Single-Source Thermal Ablation of halide perovskites, limitations and opportunities: The lesson of MAPbBr3

Article abstract of DOI:10.1016/j.jallcom.2021.159954

Single-Source Thermal Ablation (SSTA) is a vacuum-based evaporation method that was only minorly employed in the preparation of hybrid perovskite materials for solar cells despite some promising early achievements. We study the preparation of MAPbBr₃ films, showing that this material does not allow the reproducible preparation of thin films suitable for integration in multijunction devices due to the low adhesion of MABr to the substrate, on one hand, and to the proximity of decomposition and melting temperatures of MAPbBr₃ combined with the high vapor pressure of MABr on the other. Based on the insights obtained on MAPbBr₃ processing we demonstrate that, conversely, homogeneous fully-inorganic CsPbI₂Br thin films can be prepared by SSTA with a stability comparable to that reported for films obtained by spin coating. This work provides guidelines for the selection of halide perovskites that can be successfully prepared by SSTA for thin film application.

Full text of DOI:10.1016/j.jallcom.2021.159954

Journal of Alloys and Compounds 875 (2021) 159954
Contents lists available at ScienceDirect
Journal of Alloys and Compounds
journal homepage: www.elsevier.com/locate/jalcom
Single-Source Thermal Ablation of halide perovskites, limitations and
opportunities: The lesson of MAPbBr3
Davide Calestania,1, Lucia Nasi , Francesco Mezzadria,b, Francesco Fracassic,d
,
a,1
c
a
a,⁎
Andrea Listorti , Patrizia Ferro , Roberto Mosca
a
CNR-IMEM Institute of Materials for Electronics and Magnetism, Parma, Italy
Department of Chemistry, Life Sciences and Environmental Sustainability, University of Parma, Parma, Italy
Department of Chemistry, University of Bari “Aldo Moro”, Bari, Italy
b
c
d
CNR-NANOTEC Institute of Nanotechnology, c/o Department of Chemistry, University of Bari ‘Aldo Moro’, Bari, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Single-Source Thermal Ablation (SSTA) is a vacuum-based evaporation method that was only minorly
employed in the preparation of hybrid perovskite materials for solar cells despite some promising early
achievements. We study the preparation of MAPbBr films, showing that this material does not allow the
3
reproducible preparation of thin films suitable for integration in multijunction devices due to the low
adhesion of MABr to the substrate, on one hand, and to the proximity of decomposition and melting
Received 23 December 2020
Received in revised form 27 March 2021
Accepted 9 April 2021
Available online 15 April 2021
temperatures of MAPbBr
insights obtained on MAPbBr
CsPbI Br thin films can be prepared by SSTA with a stability comparable to that reported for films obtained
by spin coating. This work provides guidelines for the selection of halide perovskites that can be success-
fully prepared by SSTA for thin film application.
3
combined with the high vapor pressure of MABr on the other. Based on the
Keywords:
3
processing we demonstrate that, conversely, homogeneous fully-inorganic
Flash evaporation
Halide perovskite
Thin film
2
3
CH NH
3
PbBr
3
CsPbI
2
Br
© 2021 Elsevier B.V. All rights reserved.
1. Introduction
opportunities into complex devices such as monolithic tandem solar
cells and LEDs. Most of the works on vacuum-based methods for
Organic-inorganic metal-halide perovskites are currently con-
3
ABX halide perovskites exploit the dual source evaporation method
sidered a very interesting class of materials mainly because of their
outstanding impact in solar cell application [1] that fostered also
strong developments in perovskite-based optoelectronic applica-
tions exploiting not only thin films [2], but also nanocrystals [3,4].
Among PV applications, halide perovskites appear very appealing
also for their potential exploitation in tandem with silicon to im-
prove the performance of commercially available solar cells and
reduce the cost of photovoltaic modules [5]. In most of these ap-
plications, poor morphology and bad substrate coverage limit the
performance of the devices. In order to achieve smooth and pinhole-
free films from solution methods, different approaches have been
proposed [6,7]. Compared to solution-processing methods, vacuum
evaporation naturally provides uniform and compact films with high
reproducibility and fine thickness control, while offering integration
[8–10], even if the use of different precursor vapors makes it difficult
to keep control over the precursor ratio. Single-Source Thermal
Ablation (SSTA, also named Flash Evaporation) is a simple alternative
approach to the vacuum deposition of halide perovskites films,
based on a single thermal source consisting in a metal foil heater
used to vaporize rapidly the perovskite precursor by passing a large
current. This method was used to prepare hybrid perovskites in the
1990s and 2000s for different applications [11,12], but in the recent
season of halide perovskites for PVs, to the best of our knowledge,
only very few works exploited this deposition technique. The first
were Longo et al. [13], who performed the flash evaporation of
MAPbI in low vacuum (∼0.1 mbar), obtaining films having a high
3
degree of crystallinity and a smooth surface morphology. These films
allowed power conversion efficiencies exceeding 12% in planar het-
erojunction solar cells. Later on, Xu et al. [14] showed that MAPbI
3
films prepared in of 0.1 mbar Ar pressure from stoichiometric pre-
cursors have pure perovskite phase but a rough surface with voids.
⁎
Corresponding author.
E-mail address: roberto.mosca@imem.cnr.it (R. Mosca).
These authors have contributed equally to this work
On the contrary, film deposited under a low growth pressure of 5
−5
10
2
mbar are dense and without pinholes, but PbI -free films can
1
https://doi.org/10.1016/j.jallcom.2021.159954
0925-8388/© 2021 Elsevier B.V. All rights reserved.

D. Calestani, L. Nasi, F. Mezzadri et al.
Journal of Alloys and Compounds 875 (2021) 159954
only be achieved by using a precursor with an MAI-to-PbI
2
molar
The precursor layer for evaporation was created by spreading
32 μl of the precursor solution onto a tantalum boat (Testbourne
S46-.005Ta) that was then placed on a hotplate at 100 °C, while
maintaining the precursor distribution as even as possible. For
ratio of 2, which is necessary to compensate for the MAI loss during
the transport of material from the source crucible to the substrate.
They also showed that grain growth is limited by the presence of the
PbI
using PbI
SSTA of formamidinium lead iodide (FAPI) [15], showing that PbI
free films can only be obtained by carrying out the deposition on
substrates heated at 105 °C and using a FAI to PbI molar ratio of 1.5
to compensate for the loss of FAI. The perovskite solar cell they
obtained by using these films showed a 12.55% efficiency, but sta-
bility was lower than that of cells prepared by solution methods.
In spite of these rather promising results and the potential ad-
vantages deriving from being a relatively simple and quick process,
SSTA was left behind without a clearly defined reason. Indeed, we
2
phase in the films and obtained a ∼10% maximum efficiency by
2
CsPbI Br 90 μl were used.
2
-free films. Recently, the same group reported about the
The boat was then clamped between two electrodes in the eva-
poration chamber that was immediately pumped to vacuum. When
the pressure in the chamber was below 2 10−5 mbar, electrical cur-
rent was passed through the boat to promote the precursor eva-
poration. Argon was finally inlet and samples removed from the
chamber with the ambient humidity maintained below 30%. Unless
stated otherwise, films were then stored into a desiccator in
the dark.
2
-
2
Under these conditions, for a vertical distance of 8 cm between
the crucible and the substrate, a film thickness of 240 nm was
measured by a standard stylus profilometer.
3
recently showed that CsPbBr films prepared by SSTA allow PeLED
preparation, which confirms the suitability of SSTA for thin film
application.
When aiming to investigate the evolution of the evaporation
process, deposition was performed in two steps: in the former one, a
90 W power preset was applied until the color of the precursor layer
was seen to change from orange to whitish (step 1). The chamber
was then opened, substrates were replaced by virgin others and the
chamber was pumped again to vacuum. Then, a second evaporation
was performed by applying the same 90 W power until the boat
became incandescent (step 2).
Unless otherwise stated, all the film characterizations were car-
ried out in low humidity atmosphere (RH≤30%) after having aged the
films for a few hours in desiccator. In particular, the structural
quality of the films was studied by X-ray diffraction (XRD) mea-
In this work we consider the deposition of MAPbBr
SSTA, showing that SSTA provides pure phase MAPbBr
without any MABr excess sufficiently when high powers are used,
but the proximity of the MAPbBr decomposition and melting
3
films by
3
films
3
temperatures combined with the high vapor pressure of MABr
produces spits on the film surface. Reduced power allows spit-free
films, but requires the use of a MABr excess to compensate for the
MABr losses during evaporation. This MABr excess yields pure phase
MAPbBr
reproducibility. These achievements point out that MAPbBr
suitable for thin film application can hardly be obtained by SSTA.
Based on the results obtained on MAPbBr we demonstrate fully-
inorganic CsPbI Br films which are instead suitable for applications
such as light emitting devices and tandem solar cells.
3
films, but at the cost of a significant loss of stability and
3
films
surements in a Siemens (D500) powder diffractometer, with CuK
α
3
radiation. The refinement of XRD data was carried out using the
GSAS II software [17].
2
Film absorbance spectra were measured by a Jasco UV–vis V-530
spectrometer. Morphological characterization was performed by
atomic force microscopy using a Veeco Dimensions 3100 SPM and by
scanning electron microscopy (SEM) using a Zeiss Auriga field
emission microscope (FESEM) operated at 5 kV. Selected Area
Electron Diffraction (SAED) and Energy Dispersive X-ray
Spectrometry (EDS) were performed in a Jeol 2200FS Transmission
Electron Microscope (TEM). To minimize any damage caused by the
electron beam, the TEM measurements were carried out by using an
acceleration voltage of 80 kV and a low beam current density.
Steady state and time resolved photoluminescence were mea-
sured by an Edinburgh FLS920 spectrometer equipped with a Peltier-
cooled Hamamatsu R928 photomultiplier tube (185–850 nm). An
Edinburgh Xe900 450 W Xenon arc lamp was used as exciting light
source. Corrected spectra were obtained via a calibration curve
supplied with the instrument (lamp power in the steady state PL
experiments 0.6 mW cm−2, spot area 0.5 cm ). Emission decay time
were determined with the single photon counting technique by
means of the same Edinburgh FLS980 spectrometer using a laser
diode as excitation source (1 MHz, exc = 635 nm, 67 ps pulse width
2
. Methods
Hybrid perovskite precursor solutions were prepared by using
commercially available PbBr
CH NH (41% ethanol solution) (Fluka) without further purification.
2
(Aldrich), HI (57%) (Aldrich), and
3
2
MABr was prepared by reacting amine solutions with HBr, as de-
scribed previously [16]. In particular, methylamine (40% in water,
Aldrich) was mixed with hydrobromic acid (48% in water, Aldrich) in
a 1:1 molar ratio. After continuous stirring in the ice bath for 2 h,
MABr was crystallized by removing the solvent at 65 °C for 2 h. The
resultant precipitates, in the form of white crystals, were washed
three times in diethyl ether, dried overnight under vacuum, and then
stored in a dark, dry box.
2
MAPbBr
3
precursor solution (0.785 M) was prepared by stirring
in N,N-
overnight at 60 °C equimolar mixtures of MABr and PbBr
dimethylformamide (Fluka) and then naturally cooling down the
solution, that appeared clear at room temperature. In order to
2
achieve precursor solutions with a MAPbBr
0 ≤ x ≤ 2), proper amounts of MABr were added to the perovskite
solution at room temperature while stirring.
For CsPbI Br films, a 0.45 M precursor solution was prepared by
dissolving stoichiometric amounts of CsBr and PbBr in dimethyl
sulfoxide (Aldrich) at room temperature under stirring.
3
:MABr molar ratio of 1:x
and about 30 ps time resolution after deconvolution) and a
(
Hamamatsu MCP R3809U-50 (time resolution 20 ps) as detector.
(Laser power in the TRPL experiment 1.6 W cm−2
0.3 mm ) [18].
, spot area
2
2
2
3. Results and discussion
Glass substrates were cleaned by sequential rinsing in 1%
Hellmanex, de-ionized water, hot acetone, and isopropanol for
In order to investigate the influence of the MABr excess, films
were prepared using a power preset of 440 W. Unfortunately, the
elemental composition could not be assessed by EDS because the Br/
Pb ratio was found to rapidly drop upon electron irradiation, which
limits the measurement accuracy. The XRD patterns of all the films
are dominated by two peaks at 14.93° and 30.12°, respectively,
15 min. After drying in blowing nitrogen, up to four substrates were
placed on a sample holder that was mounted into the evaporation
chamber at a fixed vertical distance from the evaporation source.
The evaporation system was tailor made [12] following the sche-
matization given by Mitzi et al. [11] and allowed power to be in-
creased from 0 to the preset value with a rate of approximately
(Fig. 1) corresponding to the 100 and 200 reflections of the MAPbBr
cubic phase [19]. However, pure phase films were achieved only
3
6
00 W/s, maximum power being 1000 W.
2

D. Calestani, L. Nasi, F. Mezzadri et al.
Journal of Alloys and Compounds 875 (2021) 159954
Ref. [14], we infer that the power used in that case increased from
~
48 W to ~130 W during evaporation. Here, when x = 1 was used,
pure phase MAPbBr films were obtained only for power presets of
at least 400 W (Fig. 2), while lower powers systematically produce
PbBr peaks in the XRD pattern. This means that, achieving pure
phase MAPbBr films requires a steeper temperature rise compared
to MAPbI [14], likely because of the higher vapor pressure of MABr.
3
2
3
3
Films obtained with x = 1 and 400 W power preset show a
compact and smooth surface with roughness of about 3.2 nm
(Fig. 3A). The absorption spectra show an onset at ~535 nm (Fig. 3B)
that is typical of MAPbBr films prepared by spin coating [16,20]. The
3
bandgap of ∼2.33 eV determined by the Tauc plot may be considered
in good agreement with the 2.31 eV value reported in the literature
[
16,20,21]. Indeed, the presence of interference fringes pointed out
by the broad maximum at 590 nm, while confirming the good sur-
face quality of the films, prevents a more accurate determination of
the bandgap.
Fig. 1. XRD patterns of MAPbBr
MABr excesses. Stars indicate the PbBr
obtained from Ref. [19] is shown at the bottom.
3
films prepared by 440 W power preset with different
2
peaks. The reference pattern of MAPbBr
3
To rationalize how film properties would impact its optoelec-
tronic performance, we recorded the steady state photo-
luminescence, measured its quantum yield (PLQY) and analyzed the
PL decay (TRPL) of as-deposited perovskite films. The PL band is
practically superimposed to the absorption band with minimum
stoke shift. This is in line with what usually observed for MAPbBr
perovskites suggesting an emission from direct-gap recombination
22]. Noticeably, PLQY recorded at low excitation fluency, being
.23% implies reasonably good optoelectronic properties of the
material, as without special treatments polycrystalline film of
MAPbBr usually show PLQY below 1% [23]. Halide perovskites de-
3
[
0
3
spite the straightforward deposition methodologies often exhibit
superior optical properties than many other materials, but they still
suffer from great loss of PL intensity and lifetime induced by trap-
states. The emission decay time mirror what suggested by PLQY
analysis as relatively long emission lifetime are retrieved. By fitting
the PL decay with a multi-exponential decay law [24], we determine
an average lifetime ( ) of 16.0 ns (τ
= 49.70; τ = 20.99 ns, α = 31.13).
These results point out that film deposition performed by using a
1
= 1.06 ns, α
1
= 19.17; τ
2
= 5.87 ns,
α
2
3
3
Fig. 2. XRD patterns of films obtained with (1:1) MAPbBr
using different power presets, as indicated. Stars indicate the PbBr
3
to MABr molar ratio by
2
peaks. The re-
MABr excess of x = 1 and a 400 W power preset provides good
quality material. Unfortunately, we found that, except for the posi-
tion of the absorbance onset, achieving an acceptable reproducibility
of the film properties was a real issue, which forced us to investigate
the evaporation process better. To this aim, we exploited a low
power preset (90 W) to have a temperature rise slow enough to
allow us to interrupt the deposition immediately after the orange
precursor layer starts to change color to off-white. This way we
found that at low powers the precursor decomposes so that the
evaporation is made of two distinct steps: in the former one mainly
MABr evaporates, while in the latter step, corresponding to the
ference pattern of MAPbBr obtained from Ref. [19] is shown at the bottom.
3
when x ≥ 1 was used, while for x < 1 peaks appear at 18.60°, 21.65°
and 37.72°, that can be assigned to the 020, 120 and 040 reflections
of the PbBr orthorhombic phase, respectively (ICSD 202134). The
2
observation that an excess of MABr is necessary to achieve pure
phase films suggests that MABr is lost during evaporation. A similar
result was reported for MAPbI
lammonium halide excess was necessary to avoid the presence of
PbI phase in the perovskite films. In that case, a current of 180 A was
considered high enough to yield the instantaneous evaporation of
both PbI and MAI. Based on the experimental details given in
3
[14], where the same methy-
2
2
evaporation of the off-white layer, only PbBr is deposited, in close
2
similarity with what addressed for MAPbI films [14]. Increasing the
3
Fig. 3. Typical A) AFM image, B) absorbance and photoluminescence spectra (the relevant time-resolved PL decay is shown in the inset) of films prepared with (1:1) MABr-to-
MAPbBr molar ratio and a power preset of 400 W.
3
3

D. Calestani, L. Nasi, F. Mezzadri et al.
Journal of Alloys and Compounds 875 (2021) 159954
3
Fig. 4. Typical photographs of a film prepared by using a preset power of 440 W with a precursor containing a x = 1 MABr excess (i.e. with a 1:1 MAPbBr :MABr molar ratio).
Images were taken A) soon after extraction and B) after 20 min.
power preset makes the two steps gradually less distinct, so that at
4
00 W the formation of an off-white layer becomes less obvious than
at low powers and is barely visible at 800 W. However, we cannot
rule out that the MAPbBr precursor decomposes at least partially
even at high powers [25]. This two-step deposition process means
that achieving pure phase MAPbBr films requires that, after de-
position, ion interdiffusion occur between the MABr-rich layers near
the substrate and the PbBr -rich upper layers. Indeed, after their
3
3
2
extraction from the deposition chamber films change color from pale
yellow to orange in a few minutes through the formation of orange
spots that appear soon after extraction and then extend up to cover
the whole sample (Fig. 4). This evolution, that becomes more evident
when the power preset is reduced and the MABr excess increased, is
3
similar to the one observed in CsPbBr films deposited by SSTA,
where it was ascribed to a vertical composition gradient due to the
precursor decomposition [25]. It is worth noting that this evolution
takes place immediately after the exposure to ambient air, while
keeping films in Ar or dry air flux for up to an hour before extraction
did not prevent this evolution. This suggests that ambient humidity
plays a role. Moisture is known to play a significant role during the
formation of methylammonium lead halide perovskite films [26],
although the details of involved mechanisms are still subject of
debate [27]. In particular, water was proposed to enhance the re-
construction process during film formation by partially dissolving
the reactants and accelerating mass transport within the film [28],
while the exposure to humidity improves perovskite crystal-
linity [29].
Fig. 5. XRD patterns of two films deposited in the same run (with x = 1 and 440 W
power preset). Films were exposed to X-rays for 30 min a) soon after extraction from
the deposition chamber, b) after 2 h storage in the same atmosphere but in the dark.
Relevant XRD measurements were performed immediately after this X-ray irradiation
2 3
Stars point out peaks ascribed to PbBr . The reference pattern of MAPbBr obtained
from Ref. [19] is shown at the bottom.
The limited stability of as prepared films was confirmed by the
observation that films placed under vacuum soon after extraction
systematically show the presence of PbBr
connected to the formation of features (Fig. 6B) that are similar to
the PbBr clusters found in cleaved single crystals stored in vacuum
for weeks [30]. The mechanism of PbBr formation is likely the same
proposed for MAPbI films prepared by spin coating, where PbI
appears after 1 month vacuum storage due to the loss of volatile
parts, likely HI and CH NH [31]. Interestingly, twin samples pre-
pared in the same deposition run do not exhibit PbBr signatures
after up to 14 days storage in desiccator (Fig. 6, patterns a and b). In
addition, PbBr formation was not observed in MAPbBr films pre-
2
(Fig. 6A, pattern c), likely
2
We tried to investigate how this initial evolution affects the
crystalline quality of the films. The XRD pattern of a film exposed to
X-rays soon after deposition (Fig. 5) exhibits not only the peaks ty-
2
3
2
pical of the MAPbBr
3
cubic phase [19], but also peaks belonging the
signatures were
3
2
PbBr orthorhombic phase. On the contrary, PbBr
2
2
2
not observed on a film that, coming from the same deposition run,
was kept for about two hours in the same RH = 30% atmosphere, but
not exposed to X rays. In addition, soon after extraction the height of
2
3
pared by spin coating [16] that remain pure phase after more than
one-year storage in vacuum (Fig. 6, pattern d). These results confirm
that soon after extraction, films are prone to degradation likely due
to the highly volatile and hygroscopic character of MABr. This means
2
the XRD peaks of both perovskite and PbBr change with time,
whereas negligible pattern changes were observed under X-ray ir-
radiation when measurements were performed after having stored
films for 2–4 h in desiccator or in low humidity (RH≤ 30%) atmo-
sphere (Fig. 6A). These results highlight that X-ray irradiation in-
terferes with the initial film evolution, which limits the possibility of
studying the influence of the deposition parameters on the film
crystallization soon after deposition as well as the initial evolution
itself.
3
that the achievement of MAPbBr films comes after an initial evo-
lution phase that is influenced by not only the evaporation para-
meters (e.g. vacuum level, power preset) but also the environmental
conditions the films experience during and immediately after its
extraction from the evaporation chamber. This film delicacy is a real
issue because it makes the extremely accurate control of both
4

D. Calestani, L. Nasi, F. Mezzadri et al.
Journal of Alloys and Compounds 875 (2021) 159954
Fig. 6. A) XRD patterns of samples coming from the same deposition run as measured after a) 2 h storage in desiccator, b) 14 days in desiccator, and c) 2 days in vacuum. (the star
indicates the 002 reflection of orthorhombic PbBr ). Pattern d) refers to a film deposited by spin coating and measured after storage in vacuum for 390 days. The reference pattern
of MAPbBr obtained from Ref. [19] is shown at the bottom. B) AFM image taken on a film stored for 2 days in vacuum.
2
3
when power used is high enough, no MABr excess is necessary to
achieve PbBr -free films.
2
Unfortunately, while on the micrometric scale surface mor-
phology appears flat and compact, similar to Fig. 3a, a larger view
points out that bumps are present on the film surface with dimen-
sions ranging from a few hundred nanometers to a few tens of mi-
crometers (Fig. 8). EDS measurements performed on these bumps
pointed out Br/Pb ratios that change with electron irradiation time,
gradually approaching that of PbBr
was observed in the regions where the film is flat, which suggests
that MAPbBr perovskite films are unstable under an electron beam,
as reported for MAPbBr nanocrystals [32]. This instability hindered
2
. However, the same behavior
3
3
also TEM measurements, so that bump composition finally could not
be assessed. Nevertheless, we observed that there is no correlation
between the density and size of bumps and the presence of PbBr
the XRD patterns, as we found bumps (Fig. 8B) even in films showing
pure MAPbBr perovskite phase (pattern c in Fig. 8F). Besides, we
observed that bump density decreases with decreasing power
(Fig. 8C-E), which suggests that these are MAPbBr bumps likely due
2
in
3
Fig. 7. XRD patterns of MAPbBr films prepared at 800 W by using different MABr
excesses, as indicated. The reference pattern of MAPbBr
shown at the bottom.
3
obtained from Ref. [19] is
3
atmosphere and timings soon after extraction a prerequisite for
achieving acceptable reproducibility of film properties.
3
to spitting. Indeed, SSTA combines a rapid temperature rise, which is
necessary to prevent the decomposition of the organic component,
Since the initial evolution comes from the vertical composition
gradient due to the delay between MABr and PbBr
2
evaporation, we
2
with high powers required to make PbBr and MABr evaporations as
further increased power in order reduce this delay. When a power
preset of 800 W was used, the initial color change was not observed,
simultaneous as possible. The rapid achievement of high tempera-
tures is generally considered as one of the main causes of spitting
because when the melting temperature is reached the molten pre-
cursor spreads over the crucible that continues heating very rapidly.
3
while pure phase MAPbBr films were obtained even for a stoi-
chiometric precursors composition (x = 0) (Fig. 7). This means that,
Fig. 8. A) and B) SEM images of bumps. C)-E) optical microscope images of film prepared without any MABr excess by using different power presets, namely C) 800 W, D) 400 W,
and E) 90 W (markers indicate 250 µm). The relevant XRD patterns are shown in F) with the MAPbBr
peak positions).
3 2
reference pattern obtained from Ref. [19] (empty diamonds indicate PbBr
5

D. Calestani, L. Nasi, F. Mezzadri et al.
Journal of Alloys and Compounds 875 (2021) 159954
This creates vapor bubbles that can explode through the surface
giving the spitting of molten material. Materials having high vapor
pressures at their melting points are more likely to give spits [33]. In
3
the present case, the formation of MAPbBr spits can be justified
considering that this material starts decomposing at ~220 °C [34].
While 3D lead perovskites are known to melt at temperatures near
or above their decomposition temperature [6]. High heating rates are
expected to shift the decomposition reaction to higher temperatures,
so that decomposition and melting processes can overlap due to the
fast crucible heating. Consequently, decomposition may take place in
the presence of molten perovskite, which combined with the high
vapor pressure of MABr may cause perovskite spits. This would be
consistent with the observation that only the slow temperature raise
obtained by power presets as small as 90 W prevents bump forma-
tion (Fig. 8D). Unfortunately, at these low powers the presence of
PbBr
of at least x = 2 allows pure phase MAPbBr
reduced film reproducibility and stability, as discussed above. The
limited suitability for thin film application of MAPbBr films ob-
2
is unavoidable (pattern d in Fig. 8E), so that only a MABr excess
3
films, which implies
3
tained at high powers was confirmed by the lack of any electro-
luminescence in PeLEDs prepared by using these films.
We can conclude that this compound presents us with an in-
soluble problem, as the poor MABr adhesion on the substrate forces
us to use high powers, that make films unsuitable for thin film ap-
plication due to the spitting caused by the proximity of the de-
composition temperature to the melting temperature, combined
with the high vapor pressure of MABr. On the other hand, achieving
pure phase MAPbBr films by exploiting low powers to avoid spitting
3
requires the use of MABr excesses that limit film stability, making
reproducibility a real issue.
From the above results, we gather that, whenever the precursor
compound does not sublimate before decomposing, SSTA deposition
starts with the decomposition of the perovskite precursor followed
by the evaporation of the organic halide first and then of the lead
halide. Although perovskite synthesis reaction is extremely favor-
able, achieving pure phase thin films becomes problematic if any of
the evaporating species shows a low sticking coefficient while pre-
cursor decomposition temperature is close to the melting tem-
perature, because of spitting phenomena, especially if any of the
decomposition products has a high vapor pressure. This allows us to
Fig. 9. (A) XRD patterns measured (a) soon after sample extraction from the de-
position chamber, (b) soon after an annealing at 300 °C for 60 s (c) after having stored
an annealed sample in desiccator for one month, and d) after having exposed an
annealed sample to ambient air (RH = 35%) for 180 min. The stick patterns for the
orthorhombic perovskite and orthorhombic non-perovskite structures are also
rationalize the preparation of CsPbBr
previously reported [25]. Indeed, we obtained CsPbBr
stoichiometric precursors independently of the power used, con-
firming the absence of sticking problems for either PbBr or CsBr.
Besides, the melting temperature of CsPbBr (567 °C) [35] is much
3
films by SSTA that we have
shown. (B) Typical SEM image of films deposited by SSTA at 400 W from the CsPbI
2
Br
3
films from
precursor and annealed at 300 °C for 60 s.
2
3
The XRD patterns of freshly annealed brown films (Fig. 9A pat-
tern b), despite being affected by (101) preferential orientation, are
consistent with the orthorhombic perovskite polymorph of CsPbI Br
2
[41]: Rietveld refinement of the PXRD data allowed us to determine
the lattice parameters as a = 8.4873(19) Å, b = 11.981(6) Å,
c = 8.558(2) Å. The SAED patterns (not shown) taken by scratching
part of the brown film just before performing XRD, are in good
agreement with the XRD patterns reported in Fig. 9A. Moreover,
Quantitative EDS-TEM analysis showed that the Cs:Pb:I:Br is in good
agreement with the 1:1:2:1 ratio, with a variation of about ± 2 atom
larger than the decomposition one (450 °C) [36], which allowed us to
take full advantage of the high evaporation rate of SSTA by in-
creasing power while avoiding spitting, also because of the low
vapor pressure of both CsBr and PbBr
Based on the achievements on MAPbBr
that also CsPbI Br appears suitable to be prepared by SSTA since its
2
compared to MABr.
3
and CsPbBr , we infer
3
2
melting temperature (463 °C) [37] is significantly higher than its
decomposition temperature (330 °C) [38]. In addition, the vapor
pressure of PbI
2
is only slightly larger than that of PbBr
2
[39] and we
%
evaluated as the standard deviation in the fit. The EDS maps ob-
did not observe any adhesion issue. CsPbI
2
Br combines the expected
tained in scanning TEM mode, shows a uniform distribution of Cs,
Pb, I, and Br.
Absorbance spectra taken on brown films point out the presence
of interference fringes (Fig. 10) that confirm the good surface quality
of films. The absorbance onset at ~650 nm corresponds to a bandgap
of 1.92 eV, as determined by the Tauc plot, that is in good agreement
stability advantages deriving from the use of cesium in place of
methylammonium cations with a great potential in tandem solar cell
application due to the bandgap of about 1.9 eV [37]. For these rea-
sons, we investigated the deposition of CsPbI Br films by SSTA. XRD
2
patterns taken immediately after film extraction from the evapora-
tion chamber show poor crystallization (Fig. 9A, pattern a), but, si-
with the value reported for CsPbI
coating [37,40]. The above results confirm that SSTA provides good
quality CsPbI Br films.
We finally assessed the stability of the films obtained by SSTA,
because the CsPbI Br brown perovskite phase, although used in PV
2
Br thin films prepared by spin
2
milar to what reported for CsPbI Br films prepared by solution
methods [37,40], films turn dark brown after an annealing at tem-
peratures larger than 250 °C. The surface of films annealed at 300 °C
for 60 s appears compact and free of bumps, regardless of the power
preset used (400 W in Fig. 9B).
2
2
6

D. Calestani, L. Nasi, F. Mezzadri et al.
Journal of Alloys and Compounds 875 (2021) 159954
Fig. 10. Absorbance spectra of films annealed at 300 °C for 1 min (A) spectra measured either soon after annealing or after storage in desiccator, as indicated; (B) spectra measured
after exposure to ambient air (RH = 35%) for times indicated. The inset shows the conversion from the brown phase to the yellow one for a CsPbI
air (RH = 35%).
2
Br film exposed to ambient
applications, is known to be metastable at room temperature, the
thermodynamically stable structure being an orthorhombic non-
perovskite structure (yellow phase). After having stored the film in
desiccator for up to one month, we found that the XRD pattern of the
aged film is consistent with the just annealed one (Fig. 9A, pattern b
and c). Besides, storing films in desiccator for up to one month does
not affect significantly its absorbance spectrum (Fig. 10A), whereas
exposing the film to ambient air progressively decreases its absor-
bance at a rate that depends on humidity. Indeed, in RH≤ 27% the
absorption intensity at the peak of onset (624 nm) decreases by less
than 0.5% after 120 min, whereas at RH~35% the decrease is already
3
powers produce MAPbBr spits on the film surface, thus hindering
their thin film application. Spitting is due to the proximity of the
decomposition and melting temperatures, combined with the high
vapor pressure of MABr. Using intermediate powers allows stoi-
chiometric films to be obtained only when a precursor containing a
proper MABr excess is used, but the presence of an initial evolution
makes these films prone to degradation, while spits persist. Low
powers yield spit-free films, but pure phase MAPbBr cannot be
3
achieved even by increasing the MABr excess to x = 2, which, in turn,
affects reproducibility and stability. Therefore, the combination of
the delay between MABr and PbBr
adhesion on the substrate, the proximity of the MAPbBr
position and melting temperatures, and the high vapor pressure of
MABr prevents us from obtaining by SSTA pure phase MAPbBr films
suitable for thin film application. Based on these results, we have
however rationalized that CsPbBr films can be successfully pre-
pared by SSTA because the decomposition and melting temperatures
are far enough, while the decomposition products (CsBr and PbBr
2
evaporation, the poor MABr
~
3.5% after 20 min and continues until the perovskite absorbance
3
decom-
step completely disappears after about 3 h (Fig. 10B). During the
exposure to ambient air, pale spots appear and then increase in size
to cover the entire film (inset in Fig. 10B). After the color change has
come to completion, in addition to interference fringes the absor-
bance spectrum points out a step at ~430 nm, that corresponds to a
3
3
2
.94 eV bandgap, as obtained from the Tauc plot. The XRD pattern of
2
)
converted films (Fig. 10 pattern d) is consistent with a [100] oriented
orthorhombic non perovskite phase [42] having lattice parameters
a = 10.191(2) Å b = 5.016(3) Å c = 17.462(12) Å, thus confirming that
the brown phase has converted to the yellow one. This behavior is
have low enough vapor pressures and do not show sticking pro-
blems. Moving from this analysis, for the first time we successfully
applied SSTA to CsPbI Br achieving films that not only are suitable
2
for thin film application, but have also a stability comparable to that
fully consistent with what previously reported for CsPbI
obtained by spin coating [37], which points out that SSTA allows us
to obtain CsPbI Br films that are not only suitable for thin film ap-
2
Br films
of films of the same material prepared by solution methods. These
results shine new light on the SSTA technique for the preparation of
fully inorganic halide perovskite thin films.
2
plication, but also have stability properties comparable to those of
films obtained by solution methods.
CRediT authorship contribution statement
2
This first demonstration of CsPbI Br films obtained by SSTA
confirms that the high deposition rate of this technique can be fully
exploited for thin film application when 1) the decomposition and
melting temperatures of the precursor material are not close to each
other and 2) the decomposition products have relatively low vapor
pressure. This allows us to raise some doubts on the efficacy of SSTA
for MA-based halide perovskites, but at the same time to envisage
the suitability of the SSTA technique for the preparation of fully
inorganic halide perovskite thin films.
Davide Calestani: Validation, Investigation, Writing - Review &
Editing. Lucia Nasi: Validation, Investigation, Writing - Review &
Editing. Francesco Mezzadri: Investigation, Writing - Review &
Editing. Francesco Fracassi: Writing - Review & Editing. Andrea
Listorti: Investigation, Writing - Review & Editing. Patrizia Ferro:
Investigation, Resources. Roberto Mosca: Validation, Investigation,
Writing - Original Draft, Writing - Review & Editing, Supervision.
4. Conclusions
Declaration of Competing Interest
We have investigated the preparation of hybrid MAPbBr
ovskite films by SSTA, showing that sufficiently high powers allow
for pure phase films withstanding no MABr excess. However, high
3
per-
The authors declare that they have no known competing fi-
nancial interests or personal relationships that could have appeared
to influence the work reported in this paper.
7

D. Calestani, L. Nasi, F. Mezzadri et al.
Journal of Alloys and Compounds 875 (2021) 159954
Acknowledgments
fast response and small J-V hysteresis, Nano Energy 17 (2015) 269–278, https://
doi.org/10.1016/j.nanoen.2015.08.023
[
21] H.M. Kim, Y.C. Kim, H.J. An, J.M. Myoung, Highly stretchable and contact-re-
LN, DC, FM, PF, and RM thank T. Besagni for technical assistance.
Thanks are due to M. Mazzeo for his availability in the preparation of
PeLEDs. The work was supported by the project entitled “Tecnologia
per celle solari bifacciali ad alta Efficienza a 4 terminali per utility
scale” (BEST4U), Research Project ARS01 _00519 funded under the
PON "R&I" 2014-2020 Program - Action II - D.D. Prot. N. 2059 of 02/
3
sponsive light-emitting diodes based on MAPbBr –PEO composite film, J. Alloy.
Compd. 819 (2020) 153360, https://doi.org/10.1016/j.jallcom.2019.153360
22] Z.Y. Zhang, H.Y. Wang, Y.X. Zhang, Y.W. Hao, C. Sun, Y. Zhang, B.R. Gao, Q.D. Chen,
H.B. Sun, The role of trap-assisted recombination in luminescent properties of
[
3 3 3
organometal halide CH NH PbBr perovskite films and quantum dots, Sci. Rep. 6
(2016) 1–7, https://doi.org/10.1038/srep27286
[
23] K. Lin, J. Xing, L.N. Quan, F.P.G. de Arquer, X. Gong, J. Lu, L. Xie, W. Zhao, D. Zhang,
C. Yan, W. Li, X. Liu, Y. Lu, J. Kirman, E.H. Sargent, Q. Xiong, Z. Wei, Perovskite
light-emitting diodes with external quantum efficiency exceeding 20 per cent,
Nature 562 (2018) 245–248, https://doi.org/10.1038/s41586-018-0575-3
24] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, 3rd ed., Springer, US,
0
8/2018, financed by the Italian Ministry MUR.
[
[
References
2
006.
25] L. Nasi, D. Calestani, F. Mezzadri, F. Mariano, A. Listorti, P. Ferro, M. Mazzeo,
R. Mosca, All-Inorganic CsPbBr perovskite films prepared by single source
[
[
[
[
[
[
[
1] J.Y. Kim, J.W. Lee, H.S. Jung, H. Shin, N.G. Park, High-efficiency perovskite solar
cells, Chem. Rev. 120 (2020) 7867–7918, https://doi.org/10.1021/acs.chemrev.
3
thermal ablation, Front. Chem. 8 (1–10) (2020) 313, https://doi.org/10.3389/
fchem.2020.00313
0c00107
2] H. Cho, C. Wolf, J.S. Kim, H.J. Yun, J.S. Bae, H. Kim, J.M. Heo, S. Ahn, T.W. Lee, High-
efficiency solution-processed inorganic metal halide perovskite light-emitting
diodes, Adv. Mater. 29 (2017), https://doi.org/10.1002/adma.201700579
3] G. Qiao, Z. Zeng, J. Gao, Y. Tang, Q. Wang, An efficient route to assemble novel
organometal halide perovskites and emission evolution performance, J. Alloy.
Compd. 771 (2019) 418–423, https://doi.org/10.1016/j.jallcom.2018.08.322
4] X. Liu, J. Gao, W. Liu, Q. Wang, Reinforcing effects of waterproof substrate on the
photo-, thermal and pH stabilities of perovskite nanocrystals, J. Alloy. Compd.
[26] G.E. Eperon, S.N. Habisreutinger, T. Leijtens, B.J. Bruijnaers, J.J. Van Franeker,
D.W. Dequilettes, S. Pathak, R.J. Sutton, G. Grancini, D.S. Ginger, R.A.J. Janssen,
A. Petrozza, H.J. Snaith, The importance of moisture in hybrid lead halide per-
ovskite thin film fabrication, ACS Nano 9 (2015) 9380–9393, https://doi.org/10.
1021/acsnano.5b03626
[27] A. Dubey, N. Adhikari, S. Mabrouk, F. Wu, K. Chen, S. Yang, Q. Qiao, A strategic
review on processing routes towards highly efficient perovskite solar cells, J.
Mater. Chem. A 6 (2018) 2406–2431, https://doi.org/10.1039/c7ta08277k
[28] H. Zhou, Q. Chen, G. Li, S. Luo, T.-B. Song, H.-S. Duan, Z. Hong, J. You, Y. Liu,
Y. Yang, Interface engineering of highly efficient perovskite solar cells, Science
(80-.) 345 (2014) 542–546, https://doi.org/10.1126/science.1254050
8
17 (2020) 152693, https://doi.org/10.1016/j.jallcom.2019.152693
5] J. Werner, B. Niesen, C. Ballif, Perovskite/silicon tandem solar cells: marriage of
convenience or true love story? – An overview, Adv. Mater. Interfaces 5 (2018)
1
–19, https://doi.org/10.1002/admi.201700731
[29] K.K. Bass, R.E. McAnally, S. Zhou, P.I. Djurovich, M.E. Thompson, B.C. Melot,
Influence of moisture on the preparation, crystal structure, and photophysical
properties of organohalide perovskites, Chem. Commun. 50 (2014)
15819–15822, https://doi.org/10.1039/C4CC05231E
6] W.A. Dunlap-Shohl, Y. Zhou, N.P. Padture, D.B. Mitzi, Synthetic approaches for
halide perovskite thin films, Chem. Rev. 119 (2019) 3193–3295, https://doi.org/
10.1021/acs.chemrev.8b00318
7] A. Giuri, S. Masi, A. Listorti, G. Gigli, S. Colella, C. Esposito Corcione, A. Rizzo,
Polymeric rheology modifier allows single-step coating of perovskite ink for
highly efficient and stable solar cells, Nano Energy 54 (2018) 400–408, https://
doi.org/10.1016/j.nanoen.2018.10.039
[30] J.I.J. Choi, M.E. Khan, Z. Hawash, K.J. Kim, H. Lee, L.K. Ono, Y. Qi, Y.H. Kim, J.Y. Park,
3
Atomic-scale view of stability and degradation of single-crystal MAPbBr surfaces, J.
Mater. Chem. A 7 (2019) 20760–20766, https://doi.org/10.1039/c9ta05883d
[31] I. Deretzis, A. Alberti, G. Pellegrino, E. Smecca, F. Giannazzo, N. Sakai, T. Miyasaka,
[
[
8] L.K. Ono, M.R. Leyden, S. Wang, Y. Qi, Organometal halide perovskite thin films
and solar cells by vapor deposition, J. Mater. Chem. A 4 (2016) 6693–6713,
https://doi.org/10.1039/c5ta08963h
9] M. Sessolo, C. Momblona, L. Gil-Escrig, H.J. Bolink, Photovoltaic devices em-
ploying vacuum-deposited perovskite layers, MRS Bull. 40 (2015) 660–666,
https://doi.org/10.1557/mrs.2015.170
A. La Magna, Atomistic origins of CH
Appl. Phys. Lett. 106 (2015) 3–8, https://doi.org/10.1063/1.4916821
[32] F. Zhu, L. Men, Y. Guo, Q. Zhu, U. Bhattacharjee, P.M. Goodwin, J.W. Petrich,
E.A. Smith, J. Vela, Shape evolution and single particle luminescence of orga-
nometal halide perovskite nanocrystals, ACS Nano 9 (2015) 2948–2959, https://
doi.org/10.1021/nn507020s
3 3 3 2
NH PbI degradation to PbI in vacuum,
[
[
[
[
[
10] F. Mariano, A. Listorti, A. Rizzo, S. Colella, G. Gigli, M. Mazzeo, Thermally eva-
porated hybrid perovskite for hetero-structured green light-emitting diodes,
Appl. Phys. Lett. 111 (2017) 163301, https://doi.org/10.1063/1.5001828
11] D.B. Mitzi, M.T. Prikas, K. Chondroudis, Thin film deposition of organic-inorganic
hybrid materials using a single source thermal ablation technique, Chem. Mater.
[33] D.M. Mattox, Handbook of Physical Vapor Deposition (PVD) Processing, 2nd ed.,
Elsevier, United Kingdom, 2010, https://doi.org/10.1016/B978-0-8155-2037-5.
00025-3
[34] M. Kulbak, S. Gupta, N. Kedem, I. Levine, T. Bendikov, G. Hodes, D. Cahen, Cesium
enhances long-term stability of lead bromide perovskite-based solar cells, J.
Phys. Chem. Lett. 7 (2016) 167–172, https://doi.org/10.1021/acs.jpclett.5b02597
[35] C. Zhang, J.F.S. Fernando, K.L. Firestein, J.E. Von Treifeldt, D. Siriwardena, X. Fang,
1
1 (1999) 542–544, https://doi.org/10.1021/cm9811139
12] R. Mosca, P. Ferro, T. Besagni, D. Calestani, F. Chiarella, F. Licci, Effect of humidity
on the a.c. Impedance of CH NH SnCl hybrid films, Appl. Phys. A Mater. Sci.
3
3
3
3
D. Golberg, Thermal stability of CsPbBr perovskite as revealed by in situ
transmission electron microscopy, APL Mater. 7 (2019) 1–7, https://doi.org/10.
1063/1.5108849
Process. 104 (2011) 1181–1187, https://doi.org/10.1007/s00339-011-6407-z
13] G. Longo, L. Gil-Escrig, M.J. Degen, M. Sessolo, H.J. Bolink, Perovskite solar cells
prepared by flash evaporation, Chem. Commun. 51 (2015) 7376–7378, https://
doi.org/10.1039/C5CC01103E
14] H. Xu, Y. Wu, F. Xu, J. Zhu, C. Ni, W. Wang, F. Hong, R. Xu, F. Xu, J. Huang, L. Wang,
Grain growth study of perovskite thin films prepared by flash evaporation and its
effect on solar cell performance, RSC Adv. 6 (2016) 48851–48857, https://doi.org/
[36] M. Liao, B. Shan, M. Li, In situ Raman spectroscopic studies of thermal stability of
all-inorganic cesium lead halide (CsPbX , X˭Cl, Br, I) perovskite nanocrystals, J.
3
Phys. Chem. Lett. 10 (2019) 1217–1225, https://doi.org/10.1021/acs.jpclett.
9b00344
[37] S. Mariotti, O.S. Hutter, L.J. Phillips, P.J. Yates, B. Kundu, K. Durose, Stability and
1
0.1039/c6ra07549e
15] F. Xu, Y. Tian, W. Wang, Y. Zhu, L. Zeng, B. Yao, Z. Fang, H. Xu, R. Xu, F. Xu, F. Hong,
films by vacuum flash evaporation
2
performance of CsPbI Br thin films and solar cell devices, ACS Appl. Mater.
[
Interfaces 10 (2018) 3750–3760, https://doi.org/10.1021/acsami.7b14039
[38] J.K. Nam, M.S. Jung, S.U. Chai, Y.J. Choi, D. Kim, J.H. Park, Unveiling the crystal
formation of cesium lead mixed-halide perovskites for efficient and stable solar
cells, J. Phys. Chem. Lett. 8 (2017) 2936–2940, https://doi.org/10.1021/acs.jpclett.
7b01067
[39] J.G. Speight, Chapter Four - Properties of Inorganic Compounds, in: J.G. Speight
(Ed.), Environ. Inorg. Chem. Eng. Butterworth-Heinemann, 2017, pp. 171–229,
https://doi.org/10.1016/B978-0-12-849891-0.00004-7
[40] R.J. Sutton, G.E. Eperon, L. Miranda, E.S. Parrott, B.A. Kamino, J.B. Patel,
M.T. Hörantner, M.B. Johnston, A.A. Haghighirad, D.T. Moore, H.J. Snaith,
Bandgap-tunable cesium lead halide perovskites with high thermal stability for
efficient solar cells, Adv. Energy Mater. 6 (2016) 1–6, https://doi.org/10.1002/
aenm.201502458
L. Wang, In situ deposition of black α-FAPbI
for solar cells, J. Mater. Sci. Mater. Electron. 30 (2019) 8381–8389, https://doi.
org/10.1007/s10854-019-01155-w
3
[
16] P. Fedeli, F. Gazza, D. Calestani, P. Ferro, T. Besagni, A. Zappettini, G. Calestani,
E. Marchi, P. Ceroni, R. Mosca, Influence of the synthetic procedures on the
structural and optical properties of mixed-halide (Br, I) perovskite films, J. Phys.
Chem. C 119 (2015) 21304–21313, https://doi.org/10.1021/acs.jpcc.5b03923
17] B.H. Toby, R.B. Von Dreele, GSAS-II: the genesis of a modern open-source all
purpose crystallography software package, J. Appl. Crystallogr. 46 (2013)
[
[
5
44–549, https://doi.org/10.1107/S0021889813003531
18] S. Masi, F. Aiello, A. Listorti, F. Balzano, D. Altamura, C. Giannini, R. Caliandro,
G. Uccello-Barretta, A. Rizzo, S. Colella, Connecting the solution chemistry of PbI
2
and MAI: a cyclodextrin-based supramolecular approach to the formation of
hybrid halide perovskites, Chem. Sci. 9 (2018) 3200–3208, https://doi.org/10.
[41] C.C. Stoumpos, C.D. Malliakas, J.A. Peters, Z. Liu, M. Sebastian, J. Im, T.C. Chasapis,
A.C. Wibowo, D.Y. Chung, A.J. Freeman, B.W. Wessels, M.G. Kanatzidis, Crystal
1039/c7sc05095j
3
growth of the perovskite semiconductor CsPbBr : a new material for high-en-
[
[
19] H. Mashiyama, Disordered cubic perovskite structure of CH
3
NH
3
PbX
3
(X= Cl, Br,
ergy radiation detection, Cryst. Growth Des. 13 (2013) 2722–2727, https://doi.
org/10.1021/cg400645t
[42] D.M. Trots, S.V. Myagkota, High-temperature structural evolution of caesium and
rubidium triiodoplumbates, J. Phys. Chem. Solids 69 (2008) 2520–2526, https://
doi.org/10.1016/j.jpcs.2008.05.007
I), J. Korean Phys. Soc. 32 (1998) S156–S158 〈https://www.jkps.or.kr/journal/
view.html?uid=2518&vmd=Full〉.
3 3 3
20] X. Zheng, B. Chen, C. Wu, S. Priya, Room temperature fabrication of CH NH PbBr
by anti-solvent assisted crystallization approach for perovskite solar cells with
8