCl6	Pt^2+ at%	Pt^4+ at%
0.0002 mM	0.06	0.01
0.001 mM	0.13	0.01
0.005 mM	0.65	0.05
0.05 mM	1.54	1.72
0.1 mM	1.63	3.68
1 mM	1.14	1.31
2 mM	0.79	0.31
5 mM	0.37	0.04
10 mM	0.15	0.04

---

To evaluate the origin of the Pt loading distribution as a function of precursor concentration (in Fig. 2(a)) and its compliance with a basic strong electrostatic adsorption (SEA) concept,²⁴ the pH and conductivity of the precursor solutions were determined (Fig. 3(a)). It is evident that, as expected, with increasing concentration of the H₂PtCl₆ precursor in the aqueous medium, the pH of solutions gradually decreases, from ∼5.5 to 2.0, while the conductivity of the solutions increases accordingly from ∼0.01 to 7.5 mS cm⁻¹. Considering the point of zero charge (PZC) for anatase is in the range of 5.5–6.5,³³ all H₂PtCl₆ solutions should be able to provide electrostatically absorbable species as long as their hydrolysis leads to a negatively charged species (e.g., [PtCl₆]²⁻, [PtCl₅(OH)]²⁻, [PtCl₄(OH)₂]²⁻, etc.). In general, the lower the pH, the higher the positive charge on the TiO₂ surface, but the speciation of the hydrolyzed precursor ([PtCl_x(H₂O)_y]ⁿ⁻) may be less favorable for adsorption.³³ It is reported that partial hydrolysis of [PtCl₆]²⁻ increases the possibility of adsorption and subsequent anchoring of Pt species.³³ At lower pH, near ∼2.6, partial hydrolysis of [PtCl₆]²⁻ occurs with complexes containing 2–3 Pt–O bonds and a few Pt–Cl bonds,³⁴ while for higher pH, the hydrolysis increases OH coordination (mainly OH and H₂O become dominant ligands).³⁵ The speciation of Pt to partially hydrolyzed species, i.e., mixed Cl and O coordinates, reaches its peak at pH ∼ 3 to 4, which often makes this range ideal for partial hydrolysis-enabled adsorption and complexation via electrostatic and possible coordination mechanisms.^34,36 This is in line with our results, i.e., the highest Pt loading from the 0.1 mM (pH ∼ 3.5) H₂PtCl₆ solution and the highest photocatalytic H₂ evolution activity of the TiO₂ layer. Such behavior is consistent with SEA, where controlled precursor hydrolysis and charge complementarity guide preferential deposition of metal species on oxide supports. However, in this pH range (∼3.5 to ∼4.5), speciation also leads to neutrally charged complexes. Particularly, the neutral species, [PtCl₄(H₂O)₂]⁰, dominates near pH ∼ 4,³⁶ which reduces the driving force for coulombic adsorption since electrostatic attraction to the positively charged TiO₂ surface is absent. In fact, the odd point in Fig. 2(c) for 0.05 mM corresponds to pH ∼ 4.0 and thus the presence of abundant zero-charged complexes. Due to the lack of repulsive charge on the complexes, under these conditions, these Pt⁴⁺ species are prone to uncontrolled deposition and aggregation rather than well-dispersed anchoring, which compromises the formation of tight bonds and isolated SAs. Below pH ∼ 2, it is more likely to observe a speciation in the fully chlorinated form [PtCl₆]²⁻ as hydrolysis is strongly suppressed. In such cases, the combination of limited hydrolysis and high ionic strength diminishes adsorption efficiency, and the process may even become dominated by ionic shielding effects rather than direct surface–species interaction, which leads to very low Pt loading.³⁶ Importantly, Pt loading exhibits an upper limit under the studied deposition conditions. Although isolated cases show relatively high loadings, the uptake of Pt SAs saturates at approximately 1.2 at% Pt. At the higher precursor concentrations, particularly at pH < 2, adsorption may be controlled by ionic screening effects (of the surface charge), i.e., the dominant factor is the ionic strength of the precursor solution.³⁶ Furthermore, the variation in Pt oxidation states (Table 1) primarily reflects the extent of precursor–surface interaction. The catalytically active Pt²⁺ species forms through surface-mediated reduction during deposition, where partially hydrolyzed Pt(IV) complexes are reduced via charge transfer with Ti³⁺ sites to yield stable Pt–O–Ti linkages, as previously reported.^37,38 Minor Pt⁴⁺ contributions at most concentrations likely arise from residual Pt(IV) complex adsorption. At intermediate precursor concentrations (∼0.05–0.1 mM, pH ∼ 3.5–4.0), neutral hydrolyzed Pt(IV) species (partially agglomerated) can attach non-selectively, causing partial aggregation and increased Pt⁴⁺ content. Their accumulation explains the bell-shaped trend in Pt loading and H₂ evolution, as these aggregated Pt(IV)-rich domains are largely inactive compared to isolated Pt²⁺ sites. Additionally, at higher precursor concentration (low pH), the limited number of high-affinity anchoring sites on the TiO₂ surface may play a limiting role in the saturation behavior of the Pt loading. Once high-affinity sites are occupied, further deposition occurs on less reactive or weakly coordinated regions, leading to a decline in intrinsic activity per Pt atom.

Fig. 3 (a) pH and conductivity vs. concentration of H₂PtCl₆. (b) H₂ evolution vs. duration from the Pt SAs decorated TiO₂ layer from 10 mM H₂PtCl₆ with modified pH using NH₃ and NaOH. (c) H₂ evolution vs. duration from the Pt SAs decorated TiO₂ layer from 0.005 mM H₂PtCl₆ with modified pH using HCl and HClO₄. (d) H₂ evolution normalized to Pt at% comparison chart for Pt SAs loaded TiO₂ layers deposited from various H₂PtCl₆ solutions with adjusted pH.

Based on these findings, in further experiments, we adjusted the pH of the precursor solution to an optimized speciation region that is pH ∼ 3.5. Solutions with pH < 3.5 were adjusted using NaOH and NH₃, respectively, and solutions with pH > 3.5 using HCl and HClO₄, respectively. Fig. 3(b) shows the Pt loading and the resulting photocatalytic H₂ evolution for 10 mM H₂PtCl₆ solution with an adjusted pH at ∼3.5 using NH₃ and NaOH, respectively. Compared with the non-pH corrected sample (0.2 at% Pt), XPS analysis for these (Fig. S5(a)) reveals an increase in the Pt loading onto the TiO₂ layer surface using NH₃ (to 0.9 at%) and NaOH (to 1.3 at%) for the modified H₂PtCl₆ solution, but still maintaining full SA signatures. The corresponding SEM images of the Pt SA decorated TiO₂ layers from the pH-modified (pH ∼ 3.5) 10 mM precursor solutions (NH₃ in Fig. S6(a) and NaOH in Fig. S6(b)) show the absence of any Pt agglomerates on the surfaces. Evidently, the photocatalytic H₂ evolution increased by a factor of ∼6 and ∼5, after adjusting the pH of the 10 mM H₂PtCl₆ solution using NH₃ and NaOH, respectively, demonstrating the strong influence of pH on the adsorption behavior of the Pt species. More importantly, with a very low Pt loading (∼1 at%) in comparison to the 0.1 mM H₂PtCl₆ solution (∼5.4 at%), a maximum photocatalytic H₂ evolution activity, equivalent to that of the 0.1 mM H₂PtCl₆ solution, was obtained from the pH-modified (pH ∼ 3.5) 10 mM precursor solutions. These results confirm that for the adsorption of complexes from H₂PtCl₆ onto TiO₂, an optimum pH is beneficial. Under these conditions, the pH effect dominates the process, strongly influencing the surface–complex interaction and thereby controlling the extent of adsorption. Consequently, with optimized pH adjustment using NH₃ and NaOH, also a very high photocatalytic H₂ evolution activity was achieved (comparable to or even higher than the best non-pH-adjusted case).

In a control experiment, we adjusted the pH of the deposition solution using HCl and HClO₄, respectively, with a 0.005 mM H₂PtCl₆ solution (pH ∼ 5.1) and adjusted the pH value to ∼3.5. Unlike the case with the 10 mM solutions, here the pH adjustment did not enhance the Pt uptake or the photocatalytic activity. Instead, a clear reduction in Pt loading was observed, accompanied by a corresponding decrease in photocatalytic H₂ evolution performance. Fig. 3(c) shows the results of normalized Pt loading and H₂ generation. In this case, a detrimental effect of the pH adjustment is detected: less Pt is deposited, and consequently, less activity is achieved.

From above discussion, the deterioration in the photocatalytic performance can be ascribed to the combined effects of ionic strength and counter-ion chemistry introduced by the use of strong acids such as HCl and HClO₄. Specifically, the high concentration of counter-ions (Cl⁻ or ClO⁴⁻) leads to shielding at the TiO₂/solution interface that reduces the effective interaction between the negatively charged Pt complexes and the TiO₂ surface. Furthermore, the chloride ions play a reactive role in the hydrolysis and speciation pathways of H₂PtCl₆, i.e., the speciation of the Pt complexes in solution and the equilibrium with adsorbed species on the TiO₂ surface.

Please note, the conductivity of the solution also increases from ∼0.01 mS cm⁻¹ to ∼0.36 mS cm⁻¹ (for HCl) and to ∼0.46 mS cm⁻¹ (for HClO₄) upon the introduction of the strong acids, reflecting a stronger coulombic shielding of the surface. In contrast, the use of bases such as NH₃ and NaOH for pH adjustment promoted not only a more favorable speciation, but also did not strongly affect the conductivity, i.e., ∼2.62 mS cm⁻¹ (for NH₃) and ∼2.44 mS cm⁻¹ (for NaOH) relative to the unmodified 10 mM solution (∼7.42 mS cm⁻¹).

These results demonstrate the overlap of factors that dictate the uptake and activity of Pt on a TiO₂ surface exposed to H₂PtCl₆.

The photocatalytic H₂ evolution performances from the Pt SA loaded TiO₂ layers deposited from various H₂PtCl₆ solutions with different concentrations and pH are presented in Fig. 3(d), which demonstrates the strong and interconnected effect of the parameters, such as precursor concentration, solution pH, ionic strength, and counter-ion chemistry of the Pt precursor solution in controlling Pt adsorption, dispersion, and ultimately the photocatalytic activity. Importantly, the plot reveals an efficient photocatalytic activity from the dilute precursor solutions for the range of 0.001 mM–0.005 mM concentrations, suggesting the high activity with minimal utilization of Pt.

This interplay between precursor speciation, surface adsorption, and solution conductivity provides a more complete mechanistic understanding of the observed trends.

In summary, the observed trends reflect a surface electrostatic adsorption (SEA)-driven and aquation-assisted grafting process that defines the pre-reaction state of Pt. In ultra-dilute, additive-free H₂PtCl₆ solutions, anionic hydrolyzed Pt(IV) species (e.g., [PtCl₂(OH)₄]²⁻ and other partially aquated forms) are accordingly selectively attracted to positively charged Ti–OH₂⁺ sites on the sputtered TiO₂ surface. The low ionic strength preserves a long Debye length, restricting uptake to spatially separated, high-affinity or highly reactive surface sites. In particular, both solution-phase hydrolysis and surface-mediated reactions play crucial roles in determining the final activity of the deposited Pt; however, their contributions differ in terms of timing and extent. The solution-phase speciation, controlled by precursor concentration, pH, and ionic strength, primarily dictates the coordination environment of Pt complexes available for adsorption. Within the optimal pH range (∼3.5–4.0), partially hydrolyzed Pt(IV) species predominate and can readily interact with the positively charged TiO₂ surface.^35,36 Once adsorbed, surface-mediated ligand exchange and charge-transfer processes promote partial reduction of Pt(IV) to Pt²⁺, which becomes stabilized through Pt–O–Ti bond formation. Therefore, while surface reactions define the final stabilization of Pt²⁺ on TiO₂, the solution-phase hydrolysis largely governs the precursor-state of the Pt species available for deposition. Notably, XPS confirms that the anchored species are present in a reduced Pt²⁺ state prior to photocatalytic operation, indicating that the speciation/adsorption/reactive deposition sequence determines both the oxidation state and dispersion. An overall optimum is present at a pH of 3.5 solution with an ionic strength as low as possible. Increasing precursor concentration or modifying pH perturbs this pathway: high ionic strength and altered speciation (e.g., the occurrence of zero-charged complexes) reduce site selectivity, leading to aggregation and the formation of large (SEM visible >1 nm) 2D rafts. Consequently, various factors govern the activity per Pt atom. In general, maintaining dilute, excess precursor-free but pH-optimized conditions is key to provide reproducible Pt decoration in order to access the highly efficient SA regime.

4. Conclusions

This study identifies and isolates the solution chemistry conditions required for reproducible formation of high-efficiency Pt single-atom co-catalysts on sputtered TiO₂ thin films. Using a structurally defined and morphologically flat substrate, we show that deposition from highly dilute, additive-free aqueous H₂PtCl₆ selectively yields isolated Pt²⁺ species with maximal photocatalytic H₂ evolution rates per Pt. Systematic variation of precursor concentration and solution composition reveals that even small changes in ionic strength or speciation—such as those induced by pH adjustment—disrupt selective anchoring and promote the formation of Pt(IV)-rich surface rafts or clusters before any catalytic reaction takes place. These aggregated forms lower the intrinsic activity per Pt and cannot be rescued by subsequent reduction or illumination. XPS and STEM data demonstrate that under optimized conditions, the Pt remains atomically dispersed and in the +2 oxidation state prior to catalysis. Taken together, our results demonstrate that performance is controlled not by total Pt content but by the solution-mediated anchoring step, which dictates the local coordination environment and surface dispersion of Pt. These insights establish a clear and actionable protocol—ultra-dilute precursor, unmodified pH, and low ionic strength—for preparing supported Pt single-atom catalysts with high reproducibility and superior catalytic performance.

Author contributions

Dr Sutapa Dey: investigation, writing – original draft, review & editing, validation, visualization. Dr Hyesung Kim: writing – review & editing, investigation, visualization. Dr Xin Zhou: investigation. Prof. Patrik Schmuki: conceptualization, supervision, project administration, funding acquisition, writing – review & editing.

Conflicts of interest

The authors declare that there is no conflict of interest.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Supplementary information (SI) is available. See DOI: https://doi.org/10.1039/d5nr03976b.

Acknowledgements

The authors acknowledge DFG and the Operational Program Research, Development and Education (European Regional Development Fund, project no. CZ.02.1.01/0.0/0.0/15_003/0000416 of the Ministry of Education, Youth and Sports of the Czech Republic) for financial support. The authors would also like to acknowledge the support from the Czech Science Foundation project GA CR-EXPRO (23-08019X). The authors would like to acknowledge the support of the Center for Nanoanalysis and Electron Microscopy (CENEM, Friedrich-Alexander-Universität Erlangen-Nürnberg).

References

1. D. Liu, Q. He, S. Ding and L. Song, Adv. Energy Mater., 2020, 10, 2001482.
2. S. Mitchell and J. Pérez-Ramírez, Nat. Commun., 2020, 11, 4302.
3. X. Li, X. Yang, Y. Huang, T. Zhang and B. Liu, Adv. Mater., 2019, 31, 1902031.
4. H. Zhang, G. Liu, L. Shi and J. Ye, Adv. Energy Mater., 2018, 8, 1–24.
5. J. Su, R. Ge, Y. Dong, F. Hao and L. Chen, J. Mater. Chem. A, 2018, 6, 14025–14042.
6. Q. Zhang and J. Guan, Adv. Funct. Mater., 2020, 30, 2000768.
7. W. Zang, Z. Kou, S. J. Pennycook and J. Wang, Adv. Energy Mater., 2020, 10, 1903181.
8. C. Zhu, S. Fu, Q. Shi, D. Du and Y. Lin, Angew. Chem., Int. Ed., 2017, 56, 13944–13960.
9. Q. Zhang and J. Guan, Sol. RRL, 2020, 4, 2000283.
10. H. Zeng, Y. Chen, J. Xu, J. Li, D. Li and J. Zhang, J. Environ. Chem. Eng., 2023, 11, 110681.
11. Z. Shang, X. Feng, G. Chen, R. Qin and Y. Han, Small, 2023, 19, 2304975.
12. X. Chen, S. Shen, L. Guo and S. S. Mao, Chem. Rev., 2010, 110, 6503–6570.
13. S. Dey and S. C. Roy, J. Alloys Compd., 2021, 881, 160481.
14. S. M. Wu and P. Schmuki, Adv. Mater., 2024, 37, 2414889.
15. B. Liu, Z. Hu, Y. Li, X. Tan, J. Ye and T. Yu, Nat. Commun., 2025, 16, 6014.
16. A. Kumar, P. Choudhary, A. Kumar, P. H. C. Camargo and V. Krishnan, Small, 2022, 18, 2101638.
17. D. Benz, K. M. Felter, J. Köser, J. Thöming, G. Mul, F. C. Grozema, H. T. Hintzen, M. T. Kreutzer and J. R. Van Ommen, J. Phys. Chem. C, 2020, 124, 8269–8278.
18. J. Ma, X. Tan, Q. Zhang, Y. Wang, J. Zhang and L. Wang, ACS Catal., 2021, 11, 3352–3360.
19. I. Vamvasakis, B. Liu and G. S. Armatas, Adv. Funct. Mater., 2016, 26, 8062–8071.
20. C. Dong, C. Lian, S. Hu, Z. Deng, J. Gong, M. Li, H. Liu, M. Xing and J. Zhang, Nat. Commun., 2018, 9, 1252.
21. K. Maiti, S. Maiti, M. T. Curnan, H. J. Kim and J. W. Han, Adv. Energy Mater., 2021, 11, 2101670.
22. S. Qin, J. Will, H. Kim, N. Denisov, S. Carl, E. Spiecker and P. Schmuki, ACS Energy Lett., 2023, 8, 1209–1214.
23. S. Qin, N. Denisov, B. B. Sarma, I. Hwang, D. E. Doronkin, O. Tomanec, S. Kment and P. Schmuki, Adv. Mater. Interfaces, 2022, 9, 2200808.
24. Y. Wang, S. Qin, N. Denisov, H. Kim, Z. Bad'ura, B. B. Sarma and P. Schmuki, Adv. Mater., 2023, 35, 2211814.
25. Y. Wang, N. Denisov, S. Qin, D. S. Gonçalves, H. Kim, B. B. Sarma and P. Schmuki, Adv. Mater., 2024, 36, 2400626.
26. I. Hwang, A. Mazare, J. Will, T. Yokosawa, E. Spiecker and P. Schmuki, Adv. Funct. Mater., 2022, 32, 2207849.
27. Z. Wu, I. Hwang, G. Cha, S. Qin, O. Tomanec, Z. Badura, S. Kment, R. Zboril and P. Schmuki, Small, 2022, 18, 2104892.
28. X. Zhou, I. Hwang, O. Tomanec, D. Fehn, A. Mazare, R. Zboril, K. Meyer and P. Schmuki, Adv. Funct. Mater., 2021, 31, 2102843.
29. G. Cha, A. Mazare, I. Hwang, N. Denisov, J. Will, T. Yokosawa, Z. Badura, G. Zoppellaro, A. B. Tesler, E. Spiecker and P. Schmuki, Electrochim. Acta, 2022, 412, 140129.
30. G. Cha, I. Hwang, S. Hejazi, A. S. Dobrota, I. A. Pašti, B. Osuagwu, H. Kim, J. Will, T. Yokosawa, Z. Badura, Š. Kment, S. Mohajernia, A. Mazare, N. V. Skorodumova, E. Spiecker and P. Schmuki, iScience, 2021, 24, 102938.
31. S. Qin, N. Denisov, J. Will, J. Kolařík, E. Spiecker and P. Schmuki, Sol. RRL, 2022, 6, 2101026.
32. H. Kim, Y. Wang, N. Denisov, Z. Wu, Š. Kment and P. Schmuki, J. Mater. Sci., 2022, 57, 12960–12970.
33. K. Wenderich and G. Mul, Chem. Rev., 2016, 116, 14587–14619.
34. Y. Zhao, L. Wang, A. Kochubei, W. Yang, H. Xu, Y. Luo, A. Baiker, J. Huang, Z. Wang and Y. Jiang, J. Phys. Chem. Lett., 2021, 12, 2536–2546.
35. W. A. Spieker, J. Liu, J. T. Miller, A. J. Kropf and J. R. Regalbuto, Appl. Catal., A, 2002, 232, 219–235.
36. W. A. Spieker, J. Liu, X. Hao, J. T. Miller, A. J. Kropf and J. R. Regalbuto, Appl. Catal., A, 2003, 243, 53–66.
37. S. M. Wu, I. Hwang, B. Osuagwu, J. Will, Z. Wu, B. B. Sarma, F. F. Pu, L. Y. Wang, Z. Badura, G. Zoppellaro, E. Spiecker and P. Schmuki, ACS Catal., 2023, 13, 33–41.
38. S. Qin, N. Denisov, H. Kim and P. Schmuki, Angew. Chem., Int. Ed., 2024, 63, e202316660.

---