Journal of

Materials Chemistry A
View Article Online

Published on 27 March 2015. Downloaded by Sungkyunkwan University on 28/05/2015 12:40:54.

PAPER

Cite this: J. Mater. Chem. A, 2015, 3,

9999

View Journal | View Issue

Interface engineering of perovskite solar cells with

PEO for improved performance
H. P. Dong,a Y. Li,b S. F. Wang,b W. Z. Li,a N. Li,a X. D. Guoa and L. D. Wang*a
Interface engineering is an important and ecient way to further improve the conversion eciency of
perovskite solar cells. In this study, we report the modication of the electron transport layer (ETL) using
a thin layer of PEO. Characterizations showed that PEO was uniformly coated on top of the original TiOx
ETL, without resulting in an evident change of the surface morphology, hydrophilic ability or
transparency. With the interface dipole formed at the interface, the work function of the ETL greatly

Received 18th January 2015

Accepted 26th March 2015

decreased. Compared with devices with TiOx only, devices based on the modied ETL gave a nearly 15%
enhancement to the overall conversion eciency, with both Voc and Jsc improved. Further studies

DOI: 10.1039/c5ta00407a

showed that the improved performance could mainly be attributed to the better retardation of back

www.rsc.org/MaterialsA

recombination and the enhanced electron collection eciency by means of the PEO thin layer modication.

1. Introduction
Organicinorganic hybrid perovskite solar cells have been a
quite hot research topic recently, due to their rapid and
improved power conversion eciency (PCE).14 The latest record
certied by the National Renewable Energy Laboratory (NREL)
was 20.1%, which is very close to its theoretical PCE. With the
increase in the PCE, also new structures with this kind of solar
cell have been developed. At the beginning, a classical structure
of dye-sensitized solar cells was adopted, with mesoporous TiO2
as the electron collection and transport layer,5,6 but now devices
with insulating nanoparticles (such as Al2O3, ZrO2) as scaolds
have also been reported and have achieved great success.7,8
Planar structured perovskite solar cells have also been widely
researched and have given quite good results.9,10
For most of the structures, a thin layer of TiO2 or ZnO is
commonly used on a transparent conductive electrode (usually
FTO or ITO) as the electron transport layer (ETL). Matching the
energy level could prevent holes formed in the perovskite or
hole transport layer (HTL) from reaching the conductive electrode, which would otherwise lead to a back recombination and
reduced short-circuit current. Optimization of this layer is an
eective means to further improve the performance of the solar
cells. Snaith et al. developed a low temperature route for

Key Lab of Organic Optoelectronics and Molecular Engineering of Ministry of

Education, Department of Chemistry, Tsinghua University, Beijing, P. R. China.
E-mail: chldwang@mail.tsinghua.edu.cn

b
Institute of Modern Optics & State Key Laboratory for Articial Microstructure and
Mesoscopic Physics, School of Physics, Peking University, Beijing, P. R. China

Electronic supplementary information (ESI) available: Contact angle test, cross

section image, PEO dissolving situation in solvents, distribution of photovoltaic
parameters, JV curves tested with dierent scanning direction and steady-state
PL results. See DOI: 10.1039/c5ta00407a

This journal is The Royal Society of Chemistry 2015

creating a TiO2 compact layer with decreased series resistance,

which, hence, resulted in an improved ll factor and eciency.11 Yang's group doped TiO2 with yttrium to enhance this
layer's electron extraction and transport ability. The highest
eciency of 19.3% in a planar geometry was nally achieved.9
Except for a doping strategy, modifying this layer with other
materials is also a good choice to improve the device's overall
performance.12 This is the usual method adopted by
researchers for organic solar cells and polymer solar cells. For
the modication, a self-assembled functional molecular layer
or a thin layer of polymer (such as polyoxyethylene, polyethylenimine) have been reported.1317 The work function of the
ETL could be tuned and back electronhole recombination
could be retarded due to the presence of the interfacial dipole;
additionally, the series resistance could also be reduced with
modication. All these should contribute to improved
performance.
Herein, we report using a thin layer of polyoxyethylene (PEO)
to modify the TiOx blocking layer for perovskite solar cells with
Al2O3 mesoporous lms as scaolds. As a type of widely used
polymer, PEO has various advantages as an ideal material for
the interfacial engineering of ETL, such as low cost, environmentally friendly, high transmittance, high polarity and easy to
form uniform lm. Ultraviolet photoelectron spectroscopy
(UPS) results showed that with the PEO modication thin layer,
the work function of the ETL greatly decreased. Devices with
PEO modied TiOx as the ETL had improved Voc and Jsc, and
gave a nearly 15% enhanced overall power conversion eciency
(PCE) compared with control devices with bare TiOx as the ETL.
Further characterizations showed that the electronhole
recombination was signicantly retarded with the PEO modication, leading to higher electron collection eciency and
ultimately to an improved performance.

J. Mater. Chem. A, 2015, 3, 999910004 | 9999

View Article Online

Journal of Materials Chemistry A

2.

Published on 27 March 2015. Downloaded by Sungkyunkwan University on 28/05/2015 12:40:54.

2.1

Paper

Experimental section
Materials and reagents

All the solvents and reagents were used as received. PbCl2

(99.999%), Al2O3 nanoparticles dispersion in isopropanol
(<50 nm, 20 wt%), and poly(ethylene oxide) (average Mv
2 000 000) were all purchased from Sigma Aldrich. Super
dehydrated solvents of dimethyl formamide, toluene and
acetonitrile were purchased from J&K. CH3NH3I (MAI) was
purchased from Xi'an Polymer Light Technology Corp., China.
2.2

Device fabrication

TiOx ETL was fabricated on FTO glass through the atomic layer
deposition (ALD) method. The perovskite solar cells were fabricated using a similar process to Snaith's group, with the optimized parameters, whereby a thin layer of Al2O3 scaold was
spin-coated on the substrate at a rotation speed of 6000 rpm for
60 s, then dried at 150
C on a hotplate for 30 min. The perovskite
layer was prepared by spin-coating from a homogeneous 40 wt%
precursor solution (PbCl2 : MAI 1 : 3, in DMF) and annealed at
100
C for 90 min. Aer cooling down to room temperature, the
HTL layer was spin-coated on top with a spinning speed of
2500 rpm, for 60 s. Finally, a 50 nm thick gold electrode was
deposited by thermal evaporation through a shadow mask. The
active area of the studied devices was 10 mm2. For the devices
with PEO modication, a solution of PEO/acetonitrile was spincoated on to the ALD-TiOx layer at 5000 rpm for 60 s and then
dried at 60
C for 15 min to remove the residual solvent.
2.3

Characterizations

The currentvoltage curves (JV curves) and electrochemical

impedance spectroscopy (EIS) were measured by a ZAHNER
CIMPS electrochemical workstation (Germany) under solar
simulation (Xenon lamp, Oriel, AM 1.5). The incident light
intensity was calibrated with a standard crystalline silicon solar
cell. The total active area of the device was 10 mm2. The IPCE
spectra were measured in air with equipment developed by the
Institute of Physics, Chinese Academy of Sciences. The UV-vis
absorption spectra were tested with the Hitachi Model U-3010
UV-vis spectrophotometer. XPS spectra were tested with the PHI
5300 ESCA instrument and the spectra were calibrated with the
C1s peak (284.8 eV). PL decay spectra were recorded by a high
resolution streak camera system (Hamamatsu C10910). All the
samples were excited by 517 nm at room temperature with
135 nJ per cm2 per pulse. Steady-state PL spectra were tested
with a uorospectrophotometer (Jobin Yvon, FluroMax-3). The
AFM images were taken with a SPA-400, Seiko.

3.

Results and discussion

3.1

Physical characterizations of the lms

AFM was carried out to characterize the surface morphology of

the ETL without and with the PEO thin layer. As shown in Fig. 1,
no clear morphology change was noted. The surface root-meansquare (RMS) roughness decreased a little aer spinning a thin
layer of PEO. The result was as we expected, because both the

10000 | J. Mater. Chem. A, 2015, 3, 999910004

Fig. 1 Atomic force microscopy images of (a) FTO/TiOx and (b) FTO/
TiOx/PEO.

ALD TiOx and PEO should form a uniform coating on the

surface of FTO. Fig. S1 shows the contact angle tests for the
surfaces. Both the lms showed a quite good hydrophilic ability.
This is quite important for high conversion eciency perovskite
solar cells, because it ensures that the perovskite precursor
solution can inltrate into the mesoporous Al2O3 scaolding
and form a good and compact contact with the ETL. We tested
the dissolving ability of PEO in isopropanol (IPA, solvent for
Al2O3 nanoparticles) and dimethyl formamide (DMF, solvent for
perovskite precursor) to indirectly check the solvent resistance
of the PEO layer, and the results are shown in Fig. S2. PEO with
high molecular weight had a quite poor solubility in IPA and
DMF. As a result, the PEO layer should not be changed much by
the process followed. The cross-section image of the device with
PEO is shown in Fig. S3. No clear PEO layer was noted, indicating that this layer is thin.
Fig. 2 shows the UV-vis absorption spectra of the TiOx and
TiOx/PEO lms deposited on FTO glass. In wavelengths longer
than 400 nm, both the lms showed little absorption, allowing
most of the incident light to reach the active layer. The
absorption onset of the TiOx/PEO lm exhibited a blue-shi
compared with the bare TiOx lm, which might be due to the
change in the scattering cross-section of the TiOx/PEO layer.

Fig. 2 UV-vis absorption curves of TiOx and TiOx/PEO spun on FTO

glass as a hole blocking layer.

This journal is The Royal Society of Chemistry 2015

View Article Online

Paper

Journal of Materials Chemistry A

Published on 27 March 2015. Downloaded by Sungkyunkwan University on 28/05/2015 12:40:54.

the PEI-modied TiOx lms.15 The Ti2p peak was clearly

observed from the TiOx/PEO lm. This indicated that the PEO
layer was quite thin, and should only be several nanometers.
Ultraviolet photoelectron spectroscopy (UPS) was further
carried out to measure the energy level change of the ETL. The
results (Fig. 3(c)) showed that the work function of ETL dropped
from 4.15 to 3.25 eV with a thin layer of PEO spun on. This work
function change could be contributed to the dipole layer formed
by the PEO. Similar phenomena have been extensively observed
by others.16,18 The decrease of the work function of the ETL
could increase the built-in potential inside the devices, which
could not only lead to the increased Voc but might also enhance
the extraction of the charge carriers from the absorption layer,
and therefore increase Jsc.15

3.2

Photovoltaic performances

The currentvoltage (JV) characteristics of the devices based on

ETL without and with PEO are shown in Table 1 and Fig. 4. The
results were obtained in the reverse scanning direction with a
50 mV s1 scanning speed, from 1.1 V to 0 V. To optimize the
thickness of the PEO lm, solutions with dierent concentrations were used for the spin coating. The control device with
bare TiOx showed an open-circuit voltage (Voc), a short-circuit
photocurrent density (Jsc), and a ll factor (FF) of 1.00 V,
17.9 mA cm2 and 0.66, respectively, leading to a PCE of 11.8%.
With a thin layer of PEO, both the Voc and Jsc of the solar cells
increased. For the optimized thickness (from 2.5 mg mL1
solution), a PCE of 13.8% was obtained, which is enhanced by
more than 15% compared with the control device. Further
increasing the thickness of PEO gave a reduced PCE. We
thought that was due to the decreased electron transport ability
of the ETL. The distribution of the photovoltaic parameters
from 18 devices is shown in Fig. S5. JV curves tested under
dierent scanning direction are shown in Fig. S6. For both
devices with TiOx and TiOx/PEO used as the ETL, a hysteretic
behaviour was observed. Though this was a common problem
for perovskite solar cells, the reasons for it are still under
research.1,19,20
The incident photo-to-electron conversion eciency (IPCE)
spectrum of the control device and optimized device are

Fig. 3 XPS focus on the (a) Ti2p and (b) O1s peaks for FTO/TiOx and
FTO/TiOx/PEO lms. (c) UPS spectra of TiOx and TiOx/PEO lms. The
light source of UPS was a He1 discharge lamp (hn 21.2 eV).

The X-ray photoelectron spectroscopy (XPS) of the TiOx and

TiOx/PEO are shown in Fig. 3(a) and (b), and the survey of the
XPS spectra is shown in Fig. S4. To check the position of the
O1s peak from PEO, a thick PEO lm deposited on FTO with a
10 mg mL1 concentration at 5000 rpm was used. We could
notice that, with a thin layer of PEO on top, the Ti2p and O1s
peaks belonging to the TiOx lms were shied to a lower
binding energy, indicating an electronic interaction between
the TiOx and PEO. The similar shi of peaks was also found for

This journal is The Royal Society of Chemistry 2015

Table 1 Performance summary of the perovskite solar cells with

dierent thicknesses of PEO lms as a surface modication layer

Control
1 mg mL1
2.5 mg mL1
5 mg mL1
Control
(average  s.d.)a
2.5 mg mL1
(average  s.d.)a

Voc [V]

Jsc [mA cm2] FF

PCE [%]

1.00
1.01
1.02
1.01
1.03  0.03

17.9
18.6
20.7
18.7
17.1  1.1

11.8
12.5
13.8
11.4
10.8  0.90

1.05  0.03 19.0  0.8

0.66
0.67
0.65
0.60
0.61  0.03

0.62  0.02 12.3  0.64

The average and standard deviation (s.d.) of the 18 cells from 3

dierent batches processed under the same conditions.

J. Mater. Chem. A, 2015, 3, 999910004 | 10001

View Article Online

Published on 27 March 2015. Downloaded by Sungkyunkwan University on 28/05/2015 12:40:54.

Journal of Materials Chemistry A

Paper

Fig. 4 JV characteristics of solar cells based on TiOx and optimized

TiOx/PEO ETL under one sun irradiance (100 mW cm2 photon ux).

Fig. 6 (a) UV-vis absorption of the perovskite active layers for solar
cells; (b) PL decay spectra of lms: FTO/TiOx/MAPbIxCl3x (black
curve) and FTO/TiOx/PEO/MAPbIxCl3x (red curve).

IPCE spectra of solar cells based on TiOx and optimized TiOx/

PEO ETL.
Fig. 5

presented in Fig. 5. In the range of 400750 nm, the device with

PEO-modied ETL gave an enhanced IPCE, which is consistent
with the improved Jsc.
3.3

Discussion about the improved IPCE and performance

We tried to better understand the improved IPCE, which could

be given by the following equation:
IPCE(l) LHE(l)4injhcol
where LHE is the light-harvesting eciency by the active layer,
4inj is the electron injection eciency from the perovskite layer
to the electron collection layer, and hcol is the charge collection
eciency at the electrodes.21
We rstly measured the absorption spectrum of the perovskite lms to check the active layers' LHE. As shown in Fig. 6(a),
no dierence was found for the lms without or with the PEO
layer. This is not hard to understand. As mentioned before, both
the morphology and the surfaces' hydrophilicity do not change

10002 | J. Mater. Chem. A, 2015, 3, 999910004

a lot aer spinning a layer of PEO, so the active layer should not
be aected by the inserted PEO layer. The 4inj was positively
related to the ETL's electron extraction ability from the perovskite layer, which could be measured by the photoluminescence
spectroscopy (PL) decay.22 As shown in Fig. 6(b), with a thin layer
of PEO, the PL quenching was slightly reduced, indicating it was
a little hard for the electrons to inject into the hole-blocking
layer. This suggested that the PEO reduced the 4inj mildly due to
its insulating property. This was also conrmed by the
enhanced steady state PL for the TiOx/PEO/perovskite sample,
as shown in Fig. S7. Now that the LHE remains constant and
the 4inj is even a little decreased, the enhancement of IPCE
should contribute to the improved hcol.
The charge collection eciency (hcol) cumulatively takes into
account the transport and recombination dynamics and could
be dened by the following equation:23
hcol transport rate/(transport rate + recombination rate)

We measured the JV curves and EIS spectra to check the

charge collection situations inside the devices. Fig. 7(a) shows
the JV curves of the devices measured in the dark. The onset of

This journal is The Royal Society of Chemistry 2015

View Article Online

Published on 27 March 2015. Downloaded by Sungkyunkwan University on 28/05/2015 12:40:54.

Paper

the dark current of the device based on TiOx/PEO ETL shied to

a higher voltage with respect to the device based on TiOx ETL. As
reported, the shi in the onset of the dark current and the
corresponding increase in the Voc could have resulted from the
reduced electronhole recombination at the interfaces, mainly
due to the back ow of electrons from the blocking layer to the
perovskite layer or hole transport layer.24,25 Electrochemical
impedance spectroscopy (EIS) was carried out to further characterize the blocking eect of the PEO layer inside the solar
cells. Fig. 7(b) shows the EIS spectra measured in the dark at an
applied potential of 0.9 V, from 105 to 0.1 Hz. Most of the
spectra correspond to the lower frequency. As reported, spectra
in the low-frequency range are related to the interfacial
recombination between electrons and holes, which has a positive correlation with the recombination resistance of the solar
cells (Rrec).26,27 It could be clearly seen that Rrec increased largely
for the TiOx/PEO blocking layer. As was found in the dyesensitized solar cells, the increase of Rrec reduces the recombination kinetics by about 2 orders of magnitude, so the charge
recombination at the interface would be signicantly retarded,
which is consistent with the conclusion from the dark JV

Journal of Materials Chemistry A

curves. Anyway, both the dark current and the EIS results
indicated that inside the devices, the back recombination was
retarded with a thin layer of PEO, leading to a higher hcol.
Thus, out of the three parameters dening the IPCE, LHE
remained constant and 4inj decreased a little upon addition of
the PEO modication overlayer on the TiOx blocking layer. In
contrast, the appreciable gain in charge collection eciency
oset the loss in injection eciency and served to explain the
enhanced IPCE and Jsc observed for devices with a PEO modication layer.

4. Conclusions
In summary, we introduced a PEO thin layer into perovskite
solar cells for surface engineering of the electron transport
layer. This thin layer was uniformly coated on top of the TiOx,
without changing the ETL's morphology, transparency or
hydrophilicity. Due to the dipole formed at the interface, the
work function of the ETL decreased with a thin layer of PEO.
Compared with devices with TiOx only, devices based on the
modied ETL gave a nearly 15% enhancement on the overall
conversion eciency, with both Voc and Jsc improved too. This
could mainly be attributed to the better retardation of the back
recombination and the enhanced electron collection eciency
by means of the PEO thin modication layer. Our results
suggest that interface engineering could be an eective
approach to further improve the performance of the perovskite
solar cells.

Acknowledgements
This work was supported by the National Natural Science
Foundation of China under Grant no. 51273104 and the
National Natural Science Foundation of China under Grant no.
91433205.

Notes and references

Fig. 7 (a) JV characteristics of solar cells based on TiOx and optimized TiOx/PEO ETL tested in the dark; (b) EIS spectra of solar cells
based on TiOx and optimized TiOx/PEO ETL tested in the dark with a
0.9 V bias voltage.

This journal is The Royal Society of Chemistry 2015

1 F. Service, Science, 2014, 344, 458.

2 Y. Z. Ma, S. F. Wang, L. L. Zheng, Z. L. Lu, D. F. Zhang,
Z. Q. Bian, C. H. Huang and L. X. Xiao, Chin. J. Chem.,
2014, 32, 957963.
3 P. Gao, M. Gratzel and M. K. Nazeeruddin, Energy Environ.
Sci., 2014, 7, 24482463.
4 S. Q. Luo and W. A. Daoud, J. Mater. Chem. A, 2015, DOI:
10.1039/c4ta04953e.
5 A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am.
Chem. Soc., 2009, 131, 60506051.
6 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. Gratzel and N.-G. Park, Sci. Rep., 2012, 2, 591.
7 M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and
H. J. Snaith, Science, 2012, 338, 643647.
8 D. Bi, S.-J. Moon, L. Haggman, G. Boschloo, L. Yang,
E. M. J. Johansson, M. K. Nazeeruddin, M. Gratzel and
A. Hagfeldt, RSC Adv., 2013, 3, 1876218766.

J. Mater. Chem. A, 2015, 3, 999910004 | 10003

View Article Online

Published on 27 March 2015. Downloaded by Sungkyunkwan University on 28/05/2015 12:40:54.

Journal of Materials Chemistry A

9 H. P. Zhou, Q. Chen, G. Li, S. Luo, T. Song, H. Duan,

Z. R. Hong, J. B. You, Y. S. Liu and Y. Yang, Science, 2014,
345, 542546.
10 D. Liu and T. L. Kelly, Nat. Photonics, 2014, 8, 133138.
11 J. M. Ball, M. M. Lee, A. Hey and H. J. Snaith, Energy Environ.
Sci., 2013, 6, 17391743.
12 Q. Hu, J. Wu, C. Jiang, T. H. Liu, X. L. Que, R. Zhu and
Q. H. Gong, ACS Nano, 2014, 8, 1016110167.
13 H.-L. Yip, S. K. Hau, N. S. Baek, H. Ma and A. K.-Y. Jen, Adv.
Mater., 2008, 20, 23762382.
14 A. Y. Li, R. M. Nie, X. Y. Deng, H. X. Wei, S. Z. Zheng, Y. Q. Li,
J. X. Tang and K.-Y. Wong, Appl. Phys. Lett., 2014, 104,
123303.
15 D. Yang, P. Fu, F. J. Zhang, N. Wang, J. Zhang and C. Li,
J. Mater. Chem. A, 2014, 2, 1728117285.
16 Y. H. Zhou, F. H. Li, S. Barrau, W. J. Tian, O. Inganas and
F. L. Zhang, Sol. Energy Mater. Sol. Cells, 2009, 93, 497500.
17 C. He, C. M. Zhong, H. B. Wu, R. Q. Yang, W. Yang, F. Huang,
G. C. Bazan and Y. Cao, J. Mater. Chem., 2010, 20, 26172622.
18 Y. Zhou, L. Y. Yang, W. J. Qin and S. G. Yin, J. Renewable
Sustainable Energy, 2014, 6, 043108.

10004 | J. Mater. Chem. A, 2015, 3, 999910004

Paper

19 T. Salim, S. Sun, Y. Abe, A. Krishna, A. C. Grimsdale and

Y. M. Lam, J. Mater. Chem. A, 2015, DOI: 10.1039/c4ta05226a.
20 H.-S. Kim and N.-G. Park, J. Phys. Chem. Lett., 2014, 5, 2927
2934.
21 M. Gr
atzel, Acc. Chem. Res., 2009, 42, 17881798.
22 A. K. Chandiran, M. K. Nazeeruddin and M. Gratzel, Adv.
Funct. Mater., 2014, 24, 16151623.
23 A. K. Chandiran, N. Tetreault, R. Humphry-Baker, F. Kessler,
E. Barano, C. Yi, M. K. Nazeeruddin and M. Gratzel, Nano
Lett., 2012, 12, 39413947.
24 A. K. Chandiran, A. Yella, M. T. Mayer, P. Gao,
M. K. Nazeeruddin and M. Gratzel, Adv. Mater., 2014, 26,
43094312.
25 Y. X. Zhao and K. Zhu, J. Phys. Chem. Lett., 2013, 4, 2880
2884.
26 X. B. Xu, H. Zhang, K. Cao, J. Cui, J. F. Lu, X. W. Zeng, Y. Shen
and M. K. Wang, ChemSusChem, 2014, 7, 30883094.
27 J. Dong, Y. Zhao, J. Shi, H. Wei, J. Xiao, X. Xu, J. Luo, J. Xu,
D. Li, Y. Luo and Q. Meng, Chem. Commun., 2014, 50,
1338113384.

This journal is The Royal Society of Chemistry 2015