DOI: 10.1002/cssc.

201600949 Full Papers

Ionic and Optical Properties of Methylammonium Lead

Iodide Perovskite across the Tetragonal–Cubic Structural
Phase Transition
Md Nadim Ferdous Hoque,[a] Nazifah Islam,[a] Zhen Li,[b] Guofeng Ren,[a] Kai Zhu,*[b] and
Zhaoyang Fan*[a]

Practical hybrid perovskite solar cells (PSCs) must endure tem- whereas the optical properties exhibit no particular transition
peratures above the tetragonal–cubic structural phase transi- except for the steady increase of the bandgap with tempera-
tion of methylammonium lead iodide (MAPbI3). However, the ture. These observations can be explained by the “continuous”
ionic and optical properties of MAPbI3 in such a temperature nature of perovskite phase transition. We speculate that the
range, and particularly, dramatic changes in these properties critical temperature at which the ionic/electrical properties
resulting from a structural phase transition, are not well stud- change, although related to crystal symmetry variation, is not
ied. Herein, we report a striking contrast at approximately necessarily the same temperature as when tetragonal–cubic
45 8C in the ionic/electrical properties of MAPbI3 owing to structural phase transition occurs.
a change of the ion activation energy from 0.7 to 0.5 eV,

Introduction

Organic–inorganic halide perovskites, particularly methylam- of the active photo-absorption layer in a well-encapsulated
monium lead iodide (MAPbI3), are promising for developing solar cell can easily rise up to over 70 8C, particularly consider-
low-cost high-efficiency perovskite solar cells (PSCs). With giant ing the low thermal conductivity of MAPbI3.[3] Compared with
leaps in efficiency of PSCs to over 22 % in just a few years conventional semiconductors used for photovoltaics, the or-
based on a cheap solution-processing method,[1] more atten- ganic–inorganic hybrid perovskites in general, and MAPbI3 in
tion is now focused on understanding the fundamental materi- particular, have a considerably low degradation temperature. It
al properties and device mechanisms to develop PSCs with has been shown that PSC performance is largely dependent
good stability, as well as elimination of the current–voltage on the ambient temperature.[4] Thus, accounting for the tem-
hysteresis phenomenon, for practical applications. In terms of perature dependence of perovskites is a major issue now
environmental factors that impact PSC stability, it is well- along with other stability-related performance effects.[5]
known that humidity and possibly oxygen have detrimental ef- The temperature-dependent performance becomes more
fects by promoting decomposition of perovskite active materi- complex considering that the MAPbI3 perovskite could under-
al.[2] Device degradation through this path could be prevented go a structural phase transition from tetragonal to cubic phase
by suitable encapsulation techniques. However, the influence in the practical operating temperature range of PSCs. There-
of another environmental factor, temperature, on the stability fore, understanding how the electrical and optical properties
of perovskite layers and PSC performance will be more intrin- of the perovskite material itself change across this phase tran-
sic, as PSCs often work at a temperature above the room tem- sition will provide useful information for practical PSC develop-
perature when under solar irradiation. In fact, the temperature ment, as well as enriching our fundamental understanding on
this material system. Particularly, ionic migration[6] in MAPbI3
[a] M. N. F. Hoque, N. Islam, G. Ren, Z. Fan has been regarded as a serious issue that is related to the hys-
Department of Electrical and Computer Engineering and Nano Tech Center teresis problem in PSCs, while the thermally activated ionic
Texas Tech University transportation process will become significant with tempera-
Lubbock, Texas, 79409 (USA)
ture. The activation energy, a critical parameter of ion conduc-
E-mail: [email protected]
tion, has been reported covering a wide range from approxi-
[b] Z. Li, K. Zhu
Chemistry and Nanoscience Center mately 0.2 eV to more than 0.6 eV, which suggests a lack of
National Renewable Energy Laboratory general consensus on the energy barrier determining ion trans-
Golden, Colorado 80401 (USA) port.[6a] Related to the structural phase transition, another in-
E-mail: [email protected]
teresting question arises: does the energy barrier/mechanics
Supporting Information for this article can be found under:
for ion diffusion change subsequently? Therefore, studying the
http://dx.doi.org/10.1002/cssc.201600949.
ionic properties of MAPbI3 in this high-temperature range, and
This publication is part of a Special Issue focusing on the “Stability of
Perovskite Solar Cells & Devices”. A link to the issue’s Table of Contents especially across the phase transition boundary, is necessary.
will appear here once it is complete.

ChemSusChem 2016, 9, 1 – 8 1  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

These are not the final page numbers! ÞÞ

 Full Papers
It is well-known that at room temperature MAPbI3 crystalli- formation. As presented by the Nyquist plot in Figure 1 a,
zes in the tetragonal phase with the neighboring PbI6 octahe- a clear change of impedance with temperature is observed,
dra rotating alternatively around the c-axis. With increasing and all the spectra exhibit two distinct charge-transport re-
temperature, this rotation angle decreases monotonically until gimes, corresponding to two semicircles whose radiuses
it becomes zero, and simultaneously the lattice constants of become smaller as temperature increases. Particularly, the
c and a becomes the same, resulting in a high-temperature second semicircle in the low frequency range becomes more
cubic structure.[7] This tetragonal–cubic phase transition tem- apparent at higher temperatures. It starts forming at 7 Hz for
perature (Tc) of MAPbI3 has been reported with a broad range 15 8C and increases to 403 Hz for 70 8C. The starting frequen-
by various groups measured by different characterization tech- cies at different temperatures were extracted and plotted in
niques.[4c] Temperature-dependent X-ray diffraction (XRD) anal- Figure S1 in the Supporting Information. Two distinct regions
ysis of the crystal structure yielded a Tc range of 54–60 8C.[8] are found, with a transition temperature of approximately
However, a larger temperature range of 37–57 8C has been re- 45 8C.
ported from various optical and electrical measurements.[4c, 6b, 9] In the literature, the first semicircle close to the origin has
It is unclear whether the disparity of the reported Tc is a result been ascribed to be predominately free-electron transporta-
of different material quality, measurement method, or more tion; however, different mechanisms, related to the hysteresis
fundamental reasons. Whether and how the structural phase of PSCs, have been proposed to explain formation of the low
transition will cause a salient contrast in their ionic/electrical frequency component. Some studies attribute this second
and optical properties is unclear. Such a question is interesting semicircle to the double-layer capacitance arising from elec-
considering that the structural phase transition of MAPbI3 per- tron trapping and accumulation at the interface,[12] while other
ovskite has a “continuous” characteristic, and particularly for studies[13, 14] have proposed charge accumulation by slow ionic
the high temperature cubic phase, the local atomic structure is migration to the contact. However, as pointed out by Bag
on average significantly distorted from the ideal cubic struc- et al.,[9e] in these previous models, a critical feature, the linear
ture on the sub-ps time scale.[10] region between the two semicircles in EIS spectra, was not ad-
In this work, we studied temperature-dependent electro- dressed. Considering this obvious feature in Figure 1 a, the
chemical impedance spectroscopy (EIS) and capacitance spec- charge accumulation could be ascribed to ionic diffusion (War-
troscopy for understanding the ionic/electrical properties of burg diffusion) as well as ionic migration.
MAPbI3 together with vis-NIR absorption spectroscopy to As discussed in the Supporting Information (Figure S2), the
follow the optical properties, with a particular emphasis on the Warburg diffusion component in the impedance spectrum
phase transition. By considering the Warburg diffusion, which comes from the mass transfer of ionic diffusion. The frequen-
has been largely neglected as noted by Bag et al.,[9e] we found cy-dependent Warburg impedance has the same amplitude for
two different activation energies (i.e., 0.7 and 0.5 eV) for ion its resistance and capacitive components (Equation S1), result-
diffusion at low (< 45 8C) and high (> 45 8C) temperature ing in the characteristic 458 linear Warburg diffusion region in
ranges, respectively. Temperature-dependent ion conductivity the impedance spectra. Such a Warburg diffusion feature is fre-
and ion diffusion coefficients as well as ion mobility in the dark quently observed in liquid electrolytes resulting from a non-
were reported, with values at room temperature of 2.4  uniform ionic concentration, while in ionic and mixed conduc-
109 S cm1, 2.0  109 cm2 s1, and 7.6  107 cm2 V1 s1, respec- tive solids,[15] a number of examples have also been report-
tively. The ion hopping rate was estimated to be around 8 s1. ed.[16] Considering the obvious Warburg diffusion feature in the
All ionic and related electrical studies unambiguously indicate impedance spectra (Figure 1 a), here we incorporate this diffu-
a property transition at approximately 45 8C, associated with sion component in the model to fit the measured EIS spectra,
the change of the ion energy barrier. In contrast, optical stud- with the equivalent circuit[9e] shown in the inset of Figure 1 a.
ies do not show such a clear variation of properties, including In this model, the low frequency components consist of RCT as
absorbance and bandgap, across the structural phase transi- interfacial charge-transfer resistance at contacts, Cdl as double-
tion. These findings are useful for understanding the funda- layer capacitance owing to ionic blocking at the contact, Ws as
mental MAPbI3 perovskite properties, PSC hysteresis, and tem- Warburg element, and Rtr as coupled electron, and ion trans-
perature-related stability. port resistance. In addition, Relectr represents free carrier recom-
bination or transport that is not coupled to ion diffusion/accu-
mulation, Rs represents the series contact resistance, and Cm is
Results and Discussion chemical capacitance of the perovskite layer. We notice that in
the literature, simpler models, without considering the War-
Ionic and electrical properties
burg diffusion, have been used to capture the major features
MAPbI3 thin films were coated on fluorine-doped tin oxide of EIS spectra by disclosing the ionic-migration nature.[14] How-
(FTO) glass substrates, using anti-solvent method in a N2-filled ever, as compared in Figure S3, the fitting accuracy of these
glovebox,[11] with Au deposited as the top contact for ionic/ models can be significantly different, particularly at the War-
electrical studies. EIS was measured from 1 MHz to 0.1 Hz with burg region. The current model with the Warburg diffusion
100 mV AC signal under 0 V DC bias in the dark, and the component fits the measured spectra very well with goodness
sample temperature was changed from 15 to 70 8C. Details of of fit parameter c2 < 0.004 for frequency range of 1 MHz to
the experimental methods can be found in the Supporting In- 1 Hz.

& ChemSusChem 2016, 9, 1 – 8 www.chemsuschem.org 2  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

 Full Papers

Figure 1. (a) The Nyquist plot from impedance spectroscopy of MAPbI3 film measured at temperatures of 15–70 8C in the dark. The inset shows the model
used for fitting the impedance data and the fitting is shown using solid lines. (b) The Arrhenius plot of Warburg time constant tw showing two slopes corre-
sponding to Ea of ion diffusion at low and high temperature ranges. (c) Temperature-dependent ionic conductivity and the derived Ea. (d) Diffusion coefficient
and ionic mobility showing two different trends in their temperature dependence. (e) The Bode plot of EIS showing the shift of the peak frequency. (f) The Ar-
rhenius plot of the ion hopping rate showing similar Ea.

The Warburg element Ws is a function of Warburg time con- Di varies from 6.5  1010 to 5  108 cm2 s1, and mi changes be-
stant tw,[16b] which is related to the activation energy Ea for the tween 2.6  107 and 1.7  105 cm2 V1 s1. Particularly, at room
ionic diffusion by 1=tw ¼ ATeE a =kT [Eq. (S3)]. With the EIS spec- temperature, Di and mi are found to be 2.0  109 cm2 s1 and
tra fitting to obtain Ws and then tw at different temperature T, 7.6  107 cm2 V1 s1, respectively. In the literature, Yang et al.[14]
lnðT  tw Þ1 vs. 1/T is plotted in Figure 1 b. Two obvious regions reported Di as 2.4  108 cm2 s1, while Delugas et al.[17] calculat-
with different activation energies were found based on the ed a mobility of 2.8  105 cm2 V1 s1. The diffusion and mobili-
fitted slope. The fitted curves of the two regions intersect at ty are expected to be affected by the material processing tech-
a temperature of approximately 48 8C. Thus, the ion activation niques that result in different material quality and defect densi-
energy in low-temperature and high-temperature regions is ty.
0.72  0.05 and 0.53  0.02 eV, respectively. The EIS spectra can also be presented in the format of the
Another way to find the ionic activation energy is to study Bode plot, as shown in Figure S5. The two frequency peaks in
the temperature-dependent ion conductivity from Warburg im- the phase–frequency Bode plot clearly indicate two different
pedance Aw.[9e] Aw was fitted using the j Z j vs. w1/2 curve (Fig- transport regimes, especially at high temperature. The lower
ure S4). From Equation (S4), we can get the ionic conductivity frequency response is zoomed in Figure 1 e, from which we
sion for different temperatures. As Figure 1 c shows, from the obtained the low frequency peak (f0), corresponding to the re-
Arrhenius plot of this temperature-dependent ionic conductivi- laxation frequency of ionic transportation in a hopping pro-
ty, two different activation energies were found with values of cess, while the relaxation time constant can be determined by
1
0.71  0.03 and 0.51  0.01 eV, respectively, similar to those de- t ¼ 2pf 0 , as shown in Figure S6, with two different trends of re-
rived from tw. The ionic diffusion coefficient (Di) and ionic mo- sponse in low and high temperature ranges. Finally, the ion
bility (mi) can be further derived using Equation (S5) and (S6) hopping rate constant can be evaluated using the formula,
(Figure 1 d). Clearly two different trends with temperature are k ¼ 1=t as shown in Figure 1 f. Obviously, similar activation en-
found, related to different energy barriers. From 15 8C to 70 8C, ergies of 0.51  0.02 and 0.73  0.04 eV in different regimes

ChemSusChem 2016, 9, 1 – 8 www.chemsuschem.org 3  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim &

These are not the final page numbers! ÞÞ

 Full Papers
were found from relaxation frequency measurement, which is sample was large. The C–f curve in the frequency range of
comparable to activation energy derived from Warburg time 1 MHz to 100 Hz was measured under dark condition, with
constant (Figure 1 b) and conductivity (Figure 1 c). The two 30 mV AC amplitude and no DC bias, to find out the dielectric
fitted lines crossed each other at approximately 43 8C in this properties. The C–f response from 15–70 8C is shown in Fig-
case. ure 2 a where changes in capacitance can be seen with tem-
In the literature, point-defect mediated ion migration in peratures similar to previous studies.[21] The directly measured
MAPbI3 has been ascribed to iodine[14, 18] or MA,[19] with differ- capacitance spectra are in fact in line with those derived from
ent activation energies, calculated or experimentally measured, EIS, which are shown in Figure S8. At low frequency, higher
ranging from 0.1,[17] 0.2,[18b] and 0.4[19] to 0.6 eV.[20] These studies temperature shows higher capacitance, whereas at high fre-
did not report any abrupt change of activation energy at low quency the trend flips as is clear from the inset of Figure 2 a.
and high temperatures. And our activation energy of 0.7 eV in To gain better insight, the dielectric constant (er’) was calculat-
the low-temperature regime is at the high end of the range. ed from the capacitance values at different temperatures as
Bag et al.[9e] also reported activation energies of 0.58 and shown in Figure 2 b. At low frequencies, er ’ increases with tem-
0.23 eV below and above 47 8C, respectively, by fitting EIS perature, whereas at high frequencies the trend is opposite
spectra measured under light, and attributed them to tetrago- but at a much smaller rate. This boundary is at 2  103 Hz.
nal and cubic phases, respectively arisen from a discontinuous Three frequencies, namely 102, 103, and 104 Hz, were chosen as
volume change in the unit cell. Material crystalline quality, examples to illustrate the change of er ’ as a function of tem-
measured methods, and other factor might all have impact on perature (Figure 2 c). Again, a clear distinction can be seen be-
the reported values. We would further mention that Ref. [9e] tween two regions: below 45 8C and above 45 8C. The conduc-
only reported ion conduction under light irradiation, while we tivity (s) spectrum also shows a temperature-dependent re-
observed it also in the dark. It is shown in Figure S7 that even sponse (Figure S9). Similar to the capacitance, the conductivity
under dark conditions, an n-i-p planar PSC shows two different is also higher with increasing temperature at low frequency,
transport regions, with the low frequency region assigned to and the trend becomes opposite at higher frequencies with
ion diffusion. the boundary at ~ 105 Hz. Dielectric loss (er’’) was calculated
Capacitance spectroscopy is another electrical characteriza- from the conductivity curves (Figure 2 b). The dielectric loss
tion that can tell the change in dielectric characteristics across shows 1/f behavior for frequencies below tens of kHz following
the phase transition temperature. Parallel circuit model was the so-called Jonscher’s law, which further reveals the ionic
chosen to measure the capacitance value using a inductance– conductivity of the material.[22] Again, three different frequen-
capacitance–resistance meter, because the impedance of the cies (i.e., 102, 103, and 104 Hz) were chosen to plot the temper-

Figure 2. (a) C–f curve of the MAPbI3 film sandwiched between two contacts measured at different temperatures. The inset shows the zoomed in area of the
C–f curve with a clear transition of capacitance values with temperature at different frequencies. (b) The spectra of dielectric constant and dielectric loss at dif-
ferent temperatures. (c) Temperature dependence of dielectric constant at represented frequencies showing different trends with a boundary at approximate-
ly 45 8C. (d) Temperature dependence of the dielectric loss at represented frequencies showing different trends with a boundary at approximately 47 8C.

& ChemSusChem 2016, 9, 1 – 8 www.chemsuschem.org 4  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

 Full Papers
ature dependence of er’’ (Figure 2 d), and a clear temperature For MAPbI3, the positive temperature coefficient of the bandg-
boundary of ~ 47 8C was observed. Finally, conducting the ap (dEg/dT) was explained by lowering the valence band maxi-
same experiment on a planar PSC cell [FTO/TiO2/MAPbI3/hole- mum with increasing temperature.[9b] Notably, the slope (dEg/
transport material (HTM)/Au] results in similar trend in C–f dT) change in the low (tetragonal phase) and high (cubic
spectra (Figure S10). phase) temperature ranges is not obvious, although a slight
change of slope between 50–60 8C might be discernible. Simi-
lar observation was noticed by Foley et al.[9b] and explained by
Optical properties Quarti[10] noting that the cubic MAPbI3 phase has strong instan-
Absorption spectroscopy was studied to characterize the tem- taneous deviations that define the time-average band gap.
perature-dependent optical properties of MAPbI3 perovskite in The onset and width of absorption edge might also be influ-
the normal solar cell working temperature range and especially enced by exciton binding energy, lattice relaxation, and crystal
across its phase transition. To accurately account for light scat- structure symmetry,[9b, 25] and therefore could be influenced by
tering, a spectrophotometer equipped with an integrating the crystal structural phase transition. Thus, the spectral width
sphere was used for the absorbance measurement (Figure S11), and peak position of the absorption edge of MAPbI3 were ob-
with the calculated absorption coefficient in Figure S12. The tained by calculating the derivative of the absorption coeffi-
Tauc-plot of the absorption coefficient is shown in Figure 3 a cient (da/dE), and the normalized values with respect to that
and S13 for temperature range of 20 to 80 8C, where a clear at 20 8C (Figure 3 c). It is apparent that, as the temperature in-
blue-shift of absorbance spectra edge with increasing tempera- creases, the peak value of da/dE decreases along with peak
ture can be observed. It is interesting to also note that at broadening. A full width at half maximum (FWHM) of
wavelengths below 625 nm, the absorbance at a higher tem- 40.38 meV and a peak position of 1.621 eV were obtained at
perature is slightly stronger than at a lower temperature (Fig- 20 8C, whereas they are 50.29 meV and 1.635 eV at 80 8C, re-
ure S11). spectively. For quantitative analysis, the FWHM values are plot-
The temperature-dependent bandgap of MAPbI3, derived ted against the temperature in Figure 3 d, which again do not
from the linear fitting of the Tauc-plot, is shown in Figure 3 b. show a clear difference in the slope for the tetragonal versus
As the temperature is increased from 20 to 80 8C, the bandgap cubic phase. All these optical data show that the structural
Eg of MAPbI3 changes from 1.595 to 1.605 eV, with a blue-shift change in perovskite from tetragonal to cubic does not have
by 0.01 eV. This shift is reversible as the temperature is lowered very obvious effect on the optical properties.
again as our sample surface was properly protected in the Based on vis-NIR absorption spectroscopy, the tetragonal–
measurement. The bandgap increases with temperature, which cubic structural phase transition of MAPbI3 does not introduce
is in contrast to conventional semiconductors.[23] This blue-shift a salient transition for the temperature-dependent characteris-
phenomenon is not uncommon in Pb-based compounds.[24] tics of optical bandgap and absorption edge, whereas the

Figure 3. (a) Tauc plot of the absorption coefficient showing a blue-shift in the absorption edge with temperature. (b) Temperature-dependent bandgap mea-
sured from the Tauc plot. (c) Differentiated absorption coefficient with energy for different temperatures showing change in both peak position and width.
(d) FWHM of the normalized da/dE at different temperatures. The error bars represent the standard deviation of the Gaussian fitting of the peak of normal-
ized da/dE for calculating the FWHM.

ChemSusChem 2016, 9, 1 – 8 www.chemsuschem.org 5  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim &

These are not the final page numbers! ÞÞ

 Full Papers
changing rates of temperature-dependent ionic-related electri- ture-dependent electrochemical impedance spectroscopy and
cal properties do exhibit different behaviors at low and high independent capacitance spectroscopy. It was discovered that
temperature regions. Both electrical and optical properties do both optical and ionic/electrical properties have no abrupt
not exhibit an abrupt change across the structural phase tran- change across the structural phase transition temperature. The
sition. These observations might be ascribed to the inherently changing rate of optical properties with temperature also has
“soft” nature of the MAPbI3 crystal structures. no abrupt transition. These could be explained by the continu-
In the high-temperature cubic phase, MA + cations in ous nature of the MAPbI3 perovskite crystal symmetry transi-
MAPbI3 are dynamically disordered rotating in an isotropic po- tion and the strong instantaneous deviation from the cubic
tential, whereas in the room-temperature stable tetragonal symmetry at high temperature. However, the changing rate of
phase, MA + cations are partially constrained to specific orien- ionic/electrical properties, with respect to temperature, do
tations by hydrogen bonding to iodine, resulting in the tilting have an abrupt transition across a critical temperature of ap-
of the PbI6 octahedra.[26] The structural phase transition is char- proximately 45 8C, owing to the change of ion activation
acterized by a continuous tilting of the PbI6 octahedra as well energy. It is speculated that the ion energy barrier change is
as the continuously changing of PbI bond lengths.[27] This related to crystal symmetry change, but not necessarily at the
“continuous” nature of structural transition suggests the ab- same temperature when the tetragonal–cubic transition
sence of an abrupt change of the electrical and optical proper- occurs.
ties across Tc. It was recently pointed out[10] that the high-
temperature stable cubic symmetry structures assigned from
XRD studies are only true in the time-average sense, whereas
Acknowledgements
at a sub-ps time scale, the dynamics of PbI6 octahedra rotation
Funding for the work at Texas Tech University from the National
leads to strong instantaneous deviation from the cubic sym-
Science Foundation (CBET-1438681) is greatly acknowledged. The
metry, with the time-averaged bandgap Eg inheriting the signa-
work at the National Renewable Energy Laboratory was support-
ture of tetragonal phase. Therefore, not only Eg, but also dEg/
ed by the U.S. Department of Energy under Contract No. DE-
dT does not display obvious transition across Tc. In contrast to
AC36-08-GO28308. K.Z. and Z.L. acknowledge the support by the
Eg, which is determined by the lowest energy necessary for
hybrid perovskite solar cell program of the National Center for
band edge optical transitions, the ionic conductivity and all
Photovoltaics funded by the U.S. Department of Energy, Office of
ionic-related electrical properties are mainly determined by
Energy Efficiency and Renewable Energy, Solar Energy Technolo-
defect-mediated ion hopping rates or the energy barrier. We
gies Office.
tentatively ascribe the different activation energy at low and
high temperature ranges to different hopping mechanics, relat-
ed to crystal symmetry change, but with different energy barri- Keywords: cubic · optoelectronic properties · tetragonal ·
ers. For example, both vacancies and interstitial defects can perovskite · phase transition
mediate the ion hopping. As suggested by molecular dynamics
simulation,[17] at room temperature, I vacancies are more dif- [1] Y. Zhou, K. Zhu, ACS Energy Lett. 2016, 1, 64 – 67.
[2] a) S. Pathak, A. Sepe, A. Sadhanala, F. Deschler, A. Haghighirad, N. Sakai,
fusive than its interstitial defects, but the anisotropic interstitial
K. C. Goedel, S. D. Stranks, N. Noel, M. Price, ACS Nano 2015, 9, 2311 –
dynamics increases more rapidly with temperature, and it be- 2320; b) M. K. Gangishetty, R. W. Scott, T. L. Kelly, Nanoscale 2016, 8,
comes dominant at high temperatures. Therefore, above this 6300 – 6307.
temperature, a smaller activation energy based on different [3] T. Hata, G. Giorgi, K. Yamashita, Nano Lett. 2016, 16, 2749 – 2753.
[4] a) T. J. Jacobsson, W. Tress, J.-P. Correa-Baena, T. Edvinsson, A. Hagfeldt,
ion hopping mechanics will set in, leading to different activa-
J. Phys. Chem. C 2016, 120, 11382 – 11393; b) D. Bryant, S. Wheeler, B. C.
tion energy and all other ion-related electrical property O’Regan, T. Watson, P. R. Barnes, D. Worsley, J. Durrant, J. Phys. Chem.
changes, as observed in our ionic/electrical studies. However, Lett. 2015, 6, 3190 – 3194; c) R. L. Milot, G. E. Eperon, H. J. Snaith, M. B.
this temperature, although closely related to the crystal sym- Johnston, L. M. Herz, Adv. Funct. Mater. 2015, 25, 6218 – 6227.
[5] T. Leijtens, G. E. Eperon, N. K. Noel, S. N. Habisreutinger, A. Petrozza,
metry change, is not necessarily Tc, which is again a result of
H. J. Snaith, Adv. Energy Mater. 2015, 5, 1500963.
the “continuous” symmetry transition nature of the MAPbI3 [6] a) E. Mosconi, F. De Angelis, ACS Energy Lett. 2016, 1, 182 – 188;
perovskite. If this speculation is correct, it may also explain b) M. N. F. Hoque, M. Yang, Z. Li, N. Islam, X. Pan, K. Zhu, Z. Fan, ACS
why the critical temperature observed in the ionic property Energy Lett. 2016, 1, 142 – 149.
[7] H. Mashiyama, Y. Kawamura, E. Magome, Y. Kubota, J. Korean Phys. Soc.
study (45 8C) is smaller than the literature-reported crystal
2003, 42, S1026 – S1029.
phase transition temperature 54–60 8C, which was determined [8] a) T. J. Jacobsson, L. J. Schwan, M. Ottosson, A. Hagfeldt, T. Edvinsson,
by XRD. In addition, we also pointed out a recent work report- Inorg. Chem. 2015, 54, 10678 – 10685; b) A. Poglitsch, D. Weber, J. Chem.
ing the possibility of an intermediate phase of MAPbI3 in a tem- Phys. 1987, 87, 6373 – 6378; c) C. C. Stoumpos, C. D. Malliakas, M. G. Ka-
natzidis, Inorg. Chem. 2013, 52, 9019 – 9038; d) Y. Kawamura, H. Mashiya-
perature range of 40–80 8C.[28]
ma, K. Hasebe, J. Phys. Soc. Jpn. 2002, 71, 1694 – 1697; e) E. Smecca, Y.
Numata, I. Deretzis, G. Pellegrino, S. Boninelli, T. Miyasaka, A. La Magna,
A. Alberti, Phys. Chem. Chem. Phys. 2016, 18, 13413 – 13422.
Conclusions [9] a) H. Zhang, X. Qiao, Y. Shen, T. Moehl, S. M. Zakeeruddin, M. Grtzel, M.
Wang, J. Mater. Chem. A 2015, 3, 11762 – 11767; b) B. J. Foley, D. L. Mar-
The ionic/electrical properties and the optical properties of
lowe, K. Sun, W. A. Saidi, L. Scudiero, M. C. Gupta, J. J. Choi, Appl. Phys.
methylammonium lead iodide (MAPbI3) perovskite across its Lett. 2015, 106, 243904; c) T. Dittrich, C. Awino, P. Prajongtat, B. Rech,
tetragonal–cubic phase transition were studied using tempera- M. C. Lux-Steiner, J. Phys. Chem. C 2015, 119, 23968 – 23972; d) N.

& ChemSusChem 2016, 9, 1 – 8 www.chemsuschem.org 6  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

 Full Papers
Onoda-Yamamuro, T. Matsuo, H. Suga, J. Phys. Chem. Solids 1992, 53, son, J. Kießling, A. Kçhler, Y. Vaynzof, S. Huettner, Adv. Mater. 2016, 28,
935 – 939; e) M. Bag, L. A. Renna, R. Adhikari, S. Karak, F. Liu, P. M. Lahti, 2446 – 2454.
T. P. Russell, M. T. Tuominen, D. Venkataraman, J. Am. Chem. Soc. 2015, [19] Y. Yuan, J. Chae, Y. Shao, Q. Wang, Z. Xiao, A. Centrone, J. Huang, Adv.
137, 13130 – 13137. Energy Mater. 2015, 5, 1500615.
[10] C. Quarti, E. Mosconi, J. M. Ball, V. D’Innocenzo, C. Tao, S. Pathak, H. J. [20] C. Eames, J. M. Frost, P. R. Barnes, B. C. O’Regan, A. Walsh, M. S. Islam,
Snaith, A. Petrozza, F. De Angelis, Energy Environ. Sci. 2016, 9, 155 – 163. Nat. Commun. 2015, 6, 7497.
[11] N. Islam, M. Yang, K. Zhu, Z. Fan, J. Mater. Chem. A 2015, 3, 24315 – [21] a) H.-S. Duan, H. Zhou, Q. Chen, P. Sun, S. Luo, T.-B. Song, B. Bob, Y.
24321. Yang, Phys. Chem. Chem. Phys. 2015, 17, 112 – 116; b) M. Samiee, S. Kon-
[12] a) A. R. Pascoe, N. W. Duffy, A. D. Scully, F. Huang, Y.-B. Cheng, J. Phys. duri, B. Ganapathy, R. Kottokkaran, H. A. Abbas, A. Kitahara, P. Joshi, L.
Chem. C 2015, 119, 4444 – 4453; b) Y. Shao, Z. Xiao, C. Bi, Y. Yuan, J. Zhang, M. Noack, V. Dalal, Appl. Phys. Lett. 2014, 105, 153502.
Huang, Nat. Commun. 2014, 5, 5784. [22] A. Jonscher, J. Mater. Sci. 1978, 13, 553 – 562.
[13] J. Beilsten-Edmands, G. Eperon, R. Johnson, H. Snaith, P. Radaelli, Appl. [23] K. O’Donnell, X. Chen, Appl. Phys. Lett. 1991, 58, 2924 – 2926.
Phys. Lett. 2015, 106, 173502. [24] N. W. Ashcroft, N. D. Mermin, Solid state physics, Holt, Rinehart and Win-
[14] T. Y. Yang, G. Gregori, N. Pellet, M. Grtzel, J. Maier, Angew. Chem. 2015, ston, New York, 1976.
127, 8016 – 8021. [25] a) C. Wehrenfennig, M. Liu, H. J. Snaith, M. B. Johnston, L. M. Herz, J.
[15] a) J. Jamnik, J. Maier, J. Electrochem. Soc. 1999, 146, 4183 – 4188; b) W. Phys. Chem. Lett. 2014, 5, 1300 – 1306; b) S. A. Kandjani, S. Mirershadi, A.
Lai, S. M. Haile, J. Am. Ceram. Soc. 2005, 88, 2979 – 2997. Nikniaz, Inorganic-Organic Perovskite Solar Cells, InTech, 2015.
[16] a) R. Andreaus, W. Sitte, J. Electrochem. Soc. 1997, 144, 1040 – 1044; b) P. [26] C. Quarti, E. Mosconi, F. De Angelis, Chem. Mater. 2014, 26, 6557 – 6569.
Jasinski, V. Petrovsky, T. Suzuki, H. U. Anderson, J. Electrochem. Soc. [27] M. T. Weller, O. J. Weber, P. F. Henry, A. M. Di Pumpo, T. C. Hansen, Chem.
2005, 152, J27 – J32. Commun. 2015, 51, 4180 – 4183.
[17] P. Delugas, C. Caddeo, A. Filippetti, A. Mattoni, J. Phys. Chem. Lett. 2016, [28] X. Guo, C. McCleese, C. Kolodziej, A. C. Samia, Y. Zhao, C. Burda, Dalton
7, 2356 – 2361. Trans. 2016, 45, 3806 – 3813.
[18] a) M. De Bastiani, G. Dell’Erba, M. Gandini, V. D’Innocenzo, S. Neutzner,
A. R. S. Kandada, G. Grancini, M. Binda, M. Prato, J. M. Ball, Adv. Energy Received: July 16, 2016
Mater. 2016, 6, 1501453; b) C. Li, S. Tscheuschner, F. Paulus, P. E. Hopkin- Published online on && &&, 0000

ChemSusChem 2016, 9, 1 – 8 www.chemsuschem.org 7  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim &

These are not the final page numbers! ÞÞ

 FULL PAPERS
M. N. F. Hoque, N. Islam, Z. Li, G. Ren,
K. Zhu,* Z. Fan*
&& – &&

Ionic and Optical Properties of

Methylammonium Lead Iodide
Perovskite across the Tetragonal–
Cubic Structural Phase Transition Cause & effect: The ionic/electrical and ing contrast in the ionic/electrical prop-
optical properties of methylammonium erties around 45 8C, attributed to
lead iodide perovskite are studied over a change in the ion activation energy;
a temperature range where a tetrago- however, the optical properties exhibit
nal–cubic phase transition is expected no particular transition around this tem-
to occur in the material. There is a strik- perature.

& ChemSusChem 2016, 9, 1 – 8 www.chemsuschem.org 8  2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!