Solar Energy 198 (2020) 454–460

Contents lists available at ScienceDirect

Solar Energy
journal homepage: [Link]/locate/solener

Simulated development and optimized performance of CsPbI3 based all- T

inorganic perovskite solar cells
⁎
Lingyan Lin, Linqin Jiang, Ping Li, Hao Xiong, Zhenjing Kang, Baodian Fan, Yu Qiu
Key Laboratory of Green Perovskites Application of Fujian Province Universities, Fujian Jiangxia University, Fuzhou 350108, China
College of Electronic Information Science, Fujian Jiangxia University, Fuzhou 350108, China

A R T I C LE I N FO A B S T R A C T

Keywords: Cesium lead iodide (CsPbI3), as one promising inorganic halide perovskite for higher stability, has received
Device modeling extensive attention in recent years. However, CsPbI3 based perovskite solar cells (PSCs) have yet to realize the
CsPbI3 high efficiency achieved in organic-inorganic hybrid PSCs. In this work, we perform a device modeling by
All-inorganic perovskite solar cell SCAPS-1D to investigate the limitation of CsPbI3 all-inorganic PSC (CsPbI3 i-PSCs) and improve its performance.
For i-PSC, not only the absorber but all the layers should be composed of inorganic materials only. Therefore,
several potential inorganic hole and electron transport layers (i-HTL and i-ETL) are compared firstly and the
results reveal that Cu2O HTL and SnO2 ETL are the most suitable materials among them. Moreover, the device
performance is further improved by optimizing the work function of back electrode, absorber thickness, doping
density as well as defect density. Under optimized conditions, a conversion efficiency of 21.31% is obtained for
the FTO/SnO2/CsPbI3/Cu2O/Au i-PSC, indicating that there is much room for further performance enhance-
ment.

1. Introduction band gap solar cells (Ahmad et al., 2017; Beal et al., 2016). Since its
first report in 2015 to now, the device efficiency for CsPbI3-based PSC
As a novel third-generation solar cell, hybrid organic-inorganic has enhanced from 2.9% (Eperon et al., 2015) to 19.03% (Zhao et al.,
perovskite solar cells have drawn a great deal of attentions in photo- 2019) together with improved stability, indicating its great potential for
voltaic research due to their simpler fabrication process and lower cost high-efficiency inorganic PSCs. Despite its rapid development, the
compared to conventional crystalline silicon solar cell (Gao et al., conversion efficiency for the CsPbI3 based perovskite solar cells is still
2014), and the conversion efficiency for the PSCs has enhanced from inferior to the organic-inorganic hybrid perovskite solar cells. Com-
3.8% (Kojima et al., 2009) to 25.2% ([Link] paring to the numerous efforts in the research of organic-inorganic
pdfs/best-research-cell-effi[Link]) within the past hybrid PSCs, study of CsPbI3 inorganic perovskite solar cells is still in
decade. However, the volatile organic components induced thermal and the infancy stage. Therefore, further efforts should be made to explore
(photo) chemical instability in hybrid organic-inorganic perovskite can the potential of CsPbI3 solar cells. The ultimate goal is to achieve highly
be an inevitable obstacles for its further commercialization (Leijtens stable all-inorganic PSCs (i-PSCs), which is composed of not only the
et al., 2015). Recently, inorganic halide perovskite obtained by sub- inorganic halide perovskite absorber and also inorganic hole and elec-
stituting the volatile organic components with cesium (Cs) have at- tron transporting layers (i-HTL and i-ETL).
tracted great attentions due to their intrinsic inorganic stability and In order to enhance the solar cell performance, a deep under-
prominent photovoltaic performance (Wang et al., 2016; Liu et al., standing of the internal physical mechanisms is indispensable. Device
2018; Kulbak et al., 2015). For instance, the inorganic halide per- simulation can provide deep insight into the device physics and eluci-
ovskites CsPbX3 (X = Cl, Br, I) were reported to exhibit higher thermal date the relationship between material properties and device perfor-
stability than the organic-inorganic hybrid perovskite (Sutton et al., mance. To the best of our knowledge, there has been no report on
2016; Frolova et al., 2017). Among these inorganic perovskites, CsPbI3 numerical analysis of CsPbI3-based all-inorganic PSCs. Therefore, in this
with the cubic phase exhibits the most suitable bandgap of 1.73 eV for work, we conducted a detailed investigation on the CsPbI3-based i-
photovoltaic applications (Eperon, 2014; Chen et al., 2019), which is PSCs. Firstly, the validity of our simulation is verified through com-
also an ideal material to configure tandem solar cells with other low paring the real experimental results with our simulation results. Then

⁎
Corresponding author at: Key Laboratory of Green Perovskites Application of Fujian Province Universities, Fujian Jiangxia University, Fuzhou 350108, China.
E-mail address: yuqiu@[Link] (Y. Qiu).

[Link]
Received 12 November 2019; Received in revised form 31 December 2019; Accepted 29 January 2020
0038-092X/ © 2020 International Solar Energy Society. Published by Elsevier Ltd. All rights reserved.
L. Lin, et al. Solar Energy 198 (2020) 454–460

Fig. 1. Device configuration of i-PSC used in the simulation.

several inorganic materials are investigated for possible i-HTL and i-

ETL. For the ETL, the proposed materials are: SnO2, ZnO andTiO2 with Fig. 2. J-V characteristic calculated with the physical parameters in Table 1.
The experimental values are taken from Ref. Minemoto and Murata (2014).
widely-used PCBM as the benchmark. CuSCN, CuI and CuO2 are tested
as potential HTL with Spiro-OMeTAD as the benchmark. Furthermore,
we analyze the how the work function of back contact affects the per- and CsPbI3/ETL to simulate a more realistic situation. The physical
formance of PSCs. In addition, the thickness, doping density and defect parameters of IDL1 and IDL2 are set the similar to the absorber except
density of CsPbI3 absorber are also optimized. The research results will with high level defect states. Neutral Gaussian distribution defect with
provide a helpful guidance for designing and fabricating high perfor- characteristic energy of 0.1 eV is selected and is located at mid-gap.
mance all-inorganic PSCs. Captured cross section of electron and hole is 2 × 10−14 cm2. The
absorption constant Aα is set as 105 to obtain absorption coefficient (α)
calculated through α = Aα(hv-Eg)1/2. Work functions of back and front
2. Device simulation parameters
contact are 5.1 eV (Au) and 4.7 eV (FTO), respectively.
The numerical modelling carried out in the work is based on SCAPS-
1D software that developed by Gent University (Burgelman et al., 3. Results and discussion
2000). SCAPS-1D is based on the three coupled differential equations
namely continuity equations for electrons and holes as well as Poisson’s 3.1. Device model verification
equation. Through solving the three equations, it can calculate con-
centrations and currents, Current-Voltage characteristics, energy bands To validate the physical parameters employed in our study, SnO2 as
and spectral response. The device configuration of CsPbI3 based i-PSC ETL and Spiro-OMeTAD as HTL are firstly investigated as a reference.
consists of FTO/i-ETL/CsPbI3/i-HTL/Au, as depicted in Fig. 1. Table 1 The current density-voltage (J-V) curve simulated sing the parameters
lists the primary physical parameters employed in our simulation, in Table 1 is presented in Fig. 2. The device shows a photoconversion
which are collected from published experimental and theoretical stu- efficiency (PCE) of 14.67% with a fill factor (FF) of 75.18%, a short
dies. In the table, NC and NV correspond to effective conduction and current density (JSC) of 18.28 mA/cm2 and an open-circuit voltage
valence band density, Eg denotes band gap, NA and ND denote acceptor (VOC) of 1.07 V, which is quite similar to the reported experimental
and donor density, μp (μn) represents hole (electron) mobility, Nt is the result (Minemoto and Murata, 2014). Therefore, to some extent, the
defect density, χ represents electron affinity and εr represents relative device configuration and the present values of materials are reasonable.
permittivity. In this study, the initial defect density of CsPbI3 absorber Further research can be performed based on this model.
is assumed as 8 × 1019 cm−3, which gives carrier diffusion length of
1 μm, which is a similar value to the reported work (Dastidar et al., 3.2. Effect of the ETL
2017). Thermal velocity of hole and electron is 107 cm/s. To consider
the influence of interface recombination, we insert a very thin virtual In this section, several inorganic materials suggested for ETL are
layer (IDL1 and IDL2, not indicated in the figure) between HTL/CsPbI3 studied. The material parameters of the suggested ETL are exhibited in

Table 1
Material parameters used in the simulation.
Spiro-OMeTAD IDL1 CsPbI3 IDL2 SnO2

Thickness (nm) 170 5 350 5 25

NA (cm−3) 2 × 1018 1015 1015 1015
ND (cm−3) – – – – 1018
εr 3 6 6 6.5 9
χ (eV) 2.45 3.95 3.95 3.95 4
Eg (eV) 3 1.73 1.73 1.73 3.5
μn (cm2/V/s) 2 × 10-4 16 16 16 20
μp (cm2/V/s) 2 × 10-4 16 16 16 10
Nt (cm−3) 1015 1015 2.07 × 1014 1015 1015
NC (cm−3) 2.2 × 1018 1.1 × 1020 1.1 × 1020 1.1 × 1020 4.36 × 1018
NV (cm−3) 1.8 × 1019 8 × 1019 8 × 1019 8 × 1019 2.52 × 1019
Reference Minemoto and Murata (2014) and Dastidar et al. (2017), Wang et al. (2018), and Deng et al. (2016) and Azri et al.
Huang et al. (2016) Hutter and Savenije (2018) (2019)

455
L. Lin, et al. Solar Energy 198 (2020) 454–460

Table 2 Table 3
Material parameters of the proposed ETL. Performance parameters of the PSC with different ETL.
TiO2 ZnO SnO2 PCBM VOC (V) JSC (mA/cm2) FF (%) PCE (%)

Thickness(nm) 25 25 25 25 TiO2 1.064 18.22 75.16 14.57

NA (cm−3) – – – – ZnO 1.065 18.24 75.18 14.62
ND (cm−3) 1018 1018 1018 2.93 × 1017 PCBM 1.061 16.78 74.66 13.29
εr 9 9 9 3.9 SnO2 1.067 18.28 75.18 14.67
χ (eV) 4 4 4 3.9
Eg (eV) 3.2 3.3 3.5 2
μn (cm2/V/s) 20 100 20 0.2
μp (cm2/V/s) 10 25 10 0.2
Nt (cm−3) 1015 1015 1015 1015
NC (cm−3) 1 × 1021 2.2 × 1018 4.36 × 1018 2.5 × 1021
NV (cm−3) 2 × 1020 1.8 × 1019 2.52 × 1019 2.6 × 1021
Reference Deng et al. Azri et al. Azri et al. Azri et al. (2019)
(2016) and (2019) (2019) and
Azri et al. Zhao et al.
(2019) (2019)

Table 2. Fig. 3 presents the J-V curves and the quantum efficiency (QE).
In addition, the output parameters of the PSCs are shown in Table 3. We
can see that the device with PCBM ETL exhibits the lowest efficiency
because of the lowest JSC of 16.78 mA/cm2 and lowest QE compared to
Fig. 4. Band alignment between CsPbI3 and different ETLs.
the other materials, which is due to the light absorption of PCBM be-
cause of its low Eg of 2 eV. Therefore, PCBM is not a preferential ETL for
CsPbI3 based PSCs. SnO2, TiO2 and ZnO present comparable efficiency Table 4
(14.67%, 14.56% and 14.62%, respectively) and a high QE. This can be Material parameters of the proposed HTL.
attributed to the suitable band alignment between the conduction band Cu2O CuSN CuI Spiro-
of SnO2, TiO2 and ZnO and the lowest unoccupied molecular orbital OMeTAD
(LUMO, equal to the conduction band edge in inorganic case) of the
Thickness(nm) 170 170 170 170
CsPbI3 absorber, as presented in Fig. 4. This alignment allows for a good NA (cm−3) 1018 1018 1018 2 × 1018
transport of electrons to the ETL. Instead, the slight band misalignment ND (cm−3) – –
of PCBM ETL hinders the photoelectron transport and accounts for the εr 7.11 10 6.5 3
low photocurrent and FF too. From the simulation results, the device χ (eV) 3.2 1.9 2.1 2.45
Eg (eV) 2.17 3.4 3.1 3
with SnO2 ETL shows the highest PCE, hence the SnO2 is seem to be the
μn (cm2/V/s) 200 2 × 10-4 100 2 × 10-4
best candidate for ETL in the CsPbI3 i-PSCs. μp (cm2/V/s) 80 2 × 10-1 43.9 2 × 10-4
Nt (cm−3) 1015 1014 1015 1015
NC (cm−3) 2.02 × 1017 1.7 × 1019 2.8 × 1019 2.2 × 1018
3.3. Effect of the HTL NV (cm−3) 1.1 × 1019 2.5 × 1021 1 × 1019 1.8 × 1019
Reference Hossain et al. Azri et al. Hossain et al. Minemoto and
(2015) (2019) (2015) and Murata (2014)
In this section several materials are investigated as HTL while using Kanoun et al. and Huang
SnO2 as the ETL. The physical parameters of these HTLs are listed in (2019) et al. (2016)
Table 4. Fig. 5 presents the obtained J-V characteristics and QE with
different HTL. We can find that the QE is not influenced since the HTL is
located at the back side of the solar cells hence the optical absorption of there are good band alignments between the valance band of CsPbI3
the HTL is negligible. In addition, the QE of simulate device can almost absorber and the highest occupied molecular orbital (HOMO) of all the
reach 100% is due to the fact that the reflectance at the surface and HTLs. It is noteworthy that Cu2O exhibits better performance than other
interfaces of each layer is not considered in this simulation to simply i-HTLs. Although organic Spiro-OMeTAD still shows the optimal band
the model. The energy level diagrams depicted in Fig. 6 indicate that alignment, the holes mobility in Cu2O is much larger than in Spiro-

Fig. 3. J-V curves and QE for CsPbI3 based PSCs with different ETL layer.

456
L. Lin, et al. Solar Energy 198 (2020) 454–460

Fig. 5. J-V curves and QE for CsPbI3 based PSCs with different HTL layer.

inefficient collection of photo-generated carriers. Moreover, the S-

shaped J-V curves can be observed when φ is with low values, which
might be due to the high Schottky barriers at the HTL/back-contact
interface that hinders hole transfer from HTL to the back electrode, thus
decreasing the fill factor. Therefore, it is suggested that back contact
with φ higher than 4.9 eV, such as base metals Ni and Cu and noble
metals Au and Pt, are suitable electrodes for PSCs. Within them, copper
is an excellent compromised choice between conductivity and price.

3.5. Effect of absorber thickness

The CsPbI3 absorber plays a significant part in solar cell perfor-

Fig. 6. Band alignment between CsPbI3 and different HTLs. mance since it absorbs light to generate photo-generated carriers. It
may be possible to further enhance device efficiency by increasing the
Table 5 CsPbI3 absorber thickness. In this section, the variations of the PSC
Performance parameters of the PSC with different HTL. performance with CsPbI3 thickness changing from 100 to 1000 nm are
examined. Fig. 8(a) shows that when the CsPbI3 thickness increases
VOC (V) JSC (mA/cm2) FF (%) PCE (%)
from 100 to 700 nm the PCE increases rapidly. And it almost saturated
Spiro-OMeTAD 1.067 18.28 75.18 14.67 when the thickness exceeded 700 nm. The variation tendency is con-
Cu2O 1.065 18.41 81.73 16.03 sistent with the QE results. As illustrated in Fig. 9(b), the QE of PSC is
CuSCN 1.087 18.28 76.34 15.18
enhanced remarkably when CsPbI3 thickness is increased until 700 nm,
CuI 1.087 18.29 79.44 15.79
which is because of the increased photon absorption. Then the QE stays
constant when absorber thickness is larger than 700 nm. Therefore, in
this study, we set 700 nm as the optimized value of the CsPbI3 absorber
OMeTAD. In addition, in case of spiro-OMeTAD as HTL, the kink at VOC thickness for material saving.
is mainly due to the low carrier mobility of spiro-OMeTAD in the si-
mulation. When using Cu2O as HTL the device exhibits a high FF 3.6. Effect of absorber doping density and defect density
81.73% and a PCE of 16.03% as illustrated in Table 5. Therefore, Cu2O
can be considered as a good option for i-HTL. In the following simu- It was reported that CsPbI3 exhibited a unipolar self-doping beha-
lations, the SnO2 is used as the ETL and Cu2O as the HTL. vior (p-type conductivity), which was different from CH3NH3PbI3
(Huang et al., 2018). In the present investigation, the acceptor density
(NA) in CsPbI3 is varied from 1013 to 1018 cm−3. The J-V curves and
3.4. Effect of back contact work function PCE with varying NA are exhibited in Fig. 9. As can be seen, the J-V
curves are almost identical under low acceptor density. However, when
The back contact work function (φ) is a significant factor to obtain a the NA is larger than 1016 cm−3, the JSC decreases and VOC increases.
reasonable built-in voltage (Vbi) in PSCs. In this section, the influence of This phenomenon can be explained by the energy band diagrams de-
φ on solar cell performance is analyzed. The energy band diagrams of picted in Fig. 10. When NA is at a low value of 1013 cm−3, CsPbI3 is
the CsPbI3 PSC as a function of different φ are presented in Fig. 7. As is completely depleted and a high electric field is formed across the ab-
seen, the band diagrams are overlapped as φ is between 4.9 and 5.1 eV sorber, which is benefit for the separation and collection of light-gen-
and a high Vbi is obtained. In addition, the Vbi monotonically drops with erated carriers, hence leading to a larger Jsc. However, under a high NA
decreasing of φ from 4.9 to 4.5 eV, which results in a reduction in the such as1018 cm−3, the depleted region is shrunken and the electric field
electric field across the absorber. In addition, the Schottky barrier for in the bulk is compressed to the interface of CsPbI3/ETL and CsPbI3/
hole increases with decrease of φ, as illustrated by the dotted rectangle ETL, as depicted in Fig. 10 (b). Therefore, resulting in the decrease of
box. The J-V characteristics as a function of φ values are exhibited in carrier collection probability and hence the reduction of Jsc. The en-
Fig. 7(b). It can be observed that the J-V curves are overlapped when hancement of Voc can be explained by Eqs. (1) and (2):
the φ is from 4.9 to 5.1 eV, which is consistent with the band structures
Dn Dp
in Fig. 7(a). VOC monotonically drops with decreasing φ from 4.9 to I0 = qni2 ( + )
4.5 eV due to the decline in Vbi across the absorber that results in Ln NA Lp ND (1)

457
L. Lin, et al. Solar Energy 198 (2020) 454–460

Fig. 7. Energy band diagrams (a) and J-V characteristics (b) with varied back contact work function.

Fig. 8. The variation of PCE (a) and QE (b) with different absorber thickness.

Fig. 9. (a) J-V curves and (b) PCE as a function of NA.

kT I In addition, the influence of absorber defect density Nt is also stu-

Voc = ln( L + 1)
q I0 (2) died. Fig. 11 shows the relationship between PCE and different defect
density. As can be seen, when Nt is lower than 1012 cm−3, the PCE
From the above Equations, we can find that with the increase of NA, shows few changes. However, when Nt goes beyond 1017 cm−3, the PCE
the saturation current I0 will decrease, enhancing the VOC. Since with exhibits a large decline. Therefore, controlling the Nt under ~1012
increasing NA, the JSC decreases while VOC increases, there will exist an cm−3 is a critical factor in achieving the high-efficiency CsPbI3 based
optimum NA that maximizes the device efficiency. As is seen from PSCs. At this Nt value, optimized PCE of 21.31% can be achieved for
Fig. 9(b), a maximum conversion efficiency of 17.78% can be obtained FTO/SnO2/CsPbI3/Cu2O/Cu all-inorganic PSCs.
when NA is 1016 cm−3.

458
L. Lin, et al. Solar Energy 198 (2020) 454–460

Fig. 10. Energy band diagrams of the device with different doping density (a) NA = 1013 cm−3 and (b) NA = 1018 cm−3.

Declaration of Competing Interest

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.

Acknowledgement

This work is financially supported by Natural Science Foundation of

Fujian Province (2016J01752, 2018J01642, 2019J01877), National
Natural Science Foundation of China (51402050), Educational
Department of Fujian Province of China (JAT190470), Cultivation Fund
of Young Scientific Research Talents of Fujian Jiangxia University
(JXZ2018006) and Support Program for Innovative Science Research
Group of Fujian Jiangxia University.

References

Fig. 11. PCE as a function of varied absorber defect density.

Ahmad, W., Khan, J., Niu, G., Tang, J., 2017. Inorganic CsPbI3 perovskite-based solar
cells: a choice for a tandem device. Sol. RRL 00, 1700048.
Azri, F., Meftah, A., Sengouga, N., Meftah, A., 2019. Electron and hole transport layers
4. Conclusion
optimization by numerical simulation of a perovskite solar cell. Sol. Energy 181,
372–378.
In this study, the CsPbI3 absorber based all-inorganic PSC is nu- Beal, R.E., Slotcavage, D.J., Leijtens, T., Bowring, A.R., Belisle, R.A., Nguyen, W.H.,
merical studied and optimized by SCAPS-1D software package. Firstly, Burkhard, G.F., Hoke, E.T., McGehee, M.D., 2016. Cesium lead halide perovskites
with improved stability for tandem solar cells. J. Phys. Chem. Lett. 7, 746–751.
several potential inorganic materials for HTL and ETL are examined. Burgelman, M., Nollet, P., Degrave, S., 2000. Modeling polycrystalline semiconductor
According to the simulation results, Cu2O HTL shows better perfor- solar cells. Thin Solid Films 361362, 527532.
mance than other candidates, therefore it is considered as the best Chen, K.Q., Zhong, Q.H., Chen, W., Sang, B.H., Wang, Y.W., Yang, T.Q., Liu, Y.L., Zhang,
Y.P., Zhang, H., 2019. Short-chain ligand-passivated stable α-CsPbI3 quantum dot for
candidate for HTL in CsPbI3 i-PSCs. For ETLs, SnO2, TiO2 and ZnO gives all-inorganic perovskite. Sol. Cells 29, 1900991.
similar cell efficiency when applied as HTLs due to the good bands Dastidar, Subham, Li, Siming, Smolin, Sergey Y., Baxter, Jason B., Fafarman, Aaron T.,
alignment between the LUMO of the CsPbI3 and the conduction band of 2017. Slow electron−hole recombination in lead iodide perovskites does not require
a molecular dipole. ACS Energy Lett. 2, 2239–2244.
SnO2, ZnO and TiO2. In addition, further optimization of the CsPbI3 i- Deng, Q.R., Li, Y.Q., Chen, L.A., Wang, S.G., Wang, G.M., Sheng, Y.L., Shao, G.S., 2016.
PSCs is conducted based on the Cu2O HTL and SnO2 ETL. The effect of The effects of electron and hole transport layer with the electrode work function on
the back contact work function, the absorber thickness, doping density perovskite solar cells. Mod. Phys. Lett. B 1650341.
Eperon, G.E., et al., 2014. Formamidinium lead trihalide: a broadly tunable perovskite for
as well as the defect density of absorber are analyzed. Simulation re- efficient planar heterojunction solar cells. Energy Environ. Sci. 7, 982–988.
sults reveal that metals with work function larger than 4.9 eV are sui- Eperon, G.E., Paternò, G.M., Sutton, R.J., Zampetti, A., Haghighirad, A.A., Cacialli, F.,
table back electrodes for CsPbI3 i-PSCs. The optimized absorber thick- Snaith, H.J., 2015. Inorganic cesium lead iodide perovskite solar cells. J. Mater.
ness and doping concentration is 700 nm and 1016 cm−3, respectively. Chem. A. 3, 19688–19695.
Frolova, L.A., Anokhin, D.V., Piryazev, A.A., Luchkin, S.Y., Dremova, N.N., Stevenson,
Defect density of absorber no higher than 1012 cm−3 is essential for K.J., Troshin, P.A., 2017. Highly efficient all inorganic planar heterojunction per-
good device performance. After comprehensive optimization, an op- ovskite solar cells produced by thermal coevaporation of CsI and PbI2. J. Phys. Chem.
timum efficiency of 21.31% for CsPbI3 based i-PSCs can be obtained. Lett. 8, 67–72.
Gao, P., Grätzel, M., Nazeeruddin, M.K., 2014. Organohalide lead perovskites for pho-
The simulation results are helpful for further understanding the me- tovoltaic applications. Energy Environ. Sci. 7, 2448–2463.
chanism and improving the performance of CsPbI3 based all-inorganic Hossain, Mohammad I., Alharbi, [Link].H., Tabet, Nouar, 2015. Copper oxide as in-
perovskite solar cells. organic hole transport material for lead halide perovskite based solar cells. Sol.
Energy 120, 370–380.
Huang, L.K., Sun, X.X., Li, C., Xu, R., Xu, J., et al., 2016. Electron transport layer-free
planar perovskite solar cells: further performance enhancement perspective from
device simulation. Sol. Energy Mater. Sol. Cells 157, 1038–1047.
Huang, Y., Yin, W.J., He, Y., 2018. Intrinsic point defects in inorganic cesium lead iodide

459
L. Lin, et al. Solar Energy 198 (2020) 454–460

perovskite CsPbI3. J. Phys. Chem. C. 1222, 1345–1350. 116, 054505.

Hutter, Eline M., Savenije, Tom J., 2018. Thermally activated second-order recombina- Sutton, R.J., Eperon, G.E., Miranda, L., Parrott, E.S., Kamino, B.A., Patel, J.B., Horantner,
tion hints toward indirect recombination in fully inorganic CsPbI3 perovskites. ACS M.T., Johnston, M.B., Haghighirad, A.A., Moore, D.T., Snaith, H.J., 2016. Bandgap-
Energy Lett. 3, 2068–2069. tunable cesium lead halide perovskites with high thermal stability for efficient solar
Kanoun, Ahmed-Ali, Kanoun, Mohammed Benali, Merad, Abdelkrim E., Goumri-Saidc, cells. Adv. Energy Mater. 6, 1502458.
Souraya, 2019. Toward development of high-performance perovskite solar cells Wang, Q., Dong, Q.F., Li, T., Gruverman, A., Huang, J.S., 2016. Thin insulating tunneling
based on CH3NH3GeI3 using computational approach. Sol. Energy 182, 237–244. contacts for efficient and water-resistant perovskite solar cells. Adv. Mater. 28,
Kojima, A., Teshima, K., Shirai, Y., Miyasaka, T., 2009. J. Am. Chem. Soc. 131, 6050. 6734–6739.
Kulbak, M., Cahen, D., Hodes, G., 2015. How important is the organic part of lead halide Wang, P.Y., Zhang, X.W., Zhou, Y.Q., Jiang, Q., Ye, Q.F., Chu, Z.M., Li, X.X., Yang, X.L.,
perovskite photovoltaic cells? Efficient CsPbBr 3 cells. J. Phys. Chem. lett. 6, Yin, Z.G., You, J.B., 2018. Solvent-controlled growth of inorganic perovskite films in
2452–2456. dry environment for efficient and stable solar cells. Nat. Commun. 9, 2225.
Leijtens, T., Eperon, G.E., Noel, N.K., Habisreutinger, S.N., Petrozza, A., Snaith, H.J., Zhao, P., Lin, Z.H., Wang, J.P., Yue, M., Su, J., Zhang, J.C., Chang, J.J., Hao, Y., 2019.
2015. Stability of metal halide perovskite solar cells. Adv. Energy Mater. 5, 1500963. Numerical simulation of planar heterojunction perovskite solar cells based on SnO2
Liu, C., Li, W.Z., Zhang, C.L., Ma, Y.P., Fan, J.D., Mai, Y.H., 2018. All-Inorganic CsPbI2Br electron transport layer. ACS Appl. Energy Mater. 2, 4504–4512.
perovskite solar cells with high efficiency exceeding 13%. J. Am. Chem. Soc. 140, Zhao, Y., Wang, Y., Liu, X., Zhang, T., Wang, X., Kan, M., Shi, J., Zhao, Y., 2019. The role
3825–3828. of dimethylammonium iodine in CsPbI3 perovskite fabrication: additive or dopant?
Minemoto, T., Murata, M., 2014. Device modeling of perovskite solar cells based on Angew. Chem. 131, 16844–16849.
structural similarity with thin film inorganic semiconductor solar cells. J. Appl. Phys.

460