Highly Efficient Perovskite Solar Cells Based on Zn2Ti3O8 Nanoparticles as Electron Transport Material

Article abstract of DOI:10.1002/cssc.201701779

Developing ternary metal oxides as electron transport layers (ETLs) for perovskite solar cells is a great challenge in the field of third-generation photovoltaics. In this study, a highly mesoporous Zn₂Ti₃O₈ (m-ZTO) scaffold is synthesized by ion-exchange method and used as ETL for the fabrication of methyl ammonium lead halide (CH₃NH₃PbI₃) perovskite solar cells. The optimized devices exhibit 17.21 % power conversion efficiency (PCE) with an open circuit voltage (V_oc) of 1.02 V, short-circuit current density (J_sc) of 21.97 mA cm^?2 and fill factor (FF) of 0.77 under AM 1.5G sunlight (100 mW cm^?2). The PCE is significantly higher than that based on mesoporous ST01 (m-ST01; 10 nm TiO₂ powder) layer (η=14.93 %), which is ascribed to the deeper conductive band of ZTO nanoparticles, better light absorption and smaller charge recombination. The devices stored for 100 days at ambient temperature with humidity of 10 % showed excellent stability with only 12 % reduction of the PCE. The charge transmission kinetic and long-term stability parameters of the ZTO-based perovskite film growth are discussed as well.

Full text of DOI:10.1002/cssc.201701779

Accepted Article
Title: Highly Efficient Perovskite Solar Cells Based on Zn2Ti3O8
Nanoparticles as Electron Transporting Material with Efficiency
over 17%
Authors: Zhong-Ning Chen, Aiying Pang, Deli Shen, and Mingdeng
Wei
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemSusChem 10.1002/cssc.201701779
Link to VoR: http://dx.doi.org/10.1002/cssc.201701779
A Journal of
www.chemsuschem.org

10.1002/cssc.201701779
ChemSusChem
FULL PAPER
Highly Efficient Perovskite Solar Cells Based on Zn2Ti3O8
Nanoparticles as Electron Transporting Material with Efficiency
over 17%
Aiying Pang,[a] Deli Shen,[b] Mingdeng Wei,*[b] and Zhong-Ning Chen*[a]
Abstract: Developing ternary metal oxides as electron transporting
layers (ETL) for perovskite solar cell is highly challenging in the field
of third-generation photovoltaics. In this study, highly mesoporous
Zn2Ti3O8 (m-ZTO) scaffold synthesized by ion-exchange method has
been successfully used as ETL for the fabrication of methyl
ammonium lead halide (CH3NH3PbI3) perovskite solar cells. The
optimized devices exhibit 17.21% power conversion efficiency (PCE)
with open circuit voltage (Voc) of 1.02 V, short-circuit current density
(Jsc) of 21.97 mA cm-2 and fill factor (FF) of 0.77 under AM 1.5 G
sunlight (100 mW cm−2). The PCE is obviously higher than that
based on mesoporous ST01 (m-ST01) layer (η = 14.93%), which is
ascribed to deeper conductive band of ZTO nanoparticles, better
light absorption and smaller charge recombination. The devices
stored for 100 days at ambient temperature with humidity of 10%
show excellence stability with only 12% reduction of the PCE. The
charge transmission kinetic and long-term stability parameters of the
ZTO-based perovskite film growth are discussed as well.
alkaline components.[5] Other binary oxides such as SnO2,[6]
[8]
WOx[7] and In2O3 have been employed as alternative ETLs in
PSCs, but could not keep the long-term stability of PSCs.
Compared with binary oxides, ternary oxides exhibited specific
functions and highly adjustable chemical and physical properties
by changing the compositions. SrTiO3 and TiO2 have similar
wide band gaps, but the conduction band edge of the former is
slightly higher and matches better with the band structure of
CH3NH3PbI3−xClx, thus boosting Voc by ∼120 mV and improving
distinctly FF relative to mesoporous TiO2 counterparts.[9]
Zn2SnO4 ETL has a negligible effect on the transmittance of the
incident light regardless of its thickness so that the thickest
compact layer blocks most efficiently the back-electron
transfer.[10] Noticeably, the PSCs fabricated with lanthanum
doped BaSnO3 show a steady-state PCE of 21.2% and retain 93%
of initial performance over 1000 hours of full-Sun illumination,
which is ascibed to its high electron mobility and reasonable
electronic structure.[11] As a result, there is an urgent need to
explore more wide band gap ternary oxides with suitable band
alignments and crystal structures to improve device performance
and particularly long-term stability.
Introduction
In ternary oxides study, TiO2-ZnO systems have been explored
as a new compact layers for the ETL.[12] When TiZnO12 (Ti : Zn :
O = 1 : 1 : 12) was applied as ETL for the device, a PCE of
15.10% was achieved, which is comparable to that of the
devices based on conventional TiO2 ETL. It is known that
thermally stable Zn2Ti3O8 (ZTO) as a ternary oxide with cubic
spinel type structure shows specific photocatalysis, conductivity
and dielectric properties.[13] In this work, we report the first use of
m-ZTO as electron transport layer in a regular mesoporous
structure of CH3NH3PbI3 solar cells. The corresponding devices
achieve PCE up to 17.21% with almost no hysteresis. Given the
ZTO layer's deep conduction band and high electron mobility,
the elimination of the hysteresis is ascribable to the enhanced
electron extraction caused by the reduction of energy barrier at
the perovskite/ETL interface. The photoluminescence (PL)
quenching and transient absorption spectra show efficient
photoelectron transfer from CH3NH3PbI3 to m-ZTO layer.
Furthermore, the compact layer also blocks the electrons
transporting reversely from the conducting substrate to
recombine with the holes in the light-harvesting layer. By
introducing such a layer into perovskite solar cell, the device
efficiency is dramatically enhanced.[14]
Perovskite solar cells (PSCs) have recently attracted great
attention due to the rapid growth of power conversion efficiency
(PCE) from 3.8 % to 22.1 %.[1] As is well known, the electron
transport layer (ETL) of PSCs plays an important role in
extracting electrons from perovskite as well as blocking direct
contact between transparent conducting oxide substrate and
perovskite.[2] So far, the PSCs with PCE exceeding 20% have
mainly been reported with mesoporous titanium dioxide (m-TiO2)
as an electron-transporting layer (ETL).[3] Nevertheless, TiO2-
based solar cells suffer from stability reduction under ultraviolet
light, which is attributed to the changes in the compact TiO2
and/or mesoporous TiO2 layers.[4] On the other hand, ZnO as
one of the most promising alternatives with a high PCE of 15.7 %
was chemically unstable and prone to react with acid and
[a]
Dr. A. Y. Pang, Prof. Dr. Z. N. Chen
State Key Laboratory of Structural Chemistry, Fujian Institute of
Research on the Structure of Matter, Chinese Academy of Sciences,
Fuzhou, Fujian 350002, China
E-mail: czn@fjirsm.ac.cn
[b]
Dr. D. L. Shen, Prof. Dr. M. D. Wei
Key Laboratory of Photocatalysis on Energy and Environmental,
Fuzhou University, Fuzhou, Fujian 350002, China.
E-mail: wei-mingdeng@fzu.edu.cn
Supporting information for this article is given via a link at the end of
the document.
Results and Discussion
The X-ray diffraction (XRD) patterns in Figure 1a confirm that
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701779
ChemSusChem
FULL PAPER
ZTO synthesized by a ion-exchange method displays a cubic
structural with a parameter constant a = 8.392 Å (JCPDS 01-
087-1781).[13c] As shown in Figure 1b, the TEM images show
clearly that ZTO is composed of ca. 2545 nm size particles.
The high-resolution TEM image (HRTEM) in Figure 1c shows
that the nanoparticles are highly crystalline. The lattice fringes
are around 0.46 and 0.59 nm, respectively, corresponding to
d(111) and d(110)-spacing in the XRD patterns. Moreover, the
ring-like selected area electron diffraction (SAED; the inset in
Figure 1d) pattern indicates that the as-prepared ZTO
nanoparticles exhibit polycrystalline nature. The stoichiometry of
ZTO has also been confirmed by TEM-mapping measurement
(Figure S1) and XPS (Figure S2).
nm for the CH3NH3PbI3 layer combined with mesoporous m-ZTO,
ca.180 nm for the hole-transporting layer, and 80 nm for the
back contact of thermally evaporated Au layer. It is also clearly
shown that the intimate contact was well achieved between
CH3NH3PbI3 and adjacent layers in our devices.
For the purpose of comparison, perovskite film on the m-ST01
substrate was also investigated as shown in Figure S3. Both
ETLs of m-ZTO and m-ST01 had almost no effect on the
perovskite morphology in SEM, but a higher peak (12.65°) in
XRD, a higher roughness (Rq = 12.7 nm) in AFM and a thinner
thickness (~310 nm) of CH3NH3PbI3 layer were observed indeed.
This is attributed to a coarser surface, a worse wettability and a
lower infiltrating of m-ST01. As indicated in in Figure S4, m-ZTO
film (Rq = 27.8 nm) shows a more uniform and dense surface
than that of m-ST01 film (Rq = 36.7 nm) which has an irregular
surface with some obvious valleys. The much smaller contact
angle of DMSO droplet on m-ZTO film (3.4o) than that on m-
ST01 (28.3o) owing to a good wettability of the former (Figure S4)
is beneficial for perovskite solution to infiltrate into m-ZTO and
crystallize into a stronger and thicker light-absorbing layer.
(a)
(b)
10
20
30
40
50
60
2 Theta / deg
(c)
(d)
Au
Spiro-MeOTAD
Figure 1. XRD pattern (a), TEM image (b), HRTEM image (c) and SAED
Perovskite / m-ZTO
c-TO
image (d) of ZTO nanoparticles obtained at calcined temperatures 600 oC.
FTO glass
We fabricated CH3NH3PbI3 films on FTO/c-TiO2 (abbreviated as
c-TO)/m-ZTO substrate through a one-step procedure according
to the literature.[15] The film quality of the perovskite is shown in
Figure 2. The corresponding XRD patterns of the as-prepared
CH3NH3PbI3 films based on m-ZTO substrate are depicted in
Figure 2a. A set of diffraction peaks occurring at 14.08, 28.41,
31.85, and 43.19° arise from an orthorhombic crystal structure of
CH3NH3PbI3 crystals, which are assigned to (110), (220), (310)
and (330) crystal faces, respectively.[16] The absence of the peak
of 12.65° is indicative of a complete consumption of PbI2 in the
perovskite films based on m-ZTO.[17] The as-formed perovskite
films on m-ZTO substrates (Figure 2b) display the characteristics
of uniformity, free pinhole porosity and low grain boundary with
densely microscale packed grains, thus implying its promising
advantages for PSC devices. The surface roughness of the
perovskite film on m-ZTO is very small (Rq = 8.91 nm), as
revealed by AFM measurement in Fig 2c. Furthermore, the
typical cross-sectional SEM image (Figure 2d) of the device
displays its mesoporous architecture. The thickness of each
layer is estimated to be ca. 70 nm for the compact c-TO, ca. 350
Rq=8.91 nm
Figure 2. Perovskite film on the FTO/c-TO/m-ZTO substrate. (a) XRD pattern.
(b) Top-view SEM image, scale bar 1 μm. (c) 3D AFM topography image
(1.0×1.0 μm). (d) The cross-sectional SEM image of the device.
The current density−voltage (J–V) curves and incident photon-
to-current efficiency (IPCE) spectra are shown in Figure 3 for the
champion mesoporous PSCs based on m-ZTO and m-ST01
ETLs, and its photovoltaic characteristic are summarized in
Table 1. Moreover, to optimize the photovoltaic performance of
PSCs based on ZTO ETLs, the thickness of the m-ZTO film was
optimized through repeatedly spin-coating process (Figure S5
and Table S1). To obtain superior efficiency, the optimum
thickness for the m-ZTO ETL prepared by spin-coating for two
times was ca. 200 nm. Although the devices based on m-ZTO
and m-ST01 exhibited comparable Voc values of 1.02 and 0.99 V,
respectively, the corresponding Jsc and FF values were different
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701779
ChemSusChem
FULL PAPER
as shown in Figure 3a. The cells based on m-ZTO (Jsc = 21.97
mA cm-2 and PCE = 17.21%) exhibited distinctly higher Jsc and
PCE than those of m-ST01-based cells (Jsc = 20.48 mA cm-2 and
PCE = 14.93%). The large efficiency enhancement of m-ZTO
based PSCs compared with that of m-ST01 based devices is
likely attributed to the higher light absorption and the increased
charge carrier concentration for the former. The low dark current
(J-V in Figure 3a and J-V in log scales in Figure S6) arises most
likely from the high-quality films, which may induce to decrease
recombination when carriers are separated in the light absorbing
layer. The improvement of FF is mostly attributable to the
decrease of leakage currents and the less series resistance of
the m-ZTO based CH3NH3PbI3 films with large grains and few
boundaries over a range of length scale,[18] which is consistent
with the following resistance measurements.
PSCs achieved a larger current density (21.36 mA/cm2) than
that of m-ST01 (19.86 mA/cm2) based devices under the AM
1.5G solar photon flux. Relative to Jsc from J−V measurement,
the total photocurrent collected from IPCE spectrum is slightly
lower, due probably to the surface charge traps of ETLs.10
A
contributing factor to the effective PCE is the high electron
injection power between the absorber film and m-ZTO. More
thorough and objective study is under way to get more insight to
the relationship between the film property and the device
performance.
-3.95
(b)
(a)
-4.06
-4.17
Spiro-
OMeTAD
-4.5
FTO
Gold
-5.1
c-TiO2
FTO
25
m-Zn2Ti3O8
(a)
-5.2
CH3NH3PbI3
15
-5.55
m-ZTO
m-ST01
Spiro-OMeTAD
5
Au
-7.53
-7.73
-5
m-ZTO dark
m-ST01 dark
Figure 4. Schematic diagrams of the device structure (a) and the energy level
of the mesoporous perovskite solar cell (b).
-15
0.0
0.2
0.4
0.6
0.8
1.0
Voltage / V
100
25
20
15
10
5
(b)
The PSC architecture is FTO/c-TO/m-ZTO/CH3NH3PbI3/Spiro-
OMeTAD/Au as schematically illustrated in Figure 4a. To obtain
a clear understanding of the electron transfer pathway through
the ETLs, an energy band diagram was drawn based on the
appropriate band levels (Figure S7 and Table S2). Figure 4
demonstrates that both the conduction band minima (CBM) and
valence band maxima (VBM) are shifted in a downward direction
upon using ternary oxide ZTO with respect to c-TO. This
downward shift in the CBM position correspond well with the
perovskite energy band level, which facilitates electron injection
and thus improves Jsc from 20.48 (m-ST01) to 21.97 (m-ZTO)
mA cm-2.[18] Just like semiconductor of binary oxide, ZTO exists
sites in the tail of the density states to move the quasi–Fermi
level in the material nearly to the bottom of conduction band
under illumination.[20] The higher Voc and FF indicate that fewer
surface and sub–band gap states are exhibited in the
mesoporous ZTO than those in ST01. According to a recent
study, the downward shift of ZTO VBM position results in hole
blocking and less recombination in the interface at the m-
ZTO/electrode interface.[21] The TO layer between ZTO and FTO
exhibits a higher CBM than that of both ZTO and FTO, which
benefits to retard charge-carrier recombination at the interfaces.
It is speculated that higher CBM of c-TO does not hinder the
flow of charges because the energy level of the lowest
unoccupied molecular orbital of the CH3NH3PbI3 layer is higher
than the CBM of ZTO and c-TO. Thus, efficient charge
extraction is valid in our cells as that found in other similar
structure devices.[22]
80
60
40
20
0
m-ZTO
m-ST01
0
800
300
400
500
600
700
Wavelength / nm
Figure 3 (a) J−V curves and (b) IPCE spectra and the integrated photocurrent
of the solar cells based on the as-prepared perovskite films under AM 1.5G
illumination.
Table 1. Summary of photovoltaic performances metrics of the champion
devices fabricated on m-ZTO and m-ST01.
Sample
Jsc/mA cm-2
21.97
Voc / V
1.02
FF / %
76.57
73.69
η / %
17.21
14.93
m-ZTOmax
m-ST01max
20.48
0.99
Figure 3b shows the IPCE spectra of the best performance
devices. Photocurrent generation starts at 780 nm in agreement
with the band gap of the CH3NH3PbI3 and achieves peak values
of ∼90% in the 380680 nm spectral region.[19] By integrating the
density photocurrents of the IPCE spectrum, m-ZTO based
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701779
ChemSusChem
FULL PAPER
50
(Table 1), respectively. Finally, the PCE is enhanced from 14.93%
to 17.21% when ZTO is substituted for commercially purchased
ST01.
Rs
Rco
Cco
Rrec
(a)
40
30
20
10
CPEμ
Table 2. Deconvoluted resistances from impedance analyses for m-ZTO and
m-ST01 ETLs.
m-ZTO
m-ST01
ETL
Rs / Ω·cm2
Rco / Ω·cm2
Rrec / Ω·cm2
0
0
30
60
90
120
150
m-ZTO
2.5
4.2
92.8
87.8
486.7
336.2
Z' / 
15
(b)
12
m-ST01
m-ZTO
m-ST01
RS= 2.13 
9
6
3
A = 2.61
(a)
Rs = 1.16 
A = 1.53
0
0.00
0.05
0.10
0.15
0.20
m-ZTO
m-ST01
(Jsc-J)-1 / mA-1 cm2
Figure 5. (a) EIS and (b) plots of dV/dJ vs (JSC - J)-1 derived from I–V curves
(Figure 3a) of m-ST01 (black) and m-ZTO (red) based cells.
650
104
700
750
800
850
Wavelength / nm
m-ZTO
m-ST01
(b)
Electrochemical impedance spectroscopy (EIS) is a simple but
useful method to investigate the charge recombination
resistance and interfacial properties of perovskite solar cells. As
shown in Figure 5, Nyquist plots of the devices under dark at an
applied bias voltage of 0.7 V with the frequency range of 0.1 to
100 kHz display two distinct semicircles. The Nyquist plots were
fitted using an equivalent circuit model as illustrated in the inset
of Figure 5. The curve fitting data of both the ST01 and ZTO
ETL-based devices are summarized in Table 2. The series of
resistance (RS) is associated with the wiring and FTO substrate
located in the high-frequency region. The Rco and Cco refer to the
interface resistance of the counter electrode and the
corresponding capacitance, respectively. In the Nyquist plots,
103
102
0
50
100
150
200
Time (ns)
Figure 6. The steady-state PL spectra (a) and time-resolved PL (TRPL) decay
curves (b) of CH3NH3PbI3 deposited on the top of m-ZTO and m-ST01.
the main arc is responsible for
a combination of the
To investigate the charge injection/separation behavior, we
measured steady-state photoluminescence (PL) spectra and
time-resolved PL (TRPL) decay curves of FTO/c-TO/m-
ZTO/CH3NH3PbI3 and FTO/c-TO/m-ST01/CH3NH3PbI3 films.
Figure 6a shows that the spectral band centered at 766 nm
related to the intrinsic fluorescence emission of CH3NH3PbI3 was
quenched by contact with the ETLs, indicating that electron
transfer from the perovskite to the ETLs is indeed operating.
Compared with that of m-ST01/perovskite, the PL intensity of m-
ZTO/perovskite is obviously lower, indicating that the charge
extraction from the perovskite is more efficient for ZTO ETL. The
fast decay corresponds to the quenching of carriers from the
perovskite to the ETLs and the slow decay is considered to be
the result of radiative decay within the perovskite film as shown
in Figure 6b and Table S3.[25] The PL quenching in the FTO/c-
TO/m-ZTO/CH3NH3PbI3 film was faster than that in the FTO/c-
TO/m-ST01/CH3NH3PbI3 film, implying the high speed
transformation of the charge carriers from light-harvesting layer
to the former ETL. From the deconvolution algorithm of TRPL
recombination resistance (Rrec) and the chemical capacitance
(CPE) in the ETLs/CH3NH3PbI3/spiro-OMeTAD interfaces.[23]
Figure5a shows that the Rco and Rs of the our both cells
approached to each other because the same counter electrodes
and measurement methods are applied. However, compared
with that in m-ST01-based interfaces (336.2 ·cm2), the Rrec of
ZTO-based interfaces is increased to 486.7 ·cm2. The
increased Rrec (inversely related to the interface recombination
rate) indicates that electron back flow from the ZTO to spiro-
OMeTAD is retarded and the charge transfer resistance at the
interfaces c-TO/m-ZTO/CH3NH3PbI3 increased. This is in
agreement with the measurement of dark current curves.
Moreover, agreeing with the impedance measurement, the
series of resistance of the m-ZTO based cells (1.16 ) exhibits
an obvious decrease compared with that of m-ST01 based one
(2.13 ), as shown in Figure 5b.[24] In a solar cell, a lower RS is
necessary to decrease current leakage at high bias voltages so
that the FF and Voc were increased by about 3.9% and 3.0%
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701779
ChemSusChem
FULL PAPER
decay curves with bi-exponential function, it is found that the fast
PL lifetime of FTO/c-TO/m-ZTO/CH3NH3PbI3/spiro-OMeTAD/Au
mesoporous hybrid solar cell (τ1=1.27 ns) is obviously shorter
than that of the traditional FTO/c-TO/m-ST01/CH3NH3PbI3/spiro-
OMeTAD/Au mesoporous hybrid solar cell (τ1=2.12 ns).
Moreover, the slow decay feature based on ZTO shows a higher
time coefficient (τ2=167.4 ns) than that of ST01 (τ2=97.6 ns) so
as to retard recombination and enhance PV performance. This
suggests that the charge injection from CH3NH3PbI3 to m-ZTO is
more impactful than that to m-ST01 and accordingly the charge
injection/separation efficiency is advanced in the FTO/c-TO/m-
ZTO/CH3NH3PbI3/spiro-OMeTAD/Au mesoporous perovskite
hybrid solar cells.
efficiency devices reported herein show good repeatability and
long-term stability when stored in the air under room
temperature with humidity of 10% as illustrated in Figure 7d. We
tested the stability of 20 devices under illumination with AM 1.5
G (100 mW cm-2) light in air. The PCE decreases slowly ca. 12%
in 100 days at humidity of 10%. Therefore, the photovoltaic data
of 20 devices suggest that the devices based on m-ZTO are
repeatable and stable in a long term.
Conclusions
We have successfully demonstrated that wideband-gap ZTO
can be effectively employed as a mesoporous layer between c-
TO and perovskite absorber layer to obtain high-performance
PSCs. A simple spin-coated ZTO scaffold inserted onto c-TO
results in a downward shift in conduction band energy level and
the suppression of interfacial charge recombination so as to
improve dramatically electron injection efficiency in perovskite
solar cells. An obvious increased PCE for ZTO ETL (17.21%)
with good reproducibility was achieved relative to that for m-
ST01-based counterpart (14.93%). The EIS and TRPL analyses
reveal that ZTO ETL has a decreased charge recombination and
thus contributes to the improvement of the Voc and FF. The XPS
analysis confirms a negative shift in the conductive band
potential of ZTO, thus facilitating fast and easy electron transfer
from highly mobile CH3NH3PbI3 to the conduction band of ZTO.
The devices display only 12% decrease of the PCE after 100
days at ambient temperature and 10% humidity with excellent
stability. Our findings demonstrate that the use of ternary oxide
is a promising approach to tune Eg and thereby reduce interface
charge recombination in the perovskite solar cells.
25
20
15
10
5
25
20
15
10
5
(a)
(b)
200 mV
150 mV
100 mV
50 mV
20 mV
Reverse, PCE=17.21%
Forward, PCE=16.97%
0
0.0
0
0.0
0.2
0.4
0.6
0.8
1.0
0.2
0.4
0.6
0.8
1.0
Voltage / V
Normalized PCE
Normalized JSC
Voltage / V
1.2
1.0
1.2
1.0
1.2
1.0
1.2
1.0
0.8
25
20
15
10
5
ZTO
(c)
(d)
Normalized VOC
Normalized FF
0
13
0
50
100
14
15
16
17
18
Time (d)
Power Conversion Efficiency / %
Figure 7. (a) J−V curves for hysteresis tests (open and closed circles stand for
forward and reverse scans, respectively). (b) J−V curves at various scan rates
in reverse scan conditions for the same device. (c) Statistical histogram of
PCE for 70 measured devices, fitted with a Gaussian distribution (black line).
(d) Long-term stability test of the inverted perovskite solar cells stored at room
temperature with humanity of 10%.
Experimental Section
Materials All chemicals and reagents were used as received without any
further purification. Methylamine (33 wt.
% in absolute ethanol),
hydrochloric acid (37%), lead iodide (PbI2, 99.9%), 2-propanol (99.7%),
dimethyl sulfoxide (DMSO, 99.9%) and chlorobenzene (99.8%) were
Owing to the charge traps and interface defects of the low
quality perovskite film, the photocurrent hysteresis behavior is
another critical issue in perovskite solar cells.[26] As depicted in
Figure 7a, the hysteresis was almost unobserved at 20 mv s-1
scan rate in the J-V curves of m-ZTO cells, where PCE = 17.21%
for the reverse direction and 16.97% for the forward direction.
The J−V curves with variable scan rates ranging from 20 to 200
mv s-1 further reveal that photocurrents hysteresis are almost
independent on the scan-rates (Figure 7b and Table S4). As a
results, the prepared perovskite films based on m-ZTO ETLs
had excellent characteristics, in which the defects and grain
boundary were mostly eliminated and the charge transfer highly
enhanced.[27] As depicted in Figure 7c and Table S5, the PCE
statistics from 70 cells exhibit a normal distribution with a small
discreteness and a good central tendency. On average, a 3−4%
increase in absolute PCE is observed by the use of m-ZTO with
the larger grain sizes in perovskite films. Significantly, the high-
purchased
from
Sigma–Aldrich.
2,2',7,7'-Tetrakis(N,N-di-p-
methoxyphenylamine)-9,9'-spirobifluorene (spiro-OMeTAD) was obtained
from Borun Chemical (98%, Ningbo, China).
ZTO Preparation The ZTO nanoparticles were synthesized using
titanate NaxH2-xTi3O7 nanotubes as a precursor.[28] In a typical synthetic
procedure, 1 g of TiO2 powder (ST01, 10 nm, Ishihara Sangyo Kaisha,
Ltd., Japan) and an aqueous solution of NaOH (10 M, 200 mL) were
placed into a Teflon container. The mixture was vigorously stirred in an
oil bath at 120 oC for 20 h. The white suspension of titanate nanotubes
was filtrated and washed with 1 M dilute HCl and then deionized water
until a pH  7 was obtained. Then, an aqueous ammonia solution was
slowly added to an aqueous solution of ZnCl2·6H2O. After a clear solution
formed, the precursor titanate nanotubes were dispersed into this
o
solution with stirring at 60 C for 1 d. Finally, the products were carefully
washed with deionized water for several times, dried and calcined at 600
oC for 3 h in air.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701779
ChemSusChem
FULL PAPER
CH3NH3I Synthesis The CH3NH3I was synthesized and purified by the
procedures described in literature.[29] To a methanol (33 wt. %, 24 mL)
solution of methylamine was slowly added hydroiodic acid (57 wt%, 10
mL) at 0 oC. After stirring under a N2 atmosphere for 2 h, the mixture was
evaporated at 50 oC for 1 h to remove volatile substances. The as-
obtained product was washed with absolute diethyl ether for three times
and 1 sun (100 mW cm-2) conditions. The 1 sun light intensity level was
calibrated using a standard Si reference cell certified by the Newport
Corporation. All the devices were measured in a light tight sample holder
with an active area of 0.12 cm2 for each cell and fixed using an aperture
mask. The IPCE was recorded using a quantum efficiency measurement
system (QEX10, PV Measurements, Inc.) in the wavelength range from
300 to 850 nm. The EIS was obtained using a potentiostat (IM-6, Zahner)
in the frequency range from 0.1 to 100 kHz, applying the Voc derived from
the J–V test and under dark conditions.
o
and then dried at 60 C in a vacuum oven for 12 h to afford the desired
CH3NH3I as white crystals.
Device fabrication. The ZTO and ST01 pastes were prepared by our
previously described precedures.[22] Substrate preparation was carried
out under ambient conditions. Firstly, FTO-coated glass (13 Ω/sq,
Aldrich) was patterned by etching with Zn powder and 2 M HCl. Secondly,
the substrates were cleaned with detergent (diluted to the volume ratio of
15% with deionized water), rinsed with deionized water, acetone and
isopropyl alcohol, and then dried with clean dry air. After oxygen
treatment, the clean substrates were spin-coated with 0.15 M titanium
diisopropoxide bis(acetylacetonate) at 2,000 r.p.m. for 30 s. After drying
at 120 °C for 15 min, the same spin-coating and annealing process was
used for the mesoporous electron transportation layer with the mere
difference by replacing TiO2 nanocolloidal solution with the homemade
ZTO nanoparticle paste (dilute with ethanol, v/v 1:6), which was
performed repeatedly in order to increase film thickness. The films were
Acknowledgements
This work was financially supported by National Natural Science
Foundation of China (21603230 and 91433104), the 973 project
from MSTC (2014CB845603), the CAS/SAFEA International
Partnership Program for Creative Research Teams, and the
Strategic Priority Research Program of the Chinese Academy of
Sciences (XDB20000000).
Keywords: charge transmission, electron transport layer,
o
sintered again at 500 C for 30 min. Upon cooling to room temperature,
m-ZTO were infiltrated with the perovskite precursor solution in a
perovskite solar cell, ternary metal oxide, zinc titanate oxide
nitrogen-filled glovebox, which was prepared by dissolving
a
[1]
[2]
[3]
a) A. Kojima, K. Teshima, Y. Shirai, T. Miyasaka, J. Am. Chem. Soc.
2009, 131, 6050-6051; b) H. R. Tan, A. Jain, O. Voznyy, X. Z. Lan, F. P.
G. d. Arquer, J. Z. Fan, R. Q. Bermudez, M. J. Yuan, B. Zhang, Y. C.
Zhao, F. J. Fan, P. C. Li, L. N. Quan, Y. B. Zhao, Z. H. Lu, Z. Y. Yang,
S. Hoogland, E. H. Sargent, 2017, 355, 722-726.
stoichiometric amount (1:1 molar ratio) of lead iodide and methyl
ammonium iodide in DMSO at a concentration of 1.2 M of each
component.[15] The spin coating procedures include two steps, that is, at
6000 rpm first for 20 s and then 10 s by gently dropping chlorobenzene
on the spinning substrate through a micro pipette. The substrate was
then heated at 100 oC for 1 h on a hot plate in the nitrogen-filled glovebox.
Upon drying, 60 μL of spiro-OMeTAD solution was spin-coated on the
CH3NH3PbI3 layer at 3000 rpm for 20 s. The solution was prepared by
dissolving 72 mg of spiro-OMeTAD in 1 mL of chlorobenzene, to which
28.8 μL of 4-tert-butyl pyridine and 14.4 μL of lithium
bis(trifluoromethanesulfonyl)imide (Li-TFSI) solution (520 mg Li-TFSI in 1
mL of acetonitrile) were added. Finally, 80 nm of gold was deposited at
∼10-6 bar via thermal evaporation on the spiro-OMeTAD coated film.
a) T. Salim, S. Sun, Y. Abe, A. Krishna, A.C. Grimsdale, Y.M. Lam, J.
Mater. Chem. A 2015, 3, 8943-8969; b) A. Marchioro, J. Teuscher, D.
Friedrich, M. Kunst, R.V.D Kril, T. Moehl, M. Grätzel, J.E. Moser, Nat.
photon. 2014, 8, 250-255; c) P.P. Boix, K. Nonomura, N. Mathews, S.G.
Mhaisalkar, Mater. Today 2014, 17, 16-23.
a) J. Burschka, N. Pellet, S.J. Moon, R.H. Baker, P. Gao, M.K.
Nazeeruddin, M. Grätzel, Nature 2013, 499, 316-319; b) H. P. Zhou, Q.
Chen, G. Li, S. Luo, T. Song, H.S. Duan, Z.R. Hong, J.B. You, Y.S. Liu,
Y. Yang, Science 2014, 345, 542-546; c) M. Saliba, S. Orlandi, T.
Matsui, S. Aghazada, M. Cavazzini, J.P. Correa-Baena, P. Gao, R.
Scopelliti, E. Mosconi, K.H. Dahmen, F.D. Angelis, A. Abate, A.
Hagfeldt, G. Pozzi, M. Grätzel, M.K. Nazeeruddin, Nature Energy 2016,
1, 15017.
Characterization The X-ray diffraction (XRD) patterns were recorded on
a PANalytical X’Pert spectrometer using Co Kα radiation (λ= 1.78897 Å)
and the data were converted to Cu Kα data. The scanning electron
microscopy (SEM) was measured on a Hitachi S4800 instrument. The
transmission electron microscopy (TEM) and selected area electron
diffraction (SAED) pattern images were recorded on a Tecnai G2 F20
(FEI) with an accelerated photovoltage of 200 kV. N2 adsorption–
[4]
a) T. Leijtens, G.E. Eperon, S. Pathak, A. Abate, M.M. Lee, H.J. Snaith,
Nat. Commun. 2013, 4, 2885; b) W. Li, W. Zhang, S.V. Reenen, R.J.
Sutton, J. Fan, A.A. Haghighirad, M.B. Johnston, L. Wang, H.J. Snaith,
Energy Environ. Sci. 2016, 9, 490-498.
desorption analysis was carried on
a Micromeritics ASAP 2020
[5]
[6]
D. Liu, T.L. Kelly, Nat. Photonics 2013, 8, 133-138.
instrument. The UV–vis spectra were measured on Lambda-9 (Perkin-
Elmer) spectrometer. The atomic force microscopy (AFM) was used to
characterize the texture and roughness of films on a Dimension ICON,
Bruker. The steady-state photoluminescence (PL) and time-resolved PL
spectroscopy of CH3NH3PbI3 coated on both the m-ZTO and m-ST01
ETLs were performed using Edinburgh FLS920 instrument with
picosecond pulsed diode laser of 375 nm. The ultraviolet photoelectron
spectroscopy (UPS) for both m-ZTO and c-TO were measured using
ESCALAB 250Xi (Thermo Fisher) under a background pressure of 5.0 ×
10-7 Pa.
a) W.J. Ke, G. Fang, J.Q. Liu, L.B. Xiong, P.L. Qin, H. Tao, J. Wang,
H.W. Lei, B.R. Li, J.W. Wan, G. Yang, Y.F. Yan, J. Am. Chem. Soc.
2015, 137, 6730-6733; b) H.S. Rao, X.B. Chen, W.G. Li, Y.F. Xu, H.Y.
Chen, D.B. Kuang, C.Y. Su, Adv. Funct. Mater. 2015, 25, 7200-7207.
[7] K. Wang, Y. Shi, Q. Dong, Y. Li, S. Wang, X. Yu, M. Wu, T. Ma, J. Phys.
Chem. Lett. 2015, 6, 755-759.
[8]
M. Qin, J. Ma, W. Ke, P. Qin, H. Lei, H. Tao, X. Zheng, L. Xiong, Q. Liu,
Z. Chen, J. Lu, G. Yang, G. Fang, ACS Appl. Mater, Interfaces 2016, 8,
8460-8466.
[9]
A. Bera, K. Wu, A. Sheikh, E. Alarousu, O.F. Mohammed, T. Wu, J.
Phys. Chem. C 2014, 118, 28494-28501.
The photovoltaic performance in terms of J-V characteristics was
determined using a solar simulator (Sol3A Class AAA, Oriel Instruments,
Stratford, CT, USA) and a Keithley 2440 source measurement unit
(Keithley Instruments Inc., Cleveland, OH, USA) under air mass (AM) 1.5
[10] S.S. Mali, C.S. Shim, C.K. Hong, Sci. Rep. 2015, 5, 11424.
[11] S.S. Shin, E.J. Yeom, W.S. Yang, S. Hur, M.G. Kim, J. Im, J. Seo, J.H.
Noh, S.I. Seok, Science 2017, 356, 167-171.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701779
ChemSusChem
FULL PAPER
[12] X. Yin, Z.Z. Xu, Y.J. Guo, P. Xu, M. He, ACS Appl. Mater. Interfaces
2016, 8, 29580-29587.
H.C. Kwon, W. Yang, H. Hwang, H. Lee, E. Lee, S. Ma, J. Moon, J.
Mater. Chem. A 2016, 4, 15478–15485.
[13] a) Z.T. Zhang, A.J. Rondinone, J.X. Ma, J. Shen, S. Dai, Adv. Mater.
2005, 17, 1415-1419; b) M. Zhao, B.P. Bastakoti, Y.Q. Li, H. Xu, J.H.
Ye, Z.W. Liu, Y. Yamauchi, Chem. Commun. 2015, 51,14582-14585; c)
Z.S. Hong, M.D. Wei, Q.X. Deng, X.K. Ding, L.L. Jiang, K.M. Wei,
Chem. Commun. 2010, 46, 740-742; d) Y.L. Huang, D.C. Tsai, Y.C.
Lee, D.R. Jung, F.S. Shieu, Surf. Coat. Tech. 2013, 231, 153-156.
[14] E.J. Juarez-Perez, M. Wussler, F. Fabregat-Santiago, K. LakusWollny,
E. Mankel, T. Mayer, W. Jaegermann, Mora-Sero I., J. Phys. Chem.
Lett. 2014, 5, 680-685.
[22] A.Y. Pang, X. Sun, H.C. Ruan, Y.F. Li, S.Y. Dai, M.D. Wei, Nano
Energy 2014, 5, 82-90.
[23] H. Kim, K.G. Lim, T.W. Lee, Energy Environ. Sci. 2016, 9, 12-30.
[24] H.Y. Zhang, J.J. Shi, X. Xu, L.F. Zhu, Y.H. Luo, D.M. Li, Q.B. Meng, J.
Mater. Chem. A 2016, 4, 15383-15389.
[25] a) G. Niu, W. Li, F. Meng, L. Wang, H. Dong, Y. Qiu, J. Mater. Chem. A
2014, 2, 705-710; b) B. Duan, Y.K. Ren, Y.F. Xu, W.Y. Chen, Q. Ye, Y.
Huang, J. Zhu, S.Y. Dai, Inorg. Chem. Front. 2017, 4, 473-480; c) X.M.
Wen, R. Sheng, A.W.Y. Ho-Baillie, A. Benda, S.H. Woo, Q.S. Ma, S.J.
Huang, M.A. Green, J. Phys. Chem. Lett. 2014, 5, 3849-3853.
[26] a) I.C. Smith, E.T. Hoke, D. Solis-Ibarra, M.D. McGehee, H.I.
Karunadasa, Angew. Chem. Int. Ed. 2014, 53, 11232-11235; b) D.H.
Cao, C.C. Stoumpos, O.K. Farha, J.T. Hupp, M.G. Kanatzidis, J. Am.
Chem. Soc. 2015, 137, 7843-7850.
[15] A. Abate, S. Paek, F. Giordano, J. Correa-Baena, M. Saliba, P. Gao, T.
Matsui, J. Ko, S.M. Zakeeruddin, K.H. Dahmen, A. Hagfeldt, M.
Grätzela, M.K. Nazeeruddin, Energy Environ. Sci. 2015, 8, 2946-2953.
[16] M. Liu, M.B. Johnston, H.J. Snaith, Nature 2013, 501, 395-398.
[17] J.Y. Jeng, K.C. Chen, T.Y. Chiang, P.Y. Lin, T.D. Tsai, Y.C. Chang, T.F
Guo, P. Chen, T.C. Wen, Y.J. Hsu, Adv Mater. 2014, 26, 4107-4113.
[18] Q. Chen, H.P. Zhou, Z.Z. Hong, S. Luo, H.S. Duan, H.H. Wang, Y.S.
Liu, G. Li, Y. Yang, J. Am. Chem. Soc. 2014, 136, 622-625.
[27] C.G. Wu, C.H. Chiang, Z.L. Tseng, M.K. Nazeeruddin, A. Hagfeldt, M.
Grätzel, Energy Environ. Sci. 2015, 8, 2725-2733.
[28] a) Y. Yang, X.W. Sun, B.K. Tay, J.X. Wang, Z.L. Dong, H.M. Fan, Adv.
Mater. 2007, 19, 1839-1844; b) Y. Yang, R. Scholz, H.J. Fan, D. Hesse,
U. Gösele, M. Zacharias, ACS Nano 2009, 3, 555-562.
[19] I. Jeong, H. Jung, M. Park, J.S. Park, H.J. Son, J. Joo, J. Lee, M.J. Ko,
Nano Energy 2016, 28, 380-389.
[20] M.L. Michael, J. Teuscher, T. Miyasaka, T.N. Murakami, H. J. Snaith,
Science 2012, 338, 643-647.
[29] J.H. Im, C.R. Lee, J.W. Lee, S.W. Park, N.G. Park, Nanoscale 2011, 3,
4088-4093.
[21] a) C. Huang, C. Liu, Y. Di, W. Li, F. Liu, L. Jiang, J. Li, X. Hao and H.
Huang, ACS Appl. Mater. Interfaces 2016, 8, 8520-8526; b) S.F. Shaikh,
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701779
ChemSusChem
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
FULL PAPER
Ternary metal oxide zinc titanate
oxide (Zn2Ti3O8) is first used as a
Aiying Pang,[a] Deli Shen,[b] Mingdeng
Wei,*[b] and Zhong-Ning Chen*[a]
23
13
3
mesoporous electron transporting
layer in perovskite solar cells to
achieve the highest photoconversion
efficiency of 17.21%. The efficiency
decreases ca. 12% in the 100 days
with good reproducibility and stability
in long term.
Page No.1 – Page No.7
Highly Efficient Perovskite Solar Cells
Based on Zn2Ti3O8 Nanoparticles as
Electron Transporting Material with
Efficiency over 17%
FTO
TiO2
Zn2Ti3O8
CH3NH3PbI3
Spiro-OMeTAD
Au
-7
-17
0.0 0.2 0.4 0.6 0.8 1.0
Voltage / V
This article is protected by copyright. All rights reserved.