Energy &

Environmental
Science
View Article Online
REVIEW View Journal

Organometal halide perovskite solar cells:

Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

degradation and stability

Cite this: DOI: 10.1039/c5ee02733k
Taame Abraha Berhe,a Wei-Nien Su,*a Ching-Hsiang Chen,a Chun-Jern Pan,b
Ju-Hsiang Cheng,b Hung-Ming Chen,b Meng-Che Tsai,b Liang-Yih Chen,b
Amare Aregahegn Dubaleb and Bing-Joe Hwang*bc

Organometal halide perovskite solar cells have evolved in an exponential manner in the two key areas of
efficiency and stability. The power conversion efficiency (PCE) reached 20.1% late last year. The key disquiet
was stability, which has been limiting practical application, but now the state of the art is promising, being
measured in thousands of hours. These improvements have been achieved through the application of
different materials, interfaces and device architecture optimizations, especially after the investigation of hole
conductor free mesoporous devices incorporating carbon electrodes, which promise stable, low cost
and easy device fabrication methods. However, this work is still far from complete. There are various
issues associated with the degradation of Omh-perovskite, and the interface and device instability which
must be addressed to achieve good reproducibility and long lifetimes for Omh-PSCs with high conversion
efficiencies. A comprehensive understanding of these issues is required to achieve breakthroughs in
Received 4th September 2015, stability and practical outdoor applications of Omh-PSCs. For successful small and large scale
Accepted 20th October 2015 applications, besides the improvement of the PCE, the stability of Omh-PSCs has to be improved. The
DOI: 10.1039/c5ee02733k causes of failure and associated mechanisms of device degradation, followed by the origins of degradation,
approaches to improve stability, and methods and protocols are discussed in detail and form the main
www.rsc.org/ees focus of this review article.

Broader context
Low cost, stable and efficient light harvesting crystalline materials have great promise for future cost-effective clean energy and for the environment. Solar cell
devices made of organometal halide perovskite materials are now revolutionizing the area of photovoltaics with energy conversion efficiencies of 20.1%, which
represents a fivefold jump in cell efficiency within three years. The major obstacle of bringing organometal halide perovskite solar cells to the commercial
market is the poor long term device stability, which has not been able to exceed a few thousands of hours until now. This review article presents an in-depth and
updated discussion on the progress in stability, addressing the mechanisms and origins behind degradation, approaches to improve stability, protocols and
related analytical methods to study the existing and other unknown limitations on the stability of organometal halide perovskite solar cells. Furthermore,
the review is also meant to give an overview to readers who are new to this topic.

1. Background but the evaluation and prediction of the device lifetime are not
easy as the stability is related to the fading process of individual
To become economically feasible, organometal halide perovskite solar cells.1 Because different varieties of solar devices have
solar cells (Omh-PSCs) should remain free from degradation various degradation mechanisms, the accelerated stress tests
under normal operating conditions over several years, preferably appropriate for one type of solar cell may not be useful in
tens of years. Their stability is vital for practical application, studying the degradation of other types. Organometal halide
(Omh) perovskites are a class of materials in the perovskite
a
NanoElectrochemistry Laboratory, Graduate Institute of Applied Science and family with alternating layered structures of organic and inorganic
Technology, National Taiwan University of Science and Technology, Taipei 106, constituents. Omh-perovskites are noted for their cheap produc-
Taiwan. E-mail: [email protected]
b
tion, and methylammonium lead iodide (CH3NH3PbI3) is one of
NanoElectrochemistry Laboratory, Department of Chemical Engineering,
National Taiwan University of Science and Technology, Taipei 106, Taiwan.
the most representative materials. Solar cells made of this kind of
E-mail: [email protected] material have grown so fast within a few years and have reached a
c
National Synchrotron Radiation Research Center, Hsin-Chu, 30076, Taiwan certified efficiency over 20%. Its theoretical maximum is greater

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

than 30% and this could beat the efficiency of silicon, the without encapsulation is 2000 h at room temperature, in air and
theoretical maximum of which is limited to 27%, and economically in the dark.7 The high temperature (80 1C) dark test may be too
and practically limited to B25%.2 demanding for these solar cells as it has not yet been demon-
However, the long term stability of this kind of solar cell is far strated. It is obvious that at high temperatures, i.e. above 60 1C,
from practical. Ex situ long-term stability tests for CH3NH3PbI3 either certain fading mechanisms are accelerated or new degra-
solid-state solar cells stored in air at room temperature without dation reactions occur, causing rapid absorber material and/or
encapsulation were carried out for over 500 h under one sun device degradation. The major causes responsible for this kind of
illumination3 and 1000 h under full sunlight.4 This revealed that degradation may be related to various possible external factors,
light soaking alone is not a major reason for this poor stability. such as moisture, oxygen, temperature, UV light, etc., as well as
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

Furthermore, long-term stability of Omh-PSCs for 500 h at 45 1C, internal intrinsic factors such as ion migration, electro-migrations
under constant illumination5 at approximately 100 mW cm2 and interfacial reactions. Thus, setting standard stability testing
and 1000 h at 40 1C, with continuous exposure and operation protocols requires an understanding of how and why the device
under full spectrum simulated sunlight6 has also been achieved. degrades. Such an understanding should facilitate the future
The light-soaking tests become more challenging as the tem- production of better materials for the cells.
perature increases; in fact, a light soaking test Omh-PSCs has not In order to establish typical stability testing protocols for
yet been reported for 1000 h at 80 1C. The device stability of Omh-PSCs, stability definitions should be well understood.
mesoporous TiO2/CH3NH3PbI3/C heterojunction (HJ) solar cells For example, what range of drop or raise in the performance

Taame Abraha is a PhD student Wei-Nien Su is currently an

in the Graduate Institute of assistant professor of the
Applied Science and Technology Graduate Institute of Applied
at National Taiwan University of Science and Technology at
Science and Technology, Taiwan. National Taiwan University of
He received his Bachelor in Science and Technology
Chemistry from Wollo University (NTUST), Taiwan. His research
in 2009 and master’s degree in interests are rooted in materials
Chemistry (Physical) from Bahir synthesis and characterization
Dar University, Ethiopia in 2011. of nanocatalysts for various
Then, he joined Adigrat Univer- electrochemical devices and
sity and worked as head of energy applications. Prof. Su is
Taame Abraha Berhe chemistry and Dean, College of Wei-Nien Su currently involved in various
Natural and Computational international and national
Sciences till 2014. Currently, he is conducting his PhD under research programs. He received his PhD from the Wolfson School,
Prof. Bing-Joe Hwang and Prof. Wei-Nien Su, focusing on the Loughborough University (UK) and his Diplom-Ing. in Chemical
synthesis and stability of organometal halide perovskite solar Engineering from Universität Stuttgart (Germany).
cells, structural changes and defect studies using in situ techniques.

Ching-Hsiang Chen received his Chun-Jern Pan received his PhD

PhD in Chemical Engineering degree from NTUST in 2012.
from National Taiwan University He is currently a post-doctoral
of Science and Technology researcher in Prof. Bing-Joe
(NTUST). He is currently the Hwang’s laboratory at the Depart-
global business director in Uni- ment of Chemical Engineering,
nanotech Co., Ltd. and he is also NTUST. Dr Chun-Jern Pan’s
an adjunct assistant professor in research field is mainly in
the Sustainable Energy Develop- characterizing energy materials
ment Center in NTUST. His using in situ X-ray absorption
research interests include surface spectroscopy and X-ray
enhanced spectroscopy and ultra diffraction. These techniques
Ching-Hsiang Chen fast spectroscopy in energy and Chun-Jern Pan have been adopted to investigate
biomedical science. the structure of energy related
materials, such as solar cell materials, catalysts and electrode
materials for batteries.

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

of Omh-PSCs can be accepted for the device to be taken as a another important issue which has an impact on the device
stable device? Secondly, is efficiency the only criterion in stability. The main device architectures: planar (regular TiO2),
assessing the performance? As a final point, the ultimate appli- inverted and mesoporous TiO2 are reviewed in line with their
cation should also be taken into account. Even though standard stability issues. The sequence of the different layers in the three
protocols are not well-outlined, various frequently utilized pro- architectures results in different interfacial reactions which can
tocols such as temperature (60 1C and 80 1C dark tests), light further affect the stability. The discussion of this interface is
soaking tests at 1 Sun at 60 1C or 80 1C, heat-cycle durability subdivided into ion migration into the interface and reactivity in
tests, ultraviolet (UV) tests, and humidity tests can be used to the interface. Finally, we wrap up our discussion with possible
assess stability, which commonly run for 1000 h. Hence, the aim ways toward stability improvements as a function of material,
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

of this article is to comprehensively review the causes of failure device architecture, and interface aspects, as well as protective
and associated possible mechanisms of degradation, the origins means during operation. Tests, evaluation methods and protocols
of degradations followed by approaches to improve the stability for stability studies are also suggested for future material and
as well as the methods and protocols used to study degradation. device development. This review manuscript is finalized with a
Its objectives are also to shed light on possible approaches comprehensive conclusion and prospects.
directed towards a more systematic study of degradation
phenomena in Omh-PSCs. Therefore, the scope of this article
is to critically assess stability and degradation information to 2. Overview of Omh-PSC devices
aid the commercial enterprises and research institutes that
work with these technologies. 2.1 The need for Omh-PSCs
The organization of this review originates from its back- Studies focusing on Omh-PSCs are growing in number and
ground followed by overview information focusing on the need attracting scientists from other subject areas: (1) physicists,
for Omh-PSCs and its associated mechanistic processes, as well constructing devices from new materials, typifying and opti-
as major contributions to the current status of performance mizing performance, and realizing the basic photo-physical
improvement in order to provide an overview of basic principles mechanisms; (2) chemists, developing and manufacturing suit-
and progressive improvements in efficiency and stability. Only able Omh-perovskite absorbers and dealing with structure–
promising steps have been achieved mainly in the area of device property relationships and (3) engineers, building novel device
stability from minutes to thousands of hours. Thus, the various designs. The interplay among these subjects could lead to a
causes of failure and associated mechanisms of degradation great contribution towards potential improvements of Omh-PSCs
have been comprehensively reviewed. Furthermore, the origins in the future.
of device degradation in different device components that are Solar cells with effective PCEs require materials to absorb in
capable of prohibiting the devices outside of an inert atmo- a wide spectral range, from visible to near infrared, to harvest
sphere are discussed in-depth. These components have been most of the solar photons, and to efficiently convert the photon
subdivided into Omh-perovskites, mixed Omh-perovskites, hole into free charges. In order to reduce the energy costs related to
transport layers, and counter electrodes, each contributing their electron/hole separation and charge extraction, highly crystal-
own degradation mechanisms. The device architecture is also line materials possessing good carrier mobility are required.

Meng-Che Tsai received his PhD Bing-Joe Hwang studied Chemical

degree in Chemical Engineering Engineering and received his PhD
from National Taiwan University in 1987 at the National Cheng
of Science and Technology Kung University, Taiwan. Since
(NTUST) in Taiwan in 2013, 2006, he has served as chair
and his advisor is Professor professor at National Taiwan
Bing-Joe Hwang. He is currently University of Science and Tech-
a postdoctoral fellow at NTUST. nology. His research work has
His research interests include the spanned a wide range of subjects
design and synthesis of advanced from electrochemistry to spectro-
energy materials by using a scopy, interfacial phenomena,
computational approach con- materials science and theoretical
Meng-Che Tsai taining energy conversion and Bing-Joe Hwang chemistry. He has established
storage. both experimental and computa-
tional strategies for the development of new nanoscale materials.
His theoretical work has led to a better understanding of reaction
mechanisms on nanoparticles and to an improved ability to predict
the properties of potential new materials for ion batteries, fuel cells
and solar cells.

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

behind efficient solar energy conversion is the combination of

‘carrier generation’ by light absorption and ‘charge separation’.
In the absorber perovskite (see Fig. 2(a)), electrons undergo
excitation. Fig. 2(b) shows the electron transport layer (ETL)
Fig. 1 Power conversion efficiency comparison of third generation solar
cells: a-Si, DSSC, OPV and Omh-PSCs.
and hole transport layer (HTL), both in contact with the Omh-
perovskite. These electrons and holes are separated by an ETL
that is conductive to electrons and rejects holes, and by a HTL
However, these are hampered by unreasonable costs, due to produc- that allows transport of the holes with no impedance, while
tion procedures that use vacuum based or high-temperature proces- blocking electrons, respectively. Such arrangements permit
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

sing, thereby limiting commercial application. On the other hand, the Fermi level in the metal contacts to equilibrate with the
third generation photovoltaic devices, e.g. amorphous silicon (a-Si), separate Fermi levels of the electrons and holes, generating an
dye sensitized solar cells (DSCs) and organic photovoltaic (OPVs) can external voltage, and to extract the carriers, giving a photo-
be made using simple low cost manufacturing processes.8,9 These current in the external circuit.
technologies utilize a combination of amorphous and disordered The mechanistic processes in devices using mesoporous TiO2
materials. Hence, the energy costs associated with the extraction of are still under debate,14 as Omh-perovskite materials combine both
free charges impose a fundamental limitation upon high efficiency the functions of light absorption and n-type or p-type conduction.
systems. As shown in Fig. 1, the new Omh-perovskite structure-based The present mechanistic representation is explained as follows: the
materials are revolutionizing the photovoltaic field, with the efficien- Omh-perovskite absorbs light and an excited state (electron–hole
cies of Omh-PSCs, a-Si, DSCs and OPVs being 20.1%,10 13.44%,11 pair) is produced. Charge separation can take place at the interface
13%12 and 12%13 respectively. Recent research shows that there is a through two processes: injection of a photogenerated electron into
tradeoff between a wide spectral absorption range, highly crystalline a TiO2 nanoparticle (eqn (1)), or hole injection into a HTL (eqn (2)).
structures that induce effective charge extraction, and low-cost Charge separation reactions are in kinetic competition with
fabrication techniques. other reactions, such as exciton annihilation, and lead to
The material’s organic–inorganic hybrid nature is what makes photoluminescence (eqn (3)), recombination in the perovskite
Omh-PSCs so efficient. Firstly, the organic constituent offers good (eqn (4)), as well as recombination of the charge carriers at the
solubility to the Omh-perovskite and facilitates self-assembly different interfaces (eqn (5) and (6)).14
together with its precipitation and/or deposition from solution. Electron injection:
Secondly, the inorganic part provides a comprehensive network by
eperovskite + TiO2 - eTiO2 (1)
covalent and/or ionic interactions, instead of weaker forces such
as Van der Waals or p–p interactions. Such strong interactions Hole injection:
account for the accurate crystalline structure in the deposited
films. Thirdly, powerful optical absorption is a means to attain an hperovskite+ + HTM - HTM+ (2)
excellent performance and literature reports confirmed that Omh- Luminescence:
perovskites of interest are direct-bandgap semiconductors.9
eperovskite + hperovskite+ - hn (3)
2.2 Mechanistic processes in Omh-PSCs
Recombination in perovskite:
The basic energy diagram and elements of the mechanistic
process of Omh-PSCs are represented in Fig. 2. The key idea eperovskite + hperovskite+ - r (4)

Fig. 2 (a) Schematic representation of the energy diagram in an Omh-perovskite (1) generating electrons and hole carriers that rapidly relax (2) to the
conduction band (EC) of the ETL (TiO2) and (3) to the HOMO of the HTL, respectively. (b) The scheme of the main process in Omh-PSCs, formed by the
Omh-perovskite absorber material, supplemented by two selective contacts, the ETL and HTL.

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

Recombination at interface: materials were fabricated and their optoelectronic proprieties

were exploited in the fabrication of transistor and light emitting

eTiO2 + hperovskite+ -r (5)
diodes.20 The Omh-PSC evolution efficiency is shown in Fig. 3
eperovskite + HTM+ - r (6) and Table 1. In 2009, Miyasaka and coworkers21 demonstrated
the potential of a CH3NH3PbI3 absorber in solar cell devices
In a Omh-PSC, the incoming light is absorbed by the Omh- based on TiO2 mesostructure photoanodes with a power con-
perovskite, which is anchored to the surface of TiO2 nanocrystals. version efficiency of 3.81%. The authors were attracted by its
Theoretically, for efficient Omh-PSCs, the Omh-perovskite absor- self-organizing potential in the nanoporous TiO2 layer of the solar
ber regeneration by HTM must happen more rapidly than electron cell. Later, Etgar and coworkers22 demonstrated that CH3NH3PbI3
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

recombination. can act both as a light harvester and as an HTM in a CH3NH3PbI3/

TiO2 heterojunction device. A HTM-free solid state mesoscopic
2.3 Current status of performance in Omh-PSCs CH3NH3PbI3/TiO2 heterojunction solar cell was demonstrated with
As an area of focus in Omh-PSC devices, the PCE is an important a PCE of 5.5% under standard AM 1.5 solar light of 1000 W m2
consideration when making comparisons with the more mature intensity and a PCE of 7.3% at a lower light intensity of
silicon technology. Besides the PCE and fabrication processes, 100 W m2.22 The advent of such straightforward solution-
there are at least two other factors that have an impact on the processed mesoscopic heterojunction solar cells shows the way
success of Omh-PSCs. These factors, which have been given to low-cost, high-efficiency solar cells.
relatively little consideration, are stability and cost.8,9,15,16 Inor- Among the Omh-perovskite-structure based semiconductors,
ganic silicon-based solar cells may last approximately 25 years; so CH3NH3PbI3 is the most common one, and these perovskites have
in this respect, Omh-PSC devices must be improved to become a tendency to possess a high charge carrier mobility.18,20,21,23,24
technologically attractive, economically feasible and commer- For an Omh-PSC device, adequate mobility and high charge
cially practicable. carrier lifetimes are important. Both Xing25 and Stranks26 con-
2.3.1 Efficiency. Omh-perovskite materials were pioneered and firmed that CH3NH3PbI3 has relatively large diffusion lengths
studied, at first by Mitzi and co-workers,16–20 as active layers in i.e. about 100–130 nm for both electrons and holes, which are
light-emitting diodes and field-effect transistors. Different struc- higher values compared to other solution based processed
tures, i.e. 2D and 3D spatial distributions, of the Omh-perovskite semiconductors at low temperatures.
Omh-perovskite absorbers have been used as light absorbers
in liquid electrolyte-based photoelectrochemical cells with
PCEs from 3.8121 to 6.5%.27 However, they will dissolve in the
electrolyte, resulting in a fast performance decline. This degra-
dation led to the replacement of the liquid electrolyte with a
solid hole transporter material (HTM) in 2012.15,28 The HTM of
spiro-MeOTAD was originally produced for organic light emitting
diodes (LEDs)29 but was also found to be useful for solid state dye
sensitized solar cells (ssDSSCs)30 and boosted the reported
efficiency (9.7%) for Omh-PSCs.15 In the middle of 2012, other
successes were reported using ref. 28. Firstly, mixed-halide
CH3NH3PbI3xClx, which displayed better stability and carrier
transport than its pure iodide equivalent, was successfully used.28,31
Secondly, nanoporous TiO2 surfaces were coated with a thin Omh-
perovskite layer thereby forming extremely thin absorber solar cells.
Fig. 3 Efficiency roadmap for Omh-PSC devices: efficiency values taken Thirdly, the conducting nanoporous TiO2 was replaced by a similar
from publications and NREL’s latest chart on record cell efficiencies. but non-conducting Al2O3 insulator. This increased Voc, raising the

Table 1 Summary of device structure and corresponding efficiencies for Omh-PSCs roadmap in Fig. 3

Symbol Device structure Efficiency (%) Ref.

a Liquid electrolyte/CH3NH3PbI3/TiO2 3.81 21
b Liquid electrolyte/CH3NH3PbI3 QD/TiO2 6.5 27
c Spiro-MeOTAD/CH3NH3PbI3/mesoporous TiO2 9.7 15
d Spiro-MeOTAD/CH3NH3PbI2Cl/mesoporous Al2O3 10.9 28
e PTAA/CH3NH3PbI3xBrx/mp-TiO2 12.3 33
f Spiro-MeOTAD/CH3NH3PbI3xClx/TiO2, 2-step solution deposition 15.0 5
g Spiro-MeOTAD/CH3NH3PbI3xClx/TiO2, evaporated thin film p–i–n architecture 15.4 34
h PTAA/CH3NH3PbI3xBrx/mesoporous TiO2 16.2 35
i PTAA/CH3NH3PbI3xBrx/mesoporous TiO2 17.9 35
j Spiro-MeOTAD/perovskite/yttrium-doped TiO2 19.3 36
k PTAA/(FAPbI3)1x(MAPbBr3)x/mesoporous-TiO2 20.1 10 and 37

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

cell efficiency to 10.9%. Omh-perovskites have a broader potential

due to their ambipolar properties. The fourth reported success
was the ability to form a simple planar device with the scaffolding
totally removed.31
The cooperative work of Seok, Grätzel and colleagues led to a
reported efficiency of 12.0%, using both optional layers, includ-
ing a solid Omh-perovskite capping layer overlying the scaffold-
ing (nanoporous TiO2 infiltrated by Omh-perovskite).32 Different
HTMs were investigated, including spiro-MeOTAD, with poly-
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

triarylamine proving to be the best. Seok and coworkers further

reported another efficiency improvement of 12.3% with the same
structures and CH3NH3PbI3xBrx.33 Less than 10% of bromine
gave the best initial efficiency, because of its smaller band gap, but
a greater proportion of Br than 20% offered moisture stability. This Fig. 4 Moisture effect on the stability of heterojunction solar cells based
is caused by the structural transition from tetragonal to pseudo- on MAPb(I1xBrx)3. Reprinted with permission.33 Copyright 2013 American
cubic due to the smaller ionic radius of Br.18 A PCE of 15.0% was Chemical Society.
reported from the Grätzel research group using a nanoporous
TiO2/CH3NH3PbI3/spiro-MeOTAD/Au device architecture.5 They
used TiO2 as scaffolding and a two-step method for perovskite will be necessary. In 2011, the devices were only stable for
absorber deposition, which improved the morphology. 10 minutes before degradation (about 80% degradation)27
Snaith and coworkers34 illustrated a high PCE of 15.4% because CH3NH3PbI3 QDs tend to dissolve gradually into the
with an open-circuit voltage (Voc) of 1.07 V and a short-circuit redox electrolyte (unstable in iodide-containing liquid electrolyte,
current density ( Jsc) of 21.5 mA cm2 using the mixed halide due to rapid dissolution). In 2012, the stability was remarkably
CH3NH3PbI3xClx deposited as a light absorber thin layer improved from minutes to over 500 h by using a spiro-MeOTAD
material with spiro-MeOTAD as the hole transporter. Further- based solid-state Omh-PSC device.15 This improvement is attri-
more, Seok and coworkers realized a cell efficiency of 16.2% buted to the discovery of the solid hole conductor.33 Fig. 4 shows
using CH3NH3PbI3xBrx (10–15% Br) and a poly-triarylamine the effect of moisture on the stability of heterojunction like Omh-
HTM.35 The thickness ratio of the perovskite-infiltrated TiO2 PSCs solar cells made of MAPb(I1xBrx)3. The device was main-
scaffolding relative to the continuous perovskite layer was the key tained at 35% humidity, stored in air at room temperature and
to the improved efficiency. This was increased to a PCE of intentionally exposed to a humidity of 55% for one day on the
17.9% in early 2014.35 An efficiency of 19.3% in a planar fourth day. This device showed enhanced stability at low humid-
geometry without an antireflective coating was demonstrated ity (o50%). On the other hand, the device stability began to
by Yang and coworkers.36 To improve the electron transport reduce at relatively high humidity (Z55%). This might be due to
properties, Y-doped TiO2 has been successfully used as the degradation of the MAPb(I1xBrx)3 material.33 The performance
mesoporous scaffold in perovskite solar cells. Recently, Seok of this device is more stable after 20 days at values of x = 0.2 and
and colleagues have developed a solar cell with an efficiency of x = 0.29 compared to the values of x = 0 and x = 0.06. Thus, an
20.1%10,37 by combining MA and FA (formamidinium), as con- appropriate atomic ratio of the halide anions is an important
firmed in November 2014, by the U.S. National Renewable implication for improvement of the stability.33
Energy Laboratory. Similar to MA, FA is a small, positively Since CH3NH3PbI3 films are not stable in a humid atmo-
charged molecule made of carbon, hydrogen, and nitrogen. sphere, controlling environmental conditions to a moisture
But FAPbI3 absorbs light further into the near infra-red, thereby level of less than 1% is recommended for device preparation.5
potentially boosting the cells efficiency.38 Despite accommodating Moreover, a sealed device subjected to long-term light soaking
a wide variety of cell designs and producing excellent PCEs, Omh- at a light intensity of 100 mW cm2 and a temperature of 45 1C
PSCs must first overcome a number of technical issues before showed promising stability and maintained more than 80% of
widespread commercialization is possible. These include the its initial PCE after a period of 500 h.5 The device was prepared
scale-up of the cell area, the elimination of lead toxicity and the under an argon filled glovebox and maintained at the optimal
lack of long-term device stability.39 Efforts to address these short- electric power output during the ageing process using maximum
comings have begun to show promise, with a ca. 1 cm2 Omh- power-point tracking.
perovskite device producing an efficiency value as high as 8.3%40 It is obvious that mesoporous TiO2 free solar cells are
and lead-free Omh-PSCs being now reported.41,42 significantly more flexible to UV irradiation compared to TiO2
2.3.2 Stability. The current highest PCE reached by Omh- based solar cells. The performance of TiO2 based devices is
PSCs is 20.1% and, while the stability is low, it is by no means a limited to only 5 h stability due to the fast decay in photocurrent
barrier to application in many situations. The major obstacle is and the small decay in photovoltage when tested in ambient
the relatively short lifetime achieved with current technology. conditions.6 Opposing the behavior of TiO2-based solar cells, the
To bring Omh-PSCs onto the market, in addition to higher photocurrent become stable nearly at 15 mA cm2 for over 1000 h
efficiencies and cost-effective processing, a long device lifetime in a Al2O3-based Omh-perovskite solar cell.6 Fig. 5 shows the

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

Fig. 5 Stability test of a mesoporous TiO2-free solar cell device. Reprinted

with permission.6 Fig. 6 Roadmap in stability for Omh-PSCs made from Omh-perovskite
absorbers.

stability test of a meso-super structured Al2O3-based Omh-

crystalline properties of synthesized Omh-PSCs, affecting the
perovskite solar cell (MSSC) exposed to continuous simulated AM
photoluminescence behavior.46–48 In addition, recombination
1.5 76.5 mW cm2 illumination at 40 1C. The FF and Voc started to
effects should also be expected, but further research is needed
drop in the first 200 h experienced and this initial decline may be
to correlate the observed changes in recombination with the
because of different factors such as oxygen desorption at vacancy
crystalline properties of Omh-perovskite samples.
sites on the compact TiO2 layer, partial de-doping of the spiro-
Han and co-workers4 applied a solution drop-casting method
MeOTAD,43 modifications at the spiro-MeOTAD/Au interface, or
to fabricate a novel device where its scaffolding was made of a
slight modifications within the Omh-perovskite.
double layer of TiO2 and ZrO2 enclosed in a porous carbon film.
Kwon and co-workers44 reported the ex situ long-term stabi-
The resulting mixed cation Omh-perovskite showed higher
lity of Omh-PSCs using three HTMs: spiro-MeOTAD, P3HT and
electrical charge efficiency than conventional (planar) Omh-
poly[2,5-bis(2-decyl dodecyl) pyrrolo[3,4-c]pyrrole-1,4(2H,5H)-
perovskite cells. The triple layer also resulted in a better surface
dione-(E)-1,2-di(2,20-bithiophen 5-yl)ethene] (PDPPDBTE), under
contact, leading to a considerably higher stability of 1008 h in
a 20% humidity atmosphere for over 1000 h. The PCE of the
direct sun exposure. It was also shown that devices with
spiro-MeOTAD based device slowly reduced when the ageing
mesoporous TiO2/CH3NH3PbI3/C architectures, without encap-
period was extended. The device showed a 28% decrease in the
sulation, had stabilities of over 2000 h in air in the dark.7 The
PCE in relation to the early PCE. Furthermore, the stability of the
stability evolution for Omh-PSCs is shown in Fig. 6 and Table 2.
PDPPDBTE cells improved remarkably, and the initial perfor-
The ability to process carbon electrodes at low temperatures
mance was retained at a PCE of 8.4% after 1000 h. The improved
on top of the CH3NH3PbI3 layer without destroying its structure
stability may be due to its hydrophobic behaviour, which avoided
not only reduces costs and simplifies the fabrication procedures,
water flow into the Omh-perovskites. Moreover, the solar cell
but also improves device stability.
stability of mixed halide perovskites, MAPb(BrxI1x)3yCly, grown
Although some improvements have been made to the stability
on nanostructured-TiO2 and in a thin film configuration, was
of Omh-PSCs, stability issues have not yet been well-addressed.
monitored to study cell stability as a function of the halide
The major issue pertaining to the commercialization of an Omh-
composition.45 Hence, the stability of Omh-PSC devices has been
PSC device is mainly its stability. Not only issues of ambient
limited, mainly due to ambient moisture because alkylammonium
stabilities with respect to moisture, temperature, light and
salts are, in general, highly hygroscopic. Interestingly, the insertion
oxygen, but also issues concerning intrinsic stabilities at the
of bromine into the Omh-PSC structure benefits the cell stability.34
interface, together with questions related to the device’s archi-
Those cells without bromine show an efficiency drop of 20%,
tecture, remain as major obstacles to practical application. All
30 days after preparation, which is ascribed mainly to a decrease
of these stability issues and obstacles are the main focus of this
in the FF. However, the equimolar Br/I devices showed a signifi-
review article and will be discussed separately in the following
cant enhancement of 37% PCE.45
sections.
This improvement could be a consequence of the rearrange-
ment of the Omh-PSCs 3D configuration over time. In particular,
the efficiency improvement may be attributed to the insertion of 3. Causes of failure and associated
the smaller bromine anions, leading to a more compact Omh- mechanisms of degradation
PSC structure in which the degradation of the MA cation is
prevented.45 Therefore, a deeper analysis including microstruc- The formation of Omh-perovskite materials depends on the
tural and crystallographic studies must be performed to eluci- chemical composition of the precursor and the reaction con-
date the origin of this observation. Recently, it has been pointed trolling parameters, such as the temperature and pressure of the
out that the presence of nano-structured TiO2 changes the system. Dualeh and co-workers proposed that at lower temperature,

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

Table 2 Summary of device architecture and its stability for Omh-PSCs based on the roadmap in Fig. 6

Conditions
Symbol Device structure Stability/h Temp. Atmosphere H Ref.
a TiO2/CH3NH3PbI3QD/liquid electrolyte/Ptk 0.17 RT Air — 27
b TiO2/CH3NH3PbI3/spiro-MeOTAD/Auk 500 RT Air — 15
c TiO2/CH3NH3PbI3/spiro-MeOTAD/Aum 500 45 1C Ar gas o1% 5
d Al2O3/CH3NH3PbI3xClx/spiro-MeOTAD/Aum 1000 40 1C N2 atmosphere — 6
e TiO2/CH3NH3PbI3/PDPPBTE/Auk 1000 RT Air 20% 44
f TiO2/ZrO2/(5-AVA)x(MA)1xPbI3/Ck 1008 RT Air — 4
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

g M-TiO2/CH3NH3PbI3/Ck 2000 RT Air — 10 and 37

m k
Encapsulated, non encapsulated device, Temp. = temperature, H = humidity, RT = room temperature.

the conversion to the Omh-perovskite dominates following Firstly, since the molar amount of CH3NH3I is 3 times that of
eqn (7)–(9) however, at a higher annealing temperature, there PbCl2, eqn (11) is proposed, where part of CH3NH3I reacts with
is the additional formation of PbI2 (eqn (8)).50 PbCl2 to form PbI2. As a result of eqn (11), the spun film
contains mixed phases of PbI2 (appears light yellow), CH3NH3I,
PbCl2 + 3CH3NH3I - CH3NH3PbI3 + 2CH3NH3Cl (7)
CH3NH3Cl, and possibly some unreacted PbCl2. During anneal-
PbCl2 + 3CH3NH3I - PbI2 + CH3NH3I + 2CH3NH3Cl (8) ing, two reaction processes may occur: (1) in eqn (12), CH3NH3I
reacts with PbI2 to form dark brown CH3NH3PbI3; meanwhile,
CH3NH3PbI3 - PbI2 + CH3NH2 + HI (9) the excess CH3NH3Cl breaks away from the film. The question of
how excess CH3NH3Cl escapes from the film remains unclear,
It has been confirmed that the CH3NH3PbI3 crystal lattice does
but different literatures report it as either sublimation or decom-
not decompose up to 300 1C, after which the organic constituent
position.52 (2) During the early stages of annealing or during spin
rots, governed by eqn (9).24,32 However, recent studies show that
coating, an intermediate phase may appear as indicated in
CH3NH3PbI3 begins to transform to PbI2 at lower temperatures
eqn (13). In this case, during the decomposition of this inter-
up to 140 1C.50 The excess organic CH3NH3Cl formed in this
mediate phase, the CH3NH3PbI3 crystal network can grow, during
reaction, is believed to sublime, leaving only the CH3NH3PbI3
which the driving force can be the release of gaseous CH3NH3Cl
on the mesoporous TiO2 film, as observed from XRD measure-
(or other organic chlorides).49
ments.51 The rate of sublimation of the organic species CH3NH3Cl
Despite the understanding of the formation of Omh-
increases with temperature, as does the precursor CH3NH3I,
perovskite materials for solar cell use, the full advantage of
thus driving eqn (7) and (8). From these measurements it is
these materials and their devices is still not realized due to the
evident that the formation of the Omh-perovskite films is a
presence of unsolved problems. One typical example is a lack of
multistage process, comprising solvent vaporization, perovskite
clear understanding on the mechanism of the active material
crystallization and the sublimation of excess organic CH3NH3Cl.
and device degradations which controls essential processes like
These processes occur simultaneously and their relative rates
the device life time and the mechanical stability of Omh-PSC
determine the composition and morphology of the final film.
panels. For the successful development of Omh-PSCs, it is essen-
The release of chlorine-containing compound(s) clearly sug-
tial to understand the degradation behaviour of Omh-perovskites
gests the formation of crystalline MAPbX3 films. Thus, it is vital
and devices in general. The study of Omh-perovskite degradation
to realize the potential chemical reaction mechanisms for the
behaviour is a pre-requisite for Omh-PSC applications. Such
formation process of MAPbX3. Dualeh and co-workers anti-
degradations can be linked to extrinsic (oxygen, light, moisture
cipated that eqn (10) dominates the annealing process.51 How-
and temperature) induced and intrinsic (thermal and electric
ever, Zhao noted that the initial CH3NH3I and PbCl2 are white
field) induced degradations. Fig. 7 shows a general scheme
and the solution mixture appears to be a light yellow color,
representing the recombination process at mesoporous TiO2,
indicating that a chemical reaction should take place forming
moisture dissolution of the Omh-perovskite and photooxida-
a new phase.49 Hence, eqn (10) possibly will engage numerous
tion processes that take place at the interface between HTM
intermediate steps.
and the counter electrode. The diffusion of O2 is gradually sped
PbCl2 + 3CH3NH3I - CH3NH3PbI3 + 2CH3NH3Cl (g) (10) up by UV light in the presence of TiO2. The produced oxygen
free radicals can then react with the HTM and/or active Omh-
PbCl2 + 3CH3NH3I - PbI2 + CH3NH3I + CH3NH3Cl (11) perovskite materials. H2O is also readily activated by oxygen and
light in the presence of organic molecules. Moreover, molecular
PbI2 + 3CH3NH3I + 2CH3NH3Cl - CH3NH3PbI3 + 2CH3NH3Cl (g)
oxygen and moisture may flow through the tiny pinholes present
(12)
in the counter electrodes as represented in Fig. 7. Omh-perovskite
PbI2 + xCH3NH3I + yCH3NH3Cl materials such as CH3NH3PbI3 and CH3NH3PbBr3 are prone to
chemical attack, thus devices typically degrade in a matter of
- (CH3NH3)x+yPbI2+xCly(intermediate phase) minutes to hours under 1000 W m2 illumination in the ambient
- CH3NH3PbI3 + 2CH3NH3Cl (g) (13) atmosphere. However, the impacts of these parameters that

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

presence of organic molecules,53–55 which promote the formation

of peroxide or superoxide compounds that attack and degrade the
active layers.56
The defect chemistry of the n-type semiconducting oxide
gives rise to the curious behavior of fast device decline, which is
primarily introduced by oxygen vacancies and interstitial defects,
resulting in a non-stoichiometric composition (TiO2x). Two
under-coordinated Ti(III) and one oxygen vacancy are generated
when bridging oxygen atoms become eliminated. The elimina-
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

tion of bridging oxygen atoms from the lattice frequently results

from thermal annealing. Because of their unsaturated co-ordination
the generated under-coordinated Ti(III) and oxygen vacancy
are reactive and adsorb O2 molecules to form a Ti(IV)+–O2
Fig. 7 General schematic representation of hole–electron pair recombi- complex.57 Among the various types of adsorbed oxygen (O+,
nation, moisture dissolution of Omh-perovskite and photooxidation pro- O2, O), the superoxide radical, (O2) is thermodynamically
cesses at the interface between the HTM and counter electrode. stable and an electron is transported to the adsorbed oxygen
from the TiO2x surface.58 Likewise, formation of O2 can take
decrease or increase the degradation rate are not exactly clear. place through a different reaction if electrons are available in
The objective of this section is to draw the extrinsic and intrinsic excess in sub-bandgap states or the CB,57 i.e. Ti(IV) + eCB + O2 -
degradation phenomena and associated degradation mecha- Ti(IV)+O2(a). Fig. 8 represents the formation of deep trap states
nisms of Omh-perovskite materials and Omh-PSC devices when and a depletion area as a result of oxygen adsorption and
built-in in photovoltaic systems and to review the various desorption processes at the mesoporous TiO2. In Omh-PSCs,
factors influencing the stability of Omh-perovskite materials the formation of a Ti(III) (3d1) trap state is due to the transfer of
and Omh-PSCs. one electron to the Ti(IV) (3d0) of TiO2 from the perovskite. Thus,
the electron easily transfers in to Ti(IV) from Ti(III) because of
3.1 Oxygen induced degradation the smaller bandgap between Ti(III) and Ti(IV). Those under-
An important factor in the operation of any device is for it to be coordinated Ti(III) generated due to oxygen vacancies reside in
stable in air without the use of any type of encapsulation. The sub-band gap states and work as deep trap sites. Furthermore,
rationale behind this is that every group of materials in a device they attract molecular oxygen in ambient conditions, resulting
that realizes ambient stability will address an input to future in the formation of a(Ti(IV)+O2(ads)).57,59,60 Likewise, electrons
improvement. Yang and co-workers compared devices stored in in sub-band gap states may create b(Ti(IV)+O2(ads)). These
dry air and under nitrogen atmospheres and showed that modified states do not interfere with the device’s performance
degradation of the Omh-perovskite materials occurred in ambient when illuminated. The formation of the depletion layer can be
air – thereby, highlighting the need for protection.36 The diffusion as a result of the positive potential at the TiO2 surface, caused
of molecular oxygen can be activated by UV illumination in the by the presence of O2.6,61 Devices continuously illuminated at

Fig. 8 Oxygen induced formation of deep trap states and depletion area at mesoporous TiO2 surface and interface: (a)–(c) the patterning of superoxide,
a and b(Ti(IV)+–O2). (d) Desorption of negatively charged superoxide molecules and their impact on the depth of the depletion zone (orange shaded area).

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

high temperatures deteriorate. Changing of the a superoxide perovskite into dihalogens (I2, Br2, and Cl2). Therefore, oxidation
complex to a Ti(III) and liberating O2 may take place when induced halogen elimination may be one possible reason for
photo-excitation of the mesoporous TiO2 results in the formation degradation of Omh-PSCs. Likewise, light exposure of MAPbI3
of electron–hole pairs, through which holes in the VB are leads to an irreversible breakdown into PbI2.51,79,80
trapped with O2 in the Ti(IV)+O2, as shown in Fig. 8(b). The
corresponding free electron in the conduction band (CB) is then 3.3 Moisture induced degradation
trapped at the regenerating deep trap site and will ultimately Moisture is one of the main reasons of degradation that
recombine with an excess hole in the spiro-MeOTAD, as shown diminishes the PCE of unencapsulated Omh-PSC devices. Seok
in Fig. 8(c). An adsorbed negative charge may result in an and co-workers33 suggested that moisture degrades the PCE of
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

increase in the depletion area, thereby creating an upward bend unencapsulated Omh-PSC devices causing device fading and
in the CB that diminishes interfacial charge carrier recombina- recommended that the fabrication of Omh-PSCs should be
tion. In addition, O2 desorption upon illumination diminishes carried out in a controlled atmosphere with a humidity level
this upward band bending and hence results in the disappear- of o1%.5 Besides, a proper humidity level (30  5% relative
ance of the Schottky barrier, a flattening of the CB-edge and a humidity) is essential for high quality films, enhanced opto-
lowering of the Fermi level62,63, as shown in Fig. 8(d). Oxygen- electronic properties, and improved mass transport processes.36
induced defects in TiO2 are the major reasons that cause device For instance, a CH3NH3PbI3xClx film grown at 30% relative
degradation during operation.64,65 humidity in air had improved optoelectronic properties com-
pared with a film grown in dry conditions. Moreover, You
3.2 Light induced degradation and coworkers81 reported a significant improvement in film
UV illumination can diminish the performance of Omh-PSCs morphology when Omh-perovskite precursor films were annealed
upon exposure during long-term operation due to the genera- in ambient air (humidity of 35%  5%) compared to films
tion of many possible fading processes. Electron recombination annealed in a nitrogen filled glovebox with the O2 and H2O level
thereby creates localized trapping sites and diminishes the lower than 5 ppm. The latter annealing process results in a
depletion layer and TiO2 band excitation. This may lead to pristine film with a grain size of 100–300 nm, and apparent pin
the oxidation of halogen atoms in the Omh-perovskite or HTM holes and grain boundaries.81
occurring, due to UV illumination. Of course, it should be recog- These pin holes and grain boundaries introduce ‘energetic
nized that there are other sources of degradation besides UV light, disorder’ that impedes charge transport, induces recombina-
e.g. oxidation, heat and ordinary visible light. UV illumination is in tion, and lowers photovoltaic performance.82 Omh-perovskite
many cases the major cause of degradation and thus is of great films annealed in ambient air, however, give larger individual
concern, meaning it is necessary to use stable and broadly crystal and grain sizes (beyond 500 nm), as well as reduced pin-
absorbing UV materials in order to protect both the light absorber holes and grain boundaries.81,82 This indicates that moisture-
and the substrate by either initiating a decomposition that assisted crystal growth induces grain boundary creep as a result
proceeds by another mechanism such as oxidation, or by directly of the absorption of moisture within the grain boundaries and
absorbing UV energy. then the merging of adjacent grains together. Consequently,
UV-induced degradation mechanisms can be explained by this effectively increases the grain size, avoids pinhole forma-
considering the surface chemistry of TiO2. Mesoporous TiO2 tion and finally improves the carrier lifetime.81,82 Additionally,
incorporates surface trapping states or defects (Ti3+), function- adsorbed water molecules could undergo auto-ionization at the
ing as efficient deep electron-releasing states;66,67 the electrons surface into hydroxyl and proton ions. As a result, surface OH
in these localized trapping states,68 bind with O2 molecules sites may be formed between which protons may migrate and
from the environment, adsorbing the trapping states,69–72 and act as charge carriers. Thus, it would be interesting to investi-
generating a complex of O2–Ti4+.71–73 Hence, the adsorption gate the contribution of these ions into the mass transport
process is responsible for the photocurrent degradation caused processes and optoelectronic properties. Furthermore, You and
by the adsorption of oxygen on TiO2 surfaces.66,67,69–73 Upon co-workers81 systematically studied the effect of the moisture
bandgap excitation of TiO2, a hole in the VB recombines with level (20–80% humidity) on Omh-perovskite film formation.
the electron at the oxygen adsorption site. This creates an Omh-perovskite films annealed in conditions above 80% humid-
electron–hole pair and desorbs oxygen.61,72 The excess of holes ity showed a small amount of PbI2 phase present in the crystals,
in the HTM will gradually rebind with the unbound electron indicating Omh-perovskite film decomposition to PbI2 at higher
remaining from the Eg excitation of the titania.53,74,75 The holes humidity levels. While moisture is beneficial for high quality
in spiro-MeOTAD recombine the trapped electrons.76–78 At a Omh-perovskite film growth, its relative amount should be
steady-state, the partial pressure of oxygen within deep trap sites carefully controlled.81
governs the quantity of unfilled oxygen vacancy states. This is Moisture degradation of Omh-perovskites may be correlated
because of the sluggish and reversible desorption rate of oxygen.72,73 with the transformation of MAPbI3 to its MAI salt and metal
In Omh-PSCs, the inorganic Pb2+–X creates a well ordered halides, and hence metal halide removal may also cause Omh-
inorganic matrix which successfully absorbs and re-emits light. PSC device degradation. Wang and co-workers83 reported that
Thus, UV illumination of the local environment of the Pb–X during the process of assembling and testing, moisture in the
bond could generate halogen free radicals and break down the atmosphere can directly degrade these materials, as indicated

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

Fig. 9 XRD patterns of films of (a) TiO2/CH3NH3PbI3 before and after degradation, and (b) TiO2/CH3NH3PbI3/Al2O3 before and after degradation.
Reproduced from ref. 83 with permission from The Royal Society of Chemistry.

by the XRD pattern for films of TiO2/CH3NH3PbI3 and TiO2/ given the propensity of PbI2 to crystallize and give rise to phase
CH3NH3PbI3/Al2O3 in Fig. 9. They proposed that the first step in separation, the process is unlikely to be fully reversible. The
Omh-perovskite decomposition is the reversible deprotonation formation of hydrated intermediates containing isolated PbI64
of the MA cation by water, forming methylamine, hydrated HI, octahedra as the first step in the decomposition process at
and PbI2. If this hypothesis is correct, one would expect that the 98  2% RH and further decomposition of CH3NH3I to CH3NH2
volatile CH3NH2 would be rapidly flushed from the headspace and HI would ultimately leave PbI2 as the only byproduct of the
of the in situ sample holder, meaning that the process would be reaction.86 Similarly, density functional theory analysis confirms
unlikely to be reversible. Given the apparent reversibility in the that moisture degrades the Omh-perovskite structure because
bulk powder, thin films would also display a similar reversible it weakens the hydrogen bonding between the PbI6 octa-
conversion to the hydrate phase. Furthermore, the exposure of hedra and methyl ammonium cation.87 While previous studies
CH3NH3PbI3 to moisture produces a hydrate product rather proposed (MA)4PbI62H2O as the first moisture degradation
than simply resulting in PbI2.84 product,84,86 Leguy and co-workers recently confirmed that
In the past, Vincent and coworkers were able to study the first hydrate produced is monohydrate, MAPbI3H2O rather
(CH3NH3)4PbI62H2O, which forms readily by the addition of than the dihydrate ((MA)4PbI62H2O). (MA)4PbI62H2O can be
aqueous Pb(NO3)2 to an aqueous solution of CH3NH3I.85 In produced upon additional hydration of MAPbI3H2O, as described
contrast to the perovskite crystal structure of CH3NH3PbI3, the in (eqn (15)).88
crystal structure of (CH3NH3)4PbI62H2O consists of an assembly
of PbI64 octahedra and (CH3NH3  H2O  H3NH3C)24+ dimers 4(CH3NH3)PbI3 + 4H2O $ [4(CH3NH3)PbI3H2O]
arranged on a distorted NaCl-type lattice.85 This unique structure $ (CH3NH3)4PbI62H2O + 3PbI2 + 2H2O (15)
creates a zero-dimensional network of isolated PbI6 octahedra,
as opposed to the extended three-dimensional network observed The formation of MAPbI3H2O can be a reversible process and
in CH3NH3PbI3. Similarly, (CH3NH3)4PbI62H2O is a pale yellow form crystalline MAPbI3 when it is dehydrated. Primary confir-
crystalline solid, and it has been suggested that the conversion mation recommends that even photovoltaic performance could
from CH3NH3PbI3 to (CH3NH3)4PbI62H2O can occur in humid be restored using dehydration techniques.88 In contrast, exposure
air.79 Similarly, Kelly and co-workers reported further evidence to excess water consequences in an irreversible change into PbI2
for this intermediate product using a more systematic investi- and ultimately leads to the complete dissolution of the MAPbI3
gation of Omh-perovskite degradation processes, except that layer.89 These results help elucidate the fundamental decomposi-
they found PbI2 as the ultimate product.86 The formation of a tion pathways in Omh-perovskite films, which should lead to
hydrated intermediate containing isolated PbI64 octahedra as more stable materials and more commercially viable devices.
the first step of the degradation mechanism strongly suggests It is, therefore, important to figure out the thermodynamic and
that the initial step of the Omh-perovskite decomposition pro- kinetic merits of these degradation mechanisms. Moreover,
cess is not an acid–base reaction of the MA cation, but rather the chemistry of moisture degradation still requires further
hydration of the Omh-perovskite film (as illustrated in eqn (14) more detailed computational and experimental evidences in
with (CH3NH3)4PbI62H2O, although other hydrate compositions order to rationalize and establish clear moisture degradation
may be possible):86 mechanisms.

4CH3NH3PbI3 + 2H2O $ (CH3NH3)4PbI62H2O + 3PbI2 3.4 Temperature induced degradation

(14)
Increasing the annealing temperature, from 40 1C to 100 1C
Since eqn (14) does not produce any volatile byproducts, it for 30 min, improved the performance of a perovskite QD-
would be expected to be at least partially reversible; however, sensitized TiO2 film.21 However, a further increase in the

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

annealing temperature from 100 1C to 160 1C diminished the 3.5 Thermal and electric field induced intrinsic degradation
efficiency from 4.73% to 3.71%, due mainly to the decay in the The interface is the main origin of degradation in Omh-PSCs. The
photocurrent density. External quantum efficiency (EQE) spectra question of how the interface can cause degradation is related to
showed that the decreased photocurrent density with high the intrinsic deprivation of Omh-PSCs. The origin of this degrada-
temperature treatment (160 1C) is attributed to a significant loss tion is: (1) the temperature induced flow of components due to
in EQE in the long wavelength region (above 500 nm). Both the thermal evaporation during deposition and light illumination
annealing temperatures and the storage temperature are impor- during operations. This thermal induced migration of component
tant parameters that may cause Omh-PSC device degradation, as materials at the interface results in inter-diffusion, phase segrega-
can be seen from Table 2. Though there is no literature report, it
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

tion and separation, presenting an inherent strain. (2) Electric

is clear that Omh-PSC devices need to pass tests such as thermal field induced ion migration which leads to electro-migration and
stress tests at 80 1C run for 1000 h, for the purposes of practical hysteretic effects. In the electro-migration process, an electric
application. It is obvious that the operating temperature will current flowing in a conductor may move metal ions. Electro-
increase to even higher than 80 1C and heat the sample film migration pushes anions to the anode and builds up a compres-
during operation due to continuous illumination from a solar sive stress there. The vacancy concentration in the anode becomes
simulator (e.g., 0.1 W cm2 sun light).90 In addition to light, less than the equilibrium vacancy concentration according to the
moisture and oxygen induced degradations, Omh-perovskites Nabarro Herring model of point-defect formation in a stressed
undergo thermal degradation. This fact will, therefore, lead to solid.93 In the hysteresis process, ion migration is particularly
additional primary problems in the course of Omh-perovskite sensitive to the concentration of mobile vacancies (or interstitials)
degradation processes. This degradation may produce PbI2 and depending on the mechanism. Thus, the size of crystalline
organic salt forms. The thermal stability of CH3NH3PbI3 was domains, degree of crystallinity and stoichiometry can affect the
confirmed to be even higher than 300 1C.24,32,91 However, latest transient behavior of Omh-PSC devices.94 This transient behavior
literatures confirmed that the organic decomposition tempera- is different from device to device. For instance, the larger transient
tures could be lowered up to 140 1C.51,52,83 The temperature difference between mesoporous and planar devices is most prob-
induced decomposition of methylammonium iodide may lead to ably due to variations in the composition stoichiometry or
the formation of HI and CH3NH2. CH3NH2 could stay within the morphology. Therefore, minimizing ion migration is an essential
network of the perovskite, and then disturb the photovoltaic mechanism in extending the stability of devices.
processes.52 Furthermore, the presence of metal halides, for Moreover, TiO2 has a strong ability to extract electrons from
example, PbI2, as a degradation product of Omh-perovskite, is organic materials as photocatalysts and from iodide (I) as
another issue. Due to their poor optical behaviour and poor light electrodes in DSSCs. Hence, the driving force of the decomposi-
absorption ability as well as their higher bandgap, lead halides tion mechanism may be due to the effect of electron extraction
cause device deterioration. Additionally, the mechanism of by TiO2 from an iodide anion accelerated by light illumination
expansion and compression in the dimensional changes of during operation. The possible decomposition pathway at the
the perovskite structure at low and high temperatures is not TiO2 surface may be as shown in eqn (16)–(18):95
yet clear. Moreover, high temperatures can cause the diffusion
of interlayer, diffusion of the metal counter electrode, ohmic 2I 2 I2 + 2e (16)
contact degradation and device architecture degradation, and
thus the mechanisms for these phenomena should be reported 3CH3NH3+ 2 3CH3NH2 (g) + 3H+ (17)
carefully. For this reason, new Omh-perovskites which possess
I + I2 + 3H+ + 2e 2 3HI (g) (18)
enhanced intrinsic resistance towards temperature are better
candidates for future development in areas of Omh-PSC based Decomposition of a CH3NH3PbI3 crystal under UV light illumi-
photovoltaic technologies.79,92 For instance, FAPbI3 is more nation is shown in Fig. 10. Since CH3NH3PbI3 is a combination of
temperature resistant than MAPbI3.92 CH3NH3+, Pb2+, and I ions, TiO2 can extract electrons from I,

Fig. 10 Degradation scheme of CH3NH3PbI3 on mesoporous TiO2 (mp-TiO2) during UV light exposure tests. Redrawn based on ref. 95.

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

producing I2 at the interface between TiO2 and CH3NH3PbI2 as

shown in eqn (16). This will lead to breakdown of the perovskite
crystal. Eqn (17) is in equilibrium. The electron taken by TiO2
can go back to the TiO2 surface, and eqn (18) can take place
while releasing HI (Fig. 10). As H+ is consumed by eqn (18), the
equilibrium of eqn (17) can favor the right hand side, with the
release of CH3NH2 due to its low boiling point (17 1C).96 It is,
therefore, crucial to realize three of the possible major con-
sequences of iodide ion migration that could occur in MAPbI3
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

based Omh-PSCs: (i) iodide may diffuse out into the TiO2 layer,
(ii) its distribution may be influenced by the electric field in the
Omh-perovskite materials or by any modification of it, which is
because iodide moves as a negative ion, and (iii) iodide migration
may appear to be involved in some of the observed electrical
metastability of the device. These all are pertinent to the interface
stability. The stability of the Omh-perovskite layer under light
exposure drastically vanishes, and shortens the lifetime of the
Omh-perovskite. The origin of degradation for CH3NH3PbI3 layer
is the interface between TiO2 and CH3NH3PbI3. This degradation
is as a result of the variable positioning of the CH3NH3+ cation in
Fig. 11 Pathway for migration of vacancies and defects: iodine (a), methyl-
the CH3NH3PbI3 crystal.97–99
ammonium (b), lead (c) and iodine defects (d). The colours of black, purple,
Moreover, recent computational reports100–104 overview the white, blue and brown represent Pb, I, H, N and C atoms, respectively.
defect formation energy for various defects together with the Reproduced from ref. 105 with permission from The Royal Society of
cations (MAPb and PbMA) and antisites (MAI, PbI, IMA, and IPb) Chemistry.
exchange, interstitials (MAi, Pbi, and Ii) and vacancies (VMA, VPb,
and VI). The kinetically and energetically unstable antisites
impulsively break up into the respective vacancies and inter- The chemistry of such ion/defect migration induced interfacial
stitials.104 Because of their low formation energy, interstitials and reactions, as well as its impact on the device stability, needs
vacancies are the most likely defects. Interestingly, Angelis and further detailed computational and experimental study. Above all,
coworkers105 developed a diffusion path to model vacancies in order to investigate the ion-migration mechanisms of MAPbI3,
and defect migration along the perovskite crystal for all of the studies on ion occupations are of great importance.
four explored defects, as shown in Fig. 11. Vacancies and inter-
stitial defects are referred to with dashed circles and red atoms, 4. Origins of Omh-PSCs degradation
respectively. While dashed lines show the trajectory of the
vacancies, solid lines represent migration of the ions. Fig. 11a In addition to identifying the main causes of degradation, it is
shows the formation of a vacancy, VI, in an equatorial position critical to isolate the key origin of degradation. Thus, more
and its migration towards an axial site. VI in equatorial or axial detailed discussion on device components and architecture and
sites have almost the same energy (o0.01 eV). However, the interface, including the buffer, is essential, all of which are the
axial site is preferred by 0.07 eV over the equatorial site for VBr, as focus of this section.
shown in the supplementary information of ref. 105. The inor-
ganic scaffold is responsible for the hopping of VMA between its 4.1 Effects of device components
nearby cavities that lie in the ab plane, as shown in Fig. 11b. Omh-PSCs are multilayered devices which are a combination of
Moreover, VMA diffuses across the framework of the Pb4I4 struc- an active Omh-perovskite material, n-type and p-type electron
ture. Similarly, Fig. 11c indicates an in-plane migration of VPb and hole selective layers, as well as the counter electrode. In
where VPb travels beside the square created by four I and four Pb addition, different interlayers for different functions can be
atoms. Similar to VI, the pathway in the direction of the c axis inserted. These multilayered devices are prone to component
from Fig. 11d refers to iodine interstitials (Ii). In the initial degradation. The origin of this degradation can be from
and/or last configurations, the interstitial iodine atom places components such as light harvesting active materials, electron
in between a couple of equatorial and/or axial I atoms, with and hole selective layers, anode and cathode materials, as well
almost the same lengths (3.96 and 3.87 Å), respectively.105,106 as the interlayer or buffer layer materials, which are the main
This migration of defects/ion such as iodine vacancies focus of this subsection.
across the interface can induce interface degradation, affect 4.1.1 Omh-perovskites. Omh-perovskite semiconductor
device operational mechanisms and finally cause device failure materials include CH3NH3PbI3, CH3NH3PbBr3, and mixed Omh-
during operation.105 The formation of metal oxides, metal perovskites including mixed metal cations, organic cations and
halides and molecular halogens can take place due to oxidation mixed halide anions. Omh-perovskite structures have become
and reduction processes in the presence of oxygen and light. innovative alternatives for the next-generation of high performance

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

solar cells, because they combine the advantages of both stability of MAPbBr3 films, compared with their iodide counter-
systems (organic and inorganic cages). part, may be related to differences in the bond strengths and
the crystalline forms of the two Omh-perovskites. Thus, further
4.1.1.1 Triiodide anions. A critical issue for Omh-perovskite investigation is still required on how bromine inclusion provides
materials is their stability. The highest reported stability of better stability in both material and device aspects. Despite the
Omh-PSCs with light absorbing MAPbI3 is 2000 h.7 Compared better photochemical stability of MAPbBr3 films, its larger optical
to Si solar cells, this stability is quite poor due to the ambient band gap (Eg E 2.3 vs. 1.58 eV for MAPbI3) would reduce the PV
sensitivity of the CH3NH3PbI3 semiconductor. This instability performance of MAPbBr3 single junction devices.22 Katz proposed
has been a major obstacle since Omh-perovskite-type materials that MAPbBr3 could be efficiently used as a top sub cell in future
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

contain hygroscopic amine salts. CH3NH3PbI3 undergoes a tandem Omh-perovskite-based PV architectures.122 A mixed
rapid change to hydrated CH3NH3PbI3 and then to PbI2 in halide Omh-perovskite, MAPbI3xBrx, with a larger band gap than
the presence of moisture at room temperature,84,86 resulting in MAPbBr3, may therefore have applications in Omh-perovskite
a significant decline in the device performance.107 Further- based solar cells that combine high efficiency and stability.
more, the formation of stress and microstrain, and the origin 4.1.2 Mixed Omh-perovskites. As discussed in Section 4.1.1.1,
of degradation in the CH3NH3PbI3 perovskite structure under CH3NH3PbI3 suffers from long-term ambient stability due to
severe conditions are still open questions. CH3NH3PbI3 structural its sensitivity towards moisture.84,86 In addition, Weber and
changes during operation is also another challenge. For instance, co-workers reported the temperature-dependent structure of
this can cause poor thermal and photoconductivity, and hyster- CH3NH3PbX3 (X = Cl, Br, I).99 The structure of CH3NH3PbX3
esis in the device during operation.108,109 Furthermore, when the changes from orthorhombic to cubic structures as a function of
operating temperature increases to a higher temperature temperature. The order of the methylammonium cation is also
(B58 1C) the structure of MAPbI3 changes from the ordered different in different crystal structures. It is also worth noting that
tetragonal into the disordered cubic structure. Conversely, when at room temperature, CH3NH3PbI3 forms a tetragonal structure
the operation/processing temperature decreases to room tem- below 327.4 K, whereas CH3NH3PbBr3 and CH3NH3PbCl3 form
perature, its structure returns back to its ordered tetragonal cubic structures at 236.9K or above. This order–disorder behavior
shape.110,111 Thus, how this ordered–disordered structure of of the perovskite structure caused by thermal cycles could give a
the MAPbI3 material could affect both the performance and variance in microstructure and stability. Moreover, Swainson and
long term stability is also another issue. co-workers found a phase transition of CH3NH3PbBr3 just below
Moreover, the poor resistance of CH3NH3PbI3 and a device 1 GPa and amorphorization around 2.8 GPa, without the process
based on this material towards the combined effect of all these undergoing a long-range orientation ordering of cations.123 The
operating conditions can be even more problematic. This may volume reduction under compression was attributed to the
restrict the use of these compounds for many practical applica- tilting of the PbBr6 octahedra. Suga and co-workers also gave
tions and commercialization. Interestingly, the beneficial or the pressure–temperature phase relations of CH3NH3PbX3 crys-
detrimental behavior of grain boundaries in CH3NH3PbI3 film tals in the range between 0.1 Pa and 200 MPa in detail.124 The
still remains under debate. Previous reports on perovskites crystalline phase of the Omh-perovskite can change due to
have demonstrated that grain boundaries are less harmful than environmental conditions, such as temperature and pressure,
in other semiconductors,112,113 and are rather beneficial for the which directly affects the stability of Omh-PSCs. In order to
collection of perovskite carrier efficiency.114 On the other hand, improve the stability of Omh-PSCs, emphasis should be given on
it is reported that single-crystal perovskites demonstrated a gaining a comprehensive understanding of the crystal structure
rather higher performance115–117 illustrating a progress in carrier of Omh-perovskite under different conditions such as moisture
lifetime.118,119 Conversely, current reports suggested that grain and temperature, as well as the relationship between the crystal
boundaries are connected with PL quenching, demonstrating structure and pressure. In general, the poor structural stability
that they are not as kind as have been recommended in the past. of Omh-perovskites could be improved by modifying the
Instead of this, they destruct the carrier lifetime.120 These Goldschmidt tolerance factor (tf) using appropriate substitutions,
interesting debates on the role of grain boundaries for carrier which can be defined by the following mathematical expression
lifetime and performance in perovskite materials are future in eqn (20),125,126
interesting topics of study. Despite their impact on carrier rA þ rx
lifetime and performance, the contributions of these grain tf ¼ pffiffiffi (19)
2ðrB þ rx Þ
boundaries towards operational and long term stabilities of
Omh-PSCs are not yet known. where rA, rB, and rx are the radii of cation A, cation B, and
halogen, respectively, in the ABX3 structure. That equation is
4.1.1.2 Tribromide anions. MAPbI3 reverts to its precursors applied at room temperature to the empirical ionic radii. In order
due to its hydroscopic nature.33 However, the bromide-based to stabilize the cubic structure, the tf of Omh-perovskites should
perovskites have been proved to be less moisture-sensitive.121 A be closer to unity.127 To form a stable Omh-perovskite struc-
higher stability or resistance to photobleaching and decompo- ture, the size of the ionic radius is restrained by the tolerance
sition was reported after exposing MAPbBr3 films to conditions factor. The ionic radius is the most important ionic parameter
of stress (i.e. light intensity, and temperature).122 The better that dominates the crystal structure of ionic compounds.

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

The ideal cubic structure may be seen as a network of BX6 MAPbI3 exhibits a tetragonal shape at room temperature as a
octahedra, where all the octahedra are corner sharing. The ‘A’ result of the distortion of the cubic crystal.135 The organic cation
ions occupy the cubo-octahedral holes in between the octa- was considered as it does not take part in determining the band
hedra. Many derivatives of the ideal structure are found and the structure, and works to fulfill charge neutrality within the
distortion of the structure from cubic is often considered to lattice.136 However, its size is quite vital. The size of the organic
be determined by the relative sizes of the different ions of the cation can cause the entire network to enlarge or be compressed.
compound. FAPbI3 is not sensitive, even at high temperature, compared
to MAPbI3. While MAPbI3 discolors in 30 minutes, FAPbI3 does
4.1.2.1 Metal (B) cation. The organic cation CH3NH3+ in not discolor even at 150 1C under ambient conditions. Subjecting
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

ABX3 is responsible for the structural stability of the Omh- the material to a moist atmosphere resulted in fading of the
perovskite materials, while the electronic properties are largely FAPbI3, the same as for MAPbI3.137 However, recent work on the
determined by the inorganic matrix (metal–halide hybridized long-term stability of Omh-PSCs suggests that with sufficient
orbital).102,128,129 The main problem associated with MAPbI3 is encapsulation, FAPbI3 functions continuously under illumina-
its poor ambient stability under different environmental con- tion for thousands of hours, and this is not likely to act as a
ditions.24,33 In order to avoid this limitation for both indoor serious limitation to commercialization.5,6 Furthermore, superior
and outdoor application, any modifications that can be made to temperature stability is exceptionally promising towards long
strengthen the material stability would be beneficial. One term thermal durability, and this would be an additional topic
successful technique is to combine PbI2 with SnI2 to form that needs to be studied. Stable Omh-PSCs with a cationic alloy
CH3NH3SnxPb1xI3, which has an Eg value of 1.1 eV. The edge structure have recently been fabricated in the laboratory.4 A
of the incident-photon-to-current efficiency (IPCE) curve extended 5-aminovaleric acid (5-AVA) cation replaced some of the MA
to 1060 nm.25,82,130 Nevertheless, Sn2+ easily changes to Sn4+ in the cations in the cuboctahedral site of MAPbI3, forming the new
atmosphere. Consequently, this leads to device deterioration. mixed cation Omh-perovskite, (5-AVA)x(MA)1xPbI3. A more stable
Other Omh-perovskites incorporating divalent metal cations such new Omh-perovskite, (PEA)2(MA)2[Pb3I10] (PEA = C6H5(CH2)2NH3+,
as Be2+, Mg2+, Ca2+, Sr2+, Ba2+, Zn2+, Ge2+, Fe2+, Co2+ and Ni2+ do MA = CH3NH3+) has also been reported.138 Films of these materials
not yet report on this application. Another tactic is to replace the are more moisture resistant than films of MAPbI3 and devices
organic cations, as it has been established that enlargement of can be fabricated under ambient humidity levels. This moisture
the octahedral network significantly influences the bandgap of resistance ability may be due to the more hydrophobic tail ‘R’
the Omh-perovskite. group, which may mask the hydrophilic nature of the materials.
However, the fundamental reason for alloy stabilization of the
4.1.2.2 Organic ‘A’ cation. The organic cation in ABX3 is a key structures requires further study.
part of the Omh-perovskite that determines its structure and
dimensionality, and has a direct influence on the stability and 4.1.2.3 Mixed anion halides
opto-electronic properties of the material. The cubo-octahedral MAPb(I1xBrx)3. The stability of MAPbI3 can be significantly
cavity, defined by the four edge-sharing BX6 octahedra, only improved with a small fraction of Br or Cl.33,139 Although the
permits the incorporation of small cations into the 3D pero- MAPbI3 hybrid solar cell does not show significant PCE degra-
vskite structure. The phenomenological derivation of tf allows dation at low humidity (o50%), the MAPbI3 began to decom-
an estimation of the stability based on the size of its constitu- pose at relatively high humidity (Z55%), displaying a color
ents.131 To obtain a cubic phase, tf should be unity; however, change from dark brown to yellow.33 Hence, the solar cells were
for most of the cubic structures the value of tf ranges from intentionally exposed to a relatively high humidity (55%) for
0.78–1.05, due to a slight expansion in distorted structures.132 one day while keeping the humidity to 35% on the other days.
In fact, by changing rA, tf can be varied but only in a restricted Interestingly, the MAPb(I1xBrx)3 (x = 0, 0.06) hybrid solar cells
range of values around unity (tf = 1 corresponds to a perfectly exhibited serious PCE degradation after exposure to 55%
packed Omh-perovskite structure) to have a stable and evenly humidity, whereas the other MAPb(I1xBrx)3 (x = 0.2, 0.29) cells
distorted 3D Omh-perovskite structure.133 The dynamic position maintained their PCE. A low sensitivity to humidity of the cells
conduction band of CH3NH3PbI3 films, due to the disordered based on MAPb(I1xBrx)3 (x Z 0.2) might be associated with
organic cation, plays a major role in extending the lifetime.134 their compact and stable structure, because the substitution of
So far, the most investigated organic cation counterpart for larger I atoms with smaller Br atoms in MAPb(I1xBrx)3 leads to
the substitution of MA (CH3NH3+) has been FA (HC(NH2)2+), a reduction to the lattice constant and a transition to a cubic
which gives rise to a tf of 0.99, higher than tf = 0.91 for phase. As a result of introducing Br ions into the perovskite
CH3NH3PbI3 (ion size: FA – 2.53 Å, MA – 2.17 Å, Pb2+ – 1.19 Å structure, the lattice parameter changes from 5.921 for the
and I – 2.20 Å). In a recent report, the FA cation, which is CH3NH3PbBr3, 6.144 for the MABr/MAI 2 : 1, and 6.223 for the
slightly larger than methylammonium, was confirmed to form MABr/MAI 1 : 2. The change in the lattice parameter is due to
a 3D Omh-perovskite with a lower bandgap of about 1.47 eV. the difference in the ionic radii of Br (1.96 Å) and I (2.2 Å).140
Omh-perovskites are described as any compound which crystal- The smaller ionic radius of Br is the main reason for the
lizes in the ABX3 structure, containing corner-sharing BX6 octa- formation of the cubic structure when Br is introduced into the
hedra with the cationic component neutralizing the total charge. perovskite structure.141 Thus, the role of bromine inclusion

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

in thermal, light and ambient stabilities should be identified. composition measurements, e.g. XRD, EDS may miss the residual
Such a comparative study could help towards better stability chloride remaining at the Omh-perovskite’s surface.148
improvements.
Cl + CH3NH3+ " CH3NH3Cl (s) (20)
MAPbI3xClx. In contrast to CH3NH3PbI3, this iodide–chloride CH3NH3Cl (s) " CH3NH3Cl (g) (21a)
mixed-halide Omh-perovskite was remarkably stable to processing
in air. The absorption spectra demonstrated good light-harvesting CH3NH3Cl (s) " HCl + CH3NH3 (21b)
capabilities over the visible to near-IR spectrum and were also Sublimation of CH3NH3Cl (eqn (21a)), or its decomposition
stable to prolonged light exposure, as demonstrated by 1000 h into HCl and CH3NH2 (eqn (21b)) may be responsible for the
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

of constant illumination under simulated full sunlight.6,139 The

loss of chloride during film growth. Thus, the inclusion of Cl,
improved stability can be associated with its compact and from either metal chloride (e.g., PbCl2) or organochloride such as
stable structure, due to the substitution of larger I atoms with CH3NH3-Cl, has become the most commonly applied strategy.
smaller Cl atoms in MAPbI3xClx. This leads to a reduction of In terms of how Cl additives such as CH3NH3Cl contribute
the lattice constant and a transition to a cubic phase, which is to improve stability and/or reduce degradation, several reasons
similar to the substitution of iodide using bromide.122 This com- are considered. (1) Cl inclusion could increase the binding
pact structure would decrease the sensitivity of the MAPbI3xClx constant, which may reduce the rapid dissolution or degrada-
material to moisture, light and temperature due to its higher tion of CH3NH3PbI3 in moisture. This fact might be related to
binding constant compared to the MAPbI3 materials. But these
the stable structure of CH3NH3Cl, owing to a more compact
need to be investigated. Overall, there is no clear evidence of Omh-perovskite cubic structure in which the degradation of the
the role of chlorine insertion to enhance stability. Both MAPbI3 methylammonium cation is prohibited.122,149 (2) It facilitates
and MAPbI3xClx have similar light absorption ranges, while the removal of excess CH3NH3+ ions.49 Excess CH3NH3+ ions may
MAPbI3xClx has a longer recombination lifetime than MAPbI3, act as defects or impurities at surface or interface sites, which
followed by a longer electron diffusion length. In this regard, may lead to the blocking of interfacial processes. (3) Cl-rich
MAPbI3xClx does not require a mesoporous electron transport nucleation sites lead to better crystal coalescence146 or high
layer (e.g. TiO2) unlike MAPbI3.26,112 The long electron diffusion quality film crystallinity,49 and also provide a stable morphology
length enables MAPbI3xClx to work perfectly on insulating
as well as a better coverage in Omh-PSCs fabricated by one-step
scaffolds and in bulk films.34,142 solution processes (because poor morphology stability and poor
MAPbBr3xClx. Among the family of Omh-perovskites, coverage could lead to device deterioration). (4) It enhances the
MAPbBr3xClx has a larger band gap than iodine based Omh- lifetime of the photoexcited species (the presence of chlorine
perovskites such as MAPb(I1xBrx)3 and MAPbI3xClx. MAPbBr3xClx may creep grain boundaries120 and ultimately controls pero-
has enormous capability for high energy photons in photo- vskite grain structures150). All of these factors contribute to
voltaic applications such as tandem cells, or in other devices higher stability. Despite this helpful information, there is no
which possess photon spectral splitting.143,144 But the Omh- clear evidence on the chlorine inclusion/stability relationship. In
perovskite materials suffered from a moisture-related decom- general, mixing halides in Omh-perovskites has shown benefi-
position because of the hygroscopic amine salts.33,145 In a cial effects in terms of enhancing stability. CH3NH3PbI(1xClx)3
stability test, MAPbBr3xClx was found to be less moisture- is remarkably stable during processing in air compared to
sensitive.121 The as prepared MAPbBr3xClx films were left in CH3NH3PbI3. The atomic ratio of Cl in CH3NH3Pb(I1xClx)3
the dark at room temperature with exposure to ambient air for was proposed to be up to one third, as it is in precursors.97,139
30 days. The materials did not show any apparent changes in However, recent experiments142,151 showed that its atomic ratio
the XRD patterns observed after this time, indicating the good is less than 4% and that most of the precursor Cl may be lost.
stability of MAPbBr3xClx. This significantly improved stability Therefore, the ratio of Cl into iodine should be seriously con-
may be due to a small fraction of Br or Cl substitution.33,139 It is trolled. For real practical applications, the miscibility between Cl
not clear which atom contributes more to the structural stability. and iodine mixed halides, and the fundamental contributions of
Similarly, the thermal and radiation stability of MAPbBr3xClx is Cl to stability need to be addressed. Similarly, different devices
not yet resolved. Hence, detailed evidence for the contributions of incorporating CH3NH3Pb(I1xClx)3 material showed different
bromine and chlorine to thermal and radiation stability of the performances – one has inferior performance but the other
perovskite structure and to the thermal and photoconductivity improved performance.150 Thus, how the chlorine inclusion
are required. mechanisms during fabrication processes affect device stability
is not yet well understood. Whether bromide sources could act
4.1.2.4 Chloride inclusion. The issue of chloride loss has in a similar manner to chloride sources as additives in Omh-
recently been raised.49,51,146 Currently considered loss pathways PSC fabrication is also another important open question for the
center around MACl49,51,147 sublimation, or the degradation of community.
MACl into volatile hydrochloric acid (HCl) and methylamine 4.1.3 Hole transport layer. Hole transporting layers such as
species, facilitated by residual water.83 The loss mechanism spiro-MeOTAD, PEDOT:PSS and other polymers, and inorganic
suggests that chloride must diffuse into the film’s surface to materials are significantly more stable than liquid electrolytes,
ultimately escape the system, meaning that bulk sensitive but devices based on these materials are also susceptible to

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

chemical degradation, during which photooxidation processes

and hole–electron pair recombination may occur. In this section,
we mainly focus on the most commonly used HTMs, spiro-
MeOTAD and PEDOT:PSS.

4.1.3.1 Spiro-MeOTAD. Spiro-MeOTAD is a material widely

used in Omh-PSCs for hole transporting purposes. Spiro-
MeOTAD is coated beneath the counter electrode and can easily
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

be exposed to the atmosphere. As a result, degradation can take

place due to air exposure, continuous light irradiation, elevated
temperature and dust. Qi and colleagues investigated the causes
of degradation of Omh-PSCs which use spiro-MeOTAD as the
Scheme 1 Proposed oxido-de-sulfonato substitution of PSS in PEDOT:PSS
hole selective material.152,153 Due to its amorphous properties,
layers. Reprinted from ref. 159. Copyright 2008, with permission from Elsevier.
spiro-MeOTAD is a crucial HTM; on the other hand, small
molecular species from the air could simply spread, and pass
freely through miniscule pinholes in the amorphous material. 4.1.3.2 PEDOT:PSS. Poly(ethylenedioxythiophene):poly(styrene
Unfortunately, these pinholes are too small to be seen with an sulfonic acid) (PEDOT:PSS) is generally used in the form of a
optical microscope.152,153 Moreover, the decline in Jsc and FF is water solution/suspension for solar cells as the HTM,155 and
possibly caused by Omh-perovskite layer degradation, which can has recently been used in inverted Omh-PSCs as a hole selective
be caused by (1) air molecules (O2, H2O, etc.) migrating and material.156–158 This hygroscopic property of PEDOT:PSS desta-
interacting with Omh-perovskite via the pinholes and/or (2) out- bilizes the interface and will result in device deterioration.
diffusion of mobile ions in the Omh-perovskite film via the Moreover, compared to PEDOT, the amount of PSS usually
pinholes in the spin coated spiro-MeOTAD.154 The pinholes may dominates in the polymer mixture and the diffusion of PSS
create pathways for foreign molecules such as water and other into other layers can occur and possibly result in degradation. For
gas molecules in air to diffuse through the thin film, as shown in instance, the PSS may undergo a oxido-de-sulfonato-substitution
Fig. 12. These air molecules would then become impurities forming the phenolate which can then react with PSS, forming
in the solar cells, leading to degradation and resulting in a drop two PSS chains linked together via a sulfonic ester group,
in solar cell efficiency. as shown in Scheme 1.159 The asterisk in Scheme 1 indicates
It is now clear that the presence of oxidized spiro-MeOTAD is oxidation of the carbon. Additionally, corrosion will be enhanced
required for Omh-PSCs. However, oxygen reduction reactions not only in connection with low work function metals such as Al
might also occur. The generation of O2 would change the or Ca used as a counter electrode, but also in connection with an
oxidized spiro-MeOTAD concentration and device characteris- ITO electrode.56,160
tics during the operation. For the future development of devices Dauskardt and coworkers161 verified how the hygroscopic
with long-term stability, it will be important to exclude oxygen properties and poor adhesion of PEDOT:PSS onto the active
reduction. Further work is needed to determine if this can be layer causes a general loss of device performance in a roll-to-roll
achieved by sealing the devices after preparation under ambi- processed normal geometry organic photovoltaic devices. The
ent atmosphere, or whether the devices should be fabricated objective was to show that the thermomechanical stress in a
and sealed to exclude oxygen and moisture. real life device would suffer.
4.1.4 Metal counter electrode. Gold, silver and aluminum
metals are commonly used as counter electrode materials in
mesoporous, planar and inverted Omh-PSCs.157 The drawbacks
of silver may be its corrosion in contact with the halide ions
from the Omh-perovskite light absorber, forming silver halides
such as AgCl in humid environments, and short circuits or
shunting paths with the mesoporous TiO2, leading to device
degradation.6 The high cost of a Au electrode also requires a
high-vacuum evaporation technique, thereby limiting its future
application. Low-cost carbon may be an ideal material to sub-
stitute Au as a back contact with Omh-perovskite heterojunction
(HJ) solar cells at low temperature, because its function is similar
to that of Au.7

4.2 Effect of device architecture

Both the device stability and efficiency of Omh-PSCs depend on
Fig. 12 Diffusion of foreign air molecules into the device through spiro- their architecture, which will in turn influence the choice of
MeOTAD. materials and deposition methods. Three main device architectures

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

Fig. 13 Stability comparison of three Omh-PSC device architectures.

present in Omh-PSCs: the planar (regular TiO2 based), inverted to the Omh-perovskite, as electron selective layers, respectively.167
(PEDOT:PSS based) and mesoporous structures, are schemati- The ambipolar properties, especially the superior p-type character
cally represented in Fig. 13. Since the initial evolution of Omh- of CH3NH3PbI3 which is greatly biased by the p–n heterojunction
PSCs, many research groups have been focusing on efficiency idea in organic solar cells, was the motivation for the authors’
and fabrication processes; hence, several review papers are initial material choice.168 The planar heterojunction architecture
available.96,162–165 In this section we will only focus our discus- with an inverted design achieved a higher PCE in a very short
sion on the device architecture stability. Investigation of device time, from 3.9% to over 16%.157 PEDOT:PSS has high conduc-
architecture, which remains unclear in terms of its effects tivity and is extensively employed as an electrode in organic
on the device stability, should also be included in long-term electronic devices such as organic light-emitting diodes, organic
stability tests. photovoltaics and organic field-effect transistors.169–171 However,
4.2.1 Planar (regular) TiO2. The planar (or alternatively PEDOT:PSS is extremely hygroscopic, and thus dispersing it in
regular) architecture inherits features of the most commonly water may limit the long-term stability of inverted Omh-PSC
utilized mesoscopic Omh-PSC device architecture (FTO/cl-TiO2/ devices. A direct coating of PEDOT:PSS aqueous solution would
perovskite/spiro-MeOTAD/Au). Primarily, any excellent electron degrade the CH3NH3PbI3 film.
transporting layer can be deposited on fluorine doped tin oxide Therefore, it can be proposed that inverted based Omh-PSC
or other substrate, and then after the Omh-perovskite layer, the devices may face challenges such as: (1) electrochemical reac-
HTM and finally, the counter electrode, are coated in chrono- tions occurring at the ITO and aluminum electrodes, (2) the
logical order. The application of a single compact n-type metal inclusion of moisture and impurities within the organic–metal
oxide layer makes the regular type different from the mesoporous interface enhancing ionic conduction and hence accelerating
devices which consist of both compact and meso-structured corrosion, (3) accelerated degradation due to illumination of the
scaffolds. The regular configurations have been observed to boast device, (4) the very high electron affinity of PCBM (and other
high efficiencies, similar to their mesoscopic counterparts. Their fullerenes), which would seem more prone to reaction with the
stability, however, is still in minutes94 compared to their meso- metal electrode, and finally (5) the CH3NH3PbI3 film is destroyed
scopic counterparts. This may be due to: (1) faster charge collec- by the hygroscopic PEDOT:PSS aqueous solution. Aluminum is
tion at the interface, due to ion migration resulting from an not a passive metal and it may react with organic compounds
applied bias or field, creating transport barriers in the planar producing organo-aluminum compounds or alternatively anion
architecture and (2) a strong hysteresis effect in planar devices: radicals. These compounds are highly reactive species that will
undesirable shifts in electrical conductivity that occur when react with any proton donors present (e.g. a trace of water or with
applying increasing or decreasing amounts of voltage to a oxygen). None of these challenges are confirmed, thus they should
cell.166 Thus, in detailed studies on the stability of the compact be investigated to obtain a detailed understanding of the stability
TiO2-perovskite and Omh-perovskite–HTM interface stability, of inverted Omh-PSCs. A 10–20 nm ultrathin layer of poly(N,N 0 -
pore filling properties, sensitivity towards light and moisture, bis(4-butylphenyl)-N,N 0 -bis(phenyl) benzidine)(polyTPD) was
formation of a shunting path with the HTMs and/or metal introduced between PEDOT:PSS and CH3NH3PbI3 layers using
electrodes and carrier accumulation due to excess ions, and the a meniscus-coating process.40,172 A 10–20 nm ultrathin hybrid
formation of trap states and defects at the Omh-perovskite and interfacial layer of a NiO compact layer and an insulating Al2O3
compact TiO2 surfaces, the effects of the presence of excess ions (meso-Al2O3) scaffold, which leads to little hysteresis and more
in the Omh-perovskite on the device architecture will be critical. stable power output under working conditions has been con-
4.2.2 Inverted (PEDOT:PSS based). Chen and coworkers firmed for an inverted device.156 Furthermore, in comparison
pioneered an inverted based device architecture made of to PEDOT:PSS based inverted Omh-PSCs with the best efficien-
PEDOT:PSS coated directly on the transparent conductive sub- cies (B15%),173,174 the meso-Al2O3 device showed an improved
strate and C60 derivatives such as [6,6]-phenyl C61-butyric acid stability. This shows that the potential of the NiO/meso-Al2O3
methylester (PCBM) and indene-C60 bisadduct (ICBA), coated next interfacial layer should be superior to that of PEDOT:PSS.

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

4.2.3 Mesoporous TiO2. The stability of Omh-PSCs in

liquid electrolyte is approximately 10 min because CH3NH3PbI3
tends to dissolve gradually in the redox electrolyte (I/I3). In
order to solve this problem, Grätzel, Park and coworkers inves-
tigated a new mesoporous solar cell based on a CH3NH3PbI3
semiconductor and spiro-MeOTAD as a solid HTM, and sub-
micron thick films of mesoporous titanium dioxide, showing Fig. 14 Typical architecture for Mesoporous TiO2 based HTM free
excellent long term stability of 500 h under AM 1.5 G illumina- Omh-PSCs.
tion. The mesoporous device structure incorporates active and
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

passive scaffolds such as mesoporous TiO2 and mesoporous

their conductivity. Several groups have reported Omh-PSCs
Al2O3 scaffolds, respectively. The highest stability reported with
without additives in the HTM, which simplifies the cost and
mesoporous TiO2 is 1000 h with a 28% reduction relative to its
fabrication procedures. Liu and co-workers described a tetra-
initial PCE.44 There may be many factors causing degradation
thiafulvalene derivative as a HTM (TTF-1).185 Moreover, iodine
relating to mesoporous TiO2. Some of them are:175–177 (1) the
can react with the TBP leading to an iodopyridinate complex, as
sensitivity to UV exposure, and (2) the presence of surface defects
in dye sensitized solar cells,186 resulting in iodine depletion from
(oxygen vacancies) in TiO2. The interaction of these defects or
the absorber materials. Han and coworkers introduced a TTF-1
vacancies with both water and oxygen, creating electronic defect
derivative into Omh-PSCs without the use of p-type dopants.185
states, can result in device degradation. Oxygen binds to the Ti3+
The improved stability was attributed to the avoidance of the use
sites and modifies these trap sites.72 Furthermore, lead iodide
of deliquescent and hygroscopic additives. This dopant free
can be produced upon the breakdown of Omh-perovskites at the
approach is responsible for the improved lifetime of the device
interface.51 Because of its wide bandgap (Eg = 2.3 eV) and poor
compared to the approach incorporating p-type HTM dopant
optical properties, the formation of PbI2 is unnecessary. For
additives. Moreover, these additives may act as a source of
reasons that are not yet completely known, the formation of a
moisture that could degrade the organic HTM materials and
PbI2 phase is more well-known in a mesoporous scaffold than on
shorten the device lifetime. Similarly, these HTMs are not only
a planar substrate.178 Thus, mesoporous based solar cell devices
expensive, but can also limit the long term device stability.
prepared with PbI2 have confirmed poor performance, attributa-
Although various novel HTMs have shown promising results, the
ble to its appreciably inferior transport properties (smaller diffu-
HTM-free Omh-PSC device schematically illustrated in Fig. 14 has
sion coefficient) compared with CH3NH3PbI3-based devices.179
many advantages, such as lacking a need for air-sensitive HTMs,
Moreover, the presence of PbI2 might result in energy misalign-
its simplicity, cheap cost of production and excellent stability.
ment at the TiO2 interface and the trapping of free charges created
Meng and coworkers demonstrated HTM free Omh-PSC devices
in the Omh-perovskite phase, resulting in an inferior device
with 10.5% PCE.187 The excellent carrier lifetime and ambipolar
performance. It may also give more stringent requirements for
nature25,26 enabled the pioneering of the HTM free architecture.
the processing of these thin films.180
Interestingly, HTM-free structured mesoscopic solar cells have
The two devices for mesoscopic architecture Omh-PSCs are
now been made with PCEs over 10% and stabilities of over
HTM based, and HTM-free structured solar devices. In the
2000 h.7 Etgar and co-workers showed an initial study using
former, the breakthrough for better stability of the mesoporous
TiO2 nanosheets as the mesoporous layer.22
architecture was 500 h using spiro-MeOTAD, which is now widely
used as an HTM with excellent PCEs. However, due to the limited
choice of materials, the progress in stability of inorganic HTMs has 4.3 Effect of interface
been sluggish compared to their organic counterparts. Although Although studies have revealed that device degradation is mainly
CuSCN (PCE = 12.4%),181 NiO (PCE = 11.6%)182 and CuI (PCE = attributable to moisture ingress, high temperatures, and
6%),183 have been listed as excellent, low cost and stable inorganic UV light, significant degradation of Omh-PSC devices is still
HTMs, they still have low efficiency when compared to organic observed even when they are well-encapsulated or characterized
HTMs. On the other hand, polymer HTMs, in addition to the most in an inert environment.6 These observations imply that the
commonly used spiro-MeOTAD, including poly(triarylamine) environmental effects are not the only critical factors deter-
(PTAA) and poly(3-hexylthiophene-2,5-diyl) (P3HT), have been mining device stability, but that there are also intrinsic factors
tested, because of their better hole mobility and good film- within the devices that contribute to degradation. One example
forming properties. Among all, PTAA was confirmed to have the is the unstable electronic structure of the organic/cathode inter-
highest efficiency of up to 12%, in contrast to 6.7% for P3HT.32 face, due to the gradual diffusion of metal into organic films.
This efficiency was seriously decayed in 55% relative humidity Similar phenomena are observed in ITO anode/organic inter-
over one day. This was the first most important breakthrough in faces, where reactive oxygen species, generated by UV treatment
the development of polymer HTMs, which undergo pore-filling of the indium-tin oxide (ITO) substrate, slowly degrade the
limitations, and this may lead to a diminishing of the perfor- organic layer. Other degradations may be related to the organic/
mance, an inferior stability, and could even harm the device.184 electrode interface, and time-dependent chemical reactions at
The majority of HTMs use hygroscopic additives such as Li the donor/acceptor interface can also cause device deteriora-
salts, tetrabutylpyridine (TBP) or a cobalt complex to enhance tion. Other causes may be attributed to the spatial isolation of

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

organic films from electrodes and the reduction of the energetic Snaith6,191 demonstrated that silver contacts often form shunt-
transport barrier. ing paths with mesoporous TiO2 in the absence of oxygen. As a
4.3.1 Ion migration. The rate of electromigration increases result, a Schottky barrier can be formed. Similarly, Snaith and
with temperature, electric field and illumination, and has four coworkers192 confirmed that metal migrating through the HTM
prerequisites – mobile ions, a voltage gradient, a continuous layer is partially responsible for the quick device degradation.
flow of moisture, and soluble ions. Silver is the metal most They showed that depositing the HTM within an insulating
susceptible to migration, since a low activation energy is required mesoporous ‘‘buffer layer’’ comprised of Al2O3 prevents the
to initiate the migration process.188 The severity of electro- migration, while also facilitating control of the HTM thickness.
migration may increase with the applied potential gradient and This prevents the device from degradation, even after 350 h of
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

the time to reach active materials. When Ag is under an applied operation.

electric field, or thermal evaporation, it may leave its initial
location in an ionic form and redeposit at another location.
Silver migration causes short circuits and device failure, espe- 5. Suggested approaches to improve
cially in high humidity environments. Moreover, it is well known stability of Omh-PSCs
that the cations at the A-sites and B-sites of Pb-based complex
ABX3 perovskite minerals such as PbBO3 can be replaced by Not only efficiency, but also stability is a critical requirement for
many other cations. Substitution is mainly determined by the the practical application of Omh-PSCs. It is, therefore, necessary
differences in valence and ionic radius.188 The size of Ag+ is to improve stability to an extent that will make Omh-PSC technol-
almost the same as Pb2+ (about 0.149 nm), and much larger than ogy attractive from a commercial standpoint. Based on current
B-site cations such as Nb5+ (0.064 nm) and Zn2+ (0.074 nm).189 information, device instability is the sum of all the device com-
Thus, it can be proposed that Ag+ enters into the Pb-site of ponent materials, device architecture and interfaces. Stability
MAPbX3 Omh-perovskite materials, substituting for Pb2+, though improvement therefore can be expected to result mainly from
this has not yet been reported. The nonequivalent replacement Omh-perovskite materials, device architectures and interfaces.
of Ag+ for Pb2+ may also be responsible for undesirable MAPbX3 Researchers must focus on two major issues regarding the
layer structural distortion during device operation. The intrinsic stability and longevity of Omh-PSCs: firstly, strategies to provide
low melting point of silver may account for the structural change extrinsic stability through the use of proper encapsulation to
and abnormal grain growth of the Omh-perovskites near the prevent degradation caused by ambient oxygen and water, and
interface and this has to be investigated. secondly, research to improve the understanding of the intrinsic
The role of ion migration initiated by photo-excitation in stability of different thin film perovskite materials, array of the
slower processes occurring during pre-conditioning of the devices device, and interfaces. The most suitable approaches that should
has been proposed as a cause for hysteresis.94 Hysteresis com- be considered include chemical engineering modifications to
monly takes place in metal halides such as PbI2, in which photo- active materials, design and optimization of device architecture,
excitation generates halide vacancies, facilitating halide ion and systematic interface engineering. These approaches should
migration.190 Furthermore, anions such as I ions move towards provide a means to improve both the operation and long term
the electron selective layer, i.e. TiO2, and cations such as MA+ stabilities of Omh-PSC devices.
and Li+ migrate into the HTM contact when a reverse bias is
applied during illumination.94 This causes an injection barrier 5.1 Omh-perovskite material aspects
for both electrons and holes at their respective interfaces, The stability of Omh-perovskite materials is an issue in Omh-
resulting in hysteresis. PSCs. Both chemical composition and structural engineering
4.3.2 Reactivity. Chemical reactions and inter diffusion at techniques could be considered for issues related to material
the interface can greatly affect the overall device stability. Indeed, stability. Appropriate chemical composition engineering is an
of all types of instabilities in the cell, the interface instability is essential approach to stabilize Omh-perovskite materials by the
the one where its chemistry is almost not yet investigated in chemical modification of the X site anions and of the organic
Omh-PSC devices. It is, therefore, important to consider detri- ‘A’ site cation.193 The size of the A cation is critical for the
mental interfacial reactions in detail to find out the intrinsic formation of stable close-packed Omh-perovskite structures. As
instabilities. The possible origins of degradation that could take aforementioned, the relative ionic radii of A, B and X in the
place at the interface of Omh-PSC devices, leading to interface ABX3 Omh-perovskite structures have been widely used in a
instability, are: (i) corrosion of silver forming silver halides at the method of establishing the distortion of the MX6 octahedron;
interface; (ii) silver connections forming shunting paths with the in particular, a relatively smaller ionic radius for X favors the
mesoporous TiO2 in the absence of oxygen; (iii) photo-oxidation formation of cubic structures.24 In particular, the A cation must
of the HTM organic materials degrading the electron/hole trans- fit into the space composed of four adjacent corner sharing
port properties and thus the photovoltaic stability; (iv) Omh- MX6 octahedra. The structural and opto-electrical differences
perovskite decomposition at the TiO2/CH3NH3PbI3 interface of MAPbI3 and FAPbI3 are likely to originate from the difference
through light exposure, since TiO2 has a strong ability to extract in the ionic radius of the MA (1.8 Å) and FA ions (1.9–2.2 Å).
electrons from iodide (I); and (v) formation of metal oxides Such structural instability of Omh-perovskites can be improved
which finally leads to device deterioration. Docampo and by regulating the Goldschmidt tolerance factor (tf) using

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

appropriate substitutions, which is a useful indicator of the and co-workers7 replaced the typical Ag or Au back-contact with
stability and distortion of crystal structures. Another technique carbon paste deposited by a doctor-blade. Most dramatically,
is structural engineering of the Omh-perovskite absorbers, to through the use of functionalized carbon nanotubes embedded
produce moisture resistant 3D structured analogous materials. in an insulating polymer matrix, Habisreutinger and co-workers79
The recently reported new layered (2D) structure Omh-perovskite were able to protect Omh-perovskite devices from a steady
absorbers (PEA)2(MA)2[Pb3I10] (PEA = C6H5(CH2)2NH3+, MA = stream of running water for a short time. Although these simple
CH3NH3+) may offer greater tunability at the molecular level Omh-PSCs are promising photovoltaic devices, several HTM
for material optimization and are more moisture resistant than and metal counter electrode-related drawbacks hinder their
films of commonly used (MA)[PbI3] materials.80 Devices contain- practical application.
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

ing these layered materials are stable even in moist environments. Moreover, to accomplish the necessities for commercial
As a result of the tunability behavior of the layered structure, these applications, two of the most important issues in Omh-PSCs
materials may be capable of better stability for future develop- should be addressed.15,142,198 Firstly, devices incorporating HTMs
ment. Understanding the multifaceted and interplaying chemistry require an undesired cost due to the more expensive price of
of Omh-PSCs, and developing new materials based on such HTMs such as spiro-MeOTAD and the most conserving issue of
knowledge will indisputably open up astonishing stability break- poor device stability. The second issue is the counter electrode
throughs and revolutions for this type of solar cell. made of Ag or Au, which have issues relating to: (1) silver
corrosion and the attack of perovskites forming silver halides,
5.2 Device architecture aspects (2) noble metals such as Au or Ag are not applicable for large-
While the absorber material and interface stability are essential scale production due to the deposition technique, i.e. thermal
requirements for long term stability, they are by no means evaporation, which initiates hot metal to diffuse in to different
sufficient, because the end user is ultimately interested in the layers and is highly energy consuming and complicated, and
device stability for a given application. Thus, the engineering of (3) the high cost of the Au electrode also requires a high-
suitable device architectures is required. Among the three Omh- vacuum evaporation technique, thereby limiting its future
PSC device architectures – mesoporous, regular and inverted, commercialization. Therefore, the current goal focuses on the
the mesoporous one incorporating a carbon top electrode with- removal of the HTMs and substitution of these metal electrodes
out hole conductor is stable, low cost, simple and promising for to improve stability, simplify the cost of device fabrication and
practical applications. recognize large-scale fabrication. Carbon-based electrodes have
So far, for Omh-perovskite-based solar cells, PCE values been extensively investigated, due to its high earth abundance,
ranging from 3.81% to more than 20% have been obtained by low cost, low-temperature processing (100 1C), large-scale print-
optimizing the technology and array of device structure, making it ing processing34,113,199 appropriate work function of 5.0 eV
competitive for future commercialization. Zhou and coworkers36 (5.1 eV for Au),7,200 and promising long-term stability.4,7 This
reported an 80% drop in the PCE over a 24 h period when the remarkable stability can be attributed to the thick carbon
device with planar geometry was stored under ambient condi- cathode layer, which can work as a water-retaining layer to
tions, and a 95% drop in the PCE after 6 days. While this loss in protect the perovskite material from being destroyed.201 This
performance is highly problematic, even more concerning is that result indicates the underlying potential of commercial carbon
the decomposition product (PbI2) is sparingly soluble in water paste as a promising candidate for stable and highly efficient
(1.6 mM at 25 1C);194 this would present an enormous ecotoxi- Omh-perovskite solar cells.201 Furthermore, many literatures
cological problem if a module becomes damaged in the field. have clearly established that Omh-PSCs can act as a p–i–n device
With these issues in mind, a substantial effort has gone into and thus HTMs are not the first priority for operation.112,202
improving the long-term stability of Omh-PSCs.79,83,195 Protec- Besides, Omh-perovskite has been considered as both a light
tive Al2O3 layers83 and hydrophobic oligothiophene HTMs195 harvester and a hole transporting material.34,142,167 The electron–
have both been employed as a means of protecting the under- hole diffusion length of CH3NH3PbI3 can even exceed 100 nm.25,26
lying Omh-perovskite film, while alternative Omh-perovskite All of this evidence indicates that complicated HTMs are not
compositions33,80 have also been explored as a way of improv- necessary in the fabrication of Omh-perovskite-based photovoltaic
ing the stability of the Omh-perovskite itself. Moreover, Snaith devices. HTM free Omh-PSC devices employing an impermeable
and coworkers have prepared Omh-perovskite-based MSSCs, counter electrode have the potential to overcome these drawbacks
within which the mesoporous n-type TiO2 is entirely replaced and extend long term device stability. This new device concept has
with an insulating mesoporous Al2O3 scaffold.28,142 An Al2O3 greatest potential for practical application and further improve-
scaffold, which acts as a scaffold or supporting material, showed ments, both in efficiency and stability, can be expected.
improved device stability with an almost stable photocurrent of
15 mA cm2 over a 1000 h exposure period.6 However, some of 5.3 Interface aspect
the most promising results have come from the use of carbon- The nanoscale interface properties including interface defect
based electrodes.4,79,196 Wei and co-workers197 used inkjet print- structure relaxation, grain boundary phase transformations,
ing to produce CH3NH3PbI3/C bilayer devices which retained and grain boundary size present in different layers determine
ca. 90% of their initial PCE after storage in 30% relative humidity the eventual stability of the device. Thus, understanding these
(RH) for 12 days. Similarly, Zhang and co-workers196 and Zhou interface properties, and charge separation and transfer across

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

the nanoscale interface is critical. It has been shown that the an ultra-thin layer of coating before and after perovskite coatings.
instability of Omh-PSCs is primarily a result of charge accumu- Fig. 15a represents a coating option before the Omh-perovskite
lation at the interfaces due to ion migration, resulting in a coating that is responsible for diminishing localized TiO2 surface
change in the built-in electric field of the devices and hysteretic traps and recombination from the TiO2 conduction band to the
effects.102,112,128,145 The long-term stability of Omh-PSCs can be HOMO of the perovskite and the HTM. Another problem for
improved by a proper choice of interlayer and appropriate Omh-PSCs is that the Omh-perovskite cannot totally cover
engineering of barrier layers to protect the Omh-perovskite the surface of the TiO2 photoanode completely.27,211 Wang
film beneath and avoid an unnecessary barrier in different and co-workers83 found a similar phenomenon, i.e. there are
interface layers. The thermal and moisture-sensitive nature of uncovered sites on TiO2 forming indirect contacts with the HTMs.
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

the Omh-perovskite interface is an important concern. In order These uncovered sites act as a potential cause of recombination
to overcome such challenges, a protecting polymer matrix and for electrons from the CB of TiO2 to the HTMs.
single-walled carbon nanotubes have been pioneered.79 The electrode/electrolyte (TiO2/electrolyte) interface is the
It is obvious that the mesoporous TiO2 is sensitive to UV main problem causing recombination of charge in liquid and
illumination and acts as a source of interface instability which quasi-solid DSSCs due to its susceptibility to different reac-
causes device performance loss. Moreover, this performance tions. Any modification of this interface will have a great role in
loss is not only observed for the unsealed device but also for the improving its stability. In the field of liquid DSSCs and quasi-
sealed devices. This loss in performance of the sealed device solid DSSCs,212–214 post-modification is one of the key techni-
could not be attributed to the degradation of Omh-perovskite ques. For instance, the use of 4-tert-butylpyridine is an important
and/or HTM. Instead, it could rather result from the mesoporous interface modification means to retard recombination.212 Insu-
electrode during device illumination. Therefore, the changes in lating metal oxides such as Al2O3,215,216 ZnS217,218 and TBAI219
optical and electrical behavior of the sealed electrode (TiO2) need have also proved to be effective in retarding electron recombina-
to be studied in more detail. Fast decay of the excited state and tion. Post-modification by Al2O3 has two roles. One is to protect
the appearance of a low resistance shunting path could lead to CH3NH3PbI3 and the sensitized film degradation due to moist-
the loss in Jsc and Voc for the sealed devices.59,203 There are two ure and sunlight. Another is that Al2O3 serves as an insulation
approaches to overcome the interface destabilization due to barrier between TiO2 and HTM to retard the electron recombina-
mesoporous TiO2: first, mesoporous TiO2 can be replaced by tion process. Though some activity has been directed at inves-
other mesoporous scaffolds such as aluminum-oxide. The use of tigating the TiO2/HTM interface, little effort has been directed at
insulating materials such as SnO2 with a thin layer of MgO,204 the Omh-perovskite/HTM interface through post-modification.
Al2O372 and other alternative metal oxides to replace TiO2 Recently, post-modification using aluminum oxide83 and Y2O3220
showed an improved resistance to UV light. Secondly, an insulat- in Omh-perovskite solar cells has been reported. This type of
ing metal oxide is introduced to passivate the photoinduced trap modification protects the active material from moisture degra-
states in TiO2 that are formed during illumination. Different dation, as shown in Fig. 15b. However, the thickness of the layer
techniques have been used to coat the TiO2 by a variety of metal has to be controlled. Thicker over-layers between the TiO2 and
oxides.205,206 The purpose of these metal oxides is to create Omh-perovskite block the injection efficiency, but enhance
energy barriers, change the CB potential, or diminish surface recombination or decay of the excited state. Therefore, post
traps, and finally to retard recombination. Semiconductors modification of the Omh-perovskite/HTM interface is a promis-
such as Al2O3207–210 have been employed to produce protecting ing approach for shielding the perovskite, protecting it from
barriers at the TiO2/HTM interface, raising the physical separa- external factors such as moisture and improving the lifetime of
tion of injected electrons from the oxidized Omh-perovskite/ electrons by retarding recombination.
HTM. Fig. 15 shows schematic representations of a device with Similarly, the interface between the counter electrode and
p or n-type layers ((metal/HTM or p-type layer) and metal/ETL or
n-type layer interface) could be sensitive to a chemical reaction
which can deform the interface properties. Deposition pro-
cesses, for instance, thermal evaporation techniques, can cause
the diffusion of hot reactive metals across the interface and this
normally changes the behavior of the interface.221,222 The
formation of a new aluminum–carbon bond is supposed to be
the main cause of interface instability for devices based on an
aluminum counter electrode.223,224 In order to overcome such
problems, it is better to consider: (1) the use of other techniques
like a spray-deposition technique,225,226 (2) replacing reactive
metals such as silver by other alternatives such as carbon and
(3) applying small work function metals like Ba, Ca and Mg for
Fig. 15 Schematic representation of a device with an ultra-thin layer of
application in inverted architecture devices. More interestingly, a
coating or insulating material before (a) or after (b) Omh-perovskite- blade-coating technique has been demonstrated to improve the
sensitization on the mesoporous TiO2 surface. ambient stability of the Omh-perovskite, acting as air-protection

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

patches to block oxygen/moisture infiltration by encouraging strikes the material which it will degrade. Unfortunately, it is not
the development of self-assembled crystalline domains at the always feasible to attain this in practice.
Omh-perovskite surface.227 All of these findings indicate the 5.4.2 Device encapsulation. Encapsulation enables scientists
fundamental role of interfacial engineering in improving the inter- to eliminate the most damaging components of an Omh-PSC
face stability of Omh-PSCs. from UV radiation or environmental molecules, as shown in
Fig. 16, which is an important approach to protect the device,
5.4 Device operation aspect and increase its stability.239 An encapsulation method which
5.4.1 UV filters. The exposure of Omh-PSCs to UV light can retains device flexibility is essential from the viewpoint of indus-
trial engineering.240 The most extreme illustration is perhaps a
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

retard their performance during long-term operation. For instance,

UV light can result in TiO2 direct band excitation which may device sealed completely in glass.241 These techniques result in
lead to oxidation of the perovskite and HTMs. In order to pre- inflexible solar cells free from oxygen and water diffusion.
vent direct band gap excitation of TiO2, the easiest solution Another more advanced encapsulation approach is to use a
would be to apply a UV cut-off filter, and in fact there are thermosetting epoxy glue device sealing approach.228 Moisture
reports where UV cut-off filters have been used in stability and oxygen cannot penetrate this type of encapsulation and
testing.76,228–233,241 UV cutoff filters are usually used during this extends the long term stability of the device. Yang’s group
long-term stability measurements.234–238 Interestingly, the use reported primary stability investigations on Omh-PSCs stored in
of a UV-filter has been stated as beneficial for the stability of different environments including nitrogen, dry air, and ambient
Omh-PSCs based on CH3NH3PbI3 and TiO2 nanoparticles.6 It is atmosphere.36 The Omh-PSC devices showed a rigorous decay in
then clear that the perfect way to keep products from UV depriva- ambient air compared with the devices stored in dry air and
tion is to have some way of removing all the UV light before it nitrogen atmospheres, indicating that the decomposition/dissolving
of the Omh-perovskite materials is induced by moisture or air. This
indicates that better stability can be achieved with more advanced
encapsulation methods.36
More interestingly (and as yet still unexplained), higher per-
formance loss is observed for the sealed (encapsulated) device
compared to the unsealed (unencapsulated) one, as shown in
Fig. 17. This could be due to (1) the progressive occurrence of
shunting during operation and (2) the formation of deep trap
sites in the mesoporous electrode. However, there are alter-
native approaches to overcome such performance loss during
operation. In the first approach, the combination of encapsula-
tion with a UV filter would be better for improving the stability of
the device compared to a device without a UV filter, as indicated
in Fig. 17. The second approach uses both doping and encapsula-
Fig. 16 Schematic representation of Omh-PSC encapsulation. tion assisted reduction of trap states; Fig. 18 shows the parameter

Fig. 17 Non-encapsulated (blue open circles) and encapsulated (filled black squares) with and without a o435 nm cutoff filter (open red squares) aged
Omh-perovskite solar cell device. Reprinted with permission.6

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

Fig. 18 Plot of performance parameters versus ageing time for a TiO2 based ssDSSC device doped with aluminum followed by encapsulation.
Reproduced with permission from ref. 64.

evolution for aged devices under inert atmosphere operation. help extract all the limitations and challenges for practical applica-
The cell which is not doped decayed drastically in the first tion. Establishing methods and protocols would be a critical and
30 min, but the Al doped device did not. This indicates that the essential aspect in overcoming the differences or gaps among
decrease in the sub-band gap density of states arised from Al various research reports released from different groups.
doping, assisting the reduction of short-term device deteriora-
tion. Thus, aluminum enduringly appeases these defects through 6.1 Possible methods for studying degradation of Omh-PSCs
the interstitial substitution of Ti adjacent to oxygen defect sites. The degradation process includes a multifaceted sort of process
Moreover, three optional methods6 are proposed to minimize which is not easy to illustrate. Some of these aspects are: flow of
TiO2-based solar device power loss during operation: firstly, build H2O and O2 induced device degradation, active material and
up a means to efficiently and firmly appease the TiO2 surface interface degradation, electrode and interlayer migration, reac-
states, secondly, stop the UV light from reaching the mesoporous tion of hot metal with the ETL and HTM, and macroscopic and
TiO2,64 and thirdly, use down-converting nanoparticles, efficiently morphology related modifications. It is thus a challenging task
absorb the UV light and re-emit visible light.242,243 Establishing to identify degradation mechanisms and to quantify the degree
routes to harvest all the shorter wavelength light (o435 nm) of contribution from every mechanism towards the whole loss
before it contacts the mesoporous TiO2 should allow for improved of device performance. Degradation processes can be studied in
stability without the cost of reduced photocurrents.238,243 terms of different useful properties such as (1) the bulk and
surface behaviour, (2) the interface properties and (3) the array
of device architecture. This can be achieved using both com-
6. Protocols and methods for studying putational or theoretical, and physical and chemical character-
degradation mechanisms ization techniques. Information can be obtained from specific
study methods and/or from entire devices, which is averaged
After establishing all the causes and degradation processes, as information. These methods should be considered as key ways
well as the origins of degradations and approaches to improve to obtain accurate information, while possibly also suggesting
stability, it is a prerequisite to be able to understand the under- new routes of investigation.
lying degradation mechanisms and the causes and origins for 6.1.1 Characterization of Omh-perovskite material bulk and
these degradations using appropriate physical and chemical, and surface properties. Studies of material behavior (crystal struc-
theoretical (computational) techniques. There are also various tures, phase changes, and surface traps) are ways to identify
protocols to study the fading mechanisms, such as the lifetime problems that may compromise material stability. The bulk and
and the IV curve (the most detailed source of information), isotopic surface properties of Omh-perovskite materials can be studied to
labeling, and accelerated testing conditions. There are some characterize changes in the surface chemistry, structural and
procedures which are important for lifetime testing protocols, phase changes, carrier lifetime, mass loss, thermal behavior, and
e.g. dark, laboratory weathering, thermal cycling tests, as well as mechanical properties. All of these properties provide chemical
low light testing, which is commonly carried out under accelerated and/or physical information about Omh-perovskite material
conditions of temperature, humidity and illumination. This will stability and degradation. Furthermore, degradation information

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

can be obtained from both theoretical and experimental inves- 6.1.1.3 Carrier lifetime. The carrier lifetime of materials can be
tigations. Although most of the literature reports are from studied by monitoring time-resolved photoluminescence (TRPL)
experimental analysis, the contributions from theoretical or measurements. Snaith and co-workers reported TRPL measure-
computational viewpoints using first principles, ab initio and ments for Omh-PSCs and TRPL curve decay (te = 273  7 ns,
DFT methods are also essential in investigating the structure, where te is the time taken for the PL to fall to 1/e of its initial
orientations, vibrational frequencies, interaction energy, charge intensity) was observed.26 A PL lifetime as high as 200.1 ns was
distributions and other properties. also demonstrated, and the long lifetime is tentatively ascribed to
the Cl doping effect.245 In order to investigate the feasibility and
6.1.1.1 Chemistry of surface. The study of surface chemical impact of these Lewis base treatments, Noel and coworkers
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

processes plays a significant role in basic chemical research. At the performed a spectroscopic, theoretical, and device based investi-
molecular level, the surface atoms (compared to the bulk) experi- gation.118 Nonradiative decay was inhibited significantly within
ence a different chemical environment, which is due largely to the the treated perovskite films, especially under low levels of photo-
reduced number of neighbouring atoms. As a consequence, these excitation. More research needs to be done to understand how
surface atoms with changed atomic and electronic structures exhibit this novel family of solar absorber materials gives rise to long
high chemical reactivity. This chemical reactivity can lead to corro- charge carrier lifetimes. A comparison of hysteretic effects and
sion or degradation when in contact with the environment; there- transient behavior on mesoporous titania based perovskite
fore, studying the surface changes in Omh-perovskite materials devices to regular thin film devices was done using IV curves
when exposed to different conditions such as, moisture, tempera- and slow transient external quantum efficiency measurements.94
ture, and air, is essential. At present, a vast number of experimental Currently, more advanced techniques, such as correlated scan-
and theoretical surface chemistry techniques have been developed. ning electron microscopy and confocal photoluminescence mea-
The in situ techniques are specifically of great importance. X-ray surements, have been applied to explore the structure/function
photoelectron spectroscopy (XPS),244–246 Scanning Kelvin probe relationship in Omh-perovskite films. The results obtained indi-
microscopy (SKPM),112,185 Rutherford backscattering (RBS),247 cated the presence of considerable local PL heterogeneity with
X-rays for reflectometry, Surface Enhanced Raman spectroscopy average carrier lifetime of >1005 ns for CH3NH3PbI3(Cl) films,120
(SERS), and Ambient pressure X-ray photoelectron spectroscopy implying that substantial scope remains for minimizing non-
(AP-XPS) are some of the surface sensitive techniques. radiative decay. Regardless of these wonderful advancements
in studying and understanding from both experimental and the
6.1.1.2 Structural/phase and optical properties. Omh-perovskites theoretical observations, until now, the cause for the prolonged
are intrinsically complex materials, where the presence of various carrier lifetime of Omh-perovskites remains unclear. Some theo-
types of interactions and structural disorder may play an impor- retical observations imply that the dynamic position of the
tant role on the resulting material properties. Ex situ and in situ conduction band minimum could be the reason for this pro-
measurements, such as UV-vis, X-ray diffraction, vibrational longed carrier lifetime in MAPbI3 films.134
spectroscopy (infrared248,249 and, DFT simulation assisted
Raman250 and Raman251,252), spectroscopic ellipsometry88 and 6.1.1.4 Thermal properties. Since the formation of CH3NH3PbI3
X-ray absorption spectroscopy (XAS) are essential techniques for is influenced by the release of MAX, it is important to characterize
studying optical properties, structure and phase transformations the thermal properties of CH3NH3I and CH3NH3Cl. As shown in
in materials. Light soaking has been identified to split XRD peaks Fig. 19a, the thermal gravimetric analysis (TGA) curves show
in a CH3NH3PbI3xBrx film, indicating that crystalline phases nearly 100% weight loss between 250 1C and 360 1C for CH3NH3Cl
segregate with Br as the main species and iodine as a minor and between 290 1C and 370 1C for CH3NH3I, respectively. The
species, with the latter able to act as a trap site.253 While the lower onset of the weight loss temperature of CH3NH3Cl suggests
in situ absorbance measurements provide a convenient spectro- that it is easier to turn into the gaseous phase compared to
scopic method to track film degradation kinetics, they provide no CH3NH3I. Moreover, it is important to understand in what
structural information on changes in crystallinity or phase that manner CH3NH3Cl and CH3NH3I could possibly escape from
occur during degradation. Ex situ powder X-ray diffraction the precursor films during the annealing process. As shown in
(pXRD) experiments could be carried out in order to identify Fig. 19b, the endothermic peaks in the differential scanning
both the initial Omh-perovskite phase and any crystalline degra- calorimetry curves indicate that the escape of CH3NH3Cl and
dation products. While pXRD could be set to study the crystallo- CH3NH3I is related to a sublimation process.49
graphic structure, crystallite size (grain size), and preferred Both the TGA and DSC results confirm that the release of
orientation in polycrystalline or powdered solid samples, in situ CH3NH3I occurs at a higher temperature compared to CH3NH3Cl.
grazing incidence X-ray diffraction (GIXRD) measurements are This confirms that a higher annealing temperature is required for
more useful to elucidate the reaction mechanism for the decom- the CH3NH3I/PbI2 precursor mixtures.
position process upon accelerating conditions such as moisture
and light.86 Moreover, the combination of a fast 2D area detector 6.1.1.5 Mechanical properties. The mechanical properties254 of
with a high brightness synchrotron X-ray source allows the Omh-perovskite materials and the mechanical failure mecha-
acquisition of 2D diffraction data in real time, providing impor- nisms of devices can play critical roles in the operational stability
tant insight into the film decomposition process. of Omh-PSCs for portable and outdoor applications. This section

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

Fig. 19 (a) TGA and (b) DSC curves for CH3NH3I and CH3NH3Cl powders. Reproduced with permission from ref. 49.

highlights mechanical properties-principally the stiffness and angle-resolved XPS (AR-XPS),262 Fourier transform photocurrent
brittleness of pure films of these semiconductors and component spectroscopy (FTPS) and transient absorption spectroscopy (TAS)
layers e.g. fullerenes. The ubiquitous acceptor PCBM has many measurements, charge extractions by linearly increasing voltage
deleterious effects on the mechanical stability of OPV devices. (CELIV), first-principles DFT modeling,262 transient absorption
Pure PCBM phases have low cohesive energy,255 high tensile and spectroscopy (TAS), energy dispersive X-ray spectroscopy (EDX)
low crack onset strains,256 and weak interfaces with other device and electron energy loss spectroscopy (EELS) and surface-
layers.161 Several potential determinants of mechanical proper- sensitive XPS measurements,49,112,263 spectroscopic ellipsometry
ties, including molecular structures, polymorphisms, microstruc- or X-ray reflectivity,264 and energy dispersive X-ray reflectometry,159
tures and textures have not yet been critically reported. are helpful in studying interfacial issues. Miscellaneous processes
6.1.2 Device architecture. The most in depth source of can be integrated while interfaces/surfaces are exposed to water
information is the measurement of IV-curves as a function of and oxygen, and finally, can cause segregation on the surface. The
time. From the evolution of the IV-curves, information on Jsc, Voc, best option to investigate the interior of Omh-PSCs with respect to
FF, PCE, Rsh and Rs can be extracted and the broad-spectrum of the their chemistry is to systematically get rid of each layer in the device
curve can be applied to follow the decay of the device’s perfor- and examine the exposed interfaces 2–6, as shown in Fig. 20.
mance. The principles of an equivalent circuit play an important For chemical degradation, oxygen or water-induced degrada-
role in modelling the characteristics of a diode for the extraction of tion, chemical characterization is preferred such as TOF-SIMS,
extra device parameters, such as parallel and series resistances. together with isotopic labeling using H218O and 18O2 to acquire
The IV curve for Omh-PSCs can, however, change dramatically as it in-depth information on where, and to what degree of the capture
decays. Recently, Grätzel and coworkers demonstrated that an takes place. 18O has a natural abundance of 0.2%, i.e. simply
IV-curve of the device measurement at different light intensities of remunerated through the measurements of the intensity of the
16
the light source using an array of white LEDs could be ascribed to a O signal. In order to obtain a semi-quantitative measure of 18O
decrease in the shunt resistance.35 Measuring all diode parameters incorporation, and thereby of the process, eqn (22) is used:240
at a standard interval is therefore helpful, rather than just measur-
 18   18 
ing a single diode characteristic. I O I O

Different analytical techniques can be used in addition to I ð16 OÞ hn I ð16 OÞ ref
 18   100% (22)
the IV curve, such as intensity-modulated photocurrent/photo- I O
voltage spectroscopies (IMPS/IMVS),257,258 impedance,259 time- I ð16 OÞ ref
of-flight,260 and open-circuit voltage decay.261 The different
layers in both mesoporous TiO2 and compact TiO2 device
structures affect the separation, transport, and recombination
kinetics of the photoinduced charge carriers. According to
Wang and co-workers, a mesoporous electron conductor is
necessary to optimize CH3NH3PbI3-based solar cell devices with a
high PCE.178 This mesoporous TiO2 is more likely to induce the
formation of PbI2 than compact TiO2.178 Femtosecond time-resolved
transient absorption spectroscopy probed in the visible spectrum,
when coupled with XRD, is a useful tool for this purpose.178
6.1.3 Interface characterization. Identifying issues related to
interfaces are quite complex, including concerns over transport Fig. 20 Schematic drawing of an Omh-PSC cross-section with the
barriers, charge separation, band bending, and charge transfer. configuration Ag/HTM/CH3NH3PbX3/mp-TiO2/cl-TiO2/FTO showing all
Analytical techniques/methods such as IV characterizations, the relevant interfaces, where X is I, Br, Cl or mixed halides.

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

where I(18O) is the intensity of the 18O isotope of a given species of the device, the architecture of the device itself and/or from
and I(16O) is the corresponding intensity of the 16O isotope. various interfaces connecting different layers of the device.
The degradation of different device components such as the
6.2 Accelerated testing conditions and protocols Omh-perovskite material itself, hole transporting materials and
Accelerated testing protocols are useful in analysing the effects counter electrodes can take place for different reasons. These
of artificially rapid chemical degradation or physical changes to component degradations will collectively lead to overall device
a solar cell device brought about by exaggerated storage condi- degradation and may further restrict practical applications.
tions; these studies can be used to support long-term stability Thus, identifying the contributions from each component to
the device degradation would be more beneficial. For instance,
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

studies of solar devices. The practices for organic PV materials

and devices could offer a reference point, where the most Omh-perovskites are easy to degrade under different environ-
frequent accelerated testing protocols include solar–thermal mental conditions. Primary emphasis should be given to a
cycling, solar–thermal–humidity cycling and solar–thermal– comprehensive understanding of the crystal structure of these
humidity–freeze, laboratory weathering and thermal cycling.265 materials under varying conditions. Other families of these
Moreover, concentrated sunlight may be a novel approach to materials are also available and showed promising stability. In
explore the degradation of Omh-perovskite materials.122 In addition addition, the use of HTMs such as spiro-MeOTAD and PEDOT:PSS
to the accelerating protocols, accelerated testing equipment facil- is found to be destructive to the device. It would be more useful
itates a combination of tests. For example, a temperature/humidity to avoid these materials due to their obvious drawbacks, as
test (typically 85 1C/85% RH) by a pressure cooker test/highly- mentioned. Additionally, HTM additives such as Li-TFSI and
accelerated temperature and humidity stress test (PCT/HAST) 4-tert-butylpyridine are hygroscopic, which will directly affect
chamber is frequently used for electronic components, semi- the Omh-perovskite and lead to device deterioration. It is more
conductors, and PV modules, while illumination/temperature is important to avoid the use of such deliquescent additives.
a popular test for materials. It is believed that similar environ- Similarly, the use of different metal counter electrodes such
mental stress screening can be adopted to evaluate the reli- as silver and aluminum can also be detrimental to the device
ability of Omh-PSC components accordingly. stability. Replacing such electrodes made of these metals would
be beneficial. The deposition techniques of these metal electro-
6.3 H218O and 18
O2 isotopic labeling protocols des such as thermal deposition can also have a strong influence
on the device stability. Carbon electrode printing is a promising
Characterizing the performance of the solar cell under various
deposition method compared to thermal deposition, in which
atmospheric conditions can verify the stability of the device in a
hot metal diffusion into organic layers could attack the active
particular atmosphere, but does not in any way measure the
Omh-perovskite.
degree of atmospheric oxidation.266 In addition, as each layer in
The issues of device architecture remain impenetrable and
the cell acts as a barrier, varying a layer may modify the overall
this should also be considered when performing future long-
need for encapsulation. This barrier effect could be verified by
term stability tests. Stability studies of such devices with various
dealing with oxygen transmission inside the solar cell by an
designs are insufficient; an in-depth comparative stability study
isotopically labelled dry oxygen (18O2) atmosphere or with an
for the three different device architectures (planar, inverted and
oxygen-free humid (H218O) atmosphere.
mesoporous) is still highly required. Similarly, the effect inter-
face appears to be another urgent concern and as such requires
7. Conclusion and prospects additional in-depth investigation. Destructive interfacial reac-
tions may occur in the metal/organic interface due to the
The field of Omh-PSCs is undergoing rapid development, in diffusion of hot metal during deposition, while ion migration
which the majority of research attempts are focused on fabri- can also encourage confounding hysteresis effects – all of which
cating devices with better efficiencies. An evenly vital topic that promote device deterioration. The other critical issue at the TiO2/
has gained less concentration is the improvement of the quite perovskite interface is the depletion of iodine, due to the strong
deprived stability. Only promising steps have been achieved n-type character of the TiO2 semiconductor. The surface of TiO2
from a few minutes to thousands of hours (2000 h). Thus, there is thought to play a catalytic role in Omh-perovskite degradation.
may be various causes of device failure, which need to be This depletion may be further exacerbated by high temperatures
explored in-depth. The major causes responsible for this kind or by illumination during device operation. Omh-PSC devices
of degradation may be related to various possible external factors, have complex multilayer configurations in which each part may
such as moisture, oxygen, temperature, UV light, etc., as well as not succeed for various reasons and the layers can undergo an
internal intrinsic factors such as ion migration, electro-migration interaction chemically and physically in such a way that can give
and interfacial reactions. Knowledge of degradation mecha- rise to instability.
nisms, structures and phase transformations under different In this regard, it is necessary to establish efficient solutions
operating conditions play key roles in predicting material and for improving the stability to an extent that will make Omh-PSC
device behaviour. In this review, we have reviewed various causes technology attractive and feasible from commercial perspec-
and their associated mechanisms of degradation. These degrada- tives. Based on the current information and future outlooks, we
tions could be originated mainly from the different components suggest possible stability improvement approaches from three

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

main perspectives: Omh-perovskite materials, device architec- outdoor situations, so that the inconsistency in stability reports
ture and interfacial aspects. Upon various structural modifica- from various research groups can be gradually harmonized
tions, different alternative mixed Omh-perovskites and new under standardized conditions.
layered Omh-perovskite materials with promising stability
under operating conditions have been explored. There is still
‘room to explore various families of Omh-perovskites,’ which Acknowledgements
could be more stable and better candidates for future applica-
The financial support from the Ministry of Science and Technology
tion and fabrication of stable Omh-PSC devices. To date, the
(MoST) (104-3113-E-011-001-, 104-ET-E-001-ET, 103-2221-E-011-
best stability was achieved using an HTM and metal electrode
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

156-MY3, 102-2221-E-011-157-) and the Top University Projects

free mesoporous device architecture, which used a carbon as
of Ministry of Education (MoE) (100H45140), as well as the
counter electrode. Interestingly, commercial carbon paste is now a
facilities of support from National Taiwan University of Science
promising candidate for Omh-PSC cathodes. In our opinion, the
and Technology (NTUST) are acknowledged.
mesoporous device architecture with further improvement has
great potential for practical applications not only from a stability
point of view, but also for reasons of simplicity and ease of References
fabrication. For better improvement of this device architecture,
all possible interfacial problems should be able to be solved by 1 C. R. Osterwald and T. J. McMahon, Prog. Photovoltaics,
careful interface engineering. A good choice of anode and cathode 2009, 17, 11–33.
interlayer materials may protect the active layer by means of 2 OXFORD PVs – Next generation solar power, January 2015.
blocking the diffusion of oxygen and water through the metal 3 H.-S. Kim, C.-R. Lee, J.-H. Im, K.-B. Lee, T. Moehl,
cathode/electron transport layer, or the ITO/hole-transporting A. Marchioro, S.-J. Moon, R. Humphry-Baker, J.-H. Yum,
layer. Above all, efficient perovskite structure engineering, suitable J. E. Moser, M. Grätzel and N.-G. Park, Sci. Rep., 2012,
counter electrodes and its deposition, suitable device architecture 2, 591.
and careful interface engineering, which consider all the men- 4 A. Mei, X. Li, L. Liu, Z. Ku, T. Liu, Y. Rong, M. Xu, M. Hu,
tioned environmental and intrinsic factors, as well as all other J. Chen, Y. Yang, M. Grätzel and H. Han, Science, 2014, 345,
existing challenges ought to be cautiously taken to establish an 295–298.
all-purpose method for the production of stable Omh-PSCs. It can 5 J. Burschka, N. Pellet, S.-J. Moon, R. Humphry-Baker,
be confidently summarized that there is much work to be done P. Gao, M. K. Nazeeruddin and M. Gratzel, Nature, 2013,
and the guarantee of Omh-PSCs as a future choice of solution 499, 316–319.
processed energy technology is everlasting. 6 T. Leijtens, G. E. Eperon, S. Pathak, A. Abate, M. M. Lee
The degradation of different Omh-perovskites, interfaces and and H. J. Snaith, Nat. Commun., 2013, 4, 2885.
devices remains largely unexplored. Fundamental studies on the 7 H. Zhou, Y. Shi, Q. Dong, H. Zhang, Y. Xing, K. Wang,
photochemical, chemical, physical, and mechanical properties, Y. Du and T. Ma, J. Phys. Chem. Lett., 2014, 5, 3241–3246.
interfacial processes and degradation mechanisms of various 8 G. Hodes, Science, 2013, 342, 317–318.
device components and its architecture are still needed to 9 M. Antonietta Loi and J. C. Hummelen, Nat. Mater., 2013,
advance the field. Systematic studies of the relationships between 12, 1087–1089.
Omh-perovskite materials’ properties, including chemical com- 10 M. A. Green, K. Emery, Y. Hishikawa, W. Warta and
position, phase and crystal structure, morphology, surface and E. D. Dunlop, Prog. Photovoltaics, 2015, 23, 1–9.
bulk chemistry, and material crystalline quality correlated to 11 S. Kim, J.-W. Chung, H. Lee, J. Park, Y. Heo and H.-M. Lee,
grain size are required. The degradation of Omh-perovskites Sol. Energy Mater. Sol. Cells, 2013, 119, 26–35.
caused by extrinsic and intrinsic factors should be explored to 12 S. Mathew, A. Yella, P. Gao, R. Humphry-Baker,
understand the origin of instability during long term use. F. E. CurchodBasile, N. Ashari-Astani, I. Tavernelli,
Furthermore, the interface properties (interface defect structure U. Rothlisberger, K. NazeeruddinMd and M. Grätzel, Nat.
relaxation, grain boundary phase transformations, grain boundary Chem., 2014, 6, 242–247.
size) among the different layers determine the final stability of the 13 http://mms.businesswire.com/bwapps/mediaserver/View
resulting device. In this regard, systematic cross-sectional cell Media?mgid=354751&vid=4Heliatek, press release, 2013.
configurations showing every interface profile can be very useful 14 J. C. B. A. Marchioro, J. Teuscher, M. Grätzel and J.-E.
in investigating interfacial stability problems. Taking the informa- Moser, Proc. SPIE, 2013, 8811, 881108.
tion presented as a whole, we conclude that the understanding the 15 H.-S. Kim, C.-R. Lee, J.-H. Im, K.-B. Lee, T. Moehl,
multifaceted chemistry of Omh-PSCs will be key for developing A. Marchioro, S.-J. Moon, R. Humphry-Baker, J.-H. Yum, J. E.
new materials that will bring this nascent technology to fruition. Moser, M. Gratzel and N.-G. Park, Sci. Rep., 2012, 2, 591.
It is also imperative to investigate all proposed degradations 16 D. B. Mitzi, J. Mater. Chem., 2004, 14, 2355–2365.
further and decide possible approaches for stability improve- 17 D. B. Mitzi, K. Chondroudis and C. R. Kagan, IBM J. Res.
ments. Accelerated testing protocols, such as temperature/RH Dev., 2001, 45, 29–45.
combinations, illumination and moisture must be established 18 D. B. Mitzi, C. A. Feild, W. T. A. Harrison and A. M. Guloy,
and documented for stability studies under indoor and Nature, 1994, 369, 467–469.

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

19 D. B. Mitzi, S. Wang, C. A. Feild, C. A. Chess and 43 J. Burschka, A. Dualeh, F. Kessler, E. Baranoff, N.-L. Cevey-
A. M. Guloy, Science, 1995, 267, 1473–1476. Ha, C. Yi, M. K. Nazeeruddin and M. Grätzel, J. Am. Chem.
20 K. Chondroudis and D. B. Mitzi, Chem. Mater., 1999, 11, Soc., 2011, 133, 18042–18045.
3028–3030. 44 Y. S. Kwon, J. Lim, H.-J. Yun, Y.-H. Kim and T. Park, Energy
21 A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Environ. Sci., 2014, 7, 1454–1460.
Chem. Soc., 2009, 131, 6050–6051. 45 B. Suarez, V. Gonzalez-Pedro, T. S. Ripolles, R. S. Sanchez,
22 L. Etgar, P. Gao, Z. Xue, Q. Peng, A. K. Chandiran, B. Liu, L. Otero and I. Mora-Sero, J. Phys. Chem. Lett., 2014, 5,
M. K. Nazeeruddin and M. Grätzel, J. Am. Chem. Soc., 2012, 1628–1635.
134, 17396–17399. 46 J. J. Choi, X. Yang, Z. M. Norman, S. J. L. Billinge and
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

23 D. B. Mitzi, J. Chem. Soc., Dalton Trans., 2001, 1–12, DOI: J. S. Owen, Nano Lett., 2013, 14, 127–133.
10.1039/B007070J. 47 V. Roiati, S. Colella, G. Lerario, L. De Marco, A. Rizzo,
24 C. C. Stoumpos, C. D. Malliakas and M. G. Kanatzidis, A. Listorti and G. Gigli, Energy Environ. Sci., 2014, 7,
Inorg. Chem., 2013, 52, 9019–9038. 1889–1894.
25 G. Xing, N. Mathews, S. Sun, S. S. Lim, Y. M. Lam, 48 V. Roiati, E. Mosconi, A. Listorti, S. Colella, G. Gigli and
M. Grätzel, S. Mhaisalkar and T. C. Sum, Science, 2013, F. De Angelis, Nano Lett., 2014, 14, 2168–2174.
342, 344–347. 49 H. Yu, F. Wang, F. Xie, W. Li, J. Chen and N. Zhao, Adv.
26 S. D. Stranks, G. E. Eperon, G. Grancini, C. Menelaou, Funct. Mater., 2014, 24, 7102–7108.
M. J. P. Alcocer, T. Leijtens, L. M. Herz, A. Petrozza and 50 T. Supasai, N. Rujisamphan, K. Ullrich, A. Chemseddine
H. J. Snaith, Science, 2013, 342, 341–344. and T. Dittrich, Appl. Phys. Lett., 2013, 103, 183906.
27 J.-H. Im, C.-R. Lee, J.-W. Lee, S.-W. Park and N.-G. Park, 51 A. Dualeh, N. Tétreault, T. Moehl, P. Gao, M. K.
Nanoscale, 2011, 3, 4088–4093. Nazeeruddin and M. Grätzel, Adv. Funct. Mater., 2014, 24,
28 M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and 3250–3258.
H. J. Snaith, Science, 2012, 338, 643–647. 52 A. Dualeh, P. Gao, S. I. Seok, M. K. Nazeeruddin and
29 J. Salbeck, N. Yu, J. Bauer, F. Weissörtel and H. Bestgen, M. Grätzel, Chem. Mater., 2014, 26, 6160–6164.
Synth. Met., 1997, 91, 209–215. 53 A. Seemann, H. J. Egelhaaf, C. J. Brabec and J. A. Hauch,
30 U. Bach, D. Lupo, P. Comte, J. E. Moser, F. Weissortel, Org. Electron., 2009, 10, 1424–1428.
J. Salbeck, H. Spreitzer and M. Gratzel, Nature, 1998, 395, 54 M. O. Reese, A. M. Nardes, B. L. Rupert, R. E. Larsen,
583–585. D. C. Olson, M. T. Lloyd, S. E. Shaheen, D. S. Ginley,
31 G. Giorgi, J.-I. Fujisawa, H. Segawa and K. Yamashita, G. Rumbles and N. Kopidakis, Adv. Funct. Mater., 2010,
J. Phys. Chem. Lett., 2013, 4, 4213–4216. 20, 3476–3483.
32 J. H. Heo, S. H. Im, J. H. Noh, T. N. Mandal, C.-S. Lim, 55 M. Manceau, S. Chambon, A. Rivaton, J.-L. Gardette,
J. A. Chang, Y. H. Lee, H.-J. Kim, A. Sarkar, K. NazeeruddinMd, S. Guillerez and N. Lemaı̂tre, Sol. Energy Mater. Sol. Cells,
M. Gratzel and S. I. Seok, Nat. Photonics, 2013, 7, 486–491. 2010, 94, 1572–1577.
33 J. H. Noh, S. H. Im, J. H. Heo, T. N. Mandal and S. I. Seok, 56 F. C. Krebs and K. Norrman, Prog. Photovoltaics, 2007, 15,
Nano Lett., 2013, 13, 1764–1769. 697–712.
34 M. Liu, M. B. Johnston and H. J. Snaith, Nature, 2013, 501, 57 E. Carter, A. F. Carley and D. M. Murphy, J. Phys. Chem. C,
395–398. 2007, 111, 10630–10638.
35 M. A. Green, A. Ho-Baillie and H. J. Snaith, Nat. Photonics, 58 D. C. Hurum, A. G. Agrios, K. A. Gray, T. Rajh and
2014, 8, 506–514. M. C. Thurnauer, J. Phys. Chem. B, 2003, 107, 4545–4549.
36 H. Zhou, Q. Chen, G. Li, S. Luo, T.-B. Song, H.-S. Duan, 59 H. Al-Dmour and D. M. Taylor, Appl. Phys. Lett., 2009,
Z. Hong, J. You, Y. Liu and Y. Yang, Science, 2014, 345, 94, 223309.
542–546. 60 M. Iwamoto, Y. Yoda, M. Egashira and T. Seiyama, J. Phys.
37 J. A. Christians, J. S. Manser and P. V. Kamat, J. Phys. Chem. Chem., 1976, 80, 1989–1994.
Lett., 2015, 6, 852–857. 61 U. B. Cappel, T. Daeneke and U. Bach, Nano Lett., 2012, 12,
38 N. J. Jeon, J. H. Noh, W. S. Yang, Y. C. Kim, S. Ryu, J. Seo 4925–4931.
and S. I. Seok, Nature, 2015, 517, 476–480. 62 A. Stevanovic, M. Büttner, Z. Zhang and J. T. Yates, J. Am.
39 G. Hodes and D. Cahen, Nat. Photonics, 2014, 8, 87–88. Chem. Soc., 2011, 134, 324–332.
40 O. Malinkiewicz, A. Yella, Y. H. Lee, G. M. Espallargas, 63 Z. Zhang and J. T. Yates, Chem. Rev., 2012, 112, 5520–5551.
M. Graetzel, M. K. Nazeeruddin and H. J. Bolink, Nat. 64 S. K. Pathak, A. Abate, P. Ruckdeschel, B. Roose,
Photonics, 2014, 8, 128–132. K. C. Gödel, Y. Vaynzof, A. Santhala, S.-I. Watanabe,
41 N. K. Noel, S. D. Stranks, A. Abate, C. Wehrenfennig, D. J. Hollman, N. Noel, A. Sepe, U. Wiesner, R. Friend,
S. Guarnera, A.-A. Haghighirad, A. Sadhanala, G. E. Eperon, H. J. Snaith and U. Steiner, Adv. Funct. Mater., 2014, 24,
S. K. Pathak, M. B. Johnston, A. Petrozza, L. M. Herz and 6046–6055.
H. J. Snaith, Energy Environ. Sci., 2014, 7, 3061–3068. 65 S. K. Pathak, A. Abate, T. Leijtens, D. J. Hollman, J. Teuscher,
42 F. Hao, C. C. Stoumpos, D. H. Cao, R. P. H. Chang and L. Pazos, P. Docampo, U. Steiner and H. J. Snaith, Adv. Energy
M. G. Kanatzidis, Nat. Photonics, 2014, 8, 489–494. Mater., 2014, 4, 1301667.

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

66 K. Schwanitz, U. Weiler, R. Hunger, T. Mayer and 90 M. Jørgensen, K. Norrman, S. A. Gevorgyan, T. Tromholt,

W. Jaegermann, J. Phys. Chem. C, 2006, 111, 849–854. B. Andreasen and F. C. Krebs, Adv. Mater., 2012, 24,
67 K. Schwanitz, E. Mankel, R. Hunger, T. Mayer and 580–612.
W. Jaegermann, Chimia, 2007, 61, 796–800. 91 T. Baikie, Y. Fang, J. M. Kadro, M. Schreyer, F. Wei,
68 I. Nakamura, N. Negishi, S. Kutsuna, T. Ihara, S. Sugihara S. G. Mhaisalkar, M. Graetzel and T. J. White, J. Mater.
and K. Takeuchi, J. Mol. Catal. A: Chem., 2000, 161, Chem. A, 2013, 1, 5628–5641.
205–212. 92 S. Aharon, A. Dymshits, A. Rotem and L. Etgar, J. Mater.
69 D. C. Cronemeyer, Phys. Rev., 1959, 113, 1222–1226. Chem. A, 2015, 3, 9171–9178.
70 J. Bisquert, F. Fabregat-Santiago, I. Mora-Seró, G. Garcia- 93 C. Herring, J. Appl. Phys., 1950, 21, 437–445.
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

Belmonte, E. M. Barea and E. Palomares, Inorg. Chim. Acta, 94 E. L. Unger, E. T. Hoke, C. D. Bailie, W. H. Nguyen, A. R.
2008, 361, 684–698. Bowring, T. Heumuller, M. G. Christoforo and M. D.
71 M. A. Henderson, W. S. Epling, C. L. Perkins, C. H. F. McGehee, Energy Environ. Sci., 2014, 7, 3690–3698.
Peden and U. Diebold, J. Phys. Chem. B, 1999, 103, 95 S. Ito, S. Tanaka, K. Manabe and H. Nishino, J. Phys.
5328–5337. Chem. C, 2014, 118, 16995–17000.
72 G. Lu, A. Linsebigler and J. T. Yates, J. Chem. Phys., 1995, 96 H.-S. Kim, S. H. Im and N.-G. Park, J. Phys. Chem. C, 2014,
102, 4657–4662. 118, 5615–5625.
73 G. Munuera, V. Rives-Arnau and A. Saucedo, J. Chem. Soc., 97 E. Mosconi, A. Amat, M. K. Nazeeruddin, M. Grätzel and
Faraday Trans. 1, 1979, 75, 736–747. F. De Angelis, J. Phys. Chem. C, 2013, 117, 13902–13913.
74 A. Abate, T. Leijtens, S. Pathak, J. Teuscher, R. Avolio, 98 R. E. Wasylishen, O. Knop and J. B. Macdonald, Solid State
M. E. Errico, J. Kirkpatrik, J. M. Ball, P. Docampo, Commun., 1985, 56, 581–582.
I. McPherson and H. J. Snaith, Phys. Chem. Chem. Phys., 99 A. Poglitsch and D. Weber, J. Chem. Phys., 1987, 87,
2013, 15, 2572–2579. 6373–6378.
75 T. Leijtens, J. Lim, J. Teuscher, T. Park and H. J. Snaith, 100 M. L. Agiorgousis, Y.-Y. Sun, H. Zeng and S. Zhang, J. Am.
Adv. Mater., 2013, 25, 3227–3233. Chem. Soc., 2014, 136, 14570–14575.
76 J. Bisquert, A. Zaban and P. Salvador, J. Phys. Chem. B, 101 A. Buin, P. Pietsch, J. Xu, O. Voznyy, A. H. Ip, R. Comin and
2002, 106, 8774–8782. E. H. Sargent, Nano Lett., 2014, 14, 6281–6286.
77 Y. Yu, K. Wu and D. Wang, Appl. Phys. Lett., 2011, 99, 192104. 102 W.-J. Yin, T. Shi and Y. Yan, Appl. Phys. Lett., 2014,
78 J. Weidmann, T. Dittrich, E. Konstantinova, I. Lauermann, 104, 063903.
I. Uhlendorf and F. Koch, Sol. Energy Mater. Sol. Cells, 103 A. Walsh, D. O. Scanlon, S. Chen, X. G. Gong and S.-H. Wei,
1999, 56, 153–165. Angew. Chem., Int. Ed., 2015, 54, 1791–1794.
79 S. N. Habisreutinger, T. Leijtens, G. E. Eperon, S. D. 104 M. H. Du, J. Mater. Chem. A, 2014, 2, 9091–9098.
Stranks, R. J. Nicholas and H. J. Snaith, Nano Lett., 2014, 105 J. M. Azpiroz, E. Mosconi, J. Bisquert and F. De Angelis,
14, 5561–5568. Energy Environ. Sci., 2015, 8, 2118–2127.
80 I. C. Smith, E. T. Hoke, D. Solis-Ibarra, M. D. McGehee and 106 M.-H. Du, J. Phys. Chem. Lett., 2015, 6, 1461–1466.
H. I. Karunadasa, Angew. Chem., Int. Ed., 2014, 53, 11232–11235. 107 Y. Zhao and K. Zhu, Chem. Commun., 2014, 50, 1605–1607.
81 J. You, Y. Yang, Z. Hong, T.-B. Song, L. Meng, Y. Liu, 108 A. Pisoni, J. Jaćimović, O. S. Barišić, M. Spina, R. Gaál,
C. Jiang, H. Zhou, W.-H. Chang, G. Li and Y. Yang, Appl. L. Forró and E. Horváth, J. Phys. Chem. Lett., 2014, 5,
Phys. Lett., 2014, 105, 183902. 2488–2492.
82 G. E. Eperon, V. M. Burlakov, P. Docampo, A. Goriely and 109 R. Gottesman, E. Haltzi, L. Gouda, S. Tirosh, Y. Bouhadana,
H. J. Snaith, Adv. Funct. Mater., 2014, 24, 151–157. A. Zaban, E. Mosconi and F. De Angelis, J. Phys. Chem. Lett.,
83 G. Niu, W. Li, F. Meng, L. Wang, H. Dong and Y. Qiu, 2014, 5, 2662–2669.
J. Mater. Chem. A, 2014, 2, 705–710. 110 T. Oku, M. Zushi, Y. Imanishi, A. Suzuki and K. Suzuki,
84 J. A. Christians, P. A. Miranda Herrera and P. V. Kamat, Appl. Phys. Express, 2014, 7, 121601.
J. Am. Chem. Soc., 2015, 137, 1530–1538. 111 Z. Song, S. C. Watthage, A. B. Phillips, B. L. Tompkins,
85 B. R. Vincent, K. N. Robertson, T. S. Cameron and O. Knop, R. J. Ellingson and M. J. Heben, Chem. Mater., 2015, 27,
Can. J. Chem., 1987, 65, 1042–1046. 4612–4619.
86 J. Yang, B. D. Siempelkamp, D. Liu and T. L. Kelly, ACS 112 E. Edri, S. Kirmayer, A. Henning, S. Mukhopadhyay,
Nano, 2015, 9, 1955–1963. K. Gartsman, Y. Rosenwaks, G. Hodes and D. Cahen, Nano
87 X. Dong, X. Fang, M. Lv, B. Lin, S. Zhang, J. Ding and Lett., 2014, 14, 1000–1004.
N. Yuan, J. Mater. Chem. A, 2015, 3, 5360–5367. 113 W.-J. Yin, T. Shi and Y. Yan, Adv. Mater., 2014, 26,
88 A. M. A. Leguy, Y. Hu, M. Campoy-Quiles, M. I. Alonso, 4653–4658.
O. J. Weber, P. Azarhoosh, M. van Schilfgaarde, M. T. 114 J. S. Yun, A. Ho-Baillie, S. Huang, S. H. Woo, Y. Heo,
Weller, T. Bein, J. Nelson, P. Docampo and P. R. F. Barnes, J. Seidel, F. Huang, Y.-B. Cheng and M. A. Green, J. Phys.
Chem. Mater., 2015, 27, 3397–3407. Chem. Lett., 2015, 6, 875–880.
89 B. Hailegnaw, S. Kirmayer, E. Edri, G. Hodes and D. Cahen, 115 Q. Dong, Y. Fang, Y. Shao, P. Mulligan, J. Qiu, L. Cao and
J. Phys. Chem. Lett., 2015, 6, 1543–1547. J. Huang, Science, 2015, 347, 967–970.

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

116 W. Nie, H. Tsai, R. Asadpour, J.-C. Blancon, A. J. Neukirch, 138 N. Pellet, P. Gao, G. Gregori, T.-Y. Yang, M. K.
G. Gupta, J. J. Crochet, M. Chhowalla, S. Tretiak, Nazeeruddin, J. Maier and M. Grätzel, Angew. Chem., Int. Ed.,
M. A. Alam, H.-L. Wang and A. D. Mohite, Science, 2015, 2014, 53, 3151–3157.
347, 522–525. 139 M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and
117 D. Shi, V. Adinolfi, R. Comin, M. Yuan, E. Alarousu, H. J. Snaith, Science, 2012, 338, 643–647.
A. Buin, Y. Chen, S. Hoogland, A. Rothenberger, 140 S. Aharon, S. Gamliel, B. E. Cohen and L. Etgar, Phys.
K. Katsiev, Y. Losovyj, X. Zhang, P. A. Dowben, O. F. Chem. Chem. Phys., 2014, 16, 10512–10518.
Mohammed, E. H. Sargent and O. M. Bakr, Science, 2015, 141 S. Aharon, B. E. Cohen and L. Etgar, J. Phys. Chem. C, 2014,
347, 519–522. 118, 17160–17165.
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

118 N. K. Noel, A. Abate, S. D. Stranks, E. S. Parrott, 142 J. M. Ball, M. M. Lee, A. Hey and H. J. Snaith, Energy
V. M. Burlakov, A. Goriely and H. J. Snaith, ACS Nano, Environ. Sci., 2013, 6, 1739–1743.
2014, 8, 9815–9821. 143 E. Edri, S. Kirmayer, D. Cahen and G. Hodes, J. Phys. Chem.
119 A. Abate, M. Saliba, D. J. Hollman, S. D. Stranks, Lett., 2013, 4, 897–902.
K. Wojciechowski, R. Avolio, G. Grancini, A. Petrozza and 144 E. Edri, S. Kirmayer, M. Kulbak, G. Hodes and D. Cahen,
H. J. Snaith, Nano Lett., 2014, 14, 3247–3254. J. Phys. Chem. Lett., 2014, 5, 429–433.
120 D. W. de Quilettes, S. M. Vorpahl, S. D. Stranks, 145 J. M. Frost, K. T. Butler, F. Brivio, C. H. Hendon,
H. Nagaoka, G. E. Eperon, M. E. Ziffer, H. J. Snaith and M. van Schilfgaarde and A. Walsh, Nano Lett., 2014, 14,
D. S. Ginger, Science, 2015, 348, 683–686. 2584–2590.
121 M. Zhang, H. Yu, M. Lyu, Q. Wang, J.-H. Yun and L. Wang, 146 Y. Tidhar, E. Edri, H. Weissman, D. Zohar, G. Hodes,
Chem. Commun., 2014, 50, 11727–11730. D. Cahen, B. Rybtchinski and S. Kirmayer, J. Am. Chem.
122 R. K. Misra, S. Aharon, B. Li, D. Mogilyansky, I. Visoly- Soc., 2014, 136, 13249–13256.
Fisher, L. Etgar and E. A. Katz, J. Phys. Chem. Lett., 2015, 6, 147 O. Yamamuro, M. Oguni, T. Matsuo and H. Suga, Thermo-
326–330. chim. Acta, 1986, 98, 327–338.
123 I. P. Swainson, M. G. Tucker, D. J. Wilson, B. Winkler and 148 S. T. Williams, F. Zuo, C.-C. Chueh, C.-Y. Liao, P.-W. Liang
V. Milman, Chem. Mater., 2007, 19, 2401–2405. and A. K. Y. Jen, ACS Nano, 2014, 8, 10640–10654.
124 N. Onoda-Yamamuro, O. Yamamuro, T. Matsuo and 149 M. Antoniadou, E. Siranidi, N. Vaenas, A. G. Kontos,
H. Suga, J. Phys. Chem. Solids, 1992, 53, 277–281. E. Stathatos and P. Falaras, Journal of Surfaces and Inter-
125 J. Kim, S. Choi, A. Jun, H. Y. Jeong, J. Shin and G. Kim, faces of Materials, 2014, 2, 323–327.
ChemSusChem, 2014, 7, 1669–1675. 150 Q. Chen, H. Zhou, Y. Fang, A. Z. Stieg, T.-B. Song,
126 G. Niu, X. Guo and L. Wang, J. Mater. Chem. A, 2015, 3, H.-H. Wang, X. Xu, Y. Liu, S. Lu, J. You, P. Sun, J. McKay,
8970–8980. M. S. Goorsky and Y. Yang, Nat. Commun., 2015, 6, 7269.
127 A. S. Bhalla, R. Guo and R. Roy, Mater. Res. Innovations, 151 S. Colella, E. Mosconi, P. Fedeli, A. Listorti, F. Gazza,
2000, 4, 3–26. F. Orlandi, P. Ferro, T. Besagni, A. Rizzo, G. Calestani,
128 H. J. Snaith, A. Abate, J. M. Ball, G. E. Eperon, T. Leijtens, G. Gigli, F. De Angelis and R. Mosca, Chem. Mater., 2013,
N. K. Noel, S. D. Stranks, J. T.-W. Wang, K. Wojciechowski 25, 4613–4618.
and W. Zhang, J. Phys. Chem. Lett., 2014, 5, 1511–1515. 152 Z. Hawash, L. K. Ono, S. R. Raga, M. V. Lee and Y. Qi,
129 J. Even, L. Pedesseau, J.-M. Jancu and C. Katan, J. Phys. Chem. Mater., 2015, 27, 562–569.
Chem. Lett., 2013, 4, 2999–3005. 153 L. K. Ono, P. Schulz, J. J. Endres, G. O. Nikiforov, Y. Kato,
130 P. E. Shaw, A. Ruseckas and I. D. W. Samuel, Adv. Mater., A. Kahn and Y. Qi, J. Phys. Chem. Lett., 2014, 5, 1374–1379.
2008, 20, 3516–3520. 154 M.-C. Jung, S. R. Raga, L. K. Ono and Y. Qi, Sci. Rep., 2015,
131 C. Li, X. Lu, W. Ding, L. Feng, Y. Gao and Z. Guo, Acta 5, 9863.
Crystallogr., Sect. B: Struct. Sci., 2008, 64, 702–707. 155 F. C. Krebs, S. A. Gevorgyan and J. Alstrup, J. Mater. Chem.,
132 C. A. Randall, A. S. Bhalla, T. R. Shrout and L. E. Cross, 2009, 19, 5442–5451.
J. Mater. Res., 1990, 5, 829–834. 156 W. Chen, Y. Wu, J. Liu, C. Qin, X. Yang, A. Islam, Y.-B.
133 F. S. Galasso, Structure, properties, and preparation of Cheng and L. Han, Energy Environ. Sci., 2015, 8, 629–640.
perovskite-type compounds, Pergamon Press, Oxford, 157 C.-H. Chiang, Z.-L. Tseng and C.-G. Wu, J. Mater. Chem. A,
New York, 1969. 2014, 2, 15897–15903.
134 C. Motta, F. El-Mellouhi, S. Kais, N. Tabet, F. Alharbi and 158 Y.-J. Jeon, S. Lee, R. Kang, J.-E. Kim, J.-S. Yeo, S.-H. Lee,
S. Sanvito, Nat. Commun., 2015, 6, 7026. S.-S. Kim, J.-M. Yun and D.-Y. Kim, Sci. Rep., 2014, 4,
135 D. B. Mitzi, Prog. Inorg. Chem., John Wiley & Sons, Inc., 6953.
2007, pp. 1–121, DOI: 10.1002/9780470166499. 159 M. Jørgensen, K. Norrman and F. C. Krebs, Sol. Energy
136 I. Borriello, G. Cantele and D. Ninno, Phys. Rev. B: Condens. Mater. Sol. Cells, 2008, 92, 686–714.
Matter Mater. Phys., 2008, 77, 235214. 160 M. P. de Jong, L. J. van IJzendoorn and M. J. A. de Voigt,
137 G. E. Eperon, S. D. Stranks, C. Menelaou, M. B. Johnston, Appl. Phys. Lett., 2000, 77, 2255–2257.
L. M. Herz and H. J. Snaith, Energy Environ. Sci., 2014, 7, 161 S. R. Dupont, M. Oliver, F. C. Krebs and R. H. Dauskardt,
982–988. Sol. Energy Mater. Sol. Cells, 2012, 97, 171–175.

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

162 T. C. Sum and N. Mathews, Energy Environ. Sci., 2014, 7, 187 J. Shi, J. Dong, S. Lv, Y. Xu, L. Zhu, J. Xiao, X. Xu, H. Wu,
2518–2534. D. Li, Y. Luo and Q. Meng, Appl. Phys. Lett., 2014,
163 S. P. Singh and P. Nagarjuna, Dalton Trans., 2014, 43, 104, 063901.
5247–5251. 188 R. Zuo, L. Li and Z. Gui, Ceram. Int., 2000, 26, 673–676.
164 H. J. Snaith, J. Phys. Chem. Lett., 2013, 4, 3623–3630. 189 R. Shannon, Acta Crystallogr., 1976, 32, 751–767.
165 P. Gao, M. Gratzel and M. K. Nazeeruddin, Energy Environ. 190 J. F. Verwey, J. Phys. Chem. Solids, 1970, 31, 163–168.
Sci., 2014, 7, 2448–2463. 191 P. Docampo and H. J. Snaith, Nanotechnology, 2011,
166 Y. Shao, Z. Xiao, C. Bi, Y. Yuan and J. Huang, Nat. 22, 225403.
Commun., 2014, 5, 5784. 192 S. Guarnera, A. Abate, W. Zhang, J. M. Foster, G. Richardson,
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

167 J.-Y. Jeng, Y.-F. Chiang, M.-H. Lee, S.-R. Peng, T.-F. Guo, A. Petrozza and H. J. Snaith, J. Phys. Chem. Lett., 2015, 6,
P. Chen and T.-C. Wen, Adv. Mater., 2013, 25, 3727–3732. 432–437.
168 G. Li, R. Zhu and Y. Yang, Nat. Photonics, 2012, 6, 153–161. 193 A. Dutta, C.-T. Huang, T.-C. Chen, C.-Y. Lin, C.-H. Chiu,
169 Y. Zhou, C. Fuentes-Hernandez, J. Shim, J. Meyer, Y.-C. Lin, C.-S. Chang and Y.-C. He, Nat. Commun., 2015,
A. J. Giordano, H. Li, P. Winget, T. Papadopoulos, 6, 6374.
H. Cheun, J. Kim, M. Fenoll, A. Dindar, W. Haske, 194 H. L. Clever and F. J. Johnston, J. Phys. Chem. Ref. Data,
E. Najafabadi, T. M. Khan, H. Sojoudi, S. Barlow, 1980, 9, 751–784.
S. Graham, J.-L. Brédas, S. R. Marder, A. Kahn and 195 L. Zheng, Y.-H. Chung, Y. Ma, L. Zhang, L. Xiao, Z. Chen,
B. Kippelen, Science, 2012, 336, 327–332. S. Wang, B. Qu and Q. Gong, Chem. Commun., 2014, 50,
170 W. Zhang, B. Zhao, Z. He, X. Zhao, H. Wang, S. Yang, 11196–11199.
H. Wu and Y. Cao, Energy Environ. Sci., 2013, 6, 1956–1964. 196 F. Zhang, X. Yang, H. Wang, M. Cheng, J. Zhao and L. Sun,
171 D. A. Mengistie, M. A. Ibrahem, P.-C. Wang and C.-W. Chu, ACS Appl. Mater. Interfaces, 2014, 6, 16140–16146.
ACS Appl. Mater. Interfaces, 2014, 6, 2292–2299. 197 Z. Wei, H. Chen, K. Yan and S. Yang, Angew. Chem., 2014,
172 O. Malinkiewicz, C. Roldán-Carmona, A. Soriano, 126, 13455–13459.
E. Bandiello, L. Camacho, M. K. Nazeeruddin and 198 M.-C. Hsiao, S.-H. Liao, M.-Y. Yen, P.-I. Liu, N.-W. Pu,
H. J. Bolink, Adv. Energy Mater., 2014, 4, 1400345. C.-A. Wang and C.-C. M. Ma, ACS Appl. Mater. Interfaces,
173 J. Seo, S. Park, Y. Chan Kim, N. J. Jeon, J. H. Noh, S. C. Yoon 2010, 2, 3092–3099.
and S. I. Seok, Energy Environ. Sci., 2014, 7, 2642–2646. 199 D. Liu and T. L. Kelly, Nat. Photonics, 2014, 8, 133–138.
174 P. Docampo, J. M. Ball, M. Darwich, G. E. Eperon and 200 Z. Ku, Y. Rong, M. Xu, T. Liu and H. Han, Sci. Rep., 2013,
H. J. Snaith, Nat. Commun., 2013, 4, 2761. 3, 3132.
175 J. Bisquert, J. Phys. Chem. B, 2004, 108, 2323–2332. 201 Z. Li, S. A. Kulkarni, P. P. Boix, E. Shi, A. Cao, K. Fu,
176 H. J. Snaith and M. Grätzel, Adv. Mater., 2007, 19, S. K. Batabyal, J. Zhang, Q. Xiong, L. H. Wong, N. Mathews
3643–3647. and S. G. Mhaisalkar, ACS Nano, 2014, 8, 6797–6804.
177 K. Zhu, N. R. Neale, A. F. Halverson, J. Y. Kim and 202 E. Edri, S. Kirmayer, S. Mukhopadhyay, K. Gartsman,
A. J. Frank, J. Phys. Chem. C, 2010, 114, 13433–13441. G. Hodes and D. Cahen, Nat. Commun., 2014, 5, 3461.
178 L. Wang, C. McCleese, A. Kovalsky, Y. Zhao and C. Burda, 203 M. K. I. Senevirathna, P. K. D. D. P. Pitigala, E. V. A.
J. Am. Chem. Soc., 2014, 136, 12205–12208. Premalal, K. Tennakone, G. R. A. Kumara and A. Konno,
179 Y. Zhao, A. M. Nardes and K. Zhu, Faraday Discuss., 2014, Sol. Energy Mater. Sol. Cells, 2007, 91, 544–547.
176, 301–312. 204 D. Kuang, P. Wang, S. Ito, S. M. Zakeeruddin and
180 T. Salim, S. Sun, Y. Abe, A. Krishna, A. C. Grimsdale and M. Grätzel, J. Am. Chem. Soc., 2006, 128, 7732–7733.
Y. M. Lam, J. Mater. Chem. A, 2015, 3, 8943–8969. 205 M. Lira-Cantu, K. Norrman, J. W. Andreasen, N. Casan-
181 P. Qin, S. Tanaka, S. Ito, N. Tetreault, K. Manabe, Pastor and F. C. Krebs, J. Electrochem. Soc., 2007, 154,
H. Nishino, M. K. Nazeeruddin and M. Grätzel, Nat. B508–B513.
Commun., 2014, 5, 3834. 206 A. Kay and M. Grätzel, Chem. Mater., 2002, 14, 2930–2935.
182 K.-C. Wang, P.-S. Shen, M.-H. Li, S. Chen, M.-W. Lin, 207 E. Palomares, J. N. Clifford, S. A. Haque, T. Lutz and
P. Chen and T.-F. Guo, ACS Appl. Mater. Interfaces, 2014, J. R. Durrant, J. Am. Chem. Soc., 2002, 125, 475–482.
6, 11851–11858. 208 Z. Liu, K. Pan, M. Liu, M. Wang, Q. Lü, J. Li, Y. Bai and
183 J. A. Christians, R. C. M. Fung and P. V. Kamat, J. Am. T. Li, Electrochim. Acta, 2005, 50, 2583–2589.
Chem. Soc., 2014, 136, 758–764. 209 X.-T. Zhang, I. Sutanto, T. Taguchi, K. Tokuhiro, Q.-B.
184 W. Wang, J. Yuan, G. Shi, X. Zhu, S. Shi, Z. Liu, L. Han, Meng, T. N. Rao, A. Fujishima, H. Watanabe,
H.-Q. Wang and W. Ma, ACS Appl. Mater. Interfaces, 2015, 7, T. Nakamori and M. Uragami, Sol. Energy Mater. Sol. Cells,
3994–3999. 2003, 80, 315–326.
185 J. Liu, Y. Wu, C. Qin, X. Yang, T. Yasuda, A. Islam, 210 E. Palomares, J. N. Clifford, S. A. Haque, T. Lutz and
K. Zhang, W. Peng, W. Chen and L. Han, Energy Environ. J. R. Durrant, J. Am. Chem. Soc., 2003, 125, 475–482.
Sci., 2014, 7, 2963–2967. 211 H.-S. Kim, J.-W. Lee, N. Yantara, P. P. Boix, S. A. Kulkarni,
186 H. Greijer, J. Lindgren and A. Hagfeldt, J. Phys. Chem. B, S. Mhaisalkar, M. Grätzel and N.-G. Park, Nano Lett., 2013,
2001, 105, 6314–6320. 13, 2412–2417.

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015
 View Article Online

Energy & Environmental Science Review

212 M. K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, 235 A. Hinsch, J. M. Kroon, R. Kern, I. Uhlendorf, J. Holzbock,
E. Mueller, P. Liska, N. Vlachopoulos and M. Graetzel, A. Meyer and J. Ferber, Prog. Photovoltaics, 2001, 9, 425–438.
J. Am. Chem. Soc., 1993, 115, 6382–6390. 236 H. Pettersson and T. Gruszecki, Sol. Energy Mater. Sol.
213 L. W. R. Gao, B. Ma, C. Zhan and Y. Qiu, Langmuir, 2010, Cells, 2001, 70, 203–212.
26, 2460–2465. 237 P. M. Sommeling, M. Späth, H. J. P. Smit, N. J. Bakker and
214 R. Gao, L. Wang, Y. Geng, B. Ma, Y. Zhu, H. Dong and J. M. Kroon, J. Photochem. Photobiol., A, 2004, 164, 137–144.
Y. Qiu, J. Phys. Chem. C, 2011, 115, 17986–17992. 238 M. Grätzel, J. Photochem. Photobiol., C, 2003, 4, 145–153.
215 E. Palomares, J. N. Clifford, S. A. Haque, T. Lutz and 239 S. Gevorgyan, M. V. Madsen, H. F. Dam, M. Jørgensen,
J. R. Durrant, Chem. Commun., 2002, 1464–1465, DOI: C. J. Fell, K. F. Anderson, B. C. Duck, A. Mescheloff, E. A.
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

10.1039/B202515A. Katz, A. Elschner, R. Roesch, H. Hoppe, M. Hermenau,

216 F. Luo, L. Wang, B. Ma and Y. Qiu, J. Photochem. Photobiol., M. Riede, F. C. Krebs, S. Gevorgyan, M. V. Madsen,
A, 2008, 197, 375–381. H. F. Dam, M. Jørgensen, C. J. Fell, K. F. Anderson, B. C.
217 T. Zewdu, J. N. Clifford and E. Palomares, Phys. Chem. Duck, A. Mescheloff, E. A. Katz, A. Elschner, R. Roesch,
Chem. Phys., 2012, 14, 13076–13080. H. Hoppe, M. Hermenau, M. Riede and F. C. Krebs, Sol.
218 J. Tian, R. Gao, Q. Zhang, S. Zhang, Y. Li, J. Lan, X. Qu and Energy Mater. Sol. Cells, 2013, 116, 187–196.
G. Cao, J. Phys. Chem. C, 2012, 116, 18655–18662. 240 K. Norrman and F. C. Krebs, Surf. Interface Anal., 2004, 36,
219 G. Niu, L. Wang, R. Gao, B. Ma, H. Dong and Y. Qiu, 1542–1549.
J. Mater. Chem., 2012, 22, 16914–16919. 241 J. A. Hauch, P. Schilinsky, S. A. Choulis, R. Childers,
220 Y. Ogomi, K. Kukihara, S. Qing, T. Toyoda, K. Yoshino, M. Biele and C. J. Brabec, Sol. Energy Mater. Sol. Cells,
S. Pandey, H. Momose and S. Hayase, ChemPhysChem, 2008, 92, 727–731.
2014, 15, 1062–1069. 242 J. Wu, J. Wang, J. Lin, Y. Xiao, G. Yue, M. Huang, Z. Lan,
221 B. G. Frost, N. F. van Hulst, E. Lunedei, G. Matteucci and Y. Huang, L. Fan, S. Yin and T. Sato, Sci. Rep., 2013, 3, 2058.
E. Rikkers, Appl. Phys. Lett., 1996, 68, 1865–1867. 243 J. Wu, J. Wang, J. Lin, Z. Lan, Q. Tang, M. Huang,
222 H. Oji, E. Ito, M. Furuta, K. Kajikawa, H. Ishii, Y. Ouchi and Y. Huang, L. Fan, Q. Li and Z. Tang, Adv. Energy Mater.,
K. Seki, J. Electron Spectrosc. Relat. Phenom., 1999, 101–103, 2012, 2, 78–81.
517–521. 244 Y. Chen, T. Chen and L. Dai, Adv. Mater., 2015, 27, 1053–1059.
223 M. Lögdlund and J. L. Brédas, J. Chem. Phys., 1994, 101, 245 J. You, Z. Hong, Y. Yang, Q. Chen, M. Cai, T.-B. Song,
4357–4364. C.-C. Chen, S. Lu, Y. Liu, H. Zhou and Y. Yang, ACS Nano,
224 H. Antoniadis, B. R. Hsieh, M. A. Abkowitz, S. A. Jenekhe 2014, 8, 1674–1680.
and M. Stolka, Synth. Met., 1994, 62, 265–271. 246 B. Conings, L. Baeten, C. De Dobbelaere, J. D’Haen, J. Manca
225 V. I. Annenkov, S. G. Garanin, V. A. Eroshenko, N. V. and H.-G. Boyen, Adv. Mater., 2014, 26, 2041–2046.
Zhidkov, A. V. Zubkov, S. V. Kalipanov, N. A. Kalmykov, 247 F. Lang, A. Juma, V. Somsongkul, T. Dittrich and
V. P. Kovalenko, V. A. Krotov, S. G. Lapin, S. P. Martynenko, M. Arunchaiya, Hybrid Mater., 2014, 1, DOI: 10.2478/
V. I. Pankratov, V. S. Faizullin, V. A. Khrustalev, hyma-2014-0002.
N. M. Khudikov and V. S. Chebotar, Quantum Electron., 248 N. Onoda-Yamamuro, T. Matsuo and H. Suga, J. Phys.
2009, 39, 719. Chem. Solids, 1990, 51, 1383–1395.
226 J.-Y. Lee, S. T. Connor, Y. Cui and P. Peumans, Nano Lett., 249 E. Mosconi, C. Quarti, T. Ivanovska, G. Ruani and F. De Angelis,
2008, 8, 689–692. Phys. Chem. Chem. Phys., 2014, 16, 16137–16144.
227 J. H. Kim, S. T. Williams, N. Cho, C.-C. Chueh and 250 C. Quarti, G. Grancini, E. Mosconi, P. Bruno, J. M. Ball,
A. K. Y. Jen, Adv. Energy Mater., 2015, 5, 1401229. M. M. Lee, H. J. Snaith, A. Petrozza and F. D. Angelis,
228 F. C. Krebs, Sol. Energy Mater. Sol. Cells, 2006, 90, J. Phys. Chem. Lett., 2014, 5, 279–284.
3633–3643. 251 A. Maalej, Y. Abid, A. Kallel, A. Daoud, A. Lautié and
229 C. J. Brabec, N. S. Sariciftci and J. C. Hummelen, Adv. F. Romain, Solid State Commun., 1997, 103, 279–284.
Funct. Mater., 2001, 11, 15–26. 252 M. Ledinský, P. Löper, B. Niesen, J. Holovský, S.-J. Moon,
230 K. Tennakone, G. R. R. A. Kumara, I. R. M. Kottegoda, J.-H. Yum, S. De Wolf, A. Fejfar and C. Ballif, J. Phys. Chem.
K. G. U. Wijayantha and V. P. S. Perera, J. Phys. D: Appl. Lett., 2015, 6, 401–406.
Phys., 1998, 31, 1492. 253 E. T. Hoke, D. J. Slotcavage, E. R. Dohner, A. R. Bowring,
231 P. Wang, S. M. Zakeeruddin, P. Comte, R. Charvet, H. I. Karunadasa and M. D. McGehee, Chem. Sci., 2015, 6,
R. Humphry-Baker and M. Grätzel, J. Phys. Chem. B, 613–617.
2003, 107, 14336–14341. 254 J. Feng, APL Mater., 2014, 2, 081801.
232 P. Wang, S. M. Zakeeruddin, R. Humphry-Baker and 255 V. Brand, C. Bruner and R. H. Dauskardt, Sol. Energy Mater.
M. Grätzel, Chem. Mater., 2004, 16, 2694–2696. Sol. Cells, 2012, 99, 182–189.
233 T. W. Matthew Carnie and D. Worsley, Int. J. Photoenergy, 256 S. Savagatrup, A. D. Printz, T. F. O’Connor, A. V. Zaretski
2012, 2012, 9. and D. J. Lipomi, Chem. Mater., 2014, 26, 3028–3041.
234 A. Abate, J. Martı́-Rujas, P. Metrangolo, T. Pilati, G. Resnati 257 Y. Zhao, A. M. Nardes and K. Zhu, Faraday Discuss., 2014,
and G. Terraneo, Cryst. Growth Des., 2011, 11, 4220–4226. 176, 301–312.

This journal is © The Royal Society of Chemistry 2015 Energy Environ. Sci.
 View Article Online

Review Energy & Environmental Science

258 A. J. Frank, N. Kopidakis and J. v. d. Lagemaat, Coord. 264 S. A. Gevorgyan, M. Jørgensen and F. C. Krebs, Sol. Energy
Chem. Rev., 2004, 248, 1165–1179. Mater. Sol. Cells, 2008, 92, 736–745.
259 Q. Wang, S. Ito, M. Grätzel, F. Fabregat-Santiago, I. Mora- 265 M. O. Reese, S. A. Gevorgyan, M. Jørgensen, E. Bundgaard,
Seró, J. Bisquert, T. Bessho and H. Imai, J. Phys. Chem. B, S. R. Kurtz, D. S. Ginley, D. C. Olson, M. T. Lloyd, P. Morvillo,
2006, 110, 25210–25221. E. A. Katz, A. Elschner, O. Haillant, T. R. Currier, V. Shrotriya,
260 N. Kopidakis, E. A. Schiff, N. G. Park, J. van de Lagemaat M. Hermenau, M. Riede, K. R. Kirov, G. Trimmel, T. Rath,
and A. J. Frank, J. Phys. Chem. B, 2000, 104, 3930–3936. O. Inganäs, F. Zhang, M. Andersson, K. Tvingstedt, M. Lira-
261 A. Zaban, M. Greenshtein and J. Bisquert, ChemPhysChem, Cantu, D. Laird, C. McGuiness, S. Gowrisanker, M. Pannone,
2003, 4, 859–864. M. Xiao, J. Hauch, R. Steim, D. M. DeLongchamp, R. Rösch,
Published on 20 October 2015. Downloaded by National Taiwan University on 14/01/2016 01:26:06.

262 G. Pellegrino, A. Alberti, G. G. Condorelli, F. Giannazzo, H. Hoppe, N. Espinosa, A. Urbina, G. Yaman-Uzunoglu,

A. La Magna, A. M. Paoletti, G. Pennesi, G. Rossi and J.-B. Bonekamp, A. M. van Breemen, C. Girotto, E. Voroshazi
G. Zanotti, J. Phys. Chem. C, 2013, 117, 11176–11185. and F. C. Krebs, Sol. Energy Mater. Sol. Cells, 2011, 95,
263 P. Docampo, F. C. Hanusch, S. D. Stranks, M. Döblinger, 1253–1267.
J. M. Feckl, M. Ehrensperger, N. K. Minar, M. B. Johnston, 266 H. Neugebauer, C. Brabec, J. C. Hummelen and N. S.
H. J. Snaith and T. Bein, Adv. Energy Mater., 2014, 4. Sariciftci, Sol. Energy Mater. Sol. Cells, 2000, 61, 35–42.

Energy Environ. Sci. This journal is © The Royal Society of Chemistry 2015