Influence of Defects on Excited-State Dynamics in Lead HalidePerovskites: Time-Domain ab Initio StudiesWei Li,*,† Run Long,*,‡ Jianfeng Tang,*,† and Oleg V. Prezhdo*,§

†College of Science, Hunan Agricultural University, Changsha 410128, People’s Republic of China‡College of Chemistry, Key Laboratory of Theoretical & Computational Photochemistry of Ministry of Education, Beijing NormalUniversity, Beijing 100875, People’s Republic of China§Department of Chemistry, University of Southern California, Los Angeles, California 90089, United States

ABSTRACT: This Perspective summarizes recent research into the excited-state dynamics in lead halide perovskites that are of paramount importance forphotovoltaic and photocatalytic applications. Nonadiabatic molecular dynamicscombined with time-domain ab initio density functional theory allows one tomimic time-resolved spectroscopy experiments at the atomistic level of detail.The focus is placed on realistic aspects of perovskite materials, including pointdefects, surfaces, grain boundaries, mixed stoichiometries, dopants, andinterfaces. The atomistic description of the quantum dynamics of electronand hole trapping and recombination, provided by the time-domain ab initiosimulations, generates important insights into the mechanisms of charge andenergy losses and guides the development of high-performance perovskite solarcell devices.

Hybrid organic−inorganic perovskite solar cells (PSCs)have drawn intense interest in recent years because of

the rapid evolution of the power conversion efficiency.Perovskite materials feature the chemical formula of ABX3,where A is a monovalent cation, e.g., Cs+, methylammonium(MA+), formamidinium (FA+) or guanidinium (GA+); B is adivalent cation, e.g., Pb2+ or Sn2+; and X is a halide anion suchas Cl−, Br−, or I− (Figure 1). The record efficiency of PSCsbased on the mixed (FAPbI3)0.95(MAPbBr3)0.05 perovskitereached 23.2% in 2018. Hybrid perovskites have manyadvantages over other materials, including low exciton bindingenergy, large carrier diffusion length, optimal band gap, highabsorption coefficient, long charge carrier lifetime, abundantcomponent elements, and low cost of fabrication. Thesesuperior properties motivate perovskite applications in photo-voltaics,1,2 photocatalysis,3 thermoelectrics,4 light-emittingdiodes,5,6 etc. Despite the significant progress in PSCs, therecord efficiency of 23.2% is still below the detailed balancelimit of 30% established for p−n junction solar energy deviceswith the energy gap at 1.1 eV.7 A thorough understanding ofthe mechanisms for charge and energy losses is essential forfurther progress.Defects are inevitable in solution-processed single-crystalline

and polycrystalline films, and they are frequently present bothin the bulk and at surfaces. Defects in solid materials exist inthe form of point defects, such as vacancies and interstitials,and extended defects, including grain boundaries (GBs) andinterfaces with other materials. Defects usually introducesubgap states, which are expected to trap charge carriers,facilitating nonradiative electron−hole recombination andenergy losses. Generally, long charge carrier lifetimes that

result in high carrier concentration are needed to achieveefficient solar cell devices with high voltage and large current.Defects are abundant in classical semiconductors, such ascrystalline silicon, and have relatively high concentration.Whereas nonradiative charge recombination involving defectstates is significant in the traditional semiconductors,8,9 thereported charge carrier lifetimes are exceptionally long inperovskites, ranging from a few into hundreds of nano-seconds10 and indicating that defect states in perovskitematerials are benign electrically.

Received: March 5, 2019Accepted: June 17, 2019Published: June 17, 2019

Nonradiative charge recombina-tion limits further improvementof perovskite solar cells. The

questions of whether and howdefects decrease charge carrierlifetimes are still under debate.We provide theoretical insightsinto the influence of defects oncharge recombination, in directconnection with time-resolved

experiments.

Perspective

pubs.acs.org/JPCLCite This: J. Phys. Chem. Lett. 2019, 10, 3788−3804

© 2019 American Chemical Society 3788 DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

This is an open access article published under an ACS AuthorChoice License, which permitscopying and redistribution of the article or any adaptations for non-commercial purposes.

Dow

nloa

ded

via

94.2

31.2

17.4

8 on

Jul

y 22

, 201

9 at

11:

32:0

3 (U

TC

).Se

e ht

tps:

//pub

s.ac

s.or

g/sh

arin

ggui

delin

es f

or o

ptio

ns o

n ho

w to

legi

timat

ely

shar

e pu

blis

hed

artic

les.

Many experimental and theoretical works are devoted tounderstanding the unique defect physics in PSCs.11−24 It hasbeen shown from the experimental viewpoint that defects inperovskites are correlated with such critical issues as structuralinstability,25 ion migration,25 current−voltage hysteresis,26

giant photovoltaic effect switchable by electric field,27 andformation of n−i−p structures.28 An ad hoc model based onbistable amphoteric native defects has been used to account forthese uncommon phenomena.29 Most theoretical works usefirst-principles calculations based on density functional theory(DFT) to investigate the defect properties, including formationenergies and defect energy levels. Investigation of suchproperties requires big supercells, and therefore, bare DFTfunctionals are employed for this purpose. Although suchfunctionals predict correct energy gaps because of errorcancellation, a more reliable description of perovskiteelectronic structure requires hybrid DFT functionals withspin−orbital coupling effects, or GW calculations, which arevery time-consuming.16,30,31

Several groups used first-principles calculations to studyenergy levels and formation energies of native point defects inMAPbI3, demonstrating that most defects are shallow and havelow formation energies and that defects with deep levels arehard to form.16,19,32,33 The concept of “defect tolerance” wasused to explain such particular properties of PSCs.34 Theunusual defect tolerance originates from the ionic nature andstrong antibonding coupling between Pb 6s and I 5p orbitals.Considering vacancies as an example, the dangling bond statesproduce defect levels that are either close to the band edges orresonances inside the bands. However, recent experimentssuggest that shallow defects play important roles in inducingunexpected doping and deteriorating carrier lifetimes.35−37

The majority of the theoretical studies focus on geometric andelectronic structure of defects, and detailed atomistic insightsinto the influence of defect states on charge relaxation andrecombination are generally lacking. Theoretical description ofthe nonequilibrium nature of such dynamics in condensedmatter systems is still challenging.38−41 Nonadiabatic molec-ular dynamics (NA-MD) is the method of choice to modelexcited-state processes in the time domain and at the atomisticlevel of detail, mimicking time-resolved experiments directly.

Nonradiative excited-state decay depends on the energy gap,NA coupling (NAC), and pure-dephasing/decoherence time.NAC between adiabatic orbitals j and k is defined by

ϕ ϕ− ℏ⟨ |∇ | ⟩·i j k tRRd

d. The first term characterizes how adiabatic

orbitals depend on a nuclear coordinate R, and the secondterm is the corresponding nuclear velocity. Note that R is avector. In order for the NA coupling to be large, the orbitals jand k have to be localized in the same place; that is, theyshould overlap, the orbitals should depend significantly onnuclear coordinates, and the nuclear velocities for thesecoordinates should be large, e.g. when atoms are light andphonons have high frequencies. The coherence time is long ifphonon-induced fluctuation of the corresponding energy gap issmall and if the fluctuation maintains long correlation.42

According to Fermi’s golden rule, the nonradiative decay rate isproportional to the coupling squared and inversely propor-tional to the energy gap. Quantum dynamics is faster ifcoherence time is large. Rapid decoherence leads to thequantum Zeno effect, according to which quantum dynamicsstops.43

In this Perspective, we summarize the recent progress thathas been made to understand the excited-state dynamics inhalide perovskites in the presence of defects using thesimulation tool PYXAID,44,45 which implements state-of-the-art NA-MD in the framework of time-domain densityfunctional theory (TD-DFT) developed by the Prezhdogroup.46 The discussions focus on the influences of severaltypes of defects, including point defects, surfaces, andboundaries, as well as dopants and mixed stoichiometries, onthe charge carrier trapping and relaxation processes. Exemplify-ing these practical aspects in halide perovskites, the combinedNA-MD/TD-DFT methodology demonstrates how holetrapping by the iodine interstitial defect can extend chargecarrier lifetimes; how charge recombination can be controlledby modulating oxidation states of the iodine vacancy; that alead vacancy can suppress nonradiative electron−holerecombination; why PbI2-rich perovskites show betterperformance than MAI-rich systems; how passivation ofperovskite surfaces with Lewis bases can extend charge carrierlifetime by different mechanisms; that GBs accelerate chargerecombination, whereas doping can repair the detrimental

Figure 1. (a) Crystal structure of ABX3 halide perovskites. A is a monovalent organic or inorganic cation, B a divalent metal cation, and X a halideanion. Schematics of (b) perfect lattice and various defects considered in this work: (c) iodine interstitial; (d) iodine vacancy; (e) Pb vacancy; (f)MAPbI3 with PbI surface termination; (g) MAPbI3 with MAI surface termination; (h) Σ5 (012) grain boundary (GB) in MAPbI3; (i) chlorine-doped Σ5 (012) GB; (j) MAPbI3 with some MA replaced by FA or GA; and (k) mixed halide CsPbBr1.5I1.5.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3789

effect produced by the GBs; that chemical dopants can be usedto control charge recombination at MAPbI3/TiO2 interfaces;that halide content in MAPbX3 and CsPbX3 has significantinfluences on charge recombination; and that doping MAPbI3with larger A-site organic cations can extend charge carrierlifetimes. The detailed time-domain atomistic investigationsidentify the mechanisms of energy losses, characterize thephonon modes that promote the electronic transitions andaccept the lost electronic energy, and suggest design principlesfor controlling and eliminating unfavorable nonradiativeelectron−hole recombination pathways, which are of particularimportance for design of high-performing perovskite materialsfor photovoltaic and photocatalytic applications.

NA-MD is an efficient and reliable method for simulation ofquantum dynamics of excited-state energy and charge-transferprocesses in condensed matter systems.44−48 NA-MD treatselectrons quantum mechanically and nuclei (semi)classicallyand includes transitions between electronic states. Theelectronic evolution is well-described by TD-DFT, whichprovides both good accuracy and computational efficiency.Coupling to the nuclear evolution is achieved via surfacehopping (SH), which can be viewed as a master equation withtime-dependent transition rates. Importantly, SH achievesdetailed balance between transitions upward and down inenergy, leading to thermodynamic equilibrium in the long-timelimit.49 This key feature of electron−phonon relaxation ismissing if electron−nuclear coupling is described by the mean-field (Ehrenfest) approach. Tully’s fewest switches SH (FSSH)is the most popular NA-MD approach.50 It is very well-suitedfor modeling excited-state processes involving isolated conicalintersections, e.g., photoisomerization of molecules,46 andintraband relaxation characterized by rapid hops within a densemanifold of states.51 In such cases, decoherence effects can beignored or treated in a simple way, for instance via aninstantaneous wave function collapse after a hop. Extendedsystems provide additional challenges to FSSH due to a largenumber of trivial crossings. The parameter-free correctedFSSH algorithm52 and the subspace crossing correction53 tothe standard FSSH address the issue. Slow transitions acrosslarge energy gaps, encountered for example during chargetrapping and recombination, require semiclassical correctionsfor quantum decoherence.44 In such cases, we employ thedecoherence-induced surface hopping (DISH) algorithm thatuses decoherence as the physical mechanism of wave functioncollapse leading to a surface hop.54

NA-MD requires averaging over many trajectories represent-ing both a (canonical) ensemble of initial conditions, anddifferent realizations of the stochastic hopping process.

Moreover, excited-state forces have to be calculated todetermine these trajectories. Such calculations are extremelycomputationally demanding and can be simplified greatly usingthe classical path approximation (CPA).45 CPA uses aprecomputed nuclear trajectory to determine the TD-DFTand SH dynamics. Usually, a short time ground-state trajectorycan be used, justified by the facts that excitation of one or fewelectrons in condensed matter and nanoscale systems inducesonly a small perturbation to the electron density and nuclearforces, in particular, compared to thermal nuclear fluctuations.In order to obtain quantum dynamics on a nanosecond timescale, one can extrapolate the results of picosecond simulationsusing a short-time approximation to the exponential decay.Alternatively, one can assume that the picosecond simulationprovides sufficient sampling of the excited-state energies andNA couplings, iterate the NA Hamiltonian multiple times, andperform quantum dynamics simulations on a long time scale.The extrapolation approach is suitable when consideringtransitions from initially populated states, while the Hamil-tonian iteration technique is useful to study transientpopulation and depopulation of traps. NA-MD has beenapplied to various perovskites systems, including low-dimen-sional perovskites,55,56 perovskites containing defects57−59 andGBs,60,61 perovskite surfaces passivated with organic mole-cules,62 interfaces with TiO2,

63 etc.64−75 Please refer to otherreviews for a more comprehensive description of the NA-MDmethodology and surface hopping algorithms.48,76−82

Inf luence of Point Defects on Nonradiative Charge Recombina-tion. Point defects are the most investigated defect type. Theycreate a low density of charge traps in perovskite singlecrystals.37 It has been demonstrated that the open-circuitvoltage (Voc) deficit arises from the trap-assisted electron−holerecombination.35 The Voc decrease of about 0.36 eV from themaximum value of 1.55 eV, corresponding to the band gap, israther small compared to inorganic semiconductors, indicatingthat defects play a relatively minor role in carrier losses inPSCs.83,84 At the same time, the observation of lowphotoluminescence quantum yields (less than 1%) inMAPbI3 perovskite thin films85 suggests that nonradiativeenergy losses are significant in PSCs, and some point defectsare detrimental. Understanding and controlling defects, withthe goal of suppressing the nonradiative charge recombination,are essential for further improvement of PSC performance.Defect concentration is related to their formation energy.

Experimentally, the chemical potential for equilibrium growthcan be regulated by manipulation of precursors, pressure, andtemperature to alter the free energy of defect formation. Thenature of point defects in perovskites is complicated, and thereare several reviews on this topic which provide comprehensiveaccounts of defect types and energies.31,35,84,86,87 Theoretically,p-type defects, such as the Pb vacancy (PbV) and the iodineinterstitial (Ii), and n-type defects, such the iodine vacancy(IV), are proposed to be dominant. Particularly, Ii and PbV can

Quantum dynamics involvescomplex interplay between elec-tronic and nuclear degrees offreedom. Time-domain densityfunctional theory and nonadia-batic molecular dynamics providea unique perspective on non-equilibrium processes in con-densed matter and nanoscale

materials.

Iodine-rich/Pb-poor conditionsintroduce point defects that arebenign for charge carrier lifetime.The oxidation state of the iodinevacancy strongly influences non-radiative charge recombination.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3790

be easily formed under iodine-rich/lead-poor conditionsbecause of their low formation energies.12,18 First-principlescalculations suggested that the iodine interstitial defectintroduces a deep trap, whereas the Pb vacancy creates onlya shallow trap in MAPbI3.

13,16 Using impedance spectroscopymeasurements, Yang and co-workers attributed the defect levelabout 0.16 eV above the valence band maximum (VBM) inbulk MAPbI3 to the iodine interstitial.88 The iodine interstitialdefect has a low activation energy to diffuse across theperovskite crystal under light illumination, according to DeAngelis and co-workers.18 The iodine vacancy was reported tobe an energetically favorable site for adsorption of molecularoxygen, leading to superoxide formation.57,89 A number ofworks reported measurements of carrier recombinationlifetimes in PSCs in the presence of traps using time-resolvedphotoluminescence.23,26,90−95 Experiments showed that chargecarrier lifetimes can be extended by an order of magnitude withaddition of excess iodine in precursors during synthesis of bothMAPbI3 and FAPbI3 perovskites.

1,96

Imparting point defects with a particular charge (oxidationstate) is a viable strategy to control electrical conductivity andenhance photocatalytic activity by defect engineering.97 Manyof the unusual characteristics of PSCs can be ascribed to thecharged defects formed under external and built-in electricfields or light.13,15,20 For example, Huang and co-workersreported that the giant switchable photovoltaic effect inMAPbI3-based devices is achieved under external electricfield because of drift of charged defects.27 In the absence of afield, charged defects randomly migrate along the perovskitelayer. In particular, migration of charged halide vacanciesaffects the current−voltage characteristics and can account forthe current−voltage hysteresis in PSCs at room temperature.18

Charged defects, including charged cations and anions, mayattract each other because of electrostatic Coulombinteractions and form clusters of defects that heal each other,explaining the low trap density in PSCs.15,86,98 Defects indifferent charge states have a strong influence on thenonradiative charge recombination, as discussed by Kosterand co-workers.91 Interestingly, an uncommon improvementof PSC performance in the presence of iodide vacancies underbias voltage was reported.99 The plethora of experimental dataon the influence of various defects and synthetic protocols oncharge carrier trapping and relaxation in PSCs motivatestheoretical studies. Time-domain ab initio simulations ofexcited-state dynamics provide the most direct way to modelsuch experiments and to obtain a thorough understanding ofthe defect properties at the atomistic level of detail.The influence of the iodine interstitial defect on the charge

carrier trapping and relaxation dynamics in MAPbI3 wasexamined using ab initio NA-MD.58 It was shown that aninterstitial iodine atom introduces a deep trap state (relative tokBT) near the VBM (Figure 2), in line with the previouselectronic structure investigations and available experimentalcharacterization.33,88 The charge densities of both valenceband (VB) and conduction band (CB) states are delocalized,while the defect level is localized. The CB minimum (CBM) isformed by Pb 6p orbitals, while the VBM is composed of Pb 6sand I 5p orbitals coupled in an antibonding manner. Thedefect state originates from I 5p orbitals and overlaps betterwith the charge density of the VBM than CBM. The hole andelectron trapping occur on tens of picosecond and 100 ns timescales, respectively. The hole trapping is much faster thanelectron trapping for the following reasons: First, the trap state

is closer in energy to the VBM than the CBM. Second, the NAelectron−phonon coupling for hole trapping is large, becausethe wave functions of both the defect level and the VBMoriginate from I 5p orbitals and overlap much better than thewave functions of the defect level and the CBM. Third,quantum decoherence between the defect state and the VBMis long. Most interestingly, the trapped holes are surprisinglylong-lived. Recombination of the trapped holes with CBelectrons is several times slower than recombination of VBholes with CB electrons. The trapped holes can escape into theVB prior to recombining with CB electrons. These factors canlead to reduction of charge and energy losses in MAPbI3 withthe iodine interstitial defect and contribute to the highefficiencies of PSCs.The effect of the Pb vacancy on the nonradiative electron−

hole recombination in FAPbI3 was investigated.100 Thenonradiative recombination in pristine FAPbI3 occurred on ananosecond time scale and was extended by an order ofmagnitude in the presence of the Pb vacancy (Figure 3). Thisresult agrees with the experimental observation of long chargecarrier lifetimes in FAPbI3 under iodine-rich conditions.1 The

Figure 2. Total and projected densities of states (DOS) of pristineMAPbI3 (top panel) and MAPbI3 with the iodine interstitial defect(bottom panel). The insets show charge densities of (a) VBM and (b)CBM in pristine MAPbI3 and (c) VBM, (d) CBM, and (e) trap statein MAPbI3 with the iodine interstitial defect. Adapted from ref 58.Copyright 2017 American Chemical Society.

Figure 3. Decay of excited-state population in FAPbI3 with andwithout the Pb vacancy. Adapted from ref 100. Copyright 2018American Chemical Society.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3791

decrease in the nonradiative electron−hole recombination ratein FAPbI3 with the Pb vacancy can be explained by consideringthe NA electron−phonon coupling and the energy gap. TheNA electron−phonon coupling is smaller in the presence of thePb vacancy, because the missing Pb atom localizes the VBMcharge density and thereby decreases the VBM−CBM overlap.The band gap is almost unaffected by the defect, because theshallow trap state generated by the atomic vacancy merges intothe VB. Kilin and co-workers studied the effect of the Pbvacancy on the charge carrier dynamics in MAPbI3 using aRedfield theory formalism.101 They also found that the Pbvacancy increases the nonradiative relaxation time of theexcited electron, in agreement with the NA-MD simulations.100

Point defects in halide perovskites can exist in differentoxidation states, which can be controlled by chemical orelectrical doping, and altered by flowing charges under PSCoperating conditions. The effect of the oxidation state on theelectron−hole recombination dynamics in MAPbI3 in thepresence of an iodine vacancy was investigated.59 Neutral (IV),cation (IV

+1), and anion (IV−1) iodine vacancies were considered.

Here, the charge refers to the charge of the missing iodineatom. Thus, IV

−1 corresponds to the missing iodide anion. Thecalculations show that the oxidation state of the halide vacancyhas a strong influence on the charge carrier dynamics (Figure4). The iodide vacancy IV

−1 has negligible influence on theelectron−hole recombination, because it creates no subgapstates. The neutral IV accelerates the recombination to amoderate extent, because it creates a shallow trap near theCBM, and increases the NA electron−phonon coupling. TheIV+1 defect produces an additional hole trap near the VBM,compared to the neutral IV. This is because the negative chargeoccupying the vacancy site attracts the two nearby Pb ions thatform a stable dimer and detach from iodines, leaving themchemically unsaturated. IV

+1 accelerates charge recombinationby 2 orders of magnitude, compared to pristine MAPbI3, byproviding additional relaxation pathways via both shallow anddeep trap states.The above analysis of the influence of point defects on the

charge carrier trapping and relaxation dynamics, performedusing NA-MD combined with real-time TD-DFT, demon-strates that the most easily formed defects have little influence

on charge carrier lifetimes, rationalizing the high efficiencies ofPSCs. The iodine interstitial defect in MAPbI3 even extendsthe charge carrier lifetime, because it introduces a gap statenear the VBM that couples weakly to the CBM. Holes can betrapped and detrapped many times prior to recombining withelectrons. The Pb vacancy decreases the nonradiative chargerecombination rate in FAPbI3, because it localizes the VBMcharge density and decreases the NA coupling. The iodidevacancy does not introduce midgap trap states and decreasescharge carrier lifetimes only modestly. However, the samepoint defects in other, less common oxidation states candecrease carrier lifetimes by orders of magnitude, indicatingthat PSC efficiencies can depend strongly on nonequilibriumoperation conditions involving strong currents, high carrierdensities, exposure to environment and chemicals, etc.

Charge Recombination at Surfaces and Grain Boundaries. Inaddition to the point defects in perovskite crystals, extendeddefects, such as grain boundaries, surfaces, and interfaces, caninfluence charge carrier lifetimes and transport. The observedcarrier diffusion is much longer in single crystals thanpolycrystalline MAPbI3 films,102 indicating that trap densityat perovskite surfaces and GBs is much higher than in bulk.The low thermal stability of perovskites arises in particularbecause organic cations, in the form of MAI, can be lost fromthe crystal during thermal annealing to produce nonstoichio-

Figure 4. Charge carrier trapping and relaxation dynamics in pristine MAPbI3 and MAPbI3 with iodine vacancy in different oxidation states. TheIV−1defect accelerates the recombination only slightly because no trap states are created, while the IV

+1 accelerates charge carrier losses by 2 orders ofmagnitude. Adapted from ref 59. Copyright 2018 American Chemical Society.

Removing nonradiative recombi-nation sites at surfaces, interfa-ces, and grain boundaries is

important to advance perovskiteoptoelectronic properties. PbI2-terminated surfaces maintainlong carrier lifetimes. Surfacepassivation with Lewis bases,

salts, and other additives leads tofurther improvements in chem-ical stability and photophysics.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3792

metric compositions.103 Experiments demonstrate that sig-nificant charge recombination takes place at perovskitesurfaces.104,105 Suppression of surface recombination is ofcrucial importance to further advance PSCs.Experiments show that an excess of the PbI2 precursor

during perovskite synthesis enhances PSC efficiency.105,106 Theexact role of excess PbI2 and the mechanism of the efficiencyenhancement are still matters of debate.104,107 A small excess ofPbI2 in the precursor solution improves device stability andsuppresses the current−voltage hysteresis and ion migration inPSCs.108 Surface passivation can notably improve PSCefficiency. Supasai and co-workers demonstrated a self-inducedpassivation mechanism at perovskite surfaces.109 Chen and co-workers found that upon thermal annealing of MAPbI3 films,the nonradiative recombination was greatly suppressed in thepresence of a rigid PbI2 termination layer.110 Surfacephotovoltage spectroscopy experiments revealed that a PbI2layer located at the perovskite/TiO2 interface could inhibit thecharge recombination.111 A similar effect of excess PbI2 wasobserved in other works, confirming the beneficial role of thePbI2 treatment.112

A large number of computational studies were dedicated toinvestigation of surface properties of perovskites. For example,Haruyama and co-workers examined systematically thethermodynamics properties of MAPbI3 surfaces and concludedthat the PbI2-terminated surface does not introduce subgapstates.113,114 Liu and co-workers suggested that the PbI2-terminated surface has a smaller band gap than the MAI-terminated surface115 and that structural stability andelectronic properties at the MAPbI3/TiO2 interface dependon a particular surface termination.116 Quarti and co-workersemphasized that the different surface terminations, MAI versusPbI2, affect the relative energy alignment at the perovskite/C60interface.117 Using ab initio MD simulations, Mosconi and co-workers showed that the PbI2-terminated surface is morerobust to water degradation compared with the MAItermination.118 Uratani and co-workers investigated chargecarrier trapping by surface defects and concluded that surfacedefects are more likely to form in the PbI2-terminated thanMAI-terminated surfaces.11 Whether excess PbI2 is trulyessential for high-efficiency PSCs remains an open ques-tion,104,105 especially because both MAI- and PbI2-terminatedsurfaces form and coexist under thermodynamic growthconditions.It is important to note that PbI2-terminated surfaces contain

undercoordinated Pb2+ ions, which may accumulate photo-generated charges and possibly increase recombination at theinterface with a charge transport layer. Passivation of theundercoordinated Pb2+ ions at the perovskite surface can beachieved by adding organic molecules to the precursorsolutions. Noel et al. reported that electron-rich Lewis bases,such as pyridine or thiophene, passivate perovskite surfaces,reducing the nonradiative recombination rate by an order ofmagnitude.119 Similar ideas for passivation of surface defectswere adopted by other research groups, which utilized variouspassivation agents.120−122 A comprehensive analysis of surfacepassivation in PSCs is provided in a recent review.123 Theexperimental characterization of PSC devices and electronicstructure calculations provide important information on theinfluence of surfaces and their passivation on the PSCperformance. Further key insights are generated by time-resolved spectroscopies and quantum dynamics simulations.

In addition to perovskite surfaces, interfaces betweenperovskite grains are also attracting significant attention.Many experiments show that GBs and structural disorderstrongly impact the optoelectronic quality.36,102,124−129 Thereexist conflicting reports regarding the effect of GBs on thephotovoltaic performance of perovskites. A number ofexperimental publications reported that solution-processedpolycrystalline MAPbI3 films containing GBs have excellentcharge-transporting properties, resulting in high photonconversion efficiencies, exceeding 20%.36,124 Such data implythat GBs do not open up additional channels to charge carrierlosses and may even be beneficial. This phenomenon can beexplained by the recent first-principles calculations of Yin et al.,who showed that structural disorder at GBs, resulting, forinstance, in formation of Pb−Pb bonds that are not possible inpristine perovskites, creates no subgap states because of therelatively large distance between the two Pb2+ ions.127 At thesame time, several studies found that charge carrier lifetimesgrow with increasing grain size in perovskite films,125,126

indicating that GBs are related to nonradiative chargerecombination and deteriorate device performance. Moreover,the halide ions were demonstrated to segregate spontaneouslyinto GBs and eliminate trap states, playing beneficial roles inpassivating GBs in perovskite polycrystalline films.128 Athorough understanding of the carrier dynamics at perovskiteGBs and interfaces, including mechanisms of suppression ofthe nonradiative recombination through proper passivation, isof crucial importance for further development of PSCs.Motived by the above experimental and theoretical

observations, Tong et al. investigated the nonradiativeelectron−hole recombination in MAPbI3 with PbI2- andMAI-terminated surfaces representing PbI2-rich and PbI2-poor conditions.130 The simulations demonstrated that thecharge carrier relaxation is slow in the PbI2-terminated MAPbI3compared to the MAI-terminated MAPbI3 (Figure 5),

primarily because of a smaller NA electron−phonon couplingin the PbI2-terminated surface. The MA cations exposed at theMAI-terminated surface are more mobile; couple electrostati-cally to the charge carriers; and through high-frequency modes,accelerate charge−phonon energy exchange. The MA cationsin the PbI2-terminated MAPbI3 are much less mobile andcontribute little to the charge carrier recombination. It isinteresting that even though the electrons and holes at the

Figure 5. Simulation structure and energy level diagram fornonradiative recombination at the PbI2- and MAI-terminatedperovskite surfaces. Adapted from ref 130. Copyright 2018 AmericanChemical Society. The MAI-terminated MAPbI3 surface has moremobile MA cations, which couple electrostatically to charge carriersthrough high-frequency modes.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3793

VBM and CBM are localized on the inorganic lattice, themobile organic MA species are able to participate in the chargerecombination because of their large dipole moments, whichgenerate long-range oscillating electrostatic fields, and byinfluencing motions of the inorganic lattice.The mechanism of the decreased nonradiative charge

recombination in perovskites passivated by Lewis basemolecules, such as thiophene and pyridine, was investigatedby ab initio NA-MD.62 The simulations showed that bothpyridine and thiophene passivations extend charge carrierlifetimes by reducing the NA electron−phonon coupling andquantum coherence time, while retaining the energy gap. TheNA coupling is reduced because pyridine and thiophenelocalize the charge density (Figure 6) and decrease the overlap

between the VBM and CBM wave functions. The localizationoccurs by different mechanisms, though. Thiophene pushes theelectron density away from the surface because of eliminationof the unsaturated chemical bonds. This is the traditionalpassivation mechanism. In comparison, pyridine introduces alocalized surface state by forming a chemical bond between theN atom of the molecule and the under-coordinated Pb ion ofthe perovskite surface. Importantly, this surface state is veryclose to the CB edge and does not trap electrons. Both Lewisbase molecules introduce high-frequency vibrational modesthat shorten coherence time and slow down the quantumdynamics131 of phonon-driven charge recombination.

NA-MD calculations demonstrated that GBs accelerate thenonradiative electron−hole recombination, while chlorinedoping of the boundary can repair the detrimental effectsproduced by the GBs.61 Creation of midgap trap states is themain mechanism of the accelerated recombination. By

breaking the perovskite crystalline symmetry, GBs createlocalized states which overlap less than the CBM and VBM inthe pristine perovskite (Figure 7). In the absence of midgapstates,127 such localization can reduce NA coupling and slowdown the recombination, leading to efficient PSCs.36,124 Thecalculated NA couplings were larger at the GB than in thepristine perovskite because of enhanced atomic motions withinthe less structured boundary region.61 The simulations suggestthat charge recombination depends strongly on the type ofcreated GBs and that annealing of perovskite films can heal GBdefects, eliminating midgap states, and stabilize boundarystructures, reducing atomic fluctuations, thereby extendingcharge carrier lifetimes.60 GB defects can be also healed bydoping, e.g. with Cl atoms. Cl is more electronegative than I,and its orbitals are located deeper inside the VB. Thus, Cldopants do not contribute to the VBM charge density. Becausedopants are most easily deposited at boundaries and surfaces,doping of a MAPbI3 GB with chlorines pushes the VBMcharge density away from the boundary (Figure 7), removingthe trap states. Further, by moving faster, the lighter chlorinesreduce quantum coherence time and slow down quantumdynamics, as exemplified by the quantum Zeno effect.131

The examples of the NA-MD studies of the nonradiativecharge carrier recombination in the perovskites containingoriginal and passivated surfaces and GBs demonstrate acomplex interplay of several factors, including midgap states,charge localization, and involvement of additional motions dueto both passivating species and reduced structural stability atsurfaces. Modifications of the electronic structure in theextended defect regions alter energy gaps and rates of inelasticand elastic charge−phonon scattering. A thorough under-standing of these factors allows one to engineer benign surfacesthat reduce carrier losses and maintain high PSC efficiencies.

Optimizing Composition To Control Perovskite NonradiativeElectron−Hole Recombination. The family of halide perovskiteswith the ABX3 general chemical formula contains a largenumber of compounds obtained by varying the threecomponents: the A-site monovalent cation, the B-site divalentcation, or the halide anion. Additional materials can be createdby mixing multiple components, e.g., A, A′ or X, X′, within thesame structure. Such variability provides an important meansto introduce doping impurities into the crystal lattice or tosynthesize new materials by composition engineering. Here,

Figure 6. CBM charge densities in (a) the pristine MAPbI3−xClx(001) surface and the surface passivated by (b) pyridine and (c)thiophene molecules. Adapted from ref 62. Copyright 2018 AmericanChemical Society.

The role of grain boundaries inperovskite solar cells is still underdebate. They facilitate chargeseparation while also providingcarrier recombination sites. Someboundaries are more detrimentalthan others. Boundary annealingand doping can repair the detri-

mental effects.

Tuning chemical composition, forexample, mixing of halides or A-site cations, is a viable way to

modulate optoelectronic proper-ties. Halides affect electronic

properties and carrier lifetimes,because they vary in electro-

negativity and introduce differentphonon frequencies. A-site cati-ons do not contribute to bandedges, but they influence inor-ganic lattice parameters, vibra-tional dynamics, and (thermal)

disorder.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3794

the goal is to change the optoelectronic properties of thesemiconductor and to enhance its long-term stability andefficiency. In particular, mixing the halide composition is themost common approach to tune the optoelectronic properties.It is interesting to note that the best efficiency PSCs are basedon the hybrid perovskites with mixed MA/FA cations and I/Brhalide anions.132

Halide mixing is well-illustrated by doping MAPbI3 withchlorine. Chlorine can be incorporated into MAPbI3 films atrelatively large concentration,133 leading to the release of MAand partial formation of the MACl compound. As a result,mixed halide MAPbI3−xClx perovskite are produced.134

Theoretically, the band gap is almost unaltered by incorpo-ration of Cl into MAPbI3 films.135 Recent experiments foundthat Cl doping optimizes film morphology,136 improves carriermobility,10 and delays luminescence decay.137 It is thusexpected that incorporation of Cl into perovskite materialsaffects the fundamental electronic properties. Replacement ofiodine at the MAPbI3/TiO2 interface with Cl redistributescharges and changes electronic coupling between CB states ofMAPbI3 and TiO2, which may influence the interface chargeinjection rate, according to De Angelis and co-workers.138 As aresult, Cl doping of the MAPbI3/TiO2 interface delays chargerecombination, as shown by Long and Prezhdo.63 Mixed-halideMAPbI3−xClx perovskites show a comparably better perform-ance than the pristine MAPbI3 perovskite.139 A similarbeneficial role of halide substitution was also observed for Brdoping.140 Tuning halide composition in nanocrystals of all-inorganic perovskites received much attention as well.141−143

Bulk CsPbI3 has low phase stability because it does not satisfythe Goldschmidt tolerance factor rule. It can be stabilized inthe quantum dot (QD) form because of the high surface-to-volume ratio. It was shown that replacement of iodine atomswith bromines in CsPbI3 QDs decreases the carrier diffusionlength.144 Tong et al. measured the averaged carrier lifetime inCsPbX3 QDs and demonstrated that it decreases as thechemical compositions of the halide anions changes from I toBr to Cl.145 Similar studies were reported with hybrid organic−inorganic perovskites.146,147

In addition to manipulation of the halide compositions,tuning the A-site cations for targeted properties drawsattention. It was demonstrated that mixed A-site cationscould stabilize the perovskite crystal phase because of changes

in the configurational entropy.148 First-principles calculationsrevealed that the band edge states in perovskites originate fromthe PbX3

− inorganic sublattice, suggesting that the A-sitecation should not influence the optoelectronic propertiesdirectly. In particular, the VBM is dominated by 5p orbitals ofthe halide and CBM is composed of 6p orbitals ofPb.19,33,149,150 However, band gap changes induced by A-sitecations were reported. This is likely correlated with the stericeffects: motions of the A-site cations couple to deformation ofthe inorganic lattice, leading to changes in the Pb−I bondlengths and the Pb−I−Pb bond angle.151,152 Recent photo-luminescence experiments demonstrated that the carrierlifetimes are significantly different in perovskites containingMA, FA, GA, and Cs.153 Moreover, partial replacement of MAwith the larger FA154 and GA155 cations can significantlyextend carrier lifetimes. These and other observations require afundamental understanding that is achieved by quantumdynamics simulations.Motivated by the above-mentioned experiments, Liu and

Prezhdo simulated charge carrier dynamics in Cl-dopedMAPbI3.

156 It was shown that replacement of iodine atomswith chlorines reduces the charge recombination rate,reproducing the experimental observations. The chargerecombination was slowed down by the Cl doping becauseof decreased NA electron−phonon coupling and shortenedquantum coherence time. Cl doping breaks the symmetry ofthe perovskite crystal structure, inducing localization in theVBM and CBM wave functions. The localization is furtherenhanced by the fact that chlorine is more electronegative thaniodine and contributes little to the VBM. The decreasedVBM−CBM wave function overlap weakens the NA coupling(Figure 8). Further, the quantum coherence time is shorter inthe presence of chlorines, both because the wave functionlocalization and because lighter chlorines, compared to iodines,introduce higher frequency phonon modes that lead to fasterloss of coherence within the electronic subsystem.The influence of halide composition on the nonradiative

charge recombination in CsPbBr3 perovskite QDs wasstudied.56 It was shown that replacement of Br atoms with Iatoms in CsPbBr3 QDs extends the charge carrier lifetime by afactor of almost 8, whereas half replacement extends thelifetime nearly 5 times (Figure 9). The mechanism of thereduction of the nonradiative charge recombination rate was

Figure 7. VBM and CBM charge density in (a) pristine MAPbI3, (b) Σ5 (012) GB, and (c) Cl-doped Σ5 (012) GB. VBM and CBM aredelocalized in panel a and become localized on inorganic atoms in the vicinity of the GB in panels b and c. Doping by Cl atoms pushes chargedensity away from the GB. Adapted from ref 61. Copyright 2016 American Chemical Society.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3795

rationalized by the smaller NA electron−phonon coupling.Iodines are heavier than bromines and have a slower velocity.The NA electron−phonon coupling is proportional to thenuclear velocity, and therefore, it is smaller in CsPbI3 thanCsPbBr3. The velocity argument also applies to theCsPbI1.5Br1.5 mixed perovskite, compared to CsPbI3. The NAcoupling is reduced further in the mixed system because of thelowered symmetry and the disorder and charge localizationintroduced by the mixing.NA-MD calculations demonstrated that nonradiative charge

recombination in MAPbI3 is reduced by partial replacement ofMA with the larger FA and GA cations (Figure 10).157 Thelarger cations stiffen the perovskite structure and decrease theextent of thermal atomic fluctuations. In particular, replace-ment of 25% MA with FA or GA decreases fluctuations of theinorganic PbI lattice by 10% and nearly 50%, respectively. Atthe same time, the band gap remains unchanged. Because theinorganic lattice supports the VBM and CBM wave functions,and the NA coupling depends on overlap of these wavefunctions at sequential MD time steps, the decreasedfluctuations give smaller NA coupling and slower chargerecombination. The stiffening of the perovskite structure,leading to reduction of atomic fluctuations, provides animportant design principle for extending charge carrier

lifetimes. The opposite effect was seen in the GB system61

discussed above. The looser structure around the GB regionexhibited larger thermal atomic fluctuations and increased NAelectron−phonon coupling.Similarly, another work of Long and co-workers investigated

the influence of the A-site cations on the charge carrierdynamics in lead bromide perovskites.65 The excited-statelifetime in the CsPbBr3 perovskite was compared to those inMAPbBr3 and FAPbBr3 perovskites. It was shown that thenonradiative charge recombination is slowest in FAPbBr3,followed by MAPbBr3 and CsPbBr3 (Figure 11). The

difference in the recombination rates can be mainly attributedto the smaller NA electron−phonon coupling, because theenergy gaps and coherence times are almost the same in theinvestigated systems. The larger FA and MA cations suppressmotions of the inorganic lattice, and correspondingly, the NAcoupling is reduced.The influence of doping on the charge recombination across

the MAPbI3/TiO2 interface was investigated (Figure 12).63

The recombination occurs after photoexcitation of MAPbI3and subsequent charge collection by TiO2, which is used as anelectron acceptor in many types of solar cells. Therecombination is much slower than the injection51 andconstitutes the main mechanism of carrier losses in suchheterostructures. It was found that Cl and Br doping slowdown the nonradiative charge recombination across theinterface, whereas Sn doping accelerates the recombination,

Figure 8. VBM and CBM charge densities in MAPbI3 and MAPbI2Cl.Adapted from ref 156. Copyright 2015 American Chemical Society.

Figure 9. Charge recombination dynamics in CsPbBr3, CsPbBr1.5I1.5,and CsPbI3 QDs. Adapted from ref 56. Copyright 2018 AmericanChemical Society.

Figure 10. Charge recombination dynamics in MAPbI3,MA0.75FA0.25PbI3, and MA0.75GA0.25PbI3 perovskites. Adapted fromref 157. Copyright 2018 American Chemical Society.

Figure 11. Charge recombination dynamics in CsPbBr3, MAPbBr3,and FAPbBr3. Adapted from ref 65. Copyright 2018 AmericanChemical Society.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3796

in agreement with the experiments.2,158−160 The influence ofthe Cl, Br, and Sn doping on the interfacial chargerecombination was rationalized by the following arguments:First, the Cl and Br atoms have smaller radii than I. Replacing Iatoms with Cl and Br at the interface by the doping decreasesthe bonding interactions between MAPbI3 and TiO2. A similareffect occurs when Pb is replaced with Sn. Second, Cl and Brare more electronegative than I, and therefore, their states aredeep inside the VB. The VBM is formed primarily by atomicorbitals of iodine atoms in both pristine and doped MAPbI3/TiO2. Replacing interfacial iodines with Cl and Br pushes theVBM, and hence the photogenerated hole, away from theinterfacial region, reducing the electron−hole overlap and theNA coupling. In contrast, Sn contributes to the frontierorbitals, and being lighter than Pb, it has a larger nuclearvelocity and produces a stronger NA coupling. Therecombination is driven primarily by low-frequency phononmodes which can be attributed to bending and stretching ofthe inorganic lattice.

Analysis of the NA-MD simulation of the nonradiativecharge carrier recombination in the doped and mixedperovskites reveals several design principles that can be usedto improve PSC performance. Mixed systems break the perfectsymmetries of single-component perovskites, leading to acontrolled localization and separation of electrons and holes,reducing their interaction. At the same time, the periodicstructure of the inorganic lattice is not destroyed, and efficientcharge transport along the lattice is sustained. The NAelectron−phonon coupling responsible for the nonradiativeelectron−hole recombination depends on the extent of thermalatomic motions, mostly of the atoms of the inorganic lattice.Making the perovskite structure stiffer by judicious choice ofthe perovskite components, e.g. by mixing in larger A-sitecations, the lattice fluctuations can be reduced and the NA

coupling decreased. It is important not to modify the structuresignificantly. For example, changing the lattice parameters orelectronegativity of the inorganic atoms can alter thefundamental band gap and change the light absorptioncharacteristics. Thus, doping provides an important means tofine-tune the perovskite properties.

In summary, hybrid organic−inorganic perovskites con-stitute a promising class of materials that attract a greatnumber of investigations for a variety of applications. Thecompositional complexity of perovskites increases thepossibility of introduction of point defects, boundaries,dopants, partial chemical segregation, and other imperfections.Deep understanding and identification of the defects that areresponsible for nonradiative electron−hole recombination arevery important for further advancement of PSCs and otherdevices. Many publications demonstrate that hybrid perov-skites exhibit an unusual defect science that makes manydefects benign for charge transport. Other publications showthat local and extended defects can create charge traps andaccelerate nonradiative electron−hole recombination. Chargerecombination limits the ability of carrier extraction to chargetransport layers and determines the overall PSC performance.In this Perspective, we discussed the role of defects in chargeand energy losses in hybrid perovskites, considered pointdefects, surfaces, grain boundaries, interfaces, dopants, andnonstoichiometric compositions and analyzed the chargetrapping and recombination dynamics on the basis of NA-MD and TD-DFT calculations. The results highlighted that

Figure 12. Electron (top) and hole (bottom) states for electron−hole recombination across the MAPbI3/TiO2 interface. Adapted from ref 63.Copyright 2015 American Chemical Society.

Controlled symmetry breakingwithout destroying the inorganiclattice can separate electrons andholes, reducing their interaction

while maintaining efficientcharge transport.

Time-domain ab initio nonadia-batic molecular dynamics pro-vides an atomistic description ofthe nonequilibrium processes asthey occur in real time and at theatomistic level. It is computa-

tionally demanding, and its suc-cess in applications to perov-

skites and other modern materi-als relies on efficient and reliableapproximations and algorithms.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3797

precise control of defect formation by regulating crystal growthconditions is very important for eliminating charge traps.Characterization of defects is challenging experimentally.

Usually, defect species appear in low concentrations andpresent small densities of states, making them hard to study byoptical or electrochemical means. Optical transitions betweenground and defect states have small transition dipole moments,because the ground-state wave function is delocalized, whiledefect states are localized. Thus, most defects participating inthe nonradiative relaxation are dark, and straightforwardoptical measurements cannot capture their properties. More-over, it is hard to know experimentally whether predominantdefect states trap electrons or holes, or if electrons and holesrecombine bypassing defect states. Complementary character-ization methods and more advanced techniques are requiredfor such tasks. For example, temperature-dependent admit-tance spectroscopy and combination of transient and steady-state photocurrent and optical absorption spectroscopies havebeen used to identify defect types and energy distributions inPSCs.88,161 Time-resolved two-photon photoemission spec-troscopy is useful to detect subgap states.162 Time-resolvedterahertz photoconductivity and photoluminescence character-izations have been used to measure the decay of excitedcarriers after photoexcitation in hybrid perovskites containingdefects.10,137 The diverse experimental approaches correlatedefect structure with photoluminescence properties and chargecarrier transport in the resultant film and provide data used forphenomenological models.Time-domain ab initio NA-MD simulates quantum dynam-

ics of charge carriers coupled to vibrations motions. NA-MDmimics closely time-resolved experiments and provides anatomistic description of the nonequilibrium processes as theyoccur in real time. Ab initio NA-MD and related techniquesenable accurate determination of state energies and simulationof electron-vibrational dynamics in realistic systems includingdefects, interfaces, mixed and hybrid structures, etc. Suchsimulations are particularly useful to achieve fundamentalunderstanding of nonradiative charge recombination thatconstitutes the major pathway for energy losses in photovoltaicand photocatalytic devices. However, ab initio NA-MD can beapplicable only to relatively small systems composed ofhundreds of atoms and time scales limited to a few to tensof picoseconds. The difficulty resides in the calculation ofexcited-state energies, forces, NA couplings, and decoherencetimes. Several computationally efficient approximations andalgorithms have been introduced, drastically reducing thecomputational cost. They include the Kohn−Sham represen-tation of electronic excitations,46 the classical path approx-imation for neglect of back-reaction,45 and correlation functionevaluation of coherence times under the second-ordercumulant approximation.163 Care must be taken in the propertreatment of the sign/phase of the NA coupling.164 Generally,evaluation of NA couplings requires higher accuracy of wavefunction convergence and smaller MD time steps thanevaluation of electronic energies and forces. The trivial avoidedcrossing problem has to be resolved in NA-MD simulationsinvolving long-range transport.165

Quantum dynamics methods based on semiempiricalextended Huckel theory (EHT)166−168 and tight-bindingDFT (DFTB)169 provide alternatives to ab initio simulations,allowing one to increase the simulation scale by over an orderof magnitude. Semiempirical electronic structure theoriesrequire careful parametrization, work better for chemically

perfect systems, and have to be tested carefully for defectsinvolving unsaturated chemical bonds. In many situations,nuclear dynamics is dominated by thermal fluctuations that canbe sampled within a few picoseconds. In such cases the NAHamiltonian can be iterated multiple times to extend thequantum dynamics simulation to nanoseconds.170,171 In orderto reduce the computational cost of ab initio NA-MD, Akimovdeveloped an efficient quasi-stochastic Hamiltonian NA-MDfor simulation of quantum dynamics on long time scales.172

Proper description of the electronic structure of perovskitesis among the key components of an accurate NA-MDcalculation. The Kohn−Sham description of electronicexcitations provides a very efficient and reasonably accurateway to model excited-state properties.46,173 The lowest-energyexcited states can be described by the delta self-consistent field(ΔSCF) methods.174,175 The next level of accuracy is providedby the GW and Bethe−Salpeter theories and, alternatively, bylinear-response TD-DFT.176,177 The latter is more suitable forNA-MD, because it offers explicit expressions for excited-statewave functions and NA couplings. SOC effects are significantin perovskites because of the presence of heavy elements suchas Pb and I. Fortunately, bare DFT functionals exhibit acancelation of errors associated with electron self-interactionand SOC and give sufficiently accurate results.17,19,98,178−180 Atthe same time, NA coupling may be influenced by SOC,181 andthis effect should be investigated systematically further. If SOCis explicitly included into the calculations, then either hybridDFT functionals or the many-body GW theory should beapplied,182 increasing significantly the computational effort.Despite the significant progress, further development ofaccurate, reliable, and efficient methods is required for large-scale simulation of nonequilibrium processes in perovskitesand other condensed matter and nanoscale systems.Perovskites exhibit a broad range of defects, and it is

important to sample properly various defect types. The pointdefects discussed in this Perspective are well-defined, and theirstructure fluctuates around the equilibrium geometry. Morecomplex defects, such as combinations of point defects and/ordopants, nonstoichiometric systems, surfaces, GBs, etc., cantake a large number of energetically close conformations. Theycan be sampled by combining ab initio electronic structurewith Monte Carlo simulations and modern structure-prediction tools.

Future NA-MD studies of perovskites should investigate thecharge carrier dynamics in larger, more complex systems andinclude more complex processes. A more thorough inves-tigation of the types of systems discussed in the Perspective isneeded as well. For example, the published work61 rationalizesthe detrimental influence of GBs on perovskite properties.183

However, only a single GB was considered because of the highcomputational efforts associated with NA-MD. In reality theinfluence of GBs is more complex184 and can be both negativeand benign.185 Thus, more GB systems should be analyzed.The above examples focused on 3D perovskites, and very few

Further research efforts shouldfocus on investigation of excited-state dynamics in larger, morecomplex systems and includemore complex processes.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3798

quantum dynamics studies of the now popular 2D perovskiteshave been reported thus far.186 2D systems exhibit newfeatures, such as edge states that facilitate charge separation.The reduced dimensionality can lead to higher concentrationsof charge carriers that can exchange energy by Auger-typeinteractions known to be important in 0D semiconductor QDs,1D carbon nanotubes, and 2D transition-metal dichalcoge-nides. With a notable exception,187 Auger-type effects havebeen simulated by NA-MD using NA coupling between many-particle states,188−190 and inclusion of explicit Coulombinteractions between charge carriers is desirable. While existingNA-MD approaches already provide very important insightsinto the charge carrier dynamics in perovskites and other novelmaterials, the applications motivate development of noveltechniques.

Control of quantum coherence provides an intriguing andoften overlooked means to reduce charge carrier recombina-tion.186 Transition from the excited to the ground electronicstate requires formation of quantum coherence between thetwo states. This coherence is destroyed by coupling of theelectronic subsystem to phonons. Shorter coherence timesslow down quantum dynamics,131 reducing the recombinationrate. It is important to keep in mind that faster decoherenceimplies stronger electron−phonon coupling, and strongerelectron−phonon coupling should lead to faster nonradiativerecombination. This apparent contradiction disappears whenone is reminded that nonradiative recombination correspondsto inelastic charge−phonon scattering, while decoherenceoccurs by elastic scattering (pure-dephasing). The inelasticscattering is governed by the NA coupling, which depends onthe matrix element between the VBM and CBM wavefunctions.45 No such matrix element enters expressions forthe decoherence rates.163 For example, replacing some iodineatoms with the more electronegative chlorines and bromineshas little effect on the VBM wave function, which continues toreside on the iodines. At the same time, the lighter Cl and Brintroduce faster phonons and accelerate coherence loss.Decoherence is a quantum dynamics phenomenon, requiringanalysis extending beyond the now routine electronic structurecalculations.The NA-MD calculations indicate that electronic structure

calculations should be performed not only with optimizedstructures but also for room-temperature geometries. Thermaleffects can alter significantly electronic energy levels of defectstates. Atomic motions induce symmetry breaking andlocalization of the electronic states, decreasing electron−holeinteractions. Room-temperature adiabatic MD trajectoriesallow one to estimate the stiffness and structural stability ofperovskites, the importance of polaronic effects, the extent ofstructural and electronic disorder, etc. Requiring morecomputational effort than electronic structure calculations onoptimized geometries, adiabatic MD can also be regarded as aroutine simulation and should be performed with perovskiteson a regular basis. The insights obtained with electronicstructure calculations and adiabatic MD provide the starting

point for the analysis of quantum dynamics of charge carriersby means of NA-MD that remains a state-of-the-art approach.

■ AUTHOR INFORMATIONCorresponding Authors*E-mail: weili@hunau.edu.cn (W.L.).*E-mail: runlong@bnu.edu.cn (R.L.).*E-mail: jftang@hunau.edu.cn (J.T.).*E-mail: prezhdo@usc.edu (O.V.P.).ORCIDWei Li: 0000-0002-9999-5081Run Long: 0000-0003-3912-8899Oleg V. Prezhdo: 0000-0002-5140-7500NotesThe authors declare no competing financial interest.Biographies

Wei Li received a Ph.D. from the Institute of Theoretical Chemistryfrom Jilin University in 2017 and is currently a professor of Chemistryat Hunan Agricultural University. Dr. Li’s research focuses onexcitation dynamics including charge and energy transfer in nanoscalesystems, especially hybrid perovskite materials.

Run Long obtained a Ph.D. in atomic and molecular physics fromShandong University and is currently a professor of Chemistry atBeijing Normal University. Dr. Long’s research interests focus onexcitation dynamics in nanoscale systems, in particular energy andcharge-transfer dynamics in condensed phases and at interfaces.

Jianfeng Tang obtained his Ph.D. from Hunan University and iscurrently a professor of Physics in the College of Science in HunanAgricultural University. His research interests are multiscale computersimulations of thermodynamic, mechanical, and electronic propertiesof nanoscale materials.

Oleg V. Prezhdo is a Professor of Chemistry, Physics and Astronomyat the University of Southern California. He is Editor for The Journalof Physical Chemistry Letters, Journal of Physical Chemistry, and SurfaceScience Reports. His research interests range from fundamental aspectsof semiclassical physics to excitation dynamics in nanoscale andbiological systems.

■ ACKNOWLEDGMENTSW.L. acknowledges startup funding from Hunan AgriculturalUniversity (Grant Nos. 540499818006 and 18QN02). O.V.P.acknowledges funding from the U.S. National ScienceFoundation (Grant No. CHE-1900510) for methods develop-ment support and the U.S. Department of Energy (Grant No.DE-SC0014429) for funding the applications studies. R.L.acknowledges support of the National Science Foundation ofChina, Grant Nos. 21573022 and 51861135101. R.L. alsoacknowledges financial support by the Fundamental ResearchFunds for the Central Universities, the Recruitment Programof Global Youth Experts of China, and the Beijing NormalUniversity Startup.

■ REFERENCES(1) Yang, W. S.; Park, B. W.; Jung, E. H.; Jeon, N. J.; Kim, Y. C.; Lee,D. U.; Shin, S. S.; Seo, J.; Kim, E. K.; Noh, J. H.; Seok, S. I. Iodidemanagement in formamidinium-lead-halide-based perovskite layersfor efficient solar cells. Science 2017, 356, 1376−1379.(2) Zhou, H.; Chen, Q.; Li, G.; Luo, S.; Song, T. B.; Duan, H. S.;Hong, Z.; You, J.; Liu, Y.; Yang, Y. Photovoltaics Interface engineeringof highly efficient perovskite solar cells. Science 2014, 345, 542.

Control of quantum coherenceprovides an intriguing and oftenoverlooked means to reducecharge carrier recombination.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3799

(3) Wang, W.; Tade, M. O.; Shao, Z. Research progress of perovskitematerials in photocatalysis- and photovoltaics-related energy con-version and environmental treatment. Chem. Soc. Rev. 2015, 44,5371−408.(4) Wang, M.; Lin, S. Anisotropic and Ultralow Phonon ThermalTransport in Organic-Inorganic Hybrid Perovskites: AtomisticInsights into Solar Cell Thermal Management and ThermoelectricEnergy Conversion Efficiency. Adv. Funct. Mater. 2016, 26, 5297−5306.(5) Ke, Y.; Wang, N. N.; Kong, D. C.; Cao, Y.; He, Y. R.; Zhu, L.;Wang, Y. M.; Xue, C.; Peng, Q. M.; Gao, F.; Huang, W.; Wang, J. P.Defect Passivation for Red Perovskite Light-Emitting Diodes withImproved Brightness and Stability. J. Phys. Chem. Lett. 2019, 10, 380−385.(6) Wang, Y.; Zou, R. M.; Chang, J.; Fu, Z. W.; Cao, Y.; Zhang, L.D.; Wei, Y. Q.; Kong, D. C.; Zou, W.; Wen, K. C.; Fan, N.; Wang, N.N.; Huang, W.; Wang, J. P. Tin-Based Multiple Quantum WellPerovskites for Light-Emitting Diodes with Improved Stability. J. Phys.Chem. Lett. 2019, 10, 453−459.(7) Shockley, W.; Queisser, H. J. Detailed Balance Limit ofEfficiency of p-n Junction Solar Cells. J. Appl. Phys. 1961, 32, 510−519.(8) Zhong, S.; Huang, Z.; Lin, X.; Zeng, Y.; Ma, Y.; Shen, W. High-efficiency nanostructured silicon solar cells on a large scale realizedthrough the suppression of recombination channels. Adv. Mater. 2015,27, 555−661.(9) Peaker, A. R.; Markevich, V. P.; Hamilton, B.; Parada, G.; Dudas,A.; Pap, A.; Don, E.; Lim, B.; Schmidt, J.; Yu, L.; Yoon, Y.; Rozgonyi,G. Recombination via point defects and their complexes in solarsilicon. Phys. Status Solidi A 2012, 209, 1884−1893.(10) Stranks, S. D.; Eperon, G. E.; Grancini, G.; Menelaou, C.;Alcocer, M. J.; Leijtens, T.; Herz, L. M.; Petrozza, A.; Snaith, H. J.Electron-hole diffusion lengths exceeding 1 micrometer in anorganometal trihalide perovskite absorber. Science 2013, 342, 341−344.(11) Uratani, H.; Yamashita, K. Charge Carrier Trapping at SurfaceDefects of Perovskite Solar Cell Absorbers: A First-Principles Study. J.Phys. Chem. Lett. 2017, 8, 742−746.(12) Yang, D.; Ming, W.; Shi, H.; Zhang, L.; Du, M.-H. FastDiffusion of Native Defects and Impurities in Perovskite Solar CellMaterial CH3NH3PbI3. Chem. Mater. 2016, 28, 4349−4357.(13) Mosconi, E.; Meggiolaro, D.; Snaith, H. J.; Stranks, S. D.;Angelis, F. D. Light-induced annihilation of Frenkel defects in organo-lead halide perovskites. Energy Environ. Sci. 2016, 9, 3180−3187.(14) Yang, J. H.; Yin, W. J.; Park, J. S.; Wei, S. H. Self-regulation ofcharged defect compensation and formation energy pinning insemiconductors. Sci. Rep 2015, 5, 16977.(15) Walsh, A.; Scanlon, D. O.; Chen, S.; Gong, X. G.; Wei, S.-H.Self-Regulation Mechanism for Charged Point Defects in HybridHalide Perovskites. Angew. Chem., Int. Ed. 2015, 54, 1791−1794.(16) Du, M.-H. Density Functional Calculations of Native Defects inCH3NH3PbI3: Effects of Spin−Orbit Coupling and Self-InteractionError. J. Phys. Chem. Lett. 2015, 6, 1461−1466.(17) Buin, A.; Comin, R.; Xu, J.; Ip, A. H.; Sargent, E. H. Halide-Dependent Electronic Structure of Organolead Perovskite Materials.Chem. Mater. 2015, 27, 4405−4412.(18) Azpiroz, J. M.; Mosconi, E.; Bisquert, J.; Angelis, F. D. Defectmigration in methylammonium lead iodide and its role in perovskitesolar cell operation. Energy Environ. Sci. 2015, 8, 2118−2127.(19) Yin, W.-J.; Shi, T.; Yan, Y. Unusual defect physics inCH3NH3PbI3 perovskite solar cell absorber. Appl. Phys. Lett. 2014,104, No. 063903.(20) Zhu, X.; Lee, J.; Lu, W. D. Iodine Vacancy Redistribution inOrganic−Inorganic Halide Perovskite Films and Resistive SwitchingEffects. Adv. Mater. 2017, 29, 1700527.(21) Yang, W. S.; Park, B.-W.; Jung, E. H.; Jeon, N. J.; Kim, Y. C.;Lee, D. U.; Shin, S. S.; Seo, J.; Kim, E. K.; Noh, J. H.; Seok, S. I.Iodide management in formamidinium-lead-halide−based perovskitelayers for efficient solar cells. Science 2017, 356, 1376−1379.

(22) Lorenzon, M.; Sortino, L.; Akkerman, Q.; Accornero, S.;Pedrini, J.; Prato, M.; Pinchetti, V.; Meinardi, F.; Manna, L.; Brovelli,S. Role of Nonradiative Defects and Environmental Oxygen onExciton Recombination Processes in CsPbBr3 Perovskite Nanocryst-als. Nano Lett. 2017, 17, 3844−3853.(23) Zhang, Z.-Y.; Wang, H.-Y.; Zhang, Y.-X.; Hao, Y.-W.; Sun, C.;Zhang, Y.; Gao, B.-R.; Chen, Q.-D.; Sun, H.-B. The Role of Trap-assisted Recombination in Luminescent Properties of OrganometalHalide CH3NH3PbBr3 Perovskite Films and Quantum Dots. Sci. Rep.2016, 6, 27286.(24) Steirer, K. X.; Schulz, P.; Teeter, G.; Stevanovic, V.; Yang, M.;Zhu, K.; Berry, J. J. Defect Tolerance in Methylammonium LeadTriiodide Perovskite. ACS Energy Letters 2016, 1, 360−366.(25) Kye, Y. H.; Yu, C. J.; Jong, U. G.; Chen, Y.; Walsh, A. CriticalRole of Water in Defect Aggregation and Chemical Degradation ofPerovskite Solar Cells. J. Phys. Chem. Lett. 2018, 9, 2196−2201.(26) Son, D. Y.; Kim, S. G.; Seo, J. Y.; Lee, S. H.; Shin, H.; Lee, D.;Park, N. G. Universal Approach toward Hysteresis-Free PerovskiteSolar Cell via Defect Engineering. J. Am. Chem. Soc. 2018, 140, 1358−1364.(27) Gruverman, A.; Bi, C.; Huang, J.; Sharma, P.; Wang, Q.; Dong,Q.; Yuan, Y.; Shao, Y.; Xiao, Z. Giant switchable photovoltaic effect inorganometal trihalide perovskite devices. Nat. Mater. 2015, 14, 193.(28) Zhang, Y.; Liu, M.; Eperon, G. E.; Leijtens, T. C.; McMeekin,D.; Saliba, M.; Zhang, W.; de Bastiani, M.; Petrozza, A.; Herz, L. M.;Johnston, M. B.; Lin, H.; Snaith, H. J. Charge selective contacts,mobile ions and anomalous hysteresis in organic−inorganic perovskitesolar cells. Mater. Horiz. 2015, 2, 315−322.(29) Walukiewicz, W.; Rey-Stolle, I.; Han, G.; Jaquez, M.; Broberg,D.; Xie, W.; Sherburne, M.; Mathews, N.; Asta, M. BistableAmphoteric Native Defect Model of Perovskite Photovoltaics. J.Phys. Chem. Lett. 2018, 9, 3878−3885.(30) Even, J.; Pedesseau, L.; Jancu, J.-M.; Katan, C. Importance ofSpin−Orbit Coupling in Hybrid Organic/Inorganic Perovskites forPhotovoltaic Applications. J. Phys. Chem. Lett. 2013, 4, 2999−3005.(31) Meggiolaro, D.; De Angelis, F. First-Principles Modeling ofDefects in Lead Halide Perovskites: Best Practices and Open Issues.Acs Energy Letters 2018, 3, 2206−2222.(32) Kang, J.; Wang, L.-W. High Defect Tolerance in Lead HalidePerovskite CsPbBr3. J. Phys. Chem. Lett. 2017, 8, 489−493.(33) Du, M. H. Efficient carrier transport in halide perovskites:theoretical perspectives. J. Mater. Chem. A 2014, 2, 9091−9098.(34) Brandt, R. E.; Stevanovic, V.; Ginley, D. S.; Buonassisi, T.Identifying defect-tolerant semiconductors with high minority-carrierlifetimes: beyond hybrid lead halide perovskites. MRS Commun. 2015,5, 265−275.(35) Ball, J. M.; Petrozza, A. Defects in perovskite-halides and theireffects in solar cells. Nature Energy 2016, 1, 16149.(36) Stranks, S. D. Nonradiative Losses in Metal Halide Perovskites.ACS Energy Letters 2017, 2, 1515−1525.(37) Huang, J.; Yuan, Y.; Shao, Y.; Yan, Y. Understanding thephysical properties of hybrid perovskites for photovoltaic applications.Nature Reviews Materials 2017, 2, 17042.(38) Prezhdo, O. V. Photoinduced Dynamics in SemiconductorQuantum Dots: Insights from Time-Domain ab Initio Studies. Acc.Chem. Res. 2009, 42, 2005−2016.(39) Prezhdo, O. V.; Duncan, W. R.; Prezhdo, V. V. Dynamics of thePhotoexcited Electron at the Chromophore−Semiconductor Inter-face. Acc. Chem. Res. 2008, 41, 339−348.(40) Long, R.; Prezhdo, O. V.; Fang, W. Nonadiabatic chargedynamics in novel solar cell materials. Wiley Interdisciplinary Reviews:Computational Molecular Science 2017, 7, No. e1305.(41) Wang, L.; Long, R.; Prezhdo, O. V. Time-domain ab initiomodeling of photoinduced dynamics at nanoscale interfaces. Annu.Rev. Phys. Chem. 2015, 66, 549−79.(42) Akimov, A. V.; Prezhdo, O. V. Persistent Electronic CoherenceDespite Rapid Loss of Electron−Nuclear Correlation. J. Phys. Chem.Lett. 2013, 4, 3857−3864.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3800

(43) Kilina, S. V.; Neukirch, A. J.; Habenicht, B. F.; Kilin, D. S.;Prezhdo, O. V. Quantum Zeno Effect Rationalizes the PhononBottleneck in Semiconductor Quantum Dots. Phys. Rev. Lett. 2013,110, 180404.(44) Akimov, A. V.; Prezhdo, O. V. Advanced Capabilities of thePYXAID Program: Integration Schemes, Decoherence Effects,Multiexcitonic States, and Field-Matter Interaction. J. Chem. TheoryComput. 2014, 10, 789−804.(45) Akimov, A. V.; Prezhdo, O. V. The PYXAID Program for Non-Adiabatic Molecular Dynamics in Condensed Matter Systems. J.Chem. Theory Comput. 2013, 9, 4959−4972.(46) Craig, C. F.; Duncan, W. R.; Prezhdo, O. V. Trajectory SurfaceHopping in the Time-Dependent Kohn-Sham Approach for Electron-Nuclear Dynamics. Phys. Rev. Lett. 2005, 95, 163001.(47) Akimov, A. V.; Prezhdo, O. V. Large-Scale Computations inChemistry: A Bird’s Eye View of a Vibrant Field. Chem. Rev. 2015,115, 5797−5890.(48) Wang, L.; Akimov, A.; Prezhdo, O. V. Recent Progress inSurface Hopping: 2011−2015. J. Phys. Chem. Lett. 2016, 7, 2100−2112.(49) Parandekar, P. V.; Tully, J. C. Mixed quantum-classicalequilibrium. J. Chem. Phys. 2005, 122, 094102.(50) Tully, J. C. Molecular dynamics with electronic transitions. J.Chem. Phys. 1990, 93, 1061−1071.(51) Long, R.; Fang, W. H.; Prezhdo, O. V. Strong Interaction at thePerovskite/TiO2 Interface Facilitates Ultrafast Photoinduced ChargeSeparation: A Nonadiabatic Molecular Dynamics Study. J. Phys. Chem.C 2017, 121, 3797−3806.(52) Qiu, J.; Bai, X.; Wang, L. Crossing Classified and CorrectedFewest Switches Surface Hopping. J. Phys. Chem. Lett. 2018, 9, 4319−4325.(53) Qiu, J.; Bai, X.; Wang, L. Subspace Surface Hopping with Size-Independent Dynamics. J. Phys. Chem. Lett. 2019, 10, 637−644.(54) Jaeger, H. M.; Fischer, S.; Prezhdo, O. V. Decoherence-inducedsurface hopping. J. Chem. Phys. 2012, 137, 22A545.(55) Jankowska, J.; Prezhdo, O. V. Ferroelectric Alignment ofOrganic Cations Inhibits Nonradiative Electron−Hole Recombina-tion in Hybrid Perovskites: Ab Initio Nonadiabatic MolecularDynamics. J. Phys. Chem. Lett. 2017, 8, 812−818.(56) He, J.; Vasenko, A. S.; Long, R.; Prezhdo, O. V. HalideComposition Controls Electron−Hole Recombination in Cesium−Lead Halide Perovskite Quantum Dots: A Time Domain Ab InitioStudy. J. Phys. Chem. Lett. 2018, 9, 1872−1879.(57) He, J.; Fang, W. H.; Long, R.; Prezhdo, O. V. Superoxide/Peroxide Chemistry Extends Charge Carriers’ Lifetime but Under-mines Chemical Stability of CH3NH3PbI3 Exposed to Oxygen:Time-Domain ab Initio Analysis. J. Am. Chem. Soc. 2019, 141, 5798−5807.(58) Li, W.; Liu, J.; Bai, F.-Q.; Zhang, H.-X.; Prezhdo, O. V. HoleTrapping by Iodine Interstitial Defects Decreases Free Carrier Lossesin Perovskite Solar Cells: A Time-Domain Ab Initio Study. ACSEnergy Letters 2017, 2, 1270−1278.(59) Li, W.; Sun, Y. Y.; Li, L.; Zhou, Z.; Tang, J.; Prezhdo, O. V.Control of Charge Recombination in Perovskites by Oxidation Stateof Halide Vacancy. J. Am. Chem. Soc. 2018, 140, 15753−15763.(60) Wang, Y.; Fang, W. H.; Long, R.; Prezhdo, O. V. SymmetryBreaking at MAPbI3 Perovskite Grain Boundaries Suppresses ChargeRecombination: Time-Domain ab Initio Analysis. J. Phys. Chem. Lett.2019, 10, 1617−1623.(61) Long, R.; Liu, J.; Prezhdo, O. V. Unravelling the Effects ofGrain Boundary and Chemical Doping on Electron−Hole Recombi-nation in CH3NH3PbI3 Perovskite by Time-Domain AtomisticSimulation. J. Am. Chem. Soc. 2016, 138, 3884−3890.(62) Liu, L.; Fang, W. H.; Long, R.; Prezhdo, O. V. Lewis BasePassivation of Hybrid Halide Perovskites Slows Electron-HoleRecombination: Time-Domain Ab Initio Analysis. J. Phys. Chem.Lett. 2018, 9, 1164−1171.

(63) Long, R.; Prezhdo, O. V. Dopants Control Electron−HoleRecombination at Perovskite−TiO2 Interfaces: Ab Initio Time-Domain Study. ACS Nano 2015, 9, 11143−11155.(64) Li, W.; Tang, J.; Casanova, D.; Prezhdo, O. V. Time-Domain abInitio Analysis Rationalizes the Unusual Temperature Dependence ofCharge Carrier Relaxation in Lead Halide Perovskite. ACS EnergyLetters 2018, 3, 2713−2720.(65) He, J.; Fang, W.-H.; Long, R. Unravelling the Effects of A-SiteCations on Nonradiative Electron−Hole Recombination in LeadBromide Perovskites: Time-Domain ab Initio Analysis. J. Phys. Chem.Lett. 2018, 9, 4834−4840.(66) Zhang, Z.; Long, R.; Tokina, M. V.; Prezhdo, O. V. Interplaybetween Localized and Free Charge Carriers Can Explain HotFluorescence in the CH3NH3PbBr3 Perovskite: Time-Domain AbInitio Analysis. J. Am. Chem. Soc. 2017, 139, 17327−17333.(67) Zhang, L.; Vasenko, A. S.; Zhao, J.; Prezhdo, O. V. Mono-Elemental Properties of 2D Black Phosphorus Ensure ExtendedCharge Carrier Lifetimes under Oxidation: Time-Domain Ab InitioAnalysis. J. Phys. Chem. Lett. 2019, 10, 1083−1091.(68) Akimov, A. V.; Muckerman, J. T.; Prezhdo, O. V. NonadiabaticDynamics of Positive Charge during Photocatalytic Water Splitting onGaN(10−10) Surface: Charge Localization Governs SplittingEfficiency. J. Am. Chem. Soc. 2013, 135, 8682−8691.(69) Prezhdo, O. V. Multiple excitons and the electron-phononbottleneck in semiconductor quantum dots: An ab initio perspective.Chem. Phys. Lett. 2008, 460, 1−9.(70) Long, R.; Fang, W. H.; Prezhdo, O. V. Moderate HumidityDelays Electron-Hole Recombination in Hybrid Organic-InorganicPerovskites: Time-Domain Ab Initio Simulations Rationalize Experi-ments. J. Phys. Chem. Lett. 2016, 7, 3215−3222.(71) Long, R.; Prezhdo, O. V. Asymmetry in the Electron and HoleTransfer at a Polymer-Carbon Nanotube Heterojunction. Nano Lett.2014, 14, 3335−3341.(72) Jankowska, J.; Prezhdo, O. V. Real-Time Atomistic Dynamicsof Energy Flow in an STM Setup: Revealing the Mechanism ofCurrent-Induced Molecular Emission. J. Phys. Chem. Lett. 2018, 9,3591−3597.(73) Forde, A.; Inerbaev, T.; Kilin, D. Spinor Dynamics in Pristineand Mn2+-Doped CsPbBr3 NC: Role of Spin−Orbit Coupling inGround- and Excited-State Dynamics. J. Phys. Chem. C 2018, 122,26196−26213.(74) Zhang, Z. S.; Liu, L. H.; Fang, W. H.; Long, R.; Tokina, M. V.;Prezhdo, O. V. Plasmon-Mediated Electron Injection from AuNanorods into MoS2: Traditional versus Photoexcitation Mechanism.Chem. 2018, 4, 1112−1127.(75) Li, L. Q.; Long, R.; Prezhdo, O. V. Charge Separation andRecombination in Two-Dimensional MoS2/WS2: Time-Domain abInitio Modeling. Chem. Mater. 2017, 29, 2466−2473.(76) Akimov, A. V.; Prezhdo, O. V. Theory of Nonadiabatic ElectronDynamics in Nanomaterials. In Encyclopedia of Nanotechnology;Bhushan, B., Ed.; Springer: Netherlands, 2015; pp 1−20.(77) Akimov, A. V.; Prezhdo, O. V. Theory of solar energy materials.J. Phys.: Condens. Matter 2015, 27, 130301.(78) Akimov, A. V.; Neukirch, A. J.; Prezhdo, O. V. TheoreticalInsights into Photoinduced Charge Transfer and Catalysis at OxideInterfaces. Chem. Rev. 2013, 113, 4496−4565.(79) Barbatti, M. Nonadiabatic dynamics with trajectory surfacehopping method. Wiley Interdisciplinary Reviews-ComputationalMolecular Science 2011, 1, 620−633.(80) Butler, L. J. Chemical reaction dynamics beyond the Born-Oppenheimer approximation. Annu. Rev. Phys. Chem. 1998, 49, 125−171.(81) Jasper, A. W.; Nangia, S.; Zhu, C. Y.; Truhlar, D. G. Non-Born-Oppenheimer molecular dynamics. Acc. Chem. Res. 2006, 39, 101−108.(82) Worth, G. A.; Cederbaum, L. S. Beyond Born-Oppenheimer:Molecular dynamics through a conical intersection. Annu. Rev. Phys.Chem. 2004, 55, 127−158.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3801

(83) Gottesman, R.; Zaban, A. Perovskites for Photovoltaics in theSpotlight: Photoinduced Physical Changes and Their Implications.Acc. Chem. Res. 2016, 49, 320−329.(84) Yan, Y.; Yin, W.-J.; Shi, T.; Meng, W.; Feng, C. Defect Physicsof CH3NH3PbX3 (X = I, Br, Cl) Perovskites. In Organic-InorganicHalide Perovskite Photovoltaics; Springer: Cham, 2016; pp 79−105.(85) Saba, M.; Cadelano, M.; Marongiu, D.; Chen, F.; Sarritzu, V.;Sestu, N.; Figus, C.; Aresti, M.; Piras, R.; Lehmann, A. G.; Cannas, C.;Musinu, A.; Quochi, F.; Mura, A.; Bongiovanni, G. Correlatedelectron-hole plasma in organometal perovskites. Nat. Commun. 2014,5, 5049.(86) Han, D.; Dai, C.; Chen, S. Calculation studies on point defectsin perovskite solar cells. J. Semicond. 2017, 38, No. 011006.(87) Yin, W.-J.; Yang, J.-H.; Kang, J.; Yan, Y.; Wei, S.-H. Halideperovskite materials for solar cells: a theoretical review. J. Mater.Chem. A 2015, 3, 8926−8942.(88) Duan, H.-S.; Zhou, H.; Chen, Q.; Sun, P.; Luo, S.; Song, T.-B.;Bob, B.; Yang, Y. The identification and characterization of defectstates in hybrid organic−inorganic perovskite photovoltaics. Phys.Chem. Chem. Phys. 2015, 17, 112−116.(89) Aristidou, N.; Sanchez-Molina, I.; Chotchuangchutchaval, T.;Brown, M.; Martinez, L.; Rath, T.; Haque, S. A. The Role of Oxygenin the Degradation of Methylammonium Lead Trihalide PerovskitePhotoactive Layers. Angew. Chem., Int. Ed. 2015, 54, 8208−8212.(90) Kirchartz, T.; Kruckemeier, L.; Unger, E. L. Research Update:Recombination and open-circuit voltage in lead-halide perovskites.APL Mater. 2018, 6, 100702.(91) Sherkar, T. S.; Momblona, C.; Gil-Escrig, L.; Avila, J.; Sessolo,M.; Bolink, H. J.; Koster, L. J. A. Recombination in Perovskite SolarCells: Significance of Grain Boundaries, Interface Traps, and DefectIons. ACS Energy Letters 2017, 2, 1214−1222.(92) Heo, S.; Seo, G.; Lee, Y.; Lee, D.; Seol, M.; Lee, J.; Park, J.-B.;Kim, K.; Yun, D.-J.; Kim, Y. S.; Shin, J. K.; Ahn, T. K.; Nazeeruddin,M. K. Deep level trapped defect analysis in CH3NH3PbI3 perovskitesolar cells by deep level transient spectroscopy. Energy Environ. Sci.2017, 10, 1128−1133.(93) Ghosh, S.; Pal, S. K.; Karki, K. J.; Pullerits, T. Ion MigrationHeals Trapping Centers in CH3NH3PbBr3 Perovskite. ACS EnergyLetters 2017, 2, 2133−2139.(94) Correa-Baena, J.-P.; Tress, W.; Domanski, K.; Anaraki, E. H.;Turren-Cruz, S.-H.; Roose, B.; Boix, P. P.; Gratzel, M.; Saliba, M.;Abate, A.; Hagfeldt, A. Identifying and suppressing interfacialrecombination to achieve high open-circuit voltage in perovskitesolar cells. Energy Environ. Sci. 2017, 10, 1207−1212.(95) Srimath Kandada, A. R.; Neutzner, S.; D’Innocenzo, V.;Tassone, F.; Gandini, M.; Akkerman, Q. A.; Prato, M.; Manna, L.;Petrozza, A.; Lanzani, G. Nonlinear Carrier Interactions in LeadHalide Perovskites and the Role of Defects. J. Am. Chem. Soc. 2016,138, 13604−13611.(96) Stewart, R. J.; Grieco, C.; Larsen, A. V.; Doucette, G. S.;Asbury, J. B. Molecular Origins of Defects in OrganohalidePerovskites and Their Influence on Charge Carrier Dynamics. J.Phys. Chem. C 2016, 120, 12392−12402.(97) Zhou, Z.; Liu, J.; Long, R.; Li, L.; Guo, L.; Prezhdo, O. V.Control of Charge Carriers Trapping and Relaxation in Hematite byOxygen Vacancy Charge: Ab Initio Non-Adiabatic MolecularDynamics. J. Am. Chem. Soc. 2017, 139, 6707−6717.(98) Agiorgousis, M. L.; Sun, Y.-Y.; Zeng, H.; Zhang, S. StrongCovalency-Induced Recombination Centers in Perovskite Solar CellMaterial CH3NH3PbI3. J. Am. Chem. Soc. 2014, 136, 14570−14575.(99) Frolova, L. A.; Dremova, N. N.; Troshin, P. A. The chemicalorigin of the p-type and n-type doping effects in the hybridmethylammonium−lead iodide (MAPbI3) perovskite solar cells.Chem. Commun. 2015, 51, 14917−14920.(100) He, J.; Long, R. Lead Vacancy Can Explain the SuppressedNonradiative Electron-Hole Recombination in FAPbI3 Perovskiteunder Iodine-Rich Conditions: A Time-Domain Ab Initio Study. J.Phys. Chem. Lett. 2018, 9, 6489−6495.

(101) Vogel, D. J.; Inerbaev, T. M.; Kilin, D. S. Role of LeadVacancies for Optoelectronic Properties of Lead-Halide Perovskites. J.Phys. Chem. C 2018, 122, 5216−5226.(102) Lee, J.-W.; Bae, S.-H.; De Marco, N.; Hsieh, Y.-T.; Dai, Z.;Yang, Y. The role of grain boundaries in perovskite solar cells.Materials Today Energy 2018, 7, 149−160.(103) Naikaew, A.; Prajongtat, P.; Lux-Steiner, M. C.; Arunchaiya,M.; Dittrich, T. Role of phase composition for electronic states inCH3NH3PbI3 prepared from CH3NH3I/PbCl2 solution. Appl. Phys.Lett. 2015, 106, 232104.(104) Liu, F.; Dong, Q.; Wong, M. K.; Djurisic, A. B.; Ng, A.; Ren,Z.; Shen, Q.; Surya, C.; Chan, W. K.; Wang, J.; Ng, A. M. C.; Liao, C.;Li, H.; Shih, K.; Wei, C.; Su, H.; Dai, J. Is Excess PbI2 Beneficial forPerovskite Solar Cell Performance? Adv. Energy Mater. 2016, 6,1502206.(105) Kim, Y. C.; Jeon, N. J.; Noh, J. H.; Yang, W. S.; Seo, J.; Yun, J.S.; Ho-Baillie, A.; Huang, S.; Green, M. A.; Seidel, J.; Ahn, T. K.;Seok, S. I. Beneficial Effects of PbI2Incorporated in Organo-LeadHalide Perovskite Solar Cells. Adv. Energy Mater. 2016, 6, 1502104.(106) Bi, D.; Tress, W.; Dar, M. I.; Gao, P.; Luo, J.; Renevier, C.;Schenk, K.; Abate, A.; Giordano, F.; Correa Baena, J. P.; Decoppet, J.D.; Zakeeruddin, S. M.; Nazeeruddin, M. K.; Gratzel, M.; Hagfeldt, A.Efficient luminescent solar cells based on tailored mixed-cationperovskites. Sci. Adv. 2016, 2, No. e1501170.(107) Shao, Y.; Xiao, Z.; Bi, C.; Yuan, Y.; Huang, J. Origin andelimination of photocurrent hysteresis by fullerene passivation inCH3NH3PbI3 planar heterojunction solar cells. Nat. Commun. 2014,5, 5784.(108) Roldan-Carmona, C.; Gratia, P.; Zimmermann, I.; Grancini,G.; Gao, P.; Graetzel, M.; Nazeeruddin, M. K. High efficiencymethylammonium lead triiodide perovskite solar cells: the relevanceof non-stoichiometric precursors. Energy Environ. Sci. 2015, 8, 3550−3556.(109) Supasai, T.; Rujisamphan, N.; Ullrich, K.; Chemseddine, A.;Dittrich, T. Formation of a passivating CH3NH3PbI3/PbI2 interfaceduring moderate heating of CH3NH3PbI3 layers. Appl. Phys. Lett.2013, 103, 183906.(110) Chen, Q.; Zhou, H.; Song, T. B.; Luo, S.; Hong, Z.; Duan, H.S.; Dou, L.; Liu, Y.; Yang, Y. Controllable self-induced passivation ofhybrid lead iodide perovskites toward high performance solar cells.Nano Lett. 2014, 14, 4158−4163.(111) Somsongkul, V.; Lang, F.; Jeong, A. R.; Rusu, M.; Arunchaiya,M.; Dittrich, T. Hole blocking PbI2/CH3NH3PbI3 interface. Phys.Status Solidi RRL 2014, 08, 763−766.(112) Mosconi, E.; Grancini, G.; Roldan-Carmona, C.; Gratia, P.;Zimmermann, I.; Nazeeruddin, M. K.; De Angelis, F. EnhancedTiO2/MAPbI3 Electronic Coupling by Interface Modification withPbI2. Chem. Mater. 2016, 28, 3612−3615.(113) Haruyama, J.; Sodeyama, K.; Han, L.; Tateyama, Y. SurfaceProperties of CH3NH3PbI3 for Perovskite Solar Cells. Acc. Chem. Res.2016, 49, 554−561.(114) Haruyama, J.; Sodeyama, K.; Han, L.; Tateyama, Y.Termination Dependence of Tetragonal CH3NH3PbI3 Surfaces forPerovskite Solar Cells. J. Phys. Chem. Lett. 2014, 5, 2903−2909.(115) Geng, W.; Tong, C.-J.; Tang, Z.-K.; Yam, C.; Zhang, Y.-N.;Lau, W.-M.; Liu, L.-M. Effect of surface composition on electronicproperties of methylammonium lead iodide perovskite. Journal ofMateriomics 2015, 1, 213−220.(116) Geng, W.; Tong, C.-J.; Liu, J.; Zhu, W.; Lau, W.-M.; Liu, L.-M.Structures and Electronic Properties of Different CH3NH3PbI3/TiO2

Interface: A First-Principles Study. Sci. Rep. 2016, 6, 20131.(117) Quarti, C.; De Angelis, F.; Beljonne, D. Influence of SurfaceTermination on the Energy Level Alignment at the CH3NH3PbI3Perovskite/C60 Interface. Chem. Mater. 2017, 29, 958−968.(118) Mosconi, E.; Azpiroz, J. M.; De Angelis, F. Ab InitioMolecular Dynamics Simulations of Methylammonium Lead IodidePerovskite Degradation by Water. Chem. Mater. 2015, 27, 4885−4892.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3802

(119) Noel, N. K.; Abate, A.; Stranks, S. D.; Parrott, E. S.; Burlakov,V. M.; Goriely, A.; Snaith, H. J. Enhanced photoluminescence andsolar cell performance via Lewis base passivation of organic-inorganiclead halide perovskites. ACS Nano 2014, 8, 9815−9821.(120) Li, H.; Liang, C.; Liu, Y.; Zhang, Y.; Tong, J.; Zuo, W.; Xu, S.;Shao, G.; Cao, S. Covalently Connecting Crystal Grains withPolyvinylammonium Carbochain Backbone To Suppress GrainBoundaries for Long-Term Stable Perovskite Solar Cells. ACS Appl.Mater. Interfaces 2017, 9, 6064−6071.(121) Son, D. Y.; Lee, J. W.; Choi, Y. J.; Jang, I. H.; Lee, S.; Yoo, P.J.; Shin, H.; Ahn, N.; Choi, M.; Kim, D.; Park, N. G. Self-formed grainboundary healing layer for highly efficient CH3NH3PbI3 perovskitesolar cells. Nature Energy 2016, 1, 16081.(122) Chiang, C.-H.; Wu, C.-G. Bulk heterojunction perovskite−PCBM solar cells with high fill factor. Nat. Photonics 2016, 10, 196−200.(123) Zhao, P.; Kim, B. J.; Jung, H. S. Passivation in perovskite solarcells: A review. Materials Today Energy 2018, 7, 267−286.(124) Adhyaksa, G. W. P.; Brittman, S.; Abolins, H.; Lof, A.; Li, X.;Keelor, J. D.; Luo, Y.; Duevski, T.; Heeren, R. M. A.; Ellis, S. R.;Fenning, D. P.; Garnett, E. C. Understanding Detrimental andBeneficial Grain Boundary Effects in Halide Perovskites. Adv. Mater.2018, 30, No. e1804792.(125) Reid, O. G.; Yang, M.; Kopidakis, N.; Zhu, K.; Rumbles, G.Grain-Size-Limited Mobility in Methylammonium Lead IodidePerovskite Thin Films. ACS Energy Letters 2016, 1, 561−565.(126) Adhyaksa, G. W. P.; Veldhuizen, L. W.; Kuang, Y.; Brittman,S.; Schropp, R. E. I.; Garnett, E. C. Carrier Diffusion Lengths inHybrid Perovskites: Processing, Composition, Aging, and SurfacePassivation Effects. Chem. Mater. 2016, 28, 5259−5263.(127) Yin, W.-J.; Chen, H.; Shi, T.; Wei, S.-H.; Yan, Y. Origin ofHigh Electronic Quality in Structurally Disordered CH3NH3PbI3 andthe Passivation Effect of Cl and O at Grain Boundaries. AdvancedElectronic Materials 2015, 1, 1500044.(128) Kim, G. Y.; Oh, S. H.; Nguyen, B. P.; Jo, W.; Kim, B. J.; Lee,D. G.; Jung, H. S. Efficient Carrier Separation and IntriguingSwitching of Bound Charges in Inorganic-Organic Lead Halide SolarCells. J. Phys. Chem. Lett. 2015, 6, 2355−62.(129) Wei, L. -y.; Ma, W.; Lian, C.; Meng, S. Benign InterfacialIodine Vacancies in Perovskite Solar Cells. J. Phys. Chem. C 2017, 121,5905−5913.(130) Tong, C.-J.; Li, L.; Liu, L.-M.; Prezhdo, O. V. Long CarrierLifetimes in PbI2-Rich Perovskites Rationalized by Ab InitioNonadiabatic Molecular Dynamics. ACS Energy Letters 2018, 3,1868−1874.(131) Kilina, S. V.; Neukirch, A. J.; Habenicht, B. F.; Kilin, D. S.;Prezhdo, O. V. Quantum Zeno Effect Rationalizes the PhononBottleneck in Semiconductor Quantum Dots. Phys. Rev. Lett. 2013,110, 180404.(132) Jeon, N. J.; Na, H.; Jung, E. H.; Yang, T.-Y.; Lee, Y. G.; Kim,G.; Shin, H.-W.; Il Seok, S.; Lee, J.; Seo, J. A fluorene-terminated hole-transporting material for highly efficient and stable perovskite solarcells. Nature Energy 2018, 3, 682−689.(133) Starr, D. E.; Sadoughi, G.; Handick, E.; Wilks, R. G.; Alsmeier,J. H.; Kohler, L.; Gorgoi, M.; Snaith, H. J.; Bar, M. Direct observationof an inhomogeneous chlorine distribution in CH3NH3PbI3−xClxlayers: surface depletion and interface enrichment. Energy Environ.Sci. 2015, 8, 1609−1615.(134) Yu, H.; Wang, F.; Xie, F.; Li, W.; Chen, J.; Zhao, N. The Roleof Chlorine in the Formation Process of “CH3NH3PbI3‑xClx”Perovskite. Adv. Funct. Mater. 2014, 24, 7102−7108.(135) Mosconi, E.; Ronca, E.; De Angelis, F. First-PrinciplesInvestigation of the TiO2/Organohalide Perovskites Interface: TheRole of Interfacial Chlorine. J. Phys. Chem. Lett. 2014, 5, 2619−2625.(136) Williams, S. T.; Zuo, F.; Chueh, C. C.; Liao, C. Y.; Liang, P.W.; Jen, A. K. Role of chloride in the morphological evolution oforgano-lead halide perovskite thin films. ACS Nano 2014, 8, 10640−10654.

(137) Wehrenfennig, C.; Eperon, G. E.; Johnston, M. B.; Snaith, H.J.; Herz, L. M. High Charge Carrier Mobilities and Lifetimes inOrganolead Trihalide Perovskites. Adv. Mater. 2014, 26, 1584−1589.(138) Mosconi, E.; Grancini, G.; Roldan-Carmona, C.; Gratia, P.;Zimmermann, I.; Nazeeruddin, M. K.; De Angelis, F. Enhanced TiO2/MAPbI3 Electronic Coupling by Interface Modification with PbI2.Chem. Mater. 2016, 28, 3612−3615.(139) Liu, M.; Johnston, M. B.; Snaith, H. J. Efficient planarheterojunction perovskite solar cells by vapour deposition. Nature2013, 501, 395−8.(140) Atourki, L.; Vega, E.; Marí, B.; Mollar, M.; Ait Ahsaine, H.;Bouabid, K.; Ihlal, A. Role of the chemical substitution on thestructural and luminescence properties of the mixed halide perovskitethin MAPbI3−x Brx (0 ⩽ x ⩽ 1) films. Appl. Surf. Sci. 2016, 371, 112−117.(141) Diroll, B. T.; Nedelcu, G.; Kovalenko, M. V.; Schaller, R. D.High-Temperature Photoluminescence of CsPbX3 (X = Cl, Br, I)Nanocrystals. Adv. Funct. Mater. 2017, 27, 1606750.(142) Makarov, N. S.; Guo, S.; Isaienko, O.; Liu, W.; Robel, I.;Klimov, V. I. Spectral and Dynamical Properties of Single Excitons,Biexcitons, and Trions in Cesium-Lead-Halide Perovskite QuantumDots. Nano Lett. 2016, 16, 2349−62.(143) Wang, H. C.; Lin, S. Y.; Tang, A. C.; Singh, B. P.; Tong, H. C.;Chen, C. Y.; Lee, Y. C.; Tsai, T. L.; Liu, R. S. Mesoporous SilicaParticles Integrated with All-Inorganic CsPbBr3 Perovskite Quantum-Dot Nanocomposites (MP-PQDs) with High Stability and WideColor Gamut Used for Backlight Display. Angew. Chem., Int. Ed. 2016,55, 7924−9.(144) Yettapu, G. R.; Talukdar, D.; Sarkar, S.; Swarnkar, A.; Nag, A.;Ghosh, P.; Mandal, P. Terahertz Conductivity within ColloidalCsPbBr3 Perovskite Nanocrystals: Remarkably High Carrier Mobi-lities and Large Diffusion Lengths. Nano Lett. 2016, 16, 4838−4848.(145) Tong, Y.; Bladt, E.; Ayguler, M. F.; Manzi, A.; Milowska, K.Z.; Hintermayr, V. A.; Docampo, P.; Bals, S.; Urban, A. S.;Polavarapu, L.; Feldmann, J. Highly Luminescent Cesium LeadHalide Perovskite Nanocrystals with Tunable Composition andThickness by Ultrasonication. Angew. Chem., Int. Ed. 2016, 55,13887−13892.(146) Colella, S.; Mosconi, E.; Fedeli, P.; Listorti, A.; Gazza, F.;Orlandi, F.; Ferro, P.; Besagni, T.; Rizzo, A.; Calestani, G.; Gigli, G.;De Angelis, F.; Mosca, R. MAPbl3‑x Clx Mixed Halide Perovskite forHybrid Solar Cells: The Role of Chloride as Dopant on the Transportand Structural Properties. Chem. Mater. 2013, 25, 4613−4618.(147) Kim, M. C.; Kim, B. J.; Son, D. Y.; Park, N. G.; Jung, H. S.;Choi, M. Observation of Enhanced Hole Extraction in BrConcentration Gradient Perovskite Materials. Nano Lett. 2016, 16,5756−5763.(148) Yi, C.; Luo, J.; Meloni, S.; Boziki, A.; Ashari-Astani, N.;Gratzel, C.; Zakeeruddin, S. M.; Rothlisberger, U.; Gratzel, M.Entropic stabilization of mixed A-cation ABX3 metal halide perov-skites for high performance perovskite solar cells. Energy Environ. Sci.2016, 9, 656−662.(149) Giorgi, G.; Yoshihara, T.; Yamashita, K. Structural andelectronic features of small hybrid organic−inorganic halide perov-skite clusters: a theoretical analysis. Phys. Chem. Chem. Phys. 2016, 18,27124−27132.(150) Ju, M.-G.; Sun, G.; Zhao, Y.; Liang, W. A computational viewof the change in the geometric and electronic properties of perovskitescaused by the partial substitution of Pb by Sn. Phys. Chem. Chem.Phys. 2015, 17, 17679−17687.(151) Zhu, H.; Trinh, M. T.; Wang, J.; Fu, Y.; Joshi, P. P.; Miyata,K.; Jin, S.; Zhu, X. Y. Organic Cations Might Not Be Essential to theRemarkable Properties of Band Edge Carriers in Lead HalidePerovskites. Adv. Mater. 2017, 29, 1603072.(152) Amat, A.; Mosconi, E.; Ronca, E.; Quarti, C.; Umari, P.;Nazeeruddin, M. K.; Graetzel, M.; De Angelis, F. Cation-InducedBand-Gap Tuning in Organohalide Perovskites: Interplay of Spin-Orbit Coupling and Octahedra Tilting. Nano Lett. 2014, 14, 3608−3616.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3803

(153) Jana, A.; Mittal, M.; Singla, A.; Sapra, S. Solvent-free,mechanochemical syntheses of bulk trihalide perovskites and theirnanoparticles. Chem. Commun. 2017, 53, 3046−3049.(154) Liu, J.; Shirai, Y.; Yang, X.; Yue, Y.; Chen, W.; Wu, Y.; Islam,A.; Han, L. High-Quality Mixed-Organic-Cation Perovskites from aPhase-Pure Non-stoichiometric Intermediate (FAI)1‑xPbI2 for SolarCells. Adv. Mater. 2015, 27, 4918−4823.(155) De Marco, N.; Zhou, H.; Chen, Q.; Sun, P.; Liu, Z.; Meng, L.;Yao, E. P.; Liu, Y.; Schiffer, A.; Yang, Y. Guanidinium: A Route toEnhanced Carrier Lifetime and Open-Circuit Voltage in HybridPerovskite Solar Cells. Nano Lett. 2016, 16, 1009−16.(156) Liu, J.; Prezhdo, O. V. Chlorine Doping Reduces Electron−Hole Recombination in Lead Iodide Perovskites: Time-Domain AbInitio Analysis. J. Phys. Chem. Lett. 2015, 6, 4463−4469.(157) He, J.; Fang, W.-H.; Long, R.; Prezhdo, O. V. IncreasedLattice Stiffness Suppresses Nonradiative Charge Recombination inMAPbI3 Doped with Larger Cations: Time-Domain Ab InitioAnalysis. ACS Energy Letters 2018, 3, 2070−2076.(158) Noh, J. H.; Im, S. H.; Heo, J. H.; Mandal, T. N.; Seok, S. I.Chemical management for colorful, efficient, and stable inorganic-organic hybrid nanostructured solar cells. Nano Lett. 2013, 13, 1764−1769.(159) Ogomi, Y.; Morita, A.; Tsukamoto, S.; Saitho, T.; Fujikawa,N.; Shen, Q.; Toyoda, T.; Yoshino, K.; Pandey, S. S.; Ma, T.; Hayase,S. CH3NH3SnxPb(1‑x)I3 Perovskite Solar Cells Covering up to 1060nm. J. Phys. Chem. Lett. 2014, 5, 1004−10011.(160) Quarti, C.; Mosconi, E.; Umari, P.; De Angelis, F. ChlorineIncorporation in the CH3NH3PbI3 Perovskite: Small Concentration.Inorg. Chem. 2017, 56, 74−83.(161) Leijtens, T.; E. Eperon, G.; J. Barker, A.; Grancini, G.; Zhang,W.; M. Ball, J.; Srimath Kandada, A. R.; J. Snaith, H.; Petrozza, A.Carrier trapping and recombination: the role of defect physics inenhancing the open circuit voltage of metal halide perovskite solarcells. Energy Environ. Sci. 2016, 9, 3472−3481.(162) Haight, R. Electron dynamics at surfaces. Surf. Sci. Rep. 1995,21, 275−325.(163) Akimov, A. V.; Prezhdo, O. V. Persistent ElectronicCoherence Despite Rapid Loss of Electron-Nuclear Correlation. J.Phys. Chem. Lett. 2013, 4, 3857−3864.(164) Akimov, A. V. A Simple Phase Correction Makes a BigDifference in Nonadiabatic Molecular Dynamics. J. Phys. Chem. Lett.2018, 9, 6096−6102.(165) Wang, L. J.; Prezhdo, O. V. A Simple Solution to the TrivialCrossing Problem in Surface Hopping. J. Phys. Chem. Lett. 2014, 5,713−719.(166) Li, W.; Rego, L. G.; Bai, F. Q.; Wang, J.; Jia, R.; Xie, L. M.;Zhang, H. X. What Makes Hydroxamate a Promising AnchoringGroup in Dye-Sensitized Solar Cells? Insights from TheoreticalInvestigation. J. Phys. Chem. Lett. 2014, 5, 3992−3999.(167) Rego, L. G.; Batista, V. S. Quantum dynamics simulations ofinterfacial electron transfer in sensitized TiO2 semiconductors. J. Am.Chem. Soc. 2003, 125, 7989−97.(168) Akimov, A. V.; Asahi, R.; Jinnouchi, R.; Prezhdo, O. V. WhatMakes the Photocatalytic CO2 Reduction on N-Doped Ta2O5

Efficient: Insights from Nonadiabatic Molecular Dynamics. J. Am.Chem. Soc. 2015, 137, 11517−11525.(169) Pal, S.; Trivedi, D. J.; Akimov, A. V.; Aradi, B.; Frauenheim,T.; Prezhdo, O. V. Nonadiabatic Molecular Dynamics for ThousandAtom Systems: A Tight-Binding Approach toward PYXAID. J. Chem.Theory Comput. 2016, 12, 1436−48.(170) Li, L. Q.; Long, R.; Bertolini, T.; Prezhdo, O. V. SulfurAdatom and Vacancy Accelerate Charge Recombination in MoS2 butby Different Mechanisms: Time-Domain Ab Initio Analysis. NanoLett. 2017, 17, 7962−7967.(171) Li, L. Q.; Long, R.; Prezhdo, O. V. Why Chemical VaporDeposition Grown MoS2 Samples Outperform Physical VaporDeposition Samples: Time-Domain ab Initio Analysis. Nano Lett.2018, 18, 4008−4014.

(172) Akimov, A. V. Stochastic and Quasi-Stochastic Hamiltoniansfor Long-Time Nonadiabatic Molecular Dynamics. J. Phys. Chem. Lett.2017, 8, 5190−5195.(173) Fischer, S. A.; Habenicht, B. F.; Madrid, A. B.; Duncan, W. R.;Prezhdo, O. V. Regarding the validity of the time-dependent Kohn-Sham approach for electron-nuclear dynamics via trajectory surfacehopping. J. Chem. Phys. 2011, 134, 024102.(174) Pradhan, E.; Sato, K.; Akimov, A. V. Non-adiabatic moleculardynamics with ΔSCF excited states. J. Phys.: Condens. Matter 2018,30, 484002.(175) Doltsinis, N. L.; Marx, D. Nonadiabatic Car-Parrinellomolecular dynamics. Phys. Rev. Lett. 2002, 88, 166402.(176) Tapavicza, E.; Tavernelli, I.; Rothlisberger, U. Trajectorysurface hopping within linear response time-dependent density-functional theory. Phys. Rev. Lett. 2007, 98, 023001.(177) Wu, G. F.; Li, Z.; Zhang, X.; Lu, G. Charge Separation andExciton Dynamics at Polymer/ZnO Interface from First-PrinciplesSimulations. J. Phys. Chem. Lett. 2014, 5, 2649−2656.(178) Kim, J.; Lee, S.-H.; Lee, J. H.; Hong, K.-H. The Role ofIntrinsic Defects in Methylammonium Lead Iodide Perovskite. J. Phys.Chem. Lett. 2014, 5, 1312−1317.(179) Buin, A.; Pietsch, P.; Xu, J.; Voznyy, O.; Ip, A. H.; Comin, R.;Sargent, E. H. Materials Processing Routes to Trap-Free HalidePerovskites. Nano Lett. 2014, 14, 6281−6286.(180) Sun, Y. Y.; Shi, J.; Lian, J.; Gao, W.; Agiorgousis, M. L.; Zhang,P.; Zhang, S. Discovering lead-free perovskite solar materials with asplit-anion approach. Nanoscale 2016, 8, 6284−6289.(181) Li, W.; Zhou, L.; Prezhdo, O. V.; Akimov, A. V. Spin−OrbitInteractions Greatly Accelerate Nonradiative Dynamics in LeadHalide Perovskites. ACS Energy Letters 2018, 3, 2159−2166.(182) Umari, P.; Mosconi, E.; De Angelis, F. Relativistic GWcalculations on CH3NH3PbI3 and CH3NH3SnI3 Perovskites for SolarCell Applications. Sci. Rep. 2015, 4, 4467.(183) Giesbrecht, N.; Schlipf, J.; Grill, I.; Rieder, P.; Dyakonov, V.;Bein, T.; Hartschuh, A.; Muller-Buschbaum, P.; Docampo, P. Single-crystal-like optoelectronic-properties of MAPbI3 perovskite poly-crystalline thin films. J. Mater. Chem. A 2018, 6, 4822−4828.(184) Xu, J.; Liu, J.-B.; Liu, B.-X.; Wang, J.; Huang, B. DefectEngineering of Grain Boundaries in Lead-Free Halide DoublePerovskites for Better Optoelectronic Performance. Adv. Funct.Mater. 2019, 29, 1805870.(185) Yun, J. S.; Ho-Baillie, A.; Huang, S.; Woo, S. H.; Heo, Y.;Seidel, J.; Huang, F.; Cheng, Y. B.; Green, M. A. Benefit of GrainBoundaries in Organic-Inorganic Halide Planar Perovskite Solar Cells.J. Phys. Chem. Lett. 2015, 6, 875−880.(186) Zhang, Z. S.; Fang, W. H.; Tokina, M. V.; Long, R.; Prezhdo,O. V. Rapid Decoherence Suppresses Charge Recombination inMulti-Layer 2D Halide Perovskites: Time-Domain Ab Initio Analysis.Nano Lett. 2018, 18, 2459−2466.(187) Vogel, D. J.; Kryjevski, A.; Inerbaev, T.; Kilin, D. S.Photoinduced Single- and Multiple-Electron Dynamics ProcessesEnhanced by Quantum Confinement in Lead Halide PerovskiteQuantum Dots. J. Phys. Chem. Lett. 2017, 8, 3032−3039.(188) Hyeon-Deuk, K.; Kim, J.; Prezhdo, O. V. Ab Initio Analysis ofAuger-Assisted Electron Transfer. J. Phys. Chem. Lett. 2015, 6, 244−249.(189) Hyeon-Deuk, K.; Prezhdo, O. V. Multiple Exciton Generationand Recombination Dynamics in Small Si and CdSe Quantum Dots:An Ab Initio Time-Domain Study. ACS Nano 2012, 6, 1239−1250.(190) Trivedi, D. J.; Wang, L. J.; Prezhdo, O. V. Auger-MediatedElectron Relaxation Is Robust to Deep Hole Traps: Time-Domain AbInitio Study of CdSe Quantum Dots. Nano Lett. 2015, 15, 2086−2091.

The Journal of Physical Chemistry Letters Perspective

DOI: 10.1021/acs.jpclett.9b00641J. Phys. Chem. Lett. 2019, 10, 3788−3804

3804