Introduction and Quantification of Sulfide Ion Defects in Highly Crystalline CdS for Photocatalysis Applications

To evaluate and diversify, control methods for surface defects in photocatalysts have surged because of their significant effect on carrier dynamics and reactivity. Sulfide photocatalysts, wherein anion defects act as electron traps, are not as extensively researched as oxide photocatalysts, and therefore, the knowledge of defects remains incomplete. Herein, simple treatments such as grinding, alkali immersion, and annealing are used to introduce surface defects into CdS, which is a representative sulfide photocatalyst. Moreover, Pt‐loaded CdS is prepared by reducing Pt ions using electrons trapped in the defects. The amount of defects on the CdS surface is successfully estimated by quantifying the amount of unreacted Pt ions through the absorbance of the reaction solution. The results indicate that the defect density can be increased using any of these employed methods. The surface state varies with the introduction method, leading to significant changes in photocatalytic activity. Grinding induces particle refinement and amorphization, alkali immersion induces oxidation and hydroxylation, and annealing enables the formation of a sulfurized surface. These surface conditions degrade the photocatalytic activity, and therefore, introduction of defects under relatively mild conditions is preferred. Herein, the sample calcined at 473 K shows the highest photocatalytic activity.


Introduction and Quantification of Sulfide Ion Defects in Highly Crystalline CdS for Photocatalysis Applications
Haruki Nagakawa

Introduction
Using photocatalysts in hydrogen production can help realize sustainable energy production; [1,2] however, the practical efficiency of photocatalysis is yet to be realized, and therefore, a more advanced design is required.In the field of photocatalytic materials, band and facet engineering have been investigated for controlling the bulk band structure [3,4] and exposed crystal facets, respectively. [5,6]ome recent studies have focused on crystal defects for controlling surface states. [7,8]ontrolling crystal defects is crucial for developing photocatalytic materials that surpass previous efficiency levels because crystal defects directly contribute to charge separation and electron transport in photocatalytic reactions. [9,10]rystal defects in photocatalysts can be classified as imbalances in anions and [11,12] cations [13,14] and lattice distortions. [15,16]nion defects are prone to occur on the surface of oxide and sulfide photocatalysts. [17,18]Oxygen and sulfur defects in titanium dioxide and cadmium sulfide, respectively, act as electron traps, potentially influencing the carrier dynamics and associated photocatalytic activity. [19]e previously reported an electron trapmediated deposition (ETD) method to utilize electron traps for reducing metal ions and supporting cocatalysts such as Pt or Au. [18,20]Although commonly used photocatalytic deposition methods can reduce metal ions based on light intensity and irradiation time, the ETD method relies on the number of electron traps for metal ion reduction.In this case, further reduction of metal ions ceases when the accumulated trap electrons in the photocatalyst are consumed, thereby enabling a high dispersion of cocatalysts. [18,20]Therefore, it is desirable to introduce and control defects to advance photocatalytic reactions and material synthesis using anion defects.
The total number and energy levels of electron traps can vary significantly based on particle morphology and surface states, even for photocatalysts with the same composition. [21,22]The quantitative evaluation of the electron trap density has been traditionally conducted using photochemical methods, [23] reverse double-beam excitation photoacoustic spectroscopy, [21] and diffuse reflectance spectroelectrochemistry; [24] however, its application to sulfide photocatalysts has not been reported.27] Despite its toxicity, CdS is considered a promising photocatalytic material among sulfide photocatalysts because of its strong visible light response and high reduction power. [28][36][37] Thus, quantifying defect levels DOI: 10.1002/pssa.202400213 To evaluate and diversify, control methods for surface defects in photocatalysts have surged because of their significant effect on carrier dynamics and reactivity.Sulfide photocatalysts, wherein anion defects act as electron traps, are not as extensively researched as oxide photocatalysts, and therefore, the knowledge of defects remains incomplete.Herein, simple treatments such as grinding, alkali immersion, and annealing are used to introduce surface defects into CdS, which is a representative sulfide photocatalyst.Moreover, Pt-loaded CdS is prepared by reducing Pt ions using electrons trapped in the defects.The amount of defects on the CdS surface is successfully estimated by quantifying the amount of unreacted Pt ions through the absorbance of the reaction solution.The results indicate that the defect density can be increased using any of these employed methods.The surface state varies with the introduction method, leading to significant changes in photocatalytic activity.Grinding induces particle refinement and amorphization, alkali immersion induces oxidation and hydroxylation, and annealing enables the formation of a sulfurized surface.These surface conditions degrade the photocatalytic activity, and therefore, introduction of defects under relatively mild conditions is preferred.Herein, the sample calcined at 473 K shows the highest photocatalytic activity.
in sulfide photocatalysts and establishing new defect introduction methods offer new directions for material design.
In this study, we introduce defects into highly crystalline CdS photocatalysts with exposed crystal facets [38] through simple treatments such as pressing, alkali immersion, and annealing.In addition, we employ the ETD method to deposit Pt onto photocatalysts with introduced defects using various methods.Subsequently, we identify a process that can quantify the remaining Pt ions in a solution to estimate the number of sulfur defects present on the photocatalyst surface.Further, we aim to clarify the method of defect introduction, characterize the defect states, and investigate their effect on the photocatalytic activity in hydrogen production reactions by evaluating the surface states.

Results and Discussion
In the ETD method for CdS photocatalysts, a Pt cocatalyst is deposited via the mechanism illustrated in Figure 1a. [18]Excited electrons on CdS are trapped at sulfide ion-deficient levels, which function as a Cd metal.Subsequently, galvanic replacement occurs after the addition of Pt ions (PtCl 6 2À ) in the dark, as described in Equation ( 1), leading to the deposition of Pt on CdS.
Based on the stoichiometry of this reaction, the mass of the supported Pt metal corresponded to half the number of sulfide ion defects.Therefore, it is possible to estimate the number of sulfide ion defects present on the CdS surface by quantifying the number of unreacted Pt ions and calculating the deposited amount of Pt metal.
Figure S1, Supporting Information, shows the absorption spectra of a lactic acid solution with Pt ion concentrations varying from 0 to 463 μM.Consequently, absorption peaks attributed to Pt ions were observed at 260 nm; however, the peak intensity was very high at a concentration of 308 μM during the ETD method.Therefore, calibration curves were created based on the absorbances at 300 and 360 nm (Figure S2, Supporting Information).
The absorbance of the reaction solution after ETD is measured for quantifying unreacted Pt ions in CdS samples prepared using the ETD method (Figure S3, Supporting Information).The calculated defect amounts are shown in Figure 1b and Table S1, Supporting Information.The defect amounts calculated from absorbances at 300 and 360 nm were almost identical.The untreated flux10-CdS exhibited the lowest defect amount, with a density of 41-46 μmol (g-CdS) À1 .In contrast, mortar pressing (CdS press), alkaline immersion treatment in a 1 or 10 M NaOH solution (CdS-1 M, CdS-10 M), and annealing in air at 473 or 673 K (CdS-473 K, CdS-673 K) resulted in a 2.0-2.5-foldincrease in the number of defects.
Several characterizations were performed to clarify changes in CdS induced by various treatments.Figure 2a shows the X-Ray diffraction (XRD) results; flux10-CdS exhibits a wurtzite structure, and no significant changes were observed in the crystal structure of CdS.However, the CdS press indicated a broadening of the peaks (Figure 2b).Further, pressing increases the surface distortion of the semiconductor powder and leads to particle size reduction through fragmentation. [39]The broadening of the peaks in these CdS samples resulted from a decrease in crystallinity.
The diffuse reflectance spectroscopy (DRS) profiles of the prepared CdS samples are shown in Figure 3.In addition, the bandgap for each sample was calculated using the Kubelka-Munk function (Figure S4, Supporting Information).The results show that all samples are almost similar, ranging between 2.37 and 2.39 eV.The bandgap was dependent on the crystal structure of the bulk photocatalyst, suggesting that the bandgap was hardly affected by the defect introduction treatments.However, focusing on the details of the difference in optical properties from DRS, CdS-press and CdS-10M exhibited a tailing of the absorption edge.Furthermore, CdS-673 K exhibited an increase in the baseline.These changes in optical properties are not attributed to the bulk crystal structure, but rather to surface or partial defects, amorphous phases, or bonding with different elements.The tailing in the CdS Press in conjunction with the XRD results is attributed to the distortion of the crystal lattice and presence of an amorphous phase. [40]On the other hand, CdS is partially oxidized or hydroxidized in basic aqueous solutions, [31,33] which may have led to the formation of these components on the surface in this study.The bandgap of cadmium oxide is smaller than that of CdS, and therefore, it appeared as a tail on the longerwavelength side in the DRS profile of CdS-10 M. Another possibility is that changes in the surface structure led to a decrease in crystallinity, resulting in tailing at the absorption edge.In addition, in sulfide photocatalysts, an increase in the baseline due to crystal defects has been previously reported [38] and is similar to the trend shown in Figure 1b, suggesting a relatively high defect density in CdS-673 K.
The surface states of CdS were analyzed using X-Ray photoelectron spectroscopy (XPS).The XPS spectra from the CdS samples were charge corrected by shifting all peaks to the adventitious C 1s spectral component (C─C) binding energy set to 284.6 eV.The Cd 3d spectra of the untreated flux10-CdS, CdS press, and CdS-1M showed no significant differences in peak shapes and positions (Figure 4a).However, in CdS-10 M, a shift in the peak to the lower-energy side was observed when the alkali treatment was performed at a higher concentration.This trend was observed for CdS treated with alkali solutions, and it is believed that signals originating from Cd-O were detected. [31,33]o differences in peak shape were observed in the S 2p spectrum of the alkali-treated sample (Figure 4b), which suggests that no new bonds are formed between the oxygen and sulfur atoms.In other words, partial oxidation, hydroxylation, or both occurred on the surface of CdS.Moreover, for CdS-473 K, which was obtained through annealing, the Cd 3d peak on the high-energy side was broadened, and for CdS-673 K, a significant shift was observed.In the S 2p spectrum, the peak intensity at ≈172 eV attributed to sulfate ions increased with an increase in annealing temperature.These XPS spectral features are consistent with those of CdSO 4 , [41] suggesting partial sulfation on the CdS surface.The changes in the binding states on the surface, indicated by the XPS spectra, approximately correspond to the estimated number of defects (Figure 1b).Additionally, hydroxide and sulfate ions bound to cadmium atoms readily dissociate in acidic or neutral aqueous solutions, leading to the formation of anion defects on the surface of CdS.In particular, considering the S 2p spectrum of CdS-473 K after Pt loading, the sulfate-derived peaks are reduced compared to those before loading (Figure 4b).This reduction is attributed to the formation of an electron trap by the desorption of sulfate ions into solution during the Pt loading process of ETD.Therefore, the defective regions of sulfide ions caused by surface hydroxylation and sulfation function as electron traps.
Scanning electron microscopy (SEM) was conducted to visualize the morphologies of the prepared samples (Figure 5).Flux10-CdS exhibited particles of several micrometers with smooth surfaces and hexagonal (001) facets characteristics of the wurtzite structure were observed (Figure 5a). [18]Grinding with mortar decreased the particle size because of particle crushing and surface roughening (Figure 5b).The reduction in particle size led to an increase in the specific surface area.Further, the decrease in particle size and increase in specific surface area increased the density of sulfide ion defects in the CdS press.Thin crystalline steps appeared on the surface of samples subjected to the alkali treatment, and this could be attributed to oxidized or hydroxidized Cd, as suggested by the XPS results (Figure 4a).Samples subjected to annealing showed no significant differences in surface structure compared to flux10-CdS (Figure 5e,f ).
In the SEM images obtained after Pt deposition via ETD (Figure 5), the precipitation of Pt nanoparticles ranging from several nanometers to several tens of nanometers was observed in all samples.These precipitated nanoparticles were identified as Pt, which was supported by the Pt 4f XPS spectra (Figure 4c) and SEM/energy dispersive X-ray spectroscopy images (Figure S5, Supporting Information).The highest density of the supported Pt particles was observed in CdS-673 K (Figure 5f ), similar to the results of sulfide ion defect quantification obtained from absorbance measurements (Figure 1b and Table S1, Supporting Information).
The effect of the different pretreatment methods on the hydrogen production activity of Pt-loaded photocatalysts was investigated by comparing the hydrogen production rates.Figure 6a shows the time-dependent evolution of hydrogen production, and Figure 6b presents the calculated hydrogen production rates.Among these samples, only CdS-473 K exhibited an activity higher than that of flux10-CdS without defect introduction, whereas the other samples showed a decrease in activity.The Pt loads did not vary significantly among samples with introduced defects (Table S1, Supporting Information, entry 2-6), thereby suggesting that these trends were not only attributed to the Pt loading but also attributed to the intrinsic characteristics of CdS photocatalysts.The CdS press exhibited surface amorphization.In many cases, amorphous structures led to decreased activity because they are unsuitable for electron transport. [40]lkali-treated CdS has been reported to be passivated by CdO x . [31]The XPS results (Figure 4a) indicate that a similar structure may have formed in samples.This passivation enhances the stability of CdS; however, it simultaneously increases the recombination frequency, [31] leading to a decrease in activity.After the hydrogen production reactions in flux10-CdS and CdS were subjected to annealing, similar to that in our previous study, the particles exhibited selective corrosion of the CdS(001) facets (Figure 5a,e,f ). [18]However, alkali-treated samples did not exhibit significant corrosion, even after the hydrogen production reaction, which confirmed an improvement in stability (Figure 5c,d).
Annealed samples exhibited significant differences in activity depending on the annealing temperature.The samples annealed at 673 K exhibited strong sulfation, which was evidenced by the measured optical properties (Figure 3) and XPS results (Figure 4) of these samples.This result implies the possibility of formation of partially oxidized cadmium or cadmium sulfate, which are introduced as sulfur defects on the surface as well as potentially coexist as impurities in highly crystalline CdS acting as recombination centers. [42]In contrast, CdS-473 K showed improved activity compared to flux10-CdS; this is attributed to introducing a more suitable amount of a cocatalyst through the introduction of defects.The oxidation state of CdS-473 K was weaker than that  observed at 673 K.These results suggest that under mild processing conditions, it is possible to introduce defects into CdS without significantly altering its characteristics.
To examine the influence of Pt loading on CdS-473 K, we evaluated both the Pt load and hydrogen evolution activity by varying [PtCl 6 ] 2À added during the ETD process.The amount of Pt cocatalyst deposited was calculated from the unreacted [PtCl 6 ] 2À measured via absorbance at 300 nm (Table S1, Supporting Information).The sample names shown in Figure 7a,b indicate the Pt load calculated from the absorbance at 300 nm.A sample with no Pt ions added (i.e., no Pt loading) was labeled CdS-473 K w/o Pt.The maximum Pt loading in the ETD process depends on the amount of sulfide ion defects.For the ETD process in an excess 308 μM [PtCl 6 ] 2À solution, the maximum supported amounts for flux10-CdS and CdS-473 K were ≈0.41 and 0.84 wt%, respectively (Table S1, Supporting Information).This result indicates that CdS-473 K expands the applicability range of Pt loading through the introduction of sulfide ion defects.Furthermore, varying [PtCl 6 ] 2À resulted in corresponding changes in the deposited Pt cocatalyst amount (Table S1, Supporting Information).A comparison of the hydrogen evolution activity of these photocatalysts revealed that CdS-473 K with the highest Pt load (Pt 0.84 wt%) exhibited the highest activity, which confirmed its suitability at this load (Figure 7a,b).Conversely, using a 127 μM [PtCl 6 ] 2À solution led to the formation of photocatalysts with Pt loads similar to that of flux10-CdS.Because the activity for hydrogen evolution was comparable, indicating similar crystalline states for flux10-CdS and CdS-473 K, the difference in activity was attributed to the Pt load.
The results obtained in this study demonstrate that simple treatments such as grinding, alkali immersion, and annealing of sulfide photocatalysts can increase the density of sulfide ion   defects (i.e., electron traps) on the surface.However, the surface state varies depending on the defect introduction method, significantly affecting photocatalytic activity.Annealing at a relatively low temperature (473 K) introduced defects into CdS photocatalysts without causing significant alterations in their properties.This defect introduction method can help advance our understanding of carrier dynamics in sulfide photocatalysts and improve the efficiency of charge separation through defect engineering.Further, the application range of the ETD method, in which the deposition of Pt cocatalysts depends on the number of electron traps, is expanded.Estimating defect amounts based on the ETD method provides a robust measurement approach in the current field of photocatalysis, where the defects are the focal point.This strategy is different from the existing techniques and is applicable to a diverse range of photocatalysts.

Conclusion
We developed a simple method that introduces sulfur defects into highly crystalline wurtzite CdS via pressing, alkali immersion, and annealing.We successfully applied the ETD method, which utilized electrons trapped in CdS defects, to support Pt and quantified the unreduced Pt ions, thereby estimating the amount of defects in CdS.This process is useful for estimating the defect density of sulfide-based photocatalysts, which remained challenging so far.Upon subjecting the CdS to various treatments, the density of the sulfide ion defects increased by 2-2.5 times compared to that with the untreated CdS.The introduction of defects allowed to expand the scope of reactions utilizing the defects, such as the ETD.However, the surface state of CdS varied depending on the defect-introduction method.CdS subjected to pressing showed particle size reduction and amorphization, and CdS treated with alkali immersion, which induced passivation by oxides, exhibited decreased activity.Further, CdS oxidized excessively at 673 K exhibited a significantly decreased activity.In other words, apart from the form of the processing material, careful selection of defect introduction methods is essential for realizing highly efficient photocatalytic reactions.This study demonstrates that by annealing under relatively mild conditions, defects can be introduced into CdS without causing its degradation.

Experimental Section
Materials: Highly crystalline wurtzite CdS was synthesized via the flux treatment of zincblende CdS. [38]Zincblende CdS particles were prepared using the following precipitation method.Cd(NO 3 ) 2 •4H 2 O (10 mmol) was dissolved in 100 mL of ethanol.Further, Na 2 S•9H 2 O (10 mmol) was dissolved in 20 mL of water and the solution was added dropwise to the Cd(NO 3 ) 2 solution and stirred for 30 min.The obtained particles were collected by centrifugation and washed several times with water and ethanol.The resulting product was dried at 60 °C for 1 day and ground using a mortar and pestle.The prepared zincblende CdS (1.0 g) was mixed with CaCl 2 (7.4 g) and NaCl (2.6 g) in sample tubes.The mixture was transferred to an alumina crucible with a lid, and the temperature was increased to 873 K over 1 h and held for 10 h.The mixed salt was allowed to cool to room temperature and dissolved in water, and the precipitated CdS particles were collected by centrifugation.The precipitate was washed with water and ethanol at least five times and centrifuged again.The obtained CdS was labeled flux10-CdS.
Defects were introduced into flux10-CdS by mortar pressing, alkaline immersion treatment in a NaOH solution, and annealing in air.Mortar pressing was performed by strongly grinding 0.5 g of flux10-CdS for 30 min.The resulting samples were labeled using a CdS press.The alkaline immersion treatment was performed by dispersing 1 g of flux10-CdS in 1 or 10 M NaOH solution (20 mL) and stirring for 2 h.The powder was collected by centrifugation, washed several times with pure water, and dried at 333 K for 1 d.These compounds were named CdS-1 M and CdS-10 M. Annealing was performed by placing 1 g of flux10-CdS in an alumina crucible without a lid.The temperature was increased to 473 or 673 K over 30 min and held for 2 h.These compounds were named CdS-473 K and CdS-673 K.
Electron Trap Deposition: Pt cocatalysts were deposited on CdS particles using the ETD method.In the ETD method, 30 mg of CdS particles suspended in 5 mL of 20 vol% lactic acid solution was irradiated with light for 15 min.Electrons excited from the CdS valence band to the conduction band were used to reduce Cd(II) ions at the sulfur defects to Cd metal. [18] 2 PtCl 6 aqueous solution was added to the dispersion (final concentration of H 2 PtCl 6 was 308 μM), and it was then stirred under N 2 atmosphere in the dark for 15 min for the galvanic replacement of Cd with Pt.
The number of Pt ions reduced by ETD was calculated from the absorbance of the postreaction solution.The filtered reaction solution was added to a quartz cuvette, and the absorbance was measured.A calibration curve was prepared using a 20 vol% aqueous lactic acid solution as a baseline and varying the H 2 PtCl 6 concentration up to 463 μM.
Photocatalytic Reaction: The hydrogen production experiments were conducted in test tubes sealed with rubber caps.The synthesized photocatalyst (30 mg) and 20 vol% lactic acid solution (5 mL) were placed in the reaction cell.Light emitting diode lamp (430 nm; Asahi Spectra) was used as the light source, and the evolved gas was collected using a gas-tight syringe.The sampled gas was injected into a gas chromatograph (GC-2014AT, Shimadzu) for determining the amount of evolved hydrogen.The hydrogen production rate was calculated as the amount of hydrogen produced in 60 min.

1 Figure 1 .
Figure 1.a) Schematic of Pt deposition processes conducted via the ETD method.b) Density of sulfide ion defects in the prepared CdS calculated from absorbance at 300 nm (filled bars) or 360 nm (shaded bars).

Figure 4 .
Figure 4. XPS profiles of the a) Cd 3d, b) S 2p, and c) Pt 4f regions of the prepared CdS photocatalysts.

Figure 6 .
Figure 6.a) Time courses of the amount of photocatalytically evolved hydrogen for the prepared CdS photocatalysts (30 mg) suspended in an Ar-saturated aqueous solution (5 mL) containing 20 vol% lactic acid.b) Hydrogen production rates calculated from the slope of the plots in (a).

Figure 7 .
Figure 7. a) Time courses of the amount of photocatalytically evolved hydrogen for the prepared CdS photocatalysts with different Pt loads.b) Hydrogen production rates calculated from the slope of the plots in (a).