In the OSTI Collections: Perovskites

Dr. Watson computer sleuthing scientist.

Article Acknowledgement:

Dr. William N. Watson, Physicist

DOE Office of Scientific and Technical Information

View Past "In the OSTI Collections" Articles





Alternate Text Placeholder

This month’s topic has a lot to do with the schematic figure below. 


The figure illustrates a particular arrangement of three different kinds of electrically charged atoms (ions), each illustrated by spheres of a different size and color.  The rods between the small red spheres and larger blue spheres don’t represent actual objects, but make the cubic symmetry of the arrangement easier to see.  In an actual solid material, the ions would not be so spatially separated as the spheres in the schematic are; instead, the outermost electrons of each ion would approach those of the ions closest to it.  The more nearly the relative sizes of the three kinds of ions match ideal proportions, the more closely their arrangement will match the illustrated “stacked cube” pattern, into which ions that are sized too differently cannot fit together.  The greater the deviations from ideal size ratios, the less the atomic arrangement will resemble the one illustrated.[Wikipedia]  Still, many of the chemical elements can be coordinated with others in an arrangement that resembles or exactly matches this one, which turns out to be a very common class of atomic structures for minerals. 

Figure 1.  One way to describe the perovskite structure[Wikipedia] in this illustration is as a lattice of cubic or near-cubic unit cells[Wikipedia], with small cations (positive ions) at the cell corners, large cations of different elements at the cell centers, and anions (negative ions) at the midpoints of the cell edges.  (Illustration from Wikimedia Commons[Wikimedia].) 

The mineral perovskite, made of calcium, titanium, and oxygen (CaTiO3), is a prototypical example of this class that was discovered in the Russian Ural Mountains by Gustav Rose in 1839 and named for the mineralogist Lev Aleksevich von Perovski.[Wikipedia; OhioLINK, p. 1]  This mineral’s atomic arrangement was described in 1926 by Victor Goldschmidt[Wikipedia], who applied its name to all minerals with a similar structure, synthesized the first artificial perovskites, and noted how the structure could only form with combinations of atoms that had appropriate size ratios.  This structure is the only feature that all perovskites have in common.  Because so many combinations of elements can take on the three roles in a perovskite structure, and each element has a characteristic strength of attraction between its atoms’ electrons and nuclei, the different perovskites have vastly different electromagnetic, thermal, and chemical properties, which gives the set of perovskites a huge variety of potential applications.  The different perovskites’ physical properties and uses continue to be explored today, as several recent research reports illustrate. 


Alternate Text Placeholder



One type of perovskite has been recognized for its potential to catalyze a pair of opposite reactions:  the removal of water molecules’ hydrogen atoms from their oxygen atoms, with the ionization of the hydrogens into separate protons and electrons; and the reassembly of protons, electrons, and oxygen into water molecules.  The water molecules’ constituents absorb the electrical energy used to take them apart, and their reassembly into water molecules releases the energy, so the opposing reactions provide a way to store electrical energy and recover it.  These reactions, however, are slow.  They can be sped up if the water or its constituents are in contact with a catalyst, such as an alloy of platinum, iridium, ruthenium, or palladium.  While these catalysts are initially effective, they are also expensive, and their surfaces tend to oxidize and become less effective with use. 


A paper[DoE PAGES] published in Nature Communications describes how the use of alkaline electrolytes in the storage device enables perovskites, rather than precious metals, to be used as the catalysts.  Designing perovskites that would speed up the reactions most effectively, however, requires understanding just how the perovskites interact with water and its constituents, the role of places in the perovskite structure where negative oxygen ions (oxygen anions) are missing, the chemical activity of the oxygen anions that are present, and the degree of electron sharing between the perovskite’s anions and its smaller positive ions (cations).  The paper’s authors studied the catalytic action of a set of perovskites whose small cations were cobalt and whose large cations were either all lanthanum, all strontium, or various mixtures of both.  Observations showed that the “all-lanthanum” perovskites (LaCoO3-x) had few or no missing oxygen ions (small or zero x), while those with more strontium ions in place of lanthanum ions had more vacancies among the oxygen sites.  “All-strontium” perovskites (SrCoO3-x) had the highest proportion of oxygen-site vacancies—about 10% (x ~ 0.3).  In perovskites with oxygen-site vacancies, the oxygen anions that are near those vacancies can move into them, leaving vacancies at their former locations that other neighboring oxygens can migrate into, thus making new vacancies &c. (Figure 2).  Mathematical modeling indicated that the vacancies can also serve as reaction sites for certain steps in the process of rearranging water molecules into electrons and other constituents.  

Figure 2.  Top:  Images from scanning transmission electron microscopy[Wikipedia] showing (a) the cation and anion sublattices in the perovskite La0.2Sr0.8CoO3-x and (b) layered anion-vacancy ordering in SrCoO3-x.  The upper-right corner of (a) and lower-left corner of (b) shows which columns are which elements (O – oxygen; Co – cobalt; Sr – strontium; A – lanthanum or strontium).  The brighter half of the CoO2 layers in both pictures are oxygen-deficient (e.g., those marked with white arrowheads).  As shown in the intensity profile below micrograph (a), oxygen-deficient CoO2 layers alternate with complete ones.  Bottom:  adsorbate evolution mechanism (AEM) and proposed lattice oxygen-mediated mechanism (LOM) for catalyzing the release of oxygen and electrons from water.  Red structures represent portions of the perovskite catalyst; blue ones represent molecules or anions from the water.  (From “Water electrolysis on La1-xSrxCoO3-δ perovskite electrocatalysts”[DOE PAGES], pp. 3 and 7.)


When certain materials, including some perovskites, have different temperatures at different points, they will also have electric fields between those points.  This thermoelectric effect can be used to either make a temperature gradient across such a material by applying an electric field to it, or (a matter of interest for converting heat energy to electricity) to produce an electric field that can drive electric currents by setting different points of the material to different temperatures.  A material that could easily conduct electric current, generate large electric fields to drive the current when small temperature differences are set up between different points, and simultaneously not conduct heat so readily that the temperature differences quickly even out and eliminate the electric field, would be very useful for generating electric energy from heat. 


The perovskite strontium titanate (SrTiO3), when doped with lanthanum ions to substitute for some of the strontium and add an extra conduction electron per ion, is a promising material for this purpose, though its thermal conductivity at room temperature is high enough to limit its applicability.  U.S. Patent 9,269,880, “High ZT bismuth-doped thermoelectrics”[DOepatents], describes how the thermal conductivity is less for similar materials doped with bismuth ions in place of some of the lanthanum ions.  Observations described in the patent show that replacing lanthanum with bismuth also increases the size of the voltage-difference/temperature-difference ratio.  However, the electrical conductivity is shown to decrease enough to offset these advantages, at least up to the point that lanthanum and bismuth are present in equal amounts. 


Some materials respond to electric fields by stretching.  If the material is a film on a substrate, though, its stretching response to electric fields aligned parallel to the substrate is usually hindered by the substrate’s clamping action.  The clamping can be reduced by etching a suitable pattern into the substrate, but this process is usually complicated.  A different way around this problem is reported in “Ferroelastic[Wikipedia] switching in a layered-perovskite thin film”[DOE PAGES], in which the substrate is a perovskite and the film has a perovskitelike layered structure.  The report describes its authors’ experimental demonstration that films of Bi2WO6, grown[Wikipedia] on the perovskite substrate strontium titanate, can stretch in response to electric fields without hindrance from clamping.  The report concludes that the demonstrated reversible control of significant stretching in Bi2WO6 at room temperature provides a promising framework to study how the stretching of materials’ atomic arrangements may couple to other material properties in future micrometer-scale or nanometer-scale devices. 


Small, thin perovskite films have a variety of uses in electromechanical, electrochemical, and thermodynamic devices.[SciTech Connect (p. 8)]  The design of such devices can be usefully informed by an understanding of how the films’ structure, electrical properties, and optical properties depend on how thin they are and on their temperatures.  A study of these properties for one type of perovskite, particularly as its atomic arrangement becomes less cubically symmetric with changes in temperature (particularly as it transitions between being a tetragonal lattice[Wikipedia] and an orthorhombic[Wikipedia] one), is described in the Nature Communications paper “Size-dependent phase transition in methylammonium lead iodide perovskite microplate crystals”[DOE PAGES].  The material examined in this work can be used in a wide range of optoelectronic devices (e.g., photodetectors, lasers, and LEDs[Wikipedia, OSTI]), and similar materials have been used to make solar cells that can convert 20.1% of the sunlight that falls on them into electrical energy. 


Two properties of methylammonium lead iodide (MAPbI3) in particular appear to be important for such uses:  how fast electrical charges move in it in response to a given electric field (electrical mobility[Wikipedia]), and how the size of the material’s individual crystals affects its optical and charge-transport properties.  The motion of ions in MAPbI3 has prevented observation of electrical charges’ mobility among the ions, and the effects of crystal size on their mobility had not been systematically investigated and published for perovskites, though they had been theorized about and had been studied for other materials.  Since a material’s electrical mobility is affected by the precise arrangement of the material’s ions, the paper’s authors used various experimental techniques to examine how MAPbI3’s variations with temperature from exact cubic arrangement seen in Figures 1 and 3 are affected by the size of the MAPbI3 crystals.  They found that their measurements, though inconsistent with some of the proposed theories, did increase the likelihood that transitions between tetragonal[Wikipedia] and orthorhombic[Wikipedia] atomic arrangements occur at lower temperatures for thinner MAPbI3 microplates.  Differently-arranged atoms require different amounts of energy to form surfaces instead of attaching to other atoms[Wikipedia], and the thinner plates’ greater surface/volume ratio affects the temperature at which their atomic arrangements can change. 

Figure 3.  Schematic of the structure of methylammonium lead iodide (MAPbI3).  Here, the octahedral corners are anions of iodine (I); the centers of the octahedra are cations of lead (Pb), and between the octahedra are the methylammonium (MA, or CH3NH3) cations—moleculelike groups with positive electric charge.  Atoms are not depicted to scale.  (Illustration of a generic methylammonium lead halide[Wikipedia] from “Ionic transport in hybrid lead iodide perovskite solar cells” by Christopher Eames, Jarvist M. Frost, Piers R. F. Barnes, Brian C. O’Regan, Aron Walsh & M. Saiful Islam[Nature Communications] in Nature Communications 6, article number 7497 (2015).  Image licensed under the Creative Commons Attribution 4.0 International license[CC BY 4.0].)

A particular strong response to laser pulses of MAPbI3films indicates that this form of the perovskite could be applied to spintronics (using the spins and magnetism of charged particles as ordinary electronics uses their linear motion and charge)[Wikipedia, OSTI], to ultrafast switching of signals between fiber-optic circuits[Wikipedia, OSTI], and to quantum informatics (the processing and communication of data using quantum bits)[Wikipedia, OSTI].  “Large polarization-dependent exciton optical Stark effect in lead iodide perovskites”[DOE PAGES] describes how laser pulses were found to change how strongly a film’s conduction electrons attach to “holes” in the film—states of motion that electrons in the film might occupy but don’t, which act like positively charged particles because the absent negative electrons aren’t cancelling some of the positive charge of the perovskite atoms’ nuclei.  Negative electrons bound by electrostatic attraction to positive holes form what are known as excitons[Wikipedia]


One effect of laser pulses on an exciton, if the pulses have appropriate frequency, is to change the energy levels of the exciton’s stable states.  States with opposite total electron-hole spin that have equal energies in the absence of such laser pulses will have different energies under a polarized laser pulse’s influence.  The energy changes are shown in the paper to depend on the intensity, frequency, and polarization of the laser pulses, and on how many femtoseconds ago the pulses disturbed the film, in a way that implies the pulses’ effect is more strongly associated with the film’s intrinsic properties than its preparation method.  The different effects of polarized pulses on excitons with different total spin suggests a control mechanism for spintronic-based quantum computing with MAPbI3 films.  The excitons might be used as quantum bits, with 0 and 1 represented by states of opposite total spin, if the excitons’ spins could remain undisturbed long enough to perform multiple independent operations on them.  The authors indicate that the physical mechanisms in MAPbI3 films which limit this time to about 2 picoseconds, and how much these limiting effects could be mitigated, were still unknown, but that further study to determine the films’ potential for quantum computing (and for femtosecond lasers) was warranted. 


For each of the stable ways in which a material’s charge-carrying entities can move, the entities (electrons, holes, excitons, &c.) will have some definite energy.  In most solids, the set of stable motions tends to have subsets of similar motions with very small energy differences between one motion and the next most similar one, while the distinct subsets are much more different from each other with larger energy differences between any two motions of different subsets.  Thus the stable-motion energies, plotted on a graph, show bands of energies separated by gaps.  A charge carrier that gains or loses a tiny amount of energy will have a slightly different motion whose energy is in the same band, while a charge carrier that gains or loses an appropriate larger amount of energy will move very differently and have an energy in a different band.  How useful a material is for many purposes is largely determined by how readily its charge carriers can change from one stable motion to a very different one—that is, what gaps there are between the charge carriers’ stable-motion energy bands. 


Determining a material’s bandgaps to check its usefulness for a particular purpose has generally required extensive calculation.  The most commonly used method to approximately compute a solid’s stable electron states will generally underestimate the bandgap sizes, while other methods require so much more calculation that they are impractical for screening large numbers of candidate materials in a reasonably short time.  This problem might be solved if there were a way to at least reliably estimate a material’s energy bandgaps without calculating every feature of the stable charge-carrier motions.  One possible way to do this with a particular class of perovskites was tested by researchers at Los Alamos National Laboratory and the University of Connecticut.  Their paper “Machine learning bandgaps of double perovskites”[DOE PAGES] describes an algorithm that took features of over a thousand perovskites with known bandgaps as input, “learned” which few of those features proved most informative about the perovskites’ bandgaps, and then checked how accurately the bandgaps of other perovskites were estimated from the same features. 


The particular class of perovskites considered are “double perovskites”, whose large cations are of two different elements instead of one, with the elements alternating in adjoining planes, adjoining columns, or adjoining cells, and whose small cations are likewise of two different elements whose neighboring planes, columns, or cells also alternate.  The double perovskites studied had only oxygen anions throughout.  From over a million combinations of features, the smallest useful set found for estimating double-perovskite bandgaps consisted of essentially the following two pairs: 


  • the energy of the next-to-lowest stable state for neutral atoms of both large-cation elements;
  • the tendency of neutral atoms of both small-cation elements to attract electrons to themselves (their electronegativities[Wikipedia]).  


The researchers found their method worked well on nonmagnetic double perovskites with oxygen anions that could be separated into two neutral single perovskites, but not on those that couldn’t be so separated—not surprising to them, since their input data didn’t include such “nonseparable” perovskites.  Finding features for estimating those perovskites’ bandgaps with equal accuracy would require further work, as would checking whether the same features were as useful for estimating the bandgaps of other materials, including double perovskites with quite different alternations of large- and small-cation elements. 


Alternate Text Placeholder




A similar investigation, involving some of the same researchers, is reported in “Finding new perovskite halides via machine learning”[DOE PAGES].  Here, the problem is to determine whether a material with an “ABX3” chemical formula, with A and B representing large and small cations, respectively, and X representing anions of one of the halogen elements[Wikipedia] (fluorine, chlorine, bromine, iodine, or astatine), will have the perovskite structure of Figures 1 and 3, or have a nonperovskite structure in which the halogen octahedra share edges instead of corners.  With a solution to that problem, one can avoid trying to synthesize a given structure from an unsuitable set of elements. 


Earlier attempts to predict whether a particular ABX3 compound would be a perovskite were based on ratios that involved the radii rA, rB, and rX of the A, B, and X ions respectively — the “tolerance factor”  first discussed by Victor Goldschmidt[Wikipedia], and the “octahedral factor”  — but in this work the researchers considered more features that are known to govern inorganic crystal structures.  These features of 185 different ABX3 compounds provided input data for a machine-learning algorithm to determine which subset of those features would yield the most information about which of the 455 other combinations of the same elements would be likely to have a perovskite structure.  The researchers found that the most informative pair of features about the compounds’ structure was indeed the “classical tolerance- and octahedral-factor pair”, while the most important quadruplet of features consisted of the same two factors plus the individual radii rA and rB, which improved upon the structure identification rate of the tolerance and octahedral factors alone by only about 1%.  No further improvement in structure identification came from accounting for more features. 


The results thus confirm that how closely a perovskite’s ionic arrangement matches the ideal of Figure 1 or Figure 3 depends on the sizes of the three species of ion.  But features of the ions’ arrangement, such as the anion-cornered octahedra’s orientations with respect to each other, can be changed even if the ion sizes are ideal.  This can be useful.  As the authors of a recent paper[DOE PAGES] in Scientific Reports point out, even small changes to a perovskite’s structure can induce unexpected subtle but dramatic effects on its electronic and magnetic behaviors.  One way to alter the structure is to grow a film of the perovskite on a substrate of different material whose own atoms are spaced differently, so the film’s atoms will end up spaced to match the substrate.  However, this method of controlling octahedral orientation is limited.  The spacing of the film’s atoms can only be changed by amounts made possible by the finite number of substrates the film can be grown on; also, changing a film’s atomic spacing along the substrate-film surface produces an opposing change, possibly unwanted, along the axis perpendicular to the surface[Wikipedia].  


The Scientific Reports paper, “Controlling octahedral rotations in a perovskite via strain doping”, describes a different way to fine-tune the octahedral orientations—by doping the perovskite structure with atoms of a fourth element.  The authors started with SrRuO3 films grown on substrates of SrTiO3, which would be somewhat distorted by fitting to the substrate, and doped the film’s structure with helium atoms, which made their own alterations to the structure without adding any conduction electrons or holes (Figure 4).  Specifically, they grew a 20-nanometer thick film of SrRuO3 on a single crystal of SrTiO3, capped it with a 10-nanometer thick gold film, and then implanted it with helium atoms launched toward it at about 440 kilometers per second (i.e., with kinetic energies of 4000 electron-volts each).  Helium-induced dislocations and sputtering were reduced by the gold film’s presence.  Controlling the number of helium atoms implanted let the authors control the extent to which the oxygen-anion octahedra rotated, and thus control the film’s transition from orthorhombic to tetragonal phase.  Since the helium atoms hardly diffuse at ambient temperatures, the strain they produce in the film is stable, but the strain can be reversed by heating the film for long periods to high temperatures, making the strain a mark that can be written and erased.  The authors state that their method of strain engineering with helium “is of general nature and can be extended to other perovskite oxides or even other material classes, such as garnets, apatites, or spinels”, provides a new way to precisely tailor oxides’ local structure that’s suitable for wafer-scale processing, and should be applicable to imprinting whole crystals or locally writing oxygen octahedra patterns. 


Figure 4.  “To understand how strain doping impacts orthorhombic distortions, we measure [x-ray diffraction] … around the [SrTiO3] substrate reflection with the sample successively rotated by 90° with respect to the film normal.  As expected, an orthorhombic distortion of the perovskite unit cell is clearly evidenced for the undosed [SrRuO3] film by the different spacings of the peaks with respect to the substrate peak position [a].  This result shows that the as-grown [SrRuO3] film is epitaxial to the [SrTiO3] substrate with an in-plane oriented orthorhombic unit cell, which is in agreement with previous studies.  After implanting 2.5x1015 He/cm2, we see that the orthorhombic distortion is no longer visible [b].  The blue line highlights that the [SrRuO3] peak separation to the [SrTiO3] substrate is identical for every reflection and that peak separation is increased as compared to that of the as-grown film due to the out-of-plane lattice expansion.  This trend also reiterates the fact that the in-plane lattice remains locked to the substrate and that strain uniformity throughout the film thickness is excellent.  Ultimately, these results demonstrate that the induced out-of-plane strain is sufficient to trigger a transition from an orthorhombic to a tetragonal unit cell.”  (From “Controlling octahedral rotations in a perovskite via strain doping” [DOE PAGES], pp. 4 and 11.)

According to one report[SciTech Connect] by researchers at Cornell University, Stanford University, and SLAC National Accelerator Laboratory, the ability to make and manipulate materials in a two-dimensional form “has repeatedly had a transformative impact on science and technology”.  The report notes that films grown on a substrate are often useful, or would be if they could easily be separated from the substrate, which they aren’t always.  Many methods of breaking the chemical bonds between film and substrate tend to damage the film by changing its atomic structure near the surface—a significant problem for very thin films which have little or no bulk that isn’t near the surface.  A solution to this problem for freestanding perovskite membranes is presented by the report, “Synthesis of Freestanding Single-crystal Perovskite Films and Heterostructures by Etching of Sacrificial Water-soluble Layers”:  namely, growing an easily-removed layer on the substrate and then growing the film on that, and afterward sacrificing the middle layer instead of the film to any damage from the removal process. 


“The key is the epitaxial growth of water-soluble Sr3Al2O6 on perovskite substrates, followed by in situ growth of films and heterostructures [i.e., structures made with dissimilar crystals[Wikipedia]—wnw].  Millimetre-size single-crystalline membranes are produced by etching the Sr3Al2O6 layer in water, providing the opportunity to transfer them to arbitrary substrates and integrate them with heterostructures of semiconductors and layered compounds.”


The choice of Sr3Al2O6 and water was not initially obvious.  Various sacrificial layers and etchants have been used to produce films of different materials, since for complex materials like perovskites the selection of both “is restricted by many parameters, such as:  etchant selectivity, lattice symmetry and matching for epitaxial growth [i.e., only materials with certain atomic spacings can be grown on a given substrate—wnw], and stability of the sacrificial layer during the target membrane growth (often at high temperatures under varying thermodynamic conditions)”.  What the report authors aimed for was a sacrificial layer-etchant combination that could “be applied to virtually all perovskites and their heterostructures”.  Since the repeating units of the Sr3Al2O6 atomic arrangement closely resemble those of the “most representative perovskite substrate” SrTiO3, it’s easy to grow one on the other; and if the result is submerged in water, the Sr3Al2O6 dissolves away from the SrTiO3 in seconds.  Experiments in growing various thin perovskite films on Sr3Al2O6-on-SrTiO3 combinations, removing the films by dissolving the Sr3Al2O6 , and transferring the films onto silicon wafers were successful:  the films had the intended atomic arrangement and were removed with their surfaces “uniform and intact, even down to few nanometre thicknesses … [matching] the originally grown thickness to within 2 % error”; also, the films remained single crystals throughout, with their physical properties preserved or even enhanced after release from the Sr3Al2O6-on-SrTiO3 they were grown on. 


An investigation of perovskite film growth itself is described in a slide set prepared for the 2016 Spring Meeting of the Materials Research Society and entitled “Studying Perovskite-based Solar Cells with High Resolution in situ Microscopy”[SciTech Connect].  Since the most significant contributor to improving the growth of a particular solar cell material is its growth environment, the focus of the investigation was the use of a new technique for microscopically observing how the material’s structure changes while growing in that environment instead of after removal from it.  The microscope used, like others of the same type, uses an electron beam that passes through a single spot of the sample to form an image of it; an image of the whole sample can be constructed as the beam is scanned across many individual spots.  The difference from earlier observation methods is that the microscope is set up to observe the sample in the environment that it grows in, while it’s growing.  The investigators examined how the structure of one strong candidate for high-efficiency solar cells, formamidinium[Wikipedia] lead iodide (FAPbI3 or CH(NH2)2PbI3), changed while it grew in environments of different temperatures and humidities.  They came up with a recipe for growing thin FAPbI3 films that they successfully integrated into planar heterojunction solar cells with efficiencies up to 16%, and found that the growth environment directly affected the solar cell efficiency. 


Alternate Text Placeholder


Use in solar cells


Solar cells are in fact one of the main foci of attention in perovskite research, for reasons described in a single-page summary from the National Renewable Energy Laboratory (NREL).  According to this report (“NREL Studies Carrier Separation and Transport in Perovskite Solar Cells”[SciTech Connect]), although newer perovskite-based solar cells have had successively higher efficiencies, further progress requires a deeper understanding of their physical and chemical processes.  One thing to clarify is how they actually carry electric charge.  Do the charges move as separate electrons and holes, or do the electrons and holes bind together and move as excitons?  NREL researchers scanned solar cells based on organic-inorganic hybrid perovskites like methylammonium lead iodide with a probe[Wikipedia] that measures electrostatic forces between the cells and itself.  The way the forces vary over the cells’ surfaces indicates that

  • electrons and holes not bound to each other collect near the perovskite’s junction with a titanium-dioxide layer, with most (not all) of the electrons on one side and most (not all) of the holes on the other, and
  • the carrier types that are in the minority on either side of the junction carry charge by diffusion and drift. 


“One conclusion”, according to the report, “is that improving mobility is a rational route for enhancing conversion efficiency.  Future research can compare the mobility-improved devices with the electrical field distribution and then compare it with inorganic thin-film devices.”  The report references a Nature Communications paper on this work, “Carrier Separation and Transport in Perovskite Solar Cells Studied by Nanometre-Scale Profiling of Electrical Potential”[DOE PAGES]


Complementing that experimental study, a different research group simulated charged-particle transport in small portions of methylammonium lead iodide by computer to see how the electrons and holes behaved at an atomic-scale level of detail—further information for efficient solar-cell design.  Among their findings, reported in “The nature of free-carrier transport in organometal halide perovskites”[DOE PAGES], are that 

  • the sublattice of lead cations is a pathway for the rapid transport of electrons energized and made mobile by the perovskite’s absorption of light;
  • the holes concurrently produced by the light absorption have a similar pathway in the iodine anion sublattice;
  • these electron and hole pathways minimize the amount of electron-hole recombination (loss of current due to current electrons’ “falling” into holes and rereleasing absorbed light energy into another form);
  • the sublattice of CH3NH3 cations quickly screens out electrostatic electron-hole attractions so that charge-carrying electrons and holes are generated within one picosecond. 


Thus light absorbed by the perovskite causes almost instantaneous dissociation of electrons and holes, which can travel long distances for their energy to be harvested before being dissipated as heat. 


While the experiments and simulations just described may rule out excitons as charge-carrying quasiparticles in solar cells based on methylammonium lead iodide, other researchers who did further experiments with that perovskite and the similar methylammonium lead bromide (CH3NH3PbBr3) suggest that the electrons may interact with the CH3NH3 cations to form quasiparticles[Wikipedia] of a different sort:  polarons[Wikipedia].  A polaron is a combination of a charged particle moving among ions and the pattern of distortion it induces in the ion lattice as opposite charges in the lattice are attracted toward it until the charged particle has passed.  Although opposite charges return to their equilibrium positions in the lattice, the distortion pattern follows the moving charged particle as it moves by.  The measurements that prompted this suggestion are reported in the Nature Communications paper “Extended carrier lifetimes and diffusion in hybrid perovskites revealed by Hall effect and photoconductivity measurements”[DOE PAGES].  A particular achievement was to develop new measurement methods that were sensitive enough to work on methylammonium lead halides, which have low electrical conductivities and charge-carrier mobilities when not exposed to light.  The researchers found that these measurements and others were consistent with rotations in the CH3NH3 ions, whose positive charge is predominantly on the nitrogen end, being induced by passing electrons’ attracting the positive charge and holes’ repelling it (Figure 5).


Figure 5.  Schematic diagram of polarons in a methylammonium lead halide lattice.  Negatively charged electrons and positively charged holes, which respectively attract and repel the positive ends of the methylammonium cations, accordingly reorient them (green arrows) to form dipolar polarons.  “Such polarons,” according to “Extended carrier lifetimes and diffusion in hybrid perovskites revealed by Hall effect and photoconductivity measurements”[DOE PAGES], “could account for the relatively low intrinsic charge-carrier mobilities and reduced bimolecular recombination coefficients experimentally revealed in this work.”  Upper left inset:  unit cell structure like the one in Figure 3, with lead halide octahedral and a polar methylammonium molecular cation at the center.  (After “Extended carrier lifetimes and diffusion in hybrid perovskites revealed by Hall effect and photoconductivity measurements”[DOE PAGES], p. 6.) 


Researchers have made still other observations of methylammonium lead iodide to better understand the limits on their performance.  In experiments reported in “Photo-induced halide redistribution in organic–inorganic perovskite films”[DOE PAGES], the researchers examined irregularities that show up, even in state-of-the-art CH3NH3PbI3 films used in high-efficiency solar cells, that limit charge-carrier transport.  Their combination of three imaging techniques provides visual evidence that exposing the films to light induces migration of iodine anions, and that this migration affects and is affected by interactions among the anions, the charge carriers, and impurities or imperfections in the film that restrict the charge carriers’ motion.  The same researchers showed in other work that maximum power output is reached sooner when the density of charge-carrier trapping sites is reduced, and recommend seeking ways to make perovskites with uniform iodine distributions and low trap-site densities.  Whether that uniform distribution requires having exactly three iodine anions per lead or methylammonium cation, or fewer than that, for maximum effectiveness “still remains unclear”:  “[t]he finding that potentially large fractions of iodine are moving under illumination without major changes to crystal structure … needs to be understood and remains an open question for the community”. 


These investigations, through their improvement in our understanding of perovskite phenomena, can lead future solar cells’ being more efficient.  A different investigation examined a different sort of improvement. 


“Most research on perovskite solar cells has focused on improving power-conversion efficiency and stability. However, if one could refurbish perovskite solar cells, their stability might not be a critical issue. From the perspective of cost effectiveness, if failed, perovskite solar cells could be collected and recycled; reuse of their gold electrodes and transparent conducting glasses could reduce the price per watt of perovskite photovoltaic modules.”


The report, “Selective dissolution of halide perovskites as a step towards recycling solar cells”[DOE PAGES], elaborates on the cost-effectiveness problem.  Perovskites’ raw materials and fabrication process are cheap, but the other components of a perovskite solar cell (metal electrodes, other materials to conduct electrons and holes, coated conducting-glass substrates), which are responsible for 98% of the cost of few-square-centimeter solar cells, are still too expensive for the cells to compete with existing ways to feed energy into power grids.  New methods for making the nonmetallic conductors may substantially reduce their cost, but the other materials, as limited resources, seem unlikely to become cheaper and may vary significantly in price along with worldwide demand. 


Hence the interest in recycling degraded solar cells.  The most effective method discovered by the report’s authors involves immersing the cell in a solvent whose molecules are electric dipoles with distinct positive and negative ends, with the positive end not being an acidic hydrogen atom[Wikipedia].  The metal electrodes peel away and the perovskite and nonmetallic conductor dissolve, leaving the coated glass substrate intact.  The substrate is then rinsed and dried, and the other materials are redeposited onto it in sequence.  Even after 10 recyclings, the solar cell’s performance was like new.  The authors also demonstrated an effective way to remove residual lead from the solution that came from the perovskite, which would otherwise pose a serious waste-disposal problem.  The authors tried their recycling process with different perovskites and different solar-cell fabrication methods, and found it applicable to all of them.  They concluded from their results that, while perovskite solar cell stability remains an important issue, their long-term stability isn’t as essential as with the case of commercialized cells made of other materials like silicon, cadmium telluride, or copper indium gallium selenide, and that if perovskite solar cells “can be easily refurbished, then these cells can be used as batteries that could become available in consumer products.  The combination of high efficiency, recyclability and plastic-based flexibility in [perovskite solar cell] technology will shift the paradigm” of perovskite solar cell use. 


Alternate Text Placeholder







Wikimedia Commons



Reports available through OSTI’s SciTech Connect





Journal articles available through DOE PAGES







  • “Carrier Separation and Transport in Perovskite Solar Cells Studied by Nanometre-Scale Profiling of Electrical Potential” [Metadata]
    Nature Communications, Volume 6



  • “Extended carrier lifetimes and diffusion in hybrid perovskites revealed by Hall effect and photoconductivity measurements” [Metadata]
    Nature Communications, Volume 7


Patent available through DOepatents


  • “High ZT bismuth-doped perovskite thermoelectrics” [Metadata]U.S. Patent 9,269,880


Other references


  • “Ionic transport in hybrid lead iodide perovskite solar cells” by Christopher Eames, Jarvist M. Frost, Piers R. F. Barnes, Brian C. O’Regan, Aron Walsh & M. Saiful Islam[Nature Communications] in Nature Communications 6, article number 7497 (2015) is the source of Figure 3, which is licensed under the Creative Commons Attribution 4.0 International license[CC BY 4.0].

  • “Perovskite Synthesis and Analysis Using Structure Prediction Diagnostic Software”[OhioLINK]
    Dissertation available at OhioLINKThe Ohio State University

  • The introduction of “Crystallography and Chemistry of Perovskites”[arXiv], by researchers at Stockholms universitet [Swedish] and Technische Universität Braunschweig [German], describes the importance of perovskites in the following terms:

    “The interest in compounds belonging to this family of crystal structures arise[s] in the large and ever surprising variety of properties exhibited and the flexibility to accommodate almost all of the elements of the periodic system.  …  Distorted perovskites have reduced symmetry, which is important for their magnetic and electric properties.  Due to these properties, perovskites have great industrial importance, especially the ferroelectric tetragonal form of BaTiO3.”

  • The review article “The Perovskite Structure – A Review of Its Role in Ceramic Science and Technology”[ResearchGate], published in Material Research Innovations Volume 4 (2000), describes the beginnings of human knowledge of perovskites and their potential as seen at the time of publication.  The abstract reads:

    “Starting with the history of the fundamental science of the relation of structure to composition delineated completely by Goldschmidt, we use the perovskite structure to illustrate the enormous power of crystal chemistry-based intelligent synthesis in creating new materials. The perovskite structure is shown to be the single most versatile ceramic host. By appropriate changes in composition one can modify the most significant electroceramic dielectric (BaTiO3 and its relatives) phase in industry, into metallic conductors, superconductors or the highest pressure phases in the earth. After an historical introduction of the science, detailed treatment of the applications is confined to the most recent research on novel uses in piezoelectric, ferroelectric and related applications.”


In the OSTI Collections



Alternate Text Placeholder