Realize larger grain size of CH3NH3PbI3 film with reduced non-radiative recombination for high performance perovskite solar cells via precursor colloidal size engineering

Article abstract of DOI:10.1016/j.jallcom.2021.161300

In CH₃NH₃PbI₃ perovskite solar cells, enhancement grain size of CH₃NH₃PbI₃ to reduce the non-radiative at grain boundaries is an important way to reach high performance perovskite solar cell. However, it is still a challenge to enhance the grain size of CH₃NH₃PbI₃ through a simple and low cost way. In this work, a larger precursor colloidal size is realized through tuning morphology of precursor CH₃NH₃I using a polar solvent of ethanol during purification, yielding a larger grain size of CH₃NH₃PbI₃ film, and the as-prepared perovskite solar cells are shown to be dramatically increased to 17.49% with an increase in short circuit density, fill factor and open circuit voltage, as compared to that (14.28%) in the control device with CH₃NH₃I purified by non-polar solvent of diethyl ether. The investigation result showed the increased efficiency of perovskite solar cells prepared by ethanol purification is ascribed to a faster charge transfer at CH₃NH₃PbI₃/TiO₂ interface resulting from the reduced grain boundary defects. Our work provides a route for improving the CH₃NH₃PbI₃ device efficiency through a simple yet effective approach.

Full text of DOI:10.1016/j.jallcom.2021.161300

Journal of Alloys and Compounds 886 (2021) 161300
Contents lists available at ScienceDirect
Journal of Alloys and Compounds
journal homepage: www.elsevier.com/locate/jalcom
Realize larger grain size of CH3NH3PbI3 film with reduced non-radiative
recombination for high performance perovskite solar cells via precursor
colloidal size engineering
Yan Lia,⁎, Jialu Zhenga, Xuelian Chena, Can Suna, Hao Jianga, Guangrong Lib, Xiaoyong Zhanga,⁎
a School of Materials Science and Engineering, Xi’an Shiyou University, Xi’an 710065, China
b State Key Laboratory for Mechanical Behavior of Materials, School of Materials Science and Engineering, Xi’an Jiaotong University, Xi’an, Shaanxi 710049, China
a r t i c l e i n f o
a b s t r a c t
Article history:
In CH3NH3PbI3 perovskite solar cells, enhancement grain size of CH3NH3PbI3 to reduce the non-radiative at
grain boundaries is an important way to reach high performance perovskite solar cell. However, it is still a
challenge to enhance the grain size of CH3NH3PbI3 through a simple and low cost way. In this work, a larger
precursor colloidal size is realized through tuning morphology of precursor CH3NH3I using a polar solvent of
ethanol during purification, yielding a larger grain size of CH3NH3PbI3 film, and the as-prepared perovskite
solar cells are shown to be dramatically increased to 17.49% with an increase in short circuit density, fill
factor and open circuit voltage, as compared to that (14.28%) in the control device with CH3NH3I purified by
non-polar solvent of diethyl ether. The investigation result showed the increased efficiency of perovskite
solar cells prepared by ethanol purification is ascribed to a faster charge transfer at CH3NH3PbI3/TiO2 in-
terface resulting from the reduced grain boundary defects. Our work provides a route for improving the
CH3NH3PbI3 device efficiency through a simple yet effective approach.
Received 17 April 2021
Received in revised form 4 July 2021
Accepted 22 July 2021
Available online 27 July 2021
Keywords:
Methylammonium lead iodide (CH3NH3PbI3)
Perovskite solar cells
CH3NH3I purification
Grain size
Grain boundary defect
© 2021 Published by Elsevier B.V.
1. Introduction
carriers that are recombined at the grain boundary [4]. For example,
by adding small molecule in the precursor solution to passivate the
Today, power conversion efficiencies (PCEs) of the hybrid per-
ovskite solar cells (PSCs) have exceeded 25%, and the maximum
module areas have reached 808 cm2. The rapid progress makes the
hybrid PSCs potentially appealing for fabrication of future photo-
voltaics among other types of photo-absorbers. So far, greening of
precursor solution, high throughput production technology, high
environment stability and fabrication high quality perovskite film
are all the directions for its commercialization.
The high quality perovskite film requires reduced non-radiative
recombination during the charge transport processes. The non-ra-
diative recombination via defects at the grain boundaries is mainly
focused, which are strongly linked with the composition, crystal-
linity and morphology of the perovskite film. The non-radiative re-
defects in perovskite film [5–7], and to optimal the preparation
method of TiO2 electron transport layer (CTL) to suppress defect
recombination while promote interface charge transfer at the in-
terface [8]. In addition to the aforementioned engineering methods,
though optimization of crystal growth and morphology control to
realize CH3NH3PbI3 (MAPbI3) film with large grain size are also
successful in reducing recombination loss [9,10]. Xiao et al. suc-
cessfully increased the size of MAPbI3 grains with improved crys-
tallinity via a post-treatment procedure using solvent annealing, and
they found that this method was effective in improving overall
performances of the MAPbI3 PSCs [11]. However, the thermal treat-
ment process may induce the decomposing the perovskite film
[12–14]. To get the perovskite film with large grain size without
thermal treatment, solution engineering is another low cost way. Lee
et al. showed that crystal growth and morphology of MAPbI3 can be
easily controlled owing to a weak chemical interaction in the adduct
that are formed as a result of chemical reaction between PbI2 and a
polar aprotic solvent (oxygen-donor N, N′-dimethyl sulfoxide) that
dissolves it [15]. According to the LaMer model, solution processing
of the MAPbI3 crystals are often divided into three stages, which are
supersaturation of precursors, nucleation and growth up [16,17].
combination
will
unavoidably
influence
photogeneration,
conversion and losses of charge carrier in perovskite film, and it will
ultimately affect performances of the PSCs [1–3]. Many strategies
have been proposed to effectively reduce the loss of these charge
⁎
Corresponding authors.
E-mail addresses: li1988yan@163.com (Y. Li), xyzhang@xsyu.edu.cn (X. Zhang).
https://doi.org/10.1016/j.jallcom.2021.161300
0925-8388/© 2021 Published by Elsevier B.V.

Y. Li, J. Zheng, X. Chen et al.
Journal of Alloys and Compounds 886 (2021) 161300
Yang et al. demonstrates that the number of nucleation sites, which
are tuned by chamber pressure during drying of MAPbI3, and plays a
critical role in achieving a full coverage of MAPbI3 film on substrates
as well as to prolong carrier lifetime in MAPbI3 [18]. Bi et al. reported
that desirable MAPbI3 with large grains, which has a high average
aspect ratio of 2.3–7.9, are achieved with less grain boundary defects
via MAPbI3 growth control using non-wetting underlayer. In com-
parison, non-wetting substrates suppress heterogeneous nucleation
and facilitate grain boundary migration in MAPbI3 with less dragging
force during MAPbI3 crystal growth. As a result, MAPbI3 with large
grains and improved crystallinity dramatically reduce charge re-
combination, which is comparable with single crystal MAPbI3 [19].
Further, Yan et al. revealed that the MAPbI3 precursor solution is
actually consisting of colloidal dispersions in a precursor solution
rather than a single-phase solution [20]. They also pointed out that
organic solvent with a polar nature is advantageous in opening the
layered skeleton of PbX2 to form the [PbX6]4– soft template, so that
the MAPbI3 precursor solution are stabilized with polar solvent
when MA+ incorporate [PbX6]4– through electrostatic attraction.
During nucleation, in particular, the nature of the precursor solution,
meaning the solute coordination interaction, the precursor dissolu-
tion and recrystallization as well as the electrostatic attraction and
self-assembly of charged unit etc., strongly impact by the nucleation
rate, the number of nucleates and crystallization kinetics of the
MAPbI3 precursor [21–23]. The grain size of MAPbI3 is strongly de-
pending on the chemistry nature of the MAPbI3 colloidal precursor
solution [24–26]. Meng et al. also shows that the size of
NH2CH3NH2PbI3 (FAPbI3) colloids is reduced by introducing a non-
aprotic formic acid (HCOOH) into the precursor solution, leading to a
more uniform FAPbI3 film [27]. Besides, Li et al. demonstrates that a
grain size (~ 3 µm) in MAPbI3 is achieved via manipulation of the size
of MAPbI3 colloidal crystals with introduction of MACl prior to
thermal treatment [28]. Therefore, the chemistry nature of MAPbI3
colloidal precursor indeed plays an important role in determining
the grain size of MAPbI3 film. Unfortunately, so far seldom re-
searches have been focusing on controlling of MAPbI3 colloidal size
for a high-performance MAPbI3 PSCs without additional additives.
Consequently, a simpler yet effective method is urgently needed to
further increase MAPbI3 grains with less defects at grain boundaries.
In this work, we first study nucleation and crystallization of
MAPbI3 colloids using solution processing as well as the underlying
mechanisms for growth of large grains in MAPbI3. Next, we in-
vestigate the influence of grain boundaries in MAPbI3, where defects
are anticipated to exist, on non-radiative recombination at the
MAPbI3/electron transport layer (ETL) depending on the interface
charge transfer. In particular, the MAPbI3 PSCs that are purified with
an aprotic solvent prior to nucleation of MAPbI3 colloids have shown
a higher efficiency of 17.49%, as compared to the control device
(14.28%). Last but not least, we demonstrate that the large MAPbI3
grains that are originated from the small and size-uniform pre-
cipitates of CH3NH3I (MAI) provide a new route for fabrication of
highly efficient hybrid MAPbI3 PSCs without additional additives for
the future.
purified by two solvents separately for 30 min, which is necessary to
separate the MAI powers from unreacted chemicals. Both solvents
were ethanol (C2H5OH or EtOH, 99.0%, Sigma-Aldrich) and diethyl
ether (C4H10O or Et2O, 99.5%, Sigma-Aldrich). The MAI precipitates
using EtOH and Et2O thus referred as EtOH-MAI and Et2O-MAI, re-
spectively. The purification processes were repeated three times to
ensure that unreacted residues are completely removed from the
EtOH-MAI and Et2O-MAI powders.
To obtain MAPbI3 solution, lead iodide (PbI2, 99.9%, Xi’an Polymer
Light Technology Corp.) and the aforementioned MAI powders were
stoichiometrically dissolved (35 wt% in concentration) in a mixed
solution, which was composed of γ-butyrolactone (99%, Sigma-
Aldrich) and dimethylsulfoxide (99%, Sigma-Aldrich) solvents (7:3,
vol:vol), and then stirred for 5 h (70 °C, ambient air). Subsequently,
the MAPbI3 solutions were successfully obtained with EtOH-MAPbI3
and Et2O-MAPbI3 referring to the different MAPbI3 colloids, where
the MAI were purified by EtOH and Et2O, respectively.
2.2. Preparation of MAPbI3 films
Before spin-coating to achieve MAPbI3 layers, FTO substrates
(FTO, TEC-15, LOF) were firstly treated in consecutive ultra-sonic
baths using acetone and isopropanol; 5 min each. Next, the sub-
strates were tried with a nitrogen flow. Secondly, MAPbI3 layers was
spin-coated (4000 rpm, 10 s) in ambient air, followed by a rapid
drying process under a low vacuum (1000 Pa) assisted procedure,
which we referred to as “gas pumping” method and was demon-
strated to be particularly beneficial for growth of MAPbI3 grains by
Yang et al. [29,30]. All of these procedures were carried out at room
temperature (25 °C). Accordingly, EtOH-MAPbI3 and Et2O-MAPbI3
films were obtained depending on the purification solvents during
MAI precipitation as aforementioned.
2.3. Fabrication of MAPbI3 PSCs
First, TiO2 compact layer was spin coated (3000 rpm, 6 s) on the
FTO substrates and thermally treated at 450 °C for 30 min. Next,
EtOH-MAPbI3 and Et2O-MAPbI3 were deposited separately as de-
scribed in Section 2.2 and they were also thermally treated at 100 °C
for 30 min. For hole transport layer, the hole transport material
(HTM) solution was spin coated (3000 rpm, 30 s), where HTM so-
lution consists of 2, 2′, 7, 7′
-
tetrakis(N,N-di-p-methox-
yphenylamine) - 9, 9′ - spirobifluorene (Spiro-OMeTAD) (80 mg), 4-
tert-butylpyridine (28.5 μl) and (17.5 μl) of Lithium-bis(tri-
fluoromethanesulfonyl)imide (Li-TFSI) solution (520 mg Li-TFSI in
1 ml acetonitrile) and chlorobenzene (1 ml). All the necessary che-
micals for preparation of the Spiro-OMeTAD solution were pur-
chased from Xi’an Polymer Light Technology Corp. China. Last, a
silver layer (200 nm) was thermally evaporated, so that the MAPbI3
PSCs were successfully fabricated with EtOH-MAPbI3 and Et2O-
MAPbI3 photoactive layers, coded as EtOH PSCs and Et2O PSCs re-
spectively.
2.4. Characterizations
2. Experimental section
The size of the MAPbI3 colloids was measured by dynamic light
scattering (DLS) using Brookhaven 90Plus particle size analyzer. The
contact angle of the MAPbI3 solutions was determined by surface
tension meter (SDT, KRUSS, Germany) on silicon substrates.
For MAI powders and MAPbI3 films, surface morphologies were
monitored using a field emission scanning electron microscopy
(FESEM, TESCAN, Czech Republic) and an atomic force microscopy
(AFM, Innova, America). X-ray diffraction (XRD) patterns were ob-
tained using an X-ray diffractometer (D8 Advance, Germany) with Cu
Kα radiation. For MAPbI3 films, the absorption spectra were mea-
sured using a UV–VIS spectrophotometer (U-4100, Hitachi, Japan).
2.1. Preparations of MAI powder and MAPbI3 solution
All purchased chemicals were used without purification. First,
the methylamine solution (CH3NH2, 33 wt% in ethanol, Sigma-
Aldrich) was added dropwise into the hydroiodic acid (HI, 57 wt% in
water, Sigma-Aldrich) under stirring. When it reached an equal
molar mass to HI, the mixed solution was stirred for 2 h (0 °C, am-
bient air) to allow complete reaction to MAI. Next, the precipitation
of MAI was undertaken in a rotary evaporator (50 °C, 0.1 MPa), re-
sulting in white MAI powder. Afterwards, the MAI powder was
2

Y. Li, J. Zheng, X. Chen et al.
Journal of Alloys and Compounds 886 (2021) 161300
The steady-state photoluminescence (PL) spectra were performed by
spectrophotometer (Fluoromax-4, Horiba Jobin Yvon, France) with
excitation of mono-wavelength at 560 nm. The time-resolved pho-
toluminescence (TRPL) spectra of the MAPbI3/TiO2 junctions were
measured by a labRAM HR800 setup (Horiba Jobin Yvon, France)
with a 400 nm light pulse from a HORIBA Scientific DeltaPro
fluorimeter.
For MAPbI3 devices, the current density- voltage (J–V) curves
were measured by a solar simulator (Oriel 94023A, Newport) with
light intensity of 100 mW cm–2 and equipped with a Keithley 2400
digital source meter (–0.2–1.3 V). Before measurement, the light
intensity was calibrated with a reference device using a single-
crystal silicon photovoltaic device (91150V, Oriel Instruments). The
active area of the PSC devices was controlled to 0.10 cm2 using a
metal mask. Electrochemical impedance spectroscopy (EIS) spectra
were carried out using an electrochemical workstation (IM6,
ZAHNER, Germany) at the bias condition (–VOC) with an AC probing
signal (amplitude 20 mV, frequency range 10–1 to 106 Hz) without
shielding.
aprotic and non-aprotic solvents of EtOH and Et2O were acted as
such purification solvents. The basis for the selection can be con-
cluded as: during the dissolution process, EtOH has a larger mole-
cular polarity than that in Et2O, which subsequently enhances
dissolution of polar MAI in polar EtOH; During the precipitation
process, the latent heat of evaporation, which measures the
minimum required energy for evaporation of an organic solvent of a
certain mass, in EtOH (840 kJ/kg) is two times higher than that in
Et2O (350 kJ/kg), and then the nucleation and crystallization will be
restrained for the evaporation of EtOH because of the a larger
amount of absorption of heat compared with Et2O.
Indeed, the differ purification procedures result in different mi-
crostructures in the Et2O-MAI and EtOH-MAI precipitates, shown in
Fig. 1(a) and (b), respectively. From the images, EtOH-MAI pre-
cipitates are rather small and are scattered in size (Fig. 1(a)). By
contrast, Et2O-MAI precipitates are relatively larger and appear to be
smooth and continuous in surface morphology (Fig. 1(b)). The pho-
tographic images of both powders are shown in Fig. S1(a) and (b).
The EtOH-MAI powder is shown to have different crystallographic
planes for the different intensity ratios in comparison with Et2O-MAI
(Fig. S1(c)). However, further assignment of each peak with Miller
index is challenging, owing to lack of standard Powder Diffraction
(PDF) card for MAI. The reason of the contrast in MAI surface
morphologies using different purification solvents is anticipated to
be correlated with the polarity of the organic solvents [20].
3. Results and discussions
After precipitation of MAI using rotary evaporation, the MAI
precipitates were repeatedly purified by procedures of dissolution
and precipitation, which is necessary, so that unfavorable chemicals
that are unreacted at an earlier stage are removed from MAI. Herein,
to increase the grain size of perovskite film, the commonly used
Under surface SEM images, both MAPbI3 films are uniform and
fully covered on the surface of TiO2 ETL (Fig. 1(e) and (f)). The EtOH-
Fig. 1. SEM images of MAI powders prepared with (a) EtOH and (b) Et2O. (c) Size distributions of the two MAPbI3 colloid solutions determined by DLS. The inset shows
photographs of the MAPbI3 colloidal precursor solutions. (d) XRD patterns of the two samples. Surface microstructures of (e) EtOH-MAPbI3 and (f) Et2O-MAPbI3 with grain size
distribution. The inset also shows the contact angle of the purifying agents of EtOH and Et2O on a silicon wafer.
3

Y. Li, J. Zheng, X. Chen et al.
Journal of Alloys and Compounds 886 (2021) 161300
MAPbI3 layer shows larger surface grain diameters (range of
250–750 nm, median value of 600 nm) than that those in the Et2O-
MAPbI3 (range of 150–450 nm, median value of 300 nm). However,
irrespective of purifying solvents, both MAPbI3 exhibits the tetra-
gonal perovskite with diffraction peaks of (110), (220) and (310),
observed from the XRD patterns (Fig. 1d). However, {110} peaks
accounts as much in EtOH-MAPbI3 (82.5%) as in Et2O-MAPbI3 (82.2%)
(Table S1). That means the crystallinity stay unchanged in both
samples even if EtOH-MAI has a larger grain size than Et2O-MAI.
Previous results showed the difference microstructures of
MAPbI3 layers was strongly relied on the wetting properties between
the precursor solution and the substrates [18]. In order to confirm
that the size discrepancy of the surface microstructures of the two
MAPbI3 films were affected by the precursor solutions themselves
rather than the wetting properties between the substrate and pre-
cursor solution, the wetting properties of the two MAPbI3 solutions
are examined on the same silicon wafer (Si). The results show that
both solutions have a contact angle of 43° (inset of Fig. 1(e) and (f) as
well as Fig. S2 for other locations of the silicon wafer for accuracy),
which indicates that the precursor solutions themselves, indicating
the colloidal states, would play a major role in surface grain size of
MAPbI3 layer rather than the interface between the substrate and
precursor solution.
The colloids size of the MAPbI3 precursor solution from EtOH and
Et2O purified MAI precipitates are determined using DLS. As it can be
seen in Fig. 1(c), the size of Et2O-MAPbI3 colloids are widely spanned
in the range of 0–600 nm with a median value of approximately
200 nm. Whereas, the size of EtOH-MAPbI3 colloids are concentrated
in the range of 400–700 nm with median value of 600 nm. Based on
the research results published by Espiau de Lamaestre et al. [31],
during the preparation process of the MAPbI3 film, the continuous
evaporation results in an inhomogeneous system for the nucleation
and growth up, and the final size distribution should be apparently
log-normal distribution, therefore the log-normal distribution of the
grain size for both films has been added in Fig. S3. However, based
on Fig. S3, there is almost the same deviation change tendency be-
tween the normal distribution and log-normal distribution for both
films, which illustrates there is dynamic equilibrium between
screening length (λ) and the implanted profile width (Δ), where λ
describes the concentration field interaction of neighboring pre-
cursor colloidal, and Δ is equivalent to the diffusion distance caused
by concentration difference in crystallization process of MAPbI3 film.
The above results prove that the “gas pumping” method for pre-
paring both of the EtOH-MAPbI3 film and Et2O-MAPbI3 film is a fast
mass transfer process, and the main difference is attributed to the
larger colloidal size of the EtOH-MAPbI3 precursor compared with
Et2O-MAPbI3 precursor. The dependence of colloid size of MAPbI3 on
purifying solvents suggests that the morphology of MAI powder
probably has a strong influence on nucleation and growth of MAPbI3
colloids. That is to say, smaller size of precipitates in EtOH-MAI with
larger surface area are likely to induce a rapid dissolution during
reaction, which facilitates a rapid movement of MA+ to be in-
corporated by the [PbI6]4– octahedrons [20].
In order to further investigate the underlying mechanism of the
contrast in grain size due to the size of MAPbI3 colloids, we turn to
the fundamental theories of the crystallization process of MAPbI3
colloids, as depicted in Fig. 2(a) and (b). Typically, the low pressure
preparation method for the perovskite film provides the same su-
persaturation level of the precursor solution, yielding the same
number of the nuclei in the solution at the original, and the large
EtOH-MAPbI3 colloid size is expected to rapidly reduce the driving
force of supersaturation state, as it shows large pieces of stacking,
resulting a large grain size of the film (Fig. 2(b)) [19]. By contrast,
smaller colloid size of Et2O-MAPbI3 are expected to be insufficient to
reduce the driving force of supersaturation state, as it shows small
pieces of stacking, resulting more nuclei appear in the growth up
process, therefore a small grain size of the film is reached (Fig. 2(a)).
The AFM measurements have been down and shown in Fig. S4 in the
Supporting information. As can be seen, the EtOH- MAPbI3 film
shows the larger grain size compared with the Et2O-MAPbI3 film,
which is consistent with the SEM results. The cross sectional view of
both films are shown in Fig. S5.
The planar structure and the band diagram of MAPbI3 devices is
displayed in Fig. 3(a). The maximum J-V curves of both EtOH-MAPbI3
PSC and Et2O-MAPbI3 PSC are shown in Fig. 3(b), and the champion
of EtOH-MAPbI3 PSC exhibit the short circuit density (JSC) of
22.96 mA cm−2, open circuit voltage (VOC) of 1.00 V, fill factor (FF) of
76.41 and power conversion efficiency (PCE) of 17.49%. All of the four
parameters of the highest performance of EtOH-MAPbI3 PSC are
higher than the highest performance of Et2O-MAPbI3 PSC. After 14
more devices are characterized, there are also a clearly improvement
in the PCEs of the EtOH-MAPbI3 PSCs (Figs. S6 and S7 and Table S2)
in comparison with that the device based on Et2O-MAPbI3 film (Figs.
S6 and S7 and Table S3), and the increase of the PCEs can be mainly
attributed to the improvement of the JSC for the VOC being almost
comparable of both solar cells (Fig. S7). It also can be seen, the
discrete distributions of J-V parameters of the EtOH-MAPbI3 PSCs are
obviously restrained in comparison with Et2O-MAPbI3 PSCs. Both of
the EtOH-MAPbI3 PSCs and Et2O-MAPbI3 PSCs have the stabilized
current density at the maximum output power point (Fig. S8) over a
period of 10 min, which illustrate the performance of both PSCs are
stable under continuous illumination.
Fig. 2. Schematic of nucleation and growth process of MAPbI3 (a) with small grain size and (b) with larger grain size.
4

Y. Li, J. Zheng, X. Chen et al.
Journal of Alloys and Compounds 886 (2021) 161300
Fig. 3. (a) Schematic of MAPbI3 device structure and its band diagram. (b) J–V curves and (c) EQE spectra with integrated current densities for EtOH- MAPbI3 and Et2O-
MAPbI3 PSCs.
Fig. 4. (a) Absorption spectra and (b) steady-state PL spectroscopy measurements of EtOH-MAPbI3 and Et2O-MAPbI3 layers. The inset in (a) shows a photograph of MAPbI3 layer
spin-coated on substrates (2.5 × 2.5 cm2 in size). (c) Time-resolved PL spectra of EtOH-MAPbI3/TiO2 and Et2O-MAPbI3/TiO2 interfaces.
To investigate the reasons of the increase of JSC in EtOH-MAPbI3
PSCs, the external quantum efficiency (EQE) of the champion solar
cells in the wavelength range of 300–800 nm (Fig. 3(b)) were char-
acterized. The integrated JSC are 20.47 and 22.50 mA cm–2, which
correlate well with the JSC (20.43 and 22.96 mA cm–2) measured in
the J–V curves for Et2O and EtOH devices, respectively. The EQE
performance of EtOH-MAPbI3 PSC is higher than Et2O-MAPbI3 PSC in
the wavelength region of 300–800 nm, which can be attributed to
the improvement of light absorption ability or the restrained charge
transport and recombination behaviors. In addition, the EQE per-
formance is obviously better in the short wavelength region
(300–400 nm), which is mainly contributed by the collection of
photoexcited charge carriers at the MAPbI3/TiO2 interface, rather
than the rest of the location range. This suggests that interface non-
radiative recombination is reduced owing to improved qualities of
EtOH-MAPbI3 films.
In order to further investigate the reason of an increase in JSC in
the EtOH-MAPbI3 device in comparison with the Et2O-MAPbI3
Fig. 5. Nyquist plots of EtOH-MAPbI3 and Et2O-MAPbI3 PSCs measured at a reverse
bias (amplitude of VOC).
5

Y. Li, J. Zheng, X. Chen et al.
Journal of Alloys and Compounds 886 (2021) 161300
Table 1
transmission line model to describe the Nyquist plots for the two
junctions, where Rce and Cce are resistance and capacitance at the
counter electrode, and Rct and Cct are the charge recombination
resistance and capacitance of the MAPbI3/TiO2 interfaces [34,35]. In
particular, Rct and Rce correspond to the semi-circle in the Nyquist
plots from right (the low frequency region) to left (the high fre-
quency region, too small to be identified in Fig. 5), respectively. By
fitting the Nyquist plots, the important parameters can be extracted
(Table 1). As we are more interested in the MAPbI3/TiO2 interface
charge transfer of the MAPbI3 device, in this work we have per-
formed the EIS studies of the devices at external reverse bias con-
dition (–VOC), so that the actual space charge region is strongly
localized at the MAPbI3/TiO2 interface. Because the external bias
could also alter the probing region during EIS. As can be seen, EtOH-
MAPbI3 PSC has a larger Rct (587.73 Ω) than that (144.81 Ω) in the
Et2O-MAPbI3 PSC, indicating suppressed recombination at the
MAPbI3/TiO2 interface with EtOH purified MAI, compared to Et2O
purified MAI (Table 1).
More importantly, we note that the results that are achieved in
this work might provide a new route that is simple yet effective to
enhance grain size of perovskite film for reducing non-radiative
defect recombination for future energy conversion and storage de-
vices on a broader scope. Purification of MAI using an aprotic solvent
ethanol will certainly be advantageous for synthesis of tandem and
hybrid perovskite for electrolysis cells, as well as enriching the
overall performances of the renewable solar fuel. In addition, to
meet the growing demand of energy for the future, systems that
integrate both energy conversion and storage parts such as solar
cells using different materials e.g., silicon, Cu(In,Ga)(Sn,Se)2, poly-
mers, dye molecules as well as perovskite and lithium-ion battery,
nanogenerator, supercapacitor, photoelectrosynthetic cell and elec-
trolysis cell will be urgently needed. Therefore, reducing defect re-
combination in the materials and systems is urgently needed in
these self-powered systems which are featured to be small, light-
weight, highly packed and reliable.
Lifetime components and recombination resistance extracted from time-resolved PL
measurements and Nyquist plots, respectively.
τ1 / ns
τ2 / ns
Rce / Ω
Rct / Ω
EtOH-MAPbI3
Et2O-MAPbI3
0.14
0.15
12.42
68.08
52.05
44.12
587.73
144.81
device, we performed absorption spectra in the UV–VIS spectral
range (Fig. 4(a)). As can be observed, though there is a slight increase
in absorption (the range of 400–750 nm) in the EtOH-MAPbI3 film,
both films are comparable, therefore, the increase in JSC in the EtOH-
MAPbI3 device can not be attributed to the absorption ability of the
MAPbI3 films. Furthermore, we performed PL on both samples
(Fig. 4(b)). As can be seen, the PL intensity of the major peak that is
located at 765 nm is strongly quenched in the Et2O-MAPbI3 in
comparison with EtOH-MAPbI3. The strong quenching in PL intensity
is thus ascribed to the non-radiative recombination that is due to
defects. Considering the strong body defect tolerance ability of the
organic-inorganic hybrid perovskite films, the restrained quenching
in PL intensity of the EtOH-MAPbI3 can be attributed to the decrease
of the grain boundary defects due to the large grain size of the EtOH-
MAPbI3 film. Next, we investigate charge transportation behavior of
the MAPbI3/TiO2 film using TRPL measurement (Fig. 4(c)), and the
results were fitted using a biexponential decay equation, so that two
components in different time domains (τ1 and τ2) are extracted
(Table 1). The TRPL intensities for the both interfaces are observed to
first have a rapid decay (timescale of approximately 150 ps, Table 1),
irrespective of the MAPbI3 used. The TRPL then decays more slowly
over the nanosecond timescale. For the slow decay, the EtOH-
MAPbI3/TiO2 sample shows a faster decay time τ2 of 12.42 ns, com-
pared to 68.08 ns for the Et2O-MAPbI3/TiO2 sample, which illustrate
a faster charge transfer process of the EtOH-MAPbI3/TiO2 interface
due to the restrained non-radiative recombination at the grain
boundary. Furthermore the theory proposed by Wight et al.[32] has
been used to explain the relationship between slow decay (τ2) and
grain size of both types of MAPbI3 films. A simple Eq. (1) can be
concluded based on their theory, where D is the diffusivity, and S is
the recombination velocity of the grain boundary. Take the EtOH-
MAPbI3 film as a demonstration, based on Fig. 1 and Table 1, τ2 is
12.42 ns, D can be estimate as 1.7 cm2/s [33], the median grain size
value (W) is 600 nm, then based on Eq. (1), the recombination
velocity (S) should be about 5.8 × 103 cm/s, and the D/W should be
~2.8 × 104 cm/s. Therefore, S < D/W, and it should be a recombina-
tion-controlled process in the MAPbI3 film, rather than a diffusion-
controlled process. There is no significant depression of the carrier
concentration, and the bulk and surface-controlled recombination
processes are in parallel in the perovskite film region. Based on
Wight et al. [32] theory, the lifetime of the steady-state and the
transient-state for the recombination-limited case is the same,
therefore the calculated results of the slow decay (τ2) of the EtOH-
MAPbI3 layer is 10.34 ns (i.e. W/S = 10.34 ns, where W is 600 nm and
S is 5.8 × 103 cm/s). The calculated result is comparable to the
measurement result by TRPL. It can be concluded though enlarge the
grain size of EtOH-MAPbI3 film, both the bulk and surface-controlled
recombination are restrained, therefore, it results in an enhanced
PL and a faster charge transfer process (i.e. decreased τ2) of the
EtOH-MAPbI3/TiO2 interface due to the restrained non-radiative
recombination at the grain boundary.
4. Conclusions
In summary, we demonstrate that purification of MAI with the
solution of different polarity prior to the reaction of MAI and PbI2
can effectively change the columnar grain size of MAPbI3 films. The
result show the EtOH-MAPbI3 has a larger columnar grain size
(median value of 500 nm) with less defect containing grain bound-
aries compared with the Et2O-MAPbI3 (median value of 300 nm).
The underlying cause is the increased colloidal size of EtOH-MAPbI3
compared with Et2O-MAPbI3. The PCE of MAPbI3 PSC device is in-
creased from 14.28% to 17.49% with an increase of JSC, FF and VOC
with MAI purified by an aprotic solvent (EtOH). The increasing of JSC
is strongly influenced by an improved absorption in EtOH-MAPbI3
layer and the restrained non-radiation recombination. A strong de-
fect recombination in Et2O-MAPbI3 is further clarified with a strong
quenching in photoluminescence as compared to that in EtOH-
MAPbI3. In addition, Et2O-MAPbI3/TiO2 is shown to have a slow in-
terface charge transfer reflected EtOH-MAPbI3/TiO2 due to a slower
TRPL live time and small resistance.
CRediT authorship contribution statement
Yan Li: Conceptualization, Methodology, Writing – original draft,
Writing – review & editing, Supervision. Jialu Zheng: Investigation,
Writing – review & editing. Xuelian Chen: Methodology. Can Sun:
Methodology. Hao Jiang: Investigation, Data curation. Guangrong
Li: Data curation. Xiaoyong Zhang: Conceptualization, Writing –
review & editing, Supervision.
τ2 = W2/D + W/S
(1)
We also performed electrochemical impedance spectroscopy
(EIS) measurements with Nyquist plots displayed for both PSC de-
vices (Fig. 5) to investigate the influence of grain boundary defects to
the charge transfer at MAPbI3/TiO2 interface. We used a simplified
6

Y. Li, J. Zheng, X. Chen et al.
Journal of Alloys and Compounds 886 (2021) 161300
Declaration of Competing Interest
[14] G. Tumen-Ulzii, C. Qin, T. Matsushima, M.R. Leyden, U. Balijipalli, D. Klotz,
C. Adachi, Understanding the degradation of Spiro-OMeTAD-based perovskite
solar cells at high temperature, Sol. RRL 4 (2020) 2000305.
[15] H. Xiong, G. DeLuca, U. Bach, L. Jiang, Q. Zhang, E. Reichmanis, Y. Qiu, Synergistic
effect of N,N-dimethylformamide and hydrochloric acid on the growth of
MAPbI3 perovskite films for solar cells, ACS Omega 5 (2020) 32295–32304.
[16] C. Yang, Y. Peng, T. Simon, T. Cui, Control of PbI2 nucleation and crystallization:
towards efficient perovskite solar cells based on vapor-assisted solution process,
Mater. Res. Express 5 (2018) 045507.
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.
Acknowledgment
[17] M. Jung, S.G. Ji, G. Kim, S.I. Seok, Perovskite precursor solution chemistry: from
fundamentals to photovoltaic applications, Chem. Soc. Rev. 48 (2019) 2011–2038.
[18] Y. Li, X. Li, Q. Chu, H. Dong, J. Yao, Y. Zhou, G. Yang, Tuning nucleation sites to
enable monolayer perovskite films for highly efficient perovskite solar cells,
Coatings 8 (2018) 408.
[19] C. Bi, Q. Wang, Y.C. Shao, Y.B. Yuan, Z.G. Xiao, J.S. Huang, Non-wetting surface-
driven high-aspect-ratio crystalline grain growth for efficient hybrid perovskite
solar cells, Nat. Commun. 6 (2015) 7747.
This work is also supported by State Key Laboratory for
Mechanical Behavior of Materials. Research fundings from the
Natural Science Foundation Research Project of Shaanxi Province
(Nos. 2019JQ-286, 2019JQ-184 and 2021JQ-603) are greatly ac-
knowledged by the authors.
[20] K. Yan, M. Long, T. Zhang, Z. Wei, H. Chen, S. Yang, J. Xu, Hybrid halide perovskite
solar cell precursors: colloidal chemistry and coordination engineering behind
device processing for high efficiency, J. Am. Chem. Soc. 137 (2015) 4460–4468.
[21] B. Li, D. Binks, G.Z. Cao, J.J. Tian, Engineering halide perovskite crystals through
precursor chemistry, Small 15 (2019) 1903613.
[22] Y.M. Xie, B. Yu, C. Ma, X. Xu, Y. Cheng, S. Yuan, Z.K. Wang, H.T. Chandran, C.S. Lee,
L.S. Liao, S.W. Tsang, Direct observation of cation-exchange in liquid-to-solid
phase transformation in FA1−xMAxPbI3 based perovskite solar cells, J. Mater.
Chem. A 6 (2018) 9081–9088.
[23] Y. Li, B. Ding, G.J. Yang, C.J. Li, C.X. Li, Achieving the high phase purity of
CH3NH3PbI3 film by two-step solution processable crystal engineering, J. Mater.
Sci. Technol. 34 (2018) 1405–1411.
[24] F. Zhang, Z. Ma, T. Hu, R. Liu, Q. Wu, Y. Yu, H. Zhang, Z. Xiao, M. Zhang, W. Zhang,
X. Chen, H. Yu, Ultra-smooth CsPbI2Br film via programmable crystallization
process for high-efficiency inorganic perovskite solar cells, J. Mater. Sci. Technol.
66 (2021) 150–156.
Appendix A. Supporting information
Supplementary data associated with this article can be found in
the online version at doi:10.1016/j.jallcom.2021.161300.
References
[1] F. Ji, L. Wang, S. Pang, P. Gao, H. Xu, G. Xie, J. Zhang, G. Cui, A balanced cation
exchange reaction toward highly uniform and pure phase FA1−xMAxPbI3 per-
ovskite films, J. Mater. Chem. A 4 (2016) 14437–14443.
[2] W. Kong, Z. Ye, Q. Zhen, B. Zhang, H. Wu, Characterization of an abnormal
photoluminescence behavior upon crystal-phase transition of perovskite
CH3NH3PbI3, Phys. Chem. Chem. Phys. 17 (2015) 16405–16411.
[25] H. Zhang, J. Zhuang, X. Liu, Z. Ma, H. Guo, R. Zheng, S. Zhao, F. Zhang, Z. Xiao,
H. Wang, H. Li, Defect passivation strategy for inorganic CsPbI2Br perovskite
solar cell with a high-efficiency of 16.77%, J. Mater. Sci. Technol. 82 (2021) 40–46.
[26] X. Ye, H. Cai, J. Su, J. Yang, J. Ni, J. Li, J. Zhang, Preparation of hysteresis-free
flexible perovskite solar cells via interfacial modification, J. Mater. Sci. Technol.
61 (2021) 213–220.
[27] L. Meng, Q. Wei, Z. Yang, D. Yang, J. Feng, X. Ren, Y. Liu, S. Liu, Improved per-
ovskite solar cell efficiency by tuning the colloidal size and free ion concentra-
tion in precursor solution using formic acid additive, J. Energy Chem. 41 (2020)
43–51.
[28] B. Li, M.J. Li, C.B. Fei, G.Z. Cao, J.J. Tian, Colloidal engineering for monolayer
CH3NH3PbI3 films toward high performance perovskite solar cells, J. Mater.
Chem. A 5 (2017) 24168–24177.
[29] B. Ding, L. Gao, L. Liang, Q. Chu, X. Song, Y. Li, G. Yang, B. Fan, M. Wang, C. Li, C. Li,
Facile and scalable fabrication of highly efficient lead iodide perovskite thin-film
solar cells in air using gas pump method, ACS Appl. Mater. Interfaces 8 (2016)
20067–20073.
[30] Q.Q. Chu, B. Ding, Y. Li, L. Gao, Q. Qiu, C.X. Li, C.J. Li, G.J. Yang, B. Fang, Fast drying
boosted performance improvement of low-temperature paintable carbon-based
perovskite solar cell, ACS Sustain. Chem. Eng. 5 (2017) 9758–9765.
[31] R. Lamaestre, H. Bernas, Significance of lognormal nanocrystal size distributions,
Phys. Rev. B 73 (2006) 125317.
[32] D.R. Wight, I.D. Blenkinsop, W. Harding, B. Hamilton, Diffusion-limited lifetime
in semiconductors, Phys. Rev. B 23 (1981) 5495–5510.
[33] P. Ajev, S. Miasojedovas, S. Jurnas, A carrier density dependent diffusion coeffi-
cient, recombination rate and diffusion length in MAPbI3 and MAPbBr3 crystals
measured under one- and two-photon excitations, J. Mater. Chem. C 8 (2020)
10290–10301.
[3] X. Hao, F. Zabihi, H. Wang, Q. Zhang, M. Eslamian, Grain engineering by ultra-
sonic substrate vibration post-treatment of wet perovskite films for annealing-
free, high performance, and stable perovskite solar cells, Nanoscale 10 (2018)
8526–8535.
[4] Z. Chu, M. Yang, P. Schulz, D. Wu, X. Ma, E. Seifert, L. Sun, X. Li, K. Zhu, K. Lai,
Impact of grain boundaries on efficiency and stability of organic-inorganic tri-
halide perovskites, Nat. Commun. 8 (2017) 2230.
[5] S. Ranbir, S.R. Suranagi, S.J. Yang, C. Kilwon, Enhancing the power conversion
efficiency of perovskite solar cells via the controlled growth of perovskite na-
nowires, Nano Energy 51 (2018) 192–198.
[6] X. Song, W. Wang, P. Sun, W. Ma, Z.-K. Chen, Additive to regulate the perovskite
crystal film growth in planar heterojunction solar cells, Appl. Phys. Lett. 106
(2015) 033901.
[7] J.W. Lee, H.S. Kim, N.G. Park, Lewis acid-base adduct approach for high efficiency
perovskite solar cells, Acc. Chem. Res. 49 (2016) 311–319.
[8] W. Ke, G. Fang, J. Wang, P. Qin, H. Tao, H. Lei, Q. Liu, X. Dai, X. Zhao, Perovskite
solar cell with an efficient TiO2 compact film, ACS Appl. Mater. Interfaces 6
(2014) 15959–15965.
[9] Z.S. Almutawah, S.C. Watthage, Z. Song, R.H. Ahangharnejhad, K.K. Subedi,
N. Shrestha, A.B. Phillips, Y. Yan, R.J. Ellingson, M.J. Heben, Enhanced grain size
and crystallinity in CH3NH3PbI3 perovskite films by metal additives to the single-
step solution fabrication process, MRS Adv. 3 (2018) 3237–3242.
[10] G. Dawit, I.M. Asuo, B. Daniele, B. Matteo, K. Ibrahima, S.G. Cloutier, R. Federico,
N. Riad, Solvent-antisolvent ambient processed large grain size perovskite thin
films for high-performance solar cells, Sci. Rep. 8 (2018) 12885.
[11] Z.G. Xiao, Q.F. Dong, C. Bi, Y.C. Shao, Y.B. Yuan, J.S. Huang, Solvent annealing of
perovskite-induced crystal growth for photovoltaic-device efficiency enhance-
ment, Adv. Mater. 26 (2014) 6503–6509.
[34] A. Guerrero, G. Garcia-Belmonte, I. Mora-Sero, J. Bisquert, Y.S. Kang,
T.J. Jacobsson, J.P. Correa-Baena, A. Hagfeldt, Properties of contact and bulk
impedances in hybrid lead halide perovskite solar cells including inductive loop
elements, J. Phys. Chem. C 102 (2016) 8023–8032.
[35] A. Pockett, G.E. Eperon, T. Peltola, H.J. Snaith, A. Walker, L.M. Peter, P.J. Cameron,
Characterization of planar lead halide perovskite solar cells by impedance
spectroscopy, open-circuit photovoltage decay, and intensity-modulated pho-
tovoltage/photocurrent spectroscopy, J. Phys. Chem. C 119 (2015) 3456–3465.
[12] A.F. Akbulatov, V.M. Martynenko, L.A. Frolova, N.N. Dremova, I. Zhidkov,
S.A. Tsarev, S.Y. Luchkin, E.Z. Kurmaev, S.M. Aldoshin, K.J. Stevenson, P.A. Troshin,
Intrinsic thermal decomposition pathways of lead halide perovskites APbX3, Sol.
Energy Mater. Sol. Cells 213 (2020) 110559.
[13] E. Calabrò, F. Matteocci, B. Paci, L. Cinà, L. Vesce, J. Barichello, A. Generosi,
A. Reale, A. Di Carlo, Easy strategy to enhance thermal stability of planar PSCs by
perovskite defect passivation and low-temperature carbon-based electrode, ACS
Appl. Mater. Interfaces 12 (2020) 32536–32547.
7