Solvent-Mediated Crystallization of CH3NH3SnI3 Films for Heterojunction Depleted Perovskite Solar Cells

Article abstract of DOI:10.1021/jacs.5b06658

Organo-lead halide perovskite solar cells have gained enormous significance and have now achieved power conversion efficiencies of ～20%. However, the potential toxicity of lead in these systems raises environmental concerns for widespread deployment. Here we investigate solvent effects on the crystallization of the lead-free methylammonium tin triiodide (CH₃NH₃SnI₃) perovskite films in a solution growth process. Highly uniform, pinhole-free perovskite films are obtained from a dimethyl sulfoxide (DMSO) solution via a transitional SnI₂·3DMSO intermediate phase. This high-quality perovskite film enables the realization of heterojunction depleted solar cells based on mesoporous TiO₂ layer but in the absence of any hole-transporting material with an unprecedented photocurrent up to 21 mA cm^-2. Charge extraction and transient photovoltage decay measurements reveal high carrier densities in the CH₃NH₃SnI₃ perovskite device which are one order of magnitude larger than CH₃NH₃PbI₃-based devices but with comparable recombination lifetimes in both devices. The relatively high background dark carrier density of the Sn-based perovskite is responsible for the lower photovoltaic efficiency in comparison to the Pb-based analogues. These results provide important progress toward achieving improved perovskite morphology control in realizing solution-processed highly efficient lead-free perovskite solar cells.

Full text of DOI:10.1021/jacs.5b06658

Article
pubs.acs.org/JACS
Solvent-Mediated Crystallization of CH3NH3SnI3 Films for
Heterojunction Depleted Perovskite Solar Cells
Feng Hao,† Constantinos C. Stoumpos,† Peijun Guo,‡ Nanjia Zhou,‡ Tobin J. Marks,†
Robert P. H. Chang,‡ and Mercouri G. Kanatzidis*,†
‡
†Department of Chemistry and Department of Materials Science and Engineering, and Argonne-Northwestern Solar Energy
Research (ANSER) Center, Northwestern University, 2145 Sheridan Road, Evanston, Illinois 60208, United States
S
* Supporting Information
ABSTRACT: Organo-lead halide perovskite solar cells have gained enormous
significance and have now achieved power conversion efficiencies of ∼20%. However,
the potential toxicity of lead in these systems raises environmental concerns for
widespread deployment. Here we investigate solvent effects on the crystallization of
the lead-free methylammonium tin triiodide (CH3NH3SnI3) perovskite films in a
solution growth process. Highly uniform, pinhole-free perovskite films are obtained
from a dimethyl sulfoxide (DMSO) solution via a transitional SnI2·3DMSO
intermediate phase. This high-quality perovskite film enables the realization of
heterojunction depleted solar cells based on mesoporous TiO2 layer but in the
absence of any hole-transporting material with an unprecedented photocurrent up to
21 mA cm−2. Charge extraction and transient photovoltage decay measurements
reveal high carrier densities in the CH3NH3SnI3 perovskite device which are one
order of magnitude larger than CH3NH3PbI3-based devices but with comparable
recombination lifetimes in both devices. The relatively high background dark carrier
density of the Sn-based perovskite is responsible for the lower photovoltaic efficiency in comparison to the Pb-based analogues.
These results provide important progress toward achieving improved perovskite morphology control in realizing solution-
processed highly efficient lead-free perovskite solar cells.
It is becoming increasing evident, however, that it would be
highly desirable to achieve analogous optical and photovoltaic
performance using lead-free organic−inorganic perovskite
compounds because of the anticipated toxicity from a
potentially soluble lead source. Recently, we and other groups
have reported that the methylammonium tin iodide
(CH3NH3SnI3) analogue could serve as a lead-free light-
absorbing material, and solar cell power conversion efficiencies
in the range of 5−6% have been obtained independently.10,27
However, the efficiency obtained is well below that of the lead
counterpart, and it is believed that this originates from the self-
doping of the tin perovskite due to the facile conversion of Sn2+
to the Sn4+ oxidation state. This process generates p-type
carriers and results in limited diffusion lengths for the
photoexcited carriers.27 The film quality (such as homogeneity
and coverage) of the CH3NH3SnI3 perovskite layer is another
limiting factor governing the device photovoltaic perform-
ance.10 The poor film coverage with the presence of micron-
size pinholes is related to an uncontrollable crystallization
during the solution-based deposition process, a behavior which
is notably different from the lead analogue.28 The reaction
between the SnI2 films and methylammonium iodide is
recognized to be much faster than for the Pb analogue.29 The
INTRODUCTION
■
The recent emergence of organic−inorganic halide hybrid
perovskite-based solar cells has evoked widespread interest in
both scientific and industrial communities, primarily because of
the high power conversion efficiencies coupled with a relatively
simple and scalable processing techniques.1−4 These com-
pounds are part of the ABX3 perovskite family and consist of a
network of corner-sharing BX6 octahedra, where the B atom is a
divalent metal cation (typically Ge2+, Sn2+, or Pb2+) and X is a
monovalent anion (typically halides of F−, Cl−, Br−, or I−); the
A cation balances the negative charge, and it can be a Cs+ or a
small molecular species (such as methylammonium or
formamidium).5−8 Such perovskites have excellent optical
properties that are tunable by controlling the chemical
compositions9,10 and exhibit facile ambipolar charge transport11
and long electron and hole diffusion lengths.12,13 Various device
structures have been realized ranging from the sensitized solar
cells on mesoporous TiO2 scaffolds11 or insulating Al2O3
underlayers14,15 to the planar heterojunction architectures16,17
and hole-transporting material-free depleted heterojunction
devices.18−20 Solar cells based on the benchmark CH3NH3PbI3
perovskites have reached confirmed power conversion
efficiencies up to 20%15,21−23 and thus have rapidly surpassed
the efficiencies of many other emerging and commercial
photovoltaics.24−26
Received: June 26, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b06658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
easy decomposition of the tin perovskite in isopropanol solvent
also makes the sequential solution process difficult to realize.22
Therefore, developing novel deposition methods for high-
quality CH3NH3SnI3 films which are expected to achieve an
efficiency boost in the development of lead-free perovskite solar
cells remains a challenge.
Table 1. Physical Properties of the Polar Solvents Used in
the Perovskite Film Deposition64
boiling point viscosity
dielectric
constant
vapor pressure at
20 °C (Torr)
(°C)
(cP)
DMF
GBL
152
204
202
189
0.92
1.90
1.65
2.00
36.7
39.1
33
2.7
1.5
Compared to the numerous studies on the morphological
control of CH3NH3PbI3 film growth,16,22,23,30−33 studies on the
tin counterparts are sparse because experiments are typically
hampered by rapid oxidation under ambient conditions and the
requirement of an inert dry atmosphere. Here we show that the
crystallization of the CH3NH3SnI3 perovskite films can be
greatly improved through solvent engineering employed in a
one-step spin-coating process. Solvent evaporation effects are
detrimental to obtaining high-quality CH3NH3SnI3 films given
the faster reaction rate between the two halide precursors
compared to the CH3NH3PbI3 analogue. Controlled solvent
evaporation and the convective self-assembly process during
spin-coating effectively induce the formation of well-crystallized
CH3NH3SnI3 perovskite films due to the high enthalpy of
crystallization, which is much greater than the solvation
energies of the individual ions. Here we report that very
uniform, pinhole-free CH3NH3SnI3 perovskite films can be
obtained from a strongly coordinating solvent such as dimethyl
sulfoxide (DMSO) which stabilizes a transitional SnI2·3DMSO
intermediate phase. Crystallization processes using other highly
polar solvents, N,N-dimethylformamide (DMF) and N-methyl-
2-pyrrolidone (NMP), have also been investigated and are
compared here. Taken together, these results show that
solvated intermediates are an important transitional step
toward a uniform, high-coverage perovskite films. Benefiting
from the high-quality of the present CH3NH3SnI3 films, lead-
free perovskite heterojunction depleted solar cells are fabricated
with photocurrent densities of up to 21 mA cm−2, albeit with a
relatively modest diffusion length. The underlying carrier
generation, electron transfer, and recombination dynamics are
investigated and compared by charge extraction and small
perturbation transient photovoltage analyses in both methyl-
ammonium tin and lead halide perovskite devices. We show
that the carrier density in the CH3NH3SnI3 perovskite device is
one order of magnitude larger than that of the lead analogue,
but the recombination lifetimes in both devices are comparable.
Our results hold promise for further efficiency enhancement in
lead-free perovskite solar cells, taking advantage of higher
quality perovskite films.
NMP
DMSO
0.29
0.42
46.7
One of the fundamental challenges encountered in the one-step
tin perovskite spin-coating process is that a homogeneous
perovskite layer cannot be formed over a large area, even when
a convective spreading flow due to centrifugal force is applied
to the slowly evaporating solvents.34 Note that for identical
concentrations, CH3NH3SnI3 solutions crystallize much more
rapidly than those of CH3NH3PbI3 when DMF is the solvent.
This is readily evidenced by the spontaneous formation of a
black perovskite layer after the spin-coating process in the case
of CH3NH3SnI3, whereas the complete crystallization of
CH3NH3PbI3 is much slower and is achieved only on thermal
treatment. This difference stems from the faster reaction of SnI2
with CH3NH3I to form the perovskite, owing to the greater
Lewis acidity of Sn2+ versus Pb2+.35 Therefore, retardation of
the crystallization process is considered here to be beneficial for
better CH3NH3SnI3 film quality since the perovskite is allowed
to form in a stepwise, controllable fashion. Compared to the
routinely used DMF, DMSO has a higher coordination affinity,
a relatively high boiling point (189 °C), and a low saturated
vapor pressure (0.42 Torr at 20 °C). Several recent studies have
confirmed that the strong coordination of CH3NH3PbI3 by
DMSO leads to stabilization of intermediate solvated
compounds.30,35 On the other hand, the viscosity of the
solvent is also important in controlling evaporation during the
spin-coating process and should be taken into consideration
during homogeneous thin-film fabrication. From the above
considerations, DMF and DMSO were selected to fabricate
CH3NH3SnI3 films and to compare their crystallization
processes.
A mesoporous TiO2 layer with a thickness of 200 nm was
used as the substrate for all perovskite depositions in this study.
Perovskite films were deposited using a one-step spin-coating
process from 1.0 M perovskite solutions. It was immediately
clear that the different solvents markedly affect the resultant
perovskite film formation. As stated above and shown in Figure
1a, CH3NH3SnI3 perovskite crystals form immediately during
the spin-coating process from DMF, as evidenced from the
color changing to dark brown during 4000 rpm rotation
without any thermal treatment. Switching to DMSO effectively
slows the perovskite crystallization since the color change to
dark brown is not observed during spin-coating. The coated
layer from DMSO solvent is uniform and light yellow. A black
perovskite film gradually forms when annealing the substrate at
100 °C on a preheated hot plate for 15 min. To understand the
solvent effects on the perovskite crystallization process, X-ray
diffraction (XRD) patterns of the films before and after the
thermal annealing process were analyzed. As shown in Figure
1b, the as-cast films from DMF show high crystallinity
regardless of the thermal annealing, and all the peaks can be
indexed to the tetraganol CH3NH3SnI3 phase in space group
P4mm. This high crystallinity underscores the rapid perovskite
crystallization during spin-coating from DMF solution.
However, when DMSO was used as the solvent, the as-formed
RESULTS AND DISCUSSION
■
Film Fabrication and Characterization. The selection of
solvents for dissolving tin halide perovskites is found to be
critical for obtaining high quality thin films using solution
coating techniques. In general, the choice of solvent is a balance
between several competing factors. On the one hand, the
solvent must be capable of dissolving both the inorganic and
organic halide components, and therefore a good choice of
solvent can be found among the strongly coordinating/polar
amide class. On the other hand, if the reagent solubility in the
solvent is too high, then during the deposition of successive
layers previous layers will redissolve during the deposition of
the next layer.6 Normally, γ-butyrolactone (GBL), dimethyl
sulfoxide (DMSO), N-methyl-2-pyrrolidone(NMP), and N,N-
dimethylformamide (DMF) are widely used for lead halides
and methylammonium iodide.29 The basic physical properties
of these solvents are summarized in Table 1 for comparison.
B
DOI: 10.1021/jacs.5b06658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
CH3NH3SnI3 perovskite is dramatically inhibited by the strong
solvent coordination to the metal precursor.
The presence of DMSO in the intermediate phase was
further confirmed by Fourier transform infrared spectroscopy
(FTIR). Figure 1c shows the FTIR spectra of the characteristic
C−S and S−O stretching vibrations from the Sn2+-coordinated
DMSO solvent at 960 and 1012 cm−1, respectively. Upon
annealing, these characteristic modes of the DMSO molecule
disappear, thus confirming the complete removal of the solvent
and the conversion of the intermediate compound to the
CH3NH3SnI3 perovskite. For comparison, the characteristic
CO bonding of the DMF solvent is absent in the perovskite
film fabricated from DMF, as shown in Figure S1. The crystal
structure of the solvent-induced intermediate was further
characterized by single crystal X-ray diffraction. Pale yellow
block crystals of SnI2·3DMSO were obtained by means of the a
recently reported antisolvent vapor-assisted crystallization
process,37 using 1.0 mL of a 2.0 M DMSO solution of
CH3NH3SnI3 with CH2Cl2 as the antisolvent. The crystal
structure was solved by direct methods and refined using
SHELXTL software.38 As shown in Figure 2a, the crystal
structure of SnI2·3DMSO is composed of discrete SnI-
+
(DMSO)3 ions in a seesaw coordination environment. Three
O-coordinated DMSO molecules occupy the corners of a
trigonal pyramid with the Sn ion at the apex, and a fourth
iodide ion completes the coordination sphere. The seesaw
polyhedra are linked by a nearly free I− ion which coordinates
via weak (3.522 Å) or very weak (3.987 Å) bonding
interactions. The incorporation of all three iodide ligands
changes the coordination of the Sn to octahedral with the O
atoms (DMSO) and the three loosely bound iodide atoms
occupying opposing facial positions. This self-assembly results
in the formation of 1D zigzag chains running down the
crystallographic b-axis (Figure 2b). Structural details can be
found in the Supporting Information. It is instructive to note
here that the present film formation process through solvent
engineering is markedly different in CH3NH3PbI3 and
CH3NH3SnI3 with DMSO. Whereas in the former case the
DMSO film acts as a hydrogen bond acceptor in CH3NH3PbI3·
DMSO (assuming a similar structure to CH3NH3PbI3·DMF),
in the latter case DMSO directly coordinates to the Sn ion in
Figure 1. (a) Single step spinning-coating process with the formation
of DMSO solvate intermediate. Images of the as-prepared films from a
DMSO solution before and after annealing process are shown in the
insets. (b) X-ray diffraction patterns of the perovskite layer with
different solvents before (left) and after (right) thermal annealing
process at 100 °C for 15 min. (c) FTIR spectra of the perovskite layer
deposited from DMSO solution before annealing showing the
characteristic C−S and C−O stretching vibrations from the Sn2+-
coordinated DMSO solvent at 960 and 1012 cm−1, respectively. Upon
annealing, these characteristic DMSO peaks disappear, confirming the
complete removal of the solvent and conversion of the intermediate
compound to the CH3NH3SnI3 perovskite.
+
SnI2·3DMSO, leaving the CH3NH3 cation dissolved in the
mother liquor (Figure 2c). Thus, perovskite formation takes
place on the substrate, and the perovskite film formation can be
controlled by varying the rate of DMSO evaporation.
NMP was also investigated as a solvent to further understand
the effect of the intermediate phase on the tin perovskite
formation process. As DMSO, NMP is a polar aprotic solvent
with a high boiling point of 202 °C and low vapor pressure
(0.29 Torr at 20 °C). Similar to DMSO, the NMP forms black
perovskite film only after annealing of the spin-cast film at 100
°C. The color change from pale yellow to black takes place
more slowly than from DMSO. We attribute this to the slightly
higher boiling point of NMP and slower release of the
coordinated solvent on thermal treatment. As demonstrated in
Figure 1a, X-ray diffraction of the as-cast NMP pale yellow films
shows very different features from the black CH3NH3SnI3
phase. The characteristic peak at 10° suggests a larger unit
cell for the solvated phase, which may be ascribed to the larger
volume of NMP incorporated compared to DMSO. GBL was
also examined for comparison in this study. It was noticed the
crystallization rate of the perovskite film from GBL solvent is
faster than from DMSO and NMP, but slower than DMF. The
films show a totally different set of reflections, which cannot be
assigned either to the starting precursor or the target perovskite
compound. We therefore assume this is an intermediate
solvated phase formed during the perovskite deposition
process. Previous reports have shown that DMSO forms a
CH3NH3PbI3·DMSO complex.32,36 In the present system, the
formation of the intermediate phase can be observed from the
XRD features at lower angles. Unlike the case of CH3NH3PbI3·
DMSO, however, where reflections analogous to our previously
reported CH3NH3PbI3·DMF and CH3NH3PbI3·H2O solvates
were observed,28 in the case of the tin perovskite an
+
intermediate solvated phase does not contain the CH3NH3
cation and has been identified as SnI2·3DMSO. Note that the
Sn solvate has a completely different crystal structure than that
of formally analogous PbI2·2DMSO.36 The presence of the
DMSO solvate phase indicates that the formation of the
C
DOI: 10.1021/jacs.5b06658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 2. Crystal structure of SnI2·3DMSO, an intermediate compound in the CH3NH3SnI3 film fabrication process. (a) The dimeric structure of
the SnI(DMSO)3+ ions linked through the lone I− ions. (b) The unit cell of SnI2·3DMSO in parallel view. Space group C2/c; a = 21.9920(10) Å, b =
10.7497(6) Å, c = 16.0302(7) Å β = 111.838(3)°. Sn−I(1) = 3.285 Å, Sn−I(2) = 3.522 Å, Sn···I(2)′ = 3.987 Å (structure details are given in the
Supporting Information). (c) A schematic on the film formation of the CH3NH3SnI3 perovskite film starting from SnI2 through the SnI2·3DMSO
intermediate.
Figure 3. Representative SEM images of the as-obtained CH3NH3SnI3 perovskite layer on mesoporous TiO2 from different solvents, (a) DMF, (b)
NMP, and (c, d) DMSO with different magnifications, and the corresponding optical absorption spectra (e). Apparently, films deposited from NMP
and DMSO solvents showed complete surface coverage, whereas obvious uncovered TiO2 surface is exposed when cast from the DMF as the solvent.
black perovskite film forms in seconds after finishing the
spinning process without any heat treatment. The above trends
appear to originate from the differences in evaporation rates of
these three solvents during the spin-coating process. The vapor
pressure at 20 °C for GBL is 1.5 Torr as listed in Table 1, which
lies in between DMF (2.7 Torr) and DMSO (0.42 Torr) or
NMP (0.29 Torr).
On the basis of the above observations, we propose a
plausible mechanism for the formation of the uniform, high-
coverage perovskite layer during the present single step spin-
coating procedure. As shown in the Figure 2c, the formation of
the CH3NH3SnI3 proceeds in a controllable manner through
the SnI2·3DMSO intermediate which allows the iodide ions to
react with SnI2 in a controlled manner that is governed by the
rate of the DMSO removal as a function of temperature.
Therefore, a close relationship appears to exist between the
solvent-induced intermediate phase and the control of the
resultant perovskite film crystallization process.
Analysis of these perovskite films prepared from different
solvents by scanning electron microscopy (SEM) reveals
strikingly different morphologies. As compared in Figure 3,
films produced from DMF solution are quite discontinuous and
composed of large CH3NH3SnI3 grains and many unfilled
pinholes. This morphology arises from relatively fast crystal-
lization rate resulting from the high nucleation rate during the
drying process. Employing a solvent with lower vapor pressure
than DMF, such as GBL solvent gives improved film coverage
but still has pinholes as presented in Figure S3. In contrast,
films deposited from DMSO as the solvent are homogeneous
with almost full surface coverage on the mesoporous TiO2
layer. Impressively, this very smooth CH3NH3SnI3 layer shows
large micron-sized grains, which is a highly desirable property,
as shown in recent studies where the photovoltaic performance
of the benchmark CH3NH3PbI3 solar cells was investigated as a
function of grain size.39 The normalized optical absorption
spectra of the perovskite films derived from different solvents in
D
DOI: 10.1021/jacs.5b06658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 4. (a) Device architecture of the depleted heterojunction perovskite solar cell (FTO glass/blocking TiO2 layer/mesoporous TiO2 layer/
perovskite/Au). (b) Characteristic photocurrent−voltage plots and (c) incident photon-to-current efficiency (IPCE) spectra for the CH3NH3SnI3
and CH3NH3PbI3 perovskite devices. Negative scan and positive scan direction were shown to identify the hysteresis behavior in both devices. (d)
Comparison of the photocurrent density as a function of time for the two devices held at a forward bias of maximum output power point. The cell
was placed in the dark prior to the start of the measurement.
are shown in Figure 3. Similar absorption edges at around 950
nm are observed for all samples, consistent with the same
crystalline structure of the resultant perovskite films after
thermal annealing. However, note the red-shift of the
absorption spectrum of the DMSO and NMP samples in the
long wavelength 750−950 nm range versus the DMF sample.
This red-shift of absorbance correlates well with the observed
crystal size, which is consistent with several recent studies on
the perovskite photoluminescence spectra.40,41 In these studies,
small perovskite crystallites exhibit blue-shifted emission
relative to large crystals. The optical absorption coefficient
(α/S) was calculated using reflectance data according to the
Kubelka−Munk equation,42 α/S = (1 − R)2/2R, where R is the
percentage of reflected light and α and S are the absorption and
scattering coefficients, respectively. The DMSO solvent-derived
films have a similar optical band gap edge compared to DMF-
derived films (∼1.3 eV), but with a steeper slope in the
absorption, suggesting a higher absorption coefficient.
Heterojunction Solar Cell Fabrication and Perform-
ance. The present high-quality perovskite films were
incorporated into a TiO2-based heterojunction depleted device
structure, which is free of hole transporting materials (HTMs)
such as the commonly used spiro-OMeTAD.18−20 A schematic
of the heterojunction depleted device is shown in Figure 4a, in
which the perovskite layer acts as a light absorber and also a
hole transporting material. One important reason to exclude
the HTM with these CH3NH3SnI3 perovskite films is the
incompatibility, recognized by us and other groups, of the tin
perovskite and additives generally added to increase the hole
mobility of the pristine spiro-OMeTAD, such as 4-tert-
butylpyridine and hydrophilic lithium bis(trifluoromethane)-
sulfonimide.10,27 In this study, a compact TiO2 layer with a
thickness around 50 nm was first deposited on a pre-etched
FTO substrate to reduce carrier recombination. A 200 nm
mesoporous TiO2nanocrystalline layer was then cast on top to
serve as the electron transporting layer. The perovskite layer
was next spin-coated on top of the mesoporous TiO2 layer
using a 1.0 M DMSO solution of CH3NH3SnI3. Note that 20
mol % SnF2 was added to the solution to decrease the
conductivity of the CH3NH3SnI3 layer according to our
previous work.43 Recent work also shows that SnF2 addition
is important to decrease Sn vacancies in the CsSnI3 perovskite
system.44 The device was completed by thermal evaporation of
a 100 nm gold electrode on top under a vacuum up to 10−5
Torr. All devices were assembled in a N2-filled glovebox and
were quickly transferred to the thermal evaporator, hence
minimizing exposure to air ∼5 min. Encapsulation was
subsequently performed under the same conditions using a
hot melt polymer (Surlyn film from DuPont) laminate with a
glass coverslip. This encapsulation was done to protect the
devices against O2 or moisture after fabrication. The sealed cells
were next removed from the glovebox and immediately
measured in air. Control devices using DMF solvent-derived
perovskite films were fabricated under the same conditions. A
depleted heterojunction device with CH3NH3PbI3 was also
fabricated in the absence of HTM material, without
encapsulation. The CH3NH3PbI3 perovskite layer was
deposited using the conventional sequential deposition
method.19,22 Device photovoltaic performance characteristics
are presented in Figure 4b.
E
DOI: 10.1021/jacs.5b06658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
It should be mentioned that all CH3NH3SnI3 devices
fabricated with DMF as the solvent suffered short circuits.
We attribute this to the inevitable voids or film defects formed
during the spin-coating process, as shown in Figure 3a. These
voids or pinholes would lead to direct contact between the top
electrode and the FTO substrate, causing shorting of these
devices. In contrast, the DMSO solution derived perovskite
films effectively avoid device shorting, primarily owing to the
ultrasmooth pinhole-free film morphology and low doping state
of the tin perovskite. These devices show an unprecedented
short-circuit photocurrent density of 21.4 mA cm−2, albeit with
a modest open-circuit photovoltage of 320 mV and a fill factor
of 0.46. The overall power conversion efficiency is calculated to
be 3.15%. The photovoltaic performance obtained is sub-
stantially lower than that of the CH3NH3PbI3 counterpart with
the same device structure, which shows a short-circuit
photocurrent density of 15.2 mA cm−2, an open-circuit
photovoltage of 668 mV, and a fill factor of 0.57, corresponding
to a power conversion efficiency of 5.79%. The limited
photovoltaic performance for this HTM-free heterojunction
device could originate from loss of the light harvesting in the
long wavelength region as well as the interfacial charge
recombination loss at the perovskite interface.19 The smaller
open-circuit photovoltage for the CH3NH3SnI3 cell may
originate from the lower conduction band edge (4.17 eV)
compared to that of lead analogue (3.93 eV) as measured in our
previous report.28 However, other factors may play a role such
as higher background carrier density. The higher photocurrent
density in the CH3NH3SnI3 device is consistent with the
smaller optical band gap (1.3 eV) and therefore a wider light
harvesting range than the lead analogue (1.55 eV), which is
corroborated with the incident-photon-to-current conversion
efficiency (IPCE) spectra shown in Figure 4c. The devices with
DMSO-derived CH3NH3SnI3 films show photocurrent re-
sponse up to 950 nm, which is almost 150 nm red-shifted
compared with CH3NH3PbI3 device. Integrating the overlap of
the IPCE spectrum with the AM 1.5G solar photon flux yields a
photocurrent density of 20.58 and 15.17 mA cm−2 for the tin
and lead halide perovskite films, respectively, which is in
excellent agreement with the measured photocurrent densities.
This confirms the negligible mismatch between the simulated
sunlight and the AM 1.5G standard.
Since the notorious photocurrent hysteresis presents a major
issue during characterization of device efficiency in lead
perovskites,45−47 we measured the photovoltaic response by
reversing the scanning direction in our cells. No noticeable
hysteresis is observed with different scanning directions for the
CH3NH3SnI3 perovskite solar cell in Figure 4b. A small shift in
Voc was noticed when switching the scanning direction for the
CH3NH3PbI3 device. This hysteresis agrees well with the
previous reports of lead perovskite solar cells and has been
speculated to originate from various causes including trapping/
detrapping of charge carriers,46 changes in absorber or contact
conductivity,48 ferroelectricity,49 and ion migration.50 In order
to verify the measured photocurrent density, we also measured
the steady-state photocurrent output at the maximum power
point. As shown in Figure 4d, the photocurrent rapidly
saturates, and the steady-state current is found to be stable at
16.8 mA cm−2 for the entire measurement period of 300 s.
However, a slight decay on the photocurrent density in the lead
perovskite heterojunction device is observed within the first
hundred seconds under illumination, which then stabilizes for
the entire testing period. Such a photocurrent density decay
could originate from defects near the surface of perovskite or
interface states specifically generated from the solution
processing.51 The relatively stable steady-state photocurrent
response in the CH3NH3SnI3 device further confirms the high
quality and low level of electronic disorder in the perovskite
layer.
Charge Extraction and Transient Photovoltage Meas-
urements. To further understand the photovoltaic differences
between the lead and tin perovskite solar cells, charge
extraction and transient photovoltage measurements were
performed to investigate the underlying carrier dynamics
along the entire pathway in the completed solar cells. The
carrier density in the devices under operational conditions with
different light intensities was first probed by the charge
extraction method.52−54 Figure 5a shows the charge extraction
Figure 5. (a) Representative charge extraction curves for the
CH3NH3SnI3 devices under different incident light intensity. (b)
Comparison of the integrated carrier density as a function of light
intensity for the CH3NH3SnI3 and CH3NH3PbI3 devices from the
charge extraction measurements. (c) Typical transient photovoltage
decay curves for the CH3NH3SnI3 devices as a function of light
intensity. The exponential fitted curves are also shown in both cases.
(d) Recombination lifetime derived from the transient photovoltage
decay curves as a function of the carrier density in the operational
devices for CH3NH3SnI3 and CH3NH3PbI3 devices.
curves for the CH3NH3SnI3 devices under different incident
light intensities. The exponential fitting curves are also
presented. Note that the initial steady-state starting condition
is from short circuit. The dependence of the carrier
concentration on the illuminating light intensity for the two
types of perovskite solar cells is shown in Figure 5b. Typically,
the carrier density shows an exponential increase with the
incident light intensity in both devices as an increasing number
of injected electrons are stored in the perovskite films. The
stored charge follows Boltzmann statistics, which results in
accumulation and carrier diffusion dominated dynamics. The
CH3NH3PbI3 device generates a carrier density of the order of
1016 cm−3 under white LED illumination. Impressively, the
F
DOI: 10.1021/jacs.5b06658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
light-induced carrier density for the CH3NH3SnI3 counterpart
is 1 order of magnitude higher than the lead device, which is
attributed to the higher JSC observed in the corresponding J−V
curves (Figure 4b). The photocarrier charge density obtained
can be traced back to the intrinsic properties of the
CH3NH3PbI3 and CH3NH3SnI3 perovskites, which was
shown previously by Hall effect measurements have a dark
carrier density of ∼109−1011 cm−3 and ∼1014−1017 cm−3 for
CH3NH3PbI3 and for semiconducting CH3NH3SnI3, respec-
tively. The “background” dark carrier density of the Sn-based
perovskite is thus responsible for the lower photovoltaic
efficiency in comparison with the Pb-based analogues.
lead-free perovskite heterojunction depleted solar cells without
a hole-transporting material layer. An unprecedented photo-
current density of up to 21 mA cm−2 is obtained while charge
extraction and transient photovoltage decay measurements
reveal comparable recombination lifetimes in the CH3NH3PbI3
and CH3NH3SnI3 devices. The tin perovskite device exhibits 1
order of magnitude higher photocarrier density with respect to
the lead analogue as a result of its higher dark carrier density.
These results suggest that the Sn-based devices can be used as
efficient alternatives for the Pb-based devices provided the
“background” carrier density can be reduced to intrinsic levels.
This expectation derives from the fact that semiconductors with
similar band gaps to CH3NH3SnI3, such as InP (Eg = 1.34 eV),
are capable of intrinsic “background” carrier densities as low as
107 cm−3.63 Therefore, future efforts should focus on ways to
decrease the background carrier density in CH3NH3SnI3. We
can then expect devices with efficiencies greater than 15%,
approaching those of the venerable CH3NH3PbI3 perovskite.
Small perturbation transient photovoltage measurements
were next employed to derive the recombination lifetime in the
perovskite absorber layer and to gather information on carrier
recombination rates in the working device.56,57 Representative
transient photovoltage curves for both devices are shown in
Figure 5c. The apparent lifetime of photogenerated carriers (τ)
can be evaluated by fitting the transient decay to an exponential
function. Details of the fitting curves are shown in the Figures
S6 and S7. Figure 5d shows the apparent recombination
lifetime as a function of carrier density derived from the charge
extraction. The recombination lifetime decreases with carrier
density in the device due to intercarrier collisions and the larger
driving force for charge recombination. Unexpectedly, these
two devices showed comparable recombination lifetime in the
range of 10−20 μs. However, a notable carrier density
difference is observed, which is attributable to the different
Voc in the J−V curves and possibly derives from the different
recombination mechanisms between CH3NH3PbI3 and
CH3NH3SnI3. Note that our devices showed substantially
longer recombination lifetimes compared to the several
microseconds for the recent high-efficiency CH3NH3PbI3
devices with a hole-transporting layer.58 A similar trend was
also reported in a recent study comparing the interfacial charge
transfer mechanism in both p−n heterojunction and p−i−n
device architectures.59 The additional recombination pathways
between the excited electrons in the perovskite layer and holes
in hole-transporting materials were ascribed to the shorter
carrier lifetimes in the p−i−n devices with a spiro-OMeTAD
layer.59 As for the difference between the CH3NH3SnI3 and
CH3NH3PbI3 devices in our study, we observe that the slopes
of the carrier lifetimes in Figure 5d are quite different, thus
reflecting the different recombination kinetics in these devices.
This is reasonable due to the significant mobility differences in
the two perovskite materials as shown in our previous
reports.28,55 Another reason may be associated with the
different defect distribution profiles in the two perovskites.60−62
Further investigations on this issue using electrochemical
impedance analyses and transient photoluminescence spectros-
copy are underway in our group.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.5b06658.
■
S
Detailed experimental procedures and crystalline struc-
ture information on the DMSO solvation intermediates
(SnI2·3DMSO), Figures S1−S7 and Tables S1−S5
(PDF)
Structure of SnI2·3DMSO (CIF)
AUTHOR INFORMATION
Corresponding Author
*E-mail m-kanatzidis@northwestern.edu.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This research was supported as part of the ANSER Center, an
Energy Frontier Research Center funded by the U.S Depart-
ment of Energy, Office of Science, Office of Basic Energy
Sciences, under Award DE-SC0001059. Electron microscopy
analysis was done through the use of the EPIC facility
(NUANCE Center, Northwestern University), which has
received support from the MRSEC program (NSF DMR-
1121262) at the Materials Research Center, and the Nanoscale
Science and Engineering Center (EEC-0118025/003), both
programs of the National Science Foundation, the state of
Illinois, and Northwestern University.
REFERENCES
■
(1) Nayak, P. K.; Cahen, D. Adv. Mater. 2014, 26, 1622.
(2) Kazim, S.; Nazeeruddin, M. K.; Gratzel, M.; Ahmad, S. Angew.
Chem., Int. Ed. 2014, 53, 2812.
CONCLUSIONS
■
(3) Lotsch, B. V. Angew. Chem., Int. Ed. 2014, 53, 635.
(4) Wei, Z. H.; Chen, H. N.; Yan, K. Y.; Yang, S. H. Angew. Chem.,
Int. Ed. 2014, 53, 13239.
(5) Borriello, I.; Cantele, G.; Ninno, D. Phys. Rev. B: Condens. Matter
Mater. Phys. 2008, 77, 235214.
(6) Mitzi, D. B. Prog. Inorg. Chem. 1999, 48, 1.
(7) Kagan, C. R.; Mitzi, D. B.; Dimitrakopoulos, C. D. Science 1999,
286, 945.
(8) Zhang, Q. C.; Liu, X. G. Small 2012, 8, 3711.
(9) Noh, J. H.; Im, S. H.; Heo, J. H.; Mandal, T. N.; Seok, S. I. Nano
Lett. 2013, 13, 1764.
In summary, solvent effects on the crystallization process for
CH3NH3SnI3 perovskite films has been investigated and
compared in single-step solution processes. The solvent-
induced intermediate phase plays an important role in
controlling the crystallization process and final thin film quality.
The stabilization of the intermediate SnI2 solvates prior to the
formation of the perovskite films is found to promote
homogeneous nucleation and enables an adjustable perovskite
film growth rate. High-quality, pinhole-free CH3NH3SnI3 films
are achieved using DMSO and NMP as solvents and afford
G
DOI: 10.1021/jacs.5b06658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(10) Hao, F.; Stoumpos, C. C.; Cao, D. H.; Chang, R. P. H.;
Kanatzidis, M. G. Nat. Photonics 2014, 8, 489.
(40) De Bastiani, M.; D’Innocenzo, V.; Stranks, S. D.; Snaith, H. J.;
Petrozza, A. APL Mater. 2014, 2, 081509.
(41) Zhang, W.; Saliba, M.; Moore, D. T.; Pathak, S. K.; Horantner,
M. T.; Stergiopoulos, T.; Stranks, S. D.; Eperon, G. E.; Alexander-
Webber, J. A.; Abate, A.; Sadhanala, A.; Yao, S. H.; Chen, Y. L.; Friend,
R. H.; Estroff, L. A.; Wiesner, U.; Snaith, H. J. Nat. Commun. 2015, 6,
6142.
(42) Gate, L. F. Appl. Opt. 1974, 13, 236.
(43) Chung, I.; Lee, B.; He, J. Q.; Chang, R. P. H.; Kanatzidis, M. G.
Nature 2012, 485, 486.
(11) Heo, J. H.; Im, S. H.; Noh, J. H.; Mandal, T. N.; Lim, C. S.;
Chang, J. A.; Lee, Y. H.; Kim, H. J.; Sarkar, A.; Nazeeruddin, M. K.;
Gratzel, M.; Seok, S. I. Nat. Photonics 2013, 7, 486.
(12) Xing, G. C.; Mathews, N.; Sun, S. Y.; Lim, S. S.; Lam, Y. M.;
Gratzel, M.; Mhaisalkar, S.; Sum, T. C. Science 2013, 342, 344.
(13) Stranks, S. D.; Eperon, G. E.; Grancini, G.; Menelaou, C.;
Alcocer, M. J. P.; Leijtens, T.; Herz, L. M.; Petrozza, A.; Snaith, H. J.
Science 2013, 342, 341.
(44) Kumar, M. H.; Dharani, S.; Leong, W. L.; Boix, P. P.; Prabhakar,
R. R.; Baikie, T.; Shi, C.; Ding, H.; Ramesh, R.; Asta, M.; Graetzel, M.;
Mhaisalkar, S. G.; Mathews, N. Adv. Mater. 2014, 26, 7122.
(45) Unger, E. L.; Hoke, E. T.; Bailie, C. D.; Nguyen, W. H.;
Bowring, A. R.; Heumuller, T.; Christoforo, M. G.; McGehee, M. D.
Energy Environ. Sci. 2014, 7, 3690.
(46) Snaith, H. J.; Abate, A.; Ball, J. M.; Eperon, G. E.; Leijtens, T.;
Noel, N. K.; Stranks, S. D.; Wang, J. T. W.; Wojciechowski, K.; Zhang,
W. J. Phys. Chem. Lett. 2014, 5, 1511.
(14) Wojciechowski, K.; Saliba, M.; Leijtens, T.; Abate, A.; Snaith, H.
J. Energy Environ. Sci. 2014, 7, 1142.
(15) Lee, M. M.; Teuscher, J.; Miyasaka, T.; Murakami, T. N.; Snaith,
H. J. Science 2012, 338, 643.
(16) Liu, M. Z.; Johnston, M. B.; Snaith, H. J. Nature 2013, 501, 395.
(17) Liu, D. Y.; Kelly, T. L. Nat. Photonics 2014, 8, 133.
(18) Etgar, L.; Gao, P.; Xue, Z. S.; Peng, Q.; Chandiran, A. K.; Liu,
B.; Nazeeruddin, M. K.; Gratzel, M. J. Am. Chem. Soc. 2012, 134,
17396.
(47) Wei, J.; Zhao, Y. C.; Li, H.; Li, G. B.; Pan, J. L.; Xu, D. S.; Zhao,
Q.; Yu, D. P. J. Phys. Chem. Lett. 2014, 5, 3937.
(19) Hao, F.; Stoumpos, C. C.; Liu, Z.; Chang, R. P. H.; Kanatzidis,
M. G. J. Am. Chem. Soc. 2014, 136, 16411.
(48) Staebler, D. L.; Wronski, C. R. J. Appl. Phys. 1980, 51, 3262.
(49) Dualeh, A.; Moehl, T.; Tetreault, N.; Teuscher, J.; Gao, P.;
Nazeeruddin, M. K.; Gratzel, M. ACS Nano 2014, 8, 4053.
(50) Tress, W.; Marinova, N.; Moehl, T.; Zakeeruddin, S. M.;
Nazeeruddin, M. K.; Gratzel, M. Energy Environ. Sci. 2015, 8, 995.
(51) Kim, H. S.; Park, N. G. J. Phys. Chem. Lett. 2014, 5, 3434.
(52) Shuttle, C. G.; Maurano, A.; Hamilton, R.; O’Regan, B.; de
Mello, J. C.; Durrant, J. R. Appl. Phys. Lett. 2008, 93, 183501.
(53) Duffy, N. W.; Peter, L. M.; Rajapakse, R. M. G.; Wijayantha, K.
G. U. Electrochem. Commun. 2000, 2, 658.
(54) Shuttle, C. G.; Hamilton, R.; O’Regan, B. C.; Nelson, J.;
Durrant, J. R. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 16448.
(55) Stoumpos, C. C.; Malliakas, C. D.; Kanatzidis, M. G. Inorg.
Chem. 2013, 52, 9019.
(56) Mora-Sero, I.; Dittrich, T.; Garcia-Belmonte, G.; Bisquert, J. J.
Appl. Phys. 2006, 100, 103705.
(20) Mei, A. Y.; Li, X.; Liu, L. F.; Ku, Z. L.; Liu, T. F.; Rong, Y. G.;
Xu, M.; Hu, M.; Chen, J. Z.; Yang, Y.; Gratzel, M.; Han, H. W. Science
2014, 345, 295.
(21) Kojima, A.; Teshima, K.; Shirai, Y.; Miyasaka, T. J. Am. Chem.
Soc. 2009, 131, 6050.
(22) Burschka, J.; Pellet, N.; Moon, S. J.; Humphry-Baker, R.; Gao,
P.; Nazeeruddin, M. K.; Gratzel, M. Nature 2013, 499, 316.
(23) Zhou, H.; Chen, Q.; Li, G.; Luo, S.; Song, T.-b.; Duan, H.-S.;
Hong, Z.; You, J.; Liu, Y.; Yang, Y. Science 2014, 345, 542.
(24) Mathew, S.; Yella, A.; Gao, P.; Humphry-Baker, R.; Curchod, B.
F. E.; Ashari-Astani, N.; Tavernelli, I.; Rothlisberger, U.; Nazeeruddin,
M. K.; Gratzel, M. Nat. Chem. 2014, 6, 242.
(25) Liu, Y. S.; Chen, C. C.; Hong, Z. R.; Gao, J.; Yang, Y.; Zhou, H.
P.; Dou, L. T.; Li, G.; Yang, Y. Sci. Rep. 2013, 3, 3356.
(26) Green, M. A.; Emery, K.; Hishikawa, Y.; Warta, W.; Dunlop, E.
D. Prog. Photovoltaics 2014, 22, 701.
(27) Noel, N. K.; Stranks, S. D.; Abate, A.; Wehrenfennig, C.;
Guarnera, S.; Haghighirad, A. A.; Sadhanala, A.; Eperon, G. E.; Pathak,
S. K.; Johnston, M. B.; Petrozza, A.; Herz, L. M.; Snaith, H. J. Energy
Environ. Sci. 2014, 7, 3061.
(28) Hao, F.; Stoumpos, C. C.; Chang, R. P. H.; Kanatzidis, M. G. J.
Am. Chem. Soc. 2014, 136, 8094.
(29) Liang, K. N.; Mitzi, D. B.; Prikas, M. T. Chem. Mater. 1998, 10,
403.
(30) Wu, Y. Z.; Islam, A.; Yang, X. D.; Qin, C. J.; Liu, J.; Zhang, K.;
Peng, W. Q.; Han, L. Y. Energy Environ. Sci. 2014, 7, 2934.
(31) Zhao, Y. X.; Zhu, K. J. Phys. Chem. C 2014, 118, 9412.
(32) Jeon, N. J.; Noh, J. H.; Kim, Y. C.; Yang, W. S.; Ryu, S.; Il Seol,
S. Nat. Mater. 2014, 13, 897.
(57) O’Regan, B. C.; Bakker, K.; Kroeze, J.; Smit, H.; Sommeling, P.;
Durrant, J. R. J. Phys. Chem. B 2006, 110, 17155.
(58) Zhou, H. P.; Chen, Q.; Li, G.; Luo, S.; Song, T. B.; Duan, H. S.;
Hong, Z. R.; You, J. B.; Liu, Y. S.; Yang, Y. Science 2014, 345, 542.
(59) Xu, X. B.; Zhang, H.; Cao, K.; Cui, J.; Lu, J. F.; Zeng, X. W.;
Shen, Y.; Wang, M. K. ChemSusChem 2014, 7, 3088.
(60) Yin, W. J.; Shi, T. T.; Yan, Y. F. Appl. Phys. Lett. 2014, 104,
063903.
(61) Kim, J.; Lee, S. H.; Lee, J. H.; Hong, K. H. J. Phys. Chem. Lett.
2014, 5, 1312.
(62) Agiorgousis, M. L.; Sun, Y. Y.; Zeng, H.; Zhang, S. B. J. Am.
Chem. Soc. 2014, 136, 14570.
(63) In Group IV Elements, IV-IV and III-V Compounds. Part b -
Electronic, Transport, Optical and Other Properties; Madelung, O.,
Rossler, U., Schulz, M., Eds.; Springer: Berlin, 2002; Vol. 41A1b, p 1.
̈
(33) Im, J. H.; Jang, I. H.; Pellet, N.; Gratzel, M.; Park, N. G. Nat.
Nanotechnol. 2014, 9, 927.
(34) Eperon, G. E.; Burlakov, V. M.; Docampo, P.; Goriely, A.;
Snaith, H. J. Adv. Funct. Mater. 2014, 24, 151.
(64) Reichardt, C.; Welton, T. In Solvents and Solvent Effects in
Organic Chemistry; Wiley-VCH Verlag GmbH & Co. KGaA:
Weinheim, 2010; p 549.
(35) Zhang, S. J.; Audebert, P.; Wei, Y.; Al Choueiry, A.; Lanty, G.;
Brehier, A.; Galmiche, L.; Clavier, G.; Boissiere, C.; Lauret, J. S.;
Deleporte, E. Materials 2010, 3, 3385.
(36) Miyamae, H.; Numahata, Y.; Nagata, M. Chem. Lett. 1980, 9,
663.
(37) Shi, D.; Adinolfi, V.; Comin, R.; Yuan, M.; Alarousu, E.; Buin,
A.; Chen, Y.; Hoogland, S.; Rothenberger, A.; Katsiev, K.; Losovyj, Y.;
Zhang, X.; Dowben, P. A.; Mohammed, O. F.; Sargent, E. H.; Bakr, O.
M. Science 2015, 347, 519.
(38) Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Crystallogr.
2008, 64, 112.
(39) Nie, W. Y.; Tsai, H. H.; Asadpour, R.; Blancon, J. C.; Neukirch,
A. J.; Gupta, G.; Crochet, J. J.; Chhowalla, M.; Tretiak, S.; Alam, M. A.;
Wang, H. L.; Mohite, A. D. Science 2015, 347, 522.
H
DOI: 10.1021/jacs.5b06658
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX