当前位置：文档库 › current progress and future perspectives for MAPbI3 pdf

current progress and future perspectives for MAPbI3 pdf

Current progress and future perspectives for organic/inorganic perovskite solar cells

Pablo P.Boix1,Kazuteru Nonomura1,Nripan Mathews1,2,3,*and

Subodh G.Mhaisalkar1,2,*

Energy Research Institute@NTU(ERI@N),Research Techno Plaza,X-Frontier Block,Level5,50Nanyang Drive,Singapore637553,Singapore

2School of Materials Science and Engineering,Nanyang Technological University,Nanyang Avenue,Singapore639798,Singapore

3Singapore-Berkeley Research Initiative for Sustainable Energy,1Create Way,Singapore138602,Singapore

The recent emergence of ef?cient solar cells based on organic/inorganic lead halide perovskite absorbers promises to transform the?elds of dye-sensitized,organic,and thin?lm solar cells.Solution processed photovoltaics incorporating perovskite absorbers have achieved ef?ciencies of15%[1]in solid-state device con?gurations,superseding liquid dye sensitized solar cell(DSC),evaporated and tandem organic solar cells,as well as various thin?lm photovoltaics;thus establishing perovskite solar cells as a robust candidate for commercialization.Since the?rst reports in late2012,interest has soared in the innovative device structures as well as new materials,promising further improvements.However,identifying the basic working mechanisms,which are still being debated,will be crucial to design the optimum device con?guration and maximize solar cell ef?ciencies.Here we distill the current state-of-the-art and highlight the guidelines to ascertain the scienti?c challenges as well as the requisites to make this technology market-viable.

Introduction

Achieving cost effective,easily processable,ef?cient and versa-tile solar cells has always been a challenge for the scienti?c community[2].An attractive candidate ful?lling these require-ments is the sensitized solar cell architecture.Since the devel-opment of the dye sensitized solar cell(DSC)[3]which is the ultimate expression of this con?guration,power conversion ef?ciencies have reached12.3%[4].The possibility of replacing volatile liquid electrolyte with a solid hole transport material (HTM)has been pursued[5],but low ef?ciencies(h<10%[6,7]) have been a drawback for the commercialization of this technol-ogy.In order to improve the ef?ciency of the solid state DSC (ssDSC),different approaches have been explored.This includes dye design for extending the absorption range towards the near infrared regime as well as increasing electron injection and hole regeneration.However the obtained ef?ciency remains low, since the poor optical extinction coef?cient requires a thick mesoporous?lm to absorb the light,increasing hole transport resistance and recombination.

Quantum dot solar cells(QDSC)utilize semiconductor nanocrystals as light absorbers[8].Their large intrinsic dipole moment and tunability of the bandgap by size control and shape provide an excellent tool for nanoscale design of light absorber materials for sensitized solar cells.In solid state con?guration remarkable ef?ciencies higher than8%have been reported for innovative devices structures,where PbS absorber acts as HTM[9].Similarly,extremely thin absorber (ETA)solar cells use a thin layer of inorganic absorber to sensitize a mesoporous semiconductor.Sb2S3has achieved excellent solid state ef?ciencies in this con?guration due to its low bandgap and high extinction coef?cient.Nevertheless problems such as high recombination have not been overcome yet,limiting its maximum ef?ciency to6.3%when combined with the appropriate polymer and fullerene derivative[10]. Other inorganic materials have also been employed as sensi-tizers for TiO2including CuInS2(CIS),achieving ef?ciencies of 5%[11].

RESEARCH:

Review

*Corresponding authors:.Mathews,N.(Nripan@https://www.wendangku.net/doc/956400498.html,.sg),

Mhaisalkar,S.G.(Subodh@https://www.wendangku.net/doc/956400498.html,.sg)

Since late2012,organic/inorganic halides with the perovskite structure have strongly attracted the attention of the photovoltaic community when ef?ciencies close to10%were?rst achieved in solid state cells[12,13].The excellent properties and the innova-tive device possibilities in perovskite-structured organometal halides has resulted in a frenzied increase of publications reporting high ef?ciencies[14,15],see Fig.1a.Recently15%ef?cient solar cells were reported with CH3NH3PbI3[1]target ef?ciencies of20% identi?ed as a feasible goal[16].It is therefore pertinent to evaluate the potential and analyze the prospects of this exciting technology that have galvanized the photovoltaic research community.

Here we summarize the photovoltaic studies with organometal halide perovskite compounds and propose avenues for further development.The optical and electrical characteristics of these halides are reviewed and compared to other sensitizers.The wide variety of device architectures employed so far,are evaluated. Since different architectures have diverse principles determining their performance,these insights into the working mechanisms allow the determination of the optimum approach.At the end,future perspectives with a particular focus in the improvement of ef?ciency,stability commercialization prospects are discussed.

Organic/inorganic metal halides as light absorbers Although these class of materials have been widely studied for decades[17,18],only recently have they been introduced in solar cells.The?rst reports based on hybrid organic/inorganic halides CH3NH3PbI3and CH3NH3PbBr3,published in2009,achieved 3.8%ef?ciencies in a liquid electrolyte con?guration[19]where the absorber was regarded as a QDs deposited on TiO2.The ef?ciency was further improved to6.5%[20],but the short stability of the devices caused due to dissolution of the halides in the electrolyte,appeared to be an enormous drawback.The break-through occurred in late2012with the introduction of solid state hole transporting layers within the solar cell.This resulted in stable ef?ciencies close to10%[12,21],establishing these materials as robust candidates for ef?cient solar cells.These reports of high ef?ciency coupled with the materials’excellent optical, electrical and mechanical properties[22–28]along with solution

FIGURE1

(a)Ef?ciency charts for solution processed solar cells.Data adapted from Ref.[2],(b)CH3NH3PbI3perovskite structure,(c)external quantum ef?ciency measured for a CH3NH3PbI3perovskite solar cell and AM1.5g solar spectra and(d)absorption measured for different TiO2/CH3NH3PbI3àx Br x?lms.Reprinted from Ref.[36].

17R E S E A R C H : R e v i e w

processability have triggered rapid and continuous improvements in their ef?ciencies [14,15].Furthermore,other perovskite struc-tured materials such as CsSnI 3was effectively demonstrated as HTM and absorber in ssDSC,where a combination with N719dye yielded ef?ciencies of over 10%[7].

Organic/inorganic perovskites are hybrid layered materials typi-cally with an AMX 3structure,with A being a large cation,M a smaller metal cation and X an anion from the halide series.They form an octahedral structure of MX 6,which forms a three dimen-sional structure connected at the corners [29–31]as shown in Fig.1b.The component A ?lls the coordinated space between the octahedrals that form in these three-dimensional structures.The size of the cation A is an important aspect for the formation of a closed packed perovskite structure,since this cation A must ?t into the space composed of the four adjacent octahedra which are connected together through shared corners [32].In these organic/inorganic halides,the organic cations are small and are typically restricted to methylammonium,ethylammonium and formami-dinium.The integration of larger molecules with terminal cationic groups within the inorganic framework have also been demon-strated in some cases [24,33].The metal cations are typically divalent metal ions such as Pb 2+,Sn 2+and Ge 2+while the halide anions are I à,Cl àand Br à.The optical absorption as well as photoluminescence is related to the metal halide employed,with the iodides resulting in smaller bandgaps and light emission at longer wavelengths while the bromides display higher bandgap and luminescence at shorter wavelengths [34–36].Interestingly,a perovskite structure which incorporates two halides (e.g.iodide and bromide)allows for the continuous tuning of the bandgap (Fig.1d)[19,37].

The best solar cells have been obtained from CH 3NH 3PbI 3which has a bandgap of 1.55eV,close to the optimum one for

photovoltaic performance ($1.4eV).This coupled with the good extinction coef?cient (one order of magnitude higher than standard dyes [20])enables excellent external quantum ef?-ciency spectra (EQE)in the solar cells [12,38]until 800nm,harvesting the photons in the visible range of the solar spectra and part of the near-infrared (see Fig.1c).Remarkably,when CH 3NH 3PbI 3is heated above 55–608C it undergoes a phase transition from tetragonal to cubic [31],which is expected to narrow the bandgap.

Solar cell device architectures

The huge interest of CH 3NH 3PbI 3perovskite does not only lie in the high ef?ciencies but also in the novel con?gurations made possible by the singular characteristics of the material.

Metal-halide based devices with structure similar to the classical ssDSC [12]were fabricated with the organic/inorganic halide being deposited in a nanostructured TiO 2layer by a single step spin-coating method (device structure in Fig.2a)and spiro-OMeTAD (2,20,7,70-tetrakis-(N,N-di-p-methoxyphenylamine)9,90-spirobi-?uorene)HTM deposited on top.In this report optical measure-ments show charge injection from the perovskite into both TiO 2(electrons)and HTM (holes),but the latter is the fastest one.Very recently the application of the sequential deposition process (ori-ginally developed by Mitzi and coworkers [39])whereby PbI 2is converted into CH 3NH 3PbI 3within the pores of the TiO 2have resulted in record ef?ciencies (h =15.0%,Fig.3a)[1].

An important alteration of the above architecture was the replacement of the TiO 2mesoporous by an insulating Al 2O 3scaf-fold achieving ef?ciencies of 10.9%[21].It is worth to remark that regardless of the mesoporous layer,a compact TiO 2layer is still required for both the collection of the generated electrons and hole blocking.Since alumina’s conduction band is far higher than

FIGURE 2

Devices structure for (a)mesoporous perovskite solar cell structure where no HTM interpenetration is required.In the inset the electron charge transport processes for injecting and non-injecting mesoporous materials are represented and (b)structure of a thin ?lm-like perovskite solar cell.

RESEARCH:Review

the absorber’s LUMO,no electron injection from perovskite takes place.This indicates that the electron transport occurs within CH3NH3PbI3àx Cl x perovskite,which is con?rmed by photoin-duced absorption spectroscopy(PIA)measurements of the Al2O3/CH3NH3PbI3àx Cl x layers[21].In contrast,PIA measure-ments of CH3NH3PbI3àx Cl x/spiro-OMeTAD layers indicated that the photogenerated hole was injected into spiro-OMeTAD.This approach avoids the voltage drop associated with the occupation of the TiO2band-tails[40],thus resulting in higher photovoltage, additionally small-perturbation transient photocurrent decay measurements[13]also show faster charge collection in Al2O3 based devices compared to the TiO2ones(Fig.3d).Similar results have been obtained with pure CH3NH3PbI3in combination with non-injecting ZrO2scaffolds[41].The highest photovoltage reported so far within these class of materials(1.3V)also employed this con?guration,albeit in conjunction with a higher bandgap Br analogue and N,N0-dialkyl perylenediimide HTM[42].Several other hole transporting materials such as poly-(3-hexylthio-phene-2,5-diyl)(P3HT)and poly[N-900-heptadecanyl-2,7-carba-zole-alt-5,5-(40,70-di-2-thienyl-20,10,30-benzothiadiazole)] (PCDTBT)have also been tested with interesting results[14,43], however amine based(spiro-OMeTAD,poly-triarylamine PTAA) HTMs outperform energetically equivalent thiophene derivatives indicating a better interface with the perovskite absorber.

Interestingly,it was also reported that CH3NH3PbI3perovskite functions concurrently as a hole transporting material and light absorber,with5.5%conversion ef?ciency attained with a TiO2/ CH3NH3PbI3perovskite/Au construction[44].Obviously,a good coverage of the TiO2?lm by the CH3NH3PbI3is needed in this case, resulting in a thin capping layer,this con?guration corresponding to Fig.2a without the HTM layer.Therefore,it is clear that the electrical transport in organic/inorganic metal halides can occur as an ambipolar diffusion of electrons and holes[15].This is further supported by recent reports where both electron and holes are extracted from a$350nm thick layer of CH3NH3PbI3àx Cl x absor-ber sandwiched between a compact TiO2?lm and hole transport-ing spiro-OMeTAD.Even in the absence of mesoporous?lms photocurrent densities close to15mA cmà2were achieved in an approach analogous to the classical thin?lm architecture repre-sented in Fig.2b(without scaffold)or even higher with a mini-mum scaffold holding the perovskite thin?lm(Fig.3b).Planar con?gurations of CH3NH3PbI3with other contacts as PEDOT:PSS and various fullerene derivatives also displayed photocurrents of more than10mA cmà2[45].A recent report[46]has presented a planar heterojunction perovskite solar cell fabricated by vapor deposition which matched the15%ef?ciency record of the meso-porous cell.This result con?rms the existence of relatively long range electron,hole transport in these class of materials.

FIGURE3

(a)Measured current–voltage curve and performance characteristics for the record CH3NH3PbI3solar cell.Reprinted from Ref.[1],(b)cross section measured for a thin?lm-like perovskite solar cell with thin scaffold thickness.Reprinted from Ref.[59],(c)impedance spectra measured for a nanorod/CH3NH3PbI3 solar cell.Reprinted from Ref.[38]and(d)charge transport lifetime determined by small perturbation transient photocurrent decay of perovskite sensitized TiO2(circles with black line to aid the eye)and Al2O3cells(red crosses with line to aid the eye).Inset shows normalized photocurrent transients for TiO2 (black)and Al2O3cells(red),reprinted from Ref.[13].

19R E S E A R C H : R e v i e w

Due to the wide variety of device architectures and material sets employed in conjunction with the organic/inorganic metal halides,it is important to identify the pertinent mechanisms governing each con?guration.

Photovoltaic operational mechanisms

If we consider the system to be purely analogous to an ssDSC where the absorbed photon is converted into charge by the injection of the electrons and the regeneration of the holes,(inset Fig.2a)certain features must be considered for understanding the main physical processes governing the cell behaviour.

This model involves a fast injection of carriers from the light absorber into their respective conductive media,with no carrier transport occurring within the absorber itself.In this case,a good distribution of the absorber within the mesoporous layer will ensure maximal interfacial area required to generate the photo-current.Thus the limitations of this architecture will be analogous to that of the classical ssDSC [47]which have been widely studied.One of the main considerations is the light absorption.As described before,the bandgap of CH 3NH 3PbI 3is close to the optimum for photovoltaic conversion,while the high extinction coef?cient of the material ensures a good absorption of the light at low mesoporous ?lm thickness (w.r.t.dye sensitized systems).However,in order to separate the excited state into charge carriers,an energy price has to be paid for both electron and hole injection-directly re?ected in the achievable V OC .When TiO 2and spiro-OMeTAD are used,the energetic offsets are D E $0.07eV and D E $0.21eV for electrons and holes respectively (Figs.4and 5a).Under these circumstances,the energetic disorder and defects of the absorber will only have a minor effect in the absorption [40]because both the charge separation and charge transport take place at the interface or outside the absorber material.In contrast,the distribution of energetic states in the transport materials (TiO 2,HTMs)have an effect in the splitting of the Fermi levels and in the charge transport due to the population of band-tails [40].This necessitates the development of new nanostructures for electron

separation and conduction such as TiO 2nanorods [38,48],nanosheets [49]with the objective of improving electron transport and absorber in?ltration,as well as exploration of novel HTMs with suitable band alignment and improved hole mobilities [14,43].

In addition to these voltage losses associated with energy level mismatches as well as charge separation,charge recombination can further limit the performance of these devices.In a similar solid state systems such as TiO 2/Sb 2S 3/CuSCN,the bandgap of the absorber (1.65eV)minus the offset for electron and hole injection indicates 1.30eV available as a potential difference,however the reported V OC at 1sun is only 0.60V [50].In comparison,for TiO 2/CH 3NH 3PbI 3/spiro OMeTAD devices the available potential differ-ence is 1.22eV while the V OC achieved is more than 0.9V [12].This means that $0.7V is ‘lost’in the Sb 2S 3system,while only $0.3V is lost in the perovskite system.Although this indicates that recombination in perovskite solar cells is much lower than in Sb 2S 3devices,carrier lifetime measurements do not seem to support it [43,51].Additional experiments to compare the CH 3NH 3PbI 3sys-tem against other sensitizers under similar device conditions are therefore required.

Another loss mechanisms that affects the performance is man-ifested when varying the ?lm thickness [12,14].Thicker ?lms increase the light absorption,but at the same time reduce the EQE and consequently the current,in contrast to classical liquid DSCs.This highlights the importance of studying the recombina-tion losses.Therefore,even considering the encouraging open circuit potentials,the identi?cation of the process controlling the recombination mechanism,its characterization and reduction is critical for improving the ef?ciency of perovskites solar cells.Optical measurements as photoinduced absorption spectro-scopy (PIA)[52]and transient grating (LF-HD-TG)technique [53];electrochemical impedance spectroscopy [54]as well as mixed techniques such as transient photovoltage [55]have been used to characterize these losses with good accuracy in sensitized devices.Recently some of these techniques have been applied to

FIGURE 4

Energy levels for different materials acting as electron transporting material (left),absorbers (middle)and hole transporting materials (right)in solar cells.

RESEARCH:Review

organic/inorganic halides solar cells for different absorbers and HTM[43],nevertheless further investigation is needed to under-stand which recombination process is the dominating one in this new kind of systems.

In contrast to the sensitized solar cell architecture presented above,in solar cells utilizing mesoporous Al2O3instead of meso-porous TiO2[15,56],given the energetics of alumina,the injection of electrons from the absorber is not possible(inset,Fig.2a).As a consequence the extracted electrons must?ow within the CH3NH3PbI3àx Cl x itself.As seen in previous sections,analogous cases are reported where the absorber materials transport the holes [49].This con?guration thus resembles a thin?lm solar cell with the scaffold providing roughness to load the absorber layer result-ing in ef?cient light absorption.In order to understand how this thin?lm con?guration works,the nature of the?rst excited state is a critical factor to consider.If the binding energy of the photo-generated electron–hole pair is low enough(comparable to ther-mal energy),charge generation can occur within the absorber(like in silicon solar cells[57]).This could be bene?cial for the device ef?ciency since the voltage drop due to the driving force needed to dissociate electron–hole pair can be avoided.Classical studies indicate that the generated electron–hole pair seems to behave as a Mott–Wannier exciton in the CH3NH3PbI3with low binding energies of50meV[34].This indicates the possibility of charge separation within the absorber itself.This coupled with long charge carrier diffusion lengths can explain the good performance in a thin?lm con?guration[57](schematically shown in Fig.2b, device characteristics in Fig.3c).Electron and hole carrier diffusion lengths can be measured in the perovskite layers by combining them with selective electron or hole acceptors in a bilayer con?g-uration.Primary measurements on such bilayers have shown that electron and hole transport lengths in the perovskite?lms are balanced and at least100nm[58].These results justify the excel-lent performance achieved in relatively thick layers($350nm) of CH3NH3PbI3àx Cl x[15,59],where the device con?guration was planar.Increasing the permittivity of the material(leading to low electron–hole binding energy)can enhance the charge generation.

Under this scenario the key physical mechanism of the devices are substantially different.Here the perovskite does not only absorb photons but also transports both electrons and holes, bene?ting therefore from ambipolar conduction[15].Promising conductivities have been reported for perovskite structured mate-rials as CsSnI3[60]and CH3NH3PbI3[61],but improving the carrier diffusion lengths is a pertinent way of boosting device ef?ciency.A possible route in this direction could lay in the replacement of the metal cation or the halide anion,as reported for CH3NH3PbI3 perovskite groups where replacing I or Pb modi?ed conductivity [62].Sn-based perovskites have shown good charge carrier con-ductivities[63,64],although devices based on this kind of absor-bers did not offer photovoltaic performance.While exploring newer classes of perovskite compounds,it is essential that the crystallizing nature of these compounds is not restricted.This is important both for conductivity and charge generation,since the crystallinity determines the distribution of energetic states.In addition to the solution based deposition processes,other tech-niques such as evaporative deposition may have to be pursued [65].Perovskites made from different deposition techniques will have to be compared against single crystals(either grown from solution[31]or physical vapor transport[61])in order to ensure low defect densities and less energetic disorder.

Determining the relative advantage of employing the perovskite purely as a sensitizer(electron-injection into the mesoporous semiconductor)as opposed to a thin?lm absorber is one of the key issues to be addressed.In the thin?lm approach the bulk recombination within the perovskite gains prominence,whereas interfacial recombination is dominant in the sensitized architec-ture.The reduction of the energetic defects that can act as recom-bination centres or traps in the material is then necessary for ef?ciency improvement in the thin?lm con?guration.Spatially resolved measurements are a valuable method to investigate the location of charge generation,which in combination with the optoelectrical techniques previously presented will lead to the determination of the applicable model.Impedance spectra (Fig.2c)can be?tted to extract the chemical capacitance in this kind of devices.Since the chemical capacitance(C m)directly

FIGURE5

(a)Energetics losses and possible avenues for performance improvements in perovskites-based solar cells and(b)cost per Watt peak(Wp)as a function of module cost and module ef?ciency,as a function of module cost and module ef?ciency,assuming that balance-of-systems costs can be reduced to US$100 per square metre.Reprinted from Ref.[2].

21R E S E A R C H : R e v i e w

re?ects the density of states of the material that is being electrically charged,its determination allows discernment of the perovskite’s role.Similar values obtained for C m in TiO 2(injecting)and ZrO 2(non-injecting)based devices seem to indicate that the conduction through the perovskite could play a prominent role even in the injecting case [66].

Basic questions are still open,and it is an urgent requirement to determine the optimal con?guration for these solar cells.Deeper research of this matter can open new pathways to increase ef?-ciency,contingent on a more complete understanding of the working principles of these devices.

Prospects,ef?ciency and roadmap

In the case of CH 3NH 3PbI 3,since the onset of light absorption is ca.800nm,the ideal maximum photocurrent is $27mA cm à2[47].Provided 90%of IPCE is achievable between 400nm and 800nm,24mA cm à2of photocurrent can be collected.Thin devices (200–300nm of absorber)obtain the champion ef?ciencies,but the EQE slightly drops at wavelengths higher than 550nm [1],indicating a lack of absorption in this range.Therefore,the incorporation of plasmonic absorbers could yield current enhancement due to better light harvesting [67–69]at low ?lm thicknesses.Concur-rently,a suppression of the recombination could lead to V OC s of around 1.1V (see Fig.5a)in the ssDSC con?guration with TiO 2and spiro-OMeTAD,and even higher in the thin ?lm con?guration.Under these considerations,ef?ciencies $20%becomes achiev-able with a FF of 75%(Fig.5a),surpassing ef?ciencies of amor-phous Si cells and bringing them to comparable levels of multicrystalline Si.

Thus far,much of the efforts have focused on CH 3NH 3PbI 3,however,there are reports of other candidates from this family of materials that are suitable for solar cells.Perovskite is a ?exible structure type and many elements in the periodic table (such as Co 2+,Fe 2+,Mn 2+,Pd 2+,and Ge 2+)can be incorporated through various structural adaptations.The Goldschmidt tolerance factor [28,70]can be used for guidance as to which combinations of elements may form a stable structure.There is a drive to replace Pb 2+with a less toxic element with Sn as one of the obvious candidates.Nevertheless its easy oxidation creates Sn 4+that ori-ginates a metal-like behaviour in the semiconductor which lowers the photovoltaic performance [61].Other alternatives that could be explored include other organic cations such as formamidinium.By introducing longer chained organic components at the ‘A’site,perovskite-type layer compounds comparable to the Ruddlesden–Popper series [71]can be created.Ab-initio calculations [36]are therefore needed as guidance to identify newer families of photo-voltaic perovskites.

Commercial viability of new energy generating technologies such as perovskite solar cells ultimately relies also on improved costs per Watt peak (Wp)of installation compared to existing competition in the marketplace (see Fig.5b).With light-absorbers and electrode materials amenable to application techniques at low temperatures such as spray,blade coating and roll-to-roll printing,perovskite solar cells are promising high ef?ciency,lightweight,cost-effective options.The energy payback times for these solar cells,estimated to be similar to that of DSC’s (i.e.less than one year compared with up to three years for silicon solar cells [2])are an extra advantage over the competition.In a ?rst step towards this

direction,completely low temperature processed (<908C)solar cells with ef?ciencies close to 10%have been realized [72].

While estimating lifetime costs,one also needs to account for stability and toxicity in addition to high ef?ciencies.In the lab,500h device stabilities have already been reported in dry ambient [1,12],however it has also been reported that humidity degrades the CH 3NH 3PbI 3performance [37].Halide alteration has been promis-ing in this regard with increased stability in humid environments at minor performance penalty [37]occurring when a low proportion of Br (20%versus 80%I)is introduced.Lead content is another draw-back for the viability of these cells.Although material restriction laws such as the European Union Restriction on Hazardous Sub-stances (RoHS)can make exceptions for speci?c products (e.g.Cd in solar panels),CH 3NH 3PbI 3degradation under water exposure makes encapsulation studies vital for commercialization.The roadmap to successful commercialization thus entails the comprehensive understanding of the photovoltaic principles and degradation mechanisms,new materials development,optimization of device structure and related manufacturing technologies.

The rapid strides in the development of highly ef?cient organic/inorganic halide perovskite solar cells now demands attention to be paid to fundamental studies while simultaneously pursuing new materials and device development for the widespread deploy-ment of this solution processed photovoltaic technology.

Acknowledgements

The authors want to acknowledge the productive discussions with Thomas Baikie (NTU,Singapore),Damion Milliken and Hans Desilvestro (Dyesol Limited,Australia).This work was supported by the National Research Foundation Singapore (Project #:NRF-CRP4-2008-03)and SinBeRISE CREATE programme.

References

[1]J.Burschka,et al.Nature 499(2013)316.[2]M.Graetzel,et al.Nature 488(2012)304.

[3]B.C.O’Regan,M.Gra

¨tzel,Nature 353(1991)737.[4]A.Yella,et al.Science (New York,NY)334(6056)(2011)629.[5]U.Bach,et al.Nature 395(1998)583.

[6]J.Burschka,et al.Journal of the American Chemical Society 133(2011)18042.[7]I.Chung,et al.Nature 485(2012)486.

[8]G.Hodes,D.Cahen,Accounts of Chemical Research 45(2012)705.[9]P.Maraghechi,et al.ACS Nano 7(2013)6111.[10]J.A.Chang,et al.Nano Letters 12(2012)1863.[11]M.Nanu,et al.Nano Letters 5(2005)1716.[12]H.-S.Kim,et al.Scienti?c Reports 2(2012)1.[13]M.M.Lee,et al.Science 338(6107)(2012)643.[14]J.H.Heo,et al.Nature Photonics 7(2013)486.

[15]J.M.Ball,et al.Energy &Environmental Science 6(2013)1739.[16]N.-g.Park,The Journal of Physical Chemistry Letters 4(2013)2423.[17]C.K.M?ller,Nature 180(1957)981.[18]C.K.M?ller,Nature 182(1958)1436.

[19]A.Kojima,et al.Journal of the American Chemical Society 131(2009)6050.[20]J.-H.Im,et al.Nanoscale 3(2011)4088.

[21]J.H.Heo,et al.The Journal of Physical Chemistry C 116(2012)20717.

[22]P.Mauersberger,F.Huber,Acta Crystallographica Section B 36(3)(1980)683.[23]K.Shum,et al.Applied Physics Letters 96(22)(2010)221903.[24]D.B.Mitzi,et al.Inorganic Chemistry 38(26)(1999)6246.[25]A.Kojima,et al.Chemistry Letters 41(4)(2012)397.

[26]K.Tanaka,et al.Solid State Communications 127(9/10)(2003)619.[27]T.Ishihara,Journal of Luminescence 60–61(1994)269.[28]Z.Cheng,J.Lin,CrystEngComm 12(2010)2646.

[29]H.Arend,et al.Journal of Crystal Growth 43(2)(1978)213.[30]T.Umebayashi,et al.Physical Review B 67(15)(2003)155405.

[31]

T.Baikie,et al.Journal of Materials Chemistry A 1(18)(2013)5628.

RESEARCH:Review

[32]D.B.Mitzi,Synthesis,structure,and properties of organic–inorganic perovskites

and related materials,in:Progress in Inorganic Chemistry,John Wiley&Sons, Inc.,2007.p.1.

[33]S.Zhang,et al.Acta Materialia57(2009)3301.

[34]K.Tanaka,et al.Solid State Communications127(2003)619.

[35]G.C.Papavassiliou,et al.Advanced Materials for Optics and Electronics9(6)

(1999)265.

[36]E.Mosconi,et al.The Journal of Physical Chemistry C117(2013)13902.

[37]J.H.Noh,et al.Nano Letters13(2013)1764.

[38]H.-S.Kim,et al.Nano Letters13(2013)2412.

[39]K.Liang,et al.Chemistry of Materials4756(1998)403.

[40]P.K.Nayak,et al.Energy&Environmental Science5(2012)6022.

[41]D.Bi,et al.RSC Advances3(2013)18762.

[42]E.Edri,et al.The Journal of Physical Chemistry Letters4(2013)897.

[43]D.Bi,et al.The Journal of Physical Chemistry Letters4(2013)1532.

[44]L.Etgar,et al.Journal of the American Chemical Society134(42)(2012)17396.

[45]J.-y.Jeng,et al.Advanced Materials25(2013)3727.

[46]M.Liu,et al.Nature501(2013)395.

[47]H.J.Snaith,Advanced Functional Materials20(1)(2010)13.

[48]J.Qiu,et al.Nanoscale5(2013)3245.

[49]L.Etgar,et al.Journal of the American Chemical Society134(2012)17396.

[50]P.P.Boix,et al.The Journal of Physical Chemistry C116(2011)1579.

[51]P.P.Boix,et al.ACS Nano6(2012)873.[52]G.Boschloo,A.Hagfeldt,Chemical Physics Letters370(2003)381.

[53]Q.Shen,et al.Japanese Journal of Applied Physics45(2006)5569.

[54]F.Fabregat-Santiago,et al.Journal of the American Chemical Society131

(2009)558.

[55]M.Wang,et al.Nature Chemistry2(2010)385.

[56]M.M.Lee,et al.Science(New York,NY)643(2012)1.

[57]a.V.Shah,et al.Progress in Photovoltaics:Research and Applications12(23)

(2004)113.

[58]G.Xing,et al.Science342(2013)344.

[59]G.E.Eperon,et al.Advanced Functional Materials24(2013)151.

[60]I.Chung,et al.Journal of the American Chemical Society134(2012)8579.

[61]C.C.Stoumpos,et al.Inorganic Chemistry52(2013)9019.

[62]N.Onoda-Yamamuro,et al.Journal of Physics and Chemistry53(1992)935.

[63]D.Mitzi,et al.Nature369(1994)467.

[64]D.Mitzi,et al.Journal of Solid State114(1995)159.

[65]D.B.Mitzi,et al.Chemistry of Materials11(1999)542.

[66]H.-s.Kim,et al.Nature Communications4(2013)1.

[67]H.A.Atwater,A.Polman,Nature Materials9(2010)865.

[68]B.Wu,et al.Nature Communications4(2013)2004.

[69]T.Z.Oo,et al.The Journal of Physical Chemistry C116(10)(2012)6453.

[70]V.M.Goldschmidt,IX(Oslo:Norske Vid.,1937),1926.

[71]B.V.Beznosikov,K.S.Aleksandrov,Crystallography Reports45(2000)792.

[72]H.K.Mulmudi,et al.Chemical Communications(2013).