View Article Online

View Journal

Journal of
Materials Chemistry A
Accepted Manuscript

This article can be cited before page numbers have been issued, to do this please use: N. Tripathi, M.
Yanagida, Y. Shirai, T. Masuda, L. Han and K. Miyano, J. Mater. Chem. A, 2015, DOI: 10.1039/C5TA01668A.

This is an Accepted Manuscript, which has been through the

Royal Society of Chemistry peer review process and has been
accepted for publication.

Accepted Manuscripts are published online shortly after

acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.

You can find more information about Accepted Manuscripts in the

Information for Authors.

Please note that technical editing may introduce minor changes

to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the Ethical guidelines still
apply. In no event shall the Royal Society of Chemistry be held
responsible for any errors or omissions in this Accepted Manuscript
or any consequences arising from the use of any information it
contains.

www.rsc.org/materialsA
 Page 1 of 17 Journal of Materials Chemistry A

View Article Online

Hysteresis-free and highly stable perovskite solar cells produced
DOI: 10.1039/C5TA01668A

via a chlorine-mediated interdiffusion method

Journal of Materials Chemistry A Accepted Manuscript

Neeti Tripathi,a Masatoshi Yanagida,ab Yasuhiro Shirai,*ab Takuya Masuda,a Liyuan Han,b
and Kenjiro Miyanoa
a
Global Research Center for Environment and Energy based on Nanomaterials Science
(GREEN), National Institute for Materials Science (NIMS), 1-1 Namiki, Tsukuba, Ibaraki
305-0044, Japan. *E-mail: [email protected]
b
Photovoltaic Materials Unit, National Institute for Materials Science (NIMS), 1-2-1 Sengen,
Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

Tsukuba, Ibaraki 305-0047, Japan

Table of contents entry

Perovskite solar cells fabricated using chlorine-mediated interdiffusion method showed
promising stability over 2 months with encapsulation and no immediate degradation under
continuous light exposure.

Abstract

Although lead-halide perovskite–based solar cells hold the promise of a breakthrough in the
production of next-generation photovoltaic devices, anomalous hysteresis in current–voltage
curves and inadequate stability remain as major challenges. Here, we demonstrate the
production of low-temperature solution–processed perovskite solar cells
(ITO/PEDOT:PSS/perovskite/PC61BM/Ca/Ag) with hysteresis-free current–voltage
characteristics, excellent photostability, and high reproducibility via the inclusion of methyl
ammonium chloride (MACl) using the interdiffusion method. The best-performing devices
exhibited a power conversion efficiency of over 12%. Our devices showed promising
stability by maintaining more than 90% of their initial performance over long periods of time
at ambient conditions with encapsulation using common techniques, as well as no obvious
degradation after 2 h of continuous light exposure. We statistically compared fabrication

1
 Journal of Materials Chemistry A Page 2 of 17

View Article Online

processes using the interdiffusion method with or without MACl by creating a histogram of
DOI: 10.1039/C5TA01668A

over 120 devices for each method. The results clearly indicated that including MACl gave

Journal of Materials Chemistry A Accepted Manuscript

better reproducibility and a higher average efficiency of 9.5%, as well as improved device
stabilities.

Introduction
Organic/inorganic hybrid solar cells based on metal-halide perovskite are considered
promising candidates for third-generation photovoltaics because of their efficient
Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

photovoltaic performance, simple processing conditions, and comparatively low cost.1-7

Kojima et al.2 published the first report of a perovskite-based dye-sensitized solar cell with a
power conversion efficiency (PCE) of 3.8%, but rapid progress in this area in the last 5 years
has now increased the PCE to more than 15%.3-8 Initially, state-of-the-art perovskite-based
devices used a metal-oxide porous scaffold (TiO2, Al2O3, and ZnO).4, 6, 9, 10 However, in later
reports, device processing was simplified by using a planar configuration to avoid the crucial
penetration of perovskite into the porous layers.3, 7 To date, different deposition techniques,
such as single step11 two-step4, and vapor deposition3 methods, have been employed to
fabricate these devices. Nevertheless, the use of oxide materials as electron-selective layers
generally requires high-temperature processing.
Another direction of growing research into perovskite solar devices involves low-
temperature solution–processable electron- and hole-transport materials, which have been
commonly used in organic photovoltaics (OPVs).5, 12-15 The first work on a perovskite solar
cell with structures similar to the OPV devices was reported by Jeng et al.,12 using [poly(3,4-
ethylenedioxythiophene):poly-(styrenesulfonic acid)] (PEDOT:PSS) as a hole-transporting
layer and phenyl-C61-butyric acid methyl ester (PC61BM) as an electron-selective layer in
planar heterojunction configuration. However, PCE was limited to 3.9% owing to the
difficulty in forming the uniform perovskite films on top of the PEDOT:PSS layer. Later,
Docampo et al.13 demonstrated improved PCE (up to 10%) by using an inverted device
structure on fluorine-doped tin oxide (FTO); they speculated that the smooth surface of the
indium tin oxide (ITO) substrate was unfavorable for uniform spreading of the perovskite
layer. Nevertheless, advancements in the development of ITO-based devices achieved
efficiencies of over 15% by using the interdiffusion process developed by Xiao et al.,5 which,
unlike the sequential dipping method, provides a homogenous layer of perovskite for planar
devices. Another report by Chiang et al. on the interdiffusion method16 demonstrated

2
 Page 3 of 17 Journal of Materials Chemistry A

View Article Online

efficiencies of up to 16.31%; this enhancement in efficiency was attributed to the better
DOI: 10.1039/C5TA01668A

electron mobility of PC71BM compared with that of PC61BM.

Journal of Materials Chemistry A Accepted Manuscript

Since the development of perovskite photovoltaic devices, a huge amount of attention
has been paid to increasing their efficiency. However, inadequate stability remains a central
challenge to the adoption of perovskite photovoltaics. Producing hysteresis-free devices with
sustainable performance seems to be essential for future progress in the development of
perovskite solar cells. Devices with hysteresis behaviors in J-V responses show a large
discrepancy between the forward and reverse scans, resulting in considerable errors in
Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

evaluating the solar cell efficiencies.7, 17, 18 For example, the results reported by Zhou et al.7
showed a large degree of hysteresis in device performance (16.79% efficiency in reverse scan
and 13.08% in forward scan), and device performance deteriorated after a very short time.
A study by Kim and Park19 showed that the degree of hysteresis depends strongly on
the perovskite crystal size and the presence of the mesoporous layer. Very recently, Unger et
al.20 presented a detailed study of the hysteresis effect and suggested that the transient
behavior of perovskite photovoltaic devices can be influenced by the stoichiometry of the
perovskite material, as well as by the degree of crystallinity and the size of the crystalline
domains. Although a wide number of fabrication methods4-6, 11 have been investigated, it is
difficult to find one optimized method that can ensure both high quality of the perovskite
absorber without anomalous hysteresis behavior and unperturbed device performance with
long-term stability.
Here, we demonstrate solar devices that were prepared by using a low-temperature-
solution process and have hysteresis-free current–voltage characteristics and excellent long-
term stability. Our devices show promising stability by maintaining more than 90% of their
initial performance over a long period of time at ambient conditions when encapsulated by
using a common method. Indeed, this degree of device stabilities allowed us to quantitatively
analyze the perovskite photovoltaic cells by using analytical models and impedance
spectroscopy (IS); the J-V measurements were perfectly consistent with a simple diode model,
and the IS in darkness and under illumination revealed a very slow process reminiscent of
hopping conductions.21 In this study, we compared the photovoltaic properties and stability of
two different lead-halide perovskite devices (i.e. MAPbI3 and MAPbI3-xClx) fabricated by
using the interdiffusion method.5, 16 Many studies have used mixed iodide/chloride in the
perovskite film formation to improve film qualities, and the role of the chlorine has been a
topic of much debate.22-25 We found that improved film morphology could be achieved by
adding a small amount of methyl ammonium chloride (MACl) to the methyl ammonium

3
 Journal of Materials Chemistry A Page 4 of 17

View Article Online

iodide (MAI) solution in the interdiffusion process to produce Cl-doped MAPbI3-xDOI:
Clx 10.1039/C5TA01668A
devices
instead of MAPbI3 ones. This combination of a doping approach with the interdiffusion

Journal of Materials Chemistry A Accepted Manuscript

method yields smoother perovskite films that have greater reproducibility and photostability.
By employing this method, we found that the best-performing device had a PCE of 12.3%
(device area 0.08 cm2). A statistical study of the fabrication of more than 240 devices with
and without MACl unambiguously revealed the advantage of Cl-doped MAPbI3-xClx devices.
We also report our detailed analysis of the hysteresis-free behaviors of our devices, as
evaluated by using hysteresis index approaches, and the results of stability testing under
Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

continuous light exposure (1 sun) for 2 h and in ambient conditions over 2 months.

Experimental section
Materials and characterization
All chemicals were purchased from commercial suppliers and used as received unless
stated otherwise. PbI2 (Sigma Aldrich, 99%) and methyl ammonium iodide (MAI) (Wako
Chem., 98%) were dissolved in anhydrous N, N-dimethylformamide (Wako chem.,
99.5%)(400 mg-ml-1) and anhydrous 2-propanol (Wako chem.)(50 mg-ml-1), respectively. For
chlorine doping, 5-20 wt% MACl (Wako chem.) was mixed with the MAI solution. PC61BM
(Solenne or Lumitec, 99%) solution (2 wt%) dissolved in anhydrous chlorobenzene (Wako
chem., 99%) was used for coating of the electron selective layer. All the solutions were
filtered using 0.45 µm syringe filters to avoid the risk of particle formation. UV-curable
resins (NagaseChemtex XNR5516Z) were used to seal devices with cavity glasses. The x-ray
diffraction patterns of MAPbI3 and MAPbI3-xClx were collected using Bruker D8 advanced x-
ray diffractometer (CuKα radiation, λ= 1.54050 A). The top surface and cross-sectional
images were taken by a high resolution scanning electron microscope at 5 kV accelerating
voltage (Hitachi-4800). Additionally, elemental mapping was done using energy dispersive x-
ray measurement mode at accelerating voltage of 10 kV. The current density–voltage (J-V)
characteristics under 1 sun with an AM 1.5G spectral filter (100 mW/cm2, calibrated with a Si
reference cell, emission spectrum in Figure S9) and incident monochromatic photon to
current conversion efficiency (IPCE) spectra or External quantum efficiency (EQE) were
measured with a CEP-200BX spectrometer (Bunkokeiki, Tokyo, Japan) at room temperature.
To investigate the hysteresis behaviour scan rates were varied from 0.02 to 0.3 V/s.
Histogram of over 240 devices were generated from the devices using 0.18 cm2 aperture
mask.

4
 Page 5 of 17 Journal of Materials Chemistry A

View Article Online

Device fabrication DOI: 10.1039/C5TA01668A

Solar cell devices were fabricated on pre-cleaned patterned indium tin oxide (ITO)

Journal of Materials Chemistry A Accepted Manuscript

coated glass substrates (15 Ω square–1). The ITO substrates were pre-cleaned in an ultrasonic
bath with detergent, pure water, and 2-propanol, and then treated in an ultraviolet-ozone
chamber for at least 10 min to remove the organic residuals. A thin layer (~30 nm) of
PEDOT:PSS (Clevios, AI4083) was formed by spin coating at 3000 rpm and subsequently
dried at 140˚C for 10 min on a hot plate under air. Substrates were transferred into a nitrogen-
filled glove box (<1.0 ppm O2 and H2O) for PbI2 and MAI coatings. Firstly, PbI2 was spin
Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

coated on to the top of PEDOT:PSS layer at 3000 rpm for about 30 sec. After drying at 70 ˚C
for 30 min, MAI solutions with different MACl weight percentages were spun on to the PbI2
layer at 4000 rpm, 30 sec. These PbI2/MAI samples were heated at 100˚C for 2 hours for the
complete conversion into perovskite. Finally, a PC61BM layer was spun on the perovskite
layer at 700 rpm for 60 s. Samples were then transferred into the evaporation chamber,
connected to the glove box for metal contact deposition. A 4 nm thin layer of Ca and 100 nm
of Ag were deposited by thermal evaporation at a pressure <10-4 Pa. This process provided
devices with 4.0 × 7.5 mm2, and they are sealed using the UV-curable resin before the
subsequent measurements in ambient conditions with an aperture mask of 0.08 to 0.18 cm2.

Results and discussion

Figure 1 is a simplified energy-band diagram of the solar cells used in this work. The
perovskite layer was deposited by using the interdiffusion method, and then the PC61BM
layer was deposited via a spin-coating method, instead of the vacuum evaporation technique
used previously;5 thus a more attractive solution-based fabrication process was used. An
ultra-thin layer of calcium (Ca) in conjugation with silver (Ag) was evaporated as the back
contact to complete the device structure. The combination of low-work-function Ca and Ag
was used to obtain better ohmic contacts at the cathode interface.26

5
 Journal of Materials Chemistry A Page 6 of 17

View Article Online

Figure 1. Simplified energy-band diagram of the low-temperature solution–deposited planar
DOI: 10.1039/C5TA01668A
perovskite solar cell device.

Journal of Materials Chemistry A Accepted Manuscript

To investigate the structural evolution of the device, XRD patterns of MAPbI3 and
MAPbI3-xClx films were recorded before annealing and after 10 min and 2 h of annealing at
100 °C (Fig. 2). For comparison, all the diffraction patterns were plotted at the same intensity
scale. The main diffraction peaks at 2 θ positions 14.04° and 28.42° were assigned to the
(110) and (220) diffraction planes of MAPbI3,3, 12, 27 whereas the broad hump at about
Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

20°corresponded to the ITO/glass substrate. Interestingly, before annealing, MAPbI3 films

had almost negligible features, whereas a prominent peak corresponding to the (220) plane of
the tetragonal perovskite structure was clearly apparent in the case of MAPbI3-xClx films.
Before the annealing, the presence of a small peak of PbI2 on both films suggested
incomplete reactions. After annealing at 100 °C for 10 min, samples of both MAPbI3 and
MAPbI3-xClx gave identical XRD patterns, although the patterns differed slightly in intensity
scale. When we increased the annealing time to 2 h there was no further improvement in the
case of the MAPbI3 film, whereas the XRD pattern of the MAPbI3-xClx film showed a marked
increase in intensity of the (110) and (220) planes.
The increase in the intensity ratio of the (110) to (220) planes in the case of MAPbI3-
xClx—from 2.25 to 2.62—after 2 h of annealing is a possible indication of chlorine ion–
assisted slow growth along the preferred (110) orientation. The crystallite sizes (L) of the
MAPbI3 and MAPbI3-xClx films were estimated by using the Scherrer formula,28 given by L=
(0.9λ)/(βcosθ), where λ = 0.154 nm for CuKα radiation and β is the full width at half
maximum of the diffraction peak, in radians. Crystallite sizes for MAPbI3-xClx (95 nm) were
greater than those MAPbI3 (70 nm).

6
 Page 7 of 17 Journal of Materials Chemistry A

View Article Online

DOI: 10.1039/C5TA01668A

Journal of Materials Chemistry A Accepted Manuscript

Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

Figure 2. XRD patterns of perovskite films before annealing and after annealing for different
periods of time: (a) to (c), MAPbI3;(d) to (f), MAPbI3-xClx.

These different evolutions of the XRD patterns of the MAPbI3 and MAPbI3-xClx films
suggested that there were two distinct reaction-crystallization processes. In the case of
MAPbI3, the reaction was barely initiated before annealing, whereas within 10 min of the
start of annealing the reaction crystallization had proceeded rapidly and was complete, with
no further changes between then and 2 h. On the other hand, in the case of MAPbI3-xClx, the
reaction started before annealing, but after the initial reaction the process slowed and
completed within 2 h. For the one-step coating process with PbCl2 and MAI on PEDOT:PSS
surfaces, the existence of chlorine have shown striking effect on the microstructure and
orientation of perovskite films through the templated topotactic self-assemblies.23 Althouogh
large differences can be expected between the one-step method and the interdiffusion method
employed in this study, the presence of chlorine can effectively alter the course of
crystallization processes and may improve the orientational order and degree of crystallinity.
The increased crystal size in MAPI3-xClx can reduce charge trapping and recombination at
grain boundaries, thereby improving the fill factor (FF) owing to better film formation and
hence a reduction in the number of shunting paths.29 Noticeably, the absence of any
intermediate impurity phases even before annealing suggests that a chlorine-assisted
interdiffusion method is capable of producing high-quality MAPbI3Cl3-x films. Another point
is that continuous and smooth perovskite films were obtained on the crystalline PbI2 films
(Figure S1). Therefore, the extra effort of preparing an amorphous or conformal PbI2 layer,

7
 Journal of Materials Chemistry A Page 8 of 17

View Article Online

which generally requires the PbI2 solution or ITO substrate, or both, to be heated during the
DOI: 10.1039/C5TA01668A

deposition process,5 can be avoided by using this method.

Journal of Materials Chemistry A Accepted Manuscript

We carefully inspected the morphological changes in the MAPbI3 and MAPbI3-xClx
perovskite films by using scanning electron microscopy (SEM) before and after PC61BM
coating (Figure 3). Top and cross-sectional views of the MAPbI3 layer without the PC61BM
coating are shown in Figure 3 (a) and (b), respectively. Large variations in crystallite size
with islands of sub-micron size can be seen clearly all over the surface (Figure 3 (a)).
Although the average thickness of the perovskite layer seems to be uniform, the presence of
Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

these random structures on the surface (marked by circles, Figure 3 (b)) increases the
possibility of current leakage occurring through piercing of the thin PC61BM layer to the
cathode. In contrast, we were able to produce very smooth, continuous, and pin-hole-free
perovskite films by using the interdiffusion method with a mixture of both MACl and MAI,
as seen in the top and cross-sectional views (Figure 3 (c) and (d), respectively). It was
possible that the rough features on top of the perovskite layer could be smoothened by being
covered with the slightly thicker PC61BM layer. However, this was not the case with MAPbI3,
as the sharp features on the perovskite surface were still visible even after the addition of a
70-nm-thick PC61BM layer (Figure 3 (e)). As expected, MAPbI3-xClx gave smooth films both
before and after PC61BM coating (Figure 3 (f)). These random morphological and crystallite
size features can affect the device’s performance. By performing electron beam–induced
current profiling, Edri et al.30 found a uniform current distribution throughout MAPbI3-xClx–
based solar cells but a grain-dependent current distribution in MAPbI3 cells. The enhanced Jsc
value of MAPbI3-xClx that we found here was in good agreement with these previous
observations.

8
 Page 9 of 17 Journal of Materials Chemistry A

View Article Online

Figure 3. SEM images (a) and (b): top and cross-sectional views of MAPbI3 without
DOI: 10.1039/C5TA01668A
PC61BM coating; (c) and (d): top and cross-sectional views of MAPbI3-xClx without PC61BM
coating; (e) cross-sectional view of sample MAPbI3 with PC61BM coating [inset shows top

Journal of Materials Chemistry A Accepted Manuscript

view]; (f) cross-sectional view of MAPbI3-xClx with PC61BM coating [inset shows top view].

Solution-processed perovskite solar cells are highly sensitive to crystallization

conditions such as the choice of precursors, concentrations, solvents, annealing temperatures,
and annealing times. A small variation in structural and morphological properties can lead to
Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

a huge step-down in device performance. In our experience, without the MACl, it was
difficult to obtain better control of film formation even though the synthesis process was
optimized by adjusting several parameters such as the precursor concentration, spin speed,
solution dispensing method (static or dynamic dispensing of spin coatings), and annealing
temperature or duration (or both). Although comparable PCEs were achieved in some
attempts, the reproducibility of the MAPbI3 devices was poorer than that of the Cl-doped
MAPbI3-xClx devices. The Cl-doped perovskite films showed consistently improved
crystalline structure and morphology, irrespective of small variations in PbI2 or MAI
concentrations and spin-coating methods. A cross-sectional view of a MAPbI3-xClx film, with
the estimated thickness of each layer, is shown in Figure 4 (a). The thicknesses of the
PEDOT:PSS, MAPbI3-xClx, and PC61BM layers were ~30 nm, ~245 nm, and ~70 nm,
respectively.

Figure 4. (a) Cross-sectional view and (b) elemental mapping of MAPbI3-xClx.

We also evaluated the amounts of chlorine atoms left in the MAPbI3-xClx films by
measuring the ratio of chlorine to iodine using x-ray photoemission spectroscopy (XPS). The
amount of chlorine, however, seemed to be below the sensitivity limit of the instrument. We

9
 Journal of Materials Chemistry A Page 10 of 17

also conducted a XPS depth-profiling analysis by etching the film surface step by DOI: upView
to Article
a Online
step10.1039/C5TA01668A
depth of ~50 nm (Figure S2); a small hump was observed at the Cl-2p core level position.

Journal of Materials Chemistry A Accepted Manuscript

Although the inherently weak signal from chlorine compared with that of iodine meant that
we had insufficient reliable information, we concluded that the chlorine content of the bulk
film was likely of the order of 1% and that, on the basis of the depth profiles, the chlorine
was not segregated on the surface. Considering the results of other studies, however, it
seemed reasonable to have such low chlorine concentrations in our devices because only
small amounts of chlorine (in the range of 2% to 4%) have been detected, even in films
Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

prepared from precursors with much greater chlorine contents (MAI:PbCl2, 3:1).31, 32
Furthermore, we made another attempt at elemental mapping by using energy-dispersive x-
ray (EDX) measurement (Figure 4 (b)). Traces of chlorine signals were observed in the
MACl-doped samples, and the chlorine signal consistently increased with increasing MACl
concentration in the MAI solution (Supplementary information Table S1). Note that only
qualitative information can be obtained from EDX analyses. Nevertheless, it was clear that
the amount of chlorine incorporated into the perovskite structure was far less than the
nominal composition of up to 20% MACl mixed in the MAI solution that was spread on the
surface. Indeed, recent studies clearly revealed sublimation and loss of chlorine contents
under mild annealing conditions below 100 ºC.22
In response to the very different XRD results for the non-annealed MAPbI3 and
MAPbI3-xClx films (see Figure 2 and the discussions thereafter), we compared the
morphologies of MAPbI3 and MAPbI3-xClx before annealing (Figure S3). MAPbI3 perovskite
films have a very rough surface and non-uniform thickness, possibly because of the presence
of unreacted MAI or PbI2 on the surface. In contrast, the highly compact and smooth nature
of the layer in the Cl–doped perovskite films implied that adding a small amount of MACl to
the MAI solution had a huge impact on the structure and morphology of the non-annealed
perovskite layer. SEM comparison of the features of PbI2, PbI2/MAPbI3, and PbI2/MAPbI3-
xClx films revealed the formation of a continuous film with a morphology resembling that of
the underneath PbI2 layer in the case of Cl-doped perovskite (Figure S4); this was in contrast
to the morphology of the MAPbI3 perovskite film, with its sub-micron-size random features.
It is well known that the crystallinity and morphology of the perovskite layer
determine the ultimate performance of the device, as non-uniformity causes current leakage
and defects cause the development of trapping sites for charge recombination. Figure 5(a)
shows the J-V curves for MAPbI3 and MAPbI3-xClx devices (device size, 0.18 cm2) prepared

10
 Page 11 of 17 Journal of Materials Chemistry A

View Article Online

in the same batch. We observed a clear correlation between morphology and the device
DOI: 10.1039/C5TA01668A

parameters, as evident from the substantial improvement in photovoltaic parameters in

Journal of Materials Chemistry A Accepted Manuscript

MAPbI3-xClx devices. Both interdiffusion methods—with or without the chlorine doping—
worked well, and in both cases the PCEs reached almost 12% in our hands. These results
paralleled those of original reports on the interdiffusion method using vapor-deposited
fullerenes (PCE 15.4%) or spin-coated PC61BM (PCE 9.92%).16 However, we found that,
generally, devices prepared by using the Cl-doped interdiffusion method had better
performances. We obtained statistical insight into the process of fabrication using the
Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

MAPbI3-xClx and MAPbI3 films by creating a histogram of the PCEs of 120 devices for each
film type (Figure 5(b)). The results clearly indicated that the MAPbI3-xClx devices gave better
reproducibility, with a higher average efficiency of 9.5%, whereas the MAPbI3 devices had a
broad and lower distribution of efficiencies. Furthermore, detailed analysis of the J-V curves
for each device revealed, as a general trend, a decrease in series resistance and an increase in
shunt resistance in the case of the MAPbI3-xClx devices. As a result, a marked improvement
in FF was observed in the case of MAPbI3-xClx devices; this was ascribable to the relatively
small number of grain boundaries and lack of leakage paths in the homogenous perovskite
films. The improvement in FF could also be related to the high carrier mobility of chlorine-
doped devices.1, 33 The IPCE spectra for the devices were also compared (Figure 5 (c)).
MAPbI3-xClx devices had IPCEs slightly superior to those of MAPbI3 devices. The IPCE
peaked at about 75% at wavelengths of 430 to 540 nm and 670 nm, consistent with other
reports.33, 34 Integrated Jsc values of 16.90 and 17.58 mA cm–2 were obtained for MAPbI3-xClx
and MAPbI3 , respectively; these were in good agreement with the photocurrent densities
obtained from the J-V curves.

(b)

11
 Journal of Materials Chemistry A Page 12 of 17

View Article Online

DOI: 10.1039/C5TA01668A
(c)

Journal of Materials Chemistry A Accepted Manuscript

Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

Figure 5. (a) J-V characteristics prepared in the same batch and (b) Efficiency histogram of
120 devices for each method [device size 0.18 cm2] (c) IPCE spectra of MAPbI3 and
MAPbI3-xClx devices prepared in the same batch.

We also verified the dependence of PCE on perovskite device size. It is generally

more convenient (but less reliable) to achieve higher efficiency in smaller devices, as one can
better avoid defects or pinholes using smaller devices, which also have fewer conductivity
problems than with the larger transparent electrodes in larger devices. However, there are
more errors in the smaller apertures used with smaller devices. To compare our devices (0.18
cm2) with those in other published works using smaller cells (0.04 to 0.1 cm2), the
performance of the champion cell was recorded with different apertures ranging from 0.08 to
0.18 cm2 (Figure S5). Device performances as a function of aperture size are summarized in
Table S2. The slight variation in Jsc and/or Voc with different aperture sizes may reflect
inhomogeneity of the films at a microscopic scale. As expected, smaller cells showed
increased performance, with up to 12.25% efficiency. Although our efficiencies are not
among the best reported, notably, we were able to obtain high FFs (above 0.7) for more than
half of the devices fabricated at a size of 0.18 cm2 (Figure S6).
Although substantial attention has been paid to achieving high efficiency, anomalous
J-V hysteresis behavior has been not fully disclosed on many occasions in earlier reports.
Device performance can depend markedly on voltage sweep rate or direction, or both. To
investigate the hysteresis behavior of our MAPbI3-xClx devices, we measured their J-V
characteristics at different scan rates from 0.02 to 0.3 V s–1 in the forward (–0.2 to 1.2 V) and
reverse (1.2 to –0.2 V) directions (Figure 6). The origin of hysteresis in perovskite-absorber
devices has been assigned to the slow dynamic processes originating from the trapping and
detrapping of charge carriers, as well as to changes in absorber or contact conductivity,
ferroelectricity, and ion migration.17-20 All of these effects appear to be related to the

12
 Page 13 of 17 Journal of Materials Chemistry A

View Article Online

structural quality of the perovskite layer. The absence of hysteresis in our devices DOI:
suggests
10.1039/C5TA01668A

that the fabrication method presented here is good enough to prevent such anomalous

Journal of Materials Chemistry A Accepted Manuscript

behaviors in devices. To obtain more quantitative information on J-V hysteresis, we
estimated a hysteresis index (HI) by using the equation19

𝐽𝐽𝑅𝑅𝑅𝑅 (0.8 𝑉𝑉𝑜𝑜𝑜𝑜 ) − 𝐽𝐽𝐹𝐹𝐹𝐹 (0.8 𝑉𝑉𝑜𝑜𝑜𝑜 )

𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻𝐻 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 =
𝐽𝐽𝑅𝑅𝑅𝑅 (0.8 𝑉𝑉𝑜𝑜𝑜𝑜 )
Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

where JRS (0.8 Voc) and JFS (0.8Voc) represent the photocurrent density at 80% of Voc for the
reverse scan (RS) and forward scan (FS), respectively. A histogram of 80 devices (Figure S7)
revealed that one-third of our devices had HI values below 0.1; this was a far better result
than in other reports. In fact, this degree of near-absence of hysteresis allowed us to
quantitatively analyze the perovskite photovoltaic cells by using analytical models and
impedance spectroscopy (IS).21
We further examined the photostability of our devices under illumination (1 sun)
(Figure 7 (a)). Devices were operated alternately between short and open circuit conditions
under continuous illumination. No degradation of the MAPbI3-xClx devices was found during
2 h of operation, revealing a high degree of stability. In contrast, the MAPbI3 devices showed
a gradual decrease in Jsc in the same photostability testing. Similarly, higher stability for
devices using PbCl2 as lead source, instead of non-Cl containing Pb(OAc)2, has been
reported,35 which is another example of the positive effect gained by the inclusion of chlorine
species. More interestingly (Figure S8), hysteresis-free J-V behavior lasted for more than 1
month in the case of MAPbI3-xClx devices, revealing the high degree of stability of perovskite
solar cells prepared by using the Cl-doped interdiffusion method.

13
 Journal of Materials Chemistry A Page 14 of 17

View Article Online

DOI: 10.1039/C5TA01668A

Journal of Materials Chemistry A Accepted Manuscript

Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

Figure 6. J-V curves of MAPbI3-xClx devices with different scan rates.

Halide perovskites can be very sensitive to moisture and easily decomposed because
of the highly hygroscopic nature of the amine salts.36 Therefore, the stability of perovskite
photovoltaic devices is usually poor under ambient atmosphere, limiting the use of these
devices in practical applications without encapsulation. We investigated the aging of our
devices stored under ambient atmosphere by using a common method of encapsulation with
cavity glass and UV-curable resin. The PCE of MAPbI3-xClx devices retained 90% of its
initial value for almost 2 months in some cases (Figure 7(b)). Possible mechanisms behind
the deterioration after 2 months are degradation at the PEDOOT:PSS–perovskite interface as
a result of the acidic or hygroscopic nature of PEDOT:PSS, or oxidation of the thin Ca layers
at the cathode contact. Further optimization of interface materials will be needed to replace
these materials and produce perovskite photovoltaic devices with practical durability.

14
 Page 15 of 17 Journal of Materials Chemistry A

View Article Online

DOI: 10.1039/C5TA01668A
(a)

Journal of Materials Chemistry A Accepted Manuscript

Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

Figure 7. (a) Comparison of device performances under continuous illumination (1 sun). (b)
Long-term stability of MAPbICl3-xClx devices stored under ambient atmosphere [device size
0.18 cm2].

In summary, we demonstrated low-temperature solution–processed perovskite

photovoltaic devices with hysteresis-free J-V characteristics and excellent long-term stability.
Our devices showed promising stability by maintaining more than 90% of their initial
performance over 2 months at ambient conditions when we used a common method of
encapsulation in cavity glass and UV-curable resin. We compared the photovoltaic properties
and stabilities of two different lead-halide perovskite devices (i.e. MAPbI3 and MAPbI3-xClx)
fabricated (size 0.18 cm2) by using the interdiffusion method. Inclusion of chlorine in the
interdiffusion method effectively enhanced the stability and reproducibility of the devices in
comparison with those of MAPbI3 devices. The key factors that improved the performance of
MAPbI3-xClx devices were better crystallinity and homogeneity and pin hole–free coverage of
the perovskite films, as revealed by XRD and SEM observations. Further improvements in
device performance are anticipated through rational interface engineering.
15
 Journal of Materials Chemistry A Page 16 of 17

View Article Online

DOI: 10.1039/C5TA01668A

Ackknowledgement

Journal of Materials Chemistry A Accepted Manuscript

The authors acknowledge the support under the MEXT Program for Development of
Environmental Technology using Nanotechnology.

Supplementary Information (ESI) available: XPS data for MAPbI3 and MAPbI3-xClx, SEM
images for MAPbI3-xClx, and PbI2, FF histogram and hysteresis index values.
Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

REFERENCES

1. P. Gao, M. Gratzel and M. K. Nazeeruddin, Energ Environ Sci, 2014, 7, 2448-2463.

2. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Chem. Soc., 2009, 131, 6050-6051.
3. M. Z. Liu, M. B. Johnston and H. J. Snaith, Nature, 2013, 501, 395-398.
4. J. Burschka, N. Pellet, S. J. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin and M.
Gratzel, Nature, 2013, 499, 316-319.
5. Z. Xiao, C. Bi, Y. Shao, Q. Dong, Q. Wang, Y. Yuan, C. Wang, Y. Gao and J. Huang, Energ
Environ Sci, 2014, 7, 2619-2623.
6. J. H. Im, I. H. Jang, N. Pellet, M. Gratzel and N. G. Park, Nature Nanotechnology, 2014, 9, 927-
932.
7. H. Zhou, Q. Chen, G. Li, S. Luo, T.-b. Song, H.-S. Duan, Z. Hong, J. You, Y. Liu and Y. Yang,
Science, 2014, 345, 542-546.
8. Q. Lin, A. Armin, R. C. R. Nagiri, P. L. Burn and P. Meredith, Nat Photon, 2015, 9, 106-112.
9. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Science, 2012, 338, 643-
647.
10. D. Y. Liu and T. L. Kelly, Nature Photonics, 2014, 8, 133-138.
11. J.-H. Im, H.-S. Kim and N.-G. Park, APL Mater., 2014, 2, 081510.
12. J. Y. Jeng, Y. F. Chiang, M. H. Lee, S. R. Peng, T. F. Guo, P. Chen and T. C. Wen, Adv. Mater.,
2013, 25, 3727-3732.
13. A. Abrusci, S. D. Stranks, P. Docampo, H. L. Yip, A. K. Y. Jen and H. J. Snaith, Nano Lett., 2013,
13, 3124-3128.
14. J. You, Z. Hong, Y. Yang, Q. Chen, M. Cai, T.-B. Song, C.-C. Chen, S. Lu, Y. Liu, H. Zhou and Y.
Yang, ACS Nano, 2014, 8, 1674-1680.
15. O. Malinkiewicz, A. Yella, Y. H. Lee, G. M. Espallargas, M. Graetzel, M. K. Nazeeruddin and H.
J. Bolink, Nature Photonics, 2014, 8, 128-132.
16. C.-H. Chiang, Z.-L. Tseng and C.-G. Wu, Journal of Materials Chemistry A, 2014, 2, 15897-
15903.
17. H. J. Snaith, A. Abate, J. M. Ball, G. E. Eperon, T. Leijtens, N. K. Noel, S. D. Stranks, J. T. W.
Wang, K. Wojciechowski and W. Zhang, J Phys Chem Lett, 2014, 5, 1511-1515.
18. R. S. Sanchez, V. Gonzalez-Pedro, J. W. Lee, N. G. Park, Y. S. Kang, I. Mora-Sero and J.
Bisquert, J Phys Chem Lett, 2014, 5, 2357-2363.
19. H. S. Kim and N. G. Park, J Phys Chem Lett, 2014, 5, 2927-2934.
20. E. L. Unger, E. T. Hoke, C. D. Bailie, W. H. Nguyen, A. R. Bowring, T. Heumuller, M. G.
Christoforo and M. D. McGehee, Energ Environ Sci, 2014, 7, 3690-3698.
21. K. Miyano, M. Yanagida, N. Tripathi and Y. Shirai, Appl. Phys. Lett., 2015.
22. E. L. Unger, A. R. Bowring, C. J. Tassone, V. L. Pool, A. Gold-Parker, R. Cheacharoen, K. H.
Stone, E. T. Hoke, M. F. Toney and M. D. McGehee, Chem. Mater., 2014, 26, 7158-7165.

16
 Page 17 of 17 Journal of Materials Chemistry A

23. S. T. Williams, F. Zuo, C. C. Chueh, C. Y. Liao, P. W. Liang and A. K. Y. Jen, ACS Nano, 8,Article Online
2014,View
DOI: 10.1039/C5TA01668A
10640-10654.
24. M. I. Dar, N. Arora, P. Gao, S. Ahmad, M. Grätzel and M. K. Nazeeruddin, Nano Lett., 2014, 14,

Journal of Materials Chemistry A Accepted Manuscript

6991-6996.
25. S. Colella, E. Mosconi, G. Pellegrino, A. Alberti, V. L. P. Guerra, S. Masi, A. Listorti, A. Rizzo, G.
G. Condorelli, F. De Angelis and G. Gigli, The Journal of Physical Chemistry Letters, 2014, 5,
3532-3538.
26. R. Steim, F. R. Kogler and C. J. Brabec, J. Mater. Chem., 2010, 20, 2499-2512.
27. M. Jiang, J. Wu, F. Lan, Q. Tao, D. Gao and G. Li, Journal of Materials Chemistry A, 2015, 3,
963-967.
28. A. L. Patterson, Physical Review, 1939, 56, 978-982.
29. Y. X. Zhao and K. Zhu, J Phys Chem C, 2014, 118, 9412-9418.
Published on 06 May 2015. Downloaded by Freie Universitaet Berlin on 07/05/2015 17:17:39.

30. E. Edri, S. Kirmayer, A. Henning, S. Mukhopadhyay, K. Gartsman, Y. Rosenwaks, G. Hodes and

D. Cahen, Nano Lett., 2014, 14, 1000-1004.
31. J. M. Ball, M. M. Lee, A. Hey and H. J. Snaith, Energ Environ Sci, 2013, 6, 1739-1743.
32. S. Colella, E. Mosconi, P. Fedeli, A. Listorti, F. Gazza, F. Orlandi, P. Ferro, T. Besagni, A. Rizzo,
G. Calestani, G. Gigli, F. De Angelis and R. Mosca, Chem. Mater., 2013, 25, 4613-4618.
33. J. H. Noh, S. H. Im, J. H. Heo, T. N. Mandal and S. I. Seok, Nano Lett., 2013, 13, 1764-1769.
34. P. Docampo, F. C. Hanusch, S. D. Stranks, M. Doblinger, J. M. Feckl, M. Ehrensperger, N. K.
Minar, M. B. Johnston, H. J. Snaith and T. Bein, Advanced Energy Materials, 2014, 4, 6.
35. F. K. Aldibaja, L. Badia, E. Mas-Marza, R. S. Sanchez, E. M. Barea and I. Mora-Sero, Journal of
Materials Chemistry A, 2015, 3, 9194-9200.
36. B. Suarez, V. Gonzalez-Pedro, T. S. Ripolles, R. S. Sanchez, L. Otero and I. Mora-Sero, J Phys
Chem Lett, 2014, 5, 1628-1635.

17