Ionic transport in hybrid lead iodide perovskite

Full text

Turn on search term navigation

ARTICLE

Received 9 Feb 2015 | Accepted 14 May 2015 | Published 24 Jun 2015

Solar cells based on organicinorganic halide perovskites have recently shown rapidly rising power conversion efciencies, but exhibit unusual behaviour such as currentvoltage hysteresis and a low-frequency giant dielectric response. Ionic transport has been suggested to be an important factor contributing to these effects; however, the chemical origin of this transport and the mobile species are unclear. Here, the activation energies for ionic migration in methylammonium lead iodide (CH3NH3PbI3) are derived from rst principles, and are compared with kinetic data extracted from the currentvoltage response of a perovskite-based solar cell. We identify the microscopic transport mechanisms, and nd facile vacancy-assisted migration of iodide ions with an activation energy of 0.6 eV, in good agreement with the kinetic measurements. The results of this combined computational and experimental study suggest that hybrid halide perovskites are mixed ionicelectronic conductors, a nding that has major implications for solar cell device architectures.

DOI: 10.1038/ncomms8497 OPEN

Ionic transport in hybrid lead iodide perovskite solar cells

Christopher Eames1, Jarvist M. Frost1, Piers R.F. Barnes2, Brian C. ORegan2, Aron Walsh1 & M. Saiful Islam1

1 Department of Chemistry, University of Bath, Bath BA2 7AY, UK. 2 Department of Physics, Imperial College London, London SW7 2AZ, UK. Correspondence and requests for materials should be addressed to M.S.I. (email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 6:7497 | DOI: 10.1038/ncomms8497 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8497

Solar cell materials based on organo-lead halide perovskites are attracting extraordinary attention on account of the rapid rise in their solar-to-electricity conversion efcien

cies15. Other key features of these hybrid perovskites include their ease of solution-based processing at low temperature, and their strong optical absorption, although there are signicant stability issues. Devices based on methylammonium lead iodide, CH3NH3PbI3, have dominated most research632, as so far they exhibit the highest power conversion efciencies (420%); these values are comparable to those of the best thin-lm solar cells based on Cu(In,Ga)Se2 or CdTe, but greater than those of conventional dye-sensitized or organic solar cells. The ABX3 perovskite-type structure, illustrated in Fig. 1, is comprised of an extended framework of corner-sharing PbI6 octahedra with the methylammonium cation (CH3NH3) occupying the central A site and surrounded by 12 nearest-neighbour iodide ions.

Inorganic perovskite oxides have been intensively researched for decades in connection with a rich variety of solid-state magnetic, ferroelectric and conduction properties. In particular, it is well known that numerous perovskite oxides (for example, Sr- and Mg-doped LaGaO3) exhibit high ionic conductivity mediated by defect species (typically oxide-ion vacancies), making them useful in applications such as solid oxide fuel cells33,34. In addition, ionic conductivity studies of inorganic perovskite halides (for example, CuPbI3 and CsPbCl3) have reported low activation energies for the migration of halide-ion vacancies3537.

Diffusion of intrinsic ionic defects in organo-lead halide perovskites has important implications in terms of the long-term stability and performance of perovskite solar cell devices. In these hybrid perovskites, however, the exact nature of the mobile ionic species and migration activation energies are poorly understood.

Charge transport and resistivity studies of hybrid halide perovskites have largely focussed on their semiconducting (electron/hole transport) behaviour, and indicate high carrier mobility and long minority-carrier diffusion lengths1012. Currentvoltage characteristics show a hysteresis in photovoltaic performance1317; this has been speculated to be related to ion migration17,18. Xing et al.22 suggest that the nonradiative pathways involved in electronhole recombination may include bulk defects such as vacancies. From spectroscopic impedance studies, Dualeh et al.31 suggest that the hybrid perovskites might exhibit ionic charge transport in addition to electronic conduction. More recently, Xiao et al.32 ascribe

the eld-switchable photovoltaic effect to the drift of charged ions or vacancies, although they provided no direct evidence for ion diffusion. In addition to ion migration, ferroelectric and charge-trapping effects have also been suggested as possible mechanisms13. Recent analysis of quasielastic neutron scattering data23, as well as polarization and piezoresponse force microscopy measurements38 suggests that ferroelectric behaviour in CH3NH3PbI3 is unlikely to be the dominant mechanism behind the observed hysteresis. Simple capture and release of photogenerated charges in bulk or interfacial defects is also unlikely due to the magnitude and duration of the changes in photocurrent during relaxation following biasing.

Previous electronic structure calculations have been used to examine band structures, spinorbit coupling and intrinsic defects in CH3NH3PbI3 (refs 3951). The defect studies45,46 suggest that Pb2 , I and CH3NH3 (MA ) vacancies create shallow donor or acceptor levels. Recent simulation work41 shows the prevalence of ionic over electronic disorder in CH3NH3PbI3 with low formation energies for Schottky-type defects associated with anion and cation vacancies, shown in the following reaction in the KrgerVink notation for the full Schottky defect,

nil ! V0MA V00Pb 3V I MAPbI3 1 where nil represents the perfect CH3NH3PbI3 lattice, V indicates a vacancy, subscripts the ionic species and superscripts the effective defect charge (a dot for each positive charge and prime for each negative charge); recent results41 suggest a signicant equilibrium concentration of I , Pb2 and CH3NH3 vacancies at room temperature, which could support vacancy-mediated diffusion.

Macroscopic conductivity or diffusion experiments, however, have so far not enabled the mobile defect species or the atomistic transport mechanism to be identied. Here, we investigate key issues related to intrinsic defect migration in CH3NH3PbI3 using rst-principles techniques combined with kinetic experiments monitoring the photocurrent relaxation of devices, extending our previous work on hybrid perovskites3941 and ion transport in perovskite-type oxides33,52. We have carried out a detailed examination of the pathways and relative activation energies for the vacancy-mediated migration of I , Pb2 and CH3NH3 ions, and compared these values directly with kinetic data extracted from a working CH3NH3PbI3 solar cell. The results support a mechanism of vacancy-mediated ion diffusion, pointing to mixed ionicelectronic conduction in these hybrid perovskites. The implications for perovskite solar cell operation and stability are discussed.

ResultsMigration mechanisms and energies. It is well established that the mobile ionic species in the solid state is associated with some type of vacancy or interstitial defect, the concentration of which is controlled by intrinsic Schottky and Frenkel defect reactions, non-stoichiometry or aliovalent doping33,34,53. In materials with the ABX3 perovskite structure, vacancy-mediated diffusion is the most common process, which is further supported for CH3NH3PbI3 by the ease of formation of Schottky disorder (reaction (1)), with an intrinsic concentration of I , Pb2 and

CH3NH3 vacancies predicted to exceed 0.4% at room temperature41. Interstitial migration has not been observed in inorganic perovskite oxides or halides due to the lack of interstitial space in such close-packed structures.

In this study, three vacancy transport mechanisms involving conventional hopping between neighbouring positions (illustrated in Fig. 2) were considered. These were (i) I migration along an octahedron edge; (ii) Pb2 migration along the diagonal (o1104 directions) of the cubic unit cell; (iii) CH3NH3

Figure 1 | Perovskite structure of CH3NH3PbI3. Methylammonium cation (CH3NH3) occupies the central A site surrounded by 12 nearest-neighbour iodide ions in corner-sharing PbI6 octahedra.

2 NATURE COMMUNICATIONS | 6:7497 | DOI: 10.1038/ncomms8497 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8497 ARTICLE

migration into a neighbouring vacant A-site cage. Energy proles for these mechanisms were mapped out by performing a series of transition-state calculations between adjacent equivalent sites. Particular care has to be taken to ensure the starting and end point congurations are well converged to avoid anomalously low migration energies. This approach allows the energy barrier to ion migration to be determined, and has been used successfully in previous studies on oxide-ion, proton and metal-ion migration in perovskite oxides52,54. One particular challenge is to adequately describe the polarization response to the change in position of the migrating ion, which involves a combination of electronic polarization and lattice relaxation; this required the large supercells employed here to minimize any spurious interactions.

The migration activation energies for the three mechanisms are reported in Table 1. The lowest activation energy of 0.58 eV suggests favourable vacancy-assisted diffusion of iodide ions. A higher activation energy of 0.84 eV is found for CH3NH3 migration, which involves motion through the unit cell face or bottleneck comprising four I ions (ion radius of 2.20 ); this value is probably at the lower limit since we have considered the ideal migration path normal to the unit cell face. In addition, neutron diffraction55 and quasielastic neutron scattering23 studies

conrm a high level of orientational motion of the CH3NH3 ion at room temperature, which is likely to further inhibit long-range transport. Table 1 shows a high energy barrier for Pb2 vacancy migration, suggesting an immobile Pb sublattice. This result is consistent with the thermodynamic stability of the inorganic lead iodide sublattice of the perovskite structure. The high activation energy for Pb2 transport also suggests that such cation diffusion is likely to be rate controlling for crystal growth processes.

From the calculated activation energies a diffusion coefcient can be estimated. Assuming a Boltzmann-like barrier-hopping process with a typical attempt frequency of 1012 Hz for ionic species at a temperature of 320 K, we estimate a diffusion coefcient of 10 12 cm2 s 1 for I ions, which is four orders of magnitude higher than the value of 10 16 cm2 s 1 for

CH3NH3 , suggesting negligible diffusion of CH3NH3 ions. These results clearly indicate that hybrid halide perovskites are mixed ionicelectronic conductors with iodide ions as the majority ionic carriers. A key factor for ionic conductivity will be the level of intrinsic iodide ion vacancies in the material, which will be sensitive to the synthesis conditions (equilibrium or non-equilibrium) and thermal processing routes. It is recognized that, as in binary lead halides, there are stability and photodecomposition issues with these hybrid lead perovskites1,2,56, which require further investigation.

With regard to mechanistic features, it is often assumed that the migrating ion takes the shortest path between adjacent sites, that is, a direct linear jump. However, detailed analysis of the migration paths for the iodide ion vacancy mechanism reveals a small deviation from the linear route involving a curved path between iodine sites, with the saddle point slightly bowed away from the neighbouring Pb ion (shown in Fig. 3 together with the corresponding energy prole); such an atomic-scale mechanism is difcult to extract from an experiment alone. It is worth noting that analogous curved paths have been found from both atomistic simulation52 and neutron diffraction studies33,57 of oxide-ion conduction in inorganic perovskite oxides such as doped LaGaO3.

Similar mechanisms may feature in related hybrid halide perovskites.

Kinetic measurements. Several recent studies have probed currentvoltage hysteresis in hybrid perovskite solar cells1317. However, there is currently an absence of temperature-dependent kinetic data. Hysteresis is frequently observed when performing a photocurrentvoltage scan on the device. If the voltage is held at forward bias and then swept from forward to reverse, the power conversion efciency inferred from the measurement can be as much as 90% higher than that inferred if the currentvoltage scan proceeds from reverse to forward bias. The size of the difference depends on the scan rates, the preconditioning voltage and time and the light level. Recent studies indicate that there is no direct inuence of light on the hysteresis in the photocurrent15,17. Light may have an indirect effect on hysteresis by generating a photovoltage that inuences device performance in a similar way to the application of an external bias. The effect of sweeping voltage with time can be decoupled from the hysteresis phenomenon by making stepped chronophotoampereometry measurements where the relaxation of the photocurrent towards a steady-state value at short circuit is observed following application of a constant bias in the dark.

By measuring the temperature- and time-dependent photo-current following forward and reverse biasing in the dark, the rate at which the cell relaxes to equilibrium can be determined and thus activation energies for the relaxation of the device can be estimated (see Fig. 4a,b). The temperature range covered in this study is representative of typical device-operating temperatures

Pb I plane

Vl Pb

Pb Pb

VPb

I I

MA-I plane

I I I

VMA

Figure 2 | Transport mechanisms in the CH3NH3PbI3 perovskite structure. Schematic illustration of the three ionic transport mechanisms involving conventional vacancy hopping between neighbouring positions: (a) I migration along an octahedron edge; Pb2 migration along the diagonal direction o1104; (b) CH3NH3 migration into a neighbouring vacant A-site cage involving motion normal to the unit cell face composed of four iodide ions.

Table 1 | Calculated activation energies for ionic migration in CH3NH3PbI3.

Migrating vacancy Defect notation EA (eV) I V I 0.58

Pb2 V00Pb 2.31 CH3NH3 (MA) V0MA 0.84

The migration of ion species is mediated by vacancy defects.

NATURE COMMUNICATIONS | 6:7497 | DOI: 10.1038/ncomms8497 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8497

Dark 0.5 V reverse bias

Dark 1V forward bias

Current density (mA cm2 )

1 Sun short circuit

0 100 200 300 400 Time (s)

EA= 0.62 eV

EA= 0.60 0.68 eV

6 4

Relative energy (eV)

0.6

0.5

0.4

0.3

0.2

0.1

0.0

0 1 2 3 4 Migration distance ()

Rate constant (s1 )

0.1

0.01

6 4

3.2 3.4 3.6 3.8

1,000/Temperature (K1)

Figure 3 | Iodide ion vacancy migration from density functional theory calculations. (a) Calculated migration path indicating a slightly curved path and local relaxation/tilting of the octahedra. (b) Corresponding energy prole.

but below the second-order tetragonal-to-cubic phase transition. Following dark reverse bias conditions, the photocurrent rise at short circuit was tted well by a bi-exponential. The time constant for the initial part of the photocurrent rise after reverse biasing varied between 53 s at 9.5 C and 0.36 s at 50 C. The

time constant for the slower phase of the rise was six times longer, with values very similar to those obtained by tting to single exponential functions to the tail of the photocurrent decay following forward biasing in the dark (Fig. 4a). The shape of the photocurrent relaxations remained relatively constant but scaled with time according to the temperature, implying that only the kinetics of the relaxation process had changed. The activation energies for short-circuit photocurrent relaxation from both forward (1 V) and reverse bias ( 0.5 V) preconditioning were

similar: 0.68 eV and 0.600.62 eV, respectively. This similarity, combined with the observation that the slow phases of both the rise and fall of the photocurrent have approximately the same time constants for a given temperature, suggests that the underlying process controlling the relaxation is reversible.

The measured activation energies for hysteresis, 0.600.68 eV, are very close to the value calculated for the migration of iodide ion vacancies, 0.58 eV. Our measured and calculated values are also compatible with experimental ionic conductivity studies of inorganic perovskites, which have reported similar activation energies for halide-ion vacancy conduction in perovskite halides (for example, CsPbCl3)3537 and oxide-ion vacancy transport in perovskite oxides33,34. In contrast, the activation energies

calculated for the migration of MA and Pb2 vacancies (0.84 and 2.31 eV, respectively) are higher than the measured values, conrming that signicant migration of these ions is unlikely at the temperatures and timescales examined.

The diffusion length of an iodide vacancy on the timescale of a photocurrent relaxation at room temperature (around 10 s) using our estimated diffusion coefcient (10 12 cm2 s 1) is of the order of 30 nm. The corresponding drift distance in this time due to an

Figure 4 | Chronophotoamperometry measurements of a perovskite-based cell. (a) The measurement sequence in a d-TiO2/CH3NH3PbI3/

spiro-OMeTAD/Au cell is indicated; measured temperatures (to the nearest 0.5 C) of the devices were 9.5 (dark blue), 5.5, 0.5, 5, 10.5,

15, 19.5, 24.5, 30, 40 and 50 C (dark red). The dark current under forward bias was very sensitive to uctuations in the controlled temperature; no time constants were taken from the dark current. (b) Arrhenius plot of the rates of photocurrent relaxation. Fits (purple lines) to the fast (k1) and slow (k2) components of bi-exponential to the photocurrent rise at 1 sun equivalent light intensity following reverse bias at 0.5 V in the dark (open

circles and squares, respectively). The activation energy of the fast component evaluated between 9.5 and 50 C was EA 0.62 eV. The

activation energy for the slow component evaluated between 15 and 50 C was EA 0.60 eV (the measurement duration was insufcient to reliably

estimate the slow component at lower temperatures). The photocurrent rise at 50 C did not reach a stable plateau and started to decline after its peak; we did not t to this portion of the curve. The red crosses show the rates inferred from a single exponential t to the tail of the photocurrent decay (k3) following forward bias at 1 V in the dark; the corresponding activation energy is 0.68 eV (red line) evaluated between 15 and 50 C. Given the spread of points, we consider that the range of activation energies determined is similar to the uncertainty of the estimation.

4 NATURE COMMUNICATIONS | 6:7497 | DOI: 10.1038/ncomms8497 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8497 ARTICLE

electric eld that might be found in a device of this thickness (1 V dropped uniformly across 500 nm) is around 80 nm. As an example, the change in potential across a layer of accumulated vacancies of thickness w of 10 nm yields a value of E0.6 V, using the relation qw2N/(2e0er) assuming a relative permittivity (er)

of 24 and 0.4% Schottky defects (NE1.6 1019 cm 3) at room

temperature41. Hence, a change of this magnitude in electric potential in the perovskite adjacent to electrodes would certainly result in a signicant variation in the observed photocurrent in a solar cell.

Migration of iodide ion vacancies under the inuence of an electric eld could change the photogenerated charge collection efciency of devices with time and so helps to explain hysteresis. The possible inuence of iodide ion vacancies on band energies of a perovskite thin lm device and its interfaces is described in Figure 5. We suggest the following model to explain the chronophotoamperometry measurements (see Fig. 5b). When the cell is short-circuited, a built-in electric eld is present in the perovskite layer due to the difference in the work functions of the contacts. This could result in the migration of iodide ion vacancies towards the contacts (while immobile vacancies remain xed), partially screening the eld. As discussed above, we expect the magnitude of the short-circuit photocurrent to be controlled

by the extent to which the electric eld is screened (as also recently suggested by Tress et al.17). A reduced internal eld within the device (due to ionic screening) would lead to less efcient collection of photogenerated charge carriers. Prolonged poling of the device under forward bias reduces (or reverses) the built-in eld, which could allow a dissipation of ionic charge from the contacts by diffusion. This tempering would result in more efcient collection of photogenerated charges when the device is returned to short circuit, since the built-in eld is no longer screened by accumulated ionic or vacancy charge. Conversely, holding the device under reverse bias would accentuate the migration of ions and vacancies to the interfaces, resulting in less efcient collection of photogenerated charges when the device is returned to short circuit.

Within this model, the rate at which the short-circuit photocurrent relaxes towards equilibrium following poling, under either forward or reverse bias, gives an indication of the rate of ion diffusion. The temperature dependence of iodide ion migration would then correspond to the measured activation energy. The agreement between the measured and predicted activation energies supports ionic transport as one of the primary causes of the anomalous hysteresis in currentvoltage curves1317, and the giant switchable photovoltaic effect32.

Built-in electric field

Built-in field screened in bulk

qVbi

Band energy

+ +

I vacancy drift I redistributed MA rotation MA aligned perovskite perovskite

p p n

Dark reverse bias

[afii9829]+

[afii9829]

[afii9829]+

[afii9829]

[afii9829]+

[afii9829]

current

Light short circuit

Dark forward bias

Light short circuit

time

[afii9829]+

[afii9829]+ [afii9829]+

[afii9829]+

[afii9829]

[afii9829][afii9829]

[afii9829]+

[afii9829]

Figure 5 | Inuence of iodide ion vacancies on band energies of a perovskite thin lm. (a) Schematic diagrams indicating the inuence of vacancy drift on the band energies of a p-i-n device at short circuit. EC is the conduction band energy, EV is the valence band energy and Vbi is the built-in potential. Iodide ion vacancies are represented by the squares with plus signs. Implicit in the diagram is that the vacancies with effective positive charges are balanced by immobile cation vacancies (not shown) with effective negative charges. (b) Hypothesized energy level congurations corresponding to different bias conditions and times during the chronophotoamperometry measurements. The variation in the conduction and valence bands corresponds to the redistribution of iodide ion vacancies to and from interfaces with different applied potentials and times.

NATURE COMMUNICATIONS | 6:7497 | DOI: 10.1038/ncomms8497 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8497

The detailed behaviour may be more complex owing to additional contributions from, for example, electronic charge redistribution (at faster timescales). Although this investigation is not exhaustive in terms of possible polarization phenomena and direct observation of ion migration, it does highlight a crucial area for further work on this fascinating system. Indeed, large-scale molecular dynamics simulations could be used to explore diffusion mechanisms and defect association effects, while isotope diffusion measurements would provide a direct probe for determining the extent of ion transport in these materials.

DiscussionOur combined computational and experimental study has provided new atomic-scale insights into the ion transport mechanisms operating in the hybrid perovskite CH3NH3PbI3.

We nd facile vacancy-assisted diffusion of iodide ions, with good agreement with activation energies derived from kinetic measurements on a CH3NH3PbI3-based solar cell. A key factor inuencing the conduction of iodide ions (as the majority ionic carriers) will be the levels of intrinsic iodide ion vacancies in the material, which will be sensitive to the fabrication conditions and sample thermal history.

Our study provides a framework for understanding ionic transport phenomena in organicinorganic halide perovskites, including the inuence of the migration of iodide ion vacancies to and from interfaces in solar cell devices. Such ion migration has been suggested as a factor contributing to their unusual behaviour, including currentvoltage hysteresis and a giant dielectric response at low frequencies. The picture emerging is that hybrid halide perovskites are mixed ionicelectronic conductors. This behaviour has major repercussions for interpreting solar cell device performance and degradation pathways, and for the design of future architectures.

Methods

Computational. The theoretical framework for modelling ion transport in the solid state is well developed53, and has been extensively validated for perovskite oxides52,54. The CH3NH3PbI3 perovskite structure and energies were calculated using density functional theory methods (employing the ab initio code VASP58). A 4 4 4 supercell (768 atoms) of the pseudo-cubic unit cell was modelled; a plane

wave cutoff energy of 500 eV, k-point sampling at the gamma point, PAW pseudopotentials59 and the PBEsol exchange-correlation functional were employed. For structure relaxation, forces were converged to o0.01 eV 1. For the hybrid perovskites, PBEsol reproduces the experimental crystal structure in good agreement with neutron diffraction data, and also describes the nite temperature behaviour (lattice vibrations and dynamics) in agreement with Raman and neutron scattering. We obtained an average lattice parameter of 6.28 for CH3NH3PbI3, in good agreement with the observed values of 6.286.32 from X-ray and neutron diffraction experiments (Table 2)11,55,60,61. The defect results using PBEsol are very similar to those using other functionals, for example, see discussions of LDA/PBE in ref. 50. For migration barriers, the expanded lattice volume for standard GGA functionals (such as PBE) could be a major source of error, which PBEsol avoids.

Activation energies for diffusion processes can be computed from the total energy difference between the diffusing species in their ground-state conguration

and at the saddle point of the diffusion process. Migration mediated by ion vacancies was examined using nudged elastic band and constrained energy minimization methods58; in the latter, the migrating species is propagated along the migration direction in a series of small steps with all unconstrained degrees of freedom relaxed at each step. While in a cubic lattice with disordered CH3NH3 ions, each iodine site is equivalent, for an ordered cell that we used there are distinct apical and equatorial positions. We previously conrmed41 that the vacancy formation energy is similar on both sites; here, we performed nudged elastic band calculations involving each iodine position, and nd that the resulting activation energies are within 0.01 eV, and also insensitive to the alignment of methylammonium ions.

Solar cell fabrication. Methylammonium lead iodide solar cells were fabricated using a modication of standard recipes5. Fluorine-doped tin oxide (F:SnO2)-coated glass (TEC 15, Hartford Glass Co. Inc.) was patterned by etching with zinc powder and 2 M HCl, cleaned with deionized H2O, isoproponal and heated at 450 C for 30 min. A dense titanium dioxide (B80 nm) blocking layer (d-TiO2)

was deposited on SnO2:F glass substrates via spray pyrolysis and heated at 450 C for 30 min62. A solution of 2.64 M methylammonium iodide and 0.88 M lead chloride in dry DMF was spun at 2,000 r.p.m. (acceleration 1,000 r.p.m. s 1) onto the substrate and dried at room temperature for 30 min before heating at 90 C for 3 h in a glove box. Studies suggest that negligible chloride remains in the lm following this preparation. The resulting CH3NH3PbI3 perovskite lm thickness was B500 nm. A solution of 61.6 mM spiro-OMeTAD, 55 mM tBP and 26 mM LiTFSI in dry chlorobenzene was then spun onto the perovskite, as above, to create the hole-transporting layer. A gold electrode was evaporated onto this with an active area of 0.08 cm2. The cell was sealed in the glove box using a glass coverslip and a surlyn gasket.

Kinetic measurements. Chronophotoampereometry measurements were made using the TRACER system15,63. The current and voltage were measured while bias light (from white light-emitting diodes), applied voltage and short-circuit/open-circuit status were modulated in an arbitrary sequence of steps. The modulation was accomplished with three MOSFET switches possessing o1 ms switching time and synchronicity (for both the optical and electrical steps). The temperature of the cells was controlled to within 1 C using a Peltier cooler/heater to which the devices were coupled using thermal paste. Devices were stabilized for 5 min at each temperature before measurement. Illumination by the light-emitting diodes was also found to result in o1 C temperature increase of the device. The chronophotoampereometry measurement sequence was as follows. The current through the cell was recorded with a reverse bias of 0.5 V for 102 s, the device

was then switched to short-circuit mode (0 V) and the light simultaneously turned on (with 1 sun equivalent absorbed photon ux) for a further 102 s. The light was then turned off and a forward bias of 1 V was simultaneously applied to the cell for 102 s. The cell was then returned to short-circuit mode (0 V) and the light simultaneously switched on for a nal 102 s.

Empirically, the rise in the photocurrent (Jsc) towards equilibrium (Jsc,equilibrium) after reverse bias was found to be well t by a double exponential function:

Jsc t Jsc;

equilibrium

J1 exp k1 t J2 exp k2 t ; 2 where t is the time after switching, k1 and k2 are the rate constants (units: s 1) for the fast and slow phases of the rise and J1 and J2 are constants. The decay of the photocurrent towards equilibrium was observed to have a more complicated functional form with a shoulder in the decay at early times. The tail of this decay could be well t by a single exponential function:

Jsc t Jsc;

equilibrium

J3 exp k3 t ; 3 where k3 is the rate constant for the decay and J3 is a constant. As discussed in the main text, we hypothesize that these phenomenological rates of photocurrent relaxation (k1, k2 and k3) are related to the formation or removal of energetic barriers in the material, which change the efciency of photogenerated charge carrier collection17.

Table 2 | Structural parameters of CH3NH3PbI3.

Structural parameter () DFT (this study) X-ray60 diffraction X-ray11 diffraction Neutron55 diffraction X-ray61 diffraction Cubic lattice parameter 6.28 6.28 6.31 6.32 6.29

II separation 4.45 4.46 4.47 PbI bond length 3.16 3.16 3.16 CN bond length 1.48 1.48 1.35* CH bond length 1.10 0.99* NH bond length 1.04 0.99*

DFT, density functional theory.

Calculated (DFT) parameters are compared with experimental data for the cubic phase.*Due to the manner in which orientational disorder is tted to neutron diffraction data for the cubic system, these bond lengths from ref. 55 represent an underestimate; in the renement of the orthorhombic structure they used bond lengths with soft constraints of 1.46 (CN), 1.13 (CH) and 1.00 (NH), which are comparable to our calculated values.

6 NATURE COMMUNICATIONS | 6:7497 | DOI: 10.1038/ncomms8497 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8497 ARTICLE

The barriers result from the motion of ionic charge to or from the interfaces following a change in the electric eld across the device caused by switching the bias voltage state. Measurements were taken as a function of temperature (T) from 263 to 323 K, representative of device-operating temperatures but below the tetragonalto-cubic phase transition. Associated activation energies (EA) were determined by tting the expression ln[k] CEA/(kB T) to an Arrhenius plot of the rate (k) data,

where C is a constant and kB is Boltzmanns constant. Although the activation energies are derived from the rates of photocurrent relaxation, they are likely to be related to the underlying mechanism that causes the change in photocurrent.

References

1. Green, M. A., Ho-Baillie, A. & Snaith, H. J. The emergence of perovskite solar cells. Nat. Photon. 8, 506514 (2014).

2. Graetzel, M. The light and shade of perovskite solar cells. Nat. Mater. 13, 838842 (2014).

3. Kojima, A., Teshima, K., Shirai, Y. & Miyasaka, T. Organometal halide perovskites as visible-light sensitizers for photovoltaic cells. J. Am. Chem. Soc. 131, 60506051 (2009).

4. Im, J.-H., Lee, C.-R., Lee, J.-W., Park, S.-W. & Park, N.-G. 6.5% efcient perovskite quantum-dot-sensitized solar cell. Nanoscale 3, 40884093 (2011).

5. Lee, M. M., Teuscher, J., Miyasaka, T., Murakami, T. N. & Snaith, H. J. Efcient hybrid solar cells based on meso-superstructured organometal halide perovskites. Science 338, 643647 (2012).

6. Kim, H.-S. et al. Lead iodide perovskite sensitized all-solid-state submicron thin lm mesoscopic solar cell with efciency exceeding 9%. Sci. Rep. 2, 591 (2012).

7. Burschka, J. et al. Sequential deposition as a route to high-performance perovskite-sensitized solar cells. Nature 499, 316319 (2013).

8. Liu, M., Johnston, M. B. & Snaith, H. J. Efcient planar heterojunction perovskite solar cells by vapour deposition. Nature 501, 395398 (2013).

9. Heo, J. H. et al. Efcient inorganicorganic hybrid heterojunction solar cells containing perovskite compound and polymeric hole conductors. Nat. Photon. 7, 486491 (2013).

10. Xing, G. et al. Long-range balanced electron- and hole-transport length in organicinorganic CH3NH3PbI3. Science 342, 344347 (2013).

11. Stoumpos, C. C., Malliakas, C. D. & Kanatzidis, M. G. Semiconducting tin and lead iodide perovskites with organic cations: phase transitions, high mobilities, and near-infrared photoluminescent properties. Inorg. Chem. 52, 90199038 (2013).

12. Stranks, S. D. et al. Electronhole diffusion lengths exceeding 1 micrometer in an organometal trihalide perovskite absorber. Science 342, 341344 (2013).13. Snaith, H. J. et al. Anomalous hysteresis in perovskite solar cells. J. Phys. Chem. Lett. 5, 15111515 (2014).

14. Unger, E. L. et al. Hysteresis and transient behavior in current-voltage measurements of hybrid-perovskite absorber solar cells. Energy Environ. Sci. 7, 36903698 (2014).

15. ORegan, B. C. et al. Opto-electronic studies of methylammonium lead iodide perovskite solar cells with mesoporous TiO2: separation of electronic and chemical charge storage, understanding two recombination lifetimes, andthe evolution of band offsets during JV hysteresis. J. Am. Chem. Soc 137, 50875099 (2015).

16. Kim, H. S. & Park, N. G. Parameters affecting I-V hysteresis of CH3NH3PbI3 perovskite solar cells: effects of perovskite crystal size and mesoporous TiO2 layer. J. Phys. Chem. Lett. 5, 29272934 (2014).

17. Tress, W. et al. Understanding the rate-dependent JV hysteresis, slow time component, and aging in CH3NH3PbI3 perovskite solar cells: the role of a compensated electric eld. Energy Environ. Sci. 8, 9951004 (2015).

18. Pellet, N. et al. Mixed-organic-cation perovskite photovoltaics for enhanced solar-light harvesting. Angew. Chem. Int. Ed. 53, 31513157 (2014).19. Malinkiewicz, O. et al. Perovskite solar cells employing organic charge-transport layers. Nat. Photon. 8, 128132 (2014).

20. Zhou, H. P. et al. Interface engineering of highly efcient perovskite solar cells. Science 345, 542546 (2014).

21. Wang, J. T-W. et al. Low-temperature processed electron collection layers of graphene/TiO2 nanocomposites in thin lm perovskite solar cells. Nano Lett.

14, 724730 (2014).22. Xing, G. et al. Low-temperature solution-processed wavelength-tunable perovskites for lasing. Nat. Mater. 13, 476480 (2014).

23. Leguy, A. M. A. et al. The dynamics of methyl ammonium ions in hybrid organic-inorganic perovskite solar cells. Nat. Commun. (2015doi:http://dx.doi.org/10.1038/ncomms8124

Web End =10.1038/ http://dx.doi.org/10.1038/ncomms8124

Web End =ncomms8124 .

24. Jeon, N. J. et al. Solvent engineering for high-performance inorganicorganic hybrid perovskite solar cells. Nat. Mater. 13, 897903 (2014).

25. Liu, D. & Kelly, T. L. Perovskite solar cells with a planar heterojunction structure prepared using room-temperature solution processing techniques. Nat. Photon. 8, 133138 (2014).

26. Bergmann, V. et al. Real-space observation of unbalanced charge distribution inside a perovskite-sensitized solar cell. Nat. Commun. 5, 5001 (2014).

27. DInnocenzo, V. et al. Excitons versus free charges in organo-lead tri-halide perovskites. Nat. Commun. 5, 3586 (2014).

28. Edri, E. et al. Elucidating the charge carrier separation and working mechanism of CH3NH3PbI3-xClx perovskite solar cells. Nat. Commun. 5, 3461 (2014).

29. Chen, Q. et al. Planar heterojunction perovskite solar cells via vapor-assisted solution process. J. Am. Chem. Soc. 136, 622625 (2014).

30. Marchioro, A. et al. Unravelling the mechanism of photoinduced charge transfer processes in lead iodide perovskite solar cells. Nat. Photon. 8, 250255 (2014).

31. Dualeh, A. et al. Impedance spectroscopic analysis of lead iodide perovskitesensitized solid-state solar cells. ACS Nano 8, 362373 (2014).

32. Xiao, Z. et al. Giant switchable photovoltaic effect in organometal trihalide perovskite devices. Nat. Mater. 14, 193198 (2015).

33. Malavasi, L., Fisher, C. A. J. & Islam, M. S. Oxide-ion and proton conducting electrolyte materials for clean energy applications: structural and mechanistic features. Chem. Soc. Rev. 39, 43704387 (2010).

34. Goodenough, J. B. Electronic and ionic transport properties and other physical aspects of perovskites. Rep. Prog. Phys. 67, 19151993 (2004).

35. Mizusaki, J., Arai, K. & Fueki, K. Ionic conduction of the perovskite-type halides. Solid State Ionics 11, 203211 (1983).

36. Narayan, R. L. & Suryanarayana, S. V. Transport properties of the perovskite-type halides. Mater. Lett. 11, 305308 (1991).

37. Kuku, T. A. Ionic transport and galvanic cell discharge characteristics of CuPbI3. Thin Solid Films 325, 246250 (1998).

38. Fan, Z. et al. Ferroelectricity of CH3NH3PbI3 perovskite. J. Phys. Chem. Lett. 6, 11551161 (2015).

39. Frost, J. M. et al. Atomistic origins of high-performance in hybrid halide perovskite solar cells. Nano Lett. 14, 25842590 (2014).

40. Brivio, F., Walker, A. B. & Walsh, A. Structural and electronic properties of hybrid perovskites for high-efciency thin-lm photovoltaics from rst-principles. APL Mater. 1, 042111 (2013).

41. Walsh, A., Scanlon, D. O., Chen, S., Gong, X. G. & Wei, S. H. Self-regulation mechanism for charged point defects in hybrid halide perovskites. Angew. Chem. Int. Ed. 54, 17911794 (2015).

42. Mosconi, E., Amat, A., Nazeeruddin, M. K., Graetzel, M. & De Angelis, F. First-principles modeling of mixed halide organometal perovskites for photovoltaic applications. J. Phys. Chem. C 117, 1390213913 (2013).

43. Even, J., Pedesseau, L., Jancu, J. M. & Katan, C. Importance of spin-orbit coupling in hybrid organic/inorganic perovskites for photovoltaic applications.J. Phys. Chem. Lett. 4, 29993005 (2013).44. Giorgi, G., Fujisawa, J., Segawa, H. & Yamashita, K. Small photocarrier effective masses featuring ambipolar transport in methylammonium lead iodide perovskite: a density functional analysis. J. Phys. Chem. Lett. 4, 42134216 (2013).

45. Yin, W.-J., Shi, T. & Yan, Y. Unusual defect physics in CH3NH3PbI3 perovskite solar cell absorber. Appl. Phys. Lett. 104, 063903 (2014).

46. Kim, J., Lee, S.-H., Lee, J. H. & Hong, K.-H. The role of intrinsic defects in methylammonium lead iodide perovskite. J. Phys. Chem. Lett. 5, 13121317 (2014).

47. Umari, P., Mosconi, E. & De Angelis, F. Relativistic GW calculations on CH3NH3PbI3 and CH3NH3SnI3 perovskites for solar cell applications. Sci. Rep.

4, 4467 (2014).48. Du, M. H. Efcient carrier transport in halide perovskites: theoretical perspectives. J. Mater. Chem. A 2, 90919098 (2014).

49. Agiorgousis, M.L., M.L., Sun, Y. Y., Zeng, H. & Zhang, S. Strong covalency-induced recombination centers in perovskite solar cell material CH3NH3PbI3.

J. Am. Chem. Soc. 136, 1457014575 (2014).50. Yin, W. J., Yang, J. H., Kang, J., Yan, Y. & Wei, S. H. Halide perovskite materials for solar cells: a theoretical review. J. Mater. Chem. A 3, 89268942 (2015).

51. Egger, D. A. & Kronik, L. Role of dispersive interactions in determining structural properties of organic-inorganic halide perovskites: insights from rst-principles calculations. J. Phys. Chem. Lett. 5, 27282733 (2014).

52. Islam, M. S. Ionic transport in ABO3 perovskite oxides: a computer modelling tour. J. Mater. Chem. 10, 10271038 (2000).

53. Catlow, C. R. A. Computer simulation studies of transport in solids. Ann. Rev. Mater. Sci. 16, 517548 (1986).

54. Chroneos, A., Yildiz, B., Tarancon, A., Partt, D. & Kilner, J. A. Oxygen diffusion in solid oxide fuel cell cathode and electrolyte materials: mechanistic insights from atomistic simulations. Environ. Energy. Sci. 4, 27742789 (2011).

55. Weller, M. T., Weber, O. J., Henry, P. F., Di Pumpo, A. M. & Hansen, T. C. Complete structure and cation orientation in the perovskite photovoltaic methylammonium lead iodide between 100 and 352K. Chem. Commun. 51, 41804183 (2015).

56. Schoonman, J. Organic-inorganic lead halide peroyskite solar cell materials: a possible stability problem. Chem. Phys. Lett. 619, 193195 (2015).

57. Yashima, M. et al. Conduction path and disorder in the fast oxide-ion conductor (La0.8Sr0.2)(Ga0.8Mg0.15Co0.05)O2.8. Chem. Phys. Lett. 380, 391396 (2003).

58. Kresse, G. & Furthmller, J. Efcient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B 54, 1116911186 (1996).

NATURE COMMUNICATIONS | 6:7497 | DOI: 10.1038/ncomms8497 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8497

59. Perdew, J. P. et al. Restoring the density-gradient expansion for exchange in solids and surfaces. Phys. Rev. Lett. 100, 136406 (2008).

60. Baikie, T. et al. Synthesis and crystal chemistry of the hybrid perovskite (CH3NH3) PbI3 for solid-state sensitised solar cell applications. J. Mater. Chem.

A 1, 56285641 (2013).61. Kawamura, Y., Mashiyama, H. & Hasebe, K. Structural study on cubic-tetragonal transition of CH3NH3PbI3. J. Phys. Soc. Jpn 71, 16941697 (2002).

62. Xu, W. W., Kershaw, R., Dwight, K. & Wold, A. Preparation and characterization of TiO2 lms by a novel spray pyrolysis method. Mater. Res.

Bull. 25, 13851392 (1990).63. Barnes, P. R. F. et al. Interpretation of optoelectronic transient and charge extraction measurements in dye-sensitized solar cells. Adv. Mater. 25, 18811922 (2013).

Acknowledgements

We gratefully acknowledge support from the EPSRC for the Energy Materials Programme grant (EP/K016288), Archer HPC facilities through the Materials Chemistry Consortium (EP/L000202), research grant EP/M014797/1 and P.R.F.B.s fellowship (EP/J002305/1). We thank C.H. Law for his assistance with device preparation, andP. Cameron, C.A.J. Fisher, L.M. Peter and M.T. Weller for useful discussions.

Author contributions

M.S.I. and A.W. conceived and supervised the computational project. P.R.F.B. and B.C.OR. conceived and supervised the kinetic experiments. C.E. and J.M.F. performed the DFT calculations. P.R.F.B. performed the experiments. M.S.I. drafted the rst version of the paper. All authors contributed to the analysis and writing of the nal version of the paper.

Additional information

Competing nancial interests: The authors declare no competing nancial interests.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

How to cite this article: Eames, C. et al. Ionic transport in hybrid lead iodide perovskite solar cells. Nat. Commun. 6:7497 doi: 10.1038/ncomms8497 (2015).

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

8 NATURE COMMUNICATIONS | 6:7497 | DOI: 10.1038/ncomms8497 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

Word count: 6842

Show less

Abstract

Translate

Solar cells based on organic-inorganic halide perovskites have recently shown rapidly rising power conversion efficiencies, but exhibit unusual behaviour such as current-voltage hysteresis and a low-frequency giant dielectric response. Ionic transport has been suggested to be an important factor contributing to these effects; however, the chemical origin of this transport and the mobile species are unclear. Here, the activation energies for ionic migration in methylammonium lead iodide (CH₃ NH₃ PbI₃ ) are derived from first principles, and are compared with kinetic data extracted from the current-voltage response of a perovskite-based solar cell. We identify the microscopic transport mechanisms, and find facile vacancy-assisted migration of iodide ions with an activation energy of 0.6 eV, in good agreement with the kinetic measurements. The results of this combined computational and experimental study suggest that hybrid halide perovskites are mixed ionic-electronic conductors, a finding that has major implications for solar cell device architectures.

Details

Title

Ionic transport in hybrid lead iodide perovskite solar cells

Author

Eames, Christopher; Frost, Jarvist M; Barnes, Piers R F; O'regan, Brian C; Walsh, Aron; Islam, M Saiful

Pages

7497

Publication year

2015

Publication date

Jun 2015

Publisher

Nature Publishing Group

e-ISSN

20411723

Source type

Scholarly Journal

Language of publication

English

DOI

https://doi.org/10.1038/ncomms8497

ProQuest document ID

1690661953

Ionic transport in hybrid lead iodide perovskite solar cells

Jump to:

Full text

Abstract

Details

Suggested sources