Wednesday, June 29, 2016

Perovskite Solar Cells - Some Recent Developments.

I have written the following article, which will be published later in the year in the journal Science Progress of which I am the editor. However, given the importance of its subject and the wide interest in perovskite solar cells, I am posting it here as a preview.

Perovskite solar cells were the subject of a current commentary, previously  published in this journal1 These are solar cells whose structure utilises a perovskite-type compound as the light-harvesting active layer.1 The term 'perovskite solar cell' originates from the ABX3 stoichiometry of the absorber materials, which translates to a perovskite crystal structure (Figure 1). The most extensively studied absorber material is methylammonium lead trihalide (CH3NH3PbX3; where X is a Cl, Br or I atom), with a bandgap energy in the range of 1.52.3 eV. The related material, formamidinum lead trihalide (H2NCHNH2PbX3) is also promising, with bandgaps that occupy the range 1.52.2 eV. Indeed, since the smallest bandgap approaches more closely the optimum value for a single-junction cell than does methylammonium lead trihalide, it is expected that higher efficiencies might be attained using these materials2. It has been shown that the bandgap in the methylammonium lead halides can be tuned by varying the halide content, and2,3 that the diffusion lengths for both holes and electrons in these materials are greater than 1 micron4. The latter feature means that charges can be transported over long distances in the perovskite, and the materials are effective in thin-film architectures. The charges per se have been demonstrated to comprise predominantly free electrons and holes, and are not bound excitons. This is a consequence of the fact that the binding energy for excitons is sufficiently low that charge separation can occur at room temperature5.  

The rise in power conversion efficiencies of solar cells based on perovksite materials has been little short of meteoric, considering that just 3.8% was achieved in 2009 6 but this had climbed to 22.1% in early 2016 7 (Figure 2). Working on the premise that even higher efficiencies should be attainable and that the production costs of perovskite solar cells are relatively low, they have attracted sufficient commercial interest that a number of start-up companies have expressed their intention to be selling actual photovoltaic modules by 2017 8,9. In view of the potential widespread importance of perovskite solar cells in the future, it is most appropriate to include mention of them in undergraduate courses. To this end, a very timely account has been published which demonstrates how a simple such cell can be constructed for the purposes of practical demonstration10.

Production of perovskite solar cells.
The processing of and fabrication of perovskite solar cells is considerably less demanding than is the case for the more usual silicon-based cells, for which costly multistep processes are necessary, needing to be carried out in a clean room, involving high vacuum facilities, and temperatures in excess of 1000 °C 11. In contrast, organic-inorganic perovskites can be produced using wet chemistry techniques under standard laboratory conditions. Thus, it has proved possible to synthesise the above mentioned methylammonium and formamidinium lead trihalides by means of various solvent-based and vapour deposition methods, with the indication that future production on the large scale should be possible12,13. Despite the attractive simplicity of the method, solution processing leaves voids, platelets, and other defects in the perovskite layer, which might act to impede the photovoltaic efficiency of the cell. To avoid such problems, a method has been presented14 that produces perovskite films with both a large crystal dimension and a smooth structure, with more advantageous light conversion efficiencies. An alternative approach, which involves solvent-solvent extraction at ambient temperature, enables the generation of crystalline films which cover an area of several square centimetres, in a controlled fashion, down to a thickness of 20 nanometers, and without creating pin-holes.. An alternative method15 involves annealing the spin coated, or exfoliated lead halide in the presence of methylammonium iodide vapour at temperatures close to 150 °C. This offers certain advantages over solution processing, since it may enable the creation of multi-stacked thin films over larger areas16, as might be a useful feature when it is desired to make multiple-junction cells. As a general point, we note that vapour deposition methods yield more uniform thicknesses than are created by simple solution processing. Both solution based and vapour deposition approaches have their own advantages, and should prove scalable at the manufacturing level, with much reduced costs and complexity in comparison with making silicon photovoltaic cells. Solution methods are cheaper than vapour based approaches, although the latter naturally eliminate some of the need for solvents and having to deal with solvent residues. Planar thin film layers can be created using both approaches, and either find application in mesoscopic designs, for example to place coatings on a metal oxide scaffold, as is typical for the perovskite or dye-sensitized solar cells that are currently made. However, a major uncertainty over the use of perovskite solar cells concerns their stability over the longer time, in view of the observation that such materials undergo degradation under typical environmental conditions, with concomitant losses in light conversion efficiencies.

Estimating the efficiency limit for perovskite solar cells.
The Shockley–Queisser limit (detailed balance)17 for radiative efficiency is about 31% under an AM1.5G solar spectrum at 1000 W/m2, as applied to a Perovskite with a bandgap of 1.55 eV18. This is somewhat less than the radiative limit of gallium arsenide (bandgap 1.42 eV) which can reach a radiative efficiency of 33%. A detailed balance model for the maximal output power of a solar cell that can be achieved on the proviso that certain conditions are met17:

(1) Carrier populations obey Maxwell–Boltzmann statistics. In particular, the quasi-Fermi levels of electrons and holes must be split homogeneously through the cell, where the magnitude of the “split” is equal to the applied voltage. This assumes that the mobilities of the electrons and holes (photocarriers) are sufficiently large, and indeed, charge carrier mobilities up to 10 cm2 V−1 s−1 have been observed in perovskites.

(2) The only existing recombination mechanism is that of a band-to-band (monomolecular) process, while other nonradiative recombination processes, such as Auger recombination and trap (defect) assisted recombination can be ignored.

(3) The internal light conversion efficiency is 100%, such that for each photon absorbed, one electron-hole pair is created; and in reverse, one photon is produced for each recombination of an electron-hole pair. Indeed, an internal quantum efficiency of close to 100% has been demonstrated for perovskites19.

(4) It is assumed that a photon recycling effect operates within the cell. Although one photon is created by one electron-hole pair recombination, occurring during the radiative recombination process, that photon can be reabsorbed elsewhere in the cell, such that a new electron-hole pair is generated. This photon reabsorption process can be enhanced so to achieve maximum solar cell efficiency by using particular designs to achieve light trapping and angular restriction. Nonetheless, whether such a photon recycling effect actually occurs in perovskite solar cells has yet to be convincingly demonstrated.
The detailed balance model18 for the efficiency limit of CH3NH3PbI3 solar cells can be described by:

= Jt(V) + Je(V) - Jp

where V V {\displaystyle V} is the applied voltage for the solar cell system, J p {\displaystyle J_{p}} Jp is the photocurrent (photogenerated current) which results from the absorption of incident sunlight, Je is the current density loss from radiative emission, and Jt J e {\displaystyle J_{e}} J t {\displaystyle J_{t}} Je and Jtjjis the loss of current density as a result of trap-assisted nonradiative recombination. Many approaches have been made, including analytical calculations, to identify various relevant properties of perovskites, such as bandgap, effective mass, and defect levels, while other structure-function methods have been explored as a means to interpret the more detailed properties of these materials as they might function in photovoltaic applications20.

Architectures for perovskite solar cells.
There are various different arrangements (“architectures”) possible for the design of a perovskite solar cell (Figure 3), according to the particular function of the perovskite material in the device, or the nature of the top and bottom electrode. Where positive charges are to be harvested by a transparent bottom electrode (cathode), such devices can broadly be classified as being 'sensitised' (where the perovskite functions primarily as a light absorber, and the subsequent charge transport takes place in other constituent materials), or 'thin-film', where the majority of the electron or hole transport is confined to the body of the perovskite per se. In the sensitised version, the perovskite material is coated onto a charge-conducting mesoporous scaffold as a light absorber, which is typically TiO2, and the photogenerated electrons are transferred from the perovskite layer to the mesoporous sensitised layer through which they diffuse to the electrode and are delivered into the circuit. In the case of the thin-film architecture21, both negative and positive charge carriers, generated by light absorption, are transported through the perovskite layer to charge selective contacts. Since the concept of dye-sensitized solar cells was at the root of perovskite cells, initially it was the sensitized architecture that was used, until it became recognised that better results could sometimes be achieved by placing the materials in a thin-film architecture. In terms of longer term stability, it is to be noted that sensitised architectures are likely to suffer from degradation as induced by absorption of ultraviolet light, which may limit the photovoltaic efficiency of such devices over time.

Stability of perovskite solar cells.
The stability of perovskite solar cells, under environmental operating conditions, is critical. In this regard, the solubility of the organic constituent of the absorber material in water is likely to cause a rapid degradation in humid environments22. One potential solution to this is to encapsulate the perovskite absorber as a composite along with carbon nanotubes within a matrix of a relatively unreactive polymer: this, at least, avoids an immediate deterioration of the material in moist air at elevated temperatures22,23. Nonetheless, evidence is lacking that such encapsulation methods would be successful over the longer timescale. A further consideration, which applies to devices which contain a mesoporous TiO2 layer sensitised with the perovskite absorber, is a demonstrated instability to UV-light24, which appears to be related to holes that are photogenerated within the TiO2, along with oxygen radicals that are formed on the surface of the material24. Thermal stresses are another factor, and it is of note that the very low thermal conductivity determined for CH3NH3PbI3 [0.5 W/(Km) at room temperature] may limit the rapid transfer of heat through the material, and so provide some protection against thermally induced stress and degradation over the working lifetime of the solar cell25.  Perovskite films contain a residue of PbI2 which has been shown to affect adversely the long term stability of devices based upon perovskites14. In contrast, if a metal oxide layer is used instead of the organic transport layer, it has been shown that the cell can retain 90% of its photovoltaic capacity after 60 days26. Clearly, however, this refers only to a quite short timescale and issues of longer term stability remain.

Recent Research on Lead Halide Perovskites for Solar Energy Conversion.
A representative overview of current research in the field of perovskites for solar energy conversion devices is given by a recently published special issue of Accounts of Chemical Research. The following is a summary of some of the papers contained there. The first paper27 in the issue discusses organic–inorganic semiconductors with perovskite crystal structures, which can be converted to efficient solar cells using solution-processed active layers. Since these are made from earth-abundant materials, and the necessary processing is not complicated, the likelihood is that it will prove possible to manufacture them cheaply and on a large scale. One critical difference between these materials and more traditional inorganic semiconductors is that they furnish soft solids, in which a high degree of thermally activated motion is present. The paper considers the internal motion of methylammonium lead iodide (CH3NH3PbI3) and formamidinium lead iodide ([CH(NH2)2]PbI3), with particular emphasis on: (i) rotational-librational motion within the cuboctahedral cavity; (ii) the spatial movement of large electron and hole polarons; (iii) the transport of ionic defects bearing charges. As a consequence of the above mentioned processes, various unconventional properties are produced, which include frequency-dependent permittivity, low electron–hole recombination rates, and current–voltage hysteresis27. Another paper28 explores the relationship between structural and electronic properties on a molecular to mesoscopic scale for hybrid perovskites, some of which have been demonstrated to have solar cell power conversion efficiencies exceeding 20%. It is shown that, in hybrid lead-halide perovskites, dielectric screening is a function of the local microstructure of the hybrid crystals, in addition to basic chemical composition. It is concluded that band-gap engineering is possible and the elementary photoexcitation dynamics can be controlled in order to fine-tune the performance of the perovskites as they are present in different optoelectronic devices28.

A theoretical approach29 is employed to investigate the structural stability and electronic states of the representative (110), (001), (100), and (101) surfaces of tetragonal CH3NH3PbI3, taken to be a representative model for hole-transporting-material/CH3NH3PbI3 interfaces. Calculations, made on several different PbIx polyhedron terminations, revealed the presence of two principal phases on all of the four surface facets, which can by classified as vacant- and flat-type terminations. Under conditions of thermodynamic equilibrium, the former are the more stable. Evidence was obtained for a coexistence of both terminations on the (110) and (001) surfaces. It was concluded that the surfaces may be partially responsible for the relatively long carrier lifetime which is observed experimentally for the perovskite solar cells, in which there are no mid-gap surface states. It is reasoned that the shallow surface states on the (110) and (001) flat terminations may furnish effective intermediates for hole transport to hole-transporting materials, which may help to improve the overall performance of perovskite solar cells29. Another paper considers rational strategies for the fabrication of perovksite solar cells with high efficiency, high stability and at low cost30. Important factors for the actual achievement of these aims include the architecture, processing route and the nature of the necessary materials employed in the construction of the solar cells, which are usually constructed in the form of multilayer heterostructures of light harvesting materials, of which electron and hole transporting layers are critical components30. The structure and dynamics of hybrid perovskite materials are considered31 in regard to the impressive improvement in the power-conversion efficiencies in solar cells based on these materials, which have risen above 20% in a mere 6 years. For a solution-processable material, this is outstanding, and it is accordingly essential to comprehend the physical and chemical phenomena that underpin the photovoltaic efficiency processes. In the majority of attempts to understand semiconducting materials and their optoelectronic and charge transport properties, a model is often adopted in terms of a lattice of ions which are displaced from their static positions by harmonic vibrations alone. However, many recent studies indicate that this is an oversimplification of hybrid organic-inorganic perovskite materials, where even at ambient temperature there is a range of dynamic effects operating, in addition to those of fairly minor harmonic vibrations. In their evaluation, the authors focus upon the mechanical “softness” of these particular materials in comparison with other efficient photovoltaic materials, which permits large ionic displacements to occur, even at ambient temperature. Accordingly, their analysis extends beyond the usual simple static band structure model, in regard to phase-transition behaviour and molecular/octahedral rearrangements. Atomic diffusion phenomena are considered, particularly in terms of the migration of intrinsic and extrinsic ionic species. Thus, hybrid organic-inorganic perovskite are revealed as being highly dynamic in nature, whose charge-transport and optical properties are influenced appreciably by structural fluctuations and long-range motion of ions31.

Aspects are considered32 of the so termed “inverted” planar device structure (i.e., p-i-n), for inorganic–organic hybrid perovskite solar cells, in which p-type and n-type materials are employed, respectively, as bottom and top charge transport layers. Thus, power conversion efficiencies of up to18% have been demonstrated for devices of this kind, which can be processed at lower temperatures, more flexibility, and with only minor JV hysteresis effects. A comprehensive and critical overview is made of the according mesoporous and planar structures, in addition to both the regular and inverted planar structures. Further consideration is given to how the inverted planar structure of perovskite solar cells has been developed, with particular attention being made to such effects as film growth, band alignment, stability, and hysteresis32. It has been emphasised33 that in order to optimise further hybrid perovskite solar cells, a clearer understanding is necessary of how light is converted into currents of electricity and also of how the reverse process functions. A description is given of how optical-pump–terahertz-probe (OPTP) photoconductivity spectroscopy may be used to investigate these phenomena since it permits a monitoring of the dynamics of mobile charge carriers inside the perovskite structure, which has been excited by means of a short laser pulse to deliver a photon whose energy is within the solar spectral range. It is concluded that the trapping of charges is the principal cause of mono-molecular charge-recombination, where the trap depths are quite shallow (tens of millielectronvolts). In contrast, bimolecular recombination is the result of direct band-to-band electron–hole recombination. The latter significantly violates the simple Langevin model. Values for the mobility of charge-carriers, as obtained from OPTP measurements and how these depend on both the composition and morphology of the hybrid perovskites are discussed33.

In another paper34, published experimental evidence on the photophysics of perovskites is examined within a coherent theoretical framework, based upon well established knowledge of materials optoelectronics. One critical unknown is whether photon absorption leads to a population of unbound, conductive free charges or rather excitons, formed as a result of a Coulomb interaction occurring immediately below the band gap energy. It is concluded that all hybrid halide perovskites behave as free-charge semiconductors, and it is this feature, along with a range of band gap energies that encompass the whole solar spectrum, which determines perovskites as materials from which highly efficient, single and multijunction solar cells can be made34. Also included in the series is a paper which explores the physical changes that may occur in perovskites that have undergone illumination over a prolonged period35.  Three model systems were employed: (i) a free-standing perovskite film on a glass substrate, (ii) a symmetrical system with nonselective electrical contacts, and (iii) a working perovskite solar cell (either a planar or a porous structure).The authors make a critical distinction between photoinduced effects and photo- and electric-field induced effects, and it is demonstrated that the photoinduced effects are large and of great significance to the properties of the perovskite materials35.

In another paper36 is considered the lifecycle of organic–inorganic lead halide perovskites, especially in regard to (i) the nature of the precursor solution, (ii) the formation of solid-state perovskite thin films and single crystals, and (iii) the transformation of perovskites into hydrated phases upon exposure to moisture. It is shown that information about the thermally driven creation of the perovskite structure is readily provided by a range of spectroscopic and structural characterization techniques. It is concluded that in order to make stable, high-quality perovskite thin films for future generation photovoltaic and light emitting devices, a fundamental understanding of lead halide perovskite formation and degradation pathways is advantageous36. Another paper considers the effects of light and electron-beam irradiation on hybrid lead halide perovskites and their solar cells37. Electron beams are used in the characterization of these devices but radiation damage is incurred, meaning that such exposure should be kept to a minimum. While some overall picture of the effects of light on these materials and devices is possible, it is not as yet feasible to predict the longer term stability of these solar cells, under actual working conditions37.

In its July 2016 issue, the flagship magazine of the Royal Society of Chemistry, “Chemistry World”, published an overview38 of the current state of the art for solar cell technologies, including perovskites. It is proposed that the latter can be made in thin-film forms and are sufficiently flexible to manufacture using a roll-to-roll process. It is also possible to boost the performance of conventional silicon solar cells by layering perovskites onto them. There is research being undertaken to reduce the potential environmental toxicity of perovskites by eliminating lead as one of the elemental components. A potential solution to the problem of degradation of the cell from hydrophobic lithium salts, which are used as a hole-transport dopant, is to use instead a hydrophobic hole-transporter in conjunction with a doping technique that employs a pre-oxidised salt of the respective hole-transporters, to produce perovskite cells that are efficient in their light-to-power conversion but much less moisture sensitive. A computer screening has been made of over 4 trillion possible compounds in an effort to identify the next generation of inorganic photovoltaic materials, the data from which have yet to be fully analysed.

(1) Rhodes, C.J. (2014) Sci. Prog., 97, 297.
(2) Eperon, G. E. et al. (2014) Energy & Environmental Science, 7, 982.
(3) Noh, J.H. et al. (2013) Nano Letters, 13, 1764.
(4) Stranks, S. D. et al. (2013) Science, 342, 341.
(5) D’Innocenzo, V. et al. (2014) Nature Communications, 5, 3586.
(6) Kojima, A. et al. (2009) J. Am. Chem. Soc., 131, 6050.
(10) Patwardhan, S. et al. (2015) J. Phys. Chem. Lett. 6, 251.
(12) Saidaminov, M.I. et al. (2015) Nature Communications, 6, 7586.
(13) Snaith, H.J. (2013) J. Phys. Chem. Lett., 4, 3623.
(14) Zhang, H. et al. (2015) Adv. Energ. Mater., 5, 1501354.
(15) Chen, Q. et al. (2014) J. Am. Chem. Soc. 136, 622.
(16) Liu, M., Johnston, M.B. and Snaith, H.J. (2013) Nature, 501, 395.
(17) Shockley, W. and Queisser, H.J. (1961) J. App. Phys., 32, 510.
(18) Sha, W.E. et al. (2015) Appl. Phys. Lett. 106, 221104.
(19) Yang, B. et al. (2015) J. Am. Chem. Soc. 137, 9210.
(20) Stoumpos, C.C. and Kanatzidis, M.G. (2015) Acc. Chem. Res., 48, 2791.
(21) Eperon, G. E. et al. (2014) Adv. Func. Mater., 24, 151.
(22) Habisreutinger, S.N. et al. (2014) Nano Lett., 14, 5561.
(24) Leijtens, T. et al. (2013) Nature Communications 6, 2885.
(25) Pisoni, A. et al. (2014) J. Phys. Chem. Lett., 5, 2488.
(26) You, J. et al. (2015) Nature Nano., 11, 75.
(27) Frost, J.M. and Walsh, A. (2016) Acc. Chem. Res., 49, 528.
 (28) Kandada, A.R.S. and Petrozza, A. (2016) Acc. Chem. Res., 2016, 49, 536.
(29) Haruyama, J. et al. (2016) Acc. Chem. Res., 49, 554.
(30) Seo, S., Noh, J.H. and Seok, S.I. (2016) Acc. Chem. Res., 49, 562.
(31) Egger, D.A., Rappe, A.M. and Kronik, L. (2016) Acc. Chem. Res., 49, 573.
(32) Meng, L. et al. (2016) Acc. Chem. Res., 49, 155.
(33) Johnston, M.B. and Herz, L.M. (2016) Acc. Chem. Res., 49, 146.
(34) Saba, M. et al. (2016) Acc. Chem. Res., 49, 166.
(35) Gottesman and Zaban, A. (2016) Acc. Chem. Res., 49, 320.
(36) Manser, J. et al. (2016) Acc. Chem. Res., 49, 330.
(37) Lein-Kedem, N., Cahen, D. and Hodes, G. (2016) Acc. Chem. Res., 49, 347.
(38) Hugues, E. (2016) Chemistry World, July, 58.

Captions to Figures.
Figure 1. Crystal structure of CH3NH3PbX3 perovskites (X=I, Br and/or Cl). The methylammonium cation (CH3NH3+) is surrounded by PbX6 octahedra. Credit: Christopher Eames et al. -

Figure 2.
Solar cell efficiencies of various cell technologies as tracked by the U.S. National Renewable Energy Laboratory (NREL).
Credit: U.S. Department of Energy.

Figure 3.
Schematic of a sensitized perovskite solar cell in which the active layer consist of a layer of mesoporous TiO2 which is coated with the perovskite absorber. The active layer is contacted with an n-type material for electron extraction and a p-type material for hole extraction. b) Schematic of a thin-film perovskite solar cell. In this architecture in which just a flat layer of perovskite is sandwiched between to selective contacts. c) Charge generation and extraction in the sensitized architecture. After light absorption in the perovskite absorber the photogenerated electron is injected into the mesoporous TiO2 through which it is extracted. The concomitantly generated hole is transferred to the p-type material. d) Charge generation and extraction in the thin-film architecture. After light absorption both charge generation as well as charge extraction occurs in the perovskite layer.
Credit: Sevhab