Modeling Plasmon-Induced Hot-Carrier Transfer

Abstract

In this issue of Chem, Zhang et al. use first-principle simulations (i.e., no empirically adjustable parameters) to elucidate the detailed mechanisms of plasmon-induced hot-carrier transfer at a heterojunction interface between a plasmonic gold nanomaterial and a semiconductor. Together with earlier work by Long and Prezhdo, this study shows how the donor-acceptor interaction at the interface determines the mechanism by which plasmonic excitations lead to hot-electron transfer at interfaces. In this issue of Chem, Zhang et al. use first-principle simulations (i.e., no empirically adjustable parameters) to elucidate the detailed mechanisms of plasmon-induced hot-carrier transfer at a heterojunction interface between a plasmonic gold nanomaterial and a semiconductor. Together with earlier work by Long and Prezhdo, this study shows how the donor-acceptor interaction at the interface determines the mechanism by which plasmonic excitations lead to hot-electron transfer at interfaces. Plasmons, quasi-particles that represent collective excitations of electrons, have been of great interest in chemistry and condensed-matter physics for many decades. Plasmonic excitations are central to various photonic devices, and various spectroscopic techniques, including surface-enhanced Raman spectroscopy, also take advantage of this phenomenon. In more recent years, utilization of plasmonic excitations for photovoltaic and photocatalytic applications has emerged as an exciting research field because many plasmonic metal nanomaterials show enhanced light absorption, potentially leading to high photon-to-electron conversion efficiency. When plasmonic metal nanomaterials such as gold (Au) quantum dots are interfaced with semiconductor materials, plasmonic excitations are believed to result in the transfer of hot carriers from the metal nanoparticle to the semiconductor. Developing a better understanding of this mechanism is central to making plasmon-enhanced hot-carrier transfer practical for use in future optoelectronic devices. The phenomenon of plasmonic excitation has long been studied, going back perhaps most notably to the theory by Mie on light scattering by dielectric particles. Advancements in computational capabilities have made it possible to quantitatively study plasmons in small nanomaterials by applying quantum-mechanical theory to electrons. Many nanomaterials of recent interest are relatively large in size, yet they are often too small to be described as homogeneous dielectric particles. Therefore, first-principle electronic-structure calculations based on density functional theory (DFT) have been largely used for the study of these nanomaterials. Many such studies have focused on the plasmonic excitation itself and the subsequent decay of plasmons in isolated nanomaterials. At the same time, the dynamics of plasmon-induced hot-electron transfer (PHET) at material interfaces has been scarcely studied (Figure 1). First-principle calculations, which are free of atom-specific empirical parameters, are invaluable, especially because the conventional view of PHET has been called into question in recent years. Traditionally, in the PHET process, plasmons are believed to decay rapidly into hot charge carriers (i.e., electrons and holes) inside the metallic nanomaterial within tens of femtoseconds via Landau damping, and then the generated hot electrons transfer into the interfaced semiconductor. The PHET process is often considered to be rather inefficient because the hot-carrier transfer process competes with the energy-loss process of carriers and their recombination near the Fermi energy within the metal. In 2014, Long and Prezhdo proposed that the initial plasmonic excitation has a strong charge-transfer character at the interface of a small Au nanoparticle and a TiO2 surface.1Long R. Prezhdo O.V. Instantaneous generation of charge-separated state on TiO2 surface sensitized with plasmonic nanoparticles.J. Am. Chem. Soc. 2014; 136: 4343-4354Crossref PubMed Scopus (193) Google Scholar Their first-principle simulations showed that as much as half of such cases result in hot-electron injection into the TiO2 on a sub-100-fs timescale, bypassing the carrier energy losses within the Au nanoparticle. This unconventional PHET mechanism was later demonstrated experimentally for an interface between a Au nanoparticle and a CdSe nanorod by Lian and co-workers in 2015.2Wu K. Chen J. McBride J.R. Lian T. Charge Transfer. Efficient hot-electron transfer by a plasmon-induced interfacial charge-transfer transition.Science. 2015; 349: 632-635Crossref PubMed Scopus (771) Google Scholar These two studies revealed a new PHET mechanism that is conceptually very different from the traditional mechanism in which the plasmon decay is followed by competing processes between hot-carrier transfer and carrier energy loss within the plasmonic particle. These new exciting findings called for the development of a better and more quantitative understanding at the molecular level. In this issue of Chem, the article by Zhang et al. presents a new insight into PHET mechanisms from their first-principle simulations on a Au nanorod interfaced with a two-dimensional semiconductor, MoS2.3Zhang Z. Liu L. Fang W.-H. Long R. Tokina M.V. Prezhdo O.V. Plasmon-mediated electron injection from Au nanorods into MoS2: Traditional versus photoexcitation mechanism.Chem. 2018; 4: 1112-1127Abstract Full Text Full Text PDF Scopus (62) Google Scholar The work reveals that the initial plasmon state is mostly localized within the Au nanorod. This excitation decays quickly into free electrons within 30 fs. Interestingly, the authors show that the hot-carrier transfers from the Au nanorod into the MoS2 with a timescale of ∼100 fs at this interface, and the competing mechanism of phonon-induced carrier relaxation and recombination within the Au nanorod occurs at a timescale of ∼200 fs. This is probably the first demonstration of how carrier relaxation and recombination in a metal nanomaterial can compete effectively with hot-carrier transfer in the PHET process. Unlike the case of the Au-TiO2 interface, no plasmonic charge-transfer excitations were found in this case, and the authors attribute the difference between the Au-TiO2 and Au-MoS2 interfaces to the difference in the donor-acceptor interaction strength between the Au nanomaterial and the semiconductor. In the case of the Au-MoS2 interface, the semiconductor atoms are stoichiometrically fully coordinated, whereas TiO2 possesses many under-coordinated atoms at the surface. Thus, the relatively weak donor-acceptor interaction of the Au-MoS2 interface results in plasmonic excitations localized within the Au nanorod, exhibiting no significant charge-transfer character. The first-principle simulations presented in the article by Zhang et al. also reveal how the initial plasmon excitation decays and how the energy loss of free carriers is coupled to different types of lattice and atom movements. Whereas the initial plasmonic excitations are coupled only to low-frequency acoustic phonons, the free electrons couple dominantly to high-frequency phonons of the plasmonic Au nanorod and even to those of MoS2 to some extent. In the framework of the traditional PHET mechanism, this new understanding of how different processes are affected by interface properties at the molecular level will aid in the design of optimal interfaces for an efficient PHET. The work reported by Zhang et al.3Zhang Z. Liu L. Fang W.-H. Long R. Tokina M.V. Prezhdo O.V. Plasmon-mediated electron injection from Au nanorods into MoS2: Traditional versus photoexcitation mechanism.Chem. 2018; 4: 1112-1127Abstract Full Text Full Text PDF Scopus (62) Google Scholar represents the current state-of-the-art in theoretical efforts for investigating PHET dynamics at highly complex interfaces. It is an elegant application of an advanced computational method, revealing surprising fundamental insights beyond the conventional view on the PHET mechanism. Still, the field is not without outstanding challenges, especially when it comes to making further, more quantitative predictions. Just as recent experimental findings are facilitated by the development of advanced non-linear spectroscopy techniques, advancement in computational methodologies is central to fostering theoretical efforts in the field. Several outstanding challenges remain and exciting opportunities exist for further methodology development. In studying the dynamic processes of excited electrons in nanomaterials, the challenges are twofold: one relates to the time-dependent quantum-dynamics simulation, and the other relates to the electronic structure theory calculation. With regard to the challenge of time-dependent quantum dynamics, the work by Zhang et al.3Zhang Z. Liu L. Fang W.-H. Long R. Tokina M.V. Prezhdo O.V. Plasmon-mediated electron injection from Au nanorods into MoS2: Traditional versus photoexcitation mechanism.Chem. 2018; 4: 1112-1127Abstract Full Text Full Text PDF Scopus (62) Google Scholar employs the so-called fewest switches surface hopping (FSSH) method4Tully J.C. Molecular dynamics with electronic transitions.J. Chem. Phys. 1990; 93: 1061Crossref Scopus (2864) Google Scholar for modeling the quantum dynamics of the electrons that are coupled to the classical dynamics of the nuclei. The FSSH approach is known to fall short of accurately describing the electronic decoherence because the nuclei are treated classically in this mixed quantum-classical-dynamics approach. Several remedies to correct for the shortcoming have been proposed through the study of simple models, and a better understanding of more general cases will help to overcome this limitation.5Subotnik J.E. Ouyang W. Landry B.R. Can we derive Tully’s surface-hopping algorithm from the semiclassical quantum Liouville equation? Almost, but only with decoherence.J. Chem. Phys. 2013; 139: 214107Crossref PubMed Scopus (131) Google Scholar A significant challenge is to move beyond the mixed quantum-classical-dynamics framework of the FSSH method for simulating the coupled electron-nuclear dynamics. For small molecular systems, great progress has been made. For instance, the ab initio multiple-spawning method by Martinez and co-workers models the quantum nature of nuclear dynamics by propagating Gaussian wavepackets.6Ben-Nun M. Quenneville J. Martinez T.J. Ab initio multiple spawning:c Photochemistry from first principles quantum molecular dynamics.J. Phys. Chem. A. 2000; 104: 5161-5175Crossref Scopus (656) Google Scholar More recently, Gross and co-workers introduced the so-called exact-factorization approach, in which the full time-dependent electron-nuclear wavefunction is factorized with a time-dependent external potential, alleviating the usual Born-Huang expansion.7Abedi A. Maitra N.T. Gross E.K. Exact factorization of the time-dependent electron-nuclear wave function.Phys. Rev. Lett. 2010; 105: 123002Crossref PubMed Scopus (301) Google Scholar Such approaches inherently incorporate the electronic decoherence. Although direct application of these methods would be too computationally expensive for studying large complex systems today, new insights for novel processes such as PHET might be obtained with a better description of the coupled electron-nuclear dynamics. Given that the electronic Hamiltonian drives the quantum dynamics of the excited electrons, electronic structure theory represents another important challenge.8Li L. Wong J.C. Kanai Y. Examining the Effect of Exchange-Correlation Approximations in First-Principles Dynamics Simulation of Interfacial Charge Transfer.J. Chem. Theory Comput. 2017; 13: 2634-2641Crossref PubMed Scopus (12) Google Scholar Continued development of better exchange-correlation approximations in the context of DFT remains an important research area. In recent years, the GW method and related many-body methods have emerged as promising formalisms for obtaining better descriptions of the single-particle states (i.e., charged excitations) of nanomaterials.9Onida G. Reining L. Rubio A. Electronic excitations: Density-functional versus many-body Green’s-function approaches.Rev. Mod. Phys. 2002; 74: 601Crossref Scopus (2970) Google Scholar Although treating interactions between electrons and holes by representing electronic excitations in terms of many-particle wavefunctions is desirable, having a manifold of closely spaced excited states is a great computational challenge for studying nanomaterials with thousands of electrons. An interesting and potentially promising alternative is an approach based on the real-time propagation of the so-called Kadanoff-Baym equations, which cast the time-dependent problem in terms of Greens functions. A first-principle implementation of the Kadanoff-Baym equations has been demonstrated already for simple materials such as silicon,10Marini A. Competition between the electronic and phonon-mediated scattering channels in the out–of–equilibrium carrier dynamics of semiconductors: An ab-initio approach.J. Phys. Conf. Ser. 2013; 427: 012003Crossref Scopus (43) Google Scholar and this time-evolved non-equilibrium Green-function approach might open up exciting opportunities for studying PHET dynamics in the future. This preview discusses the work by Zhang et al.3Zhang Z. Liu L. Fang W.-H. Long R. Tokina M.V. Prezhdo O.V. Plasmon-mediated electron injection from Au nanorods into MoS2: Traditional versus photoexcitation mechanism.Chem. 2018; 4: 1112-1127Abstract Full Text Full Text PDF Scopus (62) Google Scholar in the context of recent findings by Wu et al.2Wu K. Chen J. McBride J.R. Lian T. Charge Transfer. Efficient hot-electron transfer by a plasmon-induced interfacial charge-transfer transition.Science. 2015; 349: 632-635Crossref PubMed Scopus (771) Google Scholar and earlier theoretical work by Long and Prezhdo.1Long R. Prezhdo O.V. Instantaneous generation of charge-separated state on TiO2 surface sensitized with plasmonic nanoparticles.J. Am. Chem. Soc. 2014; 136: 4343-4354Crossref PubMed Scopus (193) Google Scholar The work by Zhang et al. represents the current state-of-the-art in modeling PHET at complex interfaces from first-principle theory, and the simulation has greatly advanced our understanding of PHET mechanisms. At the same time, continued progress will require further advancement in theoretical and computational methodologies, and emerging, potentially promising approaches are highlighted. The development of new computational methodologies will continue to enable more scientific discoveries and help us develop a better understanding of PHET in complex systems at the molecular scale.