Accepted Manuscript

Constructing Efficient Mixed-ion Perovskite Solar Cells Based on TiO2 Nano-

rod Array

Longkai Yang, Xin Wang, Xianmin Mai, Tan Wang, Chao Wang, Xin Li,
Vignesh Murugadoss, Qian Shao, Subramania Angaiah, Zhanhu Guo

PII: S0021-9797(18)31120-2
DOI: https://doi.org/10.1016/j.jcis.2018.09.045
Reference: YJCIS 24096

To appear in: Journal of Colloid and Interface Science

Received Date: 30 June 2018

Revised Date: 17 August 2018
Accepted Date: 12 September 2018

Please cite this article as: L. Yang, X. Wang, X. Mai, T. Wang, C. Wang, X. Li, V. Murugadoss, Q. Shao, S. Angaiah,
Z. Guo, Constructing Efficient Mixed-ion Perovskite Solar Cells Based on TiO2 Nanorod Array, Journal of Colloid
and Interface Science (2018), doi: https://doi.org/10.1016/j.jcis.2018.09.045

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Constructing Efficient Mixed-ion Perovskite Solar Cells Based on TiO2 Nanorod Array

Longkai Yang,a Xin Wang,a Xianmin Mai, b,* Tan Wang, c Chao Wang,d,e

Xin Li,a,* Vignesh Murugadoss,d,f Qian Shao,g Subramania Angaiah,f,* and Zhanhu Guod,*

a. Pen-Tung Sah Institute of Micro-Nano Science and Technology, Xiamen University,

Xiamen 361005, China

b. School of Urban Planning and Architecture, Southwest Minzu University, Chengdu 610041,

China

c. College of Chemistry and Chemical Engineering, Xiamen University, Xiamen 361005,

China

d. Integrated Composites Laboratory (ICL), Department of Chemical &Biomolecular

Engineering, University of Tennessee, Knoxville, TN 37996

e. College of Materials Science and Engineering, North University of China, Taiyuan 030051,

China

f. Electrochemical Energy Research Lab, Centre for Nanoscience and Technology,

Pondicherry University, Puducherry - 605 014, India.

g. College of Chemical and Environmental Engineering, Shandong University of Science and

Technology, Qingdao 266590, China

* Correspondence author:

Email: lixin01@xmu.edu.cn (Dr. X. Li); maixianmin@foxmail.com (Dr. X. Mai);

zguo10@utk.edu (Dr. Z. Guo); a.subramania@gmail.com (S. Angaiah)

1
 Abstract

Oriented TiO2 nanorod array (TiO2 NA) is very attractive in the fields of halide perovskite solar

cells (PSCs) due to its fewer grain boundaries and high crystallinity for effective charge

collection. The optimization of TiO2 nanostructures has been proved to be an effective approach

for efficient PSCs. On the other hand, tuning the crystallization of perovskite films on top of the

TiO2 NA isvery important for efficient TiO2-NA based PSCs. Herein, scanning electron

microscopy (SEM) and X-ray powder diffraction (XRD) were used to study the crystallization of

different mixed-ion Cs0.1(FA0.83MA0.17)0.9Pb(I0.83Br0.17)3 perovskite (in which MA=CH3NH3+,

and FA=CH(NH2)2+) films, from different perovskite precursor concentrations, on the TiO2

nanorod arrays. A mechanism was proposed to reveal the inherent connection between the

precursor concentration and the crystallite growth of the perovskite film prepared with

anti-solvent quenching process. Meanwhile, both faster charge separation at perovskite/TiO 2 NA

interface and longer charge transport were observed on thicker perovskite film with larger grains,

revealed by the time-resolved method. However, atomic force microscopy (AFM) results

indicated that too thick perovskite film impaired the charge collection owing to the increased

recombination. By balancing the charge collection and film thickness, highly efficient PSCs were

prepared with a champion power conversion efficiency (PCE) of 19.33% with little hysteresis.

The study highlights a great potential of incorporating oriented one-dimensional electron

extraction materials in high-performance PSCs and other applications.

Keywords: TiO2; nanorod array; mixed-ion; perovskite solar cell.

2
1. Introduction

Since the first report of hybrid lead halide perovskite materials in the photoelectrochemical

solar cells in 2009 [1], the rapidly increased power conversion efficiency (PCE) has been made

in the field during the past nine years[2-5]. Many device architectures, preparation methods, and

compositions of hybrid lead halide perovskite materials have been intensively studied to obtain

efficient solar cells. For perovskite solar cells, such rapid development is inseparable from the

foundation laid in the fields of dye sensitized solar cells (DSSCs) and organic solar cells (OSCs),

which can learn from each other and develop together [6-8]. TiO2 is a good example.TiO2 films

as n-type charge selective contact, including compact films [9], mesoporous structured films [5],

and ordered structure films [10], have yielded the highest efficiencies so far. Although the

nanoparticle-structured TiO2 film can collect the photon-induced electrons efficiently for

state-of-the-art perovskite solar cells, the randomly distributed nanoparticles may cause a

scattering of free electrons and reduced electron mobility. Thus, tremendous efforts have been

made to prepare ordered structured TiO2 films, such as nanotubes [11, 12], nanorods [13-15], and

nanoforests [16], to develop a potential electron transport layer with a higher electron mobility

for efficient perovskite solar cells.

Among these oriented structures, TiO2 arrays (NAs) exhibited the greatest potential for

efficient perovskite solar cells (PSCs) to date owing to their low grain boundaries, higher

electron transport mobility, and open structure for the penetration of perovskite crystals [17].

Moreover, TiO2-NA based perovskite solar cells (TiO2-NAPSCs) show a huge advantage on the

photovoltaic performance over other one-dimensional metal-oxide based PSCs [18-20]. A very

recently published work by Liu et al. has presented a small-scale PSC, based on CH3NH3PbI3

(MAPbI3) absorber and TiO2 NAs, with a PCE of 19.02%, which is comparable to that of the

3
traditional mesoporous-TiO2 based perovskite solar cells [21]. However, most of the reported

efficient PSCs using TiO2 NAs are based on MAPbI3 perovskite [17, 21], which has concerns

with respect to thermal stability [22], degradation upon contact with moisture [4], sunlight

stability [23], and the structural phase transition at 55 °C [24]. Researchers found that perovskite

absorbers based on mixed cations and halides (mixed perovskite) are more suitable for

photovoltaic applications thanking to their good photovoltaic performance, thermal and

structural stabilities [25-27]. Therefore, the fabrication of PSCs based on TiO 2 NAs and mixed

perovskite can provide a new approach for efficient and stable PSCs. To date, there are few

reports on the fabrication of efficient PSCs with the combination of TiO2 NAs as electron

transfer layer (ETL) and mixed-ion perovskite absorbers [12-16].

The preparation of high-quality mixed-ion perovskite film on TiO2 NAs by anti-solvent

quenching method is very important to obtain good photovoltaic performance [17]. The

precursor concentration and annealing temperature have been regarded as the key fabrication

parameters [28] and have been widely studied [29, 30] regardless of differences in device

structures and fabrication techniques. Even though the effect of those parameters on the film

microstructure and electronic property has been established by several groups [28, 31], the

detailed crystal nucleation and growth process of perovskite, after anti-solvent quenching and

post-annealing, are still ambiguous. Careful inspection of the perovskite crystal evolution can

provide a simple way to modulate the perovskite growth to obtain high-quality perovskite film

for efficient PSCs.

Herein, we presented an in-depth discussion of preparing highly efficient and reproducible

TiO2-NA PSCs with mixed-ion perovskite. In this work, the precursor-concentration dependent

morphology and crystallization evolution of the corresponding mixed-ion perovskite films were

4
carefully researched and modeled. Moreover, a detailed study has been carried out regarding the

effects of crystallization evolution and film thickness on the electron collection, recombination,

and photovoltaic performance. Combining the benefits of good crystallization of mixed-ion

perovskite on TiO2 NAs and high electron mobility of TiO2 NAs, the PCEs of TiO2-NA PSCs

reached a great extent to a champion PCE of 19.33% with little hysteresis. A better performance

can be expected by further optimizing the NAPSCs in the near future.

2 Experimental

2.1 Material synthesis

Titanium butoxide (97%,TBT), 2-ethyl-butyric acid (99%), formamidine acetate (99%),

methylamine solution (33wt % in methanol), hydroiodic acid (57 wt% in H2O), lead bromide

(PbBr2, 99.999%), cesiumiodide (CsI, 99.9%), chlorobenzene (99.8%) and acetonitrile (99.8%)

were purchased from Sigma-Aldrich.

2,2’,7,7’-tetrakis(N,N-di-p-methoxyphenylamine)-9,9’-spirobifluorene (Spiro-OMeTAD) was

purchased from Lumtec (Taiwan). Lead (II) iodide (99.9985%) was purchased from Alfa Aesar.

All other reagents and chemicals were purchased from Sinopharm Group Co. and used as

received without any further purification.

Formamidinium Iodide (NH=CHNH3I, FAI). Briefly, 5 g formamidine acetate was

dissolved in 60 mL methanol in a round-bottomed flask immersed in an ice bath. Then, 10 mL

hydroiodic acid (57 wt% in water) was dropped slowly into the round-bottomed flask. The

solution was further stirred for 2 hours at room temperature. The obtained light-yellow solution

was then concentrated by rotary evaporation at 50°C to obtain crude solids, which was then

dissolved by a minimum amount of methanol, recrystallized in diethyl ether, and filtered. The

5
dissolved and recrystallized procedures were repeated three times, and the resulting white solid

was collected and dried at 60°C under vacuum for 24 hours.

Methylammonium Bromide (CH3NH3Br, MABr). Typically, 11.32 mL CH3NH2 solution

was added into a round-bottomed flask immersed in an ice bath. Then, 14.67 mL hydrobromide

acid (48 wt% in water) was dropped slowly into the round-bottomed flask. The crystallization

and purification processes were similar to that of FAI.

TiO2 NA. Fluorine-doped tin oxide (FTO) coated glasses (Grene, Qingdao, China) were

cleaned with deionized water, acetone, and isopropanol, respectively. After drying, the FTO

glasses were treated with UV-ozone for 15 min. A TiO2 compact layer of 50 nm thickness was

then prepared on the FTO surface. Briefly, a drop of TBT/deionized water/HCl (37% in water)

mixture (with a volume ratio of 74:1:1:1000) was spun on the FTO surface with 2000 rpm for 60

s, and further annealed at 450 °C for 60 min. Then, the FTO substrate with the compact TiO2

layer was immersed in a Teflon-lined stainless-steel autoclave (25 mL) with 12 mL solution

containing 2-ethyl-butyric acid / HCl/ butanone (with a volume ratio of 2:4:6), and kept at

200 °C for 17 min [17]. Finally, the reactor was taken out from the oven and put into cold water

immediately. After cooling down to room temperature, the TiO2 NA-covered substrate was

immersed in ethanol for 10 min and carefully washed with deionized water to remove the

residues.

2.2 Solar cell fabrication

The PSCs with architecture of FTO/TiO2 compact layer/TiO2NA/mixed-ion

perovskite/Spiro-OMeTAD/silver were prepared as presented in Fig. 1. The typical procedure is

shown as below.

6
 FAI, MABr, PbBr2, and PbI2 were added into a mixed solvent of DMFand DMSO (4:1 volume

ratio) by a molar ratio of 1:0.2:0.2:1.1. Then CsI, pre-dissolved as a 1.5 M stock solution in

DMSO, was added to the mixed DMF/DMSO solution to achieve the perovskite

(Cs0.1(FA0.83MA0.17)0.9Pb(I0.83Br0.17)3) precursor solution. The molar concentrations were fixed as

0.55 M, 0.80 M, 1.05 M, 1.30 M, and 1.55 M, respectively.

The deposition of the perovskite onto TiO2NAs was performed in a nitrogen-filled glovebox

by spin-coating at 1000 rpm for 10 s and then at 6000 rpm for 25 s. 200 μl chlorobenzene was

poured on the spinning substrate at 3s before the end of the procedure at the second spin-coating.

The substrate was then heated at 100°C for 60 minutes on a hot-plate. A hole-transport-layer

solution was prepared by dissolving 72.3 mg Spiro-OMeTAD,

17.5μLLi-bis(trifluoromethanesulfonyl) imide (from a stock solution in 520 mg/mL in

acetonitrile) and 28.8 μLtert-butylpyridine in 1 mL chlorobenzene. The Spiro-OMeTAD solution

was subsequently spin-casted on top of the perovskite layer at 3000 rpm for 30 s. The devices

were finalized by thermal evaporation of silver layers (100 nm thickness) on the top of

Spiro-OMeTAD layers.

Fig. 1. Schematic of aperovskite solar cell based on TiO2 NA.

7
2.3 Characterization

The top-view and cross-section morphologies of the perovskite films were observed using a

field-emission scanning electron microscope (SEM, SUPRA 55, Zeiss, Germany). X-ray

diffraction (XRD) patterns of the perovskite films were recorded using a Rigaku Ultima-IV

X-ray diffractometer. The absorption spectra were obtained by using a Shimadzu UV-2550

UV-vis spectrophotometer (Japan).

The solar cells were measured using a 300 W xenon solar simulator (Newport Oriel Solar

Simulators) with a source meter (Keithley 2420, USA) under 100 mW/cm2, AM 1.5G

illumination, calibrated by a Si-reference cell. The J-V curves were scanned reversely from 1.2 V

to -0.5 V, or forwardly from -0.5 V to 1.2 V. The step voltage was fixed at 100 mV. The active

area for all devices was fixed by using a mask (0.06 cm2). The external quantum efficiency (EQE)

spectra were recorded using a solar cell quantum efficiency measurement system (QEX10, PV

measurements, USA). Prior to the measurement, a standard silicon solar cell was used as the

reference (012-2013, Pharos Technology).

Atomic force microscopy (AFM) images of the perovskite films on TiO 2NAs were tested

using an in-ORCA mode with a bias of 0.6V (Cypher S, Asylum Research, USA).

Space-charge-limited current (SCLC) measurements were used to evaluate the electron mobility

of the mixed-ion perovskite films with a source meter (Keithley 2420, USA). Time-resolved PL

(TRPL) decay spectra of the perovskite/TiO2 nanorods/compact TiO2/FTO devices were

measured by using an Edinburgh FLS 980-STM instrument (Livingston, WL, U.K.).

Electrochemical impedance spectroscopy (EIS) measurements were made in the dark as a

function of applied voltage from -0.75 V to -1.0 V on an electrochemical working station (CHI

660D, China). The FTO substrates were connected to the working electrode.

8
3. Results and Discussion

3.1 Characterization of the perovskite films from different precursor solutions

The quality and morphology of the perovskite film are reported to be a critical issue for

efficient PSCs [32]. To evaluate the effect of perovskite precursor concentrations (0.55 M, 0.80

M, 1.05 M, 1.30 M and 1.55 M) on the film morphology, top-view and cross-section-view

images are given in Fig. 2.

Fig. 2.The representative (a) morphology and (b) cross section morphology of PSCs with

different perovskite precursor concentrations. As shown in (b), the configuration of all the PSCs

is FTO/TiO2 compact layer/TiO2 NA/Perovskite/Spiro-OMeTAD/Ag from down to top. The

scale bar is 200 nm.

As shown in Fig.2, perovskite films composed of an infiltrated layer within the TiO2 NA and

an upper layer (the part over the top of the TiO2 NA) are observed to crystallize on the TiO2 NAs.

The thickness of the upper layer increases with increasing the precursor concentration, ranging

from 20-nm thick film from 0.55 M solution to 390-nm thick film from a 1.55 M solution (Fig.

S1). Certain differences in the perovskite grain can also be observed between the samples from

9
Fig. 2. Higher precursor concentration leads to larger grains. The statistics of the perovskite grain

size derived from the obtained SEM images in Fig. 2 are shown in Fig. S2. The mean grain size

of the perovskite film increases gradually from 88 to 260 nm when the concentration of solution

is increased from 0.55 to 1.55 M. No doubt such difference comes from the precursor

concentration difference. It is quite important to find out how the precursor concentration affects

the crystallization of the perovskite, which is instructive for the preparation of high-quality

perovskite film. Considering the film-preparation process, two stages can be divided as a

quenching stage and the annealing stage. Perovskite crystallites along with intermediate-phase

complex, MA(FA)I(Br)-PbI2-DMSO, are formed at the first stage. At the second annealing stage,

all the intermediate-phase complexes turn into perovskite phase and all the perovskite crystallites

grow up to some extent [33]. Actually, the second stage belongs to the Ostwald ripening process

[34], for which the driving force is the well-known curvature dependence of the atoms’ diffusion

between different-sized grains [35]. To determine which stage is the key process responsible for

the crystallization difference, further characterization is required.

10
Fig. 3.The surface morphology and cross-section morphology of the as-quenched perovskite

films from different precursor concentrations (0.55 M, 0.80 M, 1.05 M, 1.30 M, 1.55 M),

respectively. The scale bar is 200 nm.

Hence, the SEM images of the surface morphology and the cross section morphology of the

five as-quenched films are presented in Fig. 3. The statistics of the grain size derived from the

obtained SEM images in Fig. 3 are shown in Fig. S3. The mean grain size of the five

as-quenched films increases gradually from 75 to 181 nm when the concentration of the solution

is increased from 0.55 to 1.55 M. Compared with the results in Fig. S2, it is assured that different

grain morphologies would come into being at the quenching stage. Meanwhile, a slight increase

of the perovskite grain size can also be observed at the second annealing stage for all samples.

Overall, it can preliminarily concluded that the first quenching stage has much more effect on the

crystallite difference.

X-ray diffraction (XRD) measurements were also performed to investigate the crystal

structure difference of the five as-quenched films, as shown in Fig. 4. All the XRD patterns

contain several intensive peaks, belonging to the FTO, composite perovskite, and an intermediate

product. The XRD pattern of the TiO2 NA/FTO is presented in Fig. S4. The weak peaks, locating

at ~26.8° and ~38.2° in Fig. 4, are attributed to FTO[17], and the feature peaks of TiO2 are too

weak to be observed. Featuring diffraction peaks owing to the trigonal perovskite phase (P3m1)

can be seen from all patterns [36]. Moreover, the peak at ~9.2° belongs to the intermediate

product, MA(FA)I(Br)-PbI2-DMSO complex [33]. It can be concluded that perovskite

crystallites coexist with some intermediate phase after anti-solvent quenching. Herein, the

representative (101) facet was used to compare the perovskite crystal size [28]. The calculated

crystal size was listed in Table S1 based on Scherrer equation. Obviously, the crystal size

11
increases with increasing the precursor concentration, which also implies a faster crystallization

in the low precursor concentration samples. The result agrees well with the SEM result.

Fig. 4. The XRD patterns of the as-quenched perovskite films from different perovskite

precursor concentrations.

According to the SEM and XRD results, it can be inferred that the crystallite difference comes

into being, owing to different solvent extraction speeds from those wet perovskite films, at the

first quenching stage. When the chlorobenzene anti-solvent was poured on the spinning wet

perovskite film, the surface solvent molecules will be extracted immediately. A uniform and

dense thin perovskite layer, named capping layer in this context, will form on the surface

simultaneously, shown in Fig. 5 (a). Then the extraction of solvent molecules under the uniform

layer will be hindered greatly. A thicker wet film gives a slower extraction speed. As illustrated

in Fig. 5 (b) and (c), more solvent molecules will be left within the thicker wet films logically.

Though the thinner film is from lower concentrated perovskite precursor solution, the

concentration of the solution under the capping layer is higher than those of its counterparts. We

12
named this concentration conversion as a result of “capping effect”. Furthermore, more solvent

residues might lead to more intermediate phases, as proven in Fig. 4.

Fig. 5. The schematic of (a) the solvent extraction when chlorobenzene is poured down, the

forming capping layer (1) with perovskite solution (2) beneath on (b) a thinner wet perovskite

film and (c) a thicker perovskite wet film. Solvent molecules exist both within the film and on

the surface; (d) LaMer Model for the nucleation and growth of crystals in mother solution.

With the further evaporation of solvent molecules after quenching, the nucleation process

begins, which follows a LaMer theory, underneath the capping layer [37]. The LaMer model,

shown in Fig. 5 (d), is used widely for the explanation of most crystal nucleation and growth

process from liquid solutions. In LaMer model, when the precursor solution reaches a critical

13
concentration, many nucleuses come into being simultaneously and start to grow (Stage I).

However, with the continuous consumption of the solute for the crystal nucleation and growth,

the solution concentration will decrease to some extent and then the nucleation will be stopped

(Stage II). Finally, there is only the growth of the crystals (Stage III). For the thinner wet

perovskite film, the solution underneath the capping layer has a higher concentration, because of

the capping effect, and reaches the threshold crystallized concentration faster. As a result, large

number of smaller crystallites are produced based on LaMer theory. On the contrary, lager

crystals can be obtained on the thicker films from perovskite precursor solutions with higher

concentrations. Wang has observed similar phenomena [38]. By the way, if none of anti-solvent

quenching is applied, the solvent from the wet film will evaporate freely and lead to a porous and

non-uniform film, as shown in Fig. S5.

Actually, the capping effect is beneficial for the preparation of high-quality perovskite film.

For example, Zou et al. have used a heterogeneous cap or homogeneous cap for a slower

evaporation of solvent molecules within the perovskite films to deliver large-size perovskite

grains [39-41]. According to our analysis and our experience, the quenching time and controlling

the extraction of solvent molecules within the wet perovskite film are extremely important for

high-quality perovskite films with large grain sizes [39].

Post-annealing treatment was used to further remove residue solvents within the quenched

films. During that stage, all the samples go through a crystal growth and coalescence process. In

order to get further insight into the effect of post-annealing treatment, the crystal structures of the

samples were also investigated with XRD, as shown in Fig. 6. It can be seen that there are two

peaks at ~12.7° at the patterns of the perovskite films from concentrations of 0.55 M and 0.80 M.

The two peaks belong to the featuring peak of PbI 2 (001). The previous study on the interfacial

14
structure of perovskite/TiO2 has shown a spectroscopic evidence for the presence of PbI 2 and

methylamine as well when a thin perovskite-layer was coated on the TiO2 surface [42]. TiO2 is

assumed to account for the existence of PbI2 by decomposing the perovskite at the interface. The

role of PbI2 is suggested to be either positive [43] or negative [44] for the photovoltaic

performance, which depends on different conditions. As the precursor concentration is increased,

the featuring peak of PbI2 residues disappear.

Fig. 6. The XRD patterns of the post-annealed perovskite films from different perovskite

precursor concentrations.

More importantly, the peak intensities of (101)/(202)/(303) facets are increased compared with

those of the (110)/ (220) facets. Herein, we selected the representative facets (101) and (110) to

investigate the crystal growth process. The intensity ratio is calculated and shown in Fig. S6. It

can be seen that the intensity ratio of (101)/(110) facet shows a slightly increasing tendency with

the increased precursor concentration for those as-quenched samples, which is likely contrary to

the LaMer theory. The capping effect of the fastest formed perovskite layer might lead to such

result as discussed above. With further post-annealing, the residue solvents can be evaporated

15
from the films. The increased (101) facets of the films from high precursor concentrations

indicate better perovskite phase in cases of the higher precursor concentrations. The previous

study reported that the increased peak intensity of (101) facet would be beneficial for the

electron injection at the interface of perovskite/TiO2 [31], and their photovoltaic performance.

Meanwhile, the crystal size along with the normal direction of (101) facet was calculated with

Scherrer equation again to evaluate the growth of the crystals during the Ostwald ripening

process, as shown in Table S1. It can be seen that the average crystal growth rate of five films

during annealing treatment are basically the same, around 0.14 nm/min, which proves that the

difference of the five films is mainly formed at the quenching stage.

3.2 Photovoltaic performance of the different perovskite films

Since perovskite films with different grain sizes have been obtained, their impact on the

photovoltaic performance was evaluated. The typical device structures can be seen in Fig. 1. The

average photovoltaic parameters with standard deviation are listed in Table 1.

Table 1.Photovoltaic Parameters with standard deviation of Short-Circuit Photocurrent Density

(Jsc), Open-Circuit Voltage (Voc), Fill Factor (FF), and Power Conversion Efficiency (PCE) of

the perovskite solar cells evolved with perovskite precursor concentrations.

Perovskite Voc Jsc FF PCE

concentration (V) (mA/cm2) (%)
(mol/L)
0.55 0.960± 0.021 16.24 ±0.54 0.662± 0.003 10.32± 0.47

0.80 1.011± 0.023 17.91± 0.82 0.665± 0.013 12.04± 0.65

1.05 1.096± 0.032 19.38 ±0.74 0.671± 0.022 14.25 ±1.07

1.30 1.131 ±0.013 23.16± 0.60 0.714± 0.015 18.69± 0.47

16
 1.55 1.123± 0.018 21.85± 0.95 0.706± 0.03 17.32± 0.70

As shown in Table 1, the increased precursor concentration, from 0.55 M to 1.30 M, leads to

the increase of average short-circuit current density (Jsc) from 16.24 ± 0.54 mA/cm2 to 23.16 ±

0.60 mA/cm2. The increase of Jsc is mainly due to the increased light absorption, as shown in Fig.

S7. Further increasing the precursor concentration decreases the average J sc to 21.85± 0.95

mA/cm2, corresponding to 5.7% decrement. The average open-circuit voltage (Voc) shows a

similar tendency to the average short-circuit current density (Jsc), from 0.55 M to 1.30 M ,

enhanced from 0.960 ± 0.021 to 1.131 ± 0.013 V, but slightly declines to 1.123 ± 0.018 V when

the precursor concentration is 1.55 M. The fill factor (FF) also shows a similar change tendency

to those of Jsc and Voc. FF is improved gradually from 0.662 ± 0.003 to 0.714 ± 0.015, from 0.55

M to 1.30 M. Higher precursor concentration (1.55 M) decreases the FF to 0.706 ± 0.03.

Considering the change of Jsc, Voc,and FF, the optimal precursor concentration for the high

photovoltaic performance of these solar cells should be 1.30 M, exhibiting the highest average

power conversion efficiency of 18.69%. For comparison, a device with a non-quenching

perovskite film (1.30 M) was also prepared and gave a poor efficiency of 6.3%, shown in Fig. S8,

which is owing to the low-quality perovskite shown in Fig. S5.

J-V curves of one of the highest-performing devices under different scan directions under one

Sun (AM 1.5) are presented in Fig. 7 (a). In the reverse scan, the device generates a Jsc of 23.53

mA/cm2, a Voc of 1.149 V, and an FF of 0.715, which yields a PCE of 19.33%. To the best of our

knowledge, this is the highest PCE reported, by far, among the one-dimensional array-based

hybrid inorganic-organic solar cells. In addition, the PCE obtained from the forward scan

(18.81%) was close to that from reverse scan indicating a negligible J-V hysteresis for this kind

17
of cells. Meanwhile, we measured EQE to verify our measured Jsc. The EQE is the percentage of

collected electrons, obtained under short circuit current conditions, related to the number of

incident photons and can be used to determine the contribution of the photons of different energy

to the charge generation in the solar cells. The spectral response of the champion solar cell is

shown in Fig. 7 (b), which yields an integrated Jsc of 22.85 mA/cm2.

Furthermore, good stability can be achieved on the champion device. The output of J sc and

PCE at the maximum power point was recorded at a bias voltage of 0.91 V, as shown in Fig.S9.

The Jsc was stabilized at 21.1 mA cm-2 and the PCE was stabilized at 19.2% for about 200 s. The

PCE of the device remained above 96% of the initial value after 1000-hour atmospheric storage

(humidity lower than 10% in electronic box), as shown in Fig. S10. Even though the humidity

level is somewhat low and not comparable to the real environment, the TiO 2 NA based device

shows its great potential in the stability aspect. Compared with its counterpart, nanoparticle

based film, the TiO2 NA structure possesses unique highly-ordered 1D architecture which can

remarkably protect the perovskite from the moisture expected from the hole transport layer side.

With further sealing to reduce the attack of moisture from the hole transport layer side or the

grain-boundary engineering of perovskite films, outstanding stability can be achieved on the

TiO2-NAPSCs.

18
Fig. 7. (a) Current-voltage characteristics on the perovskite solar cells with a starting perovskite

precursor solution of 1.30 M under AM1.5G one sun illumination and (b) spectral response of

the champion solar cell.

3.3 Inspection of charge transport and recombination within the different perovskite films

As discussed above, the combined effect of perovskite precursor concentration on J sc, Voc and

FF resulted in an improved efficient PSCs when the optimal precursor concentration was 1.30 M.

Considering the same TiO2 NA based photoanodes, hole transport material and Ag metal

electrode are applied for all the devices, the improved photovoltaic performance of PSCs might

also come from the modified perovskite/TiO2NA interface and the electron transport within the

perovskite layer, besides the light absorption difference. To explore the effect of precursor

concentration on the perovskite/TiO2 NA interface, time-resolved photoluminescence (TRPL)

decay measurements are used. Transient photoluminescence (PL) profiles for perovskite thin

films evolved with different precursor concentrations on FTO/compact TiO 2/TiO2 NA are

exhibited in Fig. 8.

19
Fig. 8. Time-resolved PL spectra of the perovskite films on FTO/compact TiO 2 layer/TiO2NAs

prepared using different perovskite precursor solutions.

The PL lifetime can be fitted by a bi-exponential decay function:

(1)

where the fast decay lifetime τ1 can be ascribed to the quenching of free carriers from perovskite

crystals to TiO2 NAs and the slow decay process τ 2 is related to the radiative decay in perovskite

polycrystalline film [45]. The detailed data is given in Table S2. Aiming at the inspection of

varied charge transfer between TiO2 NA and perovskite layer, only the change of the fast decay

lifetime is discussed here. When the precursor concentration is as low as 0.55 M, the fastest

decay lifetime τ1, 2.90 ns, is obtained. With the increase of perovskite precursor concentration, τ 1

increases first, 5.31 ns (0.80 M), 8.79 ns (1.05 M) and then decreases gradually to 5.50 ns (1.30

M), and to 2.96 ns (1.55 M), respectively. The overall change of the electron-extraction lifetime

from perovskite film to TiO2 is similar to the trend of grain size change. It might be inferred that

the fewer grain boundaries between perovskite crystals and TiO 2 NAs has a positive effect on the

charge extraction [21]. Also, the decreased bulk traps of larger grains around TiO2 NAs will
20
reduce the trapped electrons[22]. Space-charge-limited current (SCLC) method was performed to

determine the electron mobility of the perovskite films. Devices, with architectures of FTO/TiO 2

NA/perovskite film/PCBM/Au, were fabricated and measured, as shown in Fig. S11. The ohmic

regime and SCLC regime are found at low voltages and at high voltages, respectively [39]. By

using the Mott-Geurst’s SCLC model and a trap-filled model, the electron mobility and the trap

density can be obtained. The increased precursor concentration leads to the increase of electron

mobility from 4.78E- 04 to 4.66E - 03, 1.82 E-02, 2.75 E-02, and 3.03 E-02 cm2 V-1 S-1 for

0.55-M, 0.80-M, 1.05-M, 1.30-M and 1.55-M precursors, respectively. Meanwhile, the trap

density decreases from 5.23 E+16 to 3.69 E+16, 3.40 E+16, 3.21 E+16 and 3.02 E+16 with the

increased precursor concentrations, respectively. Obviously, the increased electron mobility and

the decreased trap density with the increased precursor concentration profit from the increased

grain size.

Though the larger grains have fewer defects and traps, over-large grains might lead to

over-thick film and longer diffusion way of photon-induced electrons. As for the perovskite film

from the 1.55 M precursor solution, the weight fraction of the slow decay process increases to

99%, indicating a dominating mechanism of radiative decay. Hence, the perovskite film prepared

from the 1.30 M precursor solution exhibited the highest charge-dissociation efficiency, with a

high weight fraction of the decreased fast decay lifetime. The observed high average J sc (23.16

mA/cm2) and average FF (0.714) of the devices derived from the 1.30 M precursor solution

might be attributed to this efficient charge separation and collection. To further study the effect

of larger grain size accompanied by the thicker film on the photon-induced electron

transportation and recombination, conductive atomic force microscopy (c-AFM) analyses were

carried out to measure the local current distribution over the perovskite films. To fully evaluate

21
the electron transport through grains and grain boundaries, a bias voltage of 0.5 V was adopted

during the c-AFM measurement process.[46] Perovskite/TiO2 NA/TiO2 compact layer/FTO

devices (1.30 M and 1.55 M) were used in the c-AFM measurement. The topography and

corresponding current mapping are presented in Fig. 9.

Fig. 9.Topographic images and corresponding current images of (a) a perovskite (1.30 M) /TiO 2

NA/c-TiO2/FTO and (b) a perovskite (1.55 M) /TiO2 NA/c-TiO2/FTO typical devices.

Together with the results in Fig. 2, it is found that all the perovskite films (1.30-M, and

1.55-M based) consisting of both large grains and small grains have a relatively rougher surface.

The previous report indicated that, at a high positive bias voltage, the grain boundaries would act

as channels for current to flow rather than recombination sites [47].As can be seen in Fig. 9 (a),

at the positive bias voltage, a positive current flows through the gains and the grain boundaries

uniformly with an average magnitude of 208 pA for the perovskite film from 1.30 M precursor.

In addition, the dark area in Fig. 9 (a) may be ascribed to the impurity phases [47], generated

during the preparation process or the c-AFM measurement in an air atmosphere. In contrast, a

lower current passing through the perovskite film from higher precursor concentration (1.55 M)

with an average magnitude of 175 pA is observed in Fig. 9 (b). It can be deduced that the

lowering current might come from the increasing recombination during the transport process
22
from the top of perovskite film to the FTO substrate. Though there are fewer defects for each

larger perovskite crystal grain in the perovskite film (1.55 M), the film is thicker resulting in a

longer transport way and increased total recombination sites. The crystalline grain size and the

thickness of the perovskite film have a bifacial effect on the performance of the PSCs. These

results further demonstrate the results above.

Electronic impedance spectroscopy measurement was performed to determine the interfacial

charge recombination dynamics of different devices. Fig. 10 (a) presents the typical

corresponding Nyquist plots for those devices at a bias of -0.8 V. Two separated semicircles can

be found in the plots, from which recombination resistance within the perovskite films and their

interfaces can be fitted from the plots. Fig. 10 (b) compares the interfacial charge recombination

resistance (Rre) between perovskite and its contacts in the devices at varied bias, indicating that

Rre for device from 1.3 M is always larger than those of other devices. On the other word, the

device with the perovskite film from 1.30 M precursor solution shows much decreased interfacial

recombination, which agrees well with AFM results and IV results. The suppressed interfacial

charge recombination could be owing to a balance of the larger crystallite size and the film

thickness. This is further confirmed with the transient decay measurements in Fig. 8.

23
Fig.10. (a) Typical Nyquist plots of the devices from different concentrated perovskite precursor

solutions in the dark. The frequency range is from 1 MHz to 0.1 Hz; the bias is -0.8 V. (b)

Interfacial recombination resistance (Rre) extracted from the Nyquist plots of the devices from

different concentrated perovskite precursor solutionsrecorded at different biases in the dark.

4. Conclusions

To sum up, the effect of precursor concentration on the crystallization and photovoltaic

performance of mixed-ion perovskite films for PSCs based on TiO2 NAs has been studied. First,

higher precursor concentration was found to be favorable for the formation of larger perovskite

crystallites via anti-solvent quenching method. A detailed inspection of the as-quenched

perovskite films revealed that the quenching changed the evaporation of the solvent molecules

from the wet perovskite film and was mainly responsible for differed nucleation and growth of

the perovskite crystallites. It is the first time to disclose the relationship between the precursor

concentration and the final perovskite crystal features in the case of anti-solvent quenching.

Second, photovoltaic characterizations showed that thicker perovskite film with larger

crystallites was beneficial for highly-efficient TiO2-NA based PSCs. SCLC measurements and

time-resolved PL measurements indicated that larger crystallites had lower trap densities and

were good for photoinduced electron transport and extraction. However, PL measurements and

EIS characterizations showed that the recombination within the perovskite was increased when

the film thickness was a bit too thick. Finally, perovskite precursor solution with a concentration

of 1.30 M was suitable for highly efficient and hysteresis-free PSCs with a champion PCE of

19.33%. Our results are believed to be very instructive for the fabrication of high quality

perovskite films using anti-solvent quenching method. The key point is to avoid too fast

24
nucleation process at the quenching stage by optimizing the precursor concentration,

spin-coating parameters, and quenching time. Further investigation on the TiO2 NA based

perovskite solar cells could be turned to the practical long-term stability.

Acknowledgements

We gratefully acknowledge the support from the National Natural Science Foundation of China

(U1505243), the funding from Fujian Province of China (2016H0036, JZ160404 and

2015J01064), and the funding from Qinghai Province of China (2017-ZJ-750). Prof. Suyuan Xie

is gratefully acknowledged for his assistance in the J-V characteristic measurements. Mr. MV grateful

to the Indo-US Science and Technology Forum (IUSSTF), Department of Science and Technology

(DST), New Delhi for providing a fellowship under BASE program (IUSSTF BASE Internships

2018/ 12/ Vignesh M, dt. 09/04/2018).

References

[1] A. Kojima, K. Teshima, Y. Shirai, T. Miyasaka, Organometal Halide Perovskites as

Visible-Light Sensitizers for Photovoltaic Cells, J. Am. Chem. Soc. 131(17) (2009)

6050-6051.

[2] P. Gao, M. Gratzel, M.K. Nazeeruddin, Organohalide lead perovskites for photovoltaic

applications, Energy Environ. Sci. 7(8) (2014) 2448-2463.

[3] C. Jin, Y. Huailiang, L. Junpeng, X. Xiaobao, S. Yan, L. Hong, W. Mingkui, Recent progress

in efficient hybrid lead halide perovskite solar cells, Sci. Technol. Adv. Mater. 16(3) (2015)

036004.

25
[4] W. Chen, Y. Wu, Y. Yue, J. Liu, W. Zhang, X. Yang, H. Chen, E. Bi, I. Ashraful, M. Gratzel,

L. Han, Efficient and stable large-area perovskite solar cells with inorganic charge extraction

layers, Science. 350(6263) (2015) 944.

[5] W.S. Yang, B.-W. Park, E.H. Jung, N.J. Jeon, Y.C. Kim, D.U. Lee, S.S. Shin, J. Seo, E.K.

Kim, J.H. Noh, S.I. Seok, Iodide management in formamidinium-lead-halide-based

perovskite layers for efficient solar cells, Science. 356(6345) (2017) 1376.

[6] N. G. Park, Perovskite solar cells: an emerging photovoltaic technology[J]. Mater. Today,

2015, 18(2):65-72.

[7] N. S .Marinova, Valero, J. L. Delgado, Organic and perovskite solar cells: Working

principles, materials and interfaces.J. Colloid Interface Sci,488(2017):373.

[8] P. A.Bouit, F.Spänig , G.Kuzmanich, E. Krokos, C.Oelsner, M. A.Garcia-Garibay,

J.L.Delgado, N. Martin, D. M. Guldi Efficient utilization of higher-lying excited states to

trigger charge-transfer events, Chem.-A Eur. J.,16(31)(2010). 9638-9645.

[9] H. Tan, A. Jain, O. Voznyy, X. Lan, F.P. Garcia de Arquer, J.Z. Fan, R. Quintero-Bermudez,

M. Yuan, B. Zhang, Y. Zhao, F. Fan, P. Li, L.N. Quan, Y. Zhao, Z.-H. Lu, Z. Yang, S.

Hoogland, E.H. Sargent, Efficient and stable solution-processed planar perovskite solar cells

via contact passivation, Science. 355(6326) (2017) 722.

[10] H.-S. Kim, J.-W. Lee, N. Yantara, P.P. Boix, S.A. Kulkarni, S. Mhaisalkar, M. Gratzel,

N.-G. Park, High Efficiency Solid-State Sensitized Solar Cell-Based on Submicrometer

Rutile TiO2 Nanorod and CH3NH3PbI3 Perovskite Sensitizer, Nano Lett. 13(6) (2013)

2412-2417.

26
[11] P. Qin, M. Paulose, M.I. Dar, T. Moehl, N. Arora, P. Gao, K. Varghese Oomman, M.

Gratzel, K. Nazeeruddin Mohammad, Stable and Efficient Perovskite Solar Cells Based on

Titania Nanotube Arrays, Small 11(41) (2015) 5533-5539.

[12] H. Aibin, Z. Jingting, Z. Jianyun, Y. Yu, L. Yan, Y. Songwang, B. Shanhu, L. Lei, J. Ping,

Mesostructured perovskite solar cells based on highly ordered TiO 2 network scaffold via

anodization of Ti thin film, Nanotechnology 28(5) (2017) 055403.

[13] A. Fakharuddin, F. Di Giacomo, A.L. Palma, F. Matteocci, I. Ahmed, S. Razza, A.

D'Epifanio, S. Licoccia, J. Ismail, A. Di Carlo, T.M. Brown, R. Jose, Vertical TiO 2 Nanorods

as a Medium for Stable and High-Efficiency Perovskite Solar Modules, ACS Nano 9(8)

(2015) 8420-8429.

[14] Y.-N. Zhang, B. Li, L.-Y. Zhang, L.-W. Yin, Efficient electron transfer layer based on

Al2O3 passivated TiO2 nanorod arrays for high performance evaporation-route deposited

FAPbI3 perovskite solar cells, Sol. Energy Mater. Sol. Cells 170 (2017) 187-196.

[15] S.S. Mali, C.A. Betty, P.S. Patil, C.K. Hong, Synthesis of a nanostructured rutile TiO 2

electron transporting layer via an etching process for efficient perovskite solar cells: impact

of the structural and crystalline properties of TiO2, J. Mater. Chem. A 5(24) (2017)

12340-12353.

[16] M. Salado, M. Oliva-Ramirez, S. Kazim, A.n.R. Gonzez-Elipe, S. Ahmad, 1-dimensional

TiO2 nano-forests as photoanodes for efficient and stable perovskite solar cells fabrication,

Nano Energy 35 (2017) 215-222.

[17] X. Li, S.-M. Dai, P. Zhu, L.-L. Deng, S.-Y. Xie, Q. Cui, H. Chen, N. Wang, H. Lin,

Efficient Perovskite Solar Cells Depending on TiO2 Nanorod Arrays, ACS Appl. Mater.

Interfaces 8(33) (2016) 21358-21365.

27
[18] C. Gao, S. Yuan, B. Cao, J. Yu, SnO2 nanotube arrays grown via an in situ template-etching

strategy for effective and stable perovskite solar cells, Chem. Eng. J. 325 (2017) 378-385.

[19] Y.-Z. Zheng, E.-F. Zhao, F.-L. Meng, X.-S. Lai, X.-M. Dong, J.-J. Wu, X. Tao,

Iodine-doped ZnO nanopillar arrays for perovskite solar cells with high efficiency up to

18.24%, J. Mater. Chem. A 5(24) (2017) 12416-12425.

[20] S. Zihang, T. Guoqing, L. Huan, L. Guopeng, M. Shuai, Y. Shimeng, L. Qian, J. Yang,

Three-dimensional architecture hybrid perovskite solar cells using CdS nanorod arrays as an

electron transport layer, Nanotechnology 29(2) (2018) 025401.

[21] C. Liu, R. Zhu, A. Ng, Z. Ren, S.H. Cheung, L. Du, S.K. So, J.A. Zapien, A.B. Djurisic, D.

Lee Phillips, C. Surya, Investigation of high performance TiO 2 nanorod array perovskite

solar cells, J. Mater. Chem. A 5(30) (2017) 15970-15980.

[22] B. Conings, J. Drijkoningen, N. Gauquelin, A. Babayigit, J. D'Haen, L. D'Olieslaeger, A.

Ethirajan, J. Verbeeck, J. Manca, E. Mosconi, D. Angelis Filippo, H.G. Boyen, Intrinsic

Thermal Instability of Methylammonium Lead Trihalide Perovskite, Adv. Energy Mater.

5(15) (2015) 1500477.

[23] R.K. Misra, S. Aharon, B. Li, D. Mogilyansky, I. Visoly-Fisher, L. Etgar, E.A. Katz,

Temperature- and Component-Dependent Degradation of Perovskite Photovoltaic Materials

under Concentrated Sunlight, J. .Phys. Chem. Lett. 6(3) (2015) 326-330.

[24] C.C. Stoumpos, C.D. Malliakas, M.G. Kanatzidis, Semiconducting Tin and Lead Iodide

Perovskites with Organic Cations: Phase Transitions, High Mobilities, and Near-Infrared

Photoluminescent Properties, Inorg. Chem. 52(15) (2013) 9019-9038.

28
[25] T. Jesper Jacobsson, J.-P. Correa-Baena, M. Pazoki, M. Saliba, K. Schenk, M. Gratzel, A.

Hagfeldt, Exploration of the compositional space for mixed lead halogen perovskites for high

efficiency solar cells, Energy Environ. Sci. 9(5) (2016) 1706-1724.

[26] M. Saliba, T. Matsui, J.-Y. Seo, K. Domanski, J.-P. Correa-Baena, M.K. Nazeeruddin, S.M.

Zakeeruddin, W. Tress, A. Abate, A. Hagfeldt, M. Gratzel, Cesium-containing triple cation

perovskite solar cells: improved stability, reproducibility and high efficiency, Energy

Environ. Sci. 9(6) (2016) 1989-1997.

[27] M. Saliba, T. Matsui, K. Domanski, J.-Y. Seo, A. Ummadisingu, S.M. Zakeeruddin, J.-P.

Correa-Baena, W.R. Tress, A. Abate, A. Hagfeldt, M. Gratzel, Incorporation of rubidium

cations into perovskite solar cells improves photovoltaic performance, Science. 354(6309)

(2016) 206.

[28] D. Bi, J. Luo, F. Zhang, A. Magrez, N. Athanasopoulou Evangelia, A. Hagfeldt, M. Gratzel,

Morphology Engineering: A Route to Highly Reproducible and High Efficiency Perovskite

Solar Cells, ChemSusChem 10(7) (2016) 1624-1630.

[29] D. Bi, A.M. El-Zohry, A. Hagfeldt, G. Boschloo, Unraveling the Effect of PbI 2

Concentration on Charge Recombination Kinetics in Perovskite Solar Cells, ACS Photonics

2(5) (2015) 589-594.

[30] A. Dualeh, N. Titreault, T. Moehl, P. Gao, K. Nazeeruddin Mohammad, M. Gratzel, Effect

of Annealing Temperature on Film Morphology of Organic-Inorganic Hybrid Pervoskite

Solid-State Solar Cells, Adv. Funct. Mater. 24(21) (2014) 3250-3258.

[31] B. Zhang, M.-J. Zhang, S.-P. Pang, C.-S. Huang, Z.-M. Zhou, D. Wang, N. Wang, G.-L. Cui,

Carrier Transport in CH3NH3PbI3 Films with Different Thickness for Perovskite Solar Cells,

Adv. Mater. Interfaces 3(17) (2016) 1600327.

29
[32] E. Eperon Giles, M. Burlakov Victor, P. Docampo, A. Goriely, J. Snaith Henry,

Morphological Control for High Performance, Solution-Processed Planar Heterojunction

Perovskite Solar Cells, Adv. Funct. Mater. 24(1) (2013) 151-157.

[33] N.J. Jeon, J.H. Noh, Y.C. Kim, W.S. Yang, S. Ryu, S.I. Seok, Solvent engineering for

high-performance inorganic-organic hybrid perovskite solar cells, Nat. Mater. 13 (2014) 897.

[34] X. Li, D. Bi, C. Yi, J.-D. Decoppet, J. Luo, S.M. Zakeeruddin, A. Hagfeldt, M. Gratzel, A

vacuum flash-assisted solution process for high-efficiency large-area perovskite solar cells,

Science. 353(6294) (2016) 58.

[35] P.W. Voorhees, The theory of Ostwald ripening, J. Stat. Phys. 38(1) (1985) 231-252.

[36] T.M. Koh, K. Fu, Y. Fang, S. Chen, T.C. Sum, N. Mathews, S.G. Mhaisalkar, P.P. Boix, T.

Baikie, Formamidinium-Containing Metal-Halide: An Alternative Material for Near-IR

Absorption Perovskite Solar Cells, J. Phys. Chem. C 118(30) (2014) 16458-16462.

[37] V.K. LaMer, R.H. Dinegar, Theory, Production and Mechanism of Formation of

Monodispersed Hydrosols, J. Am. Chem. Soc. 72(11) (1950) 4847-4854.

[38] J. Chen, J. Song, F. Huang, H. Li, S. Liu, M. Wang, Y. Shen, The Role of Synthesis

Parameters on Crystallization and Grain Size in Hybrid Halide Perovskite Solar Cells, J.

Phys. Chem. C 121(32) (2017) 17053-17061.

[39] W. Zhu, L. Kang, T. Yu, B. Lv, Y. Wang, X. Chen, X. Wang, Y. Zhou, Z. Zou, Facile

Face-Down Annealing Triggered Remarkable Texture Development in CH 3NH3PbI3 Films

for High-Performance Perovskite Solar Cells, ACS Appl. Mater. Interfaces 9(7) (2017)

6104-6113.

30
[40] Y. Wang, J. Li, Q. Li, W. Zhu, T. Yu, X. Chen, L.a. Yin, Y. Zhou, X. Wang, Z. Zou, PbI 2

heterogeneous-cap-induced crystallization for an efficient CH3NH3PbI3 layer in perovskite

solar cells, Chem. Commun. 53(36) (2017) 5032-5035.

[41] W. Zhu, C. Bao, B. Lv, F. Li, Y. Yi, Y. Wang, J. Yang, X. Wang, T. Yu, Z. Zou,

Dramatically promoted crystallization control of organolead triiodide perovskite film by a

homogeneous cap for high efficiency planar-heterojunction solar cells, J. Mater. Chem. A

4(32) (2016) 12535-12542.

[42] Y.K. Kim, B. Jeon, H.J. Park, Interfacial Electronic Structure of Methylammonium Lead

Iodide Grown on a Mesoporous TiO2 Layer on F-Doped Tin Oxide Substrate, J. Phys. Chem.

C 120(39) (2016) 22460-22465.

[43] Q. Chen, H. Zhou, T.-B. Song, S. Luo, Z. Hong, H.-S. Duan, L. Dou, Y. Liu, Y. Yang,

Controllable Self-Induced Passivation of Hybrid Lead Iodide Perovskites toward High

Performance Solar Cells, Nano Lett. 14(7) (2014) 4158-4163.

[44] H. Lee Yong, J. Luo, R. Humphry-Baker, P. Gao, M. Gratzel, K. Nazeeruddin Mohammad,

Unraveling the Reasons for Efficiency Loss in Perovskite Solar Cells, Adv. Funct. Mater.

25(25) (2015) 3925-3933.

[45] X. Wang, L.-L. Deng, L.-Y. Wang, S.-M. Dai, Z. Xing, X.-X. Zhan, X.-Z. Lu, S.-Y. Xie,

R.-B. Huang, L.-S. Zheng, Cerium oxide standing out as an electron transport layer for

efficient and stable perovskite solar cells processed at low temperature, J. Mater. Chem. A

5(4) (2017) 1706-1712.

[46] J.-J. Li, J.-Y. Ma, J.-S. Hu, D. Wang, L.-J. Wan, Influence of N,N-Dimethylformamide

Annealing on the Local Electrical Properties of Organometal Halide Perovskite Solar Cells:

31
 an Atomic Force Microscopy Investigation, ACS Appl. Mater. Interfaces 8(39) (2016)

26002-26007.

[47] J.-J. Li, J.-Y. Ma, Q.-Q. Ge, J.-S. Hu, D. Wang, L.-J. Wan, Microscopic Investigation of

Grain Boundaries in Organolead Halide Perovskite Solar Cells, ACS Appl. Mater. Interfaces

7(51) (2015) 28518-28523.

32
Graphical abstract

Graphical Abstract

33