CdTe quantum dots (QDs) are attractive photosensitizers for photocatalytic proton reduction due to their broad absorbance profile that can extend from the ultraviolet to near-infrared regions, providing access to a larger portion of the solar spectrum than possible with analogous CdSe and CdS QD photosensitizers. Here, the photocatalytic hydrogen (H2) generation from various sizes of dihydrolipoic acid (DHLA)-capped CdTe QDs, ranging from 2.5 to 7.5 nm in diameter, and a molecular Ni-DHLA catalyst in aqueous solutions was evaluated, and an unusual size-dependent photocatalytic activity with CdTe QDs was observed. Under optimized conditions, using 3.4 nm CdTe-DHLA and a 1:20 ratio of QD/Ni-DHLA catalyst, as many as 38 000 turnover numbers (mol H2 per mol QD) were achieved. However, below this critical size, the H2 production efficiency decreased; this behavior is attributed to the rapid oxidation of the QD surface, resulting in detrimental surface trap states. These results are consistent with ultrafast transient absorption spectroscopic measurements, which suggest the presence of extremely fast charge-trapping processes in the oxidized CdTe-DHLA QDs. While fast electron transfer from CdTe-DHLA QDs is observed in the presence of the Ni-DHLA catalyst, the charge trapping processes occur on a competitive time scale, thus lowering the efficiency of the CdTe/Ni-DHLA H2 production system. Understanding rapid charge trapping in CdTe QDs may help suggest potential improvements for the overall CdTe photocatalytic system.
I. INTRODUCTION
Direct conversion of abundant solar energy into chemical fuels such as hydrogen (H2) by photochemical water splitting holds great promise for addressing long-term, ever-increasing energy demands on a global scale.1,2 H2 is considered to be an ideal renewable energy source owing to its clean combustion by-product and high efficiency.3,4 Over the past decade, many different forms of catalytic systems for photochemical H2 generation have been reported.5–7 Typically, these systems consist of a photosensitizer, a catalyst, and proton and electron sources. As shown in Fig. 1, after the absorption of light by the photosensitizer, the photoexcited electron is then transferred to the catalyst, where protons are reduced to H2, and the photosensitizer is reduced by the sacrificial electron donor, completing the catalytic cycle. The efficiency of photoreduction is directly dependent on the relative rates of these charge transfer processes.8,9 As photosensitizers, quantum dots (QDs) inherently possess several desirable properties for photocatalytic proton reduction reactions including size-tunable oxidation and reduction potentials,10,11 broad absorption spectra,12 rich surface-binding properties,13 high surface-to-volume ratios, and low costs.
We recently reported highly active and stable photocatalytic H2 generation systems using dihydrolipoic acid (DHLA)- or 3-mercaptopropionic acid (MPA)-capped CdSe QD photosensitizers, Ni-DHLA catalyst, and ascorbic acid (AA) as the sacrificial donor.14,15 While these systems exhibited robust proton reduction activity for weeks and achieved quantum efficiencies (QYs) for H2 generation as high as 46% in water15 and 59% in 1:1 ethanol/water,14 the absorption energies of CdSe QDs limit the available excitation wavelengths to the visible. CdTe QDs, however, offer the possibility for near-infrared (NIR) absorption,16 allowing for access to a broader range of the solar spectrum. Given the conduction and valence band energies of −1.0 and 0.54 V vs the normal hydrogen electrode (NHE), respectively, for bulk CdTe17 and their NIR absorption, CdTe QDs have attracted growing interest for solar photochemistry applications.
CdTe QDs have been previously explored as photosensitizers for proton reduction in a variety of photocatalytic systems. Several molecular cobalt cocatalysts have been investigated for use with CdTe QD photosensitizers,18,19 with the highest observed activity using thioglycolic acid-capped CdTe QDs with a macrocyclic cobalt catalyst coordinated to the QD.19 By anchoring the catalyst directly to the QD surface, improved activity and stability were observed, with the system producing 14 400 turnover numbers (TONs, mol H2 per mol catalyst). Enhanced H2 generation activity (100 000–200 000 TONs, mol H2 per mol QD) has also been observed for CdTe QDs with the addition of transition metal salts.20–22 Artificial and natural hydrogenases have been used with CdTe nanoparticles23–25 in photocatalytic H2 generation systems, as well. For example, as many as 52 800 TONs (mol H2 per mol catalyst) have been achieved, with this high activity attributed to confining the photocatalytic components using chitosan.24 While these studies provide a solid foundation for the design of CdTe QD-based photocatalytic systems, most reports have only explored a single CdTe QD size. Thus, their role in photocatalytic systems is poorly understood, and a comprehensive study of CdTe QD-based photocatalytic systems is needed.
In this work, we report studies of light-driven H2 generation using a series of CdTe-DHLA QD photosensitizers, ranging in diameter from 2.5 to 7.5 nm, and a molecular Ni-DHLA catalyst. All systems exhibited robust proton reduction activity lasting over two days (48 h), but surprisingly, under similar conditions, the CdTe-DHLA QDs exhibited much lower H2 generation activity compared with CdSe-DHLA QDs.14 Furthermore, though we expected H2 production activities to improve as we decreased the QD diameter, due to the increased driving force for electron transfer (ET) as a result of the widened bandgap for smaller diameter QDs, the size-dependent proton reduction activity decreased dramatically as the CdTe QD diameter is decreased below a critical size (3.4 nm). In order to better understand the unusual size-dependent H2 production behavior, the ET dynamics from CdTe QDs to the Ni-DHLA catalyst were measured using transient absorption (TA) spectroscopy. While subpicosecond ET to the catalyst is observed, competitive surface trapping processes seem to have a detrimental effect on the ET efficiency. We suspect that these surface trapping effects become more significant with smaller CdTe QDs, which have a higher surface-to-volume ratio, leading to the unexpected trend in the size-dependent photocatalytic activity. Overall, these results have important negative implications for the use of CdTe QDs for solar fuel generation.
II. EXPERIMENTAL DETAILS
CdTe QDs were synthesized by variations of literature methods,26,27 yielding hydrophobic nanocrystals. Synthetic details and sample characterization information are given in the supplementary material. A range of CdTe QD sizes were obtained by varying the reaction time. Following synthesis, the QDs were rendered water soluble by ligand exchange with DHLA (see Fig. S1 of the supplementary material).14 For photocatalytic studies, the Ni-DHLA catalyst was synthesized as described previously (experimental details are provided in the supplementary material).28
Photocatalytic H2 generation measurements were conducted in a temperature-controlled (15 °C), custom-built 16-sample apparatus using conditions similar to our previous studies.14,15,29 Samples, protected from both air and light before use, were prepared in 40 mL scintillation vials with 5.0 ml of solution containing varying concentrations of AA (the sacrificial electron donor), CdTe-DHLA QDs, and Ni−DHLA catalyst in 1:1 ethanol/water. The pH of the solutions was adjusted to 4.5 by the addition of sodium hydroxide. The vials were sealed with air-tight caps, fitted with pressure transducers which can monitor the headspace pressure in real time, and septa. The samples were then degassed with a 4:1 N2/CH4 gas mixture (1 atm). This gas mixture was used to retain an oxygen-free environment, while CH4 served as an internal standard for subsequent gas chromatography measurements (described in the supplementary material). The vials were irradiated from below with high power Philips LumiLED Luxeon Star Hex green (525 nm) light emitting diodes (LEDs). The light power of each LED was set to approximately 130 mW as measured with a Newport power meter (Model 1918-C). The samples were swirled using an orbital shaker, and pressure changes in the vials were recorded using a Labview program from a Freescale semiconductor sensor (MPX4259A series). All H2 evolution experiments were allowed to run for 48 h.
TA measurements were performed using a previously described standard pump-probe setup.15,28,30 Femtosecond pulses (∼100 fs) from an 800 nm Ti:sapphire oscillator (Spectra Physics, Tsunami) were amplified by a 1 kHz Ti:sapphire regenerative amplifier (Spectra Physics, Spitfire) and introduced into an optical parametric amplifier (OPA; Spectra Physics, OPA 800C). The resulting signal pulses from the OPA were tuned to 1140 nm, and a single BaB2O4 (BBO) crystal was used to double the frequency and generate a wavelength of 570 nm for the pump, probe, and reference beams. The reference beam was split from the probe before the sample and used to correct the probe for pulse-to-pulse fluctuations. A Newport motorized delay stage was used to control the optical time delay between the pump and probe pulses. Samples were placed in a 2 mm path length quartz cuvette and stirred to avoid laser-induced signal artifacts. The optical density of CdTe QDs was ∼0.2 at the pump wavelength of 570 nm. The average number of excitons absorbed per QD, ⟨N0⟩, was determined using the pump-fluence dependence of the TA signal at a delay time of 2 ps, as previously described (see Fig. S8 of the supplementary material)28,31 and was kept below one for all measurements.
III. RESULTS AND DISCUSSION
The UV-Vis absorption spectra for a series of CdTe-DHLA QDs in water are shown in Fig. 2. All the spectra show sharp and well-defined first excitonic peaks, with the transition wavelength increasing from 504 nm to 718 nm with increasing QD diameters. [Photoluminescence (PL) spectra for CdTe-DHLA QDs are shown in Fig. S2 of the supplementary material.] The corresponding transmission electron microscope (TEM) micrographs further confirm that all prepared QDs have an approximately spherical shape and narrow size distributions, with diameters varying from 2.5 ± 0.2 nm to 7.5 ± 0.7 nm as determined from the TEM (see Fig. S3 of the supplementary material). After ligand exchange, blue shifts were observed in the UV-Vis absorption spectra for all CdTe-DHLA QD samples. For example, after capping with DHLA, the lowest energy transition of 3.6 nm CdTe QDs blue shifted from 593 nm to 588 nm (see Fig. S4 of the supplementary material). This behavior may be due to the inherent instability of Te anions, which can be readily oxidized by dissolved oxygen during the ligand exchange process,32 resulting in a corresponding decrease in the size of the CdTe cores. However, this phenomenon was not observed for DHLA-capped CdSe nanocrystals examined in previous studies,14,29 which may be due to the relative stability of CdSe toward oxidation compared to CdTe.17,33
The photocatalytic H2 generation over time from DHLA-capped CdTe QDs ranging in size from 2.5 to 7.5 nm is shown in Fig. 3. For all QD sizes, DHLA-capped CdTe QD photocatalysts successfully generated H2 and retained their activity for two days, achieving an initial QY as high as 16.3% over the first 15 h of irradiation. For CdTe QDs of optimal size (i.e., 3.4 nm in Fig. 4), ∼38 000 TONs (mol H2 per mol QD) were achieved under optimized conditions (1 µM CdTe-DHLA, 20 µM Ni-DHLA, and 0.5 M AA at 15 °C). Based on the reduction potential of Ni-DHLA and the calculated conduction band energies of various CdTe sizes, the catalytic ET driving force of the system was calculated (Fig. S5 and Table SI of the supplementary material). Due to the decreased driving force and surface charge density, it is expected that QDs with larger diameters would exhibit a decrease in H2 production activity.34 However, an unusual trend in proton reduction activity was observed for smaller-sized CdTe QDs. In Fig. 4(a), the H2 production activity from Fig. 3 is plotted with respect to the driving force for ET. The plot reveals two distinct areas of activity: (1) QDs with sizes between 3.4 and 7.5 nm exhibit an increase in activity with increasing driving force and (2) QDs with diameters smaller than 3.4 nm exhibit a decrease in activity with increasing driving force. The trend in H2 generation activity in region (1) is expected due to the decreased driving force for ET and decreased surface charge density for larger diameter QDs, and similar size-dependent H2 generation activity has been observed previously for CdSe QDs.14,35,36 In contrast, the behavior observed in region (2) is quite unexpected considering that the H2 production efficiency of the previously studied CdSe QD-Ni-DHLA system only increased as the CdSe QD diameter was decreased.14 The external QY for H2 generation was plotted with respect to CdTe QD diameter, as shown in Fig. 4(b), to account for the relative amount of light absorbed by each QD sample. However, a decrease in the H2 QY was still observed below the critical QD size (3.4 nm), and this decrease in absorption at smaller QD sizes clearly does not account for the observed H2 generation behavior. It is important to note that in Fig. 3 (left), a distinct change in slope is observed for 2.9 nm and 3.4 nm CdTe QDs after extended periods of time, indicative of a change in H2 production rate. During a photocatalytic experiment, a number of factors can change the H2 production rate including photodegradation of the QDs, colloidal instability due to the photoreduction or photo-oxidation of the surface capping ligands, or photochemical changes in the molecular catalyst. This change in slope was not seen for CdSe QDs and a Ni-DHLA catalyst,14,15 and thus it is reasonable to assume that it is the easily oxidized CdTe QDs that are degrading and causing the change in H2 production rate. Nonetheless, precisely due to this change in slope, the measurements of the H2 production efficiency in Fig. 4 were performed at the 15 h mark, prior to this change in activity.
One possible explanation for the unusual size dependence of the H2 QY is that the driving force for ET is so great that the system is in the Marcus inverted regime, which would decrease the ET rate and subsequently decrease the observed H2 production activity in the photocatalytic system.37,38 As shown in Fig. 4(a) and Table SI, at QD diameters smaller than the critical size (3.4 nm), we estimate the driving force for ET to be greater than 0.75 eV. Studies of charge transfer from colloidal semiconductor QDs to molecular acceptors suggests that driving forces of this magnitude would likely fall into inverted regime behavior under classical Marcus theory.28,39,40 However, Zhu et al. and Olshansky et al. observed no inverted regime (i.e., only an increase in charge transfer efficiency with increasing driving force) due to Auger-assisted charge transfer at comparable driving forces to what we estimate here.39,40 On the other hand, Auger-assisted ET relies on hole intraband excitation coupled to the ET process. In the CdTe-DHLA system, we suspect that the hole is rapidly trapped to the surface, as discussed below, making the hole unavailable for participation in Auger-assisted ET. Indeed, in a similar manner, fast trapping of the hole (or removal of the hole) prevents rapid intraband relaxation of the electron due to an inhibition of Auger excitation of the hole.41,42 Thus, it is not unreasonable to suggest that rapid hole trapping may allow for the observation of the Marcus inverted regime in ET to Ni-DHLA. Confirmation of the presence of such behavior within our system would require extensive, QD size-dependent ET studies, which is beyond the scope of current work.
We believe that this unique size-dependent photocatalytic H2 generation activity observed for smaller diameter CdTe QDs is strongly impacted by rapid surface trapping of electrons and holes. Incomplete saturation of the dangling bonds at the QD surface achieved by conventional organic ligands leads to the formation of trap states43,44 that can localize charge carriers and has been shown to degrade photoactivity performance.45–47 For example, the organic ligands commonly used in the synthesis or ligand exchange of cadmium chalcogenide QDs, such as thiols and amines,48,49 can leave the surface chalcogenide atoms undercoordinated. In CdTe, these unsaturated Te bonds have been identified as hole traps by optically detected magnetic resonance,50 and electrochemical studies indicate that these trap states reside within the bandgap.51,52 Rapid surface trapping of electrons can also compete with ET to the catalyst, subsequently decreasing the H2 generation activity.
The H2 generation system was further evaluated by varying the concentrations of Ni-DHLA catalyst and sacrificial donor, AA, to determine their effect on the overall yield of H2 from the photocatalytic system. Figure 5 shows that for 3.5 nm diameter QDs, the total amount of H2 evolved for the system increased with the catalyst concentration up to 20 µM Ni-DHLA, at which point the highest H2 evolution activity is observed. This trend can be attributed to the increasing probability of ET from CdTe-DHLA QDs to Ni–DHLA with increasing catalyst concentration.14 However, upon further increasing the concentration of catalysts to 40 µM, a decrease in H2 production activity was observed, similar to what we have observed for CdSe QDs with the Ni-DHLA catalyst.15 Changes in the QD surface environment can impact photocatalytic activity,53,54 and it is possible that a significant excess of DHLA at high catalyst concentrations (40 µM) acts as an effective barrier for the ET between QDs and catalyst molecules, deactivating the photocatalytic system.15 In addition, the electron donor concentration dependence was studied using CdTe QDs with a diameter of 3.7 nm. As shown in Fig. S6 of the supplementary material, increasing the concentration of AA from 0.02 M to 0.5 M while keeping the concentration of QDs and Ni-DHLA constant greatly improved the efficiency of H2 production. At even higher AA concentrations, the QD colloid was unstable and precipitated over the course of the experiment. Thus, the highest H2 generation activity was observed at the maximum AA concentration of 0.5 M.
To gain insight into the charge transfer behavior of the photocatalytic system, PL quenching experiments were conducted with CdTe-DHLA QDs in the presence of either Ni-DHLA [Fig. 6(a)] or AA [Fig. 6(c)]. As expected, the band edge PL from CdTe-DHLA QDs is quenched in the presence of both Ni-DHLA and AA, indicating that ET from the QD to the catalyst and hole transfer from the QD to the electron donor are both kinetically and thermodynamically favorable. The concentration dependent PL quenching data were fit with the Stern-Volmer model [Figs. 6(b) and 6(d)],55
where I0 and I are the PL intensities in the absence and presence of quenchers, respectively. [Q] is the quencher concentration, and Kq is the quenching constant that is described by the following equation:
where kq is the quenching rate constant and τ0 is the average PL lifetime in the absence of quenchers (Fig. S7). The observed quenching behavior yields quenching constants Kq and Kq′ of 0.09 and 0.004 for Ni2+ and AA, respectively [Figs. 6(b) and 6(d)]. The values for Kq and Kq′ are comparable to the quenching constants previously reported for DHLA-capped CdSe/CdS dot-in-rod (DIR) nanoparticles with Ni−DHLA (Kq = 0.07) and AA (Kq′ = 0.007).29 Like the CdTe-DHLA QDs studied here, the DIR particles showed low H2 production activity at Ni2+/DIR ratios of 1:1. Furthermore, the quenching rate constants for CdTe-DHLA QDs were determined using Eq. (2) and the fits shown in Figs. 6(b) and 6(d), revealing that the quenching rate constant for Ni2+ (kq = 1.7 × 1013 M−1 s−1) is more than 20 times greater than for AA (kq′ = 7.5 × 1011 M−1 s−1). Under photocatalytic conditions, however, the concentration of AA (0.5 M) is much greater than that of Ni2+ (1 µM). Thus, the hole transfer kinetics from a photoexcited QD to AA are expected to be much faster in the photocatalytic system.
To better understand the CdTe/Ni-DHLA photocatalytic H2 production system, TA spectroscopy was used to study the initial ET dynamics from CdTe QDs to the Ni-DHLA catalyst. By probing the 1Se–1Sh3/2 transition of 3.5 nm QDs at 570 nm, the state filling of the 1Se–1Sh3/2 exciton can be monitored through the time-resolved photobleach recovery.31,56,57 The TA signal for DHLA-capped CdTe QDs without the addition of Ni-DHLA catalyst was fit to a multiexponential decay (Fig. 7), yielding lifetimes of 1 ps (79% of the TA signal), 13 ps (17% of the TA signal), and 1429 ps (4% of the TA signal), with an amplitude weighted average lifetime of 62 ps (Table I). Similarly, multiexponential lifetimes have been reported for thiol-capped CdTe QDs with time constants on the same order of magnitude as observed here.58–60 Rapid decay components of the 1Se–1Sh3/2 exciton, on a single picosecond time scale, are likely due to surface trapping processes in both CdTe and CdTe/CdSe QDs.59,61–65 A long lifetime component (>∼1 ns) has previously been observed in CdTe QDs60,61,65–67 and is often attributed to band edge recombination.58,59,62 It is important to note that here ∼96% of the TA signal is attributed to rapid surface trapping due to the rapid surface oxidation of the CdTe QDs in water. For comparison, the lifetimes obtained from a multiexponential fit for the TA decay of CdTe QDs prior to ligand exchange (see Fig. S9 of the supplementary material) were 4 ps (41% of the TA signal), 68 ps (33% of the TA signal), and 5000 ps (26% of the TA signal), with an amplitude weighted average lifetime of 1325 ps (see Table SII of the supplementary material). Thus, the faster decay observed for CdTe-DHLA, compared to CdTe QDs with native capping ligands in organic solutions, suggests that significantly more surface defects are introduced following ligand exchange to DHLA.57
Sample . | τQD1 (ps) (A1, %) . | τQD2 (ps) (A2, %) . | τQD3 (ps) (A3, %) . | ⟨τQD⟩ (ps) . | τET (ps) (BET, %) . |
---|---|---|---|---|---|
No catalyst | 1 ± 0.1 (79 ± 7) | 13 ± 3 (17 ± 5) | 1429 ± 0 (4 ± 2) | 60 ± 6 | |
20:1 Ni-DHLA/QDs | 1 ± 0.1 (46 ± 3) | 13 ± 3 (2 ± 3) | 1429 ± 0 (0 ± 0) | 2 ± 0.2 | 0.6 ± 0.2 (52 ± 6) |
Sample . | τQD1 (ps) (A1, %) . | τQD2 (ps) (A2, %) . | τQD3 (ps) (A3, %) . | ⟨τQD⟩ (ps) . | τET (ps) (BET, %) . |
---|---|---|---|---|---|
No catalyst | 1 ± 0.1 (79 ± 7) | 13 ± 3 (17 ± 5) | 1429 ± 0 (4 ± 2) | 60 ± 6 | |
20:1 Ni-DHLA/QDs | 1 ± 0.1 (46 ± 3) | 13 ± 3 (2 ± 3) | 1429 ± 0 (0 ± 0) | 2 ± 0.2 | 0.6 ± 0.2 (52 ± 6) |
Upon the addition of 20× Ni-DHLA, an even faster bleach recovery of the 1Se–1Sh3/2 exciton is observed, as shown in Fig. 7 (see Fig. S10 of the supplementary material for the raw data). The faster decay is indicative of ET from CdTe QDs to the Ni-DHLA catalyst, as the TA signal of pure Ni-DHLA has been shown to be negligible compared to the TA signal observed from QDs.15,28 The ET lifetimes can be estimated through a multiple rate equation model incorporating the intrinsic electron excited state lifetimes and the ET lifetimes, fixing the intrinsic lifetimes to the same values of those obtained without Ni-DHLA present, while the pre-exponential factors (Ai and Bj) are allowed to vary,28,68–71
In Eq. (3), ΣiAi exp(−t/τi) corresponds to the intrinsic electron dynamics for QDs without catalyst present and ΣjBj exp(−t/τjET) accounts for ET processes from the QDs to the Ni-DHLA catalyst. From Eq. (3), an ET lifetime of <1 ps was inferred (Table I). This ET lifetime is extremely fast but is consistent with previous reports showing ultrafast electron excited state dynamics in CdTe and CdTe/CdS QDs paired with molecular acceptors.62,64,66,72 In addition, upon addition of the Ni-DHLA catalyst, the fractional contributions to the TA signal from the intrinsic electron relaxation lifetimes decrease (to 46%) at the expense of the ultrafast ET (52%). However, while the decay due to ET accounts for majority of the TA signal, this ET efficiency (∼50%) is still significantly less than previously observed for the CdSe QDs and Ni-DHLA system (>90%),15,28 which may help explain the relatively poor performance of CdTe QDs with respect to CdSe QDs. It is worth noting that the subpicosecond ET lifetime observed here for CdTe QDs to Ni-DHLA is significantly faster than that for CdSe QDs with the Ni-DHLA catalyst.15,28 While such fast ET for freely diffusing molecular catalysts is unusual, further studies are needed to completely understand the ET process from CdTe QDs to Ni-DHLA. It is also worth noting that qualitatively similar trends were observed when probing the 1Se–1Sh3/2 transition of 3.3 nm QDs at 560 nm QDs both without and with 20× Ni-DHLA catalyst; however, the large standard deviation in CdTe surface quality precludes making strong comparisons between QD sizes.
IV. CONCLUSION
The photocatalytic H2 evolution activity was examined for CdTe-DHLA QDs, with diameters ranging from 2.5 to 7.5 nm, and a Ni-DHLA catalyst to examine the QD size-dependence on photocatalytic activity. All systems demonstrated continuous proton reduction activity over 48 h of irradiation (λex = 525 nm), and the optimal H2 production activity was observed with 3.4 nm CdTe-DHLA QDs, which exhibited a QY of 16.3% for H2 generation for the first 15 h of illumination. Above this diameter, a decrease in the H2 generation efficiency was observed and attributed, primarily, to a decrease in driving force for ET and a reduced surface charge density, consistent with CdSe QDs and Ni-DHLA catalysts.14,15,28 However, the H2 production efficiency also decreased for CdTe-DHLA with diameters less than 3.4 nm, which was very unexpected since smaller QDs have a larger driving force for charge transfer. We ascribe this behavior primarily to the rapid oxidation of the CdTe surface: undercoordinated chalcogenide atoms have been reported to influence charge trapping at the QD surface51 and spectroelectrochemical investigations of CdTe films have suggested the presence of both electron and hole traps at CdTe QD surfaces.52 Rapid electron trapping is also evident from the fast TA decays observed for CdTe-DHLA. In addition, it may be possible that rapid hole trapping allows for observation of Marcus inverted behavior for small CdTe QDs, due to the large driving force for ET, accounting for the poor H2 generation in this region.
These studies suggest that mitigating the deleterious effects of charge trapping is necessary to achieve improvements in the photocatalytic H2 production activity using CdTe QDs in this system. Though shelled QDs have been shown to negatively impact H2 generation activity29 and slow ET processes28 due to the presence of a charge transfer barrier, it is possible that by growing a very thin or patchy shell on the surface of CdTe QDs, surface trapping effects could be minimized, improving the H2 generation activity despite the presence of a small energetic barrier to charge transfer.73 Alternatively, utilizing surface ligands (such as chloride ions) that have been shown to mitigate surface trapping in CdTe QDs52,61 may also be a viable strategy for improving ET from CdTe QDs and, subsequently, improving the H2 generation activity for this system. Therefore, even though the photocatalytic proton reduction system reported in this work is presently not competitive with the previously examined CdSe systems using a Ni-DHLA catalyst,14,15 these results represent an important contribution toward understanding the reaction kinetics in nanostructured photocatalytic systems and provide direction toward improving upon this photocatalytic system and making CdTe QDs more competitive photosensitizing materials for solar fuel production.
SUPPLEMENTARY MATERIAL
Experimental details and supplementary data regarding sample preparation and characterization, calculation of QY for H2 generation, band edge energies and driving force for ET, H2 production with varying concentrations of AA, fluorescence quenching experiments, and TA measurements are provided in the supplementary material.
ACKNOWLEDGMENTS
This work was supported by the Chemical Sciences, Geosciences, and Biosciences Division, Office of Basic Energy Sciences, Office of Science, U.S. Department of Energy (Grant No. DEFG02-09ER16121). J.Y. was supported by a fellowship awarded by the China Scholarship Council. The authors thank Abigail Freyer and Sean O’Neill for assistance recording TEM micrographs of CdTe QDs.