Ente 202100560
Ente 202100560
Ente 202100560
Abstract – The efficiency of perovskite solar cells (PSCs) has risen rapidly over the last
decade, and it has already crossed the 25% mark. However, stability has long been the
to moisture, high temperature, UV light, and other environmental factors, which naturally
come in contact during operation. Moreover, degradation of the device is also associated with
the hole transport layer (HTL), electron transport layer (ETL), and buffer layers. The
mechanisms for PSCs' physical, chemical, structural, and environmental instabilities are
discussed critically in this work, along with recent efforts made by various groups to
stabilize the devices. Moreover, the lack of unified criteria for stability tests of PSCs is
discussed. This review collates and compares different degradation mechanisms and critically
evaluates recent approaches of different groups on stability analysis from a neutral point of
view. Finally, this review urges future research to focus on novel materials for different layers
which reasonably lattice-matched and stable with perovskite layer and to employ suitable
encapsulation techniques for proper sealing of the device against degrading substances.
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process, which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1002/ente.202100560
This article is protected by copyright. All rights reserved
1. Introduction
The global energy demand is constantly increasing, and the trend is estimated to raise the
Accepted Article
demand exponentially up to 778 EJ by 2035.[1] The towering pressure on already depleting
greener than any other form of energy shows the potential to have a massive role in next-
generation power demand.[2] Presently, silicon solar cells dominate the market owing to their
long-term stability, high efficiency, and availability in abundance.[3, 4] However, long energy
payback time and the use of materials in large quantities limits the use of silicon photovoltaics
for bulk power generation.[5] Additionally, reaching closer to the Shockley-Queisser limit,[6, 7]
photovoltaics is shifting towards other technologies such as thin-film solar cells, perovskite
solar cells (PSCs), organic solar cells (OSCs), etc. OSCs can offer much easier fabrication
schemes, less energy payback time, and lower material usage.[8-10] However, the efficiency is
not sufficient to employ it for bulk power generation. On the other hand, PSCs show the
conversion efficiency (PCE).[11, 12] Even though the efficiency of PSCs is more than 25%, the
Perovskites are a group of compounds having a general formula of ABX3, which possess the
organic cation (or a mixture of cations), B is a bivalent inorganic cation usually smaller in size
than A, and X is generally a halide (or a mixture of halides) anion. The crystal structure of
metal halide perovskites is illustrated in Figure 1.[16] While the history of metal halide
perovskites goes way back to 1839, the first study of this material for PV application was in
2006.[17, 18] Fabrication of perovskite solar cells was performed using dye-sensitized solar cell
This article is protected by copyright. All rights reserved
technology for the first time by Kojima et al. in 2009, achieving an efficiency of 3.1%.[10, 16, 19]
Iodine redox, which was used as a liquid electrolyte, rapidly degraded the PCE of the device.
Transport Layer (HTL) in place of the liquid electrode that saw an unexpected rise of PCE to
Accepted Article
9.7% and enhanced stability. This enhancement in PCE led to the research in PSCs holding
runner-up position in top breakthroughs picked by the editors of Science in 2013.[20] From
there, this class of solar cells quickly got the attention of many researchers, and the
documented PCE rapidly rose above 25% as of 2021.[21, 22] The progress of PSCs in terms of
PCE is the fastest by any PV technology so far. Perovskite materials have multiple suitable
properties for photovoltaic applications such as high absorption coefficient, high carrier
mobility, suitable bandgap, low-cost processing techniques, long diffusion length, high open-
circuit voltage, and others.[23-26] Such befitting properties opened doors for numerous research
Additionally, many articles have been published relating to the use of low-temperature and
step process. Lead halide is deposited in the first step, and in the second step, a reaction is
made between lead halide and organic halide, forming the perovskite under controlled
temperature.[27] Cheng et al. devised the perovskite layer by Ultrasonic Spray coating, which
involved modification of the formation of ink and different drying mechanism.[28] Luo et al.
used Chemical Vapour Deposition (CVD) for depositing a thin film of the perovskite material.
They were able to form devices having an area up to 16 cm2 and showed potential for larger
area cells.[29] Wei’s group used the inkjet printing method, which had a controlled interface
using nanocarbon as HTL.[30] Xiao et al. used an interdiffusion approach in the interface of
HTL with perovskite layer and fabricated pinhole-free PSCs.[31] In addition to various
processing techniques, different perovskite materials have also been tested to evaluate the
performance and potential for future generation PV devices. Especially, lead-free perovskites
This article is protected by copyright. All rights reserved
and mixed cation and/or anion perovskites offer the opportunity to tune the bandgap to
desirable values.[32-34] Extensive research in processing techniques are still going on since the
performance and stability are highly dependent on processing techniques, materials used, and
electrical characterization.
Accepted Article
Figure 1. Crystal structure of a typical metal halide perovskite material (CH3NH3PbI3), where
A position contains an organic metal halide cation (CH3NH3+), B is a metal cation (Pb2+), and
X is a halide anion (I-). (b) The unit cell of CH3NH3PbI3 perovskite in the cubic phase.
While considerable research has been done to achieve high PCE and improved processing
techniques that made low-cost and low-temperature devices possible, efforts to stabilize the
device have been relatively lower. Schematic illustrations of three commonly used perovskite
solar cell structures are presented in Figure 2, which shows the perovskite layer sandwiched
between two charge transport layers. There are two common structures of PSCs, mesoporous
and planar. The mesoporous structure is conventionally used employing mesoporous TiO2,
which is sensitized by the perovskite upon exposure to light. Mesoporous TiO2 is coated on
top of planar TiO2, and it acts as a scaffold for the growth of perovskite.[35] However, the
planar structure offers facile processing techniques and better PV performance owing to
temperature, UV light, O2 ingress, hysteresis, and other factors. These factors are illustrated in
Figure 3. Several reviews have discussed different degradation mechanisms in recent years.
Accepted Article
However, some of these mechanisms are debated, and some are still unknown.[22, 37, 40] Apart
from environmental factors, PSCs also suffer from various intrinsic stability issues such as
structural instability and chemical instability. While the extrinsic stability issues can be
modification of the device materials and interfaces.[41, 42] Even though the efficiency of PSCs
is good enough for commercialization, it is not expected to make a breakthrough in the market
anytime soon because of its low stability. Therefore, it is expected that extensive research in
PSCs will continue in the coming years in order to make this device viable for commercial
applications.
Figure 2. Schematic demonstration of different structured PSCs with all the essential layers
(a) n-i-p Mesoporous Structure. (b) n-i-p Planar Structure. (c) Inverted p-i-n Planar Structure.
Accepted Article
Structural Perovskite
UV Light
Instability Stability
Moisture
and Hysteresis
Oxygen
Burn in
Degradation
Aiming to improve the stability of PSCs, several techniques were implemented in the past,
which include using different electron transport layers (ETLs) and HTLs, using buffer layer,
modifying interface layer between two materials, encapsulation, and using hybrid materials
for different layers.[36, 43, 44] While none of these brought the desired stability, these techniques
improved the stability to some extent. The underlying physics of the degradation mechanisms
need to be studied more intuitively for bringing desired stability. Furthermore, stability test
standards are missing for PSCs in the literature. Due to the different nature of degradation
mechanisms, the standard stability tests should vary substantially from silicon
and comparing the stability of different works. Moreover, developing an international testing
standard and protocol for PSC stability is necessary for efficient comparison and evaluation of
responsible for degrading the performance of PSC are critically evaluated, and improvements
made over the years to overcome the degradation of PSCs are analyzed. The current state-of-
the-art devices in terms of stability against different effects are also analyzed critically.
Accepted Article
Moreover, recent works on stability improvements are discussed and evaluated. Finally, the
necessity of a unified standard protocol and standard for stability tests of PSC is presented.
Overall, this review offers an update on the literature of PSCs in terms of stability studies and
shows the gaps in the literature which need further research to bring long-term stability of
PSCs.
As discussed in the previous section, perovskite has multiple properties suitable for PV
addition to inherent morphological instability, PSC devices are also considerably degraded
when exposed to moisture, oxygen, high temperature, and UV light.[11, 22] It is widely believed
that Goldschmidt tolerance factor t shown by Equation 1 below determines the structural
Here, rA, rX, and rB are the ionic radii of the elements in positions A, X, and B, respectively.
For t between 0.8 and 1, perovskite has a cubic structure that is the most stable. Below 0.8,
perovskites generally have a tetragonal or orthorhombic structure. Hence, the inclusion Cl- or
Br- ions in a controlled amount in place of I- ions often increases the stability. The structural
stability of PSCs is also determined by the octahedral factor (µ) as shown in Equation 2,
limiting the choice of different halides and inorganic cations. It has been reported that only
Considerable deviation from this range makes the device either edge shared or face shared
octahedral structure than the conventional corner shared structure and accounts for unstable
structure.[47]
Accepted Article
𝑟
µ = 𝑟𝐵 (2)
𝑋
which perovskite decomposes to form halides of metals and gaseous HI. As a result, the
suitable material properties of perovskite are no longer retained. Moreover, a layer of metal
halides forms on the surface that causes increased recombination. All these effects together
degrade the performance parameters open-circuit voltage (VOC), short-circuit current density
(JSC), and fill factor (FF) of the cell, causing the device to lose its initial high PCE.
Moisture does not directly react with perovskite, but it works as a catalyst for the reactions
responsible for the degradation. Niu et al. proposed a series of reactions after observing the X-
ray Diffraction (XRD) patterns which involve the formation of H2 that can escape, thus
hindering the reversibility of the reactions.[48] The reactions involved in the degradation of the
perovskite layer in the presence of water are shown in Equation 3.1 - 3.4. Another possible
pathway that can degrade the perovskite layer is through deprotonation, as shown in Figure 4.
In this pathway, water being a Lewis base takes one proton from perovskite to form an
𝐶𝐻3 𝑁𝐻3 Pb𝐼3 (s) ↔ Pb𝐼2 (s) + 𝐶𝐻3 𝑁𝐻3 I (aq) (3.1)
Accepted Article
The first reaction initiates the degradation process. However, the degradation of HI by the last
two reactions makes the earlier reactions irreversible, damaging the lifetime of the perovskite
material.[50] The solubility of PbI2 formed in the first reaction is also environmentally harmful
due to the high toxicity of Pb. Yang et al. showed the effect of relative humidity on the
diffraction.[51] The gradual phase change was documented in their work, and the effect on
degrading PCE was also investigated by changing Relative Humidity (RH) by a controlled
apparatus. Yang et al. demonstrated that up to RH of 50%, the devices were relatively stable.
However, for 80% and 98% relative humidity, the degradation was very rapid, and the
absorbance of the perovskite dropped to 20% of the initial value within the first 20 hours of
operation. Interestingly, for RH of 20%, the absorbance showed an increase in the first 100
hours. Though this was not explained by the authors, it is perceived that moisture in a
controlled amount took part in the passivation of some defects in the perovskite/ETL intercept.
Moreover, stability analysis was not carried out for RH between 50% and 80%, which is often
perovskite in different RH conditions.[52] The authors divided the degradation mechanism into
four stages. While the first three of these stages can be reversed, the fourth one is not
reversible, which drove the reaction forward. From the LBIC imaging, the gradual phase
Accepted Article
change of the perovskite layer was shown. The degradation of EQE under different RH
conditions and different wavelengths was also measured from the LBIC imaging. Figure 5
shows the gradual degradation of EQE for RH of 50% and 80%. The LBIC images of stage 4
shown in Figure 5(c) also show the wavelength dependence of degradation. EQE is reduced
Figure 5. (a) External Quantum Efficiency mapping with LBIC for 532 nm wavelength of
light under 50 ± 5% Relative Humidity for MAPbI3 perovskite. (b) Degradation with time for
two different RH (50±5% and 80±5%). (c) EQE from LBIC images of stage 4 for different
improvements by passivating with PFBPA. The devices were kept in (40-75% RH) for 600
hours for stability testing. While the control device lost about half of its PCE after 600 hours,
the PFBPA passivated devices lost only 10%. Though the stability test is not long enough to
Accepted Article
find T80 or T50 of the devices, this approach seems to have a positive effect on long-term
stability as well. Surface morphology shows further evidence that improved grain boundaries
degradation.
In addition, the ingress of oxygen in the perovskite layer can induce degradation of the
perovskite material by photo-oxidation. While oxygen alone can only slightly degrade the
ions from I2 in the presence of light also forming H2 in the process. The subsequent
degradation effect of this reaction is twofold. H2 tends to escape the device, and I2 reacts with
the back electrode to form metal iodides. Besides, Pb2+ is oxidized to Pb4+ ions by photo-
oxidation, forming a layer of PbO2 on the surface of perovskite. It has been reported that the
devices do not show any considerable oxidation when kept in the dark under a wet
atmosphere.[54, 55] However, in the presence of light, oxygen seems to have an unavoidable
Bryant et al. showed the effect of light and oxygen by keeping the cells in various
conditions.[57] Their study shows that light soaking in a dry and illuminated condition in N2
filled glovebox generally leads to enhanced PCE. Besides, in dark and dry conditions, oxygen
did not degrade the devices considerably. A slight degradation in PCE was observed due to
surface oxidation of the perovskite. However, in the presence of oxygen, the devices degraded
rapidly under illumination. Therefore, it is evident from this study that oxygen and light are
combinedly responsible for the photo-oxidation of iodide ions, thus degrading the device.
Photo-oxidation within the device depends on the net rate of oxidation and electron transfer. If
This article is protected by copyright. All rights reserved
the electron transfer rate overcomes the oxidation rate, then there will ideally be no photo-
oxidation. However, the lifetime of electrons in the perovskite layer and HTL are in the range
of microseconds. Therefore, the rate of photo-oxidation is not very high to cause much
degradation to the device.[58] Proper encapsulation of the devices to eliminate oxygen ingress
Accepted Article
is one way to inhibit the degradation caused by photo-oxidation.
reactions at the TiO2/CH3NH3PbI3 interface. TiO2, which is the commonly used ETL in PSCs,
has a suitable bandgap of 3.2 eV. However, TiO2 acts as a photocatalyst to oxidize water and
organic materials in both mesoporous and planar forms.[59] Possible chemical reactions
2I − ↔ I2 + 2e− (4.1)
Here, the TiO2 interface is involved in the formation of I2 from iodide ions by oxidation. The
second reaction should shift to the left (reactant) side due to reaction kinetics. However, the
evaporation of methyl anime and reaction of H+ to form HI (which also evaporates) causes the
second reaction to proceed to the right side and cause degradation of methylamine. The
incorporation of Sb2S3 at the interface of mp-TiO2 and the perovskite layer leads to increased
stability.[60, 61] Sb2S3 is responsible for blocking I2 formation from iodide ions at the interface
of TiO2/CH3NH3PbI3.
The degradation of encapsulated and non-encapsulated PSCs was studied by Leijtens et al.,
who in their work observed the gradual decay of the PCE of the cells for 5 hours under one
This article is protected by copyright. All rights reserved
sun illumination with and without using a UV filter.[62] Surprisingly, the encapsulated devices
without UV filter decays faster than non-encapsulated devices as shown in Figure 6. The
rapid degradation is due to the collected electrons being trapped in deep trap states in the TiO2
layer. The authors replaced the mp-TiO2 scaffold with mp-Al2O3 that showed much higher
Accepted Article
stability.[62] The authors showed that mp-Al2O3 based devices showed stable PCE for 1000
hours. However, the degradation for the first 200 hours was not considered in which the
device lost more than 60% of its initial efficiency, making the stable efficiency only about 5%.
Moreover, incorporating an additional layer as a UV filter can often be costly and might lead
suitable UV blocking layer so that the device can be more stable while not trading
photovoltaic performance. It is also worth mentioning that the rate of UV degradation in Sn-
based PSCs is higher than Pb-based cells, which require better sealing for improved stability.
Figure 6. The gradual decay of PCE in the first 5 hours for Non-encapsulated, Encapsulated,
and Encapsulated + UV filter design of PSC using TiO2 as ETL under AM1.5 spectrum.
extra UV absorber layer.[63] The authors used silane coupling agents (SCA) of three different
incoming UV light of wavelength 275-400 nm. While the device performance was not
affected due to additional layers, the stability against UV light was increased. Figure 7 shows
the device performance comparison in J-V curves and XRD patterns. It is also seen in Figure
7(d) that the stability of the device with KH570 SCA interface is better than the pristine
device. However, the stability test is performed for 25 hours only, which is much shorter than
the required duration of a stability test for analysis and comparison. Moreover, the stability
Figure 7. Device performance and stability of a UV-resistant device. (a) J-V curves for
different devices. (b) PCE distribution of different devices. (c) XRD patterns of modified
devices and the conventional (control) device. (d) Stability of the control device and the
device with KH570 SCA interface. Reproduced with permission.[63] Copyright 2017, Royal
Society of Chemistry.
Wei et al. replaced TiO2 with ZnTiO3 as ETL due to a similar structure with perovskite and
good chemical stability.[64] Though the PV performance of the devices with ZnTiO3 was
similar to that of TiO2 based devices, high UV light stability was achieved by replacing TiO2
with ZnTiO3. After 100 hours of UV light soaking, the TiO2 based devices degraded
exceedingly and retained only 55% of their initial PCE, while the ZnTiO3 based devices had
retained almost 90% of their PCE. Moreover, at maximum power point (MPP), the devices
with ZnTiO3 ETL did not show any degradation for 800 hours after the initial burn-in
This article is protected by copyright. All rights reserved
degradation. However, the stability test was not performed under the standard light spectrum,
temperature, and RH. Therefore, the stability in outdoor conditions is still unknown. Though
different approaches are being considered to eliminate UV degradation, replacing TiO2 with
another suitable material as ETL is believed to be the most effective way. Other methods such
Accepted Article
as UV filters are expensive, and they can bring about additional degradation factors.
Apart from UV degradation, RGB and white light also accelerate ion migration in the device.
Xiao et al. showed the effect of red, green and blue light on the degradation of PSCs in
vacuum conditions.[65] The authors reported that the photoexcited charged carriers reduce the
energy barrier for ion migration, thus accelerating the process. As a result, deep trap states are
formed in the device, which increases the non-radiative recombination of the photogenerated
electron-hole pairs. Resulted ion migration is one of the causes of hysteresis in the device, and
deep trap states cause irreversible degradation of the device, resulting in reduced PCE. The
effect of blue light was more severe on degradation, followed by green light.
High temperature also accelerates the breakdown of the perovskite layer and causes
change of phase of perovskite. At around 160 K, the perovskite phase changes from
Orthorhombic to tetragonal. Again, at 330 K, the phase changes from tetragonal to cubic.[50]
When exposed to higher temperatures (above 85 °C or 358 K), the cubic phase
Temperature does not have a direct impact on stability. However, elevated temperature
accelerates the rate of this reaction. Degradation at 85 °C, which is well inside the range of
This article is protected by copyright. All rights reserved
operating temperature, is highly alarming. Meng et al. performed stability analysis of mixed
under temperature from 25 °C to 250 °C.[67] The authors reported that MA-based perovskites
quickly decompose into PbI2 at higher temperatures, and the PCE is degraded. Meng’s group
Accepted Article
performed SEM imaging, energy dispersive X-ray (EDX), and X-ray diffraction (XRD) on
the perovskites under different temperatures. The results suggested that the redshift of the
about 3.9% (from 1.569 eV to 1.508 eV). Increased temperature also resulted in increased
defect density.[67] From the SEM images shown in Figure 8, a white phase observed on the
surface of the perovskite was formed because of the increased concentration of PbI2. With
increasing temperature, this white phase occupied area gradually increased, and it covered
Figure 8. (a) Initial J-V characteristics of the best device. (b) EQE curve initially and
integrated short circuit current density (J). (c) Normalized PCE at different temperatures at
ambient conditions. (d) Degradation of PCE at different temperatures. (e) SEM images of the
perovskite surface (i)-(vi) represents 25 °C, 85 °C, 100 °C, 150 °C, 200 °C, 250 °C
X-ray Photoelectron Spectroscopy (PES) was used to observe the electronic structure of
MAPbI3, MAPbI3-xClx, and MAPbCl3 by Philippe et al. under various temperatures.[68] XPS
provided additional information on the chemicals formed during the decomposition process at
temperatures. At higher temperatures, both the ratios decreased due to the formation of PbI2
from the degradation of the perovskite film. The authors demonstrated that the formation of
Figure 9. (a) FESEM image of the control device and (b) 5 mg-(mL)-1 OA added device. (c)
J-V characteristics with different amounts of OA additive devices. (d) Degradation of PCE
with time for Control device and 5 mg-(mL)-1 OA added device. Reproduced with
Afroz et al. used an additive Oxalic acid having two bifacial carboxylic groups in the
perovskite solution during crystal growth.[69] The authors suggested that the addition of this
compound enhanced crystallization, which eventually produced perovskite crystals with large
grains, fewer surface traps, and reduced boundaries. The additive-enhanced devices showed
better thermal stability, retaining 90% of their initial efficiency after 9 hours in 100 °C
temperature and 60% RH, which later reduced to 70% after 19 hours. The devices with Oxalic
acid additives performed substantially better than the control device that saw a degradation of
Using Li-
410nm TFTS Not [51]
CH3NH3PbI3 22.9 50% No 400 95%
Light dopants in Mentioned
HTL
[53]
FAPbI3 ___ 40-75% No Not Surface 600 22.25% 90%
Mentioned Passivation
by PFBPA
UV blocking
510nm (UV [62]
CH3NH3PbI3–xClx ___ ___ Yes layer with 1000 11.5% 40%
aging test)
encapsulation
SCAs at
Room perovskite [63]
CH3NH3PbI3 60% No UV Light 24 17.36% 80%
Temperature ETL
interface
Cs0.05FA0.81MA0.14 [64]
Room ___ No Yes ZnTiO3 as 100 19.8% 90%
PbI2.55Br0.45 Temperature ETL with
Oxalic acid
with carboxyl [69]
CH3NH3PbI3 100 60% No Yes 19 16.47% 70%
group added
in absorber
It is illustrated in Figure 9 (c) that the addition of 5 mg-(mL)-1 OA additive results in devices
with the best performance. Figure 9 (b) further shows that the grain size and crystallinity of
the device are enhanced with OA additive, which is the possible cause of performance and
stability improvement. Moreover, the authors demonstrated that the hysteresis effect of the
Accepted Article
devices was reduced and had almost identical J-V curves in forward and reverse scans. The
authors claimed that the improvement was achieved as a result of acid-base interaction
between the carbonyl group in the oxalic acid with lead ions as well as the hydrogen bonding
of MA+ ions with the hydroxyl group of oxalic acid. However, the addition of Oxalic acid
only in a controlled amount (less than 5 mg-(mL)-1) leads to better performance and stability.
Higher amounts of additives severely reduce the performance and cause faster degradation of
The perovskite layer is the most ailing part of a PSC in terms of stability. However, other
layers also show noticeable degradation, which perturbs the regular operation of the device. It
is also established that some degradation mechanisms often start from another layer and
eventually reach the perovskite layer that is vulnerable to degradation, which deteriorates the
stability of the devices. In this section, the stability issues of other layers often used in PSCs
such as ETL, HTL, buffer layer, and electrodes will be critically evaluated. Moreover,
interfaces between two layers, which are the prominent regions of most chemical degradation,
Commonly used ETL in mesoporous PSCs is mp-TiO2 which acts as a scaffold for the
formation of the perovskite material. Though TiO2 based devices show higher initial
efficiency, it suffers from a sub-optimal filling of the scaffold, formation of deep trap states
This article is protected by copyright. All rights reserved
and excessive pinholes on the surface. These imperfections lead to fast degradation, and the
devices lose their PCE.[21, 60] Moreover, TiO2 is degraded by UV light which restricts the use
increases the cost of the device.[10, 61] While the deep trap states in the TiO2 layer can be filled
Accepted Article
with oxygen, the ingress of oxygen often leads to degradation of the perovskite layer. These
issues lead multiple groups to find alternative materials as ETL as well as engineer the device
with an interfacial layer to enhance the stability. Adding Al2O3 with TiO2 to form a bilayer
improves the efficiency and stability of the cells.[58] Al2O3 passivates the deep trap states of
the TiO2 while only TiO2 is responsible for the transport of electrons. Therefore, a bilayer of
Al2O3 and TiO2 performs better than a single layer of either material. Different materials used
as ETL includes metal oxides such as ZnO, Al2O3, Sb2S3, SnO2, and organic materials such as
Yang et al. enhanced the electron mobility and decreased recombination considerably by
forming a homogeneous bulk mixed (HBM) film with PCBM and n-type polymer poly[(9,9-
(F8TBT).[71] The HBM film improved electron mobility and showed better band alignment
with perovskite than PCBM as relative permittivity decreased from 4.73 to 3.82. As a result,
the capture cross-section electron decreased, leading to more efficient and better electron
transport. Figure 10 (a-d) demonstrates the morphology of the films by Atomic Force
Microscopy (AFM) images which showed the continuous formation of the HBM layer over
the complete surface of the perovskite while the film formed by PCBM is not uniform and
contained defects. The study by Yang et al. also showed that employing HBM as ETL
improved the PCE from 17.23% to 20.67%, along with the attainment of better stability as
shown in Figure 10 (e) and (f). The HBM based devices retained more than 80% of their
initial efficiency under illumination, while the PCBM based devices retained about 50% PCE,
demonstrating that HBM based devices performed better in multiple aspects than PCBM as
This article is protected by copyright. All rights reserved
ETL. However, a long-term stability test was not performed in that study which is more vital
for the commercialization of PSCs at this point. The absence of long-term stability tests of
such promising devices emphasizes the need for having unified stability criteria for testing
novel devices.
Accepted Article
Figure 10. (a-d) Morphological characteristics from Atomic Force Microscopy images of the
PCBM and HBM films. (e, f) Stability of PCBM and HBM based devices in dark and ambient
Inorganic metal oxide alternatives, mostly ZnO, have been used to substitute TiO2 as ETL in
recent studies due to its superior optoelectronic properties.[72-74] However, ZnO also possesses
Lewis base properties which account for deprotonation of the perovskite layer leading to rapid
degradation. Chen et al. used sulfidation on the surface of ZnO to avoid direct contact with
the perovskite layer.[75] As a result, a ZnO-ZnS layer formed, having much better interfacial
properties. ZnS acted both as a passivating layer and an ETL. Moreover, sulfur on the surface
has a strong binding with PbI2. As a result, it enhanced electron transport and retarded the
degradation of the perovskite layer. The ZnO-ZnS based champion device showed 20.7%
PCE while retaining 87% of the initial PCE after 500 hours under UV radiation. Mahmud et al.
used Al-doped ZnO as the ETL, which could be processed at low temperature (150 °C
This article is protected by copyright. All rights reserved
maximum temperature). Having similar properties with TiO2 and higher conductivity, the
devices with Al-doped ZnO ETL simultaneously had increased stability and yielded better PV
performance.[76] Their work was based on the previous work by Zhao et al., which needed
more than 500 °C for processing.[77] The devices fabricated by Mahmud et al. had a champion
Accepted Article
PCE of 14.54% and retained about 80% of its initial PCE after 570 hours in a nitrogen-filled
glovebox at 35-40% RH and ambient temperature. However, this study did not analyze the
stability of the devices in operating conditions under standard spectrum and maximum power
point biasing. Therefore, the photostability of the devices can’t be assessed or compared.
Developing a suitable ETL is one of the major milestones towards the long-term stability of
PSCs. While increasing studies are being published in recent years that show enhanced
stability, none of these have required stability performance for commercialization. However,
the devices in recent works showed better stability than the devices in the past. This improved
stability signals that through continuous research, the devices can reach commercial standards
in the future.
Spiro-OMeTAD has been the most frequently used HTL in a perovskite solar cell. However,
this HTL needs doping with bis (trifluoromethane) sulphonamide Lithium salt (Li-TFSI) for
better conductivity and carrier transport. Li-TFSI is notorious for its hydrophilic nature, which
in turn attracts water in the device and degrades the perovskite layer.[78] Different studies have
shown that the degradation mechanism of the perovskite layer at high temperature and
that the mechanism starts from the ingress of iodide ions from the perovskite layer to the HTL.
These ions cause a reduction in the Highest Occupied Molecular Orbital (HOMO) of the HTL.
As a result, a barrier is created for holes at the perovskite/HTL interface. This barrier causes
years to improve the stability and PV performance of the devices. These includes organic
(PTAA) and inorganic such as NiOX, CuSCN, CuCrO2 and others.[82-87] Jeon et al. used a
Accepted Article
fluorine terminated fine-tuned energy matched HTL (N2,N2′, N7,N7′-tetrakis(9,9-dimethyl-
- tetraamine) which is abbreviated as DM having high glass transition temperature (160 °C)
for a stable and efficient PSCs.[88] Stable devices with a PCE of 22.85% having 23.2% initial
efficiency of the champion cell were obtained. After thermal annealing at 60 °C, it retains
95% of initial PCE for more than 500 hours. The device had structure of FTO / bl-TiO2 / mp-
and DM was deposited on it by spin coating. This device produced the highest efficiency for
stable devices, certified as 20.9%. The stability of the device enhanced in both dark and
illuminated conditions. Under continuous illumination, the device retains more than 90% of
its initial PCE after 300 hours which is high stability for PSCs though not enough for
commercialization.
Jung et al. introduced the concept of an ultra-thin layer of wide bandgap halide perovskite on
top of the original absorber perovskite, which used P3HT as the HTL. In this configuration,
HTL did not need any doping.[84] This ultra-thin layer removed the common issue with P3HT,
which was interfacial defects with perovskite leading to recombination. As a result, the double
halide architecture (DLA) devices showed no hysteresis, and PCE increased substantially.
While the RH of 85% was maintained at dark conditions, RH in illuminated conditions, which
was needed to evaluate the performance with other stable devices accurately, was not
from the control device. The enhancement is caused by avoiding a direct interface between
One of the most vulnerable parts of PSCs is the boundary of the perovskite layer. Thus,
improving the interface with ETL and HTL often leads to better performance and stability.[89,
Accepted Article
90]
Therefore, an ultra-thin buffer layer is employed in the interface to avoid direct contact of
the ETL/HTL with the perovskite layer. Though the buffer layer introduces its series
matching with the perovskite and passivating the interface deep traps and defects.[91]
Sometimes the additives in the HTL/ETL are also responsible for the buffer action, which
results in improved performance of the cells. Two properties at the interface are responsible
for the degradation of the device. One is deep trap states at the perovskite surface, and another
is the mismatch of energy level at the interface. For better hole transport, materials with
suitable HOMO should be selected as an HTL.[92, 93] For reducing the interface deep traps and
defects, a buffer layer can be introduced between the HTL and perovskite, having HOMO
between the HOMO of HTL and the valence band of the perovskite layer. This buffer layer
improves band alignment and passivates the defects and the traps at the interface
In a study by Zhao et al., a series of elements containing Triphenyl amino group were used
along with HTL for passivation of surface traps and defects.[94] These materials acted both as
bromide (TPA-PEABr) as the passivating material, PCE increases from 16.69% to 18.15%.
This improvement resulted from an increase of VOC from 1.02 V to 1.09 V. Voc increased due
to the passivation of surface defects and enhanced band alignment. Additionally, the
different materials, TPA, TPA-EABr, and TPA-PEABr, are used in that study. The logic
behind the material selection is that PTAA and Spiro-OMeTAD, which are frequently used as
HTL in PSCs, are also TPA derivatives. TPA-PEABr resulted in the best performance
This article is protected by copyright. All rights reserved
between these three. The absorption edges were the same with the control device and the
device with TPA-PEABr, which supports the claim that the improvement is due to better band
alignment and reduced recombination by the traps. Doping with different amounts of TPA-
PEABr showed that the most optimal performance is achieved with 10mM density. Under
Accepted Article
operating conditions, the improved devices retained about 95% of their initial efficiency after
300 hours, while the device with Spiro-OMeTAD retained 85%, which also shows improved
stability. However, the testing conditions such as temperature and RH were not reported in the
paper.[94]
Sahin et al. used graphene oxide (GO) and modified graphene oxide (mGO) as buffer layers
between perovskite (CH3NH3PbI3-xClx) and HTL (P3HT) in order to improve the performance
of the devices.[95] Graphene oxide, also used as a buffer in other studies, is modified by adding
di-ethylamine (DEA) and 2-ethyl-hexylamine (2EHA). The addition of amine groups with
graphene oxide resulted in improved performance of the devices in terms of PCE. While no
stability analysis was performed in that study, the PCE increased from the reference device
(No buffer) by 25.5%, most of which comes from the increase of J SC and FF while the VOC
had decreased. Therefore, it is apparent that incorporating the buffer (mGO) brought some
additional series resistance while reducing the non-radiative recombination at the interface
Recently, Wang et al. used carbolong-derived organometallic complexes, namely a1, a2 and
b1, as buffer layers to tune the back electrode work function in n-i-p structure PSCs.[96] Both
a1 and b1 contain the cation BF4- while a2 contains [OFt]- cation. For the anion part, a1 and
a2 both contain Iridium (Ir), whereas b1 contains the phosphonium group. Dipoles formed on
the surface from anions and cations of the organometallic complexes reduce the work function
of the metal, enabling high work function metals to be used as back electrodes. Device
characterization showed that the devices with the buffers resulted in reduced charge transport
(c) and (d). Among the control device and three modified devices (a1, a2 and b1), with both
Au and Ag, b1 based devices performed best in terms of stability. Under an inert atmosphere
Figure 11. Characterization of the devices with and without different organometallic complex
buffer layers. (a) The Nyquist plot, showing the effect on RC and Rrec. (b) Time-resolved
photoluminescence intensity. (c), (d) PCE and normalized PCE of the control device and best
Interfaces of the perovskite layer with ETL and HTL contain defects, deep trap states and
mismatch in band alignment, all of which negatively impact the stability of the devices. While
passivation,
Strategy Duration Percentage
Relative Reference
Encapsulation Light Employed of Test Initial PCE of PCE
Humidity
Accepted Article (Hours) Retained
solved. It is believed that passivating the boundary defects and deep trap states enhances the
PV performance and improves the long-term stability of the devices.[97-99] For surface
proper band alignment can be achieved through appropriate selection of ELT, HTL and buffer
low temperature
Temperature
Temperature
Room
( °C)
___
___
___
___
25o
25o
(FAPbI3)0.95(MA
(FAPbI3)0.85(MA
FAMAPbIxBr3-x
MA0.6FA0.4PbI3
CsFAMAPbI3
CH3NH3PbI3
Perovskite
PbBr3)0.15
PbBr3)0.05
Accepted Article Layer
iso-propyl alcohol (IPA) is often used as the solvent. However, studies show that IPA and
other conventional solvents cause undesirable degradation to the perovskite layer and thus
tarnish the aim of its usage.[100, 101] Therefore, a proper solvent is required to passivate the
Mahmud et al. analyzed the effect of modifying the ETL (ZnO) and perovskite
Acetate.[102] Using Calcium Acetate, the offset of the conduction band of ZnO lowered by 50
meV resulting in a reduction in barrier seen by electrons. The band aligned devices using
Calcium Acetate showed relatively lower hysteresis and better stability. The average device
showed a PCE of 15.14%, while the efficiency of the devices with Calcium Carbonate
modification was just above half of this figure. Tested in a nitrogen-filled glovebox, the
Calcium Acetate modified devices retained 87% of their initial PCE after one month, more
than four times the PCE retained by the devices modified by Calcium Carbonate. The analysis
done by Mahmud et al. inspires to use different agents in the interface to improve the device
performance. However, the lack of stability test in higher RH and AM1.5 spectrum makes it
difficult to evaluate the long-term stability of the devices. For commercialization, stability in
Yoo et al. used a technique to deposit a layered perovskite (2D perovskite) on top of the 3D
perovskite layer using a combination of different alkyl ammonium bromides and chloroform
performance of the devices.[103] This study showed that when treated with IPA, the perovskite
surface contains a higher concentration of PbI2 at the interface, resulting in more deep trap
states. However, when chloroform is used to treat the perovskite, no visible change in the
Accepted Article
surface was observed. Different passivating agents such as C8Br, C6Br and C4Br were used
along with the control device to observe the effect. C8Br treated devices outperformed others
in terms of stability and PV performance. The champion device had a VOC of 1.17 V and an
efficiency of 23.4%. The 2D perovskite interface devices were tested under one sun
illumination using a UV filter at the maximum power point for the long-term stability test.
The devices retained 85% of their initial PCE after 500 hours at these conditions, which is
very encouraging for high-efficiency PSCs. It was observed that with the passage of time, the
VOC of the devices had increased slightly for 150 hours, whereas JSC saw a decrease.
Different buffer layers and HTLs are now selected based on their ability to passivate the
however, depends on the technique of depositing the layers as well. Passivating grain
boundaries and trap defects at the interface can take the device stability and efficiency to a
commercial standard since it will ultimately be a barrier to other degradation mechanisms like
3.5. Electrodes
Gold (Au) is the commonly used electrode in high efficiency and stable PSCs.[105] However,
life cycle assessment studies often find that the energy payback time using gold electrodes is
very high. Silver (Ag) is also used in some studies as the back electrode having sub-optimal
by the formation of AgI from HI by the decomposition reaction of MAPbI3.[107] Some other
novel metals such as Cu, Cr, Mo, Ni, W, Co, Pd have also been used in different studies as
This article is protected by copyright. All rights reserved
electrodes, and it is seen that all of them are somewhat responsible for the degradation and
contamination of the perovskite layer.[105] An ideal electrode should have the following
properties, (i) It should add minimal series resistance. Therefore, the conductivity should be
as high as possible. (ii) It should not be corroded by moisture, high temperature, oxygen, and
Accepted Article
UV light. (iii) It should be able to reflect the unabsorbed portion of light so that the
unabsorbed photons get back inside the device as much as possible. (iv) The fermi level of the
metal should be between the conduction band and valence band of the perovskite. Therefore,
with MAPbI3, it should be between -5.5 eV to -3.9 eV. (v) Finally, it should not react with any
of the substances formed in the degradation reaction of the perovskite layer like HI, I2,
CH3NH2, PbI2, and others. Therefore, it should be chemically inert while not resisting the
conduction of carriers. Table 3 summarizes the fermi level of some metals used in PSCs and
their resistivities. It is seen that only Cu has lower resistivity in the range of Au and Ag for
Cu was used in a study of the inverted structure by Zhao et al. in a device without an ETL. It
is surprising that Cu, when annealed for a longer time at 80 °C temperature, did not form CuI
and causes no contamination to the perovskite layer.[108] The device had the structure of
Table 3. Fermi level EF (eV) and Resistivity ρ (10-10 Ω-cm) of some metals used in PSCs as
electrodes.
Metal Au Ag Cu Mo W Ni Co
Fermi Level (eV) -5.10 -4.26 -4.65 -4.60 -4.55 -5.15 -5.00
Resistivity (10-10 Ω-cm) 2.27 1.63 1.73 5.52 5.44 7.20 5.60
Though Au can be used readily in lab-scale production as the back electrode in PSCs, large-
scale production will need a cheaper and more available alternative. Even though the
performance can degrade to some extent due to the mismatch of Fermi level and higher
4. Other factors
Accepted Article
4.1. Encapsulation
oxygen, UV light, and other degrading agents, thus providing additional protection for PSCs.
It has already been demonstrated in multiple pieces of research how encapsulation plays a
vital role to slow down the degradation processes.[36, 111-114] Gianmarco et al. in their study
with encapsulation, found four desirable properties of encapsulants; (i) Chemical inertness,
(ii) Low cost and facile processibility, (iii) High barrier for moisture and oxygen, and (iv)
High transmission in the visible spectrum.[115] Based on these properties, different materials
and processes have been developed for encapsulation.[116-118] The most commonly used
method is glass to glass encapsulation, in which a glass sheet surrounds the whole device. On
the contrary, the most promising candidate for encapsulant is thin-film encapsulation
consisting of an ultra-thin protective layer of different materials like Al2O3, TiO2, and ZnSO4,
which comes with additional challenges in terms of cost and processing techniques.[119-121]
Uddin et al. discussed the role and desirable properties of encapsulation materials, including
low Water Vapor Transmission Rate and Oxygen Transmission rate.[114] It is already
established that both moisture and oxygen degrade the perovskite layer. Thus, encapsulation
can provide necessary blocking towards their ingress and delay the degradation to some extent.
However, encapsulation cannot delay the decomposition permanently. Even with shallow
water and oxygen transmission rate, they will ultimately reach the perovskite layer.
Sai et al. reported a solvent-free and low-temperature processable encapsulant paraffin along
with Ultraviolet Curative Additive (UVCA), which stabilized the mixed halide based PSC
curve that the devices had very similar PV performance. Additionally, as shown in Figure 12,
they were able to show that both the devices demonstrated good stability in 65 °C temperature
and 50% RH condition. However, when it comes to stability at the maximum power point
Accepted Article
(MPP), the device without paraffin degraded rapidly while the devices with paraffin were
stable for more than 1000 hours, retaining 80% of its efficiency. The authors proposed that the
addition of paraffin with UVCA helped to passivate the defects and suppressed phase change
of the perovskite, which resulted in improved stability. Their claim was supported by a
change of colour in the devices after 40 days under illumination (AM 1.5), which showed
minimal change for UVCA and paraffin devices. The stability test was also performed at
86 °C, and the devices with paraffin showed much better stability. Moreover, the devices
Figure 12. (a) Photovoltaic performance of the devices before and after encapsulation with
Paraffin and UVCA. (b) External quantum efficiency and integrated photocurrent for Paraffin
While improving stability, encapsulants can block some portion of incident light from
Accepted Article
reaching the absorber layer that brings additional challenges. The transmittance of the
Proper encapsulation can potentially eliminate extrinsic degradation in PSCs when designed
efficiently.
Electric field induced by photovoltage facilitates ion migration in and out of the perovskite
layer, which results in a change of band structure of the absorber as well as transport layers.
dependence of the J-V curve of a solar cell on the direction and rate of voltage scan is known
performance of the devices. In the absence of an electric field, perovskite contains MA,
Halide and Pb ions. High mobility and low activation energy cause these ions to move, and
the vacancy can be filled by the migrating ions causing defects in the structure. Under electric
fields, these ions move towards ETL and HTL due to the attraction of opposite charges. In
this way, they generate their electric field opposite to the electric field that caused them to
There are several possible reasons for hysteresis, including ferroelectric effect, loss of balance
in the transport of charge carriers, ion migration, trap assisted recombination and others.[124-
127]
Almost all these factors leading to hysteresis are associated with the movement of ions,
defects or vacancies. These movements trap the photogenerated charge carriers and account
complex process to account for hysteresis rather than a single process, and also, these
processes are related to each other. Several works have been published recently demonstrating
the cause and effect of hysteresis, but there is no concrete evidence yet on the origin and long-
Accepted Article
term effect of hysteresis on the stability of PSCs.[123, 125, 128-131] Elumalai et al. divided the
origin of hysteresis into several mechanisms, including (but not limited to) ferroelectric
polarization, charge trapping, ion migration and capacitive effects.[131] Their report was one of
the earliest ones documenting the cause, effect and ways to mitigate hysteresis. However, an
independent grouping of the causes is controversial to the widely believed theory that no
single mechanism alone can cause hysteresis. Instead, hysteresis is the combined effect of
multiple mechanisms. For quantifying hysteresis, a parameter called Hysteresis Index (HI) is
used. Both Equations 6.1 and 6.2 have been used to measure HI.[132, 133]
𝑃𝐶𝐸𝑅𝑆 −𝑃𝐶𝐸𝐹𝑆
𝐻𝐼 = (6.2)
𝑃𝐶𝐸𝑅𝑆
Different strategies have been employed to overcome hysteresis, including reducing trap
density, enhanced charge transport and using structurally stable materials, all of which
shown in Figure 13 and obtained 180 °C as the optimal temperature for annealing.[134] The
authors reported that a minor energy offset between conduction band minimum (CBM) and
the Fermi level (EF) resulted in the smallest HI. While no explanation is given on this
hypothesis, they demonstrated that HI was minimum corresponding to the temperature, which
CBM and EF is 0.22 eV, giving the lowest HI about 0.12. Reproduced with permission.[134]
In their analysis of hysteresis, Zhong et al. showed the effect of different ETL used alongside
CH3NH3PbI3‐xClx.[135] They used reference perovskite layer, perovskite PCBM bilayer and
perovskite polymer-PCBM bilayer for this analysis. Hysteresis was found the least in the
devices with perovskite PCBM bilayer. The authors proposed that the reduction of hysteresis
was achieved due to the ability of PCBM molecules to infiltrate the perovskite layer and
passivate the iodine and boundary defects, thus retarding the rate of ion migration in the
Multiple cation perovskite has also been thought of as another way to fabricate hysteresis free
PSCs. Using cations like Cs that removes strain and Rb, which act as passivation centre, has
been reported to reduce the hysteresis effect. Treating the device with positive azeotropes,
which act as Lewis bases, can remove defects in the perovskite layer and suppress
hysteresis.[136] While it can be said that any means that retards the ion migration in the
perovskite device for photovoltage will reduce hysteresis, the exact cause and long-term
In different studies, tests for long-term stability are often done using different accelerated
aging tests, while burn-in degradation tests can be done in optimum condition since it
consumes sufficiently less time.[137, 138] These tests aim to estimate the performance of the
Accepted Article
cells over a long period. During these tests, the cells have to go through harsh conditions,
including higher temperature, higher RH, a large amount of UV radiation and high
mechanical stress. The cells must retain about 80% of their initial stability for individual tests
On the other hand, for short term stability tests which are frequently done in studies of PSCs,
the cells are kept in ambient conditions for limited time. However, the conditions vary widely
from study to study. Since there is no universally accepted set of conditions, authors in each
study assume the conditions.[139] While some conditions are more challenging for the cells to
sustain than others, there is no efficient comparison between different studies due to the wide
range of testing conditions found in the literature. Moreover, in some studies, some conditions
Jeffrey et al. divided the stability of PV devices into three levels based on the mechanisms of
degradation.[140, 141] They are material stability (Level 1), cell stability (Level 2) and module
stability (Level 3). At level one, no standards are needed since it is mainly based on the
materials from which the devices are fabricated. Level 2, however, needs standardized
conditions of the operating and testing environment. Level 3 needs to follow International
operations. At this moment, perovskites are only fabricated in a lab-scale, and controlled
environment and only Level 1 and Level 2 stability are measured. The authors reviewed the
first two levels of stability in this work since no work has been done on level 3 stability on
think of.
This article is protected by copyright. All rights reserved
Due to the lack of a universal testing protocol for stability, for PSCs, different research groups
assume the conditions that vary widely. These varied conditions make it harder for
researchers in this field to directly compare the results since stability highly depends on
testing conditions. It is both necessary and urgent to formulate a unified protocol for stability
Accepted Article
testing like other solar cells. Researchers on organic solar cells encountered a similar
condition, and after a summit in 2019 on organic photovoltaics (OPV), a set of strategies were
PSCs are still far away from commercialization due to their poor stability. For any solar cell
to commercialize, three factors of interest are also referred to as the golden triangle.[143] These
are the cost of the cells, the PCE of the devices and finally, the stability of the devices. At
present, the PCE of perovskite solar cells is comparable with silicon and CdTe solar cells. The
highest reported efficiency of Si solar cells is 26.7%, and the highest reported efficiency of
comparable with thin-film CdTe cells and considerably lower than Si solar cells (USD 0.24
per Watt).[145, 146] However, the highest reported lifetime of PSCs is one year, contrary to 25
years lifetime of silicon solar cells.[147] Moreover, the PSC, which had a one-year lifetime, had
and cost per watt for Silicon and Perovskite solar cells.[21, 22, 90, 144-147]
From the Golden Triangle, as shown in Figure 14, it is evident that the stability of PSCs is
lagging the stability of Si solar cells by far. All three parameters of the golden triangle are
equally important. If one of them falls behind so much, the other two can barely compensate
for that. Therefore, research on the stability of PSCs is of paramount importance since the
PCE and cost are already good enough for commercialization. The number of documents
published in recent years also shows that the trend of research on PSC has shifted towards
stability. Figure 15 demonstrates both the increase of research articles published in recent
years and the fraction of articles related to stability has increased with respect to the total
number of articles on Perovskite solar cells. It is believed that this trend will continue in the
near future until both the stability and PCE of perovskite solar cells are saturated towards a
fixed value.
Figure 15. Number of Papers published related to PSCs in different years from 2012-2020
and the number of papers with stability keywords among those. Data collected from Scopus
searching “Perovskite Solar Cell” first and then searching “Stability” within the results.
5. Effect on Environment
While solar energy is greener than any other form of energy, the materials used in the devices
can sometimes be toxic to the environment. Pb, which is often used as the inorganic cation in
the perovskite material, has alarming toxicity levels.[148, 149] Therefore, it can cause damage to
the environment in case of encapsulation failure, and the people involved in the production
are also at risk of Pb toxicity. Several studies involving the toxicity of PSCs and their impact
methylamine are also harmful to human health when exposed in a large amount.[152-154]
Inhalation is one of the routes in which MAPbI3 can go inside the body, and the material,
being soluble in organic fluids can reach the brain through the blood and other body fluids.[155,
156]
process of different PSCs. Gong et al. reported that 1kWh electricity generation by TiO2 and
ZnO based PSCs releases 82.5 g and 60.1 g of CO2 equivalent, respectively, while the energy
needed in manufacturing per square meter of PSC is 7.78 kWh.[157] In another study, Espinosa
Accepted Article
et al. estimated that 5.48 kg CO2 equivalent was generated to produce one kWh energy while
0.146 kWh energy was needed per square meter of PSC device manufacturing.[110] As seen
clearly, there is no coherence between the results. Some other works published with LCA of
PSCs also estimated the same parameters, but the result showed divergent values.[109, 158, 159]
If PSC technology is commercialized in future, another worrying thing will be the disposal of
the cells after their lifecycle. Due to toxicity, it cannot be normally disposed of and will need
a proper recycling mechanism. Zhang et al. employed an LCA approach to conducting the
environmental impact study of PSCs over their life cycle.[109] The authors of that study
identified that the production of gold as an electrode and different solvents in the production
also harms the environment, which can be minimized by substitution of gold by silver or
aluminium. It is suggested that proper burning of the materials after the life cycle having an
energy recovery system can cause minimal environmental impact. It is also reported that the
use of aluminium as the top electrode in PSCs and tandem cells can lead to less energy
payback time while eliminating substantial harm to the environment.[151] Though detailed life
cycle study and disposal techniques do not need urgent addressing, upon commercialization of
PSCs, it can cause harm if proper research is not done to maintain the standards in the
5.1. Discussion:
Stability is the major obstacle to the successful commercialization of PSCs. This review
focused on all the factors leading to the degradation of PSCs. Additionally, recent progress
This article is protected by copyright. All rights reserved
towards the long-term stability of PSCs is also analyzed critically. The authors discussed and
compared recent studies and techniques that focused on enhancing the stability of the overall
PSC devices. Recent techniques include incorporating the buffer layer, including novel ETL
or HTL, using mixed perovskites, doping different materials with different reagents, and
Accepted Article
others. Comparison of different techniques and analysis of device morphology suggest that
improved stability is obtained for band matched devices with larger grain sizes. All the
methods eventually improve the surface of the materials and interface with the adjacent
material. Therefore, a good indication of long-term stability can be obtained by observing the
interface and band alignment of the devices. Therefore, it is suggested in future research to
analyze the device physics with novel materials before device fabrication. It can be helpful to
estimate the device performance and to get insights on improving the device further.
Though lots of research work has been done to improve PSCs' stability, there are still gaps in
the literature offering opportunities for further research. From the golden triangle, it has been
shown that the PCE and cost of PSCs are outstanding, but stability is not yet good enough.
Therefore, future researchers focus in this field should be on the stability issue. It is shown in
Figure 15 that research work in PSCs has reasonably shifted towards solving the stability
misinterpretation. The authors showed that several novel studies with potentially stable
devices had no stability test in ambient conditions. It is urgent for the PSC research
community to form stability criteria as a protocol for stability testing. Unified stability testing
protocols will help researchers test the stability of novel devices or techniques and analyze the
stability of the devices. Efficient comparison of different works will lead the research in PSC
in the right direction to improve the stability of the devices. Uniform stability testing protocol
will also help future researchers to select from multiple techniques to work further.
6. Conclusion
This article is protected by copyright. All rights reserved
It is evident that multiple factors are responsible for the degradation of PSC. While some are
well documented, there is still a lack of agreement about the origin and effect of some factors
such as hysteresis and encapsulation. Some of the degradation mechanisms are also hazardous
for the environment due to the emission of lead-containing compounds. In this article,
Accepted Article
different mechanisms that affect the stability of PSCs are critically analyzed. Moreover,
different groups' recent studies to improve stability against some of these factors are critically
evaluated. Additionally, the effect on the environment and the attempt of different groups to
replace lead from perovskite are analyzed in this review. Further study is required to
characterize the behaviour of some of these factors and their origin. Research has been done
in a tremendous amount in recent years, and improvements in stability have seen a rapid rise,
which is believed to continue in the future. Finally, the lack of a universal protocol of testing
standards that hampers research and comparison of different studies is also discussed.
Several research articles have been published in recent years on PSC and PSC stability.
However, due to extensive research in recent years, the field is constantly evolving. The idea
about different degradation mechanisms, use of new materials as ETL, HTL and interface
layers and different perovskite materials are evolving so fast. This article critically discusses
the recent trend of research and literature in PSC stability which is, of course, prone to being
updated in future. The authors have discussed the degradation mechanisms of PSCs, which
are widely understandably believed in the PV community. For comparing and efficiently
evaluating, the testing conditions such as temperature, RH, the spectrum of light are essential.
It is also desired that the device's performance at different humidity conditions, temperatures,
and UV levels be reported in the stability studies for meaningful and efficient comparison in
subsequent studies.
Being the most promising PV technology, PSC still lags in terms of stability. There has been
three parameters of the Golder Triangle allow this breakthrough. Ultimately, this technology
will have to comply with IEC 61646 standards for transforming into a commercial product.
For research in the stability of PSCs, being updated with literature is crucial due to its fast-
Accepted Article
changing nature. This article critically introduces the up-to-date knowledge of degradation
engineering, and hysteresis to improve the stability further. An alternative of gold should be
found for industrial-scale manufacturing of PSC devices. As a final note, stability testing
protocols are the most important at this point for meaningful stability studies. It is urged that
the PSC research community establish standard protocols for stability studies.
Acknowledgements
The authors would like to acknowledge the financial support from Australian Government
Research Training Program Scholarship. The authors also appreciate the constructive
discussion by the OPV group members of UNSW throughout this work.
References
Supply Chains", in Polygeneration with Polystorage for Chemical and Energy Hubs, K.R.
[8] M. Wright, A. Uddin, Solar Energy Materials and Solar Cells 2012, 107, 87.
2015, 592308.
solar cells", in Solar Cells and Light Management, F. Enrichi and G.C. Righini, Eds.,
[13] D. Zhou, T. Zhou, Y. Tian, X. Zhu, Y. Tu, Journal of Nanomaterials 2018, 2018,
8148072.
[15] R. Ali, M. Yashima, Journal of Solid State Chemistry 2005, 178, 2867.
[20] P.-P. Sun, Q.-S. Li, S. Feng, Z.-S. Li, Physical Chemistry Chemical Physics 2016, 18,
14408.
[21] J. Y. Kim, J.-W. Lee, H. S. Jung, H. Shin, N.-G. Park, Chemical Reviews 2020, 120,
7867.
This article is protected by copyright. All rights reserved
[22] Q. Wali, F. J. Iftikhar, M. E. Khan, A. Ullah, Y. Iqbal, R. Jose, Organic Electronics
[25] Y. Liu, Z. Yang, D. Cui, X. Ren, J. Sun, X. Liu, J. Zhang, Q. Wei, H. Fan, F. Yu, X.
1405.
[29] P. Luo, Z. Liu, W. Xia, C. Yuan, J. Cheng, Y. Lu, Journal of Materials Chemistry A
2015, 3, 12443.
[30] Z. Wei, H. Chen, K. Yan, S. Yang, Angewandte Chemie (International ed. in English)
[31] Z. Xiao, C. Bi, Y. Shao, Q. Dong, Q. Wang, Y. Yuan, C. Wang, Y. Gao, J. Huang,
2018, 6, 5507.
[34] G. Tong, H. Li, G. Li, T. Zhang, C. Li, L. Yu, J. Xu, Y. Jiang, Y. Shi, K. Chen, Nano
[36] S. Bai, P. Da, C. Li, Z. Wang, Z. Yuan, F. Fu, M. Kawecki, X. Liu, N. Sakai, J. T.-W.
Wang, S. Huettner, S. Buecheler, M. Fahlman, F. Gao, H. J. Snaith, Nature 2019, 571, 245.
Accepted Article
[37] D. Wang, M. Wright, N. K. Elumalai, A. Uddin, Solar Energy Materials and Solar
[40] R. Wang, M. Mujahid, Y. Duan, Z.-K. Wang, J. Xue, Y. Yang, Advanced Functional
[42] Y. Lv, H. Zhang, R. Liu, Y. Sun, W. Huang, ACS Applied Materials & Interfaces
[43] F. Haque, H. Yi, J. Lim, L. Duan, H. D. Pham, P. Sonar, A. Uddin, Materials Science
[44] A. Ghosh, S. S. Dipta, S. S. S. Nikor, N. Saqib, A. Saha, J. Opt. Soc. Am. B 2020, 37,
1966.
[54] H.-S. Kim, C.-R. Lee, J.-H. Im, K.-B. Lee, T. Moehl, A. Marchioro, S.-J. Moon, R.
Accepted Article
Humphry-Baker, J.-H. Yum, J. E. Moser, M. Grätzel, N.-G. Park, Scientific Reports 2012, 2,
591.
338, 643.
2015, 3, 7219.
[60] S. Ito, S. Tanaka, K. Manabe, H. Nishino, The Journal of Physical Chemistry C 2014,
118, 16995.
[61] S.-W. Lee, S. Kim, S. Bae, K. Cho, T. Chung, L. E. Mundt, S. Lee, S. Park, H. Park,
M. C. Schubert, S. W. Glunz, Y. Ko, Y. Jun, Y. Kang, H.-S. Lee, D. Kim, Scientific Reports
2016, 6, 38150.
2013, 4, 2885.
[63] Y. Sun, X. Fang, Z. Ma, L. Xu, Y. Lu, Q. Yu, N. Yuan, J. Ding, Journal of Materials
[65] W. Xiao, J. Yang, S. Xiong, D. Li, Y. Li, J. Tang, C. Duan, Q. Bao, Solar RRL 2020,
4, 1900394.
Accepted Article
[66] B. Li, Y. Li, C. Zheng, D. Gao, W. Huang, RSC Advances 2016, 6, 38079.
[70] Z. Cao, C. Li, X. Deng, S. Wang, Y. Yuan, Y. Chen, Z. Wang, Y. Liu, L. Ding, F.
[71] D. Yang, X. Zhang, K. Wang, C. Wu, R. Yang, Y. Hou, Y. Jiang, S. Liu, S. Priya,
[72] J. Ma, Z. Lin, X. Guo, L. Zhou, J. Su, C. Zhang, Z. Yang, J. Chang, S. Liu, Y. Hao,
[74] D. Chi, S. Huang, S. Yue, K. Liu, S. Lu, Z. Wang, S. Qu, Z. Wang, RSC Advances
2017, 7, 14694.
[75] R. Chen, J. Cao, Y. Duan, Y. Hui, T. T. Chuong, D. Ou, F. Han, F. Cheng, X. Huang,
Jiang, W. Mao, X. Lin, X. Wen, J. Lu, Y.-B. Cheng, U. Bach, Advanced Energy Materials
Accepted Article
2019, 9, 1901519.
2018, 6, 2219.
2018, 6.
[82] M. Wang, H. Wang, W. Li, X. Hu, K. Sun, Z. Zang, Journal of Materials Chemistry A
2019, 7, 26421.
[83] Q. Zhao, R. Wu, Z. Zhang, J. Xiong, Z. He, B. Fan, Z. Dai, B. Yang, X. Xue, P. Cai,
[85] J. Cao, H. Yu, S. Zhou, M. Qin, T.-K. Lau, X. Lu, N. Zhao, C.-P. Wong, Journal of
[87] P.-L. Qin, Q. He, C. Chen, X.-L. Zheng, G. Yang, H. Tao, L.-B. Xiong, L. Xiong, G.
[88] N. J. Jeon, H. Na, E. H. Jung, T.-Y. Yang, Y. G. Lee, G. Kim, H.-W. Shin, S. Il Seok,
[90] W. K. Lin, S. H. Su, M. C. Yeh, C. Y. Chen, M. Yokoyama, Vacuum 2017, 140, 82.
[91] H. Lee, C. Lee, H.-J. Song, Materials (Basel) 2019, 12, 1644.
Accepted Article
[92] X. Yin, C. Wang, D. Zhao, N. Shrestha, C. R. Grice, L. Guan, Z. Song, C. Chen, C. Li,
G. Chi, B. Zhou, J. Yu, Z. Zhang, R. J. Ellingson, J. Zhou, Y. Yan, W. Tang, Nano Energy
[93] X. Liu, Y. Wang, F. Xie, X. Yang, L. Han, ACS Energy Letters 2018, 3, 1116.
[94] B. Zhao, X. Yan, T. Zhang, X. Ma, C. Liu, H. Liu, K. Yan, Y. Chen, X. Li, ACS
146, 1659.
[96] J. Wang, J. Li, Y. Zhou, C. Yu, Y. Hua, Y. Yu, R. Li, X. Lin, R. Chen, H. Wu, H. Xia,
H.-L. Wang, Journal of the American Chemical Society 2021, 143, 7759.
[98] J. Wu, J. Shi, Y. Li, H. Li, H. Wu, Y. Luo, D. Li, Q. Meng, Advanced Energy
[99] D. Yang, W. Ming, H. Shi, L. Zhang, M.-H. Du, Chemistry of Materials 2016, 28,
4349.
Zhang, C. Yan, W. Li, X. Liu, Y. Lu, J. Kirman, E. H. Sargent, Q. Xiong, Z. Wei, Nature
Gonçales, M. Wright, J. Justin Gooding, F. Haque, C. Xu, A. Uddin, Solar Energy Materials
[104] Y. Zong, Y. Zhou, Y. Zhang, Z. Li, L. Zhang, M.-G. Ju, M. Chen, S. Pang, X. C.
[105] W. Ming, D. Yang, T. Li, L. Zhang, M.-H. Du, Advanced Science 2018, 5, 1700662.
[106] L. Lin, C. Gu, J. Zhu, Q. Ye, E. Jiang, W. Wang, M. Liao, Z. Yang, Y. Zeng, J. Sheng,
W. Guo, B. Yan, P. Gao, J. Ye, Y. Zhu, Journal of Materials Science 2019, 54, 7789.
[109] J. Zhang, X. Gao, Y. Deng, Y. Zha, C. Yuan, Solar Energy Materials and Solar Cells
2017, 166, 9.
[117] Y. I. Lee, N. J. Jeon, B. J. Kim, H. Shim, T.-Y. Yang, S. I. Seok, J. Seo, S. G. Im,
[122] S. Ma, Y. Bai, H. Wang, H. Zai, J. Wu, L. Li, S. Xiang, N. Liu, L. Liu, C. Zhu, G. Liu,
X. Niu, H. Chen, H. Zhou, Y. Li, Q. Chen, Advanced Energy Materials 2020, 10, 1902472.
[123] P. Liu, W. Wang, S. Liu, H. Yang, Z. Shao, Advanced Energy Materials 2019, 9,
1803017.
[124] J. Wei, Y. Zhao, H. Li, G. Li, J. Pan, D. Xu, Q. Zhao, D. Yu, The Journal of Physical
[125] B. Chen, M. Yang, S. Priya, K. Zhu, The Journal of Physical Chemistry Letters 2016,
7, 905.
[128] Y. Zhao, W. Zhou, W. Ma, S. Meng, H. Li, J. Wei, R. Fu, K. Liu, D. Yu, Q. Zhao,
[131] N. K. Elumalai, A. Uddin, Solar Energy Materials and Solar Cells 2016, 157, 476.
Accepted Article
[132] Y. Rong, Y. Hu, S. Ravishankar, H. Liu, X. Hou, Y. Sheng, A. Mei, Q. Wang, D. Li,
M. Xu, J. Bisquert, H. Han, Energy & Environmental Science 2017, 10, 2383.
Dyakonov, M. L. Petrus, A. Baumann, P. Docampo, ACS Appl Mater Interfaces 2018, 10,
11414.
[138] F. Bella, D. Pugliese, L. Zolin, C. Gerbaldi, Electrochimica Acta 2017, 237, 87.
2018, 3, 2136.
[141] J.-W. Lee, N.-G. Park, Advanced Energy Materials 2020, 10, 1903249.
Breemen, C. Girotto, E. Voroshazi, F. C. Krebs, Solar Energy Materials and Solar Cells
Crystals 2019, 9.
[146] M. Cai, Y. Wu, H. Chen, X. Yang, Y. Qiang, L. Han, Advanced Science 2017, 4,
1600269.
[148] E. Charney, B. Kessler, M. Farfel, D. Jackson, The New England journal of medicine
[157] J. Gong, S. B. Darling, F. You, Energy & Environmental Science 2015, 8, 1953.
Table of Contents:
the market breakthrough of this technology. These factors are critically analyzed in this article.
Several groups are working to overcome these obstacles and have achieved promising results
in the past few years. Notable studies on the stability of PSCs are also critically analyzed.
Accepted Article
Table of Contents Figure (55 mm broad × 50 mm high):