Influence of donor-to-acceptor ratio on excited-state electron transfer within covalently tethered CdSe/CdTe quantum dot colloidal heterostructures

The unique size-dependent optical and electronic properties of semiconductor quantum dots (QDs), as well as their rich surface chemistry and their potential to undergo multi-exciton generation and hot carrier extraction, have driven interest in using QDs as harvesters of light and donors of excited-state charge carriers for solar energy applications.^1–6 In recent years, an increased interest in exploiting the transfer of photogenerated charge carriers from QDs to other species has propelled the use of QDs in energy conversion and storage applications, including redox photocatalysis and photoelectrochemical H₂ generation, CO₂ reduction, and water splitting.^2,7–10

Individual QDs can participate in multiple charge-transfer reactions by accommodating multiple excitons (as a donor) or injected photogenerated charge carriers (as an acceptor) due to the degeneracy of lowest energy excitonic states.¹⁰ Therefore, using QDs as both the donor and the acceptor of charge carriers within a given construct can offer advantages over donor–acceptor systems consisting of QDs and molecular acceptors. A single QD donor can facilitate the adsorption of anywhere from one to thousands molecules on its surface, depending on the QD’s size and surface chemistry, whereas molecular chromophores are much more limited in the number of bound acceptors that can be accommodated.^10,11 Dual-QD heterostructures, in particular, are intriguing architectures for improving light harvesting, excited-state charge transfer, and redox photocatalytic processes.^10,12–16 Absorption spectra and solar light-harvesting efficiency are tunable with the size and composition of constituent QDs. When two strongly absorbing QDs are interfaced with the electronic structure tuned to yield a type-II energetic offset, then bidirectional inter-QD excited-state charge-transfer mechanisms are thermodynamically favorable. This multidirectional charge transfer increases the distance between photogenerated charge carriers and yields longer lived charge-separated states.^17,18

We and others have prepared QD heterostructures using various methods, including successive ionic layer adsorption and reaction, epitaxial growth, and the tethering of colloidal QDs via adsorbed ligands.^12,18–22 Our preferred method for making QD heterostructures utilizes carbodiimide-mediated coupling to form an amide bond between the terminal functional groups of QD-adsorbed ligands.^18,19 This covalent coupling method allows for selective formation of heterostructures rather than homostructures, affords control over the extent of heterostructure formation, and is a straightforward method to systematically control the relative concentrations of donor and acceptor QDs.¹⁹ We recently reported on the synthesis of colloidal CdSe/CdTe QD heterostructures using this method.¹⁸ Successful covalent tethering between QD-adsorbed terminal functional groups was supported by ATR-FTIR spectra, ¹H NMR spectra, the quenching of steady-state emission, and the extent of agglomeration of QDs evidenced in TEM images. Bidirectional excited-state charge transfer, wherein excited-state electrons were transferred from CdTe to CdSe and excited-state holes were transferred from CdSe to CdTe, occurred on time scales of 10–100 ns, as evidenced by dynamic quenching of the band-edge emission from CdTe QDs and the trap-state emission from CdSe QDs, respectively.^18,19

Charge-transfer dynamics dictate the quantum yield of charge separation within any donor–acceptor system, which, in turn, directly affects the efficiencies of subsequent redox events underpinning photocatalysis. Therefore, it is crucial to understand and control charge-transfer dynamics within QD heterostructure systems. Notably, for QD-molecule donor–acceptor systems, the observed rate constants of charge transfer have been shown to increase with the number of molecular acceptors bound to the surfaces of QDs.^{10,11,23–26} However, for dual-QD heterostructures, the extent to which charge-transfer dynamics depends on the average number of acceptors per donor QD, as well as how to control the donor-to-acceptor ratio, remains relatively unexplored. Understanding how the donor–acceptor ratio affects the rate and efficiency of QD-to-QD charge transfer is of fundamental interest and also can inform the design of QD heterostructures for photocatalysis. For example, the anticipated acceleration of charge transfer upon increasing the number of acceptor QDs per donor QD would increase the quantum yield of charge separation and maximize the number of available charge carriers for redox catalysis. However, diluting the extracted electrons or holes among a greater number of peripheral charge-accepting QDs within heterostructures may not be optimal for promoting subsequent multi-electron reductions or multi-hole oxidations. Determining whether the donor–acceptor ratio affects charge-transfer dynamics and, if so, whether there is a threshold number of charge-accepting QDs per donor QD within heterostructures to maximize the rate of charge transfer can, thus, reveal design criteria for heterostructures with optimal charge-transfer reactivity for photocatalysis.

Primary factors expected to affect the rate constants of excited-state charge transfer within QD heterostructures include inter-QD separation, inter-QD electronic coupling, and the driving force of charge transfer, as well as the number of QD acceptors interfaced with a given light-harvesting donor QD. We set out to explore the influence of the QD donor-to-acceptor ratio on charge transfer using our model system of CdSe-amide-CdTe heterostructures (Scheme 1), for which the ratio of CdSe and CdTe QDs within heterostructures is readily adjusted with relative concentrations. This article reports our findings, highlighting the synthesis of colloidal CdSe-amide-CdTe heterostructures with varying molar ratios of electron-donor (CdTe) and acceptor (CdSe) QDs, as well as the strong influence of the donor-to-acceptor ratio on the dynamics and yield of excited-state electron transfer from photoexcited CdTe to CdSe QDs.

CdSe QDs, CdTe QDs, and CdSe-amide-CdTe heterostructures were synthesized and characterized as described previously.^18,27–31 Detailed descriptions of materials synthesis and characterization, and the spectroscopic characterization of excited-state charge transfer, are presented in Appendixes S1.1–S1.3 of the supplementary material. To prepare mixtures with variable donor–acceptor ratios, dispersions of either 4-aminothiophenol (4-ATP)-capped CdTe (4-ATP-CdTe) QDs or thiophenol (TP)-capped CdTe (TP-CdTe) QDs (0.20–0.27 µM) were combined with dispersions of N-hydroxysuccinimide ester (NHS-ester)-capped CdSe (NHS-CdSe) QDs (0.0–35 µM). Reaction mixtures were deaerated under Ar (15 min) at room temperature and were allowed to equilibrate (30 min) prior to spectroscopic characterization of excited-state charge transfer.

The syntheses and subsequent surface modification of (a) thioglycolic acid (TGA)-capped CdSe (TGA-CdSe) QDs to yield NHS-CdSe QDs and (b) tetradecylphosphonic acid (TDPA)-capped CdTe (TDPA-CdTe) QDs to yield either 4-ATP-CdTe or TP-CdTe QDs have been described in our published work on the carbodiimide-mediated formation of QD heterostructures.¹⁸ Herein, colloidal dispersions of NHS-CdSe and 4-ATP-CdTe QDs exhibited first excitonic absorption maxima centered at 470 and 580 nm, respectively, with corresponding estimated diameters of 2.1 and 3.5 nm.³² NHS-CdSe QDs exhibited broad trap-state emission ranging from 520 to 850 nm, whereas 4-ATP-CdTe QDs exhibited narrow band-edge emission centered at 597 nm (Fig. 1). The broad trap-state emission of TGA-CdSe QDs and NHS-CdSe QDs is consistent with water-dispersible QDs, especially those small in diameter, exhibiting an increased number of surface trap-states.^33,34 It has been reported that the majority of the trapped carriers within CdSe QDs, in particular, are holes.³⁴ Trap-state emission, thus, arises from the recombination of conduction band electrons with holes in distributions of trap states, from shallow to more deeply trapped. Absorbance and emission spectra for CdTe QDs with varied surface functionalizations are shown in Fig. S1 of the supplementary material.

Mixed colloidal dispersions of NHS-CdSe QDs and 4-ATP-CdTe QDs, which undergo the formation of amide bonds between terminal function groups of QD-adsorbed ligands, in situ, to yield heterostructures, are hereafter referred to as “interacting CdSe-amide-CdTe mixtures.”^18,19 We previously reported that QD-adsorbed NHS-esters were converted in high yields (∼100%) to amides upon reaction with solvated amines¹⁹ and that the reaction of NHS-CdSe QDs with 4-ATP-CdTe QDs resulted in the incorporation of ∼80% of QDs into agglomerates consisting of 20 or more QDs with an interparticle separation of 3 nm or less.¹⁸ CdSe-amide-CdTe heterostructures undergo two photoinduced charge-transfer mechanisms: (1) electron transfer from CdTe to CdSe QDs and (2) hole transfer from CdSe to CdTe QDs.¹⁸ These results suggest that varying the molar ratio of 4-ATP-CdTe QDs to NHS-CdSe QDs, within interacting mixtures, provides a straightforward approach to tuning the average CdTe-to-CdSe ratio within heterostructures, enabling us to characterize the influence of the donor–acceptor ratio on charge-transfer dynamics. Experimentally determining the actual distribution of ratios of CdTe to CdSe QDs within heterostructures, or the average number of amide linkages between QDs, is non-trivial and beyond the scope of this article. In the research reported in this article, we instead focused on the influence of the CdTe-to-CdSe molar ratio, which is readily controlled experimentally, on the rate and efficiency of charge transfer. We focused primarily on excited-state electron transfer from CdTe to CdSe, as it is relatively straightforward to characterize by monitoring the narrow band-edge emission of CdTe QDs in steady-state and time-resolved measurements. In this electron-transfer mechanism, CdTe QDs are donors of electrons and CdSe QDs are acceptors. We controlled the ratio of donor-to-acceptor QDs within interacting CdSe-amide-CdTe mixtures by adding NHS-CdSe QDs, at varying concentrations, to dispersions of 4-ATP-CdTe QDs at a fixed concentration. The molar ratio of CdTe to CdSe within the resulting mixtures ranged from infinite (1-to-0 in control samples without CdSe) to 1-to-130.

Interpreting steady-state and time-resolved emission data can be challenging when both components of the desired heterostructure absorb light, to some degree, at the selected excitation wavelengths. Thus, we performed two control experiments to help interpret the data collected from interacting CdSe-amide-CdTe mixtures. First, we compared the steady-state emission spectra of NHS-CdSe QDs alone, at varying concentrations, to the spectra of mixtures of NHS-CdSe QDs and 4-ATP-CdTe QDs. The goal of this experiment was to assess (1) whether heterostructures were, indeed, assembling when the formation of amide bonds was possible and (2) the effects of heterostructure formation on charge transfer. The lower-energy region of the trap-state emission band of NHS-CdSe QDs, at wavelengths longer than 650 nm, does not overlap with the band-edge emission of CdTe QDs (Fig. 1); thus, focusing on excited-state CdSe-to-CdTe hole transfer was an appropriate initial gauge of heterostructure formation and subsequent charge transfer. Second, we compared steady-state and time-resolved emission data of interacting CdSe-amide-CdTe mixtures with those of mixtures of TP-CdTe QDs and NHS-CdSe QDs. Because thiophenol lacks an amine group, mixtures of TP-CdTe QDs and NHS-CdSe QDs cannot form covalently tethered, amide-bridged heterostructures. These mixed dispersions, hereafter referred to as “non-interacting CdSe-/-CdTe mixtures,” thus, enabled us to evaluate whether any emission quenching or charge transfer occurred in the absence of amide-bond formation between QD-adsorbed ligands.

In this initial control experiment, we compared steady-state emission spectra of interacting CdSe-amide-CdTe mixtures with spectra of dispersions of isolated NHS-CdSe QDs having the same concentrations of CdSe QDs as in the mixtures. Representative spectra at selected concentrations of CdSe QDs (1.35 and 13.5 µM) are shown in Fig. 2, and data for a full range of molar ratio comparisons are presented in Fig. S2 of the supplementary material. At the maximum peak of trap-state emission of CdSe QDs (750 nm), the inherent emission intensity of isolated NHS-CdSe QDs was quenched greatly, by 92% or more, in the presence of 4-ATP-CdTe QDs at a fixed concentration, independent of the initial NHS-CdSe concentration (black arrows in Fig. 2).

Additionally, band-edge emission from 4-ATP-CdTe QDs was quenched with increasing concentration of NHS-CdSe QDs (green arrow in Fig. 2). Taken together, the quenching of both band-edge emission from CdTe QDs and trap-state emission from CdSe QDs, thus, indicates that two charge-transfer mechanisms are active within the heterostructures: electron transfer from CdTe to CdSe QDs and hole transfer from CdSe to CdTe QDs as a result of heterostructure formation.

Band-edge emission from 4-ATP-CdTe QDs, within interacting CdSe-amide-CdTe mixtures, was quenched to an increasing extent as the concentration of NHS-CdSe QDs was systematically increased (Fig. 3). Steady-state emission spectra of interacting CdSe-amide-CdTe mixtures were obtained following excitation at 445 nm, which is absorbed by both CdSe and CdTe QDs (see Fig. S3 of the supplementary material); thus, both electron transfer and competitive absorption by CdSe have the potential to decrease the intensity of emission from CdTe QDs. To correct for the possibility of competitive absorption, as the primary focus of this article is excited-state electron transfer, we calculated the extent of emission quenching expected to result from competitive absorption, as described in Appendix S2 of the supplementary material. After accounting for the effects of competitive absorption throughout all CdTe-to-CdSe molar ratios, the percent quenching (%Q) of band-edge emission from CdTe QDs increased with the concentration of NHS-CdSe QDs. The %Q ranged from 8.4%, for interacting CdSe-amide-CdTe mixtures with a 1-to-1 CdTe-to-CdSe molar ratio, to 76% for mixtures with a 1-to-130 molar ratio.

Trap-state emission from NHS-CdSe QDs was measurable only for the interacting CdSe-amide-CdTe mixtures with the highest concentrations of NHS-CdSe QDs (lowest CdTe-to-CdSe molar ratios). The initial absence of NHS-CdSe emission, for mixtures with relatively low concentrations of CdSe QDs, was attributed to the quenching of emission by excited-state hole transfer from CdSe to CdTe QDs.¹⁸ The subsequent (8-fold) growth of trap-state emission from CdSe QDs, within interacting CdSe-amide-CdTe mixtures with the highest CdSe concentrations, is consistent with the presence of an excess of NHS-CdSe QDs that were not incorporated into heterostructures and further suggests that surfaces of 4-ATP-CdTe QDs were ultimately saturated with NHS-CdSe QDs within interacting mixtures with the highest concentrations of CdSe.

Since our primary focus is CdTe-to-CdSe electron transfer, it would have been preferable to probe band-edge emission from CdTe QDs following selective excitation of CdTe without any competitive absorption by CdSe. In an effort to selectively excite 4-ATP-CdTe QDs, we also acquired steady-state emission spectra with a longer excitation wavelength. In order to observe any band-edge emission from CdTe QDs of this size, the longest excitation wavelength accessible was 540 nm due to the minimal Stokes shift (Fig. S1). The absorbance of NHS-CdSe QDs at 540 nm (0.34 eV lower in energy relative to the first excitonic absorption maximum) is only 3.3% of the maximum absorbance at 470 nm (Fig. 1).

Following excitation at 540 nm, CdTe-to-CdSe-ratio-dependent quenching was once again observed throughout the band-edge emission of 4-ATP-CdTe QDs. Values of %Q, after correction for competitive absorption, ranged from 8.4% for a 1-to-1 CdTe-to-CdSe molar ratio to ∼79% for mixtures with a 1-to-130 molar ratio (see Fig. S4 of the supplementary material). The contribution from competitive absorption was much less significant with excitation at 540 nm, but we, nonetheless, corrected for it. In these measurements, trap-state emission from NHS-CdSe QDs increased by approximately 5-fold for the interacting CdSe-amide-CdTe mixture with the highest concentration of CdSe QDs (Fig. S4), again suggesting that these mixtures contained isolated NHS-CdSe QDs. In summary, we attribute the nearly 10-fold net increase in quenching of band-edge emission from 4-ATP-CdTe QDs, following excitation at either 445 or 540 nm, to excited-state electron transfer from CdTe to CdSe QDs, the efficiency of which varies with the molar ratio of donor and acceptor QDs.

In our second control experiment, non-interacting CdSe-/-CdTe mixtures were characterized. We hypothesized that the photophysical properties of non-interacting mixtures would exhibit features of the two individual QD components, which to a first approximation should behave independently of one another within in situ mixtures. This control experiment can, thus, provide additional insight into (1) whether NHS-CdSe and 4-ATP-CdTe QDs, indeed, interact with each other and (2) the mechanism by which trap-state emission from NHS-CdSe QDs is quenched within interacting CdSe-amide-CdTe mixtures.

In this second control experiment, either 4-ATP-CdTe or TP-CdTe QDs were combined with NHS-CdSe QDs to yield interacting CdSe-amide-CdTe mixtures or non-interacting CdSe-/-CdTe mixtures, respectively. The difference in energy between the normalized CdTe band-edge emission bands of interacting CdSe-amide-CdTe and non-interacting CdSe-/-CdTe mixtures, both with 1-to-3 CdTe-to-CdSe molar ratio, arises from slight differences in the sizes and surface functionalizations of CdTe QDs (Fig. 4).⁵ Following ligand exchange on Cl-CdTe QDs, the band-edgeemission maximum shifted by 21 meV to higher energy for TP-CdTe QDs and by 14 meV to lower energy for 4-ATP-CdTe QDs, consistent with previous reports.¹⁸ The emission spectra of dispersions of isolated 4-ATP-CdTe QDs and interacting CdSe-amide-CdTe mixtures were superimposable (see Fig. S5 of the supplementary material); therefore, the shift of the emission spectrum of interacting CdSe-amide-CdTe mixtures, relative to the spectrum of non-interacting CdSe-/-CdTe mixtures, resulted not from the formation of amide bonds, but rather from differences in surface functionalization. Steady-state emission spectra of non-interacting CdSe-/-CdTe mixtures exhibited distinct emission features of both TP-CdTe and NHS-CdSe QDs (Fig. 4). We attribute the presence of trap-state emission from CdSe QDs in the spectra of non-interacting CdSe-/-CdTe mixtures to the lack of heterostructure formation. In contrast, the quenching of NHS-CdSe-derived trap-state emission within the spectra of interacting CdSe-amide-CdTe mixtures (Fig. 4) is logically attributed to excited-state hole transfer from CdSe to CdTe QDs.¹⁸ The marked differences in emission spectra (quenching of emission only when the formation of amide bonds is possible), thus, further confirm that NHS-CdSe QDs interact with 4-ATP-CdTe QDs, presumably through amide bond formation as previously characterized.^18,19

Time-resolved emission decay traces were acquired for interacting CdSe-amide-CdTe and non-interacting CdSe-/-CdTe mixtures at wavelengths throughout the band-edge emission of CdTe QDs. By comparing interacting CdSe-amide-CdTe mixtures (containing 4-ATP-CdTe QDs) with non-interacting CdSe-/-CdTe mixtures (containing TP-CdTe QDs), we aimed to isolate the effects of excited-state electron transfer on emission-decay dynamics for CdTe QDs. Measured emission-decay traces were fit to multiexponential decay kinetics to determine intensity-weighted average excited-state lifetimes (⟨τ⟩) as described in Appendix S3 of the supplementary material. Highest-quality fits were obtained using triexponential kinetics for interacting CdSe-/-CdTe mixtures and tetraexponential kinetics for non-interacting CdSe-/-CdTe mixtures. Parameters from multiexponential fits are listed in Table S3 of the supplementary material. Values of ⟨τ⟩ for isolated NHS-CdSe, 4-ATP-CdTe, and TP-CdTe QDs were (29 ± 3) ns, (19.4 ± 0.7) ns, and (8.9 ± 0.7) ns, respectively.

Band-edge emission from 4-ATP-CdTe QDs within interacting CdSe-amide-CdTe mixtures decayed more rapidly than emission from isolated 4-ATP-CdTe QDs [Fig. 5(a)]. This dynamic quenching of emission (accelerated excited-state decay) is consistent with a mechanism in which excited-state electron transfer from CdTe to CdSe QDs, within heterostructures, outcompetes electron–hole recombination. Importantly, for the interacting CdSe-amide-CdTe mixtures, the average excited-state lifetime of band-edge emission from CdTe QDs continued to decrease relative to that of isolated 4-ATP-CdTe QDs as the molar ratio of CdSe (acceptor) to CdTe (donor) increased, despite the increasing contribution from the longer-lived component (NHS-CdSe QDs) [Fig. 5(c)]. This result suggests that the rate of excited-state electron transfer from CdTe to CdSe QDs increased as more CdSe QDs, on average, were interfaced with each CdTe QD within the tethered amide-bridged heterostructures. Previous control experiments using 4-(methylthio)aniline in place of 4-ATP confirmed that emission quenching within interacting CdSe-amide-CdTe mixtures was not simply the result of amide bond formation, but rather a direct result of charge transfer between QDs via a super exchange mechanism.¹⁸ In contrast, for TP-CdTe QDs within non-interacting CdSe-/-CdTe control mixtures, there was no evidence of dynamic quenching of band-edge emission. As the concentration of NHS-CdSe QDs increased, the average excited-state lifetime continued to increase until the emission decay traces ultimately became dominated by the longer-lived trap-state emission from NHS-CdSe QDs [Figs. 5(b) and 5(c)]. The lack of dynamic quenching of band-edge emission from TP-CdTe QDs within non-interacting CdSe-/-CdTe mixtures is consistent with our expectation that the CdTe and CdSe QDs would behave independently within such mixtures.

Linear combinations of time-resolved emission decay traces of (1) isolated 4-ATP-CdTe and NHS-CdSe QDs or (2) isolated TP-CdTe and NHS-CdSe QDs were constructed and compared to the experimental decay traces of interacting CdSe-amide-CdTe and non-interacting CdSe-/-CdTe mixtures, respectively, for a given CdTe-to-CdSe ratio [see Figs. 5(d) and S6 of the supplementary material]. The decay traces of non-interacting CdSe-/-CdTe mixtures were well-modeled by a linear combination of the decay traces of isolated QD components, further indicating that the CdSe and CdTe QDs within these mixtures, indeed, behaved independently. In contrast, the measured time-resolved emission decay traces of interacting CdSe-amide-CdTe mixtures could not be modeled accurately by linear combinations of decay traces of isolated QDs. The measured decay traces decayed much more rapidly than the constructed linear combinations. The lack of correlation between the linear combination of the isolated QD components and the measured decay traces of the interacting CdSe-amide-CdTe mixtures indicates clearly that an additional excited-state deactivation mechanism is active for CdTe QDs within the interacting CdSe-amide-CdTe mixtures, which we assign as excited-state electron transfer from CdTe QDs to CdSe QDs.

Plots of residuals as a function of time represent the goodness of the fits (linear combinations) to the experimental time-resolved decay traces. Residual plots were constructed by subtracting the linear combinations from the measured time-resolved emission data (see Fig. S7 of the supplementary material). The residual plots likewise indicate that an alternative excited-state decay mechanism (charge transfer) is active within the interacting CdSe-amide-CdTe mixtures. Notably, as the molar ratio of CdSe (acceptor) to CdTe (donor) QDs increases, the residuals deviate further from zero for interacting CdSe-amide-CdTe mixtures, indicating an increase in the rate and extent of excited-state electron transfer.

For interacting CdSe-amide-CdTe mixtures, rate constants of electron transfer (k_et) were estimated as the difference between the average rate constant of excited-state decay, calculated as the fractional emission intensity-weighted harmonic average of individual rate constants from multiexponential fits, of interacting CdSe-amide-CdTe mixtures (⟨k⟩_{CdSe-amide-CdTe}) and free dispersed 4-ATP-CdTe QDs (⟨k⟩_4-ATP-CdTe) (see Table S4 of the supplementary material),

Equation (1) is derived in Appendix S4 of the supplementary material. This approach is equivalent to calculating the values of k_et as the difference between the reciprocals of intensity-weighted average lifetimes of CdSe-amide-CdTe mixtures (⟨τ⟩_{CdSe-amide-CdTe}) and free dispersed 4-ATP-CdTe QDs (⟨τ⟩_4-ATP-CdTe) and is consistent with previously reported calculations of estimated charge-transfer rate constants in related QD-containing donor–acceptor constructs.^18,35,36 Electron-transfer rate constants were estimated under two working assumptions: (1) that excited-state electron transfer is the only additional excited-state deactivation pathway for interacting CdSe-amide-CdTe mixtures relative to 4-ATP-CdTe QDs alone and (2) that all other rate constants are equal for interacting CdSe-amide-CdTe mixtures and 4-ATP-CdTe QDs alone.

Efficiencies of electron transfer (η_et) were calculated as follows:^37,38

where A_i and τ_i are pre-exponential factors and lifetimes from triexponential fits and the summation of the product of A_i and τ_i over all three components from triexponential fits corresponds to the integrated time-resolved emission intensity.

The estimated rate constants (k_et) of electron transfer were on the order of 10⁷ s⁻¹ for interacting CdSe-amide-CdTe mixtures with CdTe-to-CdSe ratios beyond 1-to-10, with efficiencies (η_et) of up to ∼91%. Remarkably, the estimated value of k_et, for excited-state electron transfer from CdTe to CdSe QDs, increased by 100-fold as the molar ratio of CdSe to CdTe increased; the corresponding value of η_et increased by 7.4-fold, from just (12 ± 3)% at a 1-to-1 CdTe-to-CdSe ratio to (91 ± 8)% at a 1-to-130 ratio (Fig. 6). The electron-transfer efficiency could, thus, be tuned from nearly zero to nearly unity by adjusting the donor–acceptor ratio. We attribute these increases of the rate and efficiency of electron transfer to an increase in the average number of CdSe QDs per CdTe QD within amide-bridged heterostructures. At sufficiently high concentrations of NHS-CdSe QDs within interacting CdSe-amide-CdTe mixtures, as the molar ratio of CdTe-to-CdSe decreased below 1-to-35, both k_et and η_et plateaued (Fig. 6). We attribute the plateau to the formation of heterostructures with the maximum number of interfaced NHS-CdSe QDs per 4-ATP-CdTe QD such that any additional NHS-CdSe QDs in the mixtures remained free in dispersion. This interpretation is consistent with the growth of trap-state emission from NHS-CdSe QDs within interacting CdSe-amide-CdTe mixtures having the highest CdTe-to-CdSe molar ratios (Fig. 3). If we assume that k_et and η_et increase monotonically with the number of electron-accepting CdSe QDs interfaced with a given CdTe QD, then these data imply that the upper limit of the ratio of CdSe-to-CdTe within heterostructures is ∼35-to-1. Notably, our results indicate that the extent of carbodiimide-mediated coupling between QDs, the relative number of constituent QDs within heterostructures, and the rate and efficiency of photoinduced inter-QD electron transfer can be systematically controlled with the molar ratio of dispersed donor and acceptor QDs.

The observed 100-fold variation of k_et, for our CdSe-amide-CdTe heterostructures, with the donor-to-acceptor ratio is analogous to previous reports on QD-molecular acceptor systems.¹⁰ Alivisatos and co-workers reported hole-transfer rate constants on the order of 10³–10⁷ s⁻¹ for three different ferrocene derivatives covalently linked to CdSe/CdS core/shell QDs.²⁴ Similarly, Bang and Kamat reported that the rate constants of electron transfer between CdSe QDs and an adsorbed fullerene derivative ranged from 10⁹ to 10¹⁰ s⁻¹ but also emphasized that fast charge recombination negatively affected the efficiency of the system, regardless of the rate of electron transfer.³⁹ 

Our reported type-II dual-QD CdSe-amide-CdTe heterostructures, with rate constants of electron transfer ranging from 10⁶ to 10⁷ s⁻¹, are particularly intriguing architectures for light harvesting, excited-state charge transfer, and subsequent redox photocatalysis. They undergo two active charge-transfer mechanisms, both of which further spatially separate charge carriers and decrease the rate of charge recombination, and QDs within heterostructures can accumulate multiple charge carriers potentially enabling multi-electron redox catalysis.

We selectively attached NHS-CdSe QDs to 4-ATP-CdTe QDs to yield covalently tethered heterostructures with CdTe-to-CdSe molar ratios ranging from 1-to-0 to 1-to-130. Heterostructures were synthesized via amide bond formation between terminal functional groups of QD-adsorbed ligands. Bidirectional charge transfer occurs within interacting CdSe-amide-CdTe mixtures. We focused primarily on excited-state electron transfer from CdTe to CdSe. Steady-state and time-resolved emission measurements revealed an increase of the extent of dynamic quenching of band-edge emission from 4-ATP-CdTe QDs with an increasing molar ratio of NHS-CdSe QDs to 4-ATP-CdTe QDs within interacting CdSe-amide-CdTe mixtures. Following excitation at 540 nm, band-edge emission from CdTe QDs was quenched by 79%, and the intensity-weighted average lifetime decreased by 39%. Dynamic quenching of emission did not arise from competitive absorption, but rather can logically be attributed to excited-state electron transfer from CdTe to CdSe QDs. In contrast, steady-state and time-resolved emission measurements of non-interacting CdSe-/-CdTe mixtures containing TP-CdTe and NHS-CdSe QDs exhibited features characteristic of the two independent, non-interacting QD components due to the lack of formation of amide bonds or heterostructures. The result of this control experiment thus supports our attribution of the dynamic quenching for interacting CdSe-amide-CdTe mixtures to inter-QD electron transfer.

Carbodiimide-mediated coupling chemistry enables systematic control over the donor-to-acceptor ratio within dual-QD heterostructures. As a result, charge-transfer dynamics and yields could be tuned with the concentrations of electron-accepting CdSe QDs and electron-donating CdTe QDs. Estimated values of k_et varied by 100-fold over the range of CdTe-to-CdSe molar ratios evaluated and were on the order of 10⁷ s⁻¹ for interacting mixtures with the highest concentrations of CdSe relative to CdTe. Corresponding electron-transfer efficiencies varied by 7.4-fold and reached 91% for interacting mixtures with molar ratios of CdSe to CdTe QDs of 35-to-1 and higher. Notably, these results reveal that, similarly to QD-molecule donor–acceptor systems, the rate and efficiency of excited-state charge transfer within the CdSe-amide-CdTe heterostructures can be tuned systematically with the average number of electron-accepting QDs per electron-donating QD. Specifically, the rate and efficiency of CdTe-to-CdSe electron transfer can be maximized by increasing the number of CdSe QDs per CdTe QD, up to a threshold ratio of ∼35-to-1, beyond which the electron-donating CdTe QDs cannot accommodate additional electron-accepting CdSe QDs. Photogenerated electrons can be extracted most rapidly and efficiently from CdTe QDs at this threshold molar ratio. Understanding the influence of the donor–acceptor ratio on charge-transfer dynamics enables the preparation of heterostructures with optimal excited-state charge-transfer reactivity. Given this potential to control excited-state charge-transfer dynamics and yields, these QD heterostructure systems are intriguing for redox photocatalysis.

See the supplementary material for Figs. S1–S7, Appendixes S1–S4, and Tables S1–S4: ground-state absorbance spectra, steady-state emission spectra, and time-resolved emission spectra and fits and residuals for isolated QDs, interacting CdSe-amide-CdTe mixtures, and non-interacting CdSe-/-CdTe mixtures; and materials, reported synthetic methods, spectroscopic characterization methods, competitive absorbance calculations, derivation to estimate rate constants of electron transfer, and kinetic modeling and fitting parameters of time-resolved emission data.

This work was supported by the National Science Foundation under Grant No. CHE-1306784. Additional support was provided by the University at Buffalo.

The authors have no conflicts to disclose.

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