People: Shichen Lian, Kevin McClelland, Dana Westmoreland, Zhengyi Zhang, Yishu Jiang
Quantum Dots as Photosensitizers
Due to their high photostability, their tunable absorption spectrum, large extinction coefficients (>106), and high PL quantum yields, semiconductor quantum dots (QDs) have the potential of being used as photosensitizers in photocatalysis. The goal is that a photo-excited QD transfers its photo-generated carriers to a redox-active catalyst that can then perform a useful reaction. Examples include hole transfer to a water oxidation catalyst or electron transfer to proton reduction catalyst to make molecular hydrogen as a solar fuel. However, many reactions that produce, or facilitate the production of, fuels require multiple reductions (or oxidations) of the substrate and have large kinetic overpotentials that limit the efficiencies of thermodynamically feasible reactions. We believe that QDs used in combination with molecular catalysts can be used to efficiently generate solar fuels as well as understand the fundamental processes entailed in the sensitization of the catalyst. Our goal is to take advantage of the large surface area of QDs and the chemical tunability of their cores and their surfaces to create efficient QD-Catalyst complexes to inform the creation of the next-generation of catalysts.
Quantum dots for photo-reduction of CO2 .
Photosensitization of molecular catalysts to reduce CO2 to CO is a sustainable and desireable route to production of solar fuels. Crucial to the sensitization process is the highly efficient transfer of multiple redox equivalents from sensitizer to catalyst. The highly-defective/disordered surfaces of QDs can serve as potential scaffolds for the binding of molecular catalysts without the use of covalent chemistry.
Here we demonstrate that the photosensitization of a model CO2 reduction catalyst, meso-tetraphenylporphyrin iron(III) chloride (FeTPP) catalyst is achieved by colloidal, heavy metal-free CuInS2/ZnS quantum dots (QDs) to reduce CO2 to CO using 450 nm light. The sensitization efficiency (turnover number per absorbed unit of photon energy) of the QD system is a factor of 18 greater than that of an analogous system with a fac-tris(2-phenylpyridine)iridium sensitizer. This high efficiency originates in ultrafast electron transfer between the QD and FeTPP, enabled by formation of QD/FeTPP complexes. Optical spectroscopy reveals that the electron-transfer processes primarily responsible for the first two sensitization steps (FeIIITPP → FeIITPP, and FeIITPP → FeI TPP) both occur in 200 fs.
However, photocatalysis in water (the ideal solvent for the production of solar fuels) introduces additional difficulties such as the self-aggregation of QDs. We have further optimized the earlier-shown QD-catalyst system with QDs that are capped with negatively charge ligands (3-mercaptopropionic acid, MPA) and the catalysts are functionalized with a positively charged amine group. Adding potassium modulates the electrostatic interaction between the QD sensitizers and catalysts and enhances the overall catalytic efficiency.
Quantum Dots Serving as Photocatalysts.
A desirable property of QDs as photosensitizers or as direct catalysts is their tunable surface chemistry. QDs have large potentially catalytic surfaces (as compared to bulk materials) with the ability to bind to multiple substrate molecules (whereas the typical homogeneous catalyst can only bind one or two). We focus primarily upon functionalizing QD surfaces to bind substrates in reactive geometries with decrease overpotentials.
QD Surfaces can bind substrates with high-affinity
Here, we demonstrated that CdS QDs photocatalyze the reduction of protonated nitrobenzene ([NB H]+) to aniline ([AN H]+) through six sequential proton coupled electron transfer steps, with nitrosobenzene ([NSB H]+) and phenylhydroxylamine ([PHA H]+) intermediates, and using 3-mercaptopropionic acid (MPA) and MeOH as proton and terminal electron sources. Oxidation of MPA produces the corresponding disulfide (“MPA-MPA”) and serves as the primary sacrificial electron donor. Nuclear magnetic resonance (NMR) experiments showed that the reagents were adsorbed to the QD surface throughout the catalytic cycle such that electron transfer from the QD never became diffusion-controlled (on the order of microseconds). This allows us to analyze measured rate constants for each catalytic step in terms of energetic driving force and kinetic barriers for the reaction.
Catalytic cycle for the 6-electron, 6-proton photoreduction of nitrobenzene (NB, 1) to aniline (AN) through nitrosobenzene (NSB, 2) and phenylhydroxylamine (PHA, 3) 2-electron intermediates, all of which are partially protonated at pH 3.6. 1a, 2a, and 3a denote proposed 1-electron intermediates formed during the cycle.
Ligand-exchange can increase the catalytic surface area
One advantage of QDs as photocatalysts is that simple ligand exchanges can tune the structure of the particle surface and increase the efficacy for substrate binding with minimal changes in the electronic or redox properties of the QD core. Ligand exchange can be performed to change the permeability of the ligand shell, its hydrophobicity, or its charge density to optimize interactions with the QD and the substrate.
Conditions and catalytic cycle for the coupling reaction of 1-phenyl pyrrolidine, 1, and phenyl trans-styryl sulfone, 3. The electrochemical potentials are +0.70 V vs. SCE for oxidation of PhPyr, 1, and +0.50 V vs. SCE for the reduction of PhSO2· (estimated from the PhSO2·/PhSO2Na pair). The estimated driving forces are -1.2 eV for electron transfer and -1.6 eV for hole transfer.
A) Fraction of emissive QDs that remains emissive after addition of PhPyr (PL/PL0) vs. eq. of PhPyr, for 1 μM QDs pre-treated with different eq. of OPA. All samples were stirred in the dark, under N2, for 8-10 h prior to measurement. Decreasing PL/PL0 corresponds to increasing yield of photooxidation of PhPyr by the QDs. B) GC-determined concentration of product, 5, vs. time for the first 15 minutes of illumination for the reaction in Scheme 1A, using QDs pre-treated with different eq. of OPA. C) Concentration of product, 5, vs. cumulative energy absorbed and vs. illumination time for the reaction in Scheme 1A, using QDs with their native oleate ligands (“0 eq. OPA”, black) or pre-treated with 250 eq. of OPA (red). All samples were purged with Ar(g) and illuminated with 13-mW 405-nm laser
In one of the very few examples of QD-catalyzed C-C coupling, we adjusted the surface chemistry of CdS QDs to tune their photocatalytic with regard to a C-C coupling reaction, the addition of 1-phenyl pyrrolidine and phenyl trans-styryl sulfone to form 1-phenyl-2-(2-phenylethenyl) pyrrolidine. A partial ligand exchange of the QDs’ native ligands, oleate, for octylphosphonate increase the disorder of the QD surface. We suspect the increased surface disorder lead to an increase area for substrate binding, which was supported by PL measurements, and an increased turnover rate for the system.
Ligand-exchange can also induce enantioselective C-C coupling.
We can use CdS QDs to enantioselective photocatalyze the classical α-alkylation of aldehydes, where QDs can not only serve as the photosensitizers but also provide steric hindrance for inducing chiral selectivity for the reaction. Compared to previously reported photocatalytic systems employing molecular dye sensitizers and separated enantioselective catalysts, the QD photocatalyst we have developed has three advantages:
1) the highly absorbing QDs replace previously used precious metal complexes, such as Ru(bpy).
2) the on-site enantio-inducing ligands minimize the diffusion distance for the intermediates, and hence improve the reactivity and can reduce the loading of enantioselective catalysts by two orders of magnitude (as compared to other systems in the literature)
3) the facile ligand exchange procedure leads to easier controls of enantiofacial discrimination as compared to organocatalysts that require extensive synthetic efforts..
QD-ligand systems can serve as probe for the local proton concentration for PCET reactions
As we have shown (and others have reported), QDs are efficient photocatalysts for proton-coupled electron transfer (PCET) reactions that require one or multiple photo-redox equivalents, provided that the substrate is close enough to the QD to undergo charge transfer. Tracking the kinetics of charge transfer between the QD and the substrate is easily achieve using standard time-resolved spectroscopy such as transient absorption, but tracking the kinetics of proton transfer is more analytically challenging. We have shown previously that the protonation state of phosphonate ligands affects the first exciton absorbance energy of the quantum dot. By calibrating the degree of the first exciton absorbance energy shift to the exact number of protons proximate to the surface of the quantum dot, we can develop a QD-based local proton sensor that functions in polar aprotic solvents (in which proton-coupled electron transfer reactions are typically performed). The first exciton absorbance energy of the quantum dot can then serve as analytical probe of the surface-proximate protons available for participation in quantum dot-catalyzed proton-coupled electron transfer reactions and provide insights into the timescale of proton transfer in these systems.
Tuning the size of QDs to modify the selectivity
We can also modify the size of the QDs as a method of modifying the selectivity of the QDs. Since QDs are subject to “quantum confinement effects” that increase the energies of the conduction and valence band of the QDs (as compared to bulk semiconductors), we can easily tune the energy levels of the QDs by modifying the size. This allows us to change the reactivity of the QDs for different photoredox-based reactions.
Below, we demonstrate the photocatalytic oxidation of benzyl alcohol with cadmium sulfide quantum dots (CdS QDs), and control over the selectivity of this process for either a two-electron oxidation of one alcohol molecule to benzaldehyde or two one-electron oxidations of two alcohol molecules followed by carbon-carbon coupling to hydrobenzoin. Product distributions are controlled by changing the QD:alcohol ratio and the size of the QD and result in over 80% selectivity for either pathway.