Comprehensive Study of the Two-Step Solution Processes in Ambient

Air for Lead Iodide Perovskite Solar Cells

Hojjatollah Sarvari a, Xiaohui Wang a,b, Yafei Wang b, Peng Zhang b, Shibin Li b, Hongmei Dang a, Vijay P.
Singh a, and Zhi Chen a,b*
a
Department of Electrical & Computer Engineering and Center for Nanoscale Science & Engineering,
University of Kentucky, Lexington, KY 40506, USA
b
School of Optoelectronic Information and State Key Laboratory of Electronic Thin Films & Integrated
Devices, University of Electronic Science and Technology of China, Chengdu, Sichuan 610054, China
*
Corresponding Author: [email protected]

Abstract — Organic-inorganic lead halide perovskite solar cells perovskite solar cells in higher humidity conditions [6-9]. In
have attracted a great deal of interest because of potential low cost addition, for compact titanium oxide (c-TiO2) layer deposition,
and high efficiency. However, in most publications, solar cells were spray pyrolysis and spin-coating precursor solution methods are
fabricated in a highly controlled environment (glove boxes). Only
a few papers reported fabrication of solar cells in air with the two typical methods to prepare hole blocking layer for
maximum efficiency of ~15%. We carried out comprehensive perovskite solar cells but radio frequency sputtering method can
study of the two-step solution process parameters for fabrication also be employed to fabricate a pinhole-free c-TiO2 layer for
of perovskite solar cells in ambient air including PbI2 spin coating perovskite solar cells. However, there is no report on
speed, mesoporous TiO2 spin coating speed, hole transport comprehensive studies of process parameters in ambient air in
material (HTM) spin coating speed, and HTM exposure to air etc.
We found that the higher spin speed of PbI2 is critical for high- details such as PbI2 spin coating speed, mesoporous TiO2 spin
quality PbI2 film formation. The HTM exposure to air is also coating speed, HTM spin coating speed, and HTM exposure to
important to fabricate solar cells with larger current density and air etc.
better fill factors. In this article we present comprehensive studies of two-step
spin-coating solution process of lead iodide perovskite solar
I. INTRODUCTION cells in ambient air in relative humidity of > 36% with the
maximum PCE of 11% under forward bias scan. In addition, we
Perovskite solar cells have recently demonstrated the power have used RF sputtering method for deposition of pinholes-free
conversion efficiencies (PCEs) larger to most organic solar compact thin layer of TiO2.
cells and comparable to those of commercialized Silicon solar
cells and other technologies based on inorganic
semiconductors. Perovskite with the same type of crystal II. EXPERIMENTAL
structure as CaTiO3 and general formula of ABX3 was used for The solar cell device structure is shown in Fig. 1 starting from
the first time in a solar cell in 2009 by Kojima et al. [1] which the bottom: glass/FTO substrate, compact TiO2 layer,
measured PCE was 3.8% using a thin layer of perovskite on mesoporous TiO2 layer infiltrated with the perovskite
mesoporous TiO2. Later, in 2011, Park et al modified the nanocrystals, and hole transport layer, with a final gold
structure of device to achieve a PCE of 6.5% [2]. In 2013, electrode deposited via sputtering or electron beam
Burschka et al. introduced sequential deposition method as a evaporation. The cross-section SEM image shown in Fig. 2
new route for fabrication of perovskite solar cells to optimize shows a well-defined layer-by-layer structure with obvious
the morphology of perovskite layer and the certified PCE of interfaces. The results reveal that the perovskite film is uniform
14.14% for the champion cell with the average of 12% in despite some pinholes. The thicknesses of the compact TiO2,
reverse scan measurement were achieved [3]. Since then perovskite, and HTM layers are about 60, 300, and 200 nm,
excellent works reporting high efficiencies in organic-inorganic respectively.
lead halide perovskite solar cells have been reported with All materials were purchased from Sigma-Aldrich, unless
maximum efficiencies over 20% [4, 5], showing promising otherwise specified, and they were used as received.
potential of perovskite materials towards future cost-effective Methylammonium iodide (CH3NH3I) and Co(III) TFSI salt
and high performance solar cells. The fabrication processes in were purchased from Luminescence Technology. Spiro-
most of the research works on perovskite solar cells were MeOTAD (Purity, 99.7%) was purchased from Borun New
carried out in glove boxes, it is necessary to study the Material Technology.
fabrication process in ambient air condition in order to find a
way to manufacture solar cells at low cost. There are a few A. Solar Cell Fabrication
research articles, which reported fabrication processes for

978-1-5090-5605-7/17/$31.00 ©2017 IEEE

 The device fabrication was carried out under ambient defined by the overlap of patterned FTO and gold contact using
conditions with a humidity of usually higher than 36% using the a shadow mask.
widely used two-step method [3, 10]. Fluorine doped tin oxide
B. RF Sputtering Process
(FTO) glass substrates (Pilkington, TEC-15, 15 ohms per sq.)
with dimensions of 25.4mm by 25.4mm were used. First, RF sputtering deposition method is used for a thin compact
photolithography was used to pattern the sample, then the FTO layer of TiO2 with the thickness of 40 nm to 75 nm controlled
region under the anode contact removed by etching the FTO by the deposition time. The TiO2 target has a purity of 99.99%
with zinc powder and hydrochloric acid followed by and the distance between target and samples was about 10 cm.
hydrofluoric acid diluted with deionized water (1:10), to The chamber was pumped down for 75 min to get a pressure of
prevent the leakage on the sides of the sample. Then, patterned approximately 0.01 mTorr. In order to clean the target surface,
FTO glass was cleaned by sequentially washing with detergent pre-deposition has been done for 6 min. Process was performed
solution, acetone and isopropanol in an ultrasonic bath for 15 in pure Ar gas with flow meter of 15 sccm (standard cubic
min each, rinsed with deionized water and dried by nitrogen centimeter per minute). The sputtering process of compact TiO2
followed by 2 min oxygen plasma etching to remove the last was done at power of 150 W with deposition rate of 0.7 °A/sec
traces of organic residues right before doing the next process. at a working pressure of 4.1 mTorr. For deposition of gold by
A pinhole-free dense TiO2 layer was then deposited by RF sputtering, the power was 16 W to have a deposition rate of 0.5
plasma sputtering using a TiO2 target with power of 150 W. °A/sec at a pressure of 3.6 mTorr. The film thickness adjusted
After sputtering, the samples were gradually annealed at 450℃ by the sputtering time and monitored by the film-thickness
for 1 h in air. Subsequently, 200 nm mesoporous TiO2 layer meter which is mounted on the substrate stage in the chamber.
composed of 30-nm-sized TiO2 particles was deposited using a
C. Characterizations
commercial TiO2 paste (Dyesol 30 NR-D) diluted in ethanol
with density of 0.14 mg/mL by spin coating at 2000 rpm for 20 X-ray diffraction (XRD) analysis on the thin films was
sec. After drying at 150℃ for 5 min, the samples were gradually performed to investigate the phases of the samples, using a
annealed at 550℃ for 1 h and cooled down to room temperature SIEMENS D-500 with Cu Kα , λ=1.540562 °A radiation under
for further processes. After deposition of mp-TiO2, we do the operation conditions of 40 kV and 30 mA. For the Scanning
next step quickly in order to prevent sample from absorption of Electron Microscopy (SEM) measurements a FEI Helios
air moisture into the pores because the pores filled with Nanolab 660 dual beam was used. J-V characteristic curves
moisture particles do not allow PbI2 to infiltrate to the mp-TiO2 were recorded under ambient conditions by applying an
completely and therefor leads to decrease in the absorption of external potential bias to the cell while recording the generated
light due to uncomplete formation of perovskite. photocurrent with an Agilent 4155B source meter. Illumination
PbI2 was dissolved in N,N-dimethylformamide (DMF, of the samples was made with a standard ScienceTech SS150
99.8%, Aldrich) at a concentration of 462 mg/ml (1M) under Solar Simulator. The ScienceTech SS150 solar simulator was
stirring at 70℃ for 18 hours. The solution filtered through a 0.45 calibrated to the AM1.5G standard with a standard light meter.
μm PTFE filter and then was kept at 70℃ while stirring during The layers thickness was measured using a surface profiler
the whole procedure. The mesoporous TiO 2 films were then (Dektak 6m).
infiltrated with 20 μl PbI2 solution by spin coating at 7000 rpm
for 20 sec to form a transparent yellow layer. Samples were
III. RESULTS AND DISCUSSION
dried on the hotplate at 40℃ for 3 min and 100℃ for 5 min,
respectively. After cooling to room temperature, to convert the The XRD pattern of meso-CH3NH3PbI3/c-TiO2/FTO layers is
PbI2 layer to MAPbI3, a 250 μl of solution of CH3NH3I in shown in Fig. 3. The main diffraction peaks centered at
isopropanol (10mg/ml) was spin coated on top of the PbI2 layer 2θ=14.1° (110) and 28.5° (220) can be assigned to the
at 0 rpm for 45 s (loading time) followed by 4000 rpm for 20 s, CH3NH3PbI3 phase, showing a crystal structure, and the peak
which was dried at 115℃ for 10 min. During the first spin centered at 12.7° is attributed to PbI 2 [10]. We observe
coating process, the color of sample changed from yellow to unreacted PbI2 remained, which means the conversion of PbI2
dark brown confirms the formation of perovskite. Then 80 μl of to perovskite on exposure to the MAI solution is incomplete
HTM solution was deposited on the MAPbI 3 layer by spin however a small amount of unreacted PbI2 may be useful
coating at 4000 rpm for 30 sec. It should be mentioned that the according to the other reports [3, 11].
whole process of perovskite and HTM deposition has been done Defects in the c-TiO2 layer behave like electron traps
in less than half hour to decrease the exposure of samples to the increasing the recombination of electron hole, allow direct
air moisture. The HTM solution was prepared by mixing 72.3 contact from perovskite to the FTO and even direct contact
mg Spiro-MeOTAD in 1 ml chlorobenzene, 28.8 μl 4-tert- from HTM to the FTO. Current density-voltage curves
butylpyridine, and 17.5 μl solution of Li-TFSI in acetonitrile measured for the perovskite solar cells with different mp-TiO2
(520 mg/ml) at room temperature a day before usage. Finally, spin-coating speed provided in Fig. 4. For low spin speed
80 nm of gold was deposited on top of the device using RF deposition of mp-TiO2, the thickness of deposited layer is much
sputtering to form the back contact. The active area of device is higher than that for high spin speed one which results in less

978-1-5090-5605-7/17/$31.00 ©2017 IEEE

chance for electron hole recombination between HTM and
TiO2. Therefore better fill factor is achieved for a device with
high thickness of mp-TiO2 layer. In addition, according to the
SEM images, the contact between c-TiO2 and mp-TiO2 for low
spin speed deposition is not as good as that in high spin speed
one. The bad contact decreases the electron collection rate from
perovskite to the FTO which results in the low current density.
However with high spin speed deposition, the quality of contact
between c-TiO2 and mp-TiO2 improved but lower thickness of
mp-TiO2 layer increases the chance of recombination through
pinholes and reduces the fill factor of device.
Fig. 5 shows the J-V curves for samples with PbI 2 spin-
coating speeds of 7000 and 3000 rpm. In the case of low speed,
the thickness of deposited layer is high (~350 nm) with turbid
(cloudy) yellowish color however the PbI2 layer is transparent
yellow with the thickness of 150 nm at speed of 7000 rpm for
20 seconds that results in higher Jsc. Fig. 1. Cross-sectional structure of the perovskite solar cell device.
The thickness of HTM is very critical for high performance
perovskite solar cells and can be controlled by the spin-coating
speed. As we can see from Fig. 6, the optimal spin speed is 4000
rpm for 30 sec spinning which forms a Spiro layer with
thickness of 200 nm. In samples with low spin speeds (high
HTM thickness), the holes cannot transport to the gold contact
effectively, thus the extracted current decreases. On the other
hand, at high spin speeds with low HTM thicknesses, J sc is high
however Voc decreases. The reason for that may be due to the
increase of interface recombination at perovskite/HTM
interface which causes the generated electrons in absorber layer
to recombine with holes in cathode leads to decrease in Voc.
Gold can also enter the low thickness organic HTM layer and
reach the absorber layer to increase the electron hole
recombination.
J-V curves for the perovskite solar cells with different HTM
solution preparation time has been provided in Fig. 7.
According to the results, it is important to prepare the HTM Fig. 2. Cross-sectional SEM picture of the perovskite device.
solution a day before fabrication process in order to let Spiro to
oxidize in air to increase the conductivity of solution.
In Fig. 8 we have provided the J-V curves for samples left in
vacuum or ambient air with humidity above 36% right after
deposition of HTM. Samples which left for 18 hours in ambient
air show higher Jsc,Voc, and FF compared to those left in the
vacuum. It can be attributed to the Spiro oxidation which
increases the conductivity of film and enhances the transport
process of holes toward gold contact [12].
Fig. 9 shows the forward scan J–V characteristics measured
under AM1.5G (100 mW cm-2) irradiation for the champion
device fabricated in ambient air with relative humidity above
36%. The champion cell shows a short-circuit current density
of 17.675 mA/cm2, open circuit voltage of 1V, and fill factor of
62.5%, yielding a PCE of 11.04%.

Fig. 3. XRD spectra of meso-CH3NH3PbI3/c-TiO2/FTO layers.

978-1-5090-5605-7/17/$31.00 ©2017 IEEE

Fig. 4. Current density-voltage (J-V) curves measured for the Fig. 7. J-V curves for the perovskite solar cells with different HTM
perovskite solar cells with different mp-TiO2 spin-coating speed. solution preparation time.

Fig. 5. J-V curves for the perovskite solar cells with different PbI2 Fig. 8. J-V curves for the perovskite solar cells with/without HTM
spin-coating speed. oxidation in ambient atmosphere.

Fig. 6. J-V curves for the perovskite solar cells with different HTM Fig. 9. J-V curve for the champion cell measured under AM 1.5G
spin-coating speed. solar irradiation of 100 mWcm-2 with scan-rate fixed at 20 mV s-1.

978-1-5090-5605-7/17/$31.00 ©2017 IEEE

 IV. CONCLUSION spin coating speed, HTM spin coating speed, and HTM
exposure to air etc. We found that the higher spin speed of PbI2
We have fabricated CH3NH3PbI3 perovskite solar cells using
is critical for high-quality PbI2 film formation. The HTM
RF sputtered compact TiO2 layer and modified two-step
exposure to air is also important to fabricate solar cells with
solution processes in the ambient air. We carried out
larger current density and better fill factors.
comprehensive study of the two-step solution process
parameters including PbI2 spin coating speed, mesoporous TiO2
[7] J. H. Yun, I. Lee, T.-S. Kim, M. J. Ko, J. Y. Kim, and H. J. Son,
“Synergistic enhancement and mechanism study of mechanical
REFERENCES and moisture stability of perovskite solar cells introducing
polyethylene-imine into the CH 3 NH 3 PbI 3/HTM interface,”
[1] A. Kojima, K. Teshima, Y. Shirai, and T. Miyasaka, Journal of Materials Chemistry A, vol. 3, no. 44, pp. 22176-
“Organometal halide perovskites as visible-light sensitizers for 22182, 2015.
photovoltaic cells,” Journal of the American Chemical Society, [8] Y. Xu, L. Zhu, J. Shi, X. Xu, J. Xiao, J. Dong, H. Wu, Y. Luo, D.
vol. 131, no. 17, pp. 6050-6051, 2009. Li, and Q. Meng, “The Effect of Humidity upon the
[2] J.-H. Im, C.-R. Lee, J.-W. Lee, S.-W. Park, and N.-G. Park, “6.5% Crystallization Process of Two-Step Spin-Coated Organic–
efficient perovskite quantum-dot-sensitized solar cell,” Inorganic Perovskites,” ChemPhysChem, 2015.
Nanoscale, vol. 3, no. 10, pp. 4088-4093, 2011. [9] H.-S. Ko, J.-W. Lee, and N.-G. Park, “15.76% efficiency
[3] J. Burschka, N. Pellet, S.-J. Moon, R. Humphry-Baker, P. Gao, perovskite solar cells prepared under high relative humidity:
M. K. Nazeeruddin, and M. Grätzel, “Sequential deposition as a importance of PbI 2 morphology in two-step deposition of CH 3
route to high-performance perovskite-sensitized solar cells,” NH 3 PbI 3,” Journal of Materials Chemistry A, vol. 3, no. 16,
Nature, vol. 499, no. 7458, pp. 316-319, 2013. pp. 8808-8815, 2015.
[4] W. S. Yang, J. H. Noh, N. J. Jeon, Y. C. Kim, S. Ryu, J. Seo, and [10] J. Yin, J. Cao, X. He, S. Yuan, S. Sun, J. Li, N. Zheng, and L. Lin,
S. I. Seok, “High-performance photovoltaic perovskite layers “Improved stability of perovskite solar cells in ambient air by
fabricated through intramolecular exchange,” Science, pp. controlling the mesoporous layer,” Journal of Materials
aaa9272, 2015. Chemistry A, vol. 3, no. 32, pp. 16860-16866, 2015.
[5] D. Bi, W. Tress, M. I. Dar, P. Gao, J. Luo, C. Renevier, K. [11] D. H. Cao, C. C. Stoumpos, C. D. Malliakas, M. J. Katz, O. K.
Schenk, A. Abate, F. Giordano, and J.-P. C. Baena, “Efficient Farha, J. T. Hupp, and M. G. Kanatzidis, “Remnant PbI2, an
luminescent solar cells based on tailored mixed-cation unforeseen necessity in high-efficiency hybrid perovskite-based
perovskites,” Science Advances, vol. 2, no. 1, pp. e1501170, solar cells? a),” Apl Materials, vol. 2, no. 9, pp. 091101, 2014.
2016. [12] S. Wang, W. Yuan, and Y. S. Meng, “Spectrum-Dependent Spiro-
[6] B. Lei, V. O. Eze, and T. Mori, “High-performance CH3NH3PbI3 OMeTAD Oxidization Mechanism in Perovskite Solar Cells,”
perovskite solar cells fabricated under ambient conditions with ACS applied materials & interfaces, vol. 7, no. 44, pp. 24791-
high relative humidity,” Japanese Journal of Applied Physics, 24798, 2015.
vol. 54, no. 10, pp. 100305, 2015.

978-1-5090-5605-7/17/$31.00 ©2017 IEEE