Low-Dimensional Semiconductors in Artificial Photosynthesis: An Outlook for the Interactions between Particles/Quasiparticles.

By virtue of their intriguing electronic structures and excellent surface properties, low-dimensional semiconductors hold great promise in the field of solar-driven artificial photosynthesis. However, owing to promoted structural confinement and reduced Coulomb screening, remarkable interactions between particles/quasiparticles, including electrons, holes, phonons, and excitons, can be expected in low-dimensional semiconductors, which endow the systems with distinctive excited-state properties that are distinctly different from those in the bulk counterparts. Consequently, these interactions determine not only the mechanisms but also quantum yields of photosynthetic energy utilization. In this Outlook, we review recent advances in studying the unique interactions in low-dimensional semiconductor-based photocatalysts. By highlighting the relevance of different interactions to excited-state properties, we describe the impacts of the interactions on photosynthetic energy conversion. Furthermore, we summarize the regulation of these interactions for gaining optimized photosynthetic behaviors, where the relationships between these interactions and structural factors/external fields are elaborated. Additionally, the challenges and opportunities in studying the interaction-related photosynthesis are discussed.

Introduction

Photosynthesis, which plays a crucial role in balancing ecological cycles of materials and energy, displays an excellent prototypical model for converting solar energy to chemical energy. Inspired by this process, numerous efforts have been devoted to artificial photosynthesis for pursuing solar energy utilization for decades.[1−8] Analogous to chloroplasts (in plants) or light-harvesting pigments (in some bacteria) in natural photosynthesis, catalysts in artificial photosynthesis act as a light antenna, and the transfer of energetic photoinduced species in catalysts toward substrate molecules is responsible for the relevant energy conversion. In terms of designing advanced photocatalysts, many light-harvesting materials have been proposed for their potentials in artificial photosynthesis, which can be mainly divided into two kinds: atomic/molecular systems (e.g., ions, organic species, metal complexes, etc.) and solid-state systems (mainly, noble metals and semiconductors). The two kinds of photocatalysts have their own merits and demerits in aspects like material source, molar extinction coefficient, surface property, chemical stability, biotoxicity, and so on. Besides, a major difference lies in their excited-state properties, where the different configurations and kinetics of photoinduced species result in quite distinct behaviors in certain photosynthetic reactions. Herein, exploring advanced photocatalysts for combining the advantages of both molecular and solid-state systems would be meaningful to the pursuit of plentiful solar-driven artificial photosynthesis.

Owing to their impressive surface and electronic structures, low-dimensional semiconductors have been attracting tremendous attention in the field of artificial photosynthesis.[9−12] As compared to their bulk counterparts, low-dimensional semiconductors tend to possess smaller sizes and higher specific surface areas, which facilitates reactive-site exposure and structural modifications. In addition, low-dimensional semiconductors possess novel electronic structures, where structural confinement leads to tunable size-dependent excited-state properties. The most exciting feature is that low-dimensional semiconductors appear to display some molecule-like excited-state configurations and kinetics, due to the promoted interactions between particle/quasiparticles (e.g., electrons, holes, excitons, phonon, etc.) induced by reduced dimension and Coulomb screening.[12] All these features make low-dimensional semiconductors excellent platforms combining the merits of both molecular and solid-state photocatalysts for achieving versatile photosynthetic energy utilization.

Excited-State Properties in Low-Dimensional Semiconductors

As for photosynthetic conversion of solar energy to chemical energy, the effective photoexcitation of catalysts is required. Upon illumination by light with suitable photon energy, reactive photoinduced species acting as energy carriers are generated in catalysts, followed by transferring toward substrate molecules. With regard to photosynthesis, excited-state properties such as excitation, relaxation, and recombination of these photoinduced species are vital to the relevant performance. However, owing to their solid-state features but strong confined structures, low-dimensional semiconductors exhibit fairly unique electronic states that are different from those of molecular and bulk systems. In fact, this is the fundamental cause of the rich potential of low-dimensional semiconductors in photosynthesis. In this section, we carry out a brief discussion on the excited-stated properties of low-dimensional semiconductors.

Due to their different chemical components and structures, molecular and solid-state photocatalysts possess different electronic configurations, which would lead to diverse photosynthetic behaviors. As for molecular catalysts, a discrete energy-level diagram based on molecular orbital theory is generally employed to describe the involved excited states (Figure a). Molecular orbital theory is a delocalized bonding approach, where electrons are deemed to interact with all nuclei, rather than the isolated nucleus in the system. That is, molecular orbitals, arising from the hybridization (namely, bonding and antibonding combinations) of atomic orbitals from every atom, distribute all over the molecular system, and the accommodation of electrons to these discrete (or quantized) molecular orbitals leads to energy reduction of the system. In consideration of the Coulomb and exchange interactions, the excited molecule can be regarded as a bound electron–hole state, or so-called molecular exciton.[13] As for bulk semiconductor-based photocatalysts, excited-state properties are much more complicated as compared to those of molecular systems. Typically, band theory is used to depict the involved excited states (Figure b), and a periodic potential resulting from the periodic crystal structure is proposed to simplify the electronic behaviors. In fact, semiconductors can be considered as extended, strong-couple molecular systems, where an infinite number of frontier occupied and unoccupied molecular orbitals which are closely localized in energy space constitute their valence and conduction bands, respectively. The energy separation between these molecular orbitals (in the order of 10–22–10–23 eV) is much smaller than the thermal energy (e.g., 26 meV at room temperature), which means that electronic transitions between these close-localization molecular orbitals can be easily provoked by exchanging energy between charge carriers and thermal vibrations.[14] Band theory is only an approximation, and the complicated interactions between different species including charge carriers, and excitons, are fully considered. Different electronic structures can be expected so long as these interactions are considered. In fact, it is the major factor that is responsible for the characteristic excited-state properties of low-dimensional semiconductors.

Figure 1

Schematic illustrations of (a) electronic energy levels of molecular systems, and band structures of (b) bulk and (c) low-dimensional semiconductors. LUMO, HOMO, CB, and VB denote lowest unoccupied molecular orbital, highest occupied molecular orbital, conduction band, and valence band, respectively.

As compared to their bulk counterparts, low-dimensional semiconductors possess remarkably promoted interactions between different species due to promoted structural confinement and reduced dielectric screening. These interactions endow low-dimensional systems with nontrivial excited-state properties that could hardly be observed in bulk systems. For instance, owing to size/dimension reduction, quantum confinement effects can greatly impact the band structures of low-dimensional semiconductors. That is, the hybridization of a finite number of atomic orbitals in low-dimensional semiconductors leads to a finite density of states, which thus gives rise to discrete energy levels and blue-shifted band gaps. Besides, the reduced dimension also leads to limited electronic screening in low-dimensional systems, thus resulting in promoted Coulomb interactions between electrons that can greatly impact the band structures.[15−18] As a typical result, size-/thickness-dependent band structures can be widely observed in low-dimensional semiconductors.[17,18] Moreover, as for the symmetry breaking in low-dimensional semiconductors, the potential anisotropic dispersions of electronic structures can further promote electron–electron interactions in particular orientations, thus leading to some anisotropic electronic properties.[18] In addition to the above band-structure changes, strong interactions between electrons also impact the other excited-state properties like transport behaviors and charge-carrier relaxations.[19,20] Another remarkable feature in low-dimensional semiconductors is the promoted Coulomb interactions between photoinduced electrons and holes.[21,22] In semiconductors, photoexcitation gives rise to the formation of electrons and holes in conduction and valence bands, respectively, whereas the typically high dielectric properties in bulk systems result in faint Coulomb interactions between these charged species. However, screening reduction effectively rouses such interactions and hence leads to robust excitonic effects in low-dimensional semiconductors (Figure c). As a consequence, excitons (or bound electron–hole pairs) can be expected in the systems. The energy difference between quasiparticle band gap and lowest excitonic energy level is called exciton binding energy (Eb), which can be used to evaluate the strength of excitonic effects. In addition, there are also other kinds of interactions such as electron–phonon, electron–exciton, and exciton–exciton interactions in low-dimensional semiconductors that can impact both the mechanism and efficiency of solar energy utilization (further discussion will be given later). All these interactions permit low-dimensional semiconductors to possess excited-state properties with both solid-state and molecular features, which pave the way for achieving diverse photosynthetic energy utilizations.

Impacts of the Interactions on Photosynthesis

Electron–Electron Interactions

Owing to reduced screening effects, promoted Coulomb interactions between electrons emerge in low-dimensional semiconductors. Given that electronic bands of semiconductors originate from the interplay of the electrons and nucleus, the promoted interactions between electrons lead to renormalized electronic energy levels and elevated quasiparticle band gaps. A characteristic change resulting from the promoted interactions is the size-/thickness-dependent band structures. For instance, layer-dependent band structures of two-dimensional transition metal dichalcogenides (TMDs) have been theoretically and experimentally demonstrated.[23,24] Besides, the electron–electron interaction-induced band structure renormalization widely exists in these systems, and even low-concentration electron doping can lead to giant band gap reduction (Figure a).[25,26] Since that band gap and band-edge energy levels determine the light-responsive region and charge-carrier redox capacity, respectively, the above features in low-dimensional semiconductors enable photosynthetic behaviors to be tuned by precise size/dimension control and chemical doping. Moreover, as for some low-dimensional semiconductors with unique structural factors, symmetry breaking and dimension reduction lead to anisotropic electronic band structures.[18,27,28] This anisotropic feature can give rise to pretty distinctive photosynthetic behaviors. Taking two-dimensional black phosphorus as an example, both theoretical and experimental studies confirm the highly anisotropic optical responses in the system, where the quasi-one-dimensional dispersions of electronic structures associated with the anisotropic in-plane crystal structure are crucial (Figure b).[18,28] The extraordinary band dispersions and strong electron–electron interactions induce pronounced self-energy corrections, which are responsible for the enlarged band gap and the presence of van Hove singularities. The van Hove singularities lead to nontrivial sub-band structures and therefore excitation-energy-dependent optical response in this two-dimensional system. Our group recently demonstrated that the nontrivial sub-band structures in two-dimensional black phosphorus enable optically switchable photocatalytic oxygen activation, where visible- and ultraviolet-light illuminations lead to the selective generations of singlet oxygen and hydroxyl radical, respectively (Figure c,d).[29]

Figure 2

(a) Electron doping concentration-dependent band gap evolutions, where Econ (red squares) and Eg (blue dots) denote onset energy and quasiparticle band gap, respectively. (b) Top and side views of the square of the electron wave functions of the first bound excitons in monolayer black phosphorus. (c) Schematic illustration of the sub-band structures and (d) wavelength-dependent reactive oxygen species generation of two-dimensional black phosphorus. Panel a reproduced with permission from ref (26). Copyright 2017 American Physical Society. Panel b reproduced with permission from ref (18). Copyright 2014 American Physical Society. Panels c and d reproduced with permission from ref (29). Copyright 2018 American Chemical Society.

Electron–Hole Interactions

Induced by the promoted interactions between photoinduced electrons and holes, excitonic effects could be prominent and general in low-dimensional semiconductors. As a result, excitons will be dominating photoinduced species that coexist with charge carriers (that is, electrons and holes). The excitonic aspect of photoexcitation processes is inseparable from the widely investigated charge-carrier aspect, which has quite subtle impacts on low-dimensional semiconductor-based photosynthesis.

Exciton-based energy transfer can afford an alternative pathway for solar energy utilization, beyond the traditional carrier-based charge transfer.[13,30,31] Undergoing exchange interactions or dipole–dipole coupling, excitons can resonantly transfer from photocatalysts to substrate molecules, in which the matches of both energy levels and spin multiplicities between the relevant excitonic states are required. In terms of most photosynthetic reactions, bond breaking/making via carrier-based charge transfer is more preferable to obtaining excited states of substrate molecules via exciton-based energy transfer. However, as for some special substrate molecules like oxygen and some organic molecules, exciton-based energy transfer might be much more facile, and the formation of molecular excited states with certain spin multiplicities can give rise to some intriguing behaviors. Recently, several exciting works have been done in exploring exciton-based energy-transfer-initiated photosynthesis in low-dimensional semiconductors.[29,32−35] For instance, Jiang et al. reported photocatalytic intermolecular [2 + 2] cycloadditions triggered by triplet exciton-based energy transfer from CdSe quantum-dot-based photocatalysts (Figure a).[32] Taking advantage of the tunable, triplet-like excitonic states, they regulated the involved triplet energy levels by controlling the size of CdSe quantum dots. Benefiting from these, CdSe quantum dots exhibited an excellent performance, both regioselectivity and diastereoselectivity, in photocatalytic intermolecular [2 + 2] of 4-vinylbenzoic acid derivatives. Using transient absorption spectroscopy, they deduced that triplet exciton-based energy transfer, rather than carrier-based charge transfer, between CdSe quantum dots and 4-vinylbenzoic acid occurred on the surface (Figure b). As compared to the above deduction, Castellano and co-workers have demonstrated that direct triplet exciton-based energy transfer between low-dimensional inorganic semiconductors and organic molecules is feasible, just like that in the molecular donor/acceptor system.[35]

Figure 3

(a) Schematic illustration of energy-transfer-initiated intermolecular [2 + 2] cycloadditions, and (b) exciton decay dynamics in CdSe quantum dots with the additional 4-vinylbenzoic acid. Schematic illustrations of (c) P-type delayed fluorescence originating from TTA in polymeric carbon nitride and (d) biexciton-based energy transfer from excited CdSe/ZnS nanocrystals to surface acceptor dyes. Panels a and b reproduced with permission from ref (32). Copyright 2019 Nature Publishing Group. Panel c reproduced with permission from ref (40). Copyright 2017 Royal Society of Chemistry. Panel d reproduced with permission from ref (42). Copyright 2016 American Chemical Society.

Excitonic effects determine the photosynthetic efficiencies of redox-initiated reactions like water splitting, carbon dioxide reduction, nitrogen fixation, and so on.[12,36] Since that excitonic aspect acts as an opposite of the charge-carrier aspect, excitonic and charge-carrier aspects are interrelated, and the concentrations of exciton and charge carrier are correlative. These features can be understood as follows: on one hand, the strength of excitonic effects (i.e., exciton binding energy, Eb) inherently dominates the concentrations of excitons and charge carriers, where a large Eb means an unfavorable redox-initiated photosynthesis; on the other hand, given that excitons are essential bound electron–hole pairs, desirable charge-carrier concentrations can be expected by promoting the dissociation of excitons. Based on these understandings, it is rational to consider that reducing exciton binding energy and promoting excitonic dissociation are necessary in gaining high-efficiency redox-initiated photosynthesis. We will provide detailed discussions on the relevant excitonic regulations in the next section.

Besides, excitonic effects also impact the light absorption properties. That is, as for low-dimensional semiconductors with large exciton binding energies, promoted electron–hole interactions would lead to exciton absorption, which constitutes a remarkable part of the overall light absorption. Exciton absorption enables sub-band-gap excitation to be achieved in low-dimensional semiconductors. With regard to photosynthesis, the sub-band-gap exciton absorption would facilitate solar energy utilization. However, because of the factors like spin and valley degrees of freedom, exciton absorption in some systems tends to be low-coefficient as compared to interband absorption, and brightening these excitonic states would be advantageous to photosynthetic applications. In addition, as for some semiconductors with large exciton absorption coefficients, exciton absorption would dominate the optical spectral lines. In that case, the band gap estimated from the absorption spectrum (that is, optical band gap) is in fact the energy level of the excitonic state, rather than the energy gap between conduction- and valence-band edges (that is, quasiparticle band gap). There are several effective techniques to determine quasiparticle band gap.[37−39] For instance, photoconductivity measurements have been widely employed for gaining the quasiparticle band gap, on the basis of the electrically neutral feature of the exciton.[37] Other techniques like photoluminescence excitation spectroscopy[38] and scanning tunneling spectroscopy[39] have also been demonstrated to be feasible for extracting the quasiparticle band gap. Combined with some other band-structure characterizations (like angle-resolved photoemission spectroscopy and ultraviolet photoelectron spectroscopy), quasiparticle band alignment can be deduced. However, it is worth noting that the determination of quasiparticle band structures in low-dimensional semiconductor-based photocatalysts would still be challenging, due to the fact that the band structure determined by the interactions between particles/quasiparticles can be significantly modified by the surrounding environments (detailed discussion will be given in Section ).

Quasiparticle-Involved Interactions

The interactions between quasiparticles (such as exciton and phonon) should also be considered when dealing with low-dimensional semiconductor-based photosynthesis.

In comparison with the above-discussed excitonic effects induced by two-particle electron–hole interactions, there are some high-order interactions originating from Coulomb interactions among three or more photoinduced charge carriers. A typical characteristic is the presence of the interactions between two excitons, which embodies impacts on photosynthesis in a variety of ways. For instance, exciton–exciton annihilation, a nonradiative decay process originating from the inelastic collisions between two low-lying excitons, leads to the formation of a high-lying exciton: exciton–exciton annihilation can result in fast depopulation of photoinduced species that is undoubtedly detrimental to photosynthetic energy utilization. Also, they can lead to the generation of high-lying hot excitons, which presents a potential pathway for utilizing low-energy photons.[40,41] For instance, by combining transient absorption and photoluminescence measurements, we have demonstrated the presence of triplet–triplet annihilation (TTA) in polymeric carbon nitride (Figure c), which is the major restriction of the efficiencies of both exciton- and charge-carrier-based photosynthetic reactions.[40] Then, we highlighted that hot excitons induced by TTA can be resonantly transferred to molecular ketones, thereby leading to promoted photosynthetic molecular oxygen activation under visible illumination.[41] In addition to exciton–exciton annihilation, the interactions between excitons might lead to the formation of high-order quasiparticles like the biexciton. Although these quasiparticles have been deemed to be short-lived, their abundant dipole moments permit considerable energy transfer to the single-exciton (Figure d).[42,43] Traditionally, high-order interactions closely depend on the concentrations of photoinduced species, and high-density illumination would be beneficial.

Electron–phonon interactions represent another kind of quasiparticle-involved interactions. The interactions between electronic and vibrational subsystems are intrinsic characteristics of semiconductors, which determine the transitions between different states such as photoexcitation, intraband relaxation, nonradiative/radiative transitions, and so on.[43−46] Owing to the symmetry breaking and dimension reduction, extremely rich thermal vibrations emerge in low-dimensional semiconductors, thus resulting in remarkable electron–phonon interactions. These interactions produce many subtle impacts on photosynthetic energy utilizations. For instance, elastic and inelastic electron–phonon interactions can reduce coherence between valence and conduction bands and lead to direct energy dissipation via lattice relaxation, respectively.[46] In fact, electron–phonon interactions in molecular systems have been widely investigated, leaving both theoretical and experimental understandings of their impacts on solar energy utilization.[47,48] In comparison, due to the much more complicated electronic and vibrational scenarios in semiconductors, rather limited achievements have been made in the field of photosynthesis, and the corresponding aspect deserves more attention.

Regulation of Photosynthesis-Related Properties

Extending Light Harvesting

One of the most important problems in artificial photosynthesis is to achieve effective light absorption. In view of the fact that visible and infrared light make up the majority of solar radiation, it is rational to maximize photosynthetic energy utilization by extending the light response of semiconductor-based catalysts to the visible and infrared regions.

Band structures determine light absorption properties of semiconductors, and the corresponding modifications have been the focus of photosynthetic research.[46,47] With regard to low-dimensional semiconductors, band structures are closely associated with different interactions, and this feature opens up new possibilities of extending light harvesting. Although numerous optimization strategies including defect/distortion engineering, solid solution construction, and heterojunction introduction have been proposed,[48−50] the consideration of the interactions between particles in low-dimensional semiconductors would provide some different viewpoints. For instance, given that Coulomb interactions between these charged species are related to dielectric screening, band gap modulation would be proceeded by extrinsically or intrinsically engineering the dielectric properties.[26,51−54] In this respect, Raja et al. reported band gap modulation of two-dimensional TMDC nanosheets by tuning the surrounding dielectric environment.[52] They constructed heterojunctions between TMDCs and graphene/h-BN nanosheets for achieving dielectric environment regulation. The band gaps of TMDC nanosheets exhibited high sensitivities to nanoscale dielectric change, resulting in significant band gap reduction of several hundred meV (Figure a,b). Ryou et al. demonstrated that a high dielectric environment leads to the opposite shifts of conduction- and valence-band edges that are responsible for the reduction of quasiparticle band gaps.[53] Beyond heterojunctions, chemical doping has also been demonstrated to impact band structures undergoing a Coulomb-interaction engineering mechanism.[26] Moreover, using angle-resolved photoemission/optical spectroscopies and first-principles calculations, Waldecker et al. demonstrated the rigid shifts of valence and conduction bands under dielectric environments, which could be related to the spatial structure of the changes in the Coulomb potential.[54] Such a feature endows dielectric environment-controlled strategies with extensive prospects in the field of photosynthesis. These findings hint that the impacts of traditional optimization strategies on excited-state properties of low-dimensional semiconductor-based photocatalysts should be reassessed. In fact, similar regulations have been unconsciously proceeded in many photocatalysts. For instance, Yang and co-workers have theoretically investigated the impact of buckled geometry on optical properties of g-C3N4, in which they highlighted that the changes in optical absorption would be crucial to photocatalytic water splitting.[55] According to many-body Green’s function-based quasiparticle energy calculations, they suggested that buckled structures lead to a slight increase in quasiparticle band gap but significant reduction in optical band gap linked to first bright excitonic energy, which would be responsible for the visible-light response of g-C3N4. Besides, they reported that even water adsorption can lead to a degree of reduction in excitonic energy, which also facilitates light absorption of the system.

Figure 4

(a) Schematic illustration of dielectric environment-dependent band gap engineering in the WSe2/graphene heterojunction. (b) Experimentally and theoretically obtained evolution of the band gap and excitonic energy level in the WSe2 monolayer. (c) Schematic illustration of energy transfer processes and (d) wavelength-dependent photocatalytic oxidative coupling of benzylamine in the polymeric carbon nitride/acetone system. Panels a and b reproduced with permission from ref (52). Copyright 2017 Nature Publishing Group. Panels c and d reproduced with permission from ref (41). Copyright 2020 Wiley-VCH.

The regulation of high-order interactions typifies another direction in extending light absorption. A prime example is by taking advantage of the TTA-based upconversion process in hybrid photocatalyst systems.[41,56−58] TTA enables high-lying (or hot) singlet excitons to be generated by the collisions between triplet excitons, and the utilization of high-lying singlet excitons before their cooling down would bring about rich possibilities for utilizing low-energy photons. For instance, we have proposed the utilization of ketones as molecular cocatalysts to promote the visible-light response of polymeric carbon nitride. Polymeric carbon nitride here was used as the TTA medium for its robust interactions.[41] According to spectral analyses, we demonstrated that the promoted visible-light response can be related to two factors: the sub-picosecond internal conversion (that is, cooling down of hot excitons) in polymeric carbon nitride and the near-unity intersystem crossing (ISC) efficiency in ketones. Benefiting from these features, two exciton-based energy transfer processes in opposite directions were established: the transfer of hot excitons from polymeric carbon nitride to acetone and the transfer of the triplet exciton from acetone to polymeric carbon nitride (Figure c). The resulting triplet exciton harvesting under visible light illumination promotes photocatalytic singlet oxygen generation (Figure d). It is worth noting that quantum yields of these TTA-based photosynthetic processes are usually low under solar irradiation, on account of the photoinduced species concentration-dependent feature of high-order interactions.

Charge-Carrier Accumulation

Beyond extensive light harvesting, effective accumulation of charge carriers is also an essential prerequisite of redox-initiated photosynthetic reactions. Traditional research mainly focuses on optimizing the charge-separation feature of semiconductor-based photocatalysts.[7,8,59] However, as for low-dimensional semiconductors, charge-carrier accumulation is often governed by the interaction-related factors like excitonic effects and nonradiative decays, and consequently, the regulation of these factors should be the crux of photosynthetic research.

Given the promoted excitonic effects, excitons and charge carriers jointly constitute the dominant photoinduced species in low-dimensional semiconductors. As a consequence, dissociating excitons consisting of bound electrons and holes into free charge carriers would establish an additional pathway for elevating charge-carrier accumulation. In general, energetic disorders, which can provide a reduced free energy barrier and enhanced thermodynamic driving force, are deemed to facilitate exciton dissociation.[60,61] Hence, structural factors like vacancy, heterojunction interface, and lattice distortion are generally employed for boosting exciton dissociation in low-dimensional semiconductors.[62−64] For instance, we have managed to promote charge-carrier accumulation in BiOBr nanoplates by engineering exciton dissociation via oxygen vacancy introduction.[62] According to theoretical simulations, we highlighted that band-edge charge density distributions around the oxygen vacancy are notably delocalized, implying the potential instability of excitons (Figure a). Spectral analyses indicated the reduced exciton concentration (Figure b) and promoted charge-carrier concentrations induced by oxygen vacancies. Benefiting from these features, the defective BiOBr sample exhibited an excellent performance in redox-initiated photocatalytic superoxide radical generation (Figure c). It should be emphasized that exciton dissociation is different from charge separation, and partially dissociated charge-transfer intermediate states driven by electron–phonon coupling-mediated energy dissipation are initially formed before the occurrence of charge separation (Figure d).[61]

Figure 5

(a) Products of the numbers of localized band-edge states on different atomic sites near the oxygen vacancy in BiOBr. (b) Low-temperature photoluminescence and (c) ESR-trapping tests for detecting superoxide radical generation of pristine and defective BiOBr samples. (d) Schematic illustration of exciton dissociation at a donor–acceptor interface. Panels a–c reproduced with permission from ref (62). Copyright 2018 American Chemical Society. Panel d reproduced with permission from ref (61). Copyright 2017 American Chemical Society.

Furthermore, nonradiative decays, which are induced by the interactions like electron–phonon interactions and exciton–exciton annihilation, are also detrimental to charge-carrier accumulation. For instance, electron–phonon interactions can set up an important nonradiative pathway for relaxing energy from electronic to vibrational subsystems. In this regard, suppressing these interactions, which has been unintentionally achieved in some systems, would favor charge-carrier accumulation. For instance, the construction of core–shell structures has been employed for optimizing aspects including light harvesting and charge separation,[65,66] whereas the regulation of electron–phonon interactions has been proceeded simultaneously. Woggon and co-workers have recently investigated the impact of CdS shell thickness on nonradiative decays of the CdSe core.[67] Using temperature-dependent photoluminescence measurements, they highlighted that, in addition to the promoted surface-defect passivation, the increase of shell thickness can effectively reduce electron–phonon coupling in the CdSe core, thus leading to prominent suppression in nonradiative decays. Also, some other structures like heterojunction and doping have been demonstrated to be feasible in regulating electron–phonon coupling in low-dimensional semiconductors.[68,69] As for the regulation of exciton–exciton annihilation, optimization strategies including dimension control, defect engineering, and heterojunction construction have been proposed.[40,70,71] For instance, we demonstrated that TTA in polymeric carbon nitride serves as a major reason for the limited performance of both the energy-transfer- and redox-initiated photosynthetic reactions.[40] Combining the structural characteristics of the polymeric photocatalysts, we confirmed that TTA can be suppressed by dimension control, in which two-dimensional feature-induced ISC reduction gives rise to low triplet-exciton concentration. This finding discloses the essence of the abnormally high efficiencies for both photocatalysis and photoluminescence in two-dimensional polymeric carbon nitride.[72,73]

Spin-Relaxation Optimization

Spin degrees of freedom constitute an important aspect in electronic excitation of low-dimensional semiconductors, which can also impact photosynthetic performance.

As is well-known, long-lived triplet excitons are usually required in triggering energy-transfer-mediated reactions. However, owing to spin conservation, triplet-exciton generation via the direct photoexcitation of ground states (mostly, singlet multiplicity) is forbidden. Triplet-exciton generation usually involves the spin relaxation of singlet excitons, undergoing a nonradiative process called intersystem crossing (ISC) for achieving spin flips.[74,75] Accordingly, the key of high-efficiency triplet exciton generation lies in ISC regulation. Driven by the coupling between two states with different multiplicities, the ISC rate is determined by two factors of semiconductors: spin–orbit coupling (SOC) and singlet–triplet energy gap (ΔEST). Given that strong SOC and small ΔEST favor ISC, great efforts have been devoted to the corresponding optimizations. For instance, focusing on polymeric carbon nitride with intrinsically weak SOC and large ΔEST, we have demonstrated that ISC rate can be effectively promoted by incorporating carbonyl groups into the polymeric networks.[76] The incorporation of carbonyl groups leads to simultaneous optimization of SOC and ΔEST: on one hand, carbonyl groups could mix the pristine ππ* excited states with a certain degree of nπ* configuration, thus leading to promoted SOC; on the other hand, the electrophilic carbonyl groups would reduce the overlap between band-edge charge density distributions, implying a reduced ΔEST. Using prompt fluorescence and phosphorescence measurements, we identified the optimized spin relaxation induced by the promoted SOC and reduced ΔEST in carbonylated carbon nitride samples. Benefiting from the considerable triplet exciton harvesting, the carbonylated carbon nitride samples exhibited excellent performance in energy-transfer-triggered photosynthetic singlet oxygen generation. The high-efficiency and high-selectivity singlet oxygen generation endows the photocatalyst with great potential in selective aerobic oxidation of organic chemicals. To go further, inspired by the novel coupling between semiconductors and molecules, we have proposed a molecular cocatalyst design for optimizing the spin relaxation in polymeric carbon nitride.[41] We demonstrated that the extrinsic transfer of triplet exciton from ketones to polymeric carbon nitride provides an alternative pathway for achieving apparent spin-relaxation optimization.

Note that spin relaxation can also impact the charge-transfer and charge-recombination pathways in semiconductor/molecule systems. Just recently, Wu and co-workers demonstrated the spin-controlled charge recombination pathways across the interface of CdS quantum dot/alizarin.[77] They observed that selective excitations of CdS quantum dots and alizarin molecules could lead to the formation of similar charge-separated states, whereas the recombination products appeared to be triplet alizarin and ground-state hybrid complexes, respectively (Figure ). By using dynamic spectral measurements, they highlighted that such a feature can be ascribable to the different spin-flip rates of electron and hole in CdS quantum dots as well as the different spin multiplicities of the excited states in CdS quantum dots and alizarin molecules. This intriguing phenomenon promises selective photosynthesis to be achieved by excitation light control. Migani et al. theoretically investigated the oxidation of methanol on the (110) facet of rutile TiO2 by using spin-polarized density functional theory.[78] By treating the absorption of methanol molecules at coordinatively unsaturated Ti sites as molecular excitonic states with triplet multiplicity, they analyzed the spin-density and structural evolutions in the reactions and highlighted their impacts on bond dissociation processes. In view of these results, the potential impacts of spin relaxation on photosynthesis in low-dimensional semiconductors should be widely recognized.

Figure 6

Transient absorption kinetics of different species for CdS quantum dots/alizarin under (a) 600 nm and (b) 400 nm excitation. (c, d) Corresponding illustrations of the spin-controlled charge-recombination pathways. Reproduced with permission from ref (77). Copyright 2020 American Chemical Society.

Conclusions and Outlook

Low-dimensional semiconductors exhibit great prospects in the field of artificial photosynthesis. However, the significant interactions between particles/quasiparticles induced by promoted structural confinement and reduced dielectric screening endow low-dimensional semiconductors with pretty different excited-state scenarios, as compared to those in the bulk counterparts. Therefore, it is essential to clarify the impacts of these interactions on the relevant solar energy utilization. In this Outlook, we summarize recent progress in investigating the interactions in low-dimensional semiconductors, where the impacts of different interactions on photosynthetic energy conversion are highlighted. According to our descriptions, it is clear to see that the regulation of these interactions, whether intentionally or unconsciously, has been realized in many optimization attempts, thus enabling diverse photosynthetic applications to be obtained in these low-dimensional systems.

Taken together, the interactions between particles/quasiparticles in low-dimensional semiconductors open up tremendous possibilities in photosynthetic applications. The evaluation of the impacts of these interactions not only establishes a brand new understanding of the involved catalytic mechanisms but also paves the way for gaining comprehensively optimized solar energy utilization.