DOI: 10.1002/cssc.

201403478

Full Papers

Niobium Doping Effects on TiO2 Mesoscopic Electron

Transport Layer-Based Perovskite Solar Cells
Dong Hoe Kim,[c] Gill Sang Han,[a] Won Mo Seong,[c] Jin-Wook Lee,[d] Byeong Jo Kim,[a] NamGyu Park,[d] Kug Sun Hong,[c] Sangwook Lee,*[b] and Hyun Suk Jung*[a]
doped TiO2 (NTO) and the undoped TiO2 whereas that of the
heavy doped NTO decreased by as much as ~ 0.3 eV. The lightly doped NTO-based PSCs exhibit 10 % higher efficiency than
PSCs based on undoped TiO2 (from 12.2 % to 13.4 %) and 52 %
higher than the PSCs utilizing heavy doped NTO (from 8.8 % to
13.4 %), which is attributed to fast electron injection/transport
and preserved electron lifetime, verified by transient photocurrent decay and impedance studies.

Perovskite solar cells (PSCs) are the most promising candidates

as next-generation solar energy conversion systems. To design
a highly efficient PSC, understanding electronic properties of
mesoporous metal oxides is essential. Herein, we explore the
effect of Nb doping of TiO2 on electronic structure and photovoltaic properties of PSCs. Light Nb doping (0.5 and 1.0 at %)
increased the optical band gap slightly, but heavy doping
(5.0 at %) distinctively decreased it. The relative Fermi level position of the conduction band is similar for the lightly Nb-

Introduction
Perovskite solar cells (PSCs) based on light-absorbing materials
(LAM) consisting of organometallic halides have become
a strong challenger for conventional solid-state solar cells
owing to cost-efficient processes for fabrication as well as high
power conversion efficiencies (PCEs). The record PCE of PSCs
has rapidly grown to ~ 19 %[1] since the first report of all-solidstate device structures in 2012.[2]
PSCs have various cell structures, for example, mesoscopic
structures,[3] meso-superstructures,[4] and planar junctions,[1]
based on the electron transport properties of the perovskite
LAMs. In general, mesoscopic device structures are needed for
perovskite LAMs which have short (~ 100 nm) charge diffusion

lengths (e.g., CH3NH3PbI3). There have been many studies to

match appropriate hole-transport materials (HTMs) with LAMs
in mesoscopic devices.[5, 6] However, mesoporous electron
transport materials (ETMs) have been less studied in spite of
the recent findings that the mesoporous ETM layer critically affects the charge diffusion lengths in the photoelectrode of
solar cells and the hysteresis characteristics of the device performance.[7, 8] A high performance and less hysteresis in photovoltaic properties could be obtained by modifying the mesoporous ETM.[8] Recently, metal-oxide-based mesoporous ETMs
have been studied with emphasis on modifications of morphology/structure,[9] chemical/physical properties of the ETM/
LAM interfaces,[10] the electronic band structure,[1113] and the
material itself.[14, 15] Those approaches were effective for adjusting electron transport and recombination properties and
device stability.
Doping chemistry is one of the most effective methods to
modify both the electronic band structures and surface states
of a material. Thus far, however, doping effects have not been
intensively studied in PSCs, whereas it is well studied in liquidelectrolyte-based dye-sensitized solar cells (DSCs). Especially,
aliovalent substitutions using Y3 + , Nb5 + , or Ta5 + doping into
TiO2 are known to be effective in modifying electrical properties of TiO2.[11, 16, 17] Among them, Nb5 + is regarded as an appropriate n-type dopant that enhances the charge collection properties. In DSCs, several groups demonstrated this effect by
adopting a degenerate Nb:TiO2 (NTO) compact layer[18] and
a NTO nanoparticles (NPs)-based mesoporous layer.[16] However, previous studies on NTO NPs could not see isolate the pure
effect of Nb doping on photovoltaic properties because the
NTO NPs had different sizes or different phases at different Nb
doping levels.[16, 19, 20] In PSCs, moreover, there are a few reports
on the effect of p-type dopants, such as Y3 + , Al3 + ,[11, 12] but the

[a] G. S. Han, B. J. Kim, Prof. H. S. Jung

School of Advanced Materials Science&Engineering
Sungkyunkwan University
Suwon 440-746 (Korea)
E-mail: hsjung1@skku.edu
[b] Dr. S. Lee
Department of Materials Science and Engineering
University of California at Berkeley
CA 94720 (USA)
E-mail: wook2@berkeley.edu
[c] D. H. Kim, W. M. Seong, Prof. K. S. Hong
Department of Materials Science and Engineering
Seoul National University
Seoul 151-744 (Korea)
[d] J.-W. Lee, Prof. N.-G. Park
School of Chemical Engineering
Sungkyunkwan University
Suwon 440-746 (Korea)
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201403478.
This publication is part of a Special Issue on Sustainable Chemistry at
Sungkyunkwan University. Once the full issue has been assembled,
a link to its Table of Contents will appear here.

ChemSusChem 0000, 00, 0 0

These are not the final page numbers!

 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

Full Papers
effects of n-type dopants, especially Nb5 + , have not been systemically investigated yet.
Herein, we report the effects of Nb doping of TiO2 NPs on
the electronic structure of TiO2 and on the NTO NP-based PSCs
performances. Well-crystallized NTO NPs (0, 0.5, 1.0, and
5.0 at % of Nb) were successfully synthesized using a two-step
solgel (TSSG) method. Electronic band structures of the NTOs
were investigated using UV/Vis spectroscopy and valence-band
X-ray spectroscopy (XPS). When the NTO NPs were used for
mesoporous ETM layers in PSCs, 0.5 at %-doped-NTO (05NTO)
showed faster electron transport and longer electron lifetime
than undoped TiO2 and 5.0 at %-doped-NTO (50NTO). As
a result, PSCs based on 05NTO exhibited higher PCEs than
PSCs based on undoped TiO2 (10 %) or 50NTO (52 %).

persed in the TiO2 nanoparticles, indicating successful doping.

The 10NTO and 50NTO NPs also do not show Nb agglomeration (Figure S1 in the Supporting Information). Precise Nb concentrations for the NTOs were measured using an electron
probe microanalysis (EPMA) system and XPS (Figurse S2 and S3
in the Supporting Information). The measured Nb concentration is proportional to the concentration of the initial Nb
source for the synthesis. The XPS analysis shows the existence
of the Nb5 + state in the 05NTO, 10NTO, and 50NTO NPs, implying n-type doping. The Nb concentration calculated using the
XPS spectra is consistent with the EPMA results. Figure 2 a

Results and Discussion

Characterization of NTO NPs
NTO NPs were successfully synthesized using the TSSG
method, which is well described in previous reports (see the
Experimental Section for details).[21, 22] The final oxide NPs have
a uniform size and high crystallinity as the oxide NPs are
grown in a metal hydroxide gel network, which is prepared
from a metal-alkoxide precursor solution. In this study, Nb ethoxide as the Nb source was added to the Ti-isopropoxidebased precursor solution. The Nb concentrations ([Nb]/[Ti])
were 0.5 (05NTO), 1.0 (10NTO), and 5.0 at % (50NTO). Transmission electron microscopic (TEM) images of the 05NTO NPs are
shown in Figure 1 a and b. The NPs are well dispersed and
have a mean diameter of ~ 30 nm. The d-spacing obtained
from the lattice fringes is 0.36 nm, which corresponds to the
(101) plane of anatase. The selected-area electron diffraction
(SAED) pattern (Figure 1 c) reveals that the 05NTO NPs are well
crystallized in the anatase phase without the presence of impurities (such as Nb2O5). Energy-dispersive X-ray spectroscopic
(EDS) mapping shows that Nb is not segregated but well dis-

Figure 2. a) Plane-view FESEM images of 00NTO, 05NTO, 10NTO, and 50NTO

NP-based mesoporous films, b) average particle size calculated from the
FESEM images, and c) XRD patterns of the films. Cross marks and end of
error bars represent minimum and maximum values, and the middle line in
each box represents the median value. Upper and lower ends of the box
represent the third quartile (25 %) and first quartile (75 %), respectively. Filled
squares indicate mean values.

shows scanning electron microscopic (SEM) images of the NPbased mesoporous films on fluorine-doped tin oxide (FTO)
substrates. The films were fabricated through spin-coating of
a NP-based paste and subsequent annealing at 500 8C in air.
All films made using the 00NTO, 05NTO, 10NOT, and 50NTO
NPs had very similar particle size (~ 32 nm) and porosity
(~ 24 %, pore area per total area from the SEM images). The
mean particle sizes are plotted in Figure 2 b. The particle sizes
before and after the annealing process are almost the same. In
addition, X-ray diffraction (XRD) patterns of the mesoporous
films show only the anatase phase without the presence of
any Nb oxides. Therefore, Nb doping was successful owing to
a high solubility of Nb in anatase up to ~ 20 %.[23]
Electronic structure and surface defects of the synthesized
NTOs were investigated by means of optical absorbance and
valence-band XPS analysis. Figure 3 a shows the absorbance
spectra of the NPs. Interestingly, the absorption edges shift in

Figure 1. Typical TEM images are shown (a) and (b). c) SAED pattern of
05NTO nanoparticles and d) EDS elemental mapping images showing Ti, O,
and Nb elements in the NPs.

&

ChemSusChem 0000, 00, 0 0

www.chemsuschem.org

 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

These are not the final page numbers!

Full Papers
high Jsc, Voc, and FF values,
whereas the 50NTO cells show
significant deterioration of all
parameters, leading to a decrease
in PCE by 8.8 %.
The differences in photovoltaic properties might result from
the charge injection or collection
properties of the NTO-NP-based
mesoporous films because all
PSCs have a similar amount of
LAM, which can be evidenced by
i) similar NP size and porosity of
Figure 3. a) Absorption spectra and b) valence-band XPS spectra of NTO NPs. Inset of (b) shows overall spectra for
the NP films, ii) similar thickness
binding energies ranging from 0 to 10 eV.
of LAM/ETM layer (see Figure S5
in the Supporting Information),
iii) negligible difference in optical transmittance of the ETM/
different directions depending on amount of doping; a small
FTO photoelectrode (see Figure S6 a in the Supporting Inforblue shift occurred on light doping whereas a distinctive red
shift can be observed for heavy doping. The optical band gap
energies are evaluated to be 3.23, 3.26, 3.25, and 3.17 eV for
00NTO, 05NTO, 10NTOR, and 50NTO, respectively, by fitting the
spectra using the equation ahn = A(hnEg)m, where a is the absorption coefficeient, , A is the constant, hv is the incident
photon energy, and Eg is the optical band gap energy; m = 1=2
for an indirect semiconductor. This phenomenon was observed
several times before, but the origin has not been understood
clearly yet.[24, 25]
Figure 3 b shows the valence-band XPS spectrum. The relative position of the valence band maximum (VBM) with regard
to the Fermi level (EF) can be determined by linear extrapolation of the spectrum. The extrapolated x-axis intercepts of
05NTO and 10NTO are located around that of the undoped
TiO2, whereas 50NTO shows a clear shift of the position to
a higher binding energy (as much as ~ 0.24 eV), which indicates
that EF of the 50NTO departs more from the VBM level compared to the other samples. Considering the Eg reduction
(~ 0.06 eV) of 50NTO, it is clear that the EF of 50NTO is considerably closer to the conduction band minimum (CBM-EF) than
that of the undoped TiO2 (as much as ~ 0.3 eV) whereas the
CBM-EF of the lightly doped NTOs are very similar to the undoped TiO2.
Figure 4. a) Representative JV curves of PSCs based on NTO NPs, and b) Jsc,
Voc, FF, and PCE versus Nb concentration. Cross marks and end of error bars
represents minimum and maximum values, and the middle line in each box
represents the median value. Upper and lower ends of the box represent
the third quartile (25 %) and first quartile (75 %), respectively. Filled squares
indicate mean values.

Photovoltaic properties of NTO-based PSCs

Photovoltaic properties of the NTO-NP-based PSCs (with
a 0.14 cm2 metal aperture) were measured under AM 1.5G conditions (100 mW cm2). The photocurrent densityvoltage (JV)
curves of the optimized devices are displayed in Figure 4 a; the
mean values of the short-circuit current density (Jsc), open-circuit voltage (Voc), fill factor (FF), and PCE are summarized in
Table 1 and Figure 4 b. For each doping level, ten devices were
tested. All devices show little hysteresis in the JV curves
during reverse and forward scans, but the averaged PCEs calculated from those reverse and forward scans are similar to
steady-state PCEs (see Figure S4 in the Supporting Information). It is noteworthy that the 05NTO cells exhibit a PCE by increased by 10 % compared to the undoped TiO2 cells, owing to
ChemSusChem 0000, 00, 0 0

www.chemsuschem.org

These are not the final page numbers!

Table 1. Photovoltaic parameters of PSCs fabricated using various NTObased mesoporous films.

Film

Jsc
[mA cm2]

Voc
[V]

PCE
[%]

FF

00NTO
05NTO
10NTO
50NTO

18.1
18.7
19.1
17.2

0.950
0.990
0.960
0.790

12.2
13.4
12.9
8.8

0.706
0.723
0.708
0.646

 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

Full Papers

Figure 5. a) LHE, b) EQE, and c) IQE spectra of NTO-based PSCs, where the integrated Jsc of EQE was calculated to be 15.42, 16.81, 16.64, and 13.69 mA cm2 for
00NTO, 05NTO, 10NTO, and 50 NTO, respectively.

mation) and hence iv) almost the same absorption (A) and reflection (R) spectrum of the LAM/ETM/FTO layers (see Figure S6 b in the Supporting Information). The light-harvesting
efficiency (LHE) of the NTO-based PSCs are compared in Figure 5 a by using the absorption and reflection spectrum, and
Equation (1):[26]
LHE 1R110A

The LHE of the PSCs show almost the same spectrum over
the measured wavelength range. However, external quantum
efficiency (EQE, that is, incident photon-to-current efficiency) of
the PSCs are very different for the undoped (00NTO), lightly
doped (05NTO and 10NTO), and highly doped (50NTO) NTObased PSCs, as shown in Figure 5 b. The lightly doped NTObased PSCs exhibit higher EQE than the other PSCs in the
entire wavelength region, which is consistent with the Jsc
trend, implying a superior charge injection/collection of the
lightly doped NTOs. As the internal quantum efficiency (IQE,
that is, absorbed photon-to-current conversion efficiency) excludes light harvesting, it only reflects the charge injection/collection properties. IQE of the PSCs were calculated by using
Equation (2)[26]
IQE % EQE %=LHE %

Figure 6. tc of the PSCs as a function of Jsc. For measuring tc, a probe light
(535 nm) and bias light (680 nm) were used; undoped NTO and heavily
doped NTO were measured using 0.38, 0.147, and 0.045 nW probe light and
14.99, 2.59, and 0.207 mW bias light whereas lightly doped NTO was measured using 0.3, 0.147, and 0.045 nW probe light and 9.99, 2.59, and
0.207 mW bias light. The occurred charge by probe light rapidly decreased,
showing first-order exponential decay. The time constant t was obtained by
fitting to the measured data using the relation y = y0 + A exp(x/t). The collected charge induced by the probe laser was maintained below 0.1 % of
the steady-state charge.

and the results are plotted in Figure 5 c. Clearly, the PSCs

based on lightly doped NTOs exhibit IQE considerably higher
(~ 16 %) than that of undoped-TiO2-based PSCs in the entire
wavelength region, especially at long wavelengths (600
750 nm).
Charge collection of the PSCs was investigated using transient photocurrent/photovoltage decay and impedance analyses, by which electron transport and recombination properties
can be estimated. Figure 6 shows electron transport time constants (tc) extracted from photocurrent decay curves under
short-circuit conditions. The tc of all devices show typical
power-law dependence on the photocurrent density, which is
in compliance with the well-known multiple trapping/detrapping process of electron transport in mesoscopic solar cells.[27]
The 05NTO-based PSC shows two times faster electron trans-

&

ChemSusChem 0000, 00, 0 0

www.chemsuschem.org

port than the undoped-TiO2-based PSC. On the other hand,

the electron transport becomes slower as the Nb doping concentration increases. The different effects of Nb doping on
transport with increasing doping level will be discussed later.
Figure 7 shows the Nyquist plot and electron recombination
time constants (tR) extracted from transient photovoltage
decay under open-circuit conditions. The radius of the semicircle in the Nyquist plot decreases with increasing dopant concentration, indicating that internal resistance decreases with
Nb doping. The radius of the impedance semicircle corresponds to the internal resistance of the photoelectrode including the HTM, light-harvesting materials (LHM), ETM, and transparent conducting oxide (TCO) layers and the interfaces of the
layers, although a common, exact interpretation of impedance
components has not been established yet. It was suggested in
4

 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

These are not the final page numbers!

Full Papers
the type of defects changes
from a cation vacancy to Ti3 + defects on increasing the Nb
doping level, which is evidenced
by the appearance of Ti3 + peak
in the XPS spectrum for 50NTO
(Figure S7 in the Supporting Information).
It was suggested that the
oxygen vacancies at the surface
of NPs could be reduced by light
Nb doping because the Nb5 + at
the surface site attracts additional oxygen atoms.[19] On the other
Figure 7. a) Nyquist plot of the PSCs measured at open-circuit conditions. The impedance spectra measured in
hand, the Ti3 + defect are well
the frequency range of 10 mHz to 2 MHz at open circuit potential under 1 sun illumination. b) tR of the PSCs as
known to be accompanied by
a function of Voc. The detailed measurement conditions of tR are same as those of the tc measurements (Figure 6).
oxygen vacancies; hence, the
concentration of oxygen vacancies can increase on increasing the doping level over a critical
a previous report on DSCs based on heavily doped NTO that
point. Considering that the valence band of NTO is dominantly
this reduced internal resistance mainly originates from inaffected by the O 2p orbital,[29] oxygen vacancies are likely to
creased conductivity caused by Nb doping, which increases
[16]
be responsible for the electronic band change of the NTOs. In
the electron concentration in TiO2. In general, the increase of
electron concentration in a semiconductor would increase the
this point of view, the high Voc and the fast electron transport
chance of carrier recombination. Interestingly, however, Figof the lightly doped NTO can be attributed to the reduced
ure 7 b shows that 05NTO- and 10NTO-based PSCs have almost
amount of oxygen vacancies at the NP surface, which results in
the same tR as the undoped-TiO2-based PSC, whereas the
a decrease in the electron concentration at the surface. On the
50NTO-based PSC has a smaller tR. Therefore, the electron life
other hand, the high oxygen vacancyTi3 + defect concentration
times are preserved at low Nb doping levels ( 1 at %) while it
of the heavily doped NTO might reduce the Voc and the elecsignificantly decreases at a high doping level (5 at %).
tron life time, as oxygen vacancyTi3 + is a well known recomThose distinctive Nb doping effects on the electronic strucbination center in mesoscopic solar cells.[30] In addition, these
ture and the photovoltaic properties, depending on the Nb
defects might act as scattering centers, which slows down
concentration, may be understood based on preferred defect
electron transport.[25]
types. In general, three types of defects can be created when
the Nb5 + substitutes Ti4 + in TiO2 : first, creation of one Ti vacanConclusions
cy per four Nb atoms inserted [1/4 VTi in Eq (3)]; second, stoi4+
3+
chiometric reduction of Ti to Ti [TiTi in Eq (4)]; and third,
We studied the effect of Nb doping of anatase nanoparticles
Nb5 + to Nb4 + per one Nb atom inserted [NbTix in Eq (5)] as fol(NPs) on the electronic band structure of TiO2 and on the pholows expressed in the classical KrgerVink notation, respectovoltaic properties of perovskite solar cells (PSCs) based on
tively:[28]
Nb-doped TiO (NTO). The Nb doping resulted in contradictory
2

1=2 Nb2 O5 TiTi ! NbTi C 1=4VTi

0000

TiO2 1=4 O2

1=2 Nb2 O5 TiTi x ! NbTi C TiTi 0 5=4 O2

x

NbO2 TiTi ! NbTi TiO2

effects on the electronic structure and photovoltaic properties

depending on the doping level. Lightly doped NTO NPs exhibit
band-gap (Eg) and Fermi-level energies (EF) similar to undoped
TiO2 NPs and improved electron injection/transport properties
when used in PSCs. On the other hand, heavy doping reduces
Eg, resulting in a EF closer to the conduction band minimum
(CBM), and retards the electron injection/transport of PSCs.
This contrasting result was attributed to the generation of different defect types depending on the doping level. This result
can inspire the design of appropriate n-type electron transport
materials (ETMs) of PSCs.

3
4
5

The defect type is affected by the synthesis conditions, such

as O2 partial pressure, temperature, the Nb concentration, and
the energy cost for creating such defects. In our work, the
dominant factor determining the type of defects is only the
doping level, as the NTOs were synthesized following by using
the exactly same route except for changing the amount of Nb
source in the precursor. The case presented in Equation (5) can
be excluded, as the XPS spectra (Figure S3 in the Supporting
Information) show only the presence of Nb5 + in our NTO NPs.
At a low doping level Equation (3) is likely preferred because
cations tend to remain in their higher oxidation states whereas
Equation (4) mainly takes place at high doping levels.[28] Thus,
ChemSusChem 0000, 00, 0 0

www.chemsuschem.org

These are not the final page numbers!

Experimental Section
Synthesis of the NTO NPs
A stock solution of Nb-doped Ti4 + (0.5 m) was prepared by mixing
titanium isopropoxide (TTIP; 97 %, SigmaAldrich), 2 m niobium(V)

 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

Full Papers
ethoxide, and triethanolamine (TEOA, 98 %, SigmaAldrich) at a suitable ratio ranging from 1:0:2 to 0.95:0.05:2, followed by the addition of distilled water. Afterwards, the prepared stock solution (pH
 9.3) was mixed with HClO4 to adjust the pH value of the solution
to 8.5. This solution was placed in an electric oven and aged at
100 8C for 24 h (first aging). After the first aging step, the gel was
transferred to an autoclave and aged at 160 8C for 72 h (second
aging). After completion of the aging processes, the synthesized
TiO2 NPs were washed using the same process reported in our previous study.[31] The TiO2 paste using the prepared NPs was prepared by modifying the method reported by S. Ito et al.[32]

mode X-ray photoelectron spectroscopy (XPS, Sigma Probe, ThermoVG, UK). All XPS peaks were calibrated with respect to the C 1s
level at 284.5 eV. The currentvoltage characteristics of the solar
cells were measured under solar-simulated light (Newport Oriel
Solar 3 A Class AAA, 64 023 A) by using a potentiostat (CHI 600D,
CH Instruments). The AM 1.5G sun light (100 mA cm2) was calibrated using a standard Si-solar cell (Oriel, VLSI standards). Impedance
spectroscopy was carried out at open-circuit conditions by applying an AC amplitude signal of 5 mV and frequencies ranging from
10 mHz to 2 MHz using the same potentiostat under 1 sun light illumination. External quantum efficiency (EQE) was measured using
an EQE system (PV Measurements Inc.). The source light for generating the monochromatic beam was a 75 W Xe lamp (USHIO,
Japan). EQE data acquisition was carried out at DC mode. Transient
photocurrent measurements were performed using a small light
perturbation, where probe light was incident over the steady-state
charge induced by a bias light. The detailed measurement conditions were the same previously reported by Son et al.[33]

Fabrication of the PSCs using the NTO NPs

The solar cells were fabricated on FTO substrates (TEC15, 15 W/&,
Pilkington), which were washed for 15 min each with acetone, deionized water, and ethyl alcohol in an ultrasonic bath. A compactTiO2 (c-TIO2) layer was spin-coated on an FTO substrate at
3000 rpm for 30 s using a 0.15 m and 0.3 m titanium diisopropoxide
bis(acetylacetonate) (75 wt % in isopropanol, Aldrich) solution in 1butanol (99.8 %, Aldrich), which was heated at 125 8C for 5 min, respectively, then annealed at 500 8C for 30 min. The FTO/c-TiO2 substrate was immersed in 0.04 m aqueous TiCl4 solution at 60 8C for
1 h. After washing with deionized water and drying, the substrate
was annealed at 500 8C for 30 min. Then, a 240260 nm-thick mesoporous TiO2 and NTO film was spin-coated onto the FTO/c-TiO2
substrate using lab-made pastes and calcining at 500 8C for 1 h in
air to remove organic components. The synthesis of the CH3NH2I
(MAI) and CH3NH3PbI3 perovskite light absorber layer were prepared by using a modified solvent engineering reported elsewhere.[8] The detailed process was as follows. The prepared
CH3NH3PbI3 solution (MAI/PbI2 = 1:1 molar ratio, 43 wt % in a mixture of g-butyrolactone (GBL) and dimethyl sulfoxide (DMSO) (7:3
v/v) at 60 8C for 12 h) was spin-coated at 1000 and5000 rpm for 10
and 30 s, respectively. For the second spin-coating step (5000 rpm),
the toluene was drop-casted on the substrate (around 2  2 cm2)
The substrate was dried on a hot plate at 120 8C for 10 min. Subsequently, 40 mL of a hole transport layer [56 mg spiro-MeOTAD,
8.4 mL 4-tert-butylpyridine, and 36.2 mL bis(trifluoromethane) sulfonamide lithium salt (Li-TFSI) solution (77 mg in 0.5 mL acetonitrile),
the whole mixture was dissolved in 1 mL chlorobenzene] was
formed after spin-coating at 2000 rpm for 20 s. All processes were
carried out under controlled atmospheric conditions and at a humidity of < 0.5 ppm. Finally, a 60 nm thick Au electrode was deposited by thermal evaporation at ~ 106 bar using a shadow mask.

Acknowledgements
This work was supported by the Global Frontier R&D Program at
the Center for Multiscale Energy System (2012M3A&A7054861).
This work was also supported by the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science
and Technology (2012M3A7B4049967, (2014R1A4A1008474 and
2014R1A2A2A01007722). Characterization of the crystal structure
and TEM analysis were supported by the Research Institute of Advanced Materials (RIAM) and Core-level and valence XPS analysis
was supported by the National Center for Inter-University Research Facilities (NCIRF). Especially, this paper is dedicated to my
late mentor, Prof. Kug Sun Hong.
Keywords: anatase
perovskite solar cells

The crystal structures of anatase and NTO nanoparticles were characterized using an X-ray diffractometer (New D8-Advance, Bruker
Miller Co.). The morphologies and microstructures of the prepared
TiO2 nanoparticles and the cross-sectional structure and thickness
of the solar cells were investigated using a field-emission scanning
electron microscopy (FESEM, JSM-7600F, JEOL). To analyze the morphology and crystal structure of anatase and NTO, high-resolution
transmission electron microscopy (HRTEM, JEM-2100F, JEOL) was
used. Energy-dispersive X-ray spectroscopy (EDS) images of NTO
were also obtained using the same microscope. The absorbance
and reflectance of various TiO2 powders dispersed in distilled H2O
and TiO2 and perovskite/TiO2 films were measured by means of ultraviolet-visible light (UV/Vis) spectroscopy (Cary 5000, Agilent).
The chemical states of Nb and valence-band position in TiO2 and
NTO nanoparticles were analyzed using normal and valence-band-

ChemSusChem 0000, 00, 0 0

www.chemsuschem.org

structure

niobium

[1] H. Zhou, Q. Chen, G. Li, S. Luo, T.-b. Song, H.-S. Duan, Z. Hong, J. You, Y.
Liu, Y. Yang, Science 2014, 345, 542.
[2] H. S. Kim, C. R. Lee, J. H. Im, K. B. Lee, T. Moehl, A. Marchioro, S. J. Moon,
R. Humphry-Baker, J. H. Yum, J. E. Moser, M. Grtzel, N. G. Park, Sci. Rep.
2012, 2, 591.
[3] J. Burschka, N. Pellet, S. J. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin, M. Grtzel, Nature 2013, 499, 316.
[4] M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami, H. J. Snaith, Science
2012, 338, 643.
[5] J. H. Heo, S. H. Im, J. H. Noh, T. N. Mandal, C. S. Lim, J. A. Chang, Y. H.
Lee, H. J. Kim, A. Sarkar, M. K. Nazeeruddin, M. Grtzel, S. I. Seok, Nat.
Photonics 2013, 7, 486.
[6] B. Conings, L. Baeten, C. De Dobbelaere, J. DHaen, J. Manca, H. G.
Boyen, Adv. Mater. 2014, 26, 2041.
[7] V. Gonzalez-Pedro, E. J. Juarez-Perez, W.-S. Arsyad, E. M. Barea, F. Fabregat-Santiago, I. Mora-Sero, J. Bisquert, Nano Lett. 2014, 14, 888.
[8] N. J. Jeon, J. H. Noh, Y. C. Kim, W. S. Yang, S. Ryu, S. I. Seok, Nat. Mater.
2014, 13, 897.
[9] H.-S. Kim, J.-W. Lee, N. Yantara, P. P. Boix, S. A. Kulkarni, S. Mhaisalkar, M.
Grtzel, N.-G. Park, Nano Lett. 2013, 13, 2412.
[10] G. S. Han, H. S. Chung, B. J. Kim, D. H. Kim, J. W. Lee, B. S. Swain, K. Mahmood, J. S. Yoo, N.-G. Park, J. H. Lee, J. Mater. Chem. A 2015, DOI:
10.1039/c4ta03684k.
[11] P. Qin, A. L. Domanski, A. K. Chandiran, R. Berger, H.-J. Butt, M. I. Dar, T.
Moehl, N. Tetreault, P. Gao, S. Ahmad, Nanoscale 2014, 6, 1508.

Characterization

&

electronic

 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

These are not the final page numbers!

Full Papers
[12] S. K. Pathak, A. Abate, P. Ruckdeschel, B. Roose, K. C. Gdel, Y. Vaynzof,
A. Santhala, S.-I. Watanabe, D. J. Hollman, N. Noel, Adv. Funct. Mater.
2014, 24, 6046.
[13] K. Manseki, T. Ikeya, A. Tamura, T. Ban, T. Sugiura, T. Yoshida, RSC Adv.
2014, 4, 9652.
[14] D. Bi, S.-J. Moon, L. Hggman, G. Boschloo, L. Yang, E. M. Johansson,
M. K. Nazeeruddin, M. Grtzel, A. Hagfeldt, RSC Adv. 2013, 3, 18762.
[15] L. S. Oh, D. H. Kim, J. A. Lee, S. S. Shin, J.-W. Lee, I. J. Park, M. J. Ko, N.-G.
Park, S. G. Pyo, K. S. Hong, J. Phys. Chem. C 2014, 118, 22991.
[16] X. J. Lue, X. L. Mou, J. J. Wu, D. W. Zhang, L. L. Zhang, F. Q. Huang, F. F.
Xu, S. M. Huang, Adv. Funct. Mater. 2010, 20, 509.
[17] X. J. Feng, K. Shankar, M. Paulose, C. A. Grimes, Angew. Chem. Int. Ed.
2009, 48, 8095; Angew. Chem. 2009, 121, 8239.
[18] S. Lee, J. H. Noh, H. S. Han, D. K. Yim, D. H. Kim, J. K. Lee, J. Y. Kim, H. S.
Jung, K. S. Hong, J. Phys. Chem. C 2009, 113, 6878.
[19] T. Nikolay, L. Larina, O. Shevaleevskiy, B. T. Ahn, Energy Environ. Sci.
2011, 4, 1480.
[20] A. K. Chandiran, F. Sauvage, M. Casas-Cabanas, P. Comte, S. M. Zakeeruddin, M. Grtzel, J. Phys. Chem. C 2010, 114, 15849.
[21] K. Kanie, T. Sugimoto, Chem. Commun. 2004, 1584.
[22] S. Lee, I. S. Cho, J. H. Lee, D. H. Kim, D. W. Kim, J. Y. Kim, H. Shin, J. K.
Lee, H. S. Jung, N. G. Park, K. Kim, M. J. Ko, K. S. Hong, Chem. Mater.
2010, 22, 1958.
[23] Y. Furubayashi, T. Hitosugi, Y. Yamamoto, K. Inaba, G. Kinoda, Y. Hirose,
T. Shimada, T. Hasegawa, Appl. Phys. Lett. 2005, 86, 252101.

ChemSusChem 0000, 00, 0 0

www.chemsuschem.org

These are not the final page numbers!

[24] N. A. Tsvetkov, L. L. Larina, O. Shevaleevskiy, E. A. Al-Ammar, B. T. Ahn,

Prog. Photovoltaics Res. Appl. 2012, 20, 904.
[25] P. S. Archana, R. Jose, T. M. Jin, C. Vijila, M. M. Yusoff, S. Ramakrishna, J.
Am. Ceram. Soc. 2010, 93, 4096.
[26] A. G. Agrios, A. Hagfeldt, J. Phys. Chem. C 2008, 112, 10021.
[27] N. Kopidakis, K. D. Benkstein, J. van de Lagemaat, A. J. Frank, J. Phys.
Chem. B 2003, 107, 11307.
[28] A. M. Ruiz, G. Dezanneau, J. Arbiol, A. Cornet, J. R. Morante, Chem.
Mater. 2004, 16, 862.
[29] X. D. Liu, E. Y. Jiang, Z. Q. Li, Q. G. Song, Appl. Phys. Lett. 2008, 92,
252104.
[30] J. Weidmann, T. Dittrich, E. Konstantinova, I. Lauermann, I. Uhlendorf, F.
Koch, Sol. Energy Mater. Sol. Cells 1999, 56, 153.
[31] D. H. Kim, W. M. Seong, I. J. Park, E.-S. Yoo, S. S. Shin, J. S. Kim, H. S.
Jung, S. Lee, K. S. Hong, Nanoscale 2013, 5, 11725.
[32] S. Ito, P. Chen, P. Comte, M. K. Nazeeruddin, P. Liska, P. Pchy, M. Grtzel,
Prog. Photovoltaics Res. Appl. 2007, 15, 603.
[33] D.-Y. Son, J.-H. Im, H.-S. Kim, N.-G. Park, J. Phys. Chem. C 2014, 118,
16567.

Received: December 30, 2014

Published online on && &&, 0000

 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

FULL PAPERS
D. H. Kim, G. S. Han, W. M. Seong,
J.-W. Lee, B. J. Kim, N.-G. Park, K. S. Hong,
S. Lee,* H. S. Jung*

Nb-doped TiO2 nanoparticles for perovskite solar cell: TiO2 nanoparticles

doped with n-type Nb5 + have a diameter of 30 nm and a pure anatase phase;
they are used for fabricating perovskite
solar cells. Light doping increases the
photovoltaic energy conversion efficiency by 10 % due to improved electron injection/transport properties of the
nanoparticle-based electron transport
layer.

&& &&
Niobium Doping Effects on TiO2
Mesoscopic Electron Transport LayerBased Perovskite Solar Cells

&

ChemSusChem 0000, 00, 0 0

www.chemsuschem.org

 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

These are not the final page numbers!