Effects of Mono- and Bifunctional Surface Ligands of Cu–In–Se Quantum Dots on Photoelectrochemical Hydrogen Production

Semiconductor nanocrystal quantum dots (QDs) are promising materials for solar energy conversion because of their bandgap tunability, high absorption coefficient, and improved hot-carrier generation. CuInSe2 (CISe)-based QDs have attracted attention because of their low toxicity and wide light-absorption range, spanning visible to near-infrared light. In this work, we study the effects of the surface ligands of colloidal CISe QDs on the photoelectrochemical characteristics of QD-photoanodes. Colloidal CISe QDs with mono- and bifunctional surface ligands are prepared and used in the fabrication of type-II heterojunction photoanodes by adsorbing QDs on mesoporous TiO2. QDs with monofunctional ligands are directly attached on TiO2 through partial ligand detachment, which is beneficial for electron transfer between QDs and TiO2. In contrast, bifunctional ligands bridge QDs and TiO2, increasing the amount of QD adsorption. Finally, photoanodes fabricated with oleylamine-passivated QDs show a current density of ~8.2 mA/cm2, while those fabricated with mercaptopropionic-acid-passivated QDs demonstrate a current density of ~6.7 mA/cm2 (at 0.6 VRHE under one sun illumination). Our study provides important information for the preparation of QD photoelectrodes for efficient photoelectrochemical hydrogen generation.


Introduction
Solar energy is a promising sustainable energy resource owing to its infinite supply and low environmental impact. Specifically, the Sun continuously delivers an enormous energy of 1.7 × 10 5 TW to Earth, which is several orders of magnitude larger than that produced by human civilization. It is highly desirable to develop efficient methods to convert photons into electricity, chemicals, and heat, and this has inspired tremendous interest in research on solar cells [1,2], photocatalysts [3][4][5][6], and photoelectrochemical (PEC) devices [7][8][9][10][11][12][13]. Among these various techniques, PEC hydrogen production provides sustainable and cost-effective methods for direct solar-to-chemical energy conversion to produce clean solar fuels. Previous studies on PEC hydrogen production typically used metal oxide materials such as TiO 2 , BiVO 4 , Fe 2 O 3 , and WO 3 because of their low cost and high stability during water splitting. However, their wide bandgaps (e.g., 3.2 eV for TiO 2 :) inhibit the effective utilization of the full solar spectrum.
Colloidal QDs are generally synthesized in colloidal solutions [40] and are composed of inorganic crystalline nanoparticles and organic surfactants that passivate the surface of the nanoparticles. These surface ligands have multiple functions, including controlling the synthesis process, stabilizing the QDs, regulating the solution dispersibility, and controlling the optical and electrical properties of the QDs [41][42][43]. Because of their significant impact on the properties of QDs, surface ligands are carefully controlled to fully exploit the unique properties of QDs. Surface ligands should be selected by considering the role of the QDs and the fabrication process for the target applications. For example, to enhance charge transport between QDs, the use of short-chain ligands is generally preferred for QD solar cells [44,45]. However, despite their importance, the effect of the surface ligands of QDs on their PEC applications has been less studied.
In this study, we investigated the effects of the surface ligands of colloidal CISe QDs on the fabrication of PEC photoanodes and the resulting PEC characteristics. CISe QDs passivated with monofunctional oleylamine (OAm) were synthesized by colloidal synthesis. Through a post-ligand exchange process, QDs passivated with mercaptopropionic acid (MPA, bifunctional surface ligands) were prepared for comparison. Photoanodes for PEC hydrogen production were prepared by adsorbing QDs on a mesoporous TiO 2 film, which had two different QD adsorption mechanisms according to the choice of surface ligands. Monofunctional ligand-passivated QDs were directly attached on TiO 2 by partial ligand detachment, enhancing electron transport between the QDs and TiO 2 . Bifunctional ligands acted as linkers by bridging QDs and TiO 2 , and the amount of QD adsorption was higher for MPA-passivated QDs than for OAm-passivated ones. With this trade-off, photoanodes fabricated with OAm-passivated QDs and those with MPA-passivated QDs demonstrated photocurrent densities of~8.2 and~6.7 mA/cm 2 , respectively (at 0.6 V RHE , one sun illumination). We believe that our results will contribute to the development of systems with effective PEC hydrogen generation using colloidal QDs.

Synthesis of CISe QDs
Uniform-sized CISe QDs were synthesized according to our previously reported method [31,33]. In a typical synthesis, a metal-OAm complex precursor solution was prepared by heating 0.5 mmol of CuI and 0.5 mmol of InI 3 in 15.0 mL of OAm at 120 • C under a vacuum for 30 min. An oleylammonium selenocarbamate precursor solution was prepared by heating 5.0 mmol of Se in OAm (10.0 mL) under a CO atmosphere at 80 • C. The metal-OAm complex solution was cooled to 70 • C, and 2.0 mL of the oleylammonium selenocarbamate solution was quickly injected into the solution with Ar flow. The reaction Materials 2022, 15, 6010 3 of 13 temperature was gradually increased to 180 • C and was maintained for 20 min. After the reaction, the QDs were precipitated via centrifugation using ethanol containing TOP to remove the remaining Se precursors. Finally, the QDs were dispersed in 4.0 mL of dichloromethane for further use.

Ligand Exchange Treatment of CISe QDs
For the ligand exchange of QDs from OAm to MPA, the phase-transfer ligand exchange process was used [46]. In a typical process, 1.3 mL of MPA was mixed with 4.0 mL of methanol and 40 wt% NaOH aqueous solution of the controlled amount. The total volume of the mixture was adjusted to 8.0 mL. The pH of the solution was controlled by adjusting the amount of aqueous NaOH solution. The QD solution was mixed with the MPA solution and stirred for 10 min. To remove the detached OAm, the mixture was washed several times with chloroform. Finally, the MPA-passivated QDs were dispersed in water.
For the ligand exchange of QDs from OAm to other monofunctional ligands such as OcAm, OA, and DDT, the single-phase ligand exchange process was used. In a typical process to prepare OcAm-passivated CISe QDs, 2.0 mL of OcAm was mixed with 1.0 mL of QD solution (40 mg/mL in dichloromethane). The mixture was vigorously stirred for~2 h at 25 • C. The products were precipitated via centrifugation using ethanol and re-dispersed in dichloromethane for further use. Instead of OcAm, DDT and OAc were used, respectively, for the preparation of DDT-and OAc-passivated CISe QDs.

Fabrication of CISe QD-Sensitized TiO 2 Photoanodes
Fluorine-doped tin oxide (FTO) glass (TEC-A7, Pilkington) was washed in ethanol under ultrasonication for 20 min, followed by treatment with UV/O 3 (Yuil Ultraviolet System) for 15 min to remove any contaminants. Titanium diisopropoxide-bis(acetylacetonate) (7.5 wt%, Aldrich) in n-butanol was spin-coated on the surface of the washed FTO glass and subsequently annealed at 475 • C for 10 min in air. A nanocrystalline TiO 2 paste (Ti-Nanoxide T/SP, Solaronix) was coated on the pretreated FTO glass using the doctor-blade method, followed by annealing at 525 • C for 30 min in air. Finally, the annealed FTO/mesoporous TiO 2 film was immersed in a colloidal CISe QD solution (4.0 mg/mL) for 3 h for sensitization and then rinsed with dichloromethane. The ZnS overlayers were coated on the surface of the QD-sensitized TiO 2 film by successive ionic layer adsorption and reaction (SILAR) processes, consisting of immersing the QD-sensitized TiO 2 film in a 0.05 M Zn(NO 3 ) 2 ·6H 2 O ethanol solution and 0.05 M Na 2 S in a mixed solvent of deionized water/methanol (volume ratio = 1:1) for 1 min each. The SILAR process was repeated thrice.

Material Characterization
The absorption spectra of the CISe QDs were measured using a PerkinElmer Lambda 465 instrument. Time-resolved photoluminescence measurements were performed with a Horiba Fluoromax plus a time-correlated single-photon counting system using a DeltaDiode DD-375L laser diode (peak wavelength: 371 nm). The X-ray diffraction (XRD) patterns were acquired on a Horiba Miniflex 600 X-ray diffractometer. Transmission electron microscopy (TEM) images were obtained using an FEI Tecnai G2 F20 Twin TMP microscope. Fouriertransform infrared (FT-IR) spectroscopy analysis was performed using an Agilent Cary 660 FT-IR spectrometer in attenuated total reflectance measurement mode. X-ray photoelectron spectroscopy (XPS) was performed using a Thermo Scientific ESCALAB 250XI instrument.

Photoelectrochemical Measurements
All of the PEC measurements were carried out in a quartz reactor using a potentiostat (Multi Autolab M204, Metrohm) with a three-electrode system consisting of a QD-sensitized TiO 2 film as the photoanode, a platinum mesh as the counter electrode, and a Hg/HgO (saturated calomel electrode, SCE) as the reference electrode. The electrolyte was composed of 0.25 M Na 2 S and 0.35 M Na 2 SO 3 (pH~12.9) in deionized water. Linear sweep voltammetry (LSV) measurements were performed at a scan rate of 20 mV/s under simulated light with one sun intensity (100 mW/cm 2 ) using a solar simulator (PEC-L01, Peccell) with an AM 1.5G filter. Incident photon-to-current conversion efficiency (IPCE) spectra were obtained using a xenon lamp (300 W, Oriel), monochromator (TracQBasic 6.5, Oriel), and NIST-certified Si diode. Electrochemical impedance spectroscopy (EIS) measurements were performed using a frequency-response detector in the potentiostat under a sinusoidal perturbation of ±10 mV in the frequency range of 0.1 Hz to 100 kHz.

Preparation of CISe QDs with Different Surface Ligands
To understand the effect of the surface ligand molecules of QDs on PEC hydrogen production, the CISe QDs were prepared by colloidal synthesis following our previous study [31,33] using the reaction between metal-ammine complexes and oleylammonium selenocarbamate (Materials and Method 2.2). As shown in the TEM image (Figure 1a), the synthesized QDs had an average size of~4 nm with a narrow size distribution (standard deviation: 0.5 nm). From our previous work on solar cells using CISe QDs,~4 nm was the optimum size for light absorption and electron transfer to TiO 2 , which was the main reason that 4-nm CISe QDs were used for this study. The XRD results confirm the tetragonal chalcopyrite crystal structure of the QDs ( A post-ligand-exchange treatment was carried out to obtain CISe QDs with controlled surface ligand molecules. The as-synthesized CISe QDs were solely passivated by OAm because only OAm was used as a coordinating solvent for the synthesis without adding other organic surfactants. To obtain QDs passivated by bifunctional ligands, a two-phase ligand-exchange reaction was used to replace OAm with MPA, which contains thiol and carboxyl groups (Figure 2a). The thiol group can strongly bind to the surface of QDs, and the additional carboxyl group can make QDs dispersible in polar solvents [46,[49][50][51]. As shown in Figure 2a, the ligand-exchanged QDs were dispersible in water, supporting the successful replacement of surface ligands. The absorption spectra of the These QDs had a composition of CuIn 1.5 Se 3 , as revealed by inductively coupled plasma-atomic emission spectrometry (ICP-AES) analysis. It is known that smaller CISe QDs usually have higher In/Cu ratios because the surfaces of these nanocrystals preferably have In-rich states [30,31]. The electronic state of the CISe QDs was further investigated by XPS. The main peaks of the Cu 2p 3/2 and Cu 2p 1/2 XPS spectra of these QDs are located at 932.3 and 952.1 eV, respectively (Figure 1d), corresponding to the Cu + oxidation state [47]. The estimated binding energies of In 3d 5/2 and In 3d 3/2 are 445.0 and 452.6 eV, respectively (Figure 1e), implying the In 3+ oxidation state [48]. In addition, the XPS spectrum of Se shows that the main peaks of Se 3d 5/2 and Se 3d 3/2 were located at binding energies of 54.1 and 55.0 eV, respectively, which match well with those of Se 2anions, while it does not have peaks corresponding to SeO x that can be formed by surface oxidation. These data support the successful synthesis of high-quality chalcopyrite-structured CISe QDs without severe surface oxidation.
A post-ligand-exchange treatment was carried out to obtain CISe QDs with controlled surface ligand molecules. The as-synthesized CISe QDs were solely passivated by OAm because only OAm was used as a coordinating solvent for the synthesis without adding other organic surfactants. To obtain QDs passivated by bifunctional ligands, a two-phase ligand-exchange reaction was used to replace OAm with MPA, which contains thiol and carboxyl groups (Figure 2a). The thiol group can strongly bind to the surface of QDs, and the additional carboxyl group can make QDs dispersible in polar solvents [46,[49][50][51]. As shown in Figure 2a

Properties of TiO2-CISe QD Photoanodes
To fabricate TiO2-CISe QD photoanodes, a mesoporous TiO2 film was dipped into a solution containing CISe QDs with controlled surface ligands (Materials and Methods 2.4). For both MPA-and OAm-passivated QDs, a dipping time of ~3 h in the QD solution resulted in the dense adsorption of QDs onto TiO2. A comparison between the absorption The successful preparation of CISe QDs with controlled ligand molecules was further verified by XPS. The binding energy of the C 1s XPS main peak of OAm-passivated QDs is 285 eV (Figure 2c), originating from the hydrocarbon chain [52]. In contrast, the C 1s XPS spectrum of the MPA-passivated QDs has an additional peak located at 288.4 eV, which is attributed to the presence of the carboxylic group in MPA. In addition, the S 2p 3/2 XPS spectrum of the MPA-passivated QDs has a clear main peak corresponding to the thiol group, whereas that of the OAm-passivated QDs does not have a peak at the corresponding binding energy (Figure 2d) [53]. These results support the successful preparation of OAmand MPA-passivated QDs for further studies. To gain a better understanding of the effects of surface ligands, we also prepared various monofunctional ligand-passivated QDs, including OcAm-, DDT-, and OAc-passivated QDs ( Figure S4, Supplementary Materials).

Properties of TiO 2 -CISe QD Photoanodes
To fabricate TiO 2 -CISe QD photoanodes, a mesoporous TiO 2 film was dipped into a solution containing CISe QDs with controlled surface ligands (Materials and Methods 2.4). For both MPA-and OAm-passivated QDs, a dipping time of~3 h in the QD solution resulted in the dense adsorption of QDs onto TiO 2 . A comparison between the absorption spectra of bare and CISe QD-sensitized TiO 2 films suggests a significant enhancement of absorbance after dipping (Figure 3a), implying the successful adsorption of CISe QDs onto the TiO 2 films. Cross-sectional scanning electron microscopy (SEM) images and elemental analysis results also support the sensitization of TiO 2 by CISe QDs (Figures S5 and S6, Supplementary Materials). The absorbance in the ultraviolet range was higher after QDsensitization, and the absorption wavelength was extended to the near-infrared range, suggesting the absorption of the full solar spectrum. Photographs of the TiO 2 -CISe QD photoanodes showed their deep brown and black colors, further supporting the strong adsorption of visible light by the QDs (Figure 3b). OAm-and MPA-passivated QDs were similar [55]. This enables a simple comparison of their PEC characteristics by mainly focusing on the adsorption mechanism of QDs to TiO2. We suggest two different adsorption mechanisms for the QDs with different surface ligands (Figure 3c). For QDs passivated with monofunctional ligands (e.g., OAm-passivated QDs), the functional groups of the ligands bind to the QD surface. For these QDs, some surface ligands should be detached from the QD surface before the adsorption of QDs onto the TiO2. The QDs in the QD sensitized-TiO2 films were not washed away by the original QD solvents (e.g., dichloromethane for the OAm-passivated QDs), implying strong binding between the QDs and TiO2 ( Figure S8, Supplementary Materials). If the surface of the QDs was fully covered by monofunctional ligands, the QDs could not be tightly bound to the TiO2, and they were easily removed by non-polar organic solvents. Indeed, the suggested adsorption mechanism is consistent with the results of a previous study of the adsorption mechanism of monofunctional ligand-passivated CdSe QDs on TiO2 for QD-sensitized solar cells [56]. The fact that QDs with strongly binding monofunctional ligands could not be efficiently adsorbed on TiO2 is also consistent with the proposed adsorption mechanism of monofunctional ligand-passivated QDs.
For QDs passivated with bifunctional ligands (i.e., MPA-passivated QDs in this study), one group of ligands (thiol group in this study) strongly binds to the QD surfaces, and the other group (carboxyl group in this study) can bind to the TiO2 surfaces [46,[49][50][51]. Thus, MPA acts as a linker that bridges the QDs and TiO2. Owing to this adsorption mechanism, the amount of QDs adsorbed on TiO2 can be controlled by controlling the pH of the solution. The pH of the QD solution affects the state of the carboxyl groups. When the solution pH decreases, the proportion of ionized carboxyl groups increases. This makes MPA-passivated QDs more dispersible in polar solvents but prevents their binding to TiO2 surfaces. As a result, MPA-passivated QDs were more densely adsorbed on the TiO2 films at higher pH (Figure 3b, bottom). In a pH 14 solution, the TiO2-CISe QD photoanodes showed the highest adsorption density of QDs, as confirmed by their color in photographs. With this optimization, TiO2-CISe QD photoanodes with MPA-passivated QDs had higher QD adsorption densities than those with OAm-passivated QDs, which was verified by both the absorption spectra ( Figure 3a) and photographs (Figure 3b). We additionally tested various monofunctional ligand-passivated QDs, including OcAm-, OAc-, and DDT-passivated QDs ( Figure S4, Supplementary Materials). However, for the reasons listed below, these QDs were not suitable for making PEC electrodes. The colloidal stability of the QDs was decreased by surface passivation with short-chain monofunctional ligands such as OcAm (C8) ( Figure S4, Supplementary Materials). Furthermore, thiol or carboxylic groups adhere to the QD surface too strongly compared to the amine groups [54]. These factors prevented the effective sensitization of TiO 2 films with these QDs (Figure 3a,b). Thus, OAm-and MPA-passivated QDs were mainly studied in this work. In addition, as shown by the cyclic voltammetry of OAm-QD-photoanodes and MPA-QD-photoanodes ( Figure S7, Supplementary Materials), the energy levels of the OAmand MPA-passivated QDs were similar [55]. This enables a simple comparison of their PEC characteristics by mainly focusing on the adsorption mechanism of QDs to TiO 2 .
We suggest two different adsorption mechanisms for the QDs with different surface ligands (Figure 3c). For QDs passivated with monofunctional ligands (e.g., OAm-passivated QDs), the functional groups of the ligands bind to the QD surface. For these QDs, some surface ligands should be detached from the QD surface before the adsorption of QDs onto the TiO 2 . The QDs in the QD sensitized-TiO 2 films were not washed away by the original QD solvents (e.g., dichloromethane for the OAm-passivated QDs), implying strong binding between the QDs and TiO 2 ( Figure S8, Supplementary Materials). If the surface of the QDs was fully covered by monofunctional ligands, the QDs could not be tightly bound to the TiO 2 , and they were easily removed by non-polar organic solvents. Indeed, the suggested adsorption mechanism is consistent with the results of a previous study of the adsorption mechanism of monofunctional ligand-passivated CdSe QDs on TiO 2 for QD-sensitized solar cells [56]. The fact that QDs with strongly binding monofunctional ligands could not be efficiently adsorbed on TiO 2 is also consistent with the proposed adsorption mechanism of monofunctional ligand-passivated QDs.
For QDs passivated with bifunctional ligands (i.e., MPA-passivated QDs in this study), one group of ligands (thiol group in this study) strongly binds to the QD surfaces, and the other group (carboxyl group in this study) can bind to the TiO 2 surfaces [46,[49][50][51]. Thus, MPA acts as a linker that bridges the QDs and TiO 2 . Owing to this adsorption mechanism, the amount of QDs adsorbed on TiO 2 can be controlled by controlling the pH of the solution. The pH of the QD solution affects the state of the carboxyl groups. When the solution pH decreases, the proportion of ionized carboxyl groups increases. This makes MPA-passivated QDs more dispersible in polar solvents but prevents their binding to TiO 2 surfaces. As a result, MPA-passivated QDs were more densely adsorbed on the TiO 2 films at higher pH (Figure 3b, bottom). In a pH 14 solution, the TiO 2 -CISe QD photoanodes showed the highest adsorption density of QDs, as confirmed by their color in photographs. With this optimization, TiO 2 -CISe QD photoanodes with MPA-passivated QDs had higher QD adsorption densities than those with OAm-passivated QDs, which was verified by both the absorption spectra ( Figure 3a) and photographs (Figure 3b).
The prepared TiO 2 -CISe QD photoanodes were used for PEC hydrogen generation. The cell was composed of a conventional three-electrode system with an SCE reference electrode and a Pt rod counter electrode in an electrolyte containing 0.25 M Na 2 S and 0.35 M Na 2 SO 3 at a controlled pH of 12.9 (Materials and Method 2.4), which is known to be an effective system for PEC hydrogen generation [39]. Anodic LSV scans were obtained to understand the PEC properties of the CISe QD-based photoanodes containing QDs with mono-and bifunctional ligands (Figure 4a). The bare TiO 2 anodes were also measured as the control sample. The current-density-voltage curves show that all electrodes produced an anodic photocurrent from −0.2 V RHE with a plateau from 0.2 V RHE . Both photoanodes made using MPA-passivated QDs (denoted as MPA-QD-photoanodes), and OAm-passivated QDs (denoted as OAm-QDphotoanodes) produced much higher photocurrent densities than bare TiO 2 photoanodes. The significant enhancement in the photocurrent density demonstrates that both QDs can effectively act as additional light absorbers. The introduction of QDs greatly extended the light absorption range and intensity of TiO 2 (Figure 3a), and the photoexcited electrons produced by QDs were transferred from the QDs to TiO 2 . For photoanodes made with OcAm-, OAc-, and DDT-passivated QDs, the photocurrents were lower than those of MPA-QD-photoanodes or OAm-QD-photoanodes ( Figure S9, Supplementary Information). This was attributed to the poor sensitization of the photoanodes because of either poor colloidal stability or too strong passivation of these ligands. currents were lower than those of MPA-QD-photoanodes or OAm-QD-photoanodes (Figure S9, Supplementary Information). This was attributed to the poor sensitization of the photoanodes because of either poor colloidal stability or too strong passivation of these ligands.
The stability of the photoanodes was also tested at 0.6 VRHE under light irradiation ( Figure S10, Supplementary Information). About 36% of the initial photocurrent was maintained after 1 h of operation. It should be noted that increasing the stability of the QD-photoanode is usually related to the structural engineering of QDs or photoanodes rather than surface ligands, which is beyond the scope of the current study. Although the stability of the photoanodes is not as high as those using conventional CdSe QDs with structural optimization [57], we anticipate that this will be improved with future research. Unexpectedly, the photocurrent density of an OAm-QD-photoanode (~8.2 mA/cm 2 at 0.6 VRHE) is clearly higher than that of an MPA-QD-photoanode (~6.7 mA/cm 2 at 0.6 VRHE) despite the high QD adsorption of the latter (Figure 4a and Table 1). Generally, the photocurrent density is approximately proportional to the adsorption density of the absorbers because more absorbers can produce more photoexcited electrons. The results suggest that electron transfer between QDs and TiO2 is not as effective for MPA-QD-photoanodes compared to OAm-QD-photoanodes. This is attributed to the different binding mechanisms of the QDs according to the surface ligands. The direct adsorption of Oam-QDs by The stability of the photoanodes was also tested at 0.6 V RHE under light irradiation ( Figure S10, Supplementary Information). About 36% of the initial photocurrent was maintained after 1 h of operation. It should be noted that increasing the stability of the QD-photoanode is usually related to the structural engineering of QDs or photoanodes rather than surface ligands, which is beyond the scope of the current study. Although the stability of the photoanodes is not as high as those using conventional CdSe QDs with structural optimization [57], we anticipate that this will be improved with future research.
Unexpectedly, the photocurrent density of an OAm-QD-photoanode (~8.2 mA/cm 2 at 0.6 V RHE ) is clearly higher than that of an MPA-QD-photoanode (~6.7 mA/cm 2 at 0.6 V RHE ) despite the high QD adsorption of the latter (Figure 4a and Table 1). Generally, the photocurrent density is approximately proportional to the adsorption density of the absorbers because more absorbers can produce more photoexcited electrons. The results suggest that electron transfer between QDs and TiO 2 is not as effective for MPA-QDphotoanodes compared to OAm-QD-photoanodes. This is attributed to the different binding mechanisms of the QDs according to the surface ligands. The direct adsorption of Oam-QDs by partial ligand detachment was beneficial for electron transfer between QDs and TiO 2 . However, although bifunctional ligands were helpful in increasing QD adsorption, they can prevent the effective charge transfer between QDs and TiO 2 . To gain a better understanding of the effects of surface ligands on the PEC performance, EIS analysis was performed using the TiO 2 -QD photoanodes (at 0.6 V RHE in the dark state and under simulated one sun illumination). Nyquist plots (Figure 4b) were fitted using the equivalent circuit model shown in the inset, where R S is the solution resistance, and the RC circuit represents the charge-transfer characteristics of the TiO 2 -CISe QD photoanodes and the interface between the photoanodes and electrolyte [18,58]. Consequently, R ct is the Materials 2022, 15, 6010 9 of 13 resistance related to the charge transfer between the photoanodes and the electrolyte. As listed in Table 1, our results show that the R ct of the MPA-QD-photoanodes was higher than that of the OAm-QD-photoanodes despite the high QD adsorption density of the MPA-QD-photoanodes. This was attributed to the presence of organic linkers between the MPA-QDs and TiO 2 , leading to the poor charge transfer between the QDs and redox couples in the electrolyte. We can expect that the resistance between the QDs and TiO 2 is smaller in OAm-QD-photoanodes because of the direct attachment between the inorganic parts of the QDs and TiO 2 . These results imply that the better charge separation in the OAm-QD-photoanodes compared to that in the MPA-QD-photoanodes results in enhanced hole transfer from the QDs to the redox couples in the electrolyte [18,58]. In addition, the behavior of the resistance at the interface of TiO 2 -QDs is consistent with a previous spectroscopy study on electron transfer between TiO 2 and QDs with controlled surface ligands [59]. In the literature, it was demonstrated that MPA linkers between QDs and TiO 2 can inhibit effective electron transfer.
We also analyzed the electron recombination kinetics in the TiO 2 -QD photoanodes with open-circuit voltage decay (OCVD) analysis, observing the decay of the open circuit voltage (V OC ) after turning off the illumination. The OCVD curves showed decay after 20 s in the dark because of charge recombination between the charges from photoanodes and redox couples in the electrolyte (Figure 5a). The V OC of the MPA-QD-photoanodes decayed more rapidly than that of OAm-QD-photoanodes. The electron lifetimes of the photoanodes were calculated from the OCVD data, and the electron lifetime versus voltage curves [60,61] are shown in Figure 5b. It is clear that the electron lifetime of the OAm-QD-photoanodes is longer (i.e., the charge recombination rate is higher) than that of MPA-QD-photoanodes. It is proposed that the organic linker molecules between the QDs and TiO 2 in MPA-QDphotoanodes acted as defects [62], while the inorganic cores of QDs and TiO 2 formed a direct junction in the OAm-QD-photoanodes. Considering that the major pathway of charge recombination is electron transfer from the TiO 2 conduction band to the redox couples in the electrolyte [33,49], these defects at the interface of OAm-QDs and TiO 2 can act as recombination centers, leading to inferior PEC performance. ligands [59]. This also leads to enhanced hole transfer between the QDs and redox couples in the electrolyte and an increase in the electron lifetime in the photoanode. These results explain why the OAm-QD-photoanodes produced a high photocurrent despite the lower QD adsorption density. It should also be noted that OAm may not be the optimal surface ligand for PEC using QDs. Our findings imply that electron transfer between the QDs and TiO2 is important for PEC hydrogen production, which requires the careful design of the surface states of QDs.

Conclusions
This work studied the effects of the surface ligands of colloidal CISe QDs on the fabrication of PEC photoelectrodes and their resulting PEC characteristics. In particular, OAm-and MPA-passivated CISe QDs were carefully chosen for this investigation to comprehend the effects of mono-and bifunctional ligands on PEC hydrogen production employing CISe QDs. TiO2-QDs photoanodes were prepared by adsorbing QDs onto mesoporous TiO2, and the surface ligands affected the QD adsorption process. Inorganic cores of OAm-passivated QDs were directly adsorbed on TiO2 by partial ligand detachment, which is beneficial for electron transfer from QDs to TiO2. Bifunctional ligands can act as linkers by bridging QDs and TiO2, and the amount of QD adsorption was higher for MPApassivated QDs than for OAm-passivated QDs. With this tradeoff, OAm-QD-photoanodes and MPA-QD-photoanodes showed current densities of ~8.2 mA/cm 2 and ~6.7 mA/cm 2 , respectively, at 0.6 VRHE under one sun illumination. These findings suggest that not only the QD adsorption density but also the electron transfer between QDs and TiO2 The EIS and OCVD data were consistent with the proposed QD adsorption mechanism of each photoanode: (i) A direct contact was formed between the QDs and TiO 2 in OAm-QD-photoanodes, and (ii) the QDs and TiO 2 were connected by linker molecules in the MPA-QD-photoanodes. The direct contact between the QDs and TiO 2 results in efficient electron transfer between them, which is also consistent with the results of a previous spectroscopy study on electron transfer between TiO 2 and QDs with controlled surface ligands [59]. This also leads to enhanced hole transfer between the QDs and redox couples in the electrolyte and an increase in the electron lifetime in the photoanode. These results explain why the OAm-QD-photoanodes produced a high photocurrent despite the lower QD adsorption density. It should also be noted that OAm may not be the optimal surface ligand for PEC using QDs. Our findings imply that electron transfer between the QDs and TiO 2 is important for PEC hydrogen production, which requires the careful design of the surface states of QDs.

Conclusions
This work studied the effects of the surface ligands of colloidal CISe QDs on the fabrication of PEC photoelectrodes and their resulting PEC characteristics. In particular, OAm-and MPA-passivated CISe QDs were carefully chosen for this investigation to comprehend the effects of mono-and bifunctional ligands on PEC hydrogen production employing CISe QDs. TiO 2 -QDs photoanodes were prepared by adsorbing QDs onto mesoporous TiO 2 , and the surface ligands affected the QD adsorption process. Inorganic cores of OAm-passivated QDs were directly adsorbed on TiO 2 by partial ligand detachment, which is beneficial for electron transfer from QDs to TiO 2 . Bifunctional ligands can act as linkers by bridging QDs and TiO 2 , and the amount of QD adsorption was higher for MPApassivated QDs than for OAm-passivated QDs. With this tradeoff, OAm-QD-photoanodes and MPA-QD-photoanodes showed current densities of~8.2 mA/cm 2 and~6.7 mA/cm 2 , respectively, at 0.6 V RHE under one sun illumination. These findings suggest that not only the QD adsorption density but also the electron transfer between QDs and TiO 2 are critical for PEC hydrogen production. Our results highlight the importance of surface-ligand engineering of QDs for effective PEC hydrogen production.