2

www.advancedsciencenews.com www.solar-rrl.
com
FULL PAPER
Incorporation of High-Mobility and
Room-Temperature-Deposited CuxS
as a Hole Transport Layer for
Efficient and Stable Organo-Lead
Halide Perovskite Solar Cells
Hongwei Lei, Guang Yang, Xiaolu Zheng, Zhi-Guo Zhang,* Cong Chen, Junjie Ma,
Yaxiong Guo, Zhiliang Chen, Pingli Qin, Yongfang Li, and Guojia Fang*
1. Introduction
The most widely used hole transport layer (HTL) for n-i-p perovskite solar
Clean, abundant, and renewable solar
cells (PVSCs), 2,20 ,7,70 -tetrakis(N,N-di-p-methoxyphenylamine)-9,90 -spirobi-
energy is one of the most promising energy
fluorene (spiro-OMeTAD), suffers a severe degradation from the permeation sources to meet the increasing energy
of moisture and direct contact with the metal electrode, mainly due to demand and solve the environmental prob-
hygroscopic additives and the presence of pinholes in the spiro-OMeTAD lems caused by the traditional fossil fuels.
film. To overcome this problem, we developed a stable inorganic copper- Photovoltaic solar cells directly convert solar
energy to electricity. The solution-processed,
based chalcogenide (CuxS, x ¼ 1.75) that can cooperate with spiro to serve as
inorganic–organic hybrid perovskite solar
the HTL for planar n-i-p PVSCs. The CuxS HTL has two main functions: 1) cells (PVSCs) have attracted much attention
enhancing hole transport due to its high intrinsic mobility and proper energy because of their easy fabrication, low
level alignment, resulting in a better charge transfer and reduced charge cost, high efficiency, light weight, and
recombination; 2) protecting the spiro layer from damage from both moisture flexibility.[1–7] The light absorbers for these
and the top Au anode, through the formation of a physical hydrophobic photovoltaic devices are methylammonium
lead trihalide perovskites, CH3NH3PbX3,
buffer layer. PVSCs with enhanced power conversion efficiencies (PCEs) are
where X is a halogen (I, Br, or Cl). Perovskite
realized through this simple approach, yielding the highest PCE of 18.58% materials have the advantages of long
and a steady-state PCE of 17.91%. Furthermore, benefiting from the exciton diffusion lengths, excellent charge-
hydrophobic nature of CuxS, PVSCs retained over 90% of their initial carrier transport properties, high absorption
efficiency, even after storage in air with approximately 40% humidity for coefficients, and tunable compositions and
structures.[8–10] These materials are emerg-
1000 h without encapsulation. This study demonstrates that CuxS is a
ing as the new generation of solution-
potential hole transport material for fabricating low-cost and efficient PVSCs processable and low-cost photovoltaic mate-
with long-term stability. rials. Since the first report by Miyasaka et al.
of photovoltaic performance, achieved by
using CH3NH3PbI3 nanoparticles to sensitize a mesoporous TiO2
anode,[11] efficient solid-state, meso-structured and planar perov-
skite hybrid solar cells have been realized. During the past 6 years
G. Fang, H. Lei, G. Yang, X. Zheng, C. Chen, J. Ma, tremendous improvements have been seen in the power
Y. Guo, Z. Chen, P. Qin conversion efficiency (PCE) of PVSCs. To date, both mesoscopic
Key Lab of Artificial Micro- and Nano-Structures of and planar PVSCs have achieved 16–20% PCE,[12–19] with the state-
Ministry of Education of China, Department of
of-the-art mesoscopic devices reaching a certified PCE of 22.1%.[20]
Electronic Science and Technology, School of Physics
and Technology, Wuhan University, Wuhan 430072, PVSCs were first implemented in mesoscopic dye-sensitized
China solar cells (DSSCs). However, a processing temperature of up to
E-mail: [email protected] 500 C to sinter the TiO2 scaffold is required for mesoscopic
Z.-G. Zhang, H. Lei, Y. Li DSSC fabrication, which is incompatible with flexible substrates.
Beijing National Laboratory for Molecular Sciences, Considering their long exciton diffusion lengths, perovskites can
CAS Key Laboratory of Organic Solids, Institute of
Chemistry, Chinese Academy of Sciences, Beijing also be applied to planar heterojunction solar cells. In general,
100190, China the regular device architecture of n-i-p CH3NH3PbI3 (MAPbI3)
E-mail: [email protected] PVSCs consists of a transparent conducting oxide (TCO)
DOI: 10.1002/solr.201700038 cathode, an electron transport layer (ETL), a perovskite absorber,
Sol. RRL 2017, 1, 1700038 © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim wileyonlinelibrary.com (1 of 12) 1700038
2367198x, 2017, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/solr.201700038 by KARAMANOGLU MEHMET BEY UNIVERSITY KUTUPHANE VE, Wiley Online Library on [10/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
www.solar-rrl.com www.advancedsciencenews.com
FULL PAPER
a hole transport layer (HTL), and a metal anode. As is well device stability owing to the coverage of pinholes in the spiro
known, interface engineering, by introducing interlayers layer and the high hydrophobicity of CuxS. Benefitting from the
between the electrode and active layer, can enable us to tune good hole transport properties of CuxS and the barrier protection
the electrode work function, enhance the carrier separation and from the insertion of CuxS between the spiro layer and the Au
collection efficiency, and minimize carrier trapping or accumu- electrode, the device with a configuration of FTO/SnO2/
lation, which are key to optimizing the solar cell performance. MAPbI3/spiro-OMeTAD/CuxS/Au produced a high PCE of
Therefore, both the ETL and HTL are vital for high-performing 18.58% and demonstrated enhanced stability, both of which
PVSCs. At this stage, TiO2 and spiro-OMeTAD are the most are superior to those of spiro-OMeTAD-only devices.
widely used ETL and HTL, respectively. Recently, SnO2 has been
proven to be a good ETL candidate due to its wide band-gap, high
electron mobility, and low-temperature processing proper- 2. Results and Discussion
ties.[21,22] Meanwhile, spiro-OMeTAD is the state-of-the-art
HTL material employed for conventional n-i-p PVSCs. 2.1. Vapor Deposition and Characterization of p-Type CuxS
However, spiro-OMeTAD HTLs usually show a low hole mobility
(105–104 cm2 V1 s1), even with the incorporation of addi- CuxS is one of the most important p-type semiconductors due to
tives such as Li-bis(trifluoromethanesulfonyl)-imide (Li-TFSI) its superior properties. Various mineral forms of CuxS (1 x 2)
and tert-butylpyridine (tBP).[23,24] Moreover, the dopants used in exist, in which the Cu atoms change as a function of x. The
organic HTLs are hydrophilic, which may compromise the copper sulfide family is composed of various solid-phase
device stability, considering that perovskites are susceptible to materials, such as chalcocite Cu2S, djurleite Cu1.96S, digenite
moisture. During the oxidation of the HTL, pinholes, caused by Cu1.8S, anilite Cu1.75S, covellite CuS, and others.[37–40] All of
the migration and redistribution of additives, may form in spiro- these phases are p-type semiconductors owing to Cu vacancies
OMeTAD films.[25,26] These pinholes may allow moisture to within the lattice. Our previous work demonstrated the in situ
penetrate, as well as allow a direct contact between the Au top hydrothermally fabricated divalent CuS to be an alternative hole
electrode and the perovskite layer. To fabricate highly efficient transport and interface modification layer in organic solar cells
and stable PVSCs, it is imperative to develop an HTL with a due to its high mobility, high work function, and suitable band
dense and stable morphology, a suitable energy level alignment alignment. As a preliminary examination, we introduced a p-type
between the perovskite absorber and the top electrode, a high CuxS to the conventional planar PVSCs. Vacuum vapor
charge-carrier mobility, and hydrophobicity. Considering the deposition can avoid the damage to the organic layers caused
above-mentioned problems, high-mobility stable inorganic by the spin-casting of the solution. Thus, in our study, CuxS was
semiconductors with proper band alignment show promising prepared by the thermal evaporation of CuS powder under high
applications. Copper-based semiconductors have gradually been vacuum without any additional high-temperature annealing.
employed as efficient HTLs for PVSCs. In 2014, CuI was first Figure 1(a) presents the optical properties of vapor-deposited
introduced as an inorganic HTL for PVSCs, and a 6% efficiency CuxS. The transmittance of CuxS was determined from a
was achieved.[27] Later, reports showed that CuSCN, CuPc, and relatively thick CuxS sample on a quartz glass. The thickness of
CZTS nanoparticles or films can also be good alternative the CuxS film can be easily tuned by varying the thermal
inorganic HTLs for PVSCs.[28–31] Up to now, a 16.6% PCE for evaporation time and the weight of the source powder. Since
PVSCs with n-i-p structures was reported using an inorganic both the spiro-OMeTAD and the CuxS were on top of a
CuSCN HTL.[32] Recently, our group reported a facile molecu- perovskite, their transparencies are not important, owing to the
larly engineered copper (II) phthalocyanine as a hole transport strong absorption of light by the perovskite layer. For the direct
material for planar PVSCs, achieving an efficiency of 15.73%.[33] band-gap semiconductors, we confirmed the optical band-gap
We believed that there was still room for improvement in terms (Eg) of CuxS according to the dependence of (ahy)2 on hy. The
of the PCE. Additionally, the efficiency and stability of the PVSCs variation of (ahy)2 versus hy calculated from the transparency
can also be improved by the insertion of an inorganic material, spectrum is shown in the inset of Figure 1(a), and Eg was
such as MoO3, to form an organic/inorganic hybrid bilayer evaluated by extending the linear portion of the plots to the
HTL.[34] Copper sulfide (CuxS) was proven in our previous work binding energy axis. The calculated optical band-gap of thermally
to be a superior HTL and electrode interface modification layer evaporated CuxS is 2.02 eV, which is in the reasonable range
in polymer solar cells.[35,36] CuxS shows a considerable hole compared with the values given in the literature (1–2.48 eV).[37,41]
mobility and a proper energy-band position. It is natural to A typical X-ray diffraction (XRD) pattern of the CuxS film is
consider its application in PVSCs with n-i-p structures. shown in Figure 1(b). The diffraction peaks at 13.3, 26.7, and
Herein, we report a novel high-mobility, room-temperature- 40.6 are assigned to the (1 0 1), (1 2 1), and (2 2 3) planes of the
deposited CuxS (x ¼ 1.75) polycrystalline film for use as a robust CuxS materials, respectively, which indicates that the synthe-
interfacial functional material for PVSCs. We introduced the sized CuxS is in the orthorhombic phase, from the JCPDS card
inorganic CuxS polycrystalline film on the surface of spiro- (Cu1.75S 24-0058). This fact confirms that the prepared films are
OMeTAD to form a hybrid bilayer HTL, and this HTL was proven anilite Cu1.75S. The XRD result suggests that our film was well
to perform better than the state-of-the-art HTL spiro-OMeTAD in crystallized. The TEM image presented in Figure 1(c) shows the
terms of device efficiency and long-term air stability. This bilayer aggregation of CuxS nanoparticles. Additional TEM images of
HTL shows a higher hole mobility owing to the high p-type CuxS with different scale bars are presented in Supporting
conductivity and mobility of CuxS. This bilayer HTL also Information Figure S1. The inset gives the polycrystalline
provided a better protection of the perovskite and an improved electron diffraction rings indexed according to the orthorhombic
1700038 (2 of 12) wileyonlinelibrary.com © 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Sol. RRL 2017, 1, 1700038
www.advancedsciencenews.com www.solar-rrl.com
FULL PAPER
Figure 1. (a) Transmission spectrum of the as-deposited CuxS film on quartz glass; the inset shows the calculated optical band-gap of CuxS. (b) Typical
XRD pattern of the room-temperature-deposited CuxS film. (c) TEM and SAED image of the CuxS polycrystalline film. (d) High-resolution TEM image of
the CuxS polycrystalline film and the lattice fringe attributed to the (1 1 3) plane of anilite Cu1.75S.
CuxS structure, confirming its polycrystalline nature. The samples was covered with an overlayer of cupric oxide or copper
HRTEM image is presented in Figure 1(d), from which we hydroxide species, which may affect the actual optical band-gap
calculated the interplanar spacing. The lattice distance d is as well as the environmental stability of the CuxS sample. It is
0.29 nm, attributed to the (1 1 3) crystalline plane of anilite believed that even surface overlayer species, such as cupric oxide,
Cu1.75S. are promising for hole transport, and thus, this layer is thought
Next, we examined the chemical composition of the vapor- not to be harmful, for devices at least. We then analyzed the S 2p
deposited film. Figure 2(a)–(c) presents the XPS core-shell scan shown in Figure 2(c). Peak fitting for the S 2p line shapes
spectra of CuxS. The proposed reaction mechanism involves was performed using a Gaussian–Lorentzian function with a
the copper sulfide powder first decomposing into non- spin orbit splitting of 1.2 eV while maintaining an intensity ratio
stoichiometric copper sulfide, and then, the non-stoichiometric of S 2p3/2 to S 2p1/2 of 2:1. Two well-defined peaks at 161.8 and
product is melted, sublimated, nucleated, and grown to form a 163 eV were clearly observed, which agreed well with the S 2p3/2
smooth and dense film. Elemental sulfur vapor is inevitably and S 2p1/2 binding energies reported for sulfur in anilite
generated during the evaporation process, which can be quickly Cu1.75S. No sulfate species were detected in the film. Thus, our
dislodged by a pump or tuning the baffle to prevent its contact films are mainly composed of Cu1.75S with a surface oxide layer.
with the substrate. Figure 2(a) shows the survey of XPS scan, in For convenience, our films are referred to as CuxS in the context.
which we find the signals of elements Cu and S. As shown in To elucidate the energy level of CuxS, we performed UPS, and
Figure 2(b), peak fitting of the Cu 2p scan indicated two well- the result is shown in Figure 2(d). Helium Ia was used as the
defined peaks, Cu 2p3/2 and Cu 2p1/2, which were located at ultraviolet source. The wavelength of the ultraviolet source was
binding energies of 932.4 and 952.1 eV, respectively. These peaks 58.13 nm, and the energy was 21.22 eV. Thus, the work function
were attributed to the monovalent state of copper in anilite was calculated by the equation EF ¼ 21.22 EB, where EF is the
Cu1.75S.[42–44] Note that there exists a satellite peak at a binding Fermi level and EB is the high binding energy cutoff. The valence
energy of 946 eV. This peak indicates that the surface of our band maximum (EVBM) with respect to EF can be determined by
FULL PAPER
Figure 2. XPS spectra of the CuxS film: (a) a whole survey of the XPS, (b) Cu 2p scan, and (c) S 2p scan. (d) UPS spectrum of the CuxS film. Inset is the
magnified image of the low binding energy onset.
calculating the low binding energy onset. The position of the one-step solution method described by Park et al. The typical
conduction band minimum (ECBM) was determined by adding XRD pattern of the CH3NH3PbI3 perovskite film on the FTO
the optical band-gap. The work function of CuxS is 4.81 eV, and substrate is shown in Figure 3(c). The (1 1 0), (2 2 0), (2 0 2)
the difference between EF and EVBM is 0.39 eV. Thus, we obtained planes were included. The strong peaks of the films were well
the energy levels of CuxS, whose EVBM matches well with spiro indexed to the perovskite structure. A weak PbI2 peak was found
and perovskite. Another important property is the conductivity of at 12.8 , which is beneficial to passivate the grain boundaries of
the HTL. Hall measurement demonstrated the p-type conduc- perovskite. The other peaks marked with asterisks belong to the
tivity of CuxS, with a carrier density of 1.37 1015 cm3, a FTO glass. Note that CuxS is sometimes used as an absorber in
conductivity of 0.56 S cm1 and an outstanding hole mobility of inorganic heterojunction solar cells, we examined the effect of
4.47 cm2 V1 s1. All of these parameters are superior to spiro- the insertion of CuxS on the incident light absorption of the
OMeTAD and any other organic HTL, which make CuxS a very perovskite absorber. As displayed in Figure 3(d), the CuxS thin
promising candidate for hole transport. film shows very low absorption in the visible range due to the
perovskite film. The spin-casting of spiro or the evaporation of
the thin layer of CuxS do not affect the absorption of the incident
2.2. Device Fabrication and Performance light by the perovskite film. The absorption cut-off of the
perovskite film lies at 750 nm. It should be noted that all the
As a preliminary experiment, we fabricated planar n-i-p PVSCs perovskite samples were prepared on FTO/SnO2 substrates with
via the insertion of CuxS. The schematic device structure is the same thickness.
shown in Figure 3(a), and the relative energy band alignment is Considering the p-type conductivity and suitable hole mobility
illustrated in Figure 3(b). Our perovskite solar cells consisted of of CuxS, we first tested its substitution for the best-performing
FTO-coated glass as the substrate, a solution-processed thin and spiro HTL. The J–V curve of the PVSC using a single layer of
compact SnO2 layer as the ETL, CH3NH3PbI3 as the absorber, CuxS as the HTL is shown in Supporting Information Figure S2.
hybrid bilayer spiro/CuxS as the HTL, and Au as the top Unexpectedly, a single CuxS layer does not provide superior
electrode. The energy band position of SnO2 was obtained in performance over spiro. The origins or reasons behind these
our previous work.[22] The perovskite film was deposited via the unsatisfactory results are as follows. According to the energy
FULL PAPER
Figure 3. (a) Schematic illustration of the device structure of a planar n-i-p MAPbI3 solar cell and (b) energy band diagram of the PVSC. (c) Typical XRD
pattern of the perovskite film examined in our study. (d) Absorption spectra of the CuxS film, perovskite film, and perovskite film with different HTLs.
band study, CuxS was not as effective as spiro in blocking shown in Supporting Information Figure S3. To study the effect
electrons because of its lower conductive band level. Addition- of CuxS on the device performance, we fabricated PVSCs based
ally, copper ions can react with perovskite through a replacement on the spiro/CuxS HTL with various CuxS thicknesses. The
reaction, as illustrated in the literature.[45] Finally, CuxS is too CuxS thickness was recorded by a quartz crystal thickness
conductive to act as an effective HTL. Devices using a CuxS HTL monitor (TM-400, Maxtek, USA). Typical J–V curves of the
achieved a PCE of 4.63%, with an open-circuit voltage (Voc) of planar perovskite photovoltaic devices based on the spiro/
0.65 V, a short-circuit current density (Jsc) of 14.56 mA cm2, and CuxS hybrid HTL with various CuxS thicknesses are shown in
a fill factor (FF) of 48.91%. Further improvement in the single Supporting Information Figure S4, and the parameters are
CuxS HTL device is still necessary to realize the full substitution summarized in Supporting Information Table S1. We found
of spiro. that as we increased the CuxS thickness, the device perfor-
Encouraging results were obtained when the thin layer of mance increased first and then decreased. A precise control of
CuxS cooperated with spiro. Figure 4(a) shows the top- the CuxS thickness is necessary because thicker CuxS layers
performing J–V curves of the planar perovskite photovoltaic may be too conductive or have rough surfaces, both of which
devices based on spiro or spiro/CuxS hybrid HTLs. PVSCs would cause a slight leakage or recombination. The optimized
based on the spiro HTL yielded a PCE of 17.34%, with a Voc of thickness of CuxS was 20 nm. After optimization, an obvious
1.095 V, a Jsc of 21.76 mA cm2, and an FF of 72.78%, when improvement was observed. Devices using the hybrid spiro/
measured under reverse voltage scanning. We then combined CuxS HTL achieved a PCE of 18.58%, with a Voc of 1.125 V, a Jsc
the state-of-the-art spiro HTL with a thin layer of CuxS. Note of 23.10 mA cm2, and an FF of 71.50%, when measured under
that CuxS has good Ohmic contact with the top Au electrode, the reverse voltage scanning, revealing a much improved Jsc.
and the corresponding contact I–V curve of CuxS and Au is Employing spiro/CuxS as the HTL significantly improves the
FULL PAPER
Figure 4. (a) J–V curves of the best-performing PVSCs based on the spiro/CuxS HTL and the conventional spiro HTL. (b) Corresponding IPCE curves and
the integrated current density of different samples. Steady-state efficiencies of perovskite solar cells using (c) the spiro/CuxS HTL measured at a constant
bias voltage of 0.925 V, and (d) the spiro HTL measured at a constant bias voltage of 0.905 V. Statistics of the (e) current density and (f) PCE for the two
kinds of solar cells.
device performance, particularly the Jsc. All the J–V curves were currents from the IPCE curves are in general lower than that
measured at a scan rate of 0.1 V s1. obtained from the J–V measurements. Devices based on the
Furthermore, to examine the photocurrent enhancement, spiro/CuxS hybrid HTL show higher light response in the visible
incident photon-to-electron conversion efficiency (IPCE) spectra range of 450–760 nm, suggesting that the hybrid bilayer HTL is
were measured and are presented in Figure 4(b). The integrated better at extracting photogenerated carriers or excitons from the
current densities of devices based on the spiro and spiro/ perovskite-metal contact.
CuxS HTLs were 19.31 and 20.73 mA cm2, respectively. The To verify that the hybrid spiro/CuxS is a versatile HTL for
integrated current values were different from those found in the PVSCs, devices based on a dense TiO2 ETL were also fabricated.
J–V curves because of the different measurement conditions of Enhanced short current densities and PCEs were also observed
the IPCE and J–V tests. The IPCE measurements were when CuxS was introduced in the interface between the spiro
conducted without an AM1.5G light bias. Thus, the integrated layer and the Au electrode. The corresponding representative
FULL PAPER
J–V curves of the planar perovskite photovoltaic devices are of spiro, as determined from the Hall measurement. To
shown in Supporting Information Figure S5, and the parameters investigate the hole extraction ability of spiro and spiro/CuxS,
are summarized in Supporting Information Table S2. Devices we performed steady-state PL measurement. All the perovskite
using dense TiO2 as the ETL and spiro as the HTL achieved a samples were prepared on FTO/SnO2 substrates with the same
PCE of 10.8%, with a Voc of 1 V, a Jsc of 17.97 mA cm2, and an FF thickness. As presented in Figure 5(a), the pristine perovskite
of 60.11%. For comparison, devices using dense TiO2 as the ETL film showed a strong luminescence peak at 760 nm, suggesting
and spiro/CuxS as the HTL achieved a PCE of 12.53%, with a Voc severe recombination at the interface of the perovskite and
of 1.01 V, a Jsc of 19.12 mA cm2, and an FF of 64.90%. These electrode or in the perovskite film. When we introduced the spiro
findings proved that the insertion of CuxS led to the HTL on the perovskite layer, the PL peak intensity clearly
improvement of the device performance, especially the short decayed, which indicated that carrier recombination was
current density, and this is a simple and efficient approach to suppressed at the interface or in the perovskite film. Remarkably,
further boost the efficiency of PVSCs. At this stage, planar the sample based on the spiro/CuxS HTL showed nearly no PL
devices still suffer from slight hysteresis, as shown in peak in the visible range. The results further confirmed that the
Supporting Information Figure S6 and Table S3, even when hybrid spiro/CuxS HTL was more efficient at reducing charge
we applied a thin layer of CuxS to ease the charge transport and recombination and extracting charge carriers to the metal
suppress charge accumulation at the interface or in the contact than spiro, which is the largest reason for the Jsc
perovskite films. Further work to eliminate this hysteresis is improvement. By using the space-charge-limited current (SCLC)
ongoing. model, as shown in Figure 5(b), we obtained the hole mobilities
To confirm the device reliability of the fabricated PVSCs, the of the spiro and spiro/CuxS films. The results revealed that the
stabilized photocurrents measured at constant biases of 0.925 hybrid HTL processes a mobility of 3.3 103 cm2 V1 s1,
and 0.905 V near the maximum power point for the spiro/ which is almost twice that of the spiro film. The higher mobility
CuxS and spiro devices were examined and are presented in of the HTL improves charge transport across the device, and the
Figure 4(c) and (d), respectively, showing steady-state PCEs of higher injection current from the FTO/SnO2/spiro/CuxS/Au
17.91 and 16.44%. Moreover, to ensure the repeatability of the device compared with the FTO/SnO2/spiro/Au device confirms
results, we fabricated over 30 solar cells from each material, and this argument, as shown in Figure 5(c). The higher injection
the solar cell short-current density and efficiency of the PVSCs current indicated better charge transport properties. Further-
are illustrated in Figure 4(e) and (f). The statistics of the open- more, calculations using the diode model equation, presented in
circuit voltage and FF are shown in Supporting Information Figure 5(d), revealed the lower series resistance of the hybrid
Figure S7. The dispersions of the parameters clearly indicate that HTL device under illumination. The measurement and
spiro/CuxS leads to a remarkable improvement in the perfor- calculation details are given in Section 4. In addition, Nyquist
mance, compared with the control device. The average plots of the perovskite solar cells with the spiro and spiro/
results with standard deviation for the spiro/CuxS devices are CuxS HTLs are presented in Supporting Information Figure S8
a Voc of 1.11 0.02 V, a Jsc of 22.24 0.60 mA cm2, an FF of to further understand the charge transfer and recombination
71.53 1.31%, and a PCE of 17.52 0.82%, which outperforms processes, which indicated that devices based on spiro/
spiro devices, which have a Voc of 1.10 0.02 V, a Jsc of CuxS have a slower recombination rate and better charge
21.34 0.45 mA cm2, an FF of 70.2 2.51%, and a PCE of transfer properties than the control devices based on spiro. All
16.22 1.02%. All these parameters can be found in Table 1. these properties demonstrate the benefits of the insertion of the
p-type CuxS, which resulted in the enhancement of the device
performance.
2.3. Charge Transfer Studies
To understand the cause of the improvement, we studied the 2.4. Device Stability Studies
charge transport properties. Notably, the hole mobility of the
CuxS film was three or four orders of magnitude higher than that The most critical challenge for the implementation of PVSCs in
practical applications lies in their long-term stability, particularly
their degradation during outdoor applications. The moisture
tolerance of perovskite plays a key role in determining the life-
Table 1. Summary of the best-performing and average PCE time of photovoltaic devices. To improve the conductivity and
performances of planar devices based on the spiro and spiro/
mobility of the spiro film, LiTFSI and tBP are generally doped in
CuxS HTLs, obtained from Figure 4.
the precursor solution. However, these dopants or additives
slowly degrade the perovskite, and it has been documented that
Different HTL Voc (V) Jsc (mA cm2) FF (%) PCE (%)
LiTFSI absorbs moisture in the air. Even with a very small
Spiro quantity of LiTFSI scattered on the perovskite film, the
Best 1.095 21.76 72.77 17.34 decomposition of perovskite to yellow PbI2 would be accelerated.
Average 1.10 0.02 21.34 0.45 70.2 2.51 16.22 1.02 Moreover, LiTFSI needs to be oxidized to improve the
Spiro/CuxS conductivity and mobility of the spiro layer, which causes the
migration of LiTFSI to the surface of the film. Thus, pinholes
Best 1.125 23.10 71.50 18.58
form across the film. These pinholes can cause direct contact
Average 1.11 0.02 22.24 0.60 71.53 1.31 17.52 0.82
between the Au and perovskite and form another pathway for
FULL PAPER
Figure 5. (a) Steady-state PL spectra of the perovskite/spiro, perovskite/spiro/CuxS, and bare perovskite films. (b) Dark J–V characteristics of hole-only
devices, showing the effect of the CuxS layer on the hole transport properties. The solid lines of the J–V curves represent the best fit from the SCLC model.
Inset is the J1/2–V characteristics of the SCLC single-carrier mobility. (c) Comparison of the charge transport or injection properties from the devices
using spiro and spiro/CuxS. (d) Plots of dV/dJ versus (Jsc J)1 and the linear fitting curves.
moisture penetration, which would result in high recombination film. These crystalline nanoparticles are important to enhance
losses and degradation of the perovskite. To solve this problem, the performance of the photovoltaic device. Note that the
we developed a method using a thin layer of CuxS to cover the application of CuxS on the spiro layer prevents contact between
spiro layer and simultaneously act as a buffer layer and HTL. the Au and perovskite caused by the pinholes in spiro. The
We first investigated the interfaces in our devices. Figure 6 corresponding low-magnification SEM images of the spiro and
presents the surface morphologies of different films and the spiro/CuxS hybrid HTLs are presented in Supporting Informa-
cross-sectional SEM images of our devices. Figure 6(a) displays tion Figure S9. We further used AFM to measure the surface
the top-view SEM images of CH3NH3PbI3 perovskite films morphologies of the perovskite, perovskite/spiro, and perov-
deposited on FTO/SnO2 substrates. Figure 6(b) shows the cross- skite/spiro/CuxS films. An AFM image of the perovskite layer is
sectional SEM image of our real device. The thickness of the shown in Supporting Information Figure S10, and distinct
absorber, i.e., the perovskite film, was approximately 450 nm. We perovskite grains can be found in the 2 2 mm2 region, which is
can see clearly perovskite grains with an average size of 300 nm, consistent with the SEM images. The image implied that our
which were tiled to form a dense, smooth, and pinhole-free one-step solution-deposited perovskite film was quite smooth
perovskite film. High-quality perovskite thin films are vital for and was of high quality. After spin-casting the spiro layer, we
achieving high-performance solar cells. After spin-casting the found that the film was even smoother, with an RMS roughness
spiro layer, we obtained an organic layer covered smoothly on of 0.85 nm, as shown in Figure 6(e). However, pinholes with
MAPbI3, as presented in Figure 6(c). However, pinholes were diameters of 100 nm were inevitably formed in the film, as
observed in the spiro film. We then studied the formation of mentioned above, which may cause diffusion to the metal
CuxS on the spiro layer. Figure 6(d) shows the morphology of the electrode or allow the penetration of moisture. Therefore, the
vacuum-vapor-deposited CuxS thin film. CuxS nanoparticles extra CuxS layer is vital to cover the spiro surface to prevent the
with sizes of approximately 30 nm were clearly observed in the degradation of the device. Figure 6(f) shows an AFM image of
FULL PAPER
Figure 6. (a) SEM image of a perovskite film deposited on a FTO/SnO2 substrate. (b) Cross-sectional SEM image of a real device. (c) SEM image of a
spiro film spin-casted on FTO/SnO2/perovskite. (d) SEM image of a CuxS film deposited on FTO/SnO2/perovskite/spiro. The corresponding AFM
images of (e) a spiro film on FTO/SnO2/perovskite and (f) a CuxS film on FTO/SnO2/perovskite/spiro.
the CuxS film. The CuxS film, with an RMS roughness of peak positions of the perovskites, as demonstrated in Supporting
3.28 nm, was not as smooth as spiro, which may account for the Information Figure S11. The traced results are shown in Figure 7
slightly decreased FF of our device. Nanoparticles were spread (b) and (c). The perovskite peaks were described above. The
over a small region. In other words, the thin layer of peaks located at 12.8 and 38.7 belong to PbI2. For PVSCs based
CuxS protected the perovskite films from diffusion to the Au on the spiro HTL, we observed the appearance of one PbI2 peak
metal electrode and covered the pinholes in the spiro layer to centered at 38.7 and a strong peak that was even higher than the
prevent water penetration. perovskite peak at 12.8 after 24 h of exposure to 85% humidity.
Next, we examined the 1000 h, long-term air stability of the Encouragingly, no obvious appearance or enhancement of the
PVSCs based on the newly developed hybrid HTL without PbI2 peaks was detected for the PVSCs based on the spiro/
encapsulation. For this experiment, un-encapsulated devices CuxS HTL. These results demonstrated a better preservation of
were stored in ambient air (temperature: 25 C, humidity: 40%) the perovskite crystallinity against moisture infiltration of the
and tested every 125 h. Figure 7(a) presents the long-term CuxS devices. As is well known, the surface wettability of the top
monitored results. The detailed normalized evolution of Voc, Jsc, HTL is an important factor that significantly affects the stability
FF, and PCE are shown in the graph. Both of the initial selected of the PVSC. The higher stability of the spiro/CuxS device was
samples have efficiencies of 17%. For the device based on the attributed to the high hydrophobicity of CuxS. Contact angle
spiro HTL, the PCE first increased slightly and then decreased. tests showed that spiro-OMeTAD, exhibited a water contact angle
The absorption of oxygen by spiro may account for the initial of 63.54 (inset of Figure 7(b)). In sharp contrast, after the
efficiency increase. After 1000 h of storage, the spiro device deposition of CuxS, the spiro/CuxS sample exhibited a much
retained less than half of its initial PCE, showing quite obvious larger contact angle of 91.64 (inset of Figure 7(c)). Such a
and severe performance degradation. Remarkably, we found that hydrophobic HTL would protect the perovskite from moisture in
the spiro/CuxS device retained more than 90% of its initial air and is beneficial for the long-term stability.
efficiency. All the parameters remained almost the same. The
hybrid HTL device exhibited nearly constant efficiency over
1000 h, which demonstrated its much greater stability than the
3. Conclusions
spiro device. To gain additional insight into the enhanced
stability upon the incorporation of CuxS in the PVSCs, we In summary, we successfully developed a new HTL for planar
utilized XRD measurements to illustrate the crystallinity n-i-p perovskite solar cells with a structure of FTO/SnO2/
evolution of the perovskite protected by a spiro layer or a MAPbI3/spiro-OMeTAD/CuxS/Au. A room-temperature, vapor-
spiro/CuxS layer under a very high humidity of 85%. All the deposited, inorganic CuxS (x ¼ 1.75) polycrystalline film was
samples were prepared on FTO substrates with different HTLs first introduced between the spiro-OMeTAD layer and the Au
on the perovskite surface. Note that the introduction of spiro or metal electrode to simultaneously improve the efficiency and
spiro/CuxS on the perovskite layer does not affect the original stability of the assembled PVSC. The p-type CuxS HTL plays two
FULL PAPER
Figure 7. (a) PCE, FF, Jsc, and Voc as a function of ambient storage time of devices with different HTLs. The perovskite crystallinity evolution based
on different HTLs: (b) the spiro HTL and (c) the spiro/CuxS HTL. # and represent the identified PbI2 peaks and main perovskite peaks. The insets of
(b) and (c) show the water contact angle on the spiro and spiro/CuxS surfaces, respectively.
roles in enhancing the device performance. Its first role is as a diethyl ether, dried under vacuum, and used without further purification.
hole transport material to increase the conductivity and hole Then, PbI2 (461 mg, TCI) and CH3NH3I (159 mg) were dissolved in N,N-
dimethylformamide (DMF, 723 mL) and dimethyl sulfoxide (DMSO, 81 mL).
mobility of the HTL. The second role is as a buffer layer to protect
A small amount of Pb(SCN)2 (Aldrich) was added into the precursor to
the perovskite from being damaged or decomposed by contact enlarge the perovskite grain size. The mixed solution was stirred at room
with the top metal electrode or the moisture in air. The temperature for 10 h before use. Subsequently, the perovskite film was
combination of desirable hole transport properties, hydropho- formed by spin-casting the solution on the FTO/SnO2 substrate at 4000 rpm
bicity, and mechanical protection of the underlying perovskite for 60 s. Note that diethyl ether (500 mL) was slowly dripped on the rotating
layer caused by the application of CuxS on the spiro layer allowed substrate to aid in the formation of a uniform, dense, and crystalline
us to fabricate PVSCs with higher photoelectric efficiency and perovskite film. Then, the film was thermally annealed at 60 C for 2 min and
100 C for 5 min. After the deposition of the perovskite material, the hybrid
superior life-time stability over the reference spiro-OMeTAD bilayer hole transport layer (HTL) was deposited. The organic HTL solution
PVSCs. In contrast to the spiro-OMeTAD devices, planar devices (spiro, 70 mM in chlorobenzene, ShenZhen Feiming Science and
based on the hybrid HTL achieved a superior PCE of 18.58% and Technology Co., Ltd.) was spin-coated on the perovskite layer at
showed better stability, retaining over 90% of its initial PCE after 3000 rpm for 30 s. The HTL solution was doped with tBP (29 mL) and Li-
1000 h of storage in air. Our method offers a new way to improve TFSI (17.5 mL, 520 mg mL1 in acetonitrile). The inorganic CuxS HTL was
the efficiency and stability of PVSCs by employing a spiro- prepared on top of the spiro-OMeTAD layer via the vacuum thermal
evaporation of CuS powder (Aldrich, 99.99%). Finally, the electrode was
OMeTAD/CuxS hybrid HTL. deposited via the thermal evaporation of gold under a pressure of
5 105 Torr. The active area was 0.09 cm2, as defined by a shadow mask.
4. Experimental Section
4.2. Material and Device Characterization
4.1. Device Fabrication
The crystal structure of the CuxS and perovskite films were characterized
FTO glasses were cleaned with detergent, deionized water, and acetone and by X-ray diffraction (XRD). Conventional XRD in the Bragg–Brentano
sonicated in ethanol in an ultrasonic bath for 10 min. The glasses were then configuration was performed using a Bruker D8 Advance diffraction meter
treated with UVO for 15 min prior to use. The SnO2 electron transport layer with Cu Ka radiation at 40 kV and 40 mA. Line traces were collected over 2u
(ETL) was deposited following our previous work. Typically, a values ranging from 10 to 60 . The transmittance of the films was
SnCl2 2H2O solution (0.1 M) dissolved in ethanol was spin-coated on measured by a UV-VIS-NIR spectrophotometer (CARY5000, Varian) in the
an FTO substrate at 2000 rpm for 40 s. To form a dense ETL, the substrates 300–800 nm wavelength range at RT. The surface morphologies were
were annealed at 180 C in air for 2 h. Afterwards, a PCBM solution obtained using a scanning electron microscope (SEM, FEI XL-30).
(15 mg mL1) was spin-casted at 3000 rpm for 40 s on the ETL to passivate Transmission electron microscopy (TEM) characterization was carried
the interface. Organohalide lead perovskite (CH3NH3PbI3) films were out using a JEOL JEM 2010. X-ray photoelectron spectroscopy (XPS) and
prepared on the ETL using the method described by Nam-Gyu Park et al. ultraviolet photoelectron spectroscopy (UPS) were performed using a
CH3NH3I was synthesized first. Generally, hydroiodic acid (30 mL, XPS/UPS system (Thermo Scientific, ESCLAB 250Xi, USA). The
0.227 mol, 57 wt.% in water, Aldrich) and methylamine (27.8 mL, compositions and chemical states of the CuxS film were examined by
0.273 mol, 40% in methanol, TCI) were mixed in an ice bath for 2 h. After XPS. Before testing, the samples were sputter-cleaned by lower energy
stirring at 0 C for 2 h, the resulting solution was evaporated at 50 C for 1 h to Arþ for approximately 30 s to remove the atmospheric contamination in
obtain the product (CH3NH3I). The precipitate was washed three times with the XPS chamber, and the Arþ gun was operated at 0.5 kV at a pressure of
1 107 Pa. The vacuum pressure of the analysis chamber was lower than
FULL PAPER
XRD, AFM, XPS, and UPS measurements. The authors declare no
1 108 Pa. Whole survey scans to identify the overall surface competing financial interest.
composition and chemical states were performed using a monochro-
mated Al Ka X-ray source (1486.68 eV), and the photoelectrons were
detected at a 150 eV pass energy and a channel width of 500 meV. The Received: March 21, 2017
surface carbon signal at 284.6 eV was used as an internal standard. The Revised: April 19, 2017
work function and band energy position were calculated by UPS. UPS was Published online: May 3, 2017
carried out using helium Ia radiation from a discharge lamp operated at
90 W, a pass energy of 10 eV, and a channel width of 25 meV. A 9 V bias
was applied to the samples to separate the samples and analyze the low-
[1] Lee, M. M., Teuscher, J., Miyasaka, T., Murakami, T. N., Snaith, H. J.,
kinetic-energy cutoffs. The morphologies of the perovskite and different
Science 2012, 338, 643.
HTLs spin-casted on the perovskite were characterized by atomic force
[2] Burschka, J., Pellet, N., Moon, S. J., Humphry-Baker, R., Gao, P.,
microscopy (AFM, SPM-9500J3, Shimadzu, Japan). Photoluminescence
(PL) measurements were carried out under a 488 nm laser at RT, and Nazeeruddin, M. K., Gratzel, M., Nature 2013, 499, 316.
the emissions were collected via a HORIBA Jobin-Yvon monochromator. [3] Docampo, P., Ball, J. M., Darwich, M., Eperon, G. E., Snaith, H. J.,
The current–voltage (J–V) curves of the devices were obtained using a Nat. Commun. 2013, 4, 2761.
computer-controlled Keithley 2400 Source Measure Unit, and the device [4] Marchioro, A., Teuscher, J., Friedrich, D., Kunst, M., van de Krol, R.,
test was carried out under the illumination of AM 1.5G, 100 mW cm2 (the Moehl, T., Grätzel, M., Moser, J.-E., Nat. Photon. 2014, 8, 250.
light intensity was calibrated using a Si photodiode) at RT using a solar [5] Kaltenbrunner, M., Adam, G., Glowacki, E. D., Drack, M., Schwodiauer,
simulator. The corresponding incident photon-to-electron conversion R., Leonat, L., Apaydin, D. H., Groiss, H., Scharber, M. C., White, M. S.,
efficiency (IPCE) spectrum was measured with a QE/IPCE measurement Sariciftci, N. S., Bauer, S., Nat. Mater. 2015, 14, 1032.
system (Enli Technology Co., Ltd.). SCLC measurements were conducted [6] Giordano, F., Abate, A., Correa Baena, J. P., Saliba, M., Matsui, T., Im,
by fabricating devices with structures of FTO/Au/spiro-OMeTAD/CuxS/ S. H., Zakeeruddin, S. M., Nazeeruddin, M. K., Hagfeldt, A., Graetzel,
Au or FTO/Au/spiro-OMeTAD/Au and then measuring the J–V curves of M., Nat. Commun. 2016, 7, 10379.
the devices using the Keithley 2400 Source Measure Unit. The hole [7] Yang, G., Tao, H., Qin, P., Ke, W., Fang, G., J. Mater. Chem. A 2016, 4, 3970.
mobility was extracted by fitting the J–V curves according to the modified [8] Liu, M., Johnston, M. B., Snaith, H. J., Nature 2013, 501, 395.
Mott–Gurney equation [9] Stranks, S. D., Eperon, G. E., Grancini, G., Menelaou, C., Alcocer,
M. J., Leijtens, T., Herz, L. M., Petrozza, A., Snaith, H. J., Science 2013,
9 V2 342, 341.
J ¼ er e0 mhðeÞ 3 ð1Þ [10] Xing, G., Mathews, N., Sun, S., Lim, S. S., Lam, Y. M., Gratzel, M.,
8 L
Mhaisalkar, S., Sum, T. C., Science 2013, 342, 344.
In the equation, e0 is the permittivity of free space, er is the permittivity [11] Kojima, A., Teshima, K., Shirai, Y., Miyasak, T., J. Am. Chem. Soc.
of the polymer, mh(e) is the carrier mobility, and L is the thickness of the 2009, 17, 6050.
active layer. The dielectric constant er is assumed to be 3, which is a [12] Bi, D., Tress, W., Dar, M. I., Gao, P., Luo, J., Renevier, C., Schenk, K.,
typical value for conjugated polymers. For the calculation of the Abate, A., Giordano, F., Baena, J.-P.C., Decoppet, J.-D., Zakeeruddin,
heterojunction solar cells, a diode model was used to analyze the J–V S. M., Nazeeruddin, M. K., Grätzel, M., Hagfeldt, A., Sci. Adv. 2016, 2,
curve, from which we obtained the series resistance (Rs) according to the e1501170.
linear fit of the data. [13] Liu, D., Kelly, T. L., Nat. Photon. 2013, 8, 133.
[14] Lin, Q., Armin, A., Nagiri, R. C. R., Burn, P. L., Meredith, P., Nat.
dV AK B T Photon. 2014, 9, 106.
¼ ðJsc JÞ1 þ Rs ð2Þ
dJ e [15] Zhou, H., Chen, Q., Li, G., Luo, S., Song, T. B., Duan, H. S., Hong, Z.,
You, J., Liu, Y., Yang, Y., Science 2014, 345, 542.
where A is the ideality factor of the heterojunction, KB is Boltzmann’s [16] Mahmood, K., Swain, B. S., Amassian, A., Adv. Mater. 2015, 27, 2859.
constant, T is the absolute temperature and e is the elementary charge. [17] Ke, W., Xiao, C., Wang, C., Saparov, B., Duan, H. S., Zhao, D., Xiao,
Electrochemical impedance spectroscopy (EIS) was performed on a Z., Schulz, P., Harvey, S. P., Liao, W., Meng, W., Yu, Y., Cimaroli, A. J.,
CHI660D electrochemical workstation (Shanghai, China) with a 5 mV AC Jiang, C. S., Zhu, K., Al-Jassim, M., Fang, G., Mitzi, D. B., Yan, Y., Adv.
amplitude over frequency range of 100 kHz to 1 Hz and a 0.4 V bias under Mater. 2016, 28, 5214.
darkness. [18] You, J., Meng, L., Song, T. B., Guo, T. F., Yang, Y. M., Chang, W. H.,
Hong, Z., Chen, H., Zhou, H., Chen, Q., Liu, Y., De Marco, N., Yang,
Y., Nat. Nanotechnol. 2016, 11, 75.
[19] Zhu, Z., Xu, J.-Q., Chueh, C.-C., Liu, H., Li, Z., Li, X., Chen, H., Jen,
Supporting Information A. K.-Y., Adv. Mater. 2016, 28, 10786.
Additional supporting information may be found in the online version of [20] NREL Chart, http://www.nrel.gov/pv/assets/images/efficiency_cha
this article at the publisher’s website. rt.jpg.
[21] Ke, W., Fang, G., Liu, Q., Xiong, L., Qin, P., Tao, H., Wang, J., Lei, H.,
Li, B., Wan, J., Yang, G., Yan, Y., J. Am. Chem. Soc. 2015, 137, 6730.
[22] Lei, H., Yang, G., Guo, Y., Xiong, L., Qin, P., Dai, X., Zheng, X., Ke, W.,
Acknowledgements Tao, H., Chen, Z., Li, B., Fang, G., Phys. Chem. Chem. Phys. 2016, 18,
H. W. Lei, G. Yang, and X. L. Zheng contributed equally to this work. This 16436.
work was financially supported by the National High Technology Research [23] Snaith, H. J., Grätzel, M., Appl. Phys. Lett. 2006, 89, 262114.
and Development Program (2015AA050601), the National Natural [24] Kim, G.-W., Kang, G., Kim, J., Lee, G.-Y., Kim, H. I., Pyeon, L., Lee, J.,
Science Foundation of China (11674245, 91433203), and the Fundamen- Park, T., Energy Environ. Sci. 2016, 9, 2326.
tal Research Funds for the Central Universities (20152020203). The [25] Hawash, Z., Ono, L. K., Raga, S. R., Lee, M. V., Qi, Y., Chem. Mater.
authors thank the nano center of Wuhan University for the SEM, TEM, 2015, 27, 562.
FULL PAPER
[26] Misra, R. K., Aharon, S., Li, B., Mogilyansky, D., Visoly-Fisher, I., [35] Lei, H., Fang, G., Cheng, F., Ke, W., Qin, P., Song, Z., Zheng, Q., Fan,
Etgar, L., Katz, E. A., J. Phys. Chem. Lett. 2015, 6, 326. X., Huang, H., Zhao, X., Sol. Energy Mater. Sol. C 2014, 128, 77.
[27] Christians, J. A., Fung, R. C., Kamat, P. V., J. Am. Chem. Soc. 2014, [36] Lei, H., Qin, P., Ke, W., Guo, Y., Dai, X., Chen, Z., Wang, H., Li, B.,
136, 758. Zheng, Q., Fang, G., Org. Electron. 2015, 22, 173.
[28] Qin, P., Tanaka, S., Ito, S., Tetreault, N., Manabe, K., Nishino, [37] Wei, S.-H., Xu, Q., Huang, B., Zhao, Y., Yan, Y., Noufi, R., In
H., Nazeeruddin, M. K., Gratzel, M., Nat. Commun. 2014, 5, Photovoltaic Specialists Conference 38th IEEE, 2012, 000118.
3834. [38] Grozdanov, I., Najdoski, M., J. Solid State Chem. 1995, 114, 469.
[29] Ke, W., Zhao, D., Grice, C. R., Cimaroli, A. J., Fang, G., Yan, Y., [39] Gotsis, H. J., Barnes, A. C., Strange, P., J. Phys.: Condens. Matter 1992,
J. Mater. Chem. A 2015, 3, 23888. 4, 10461.
[30] Ye, S., Rao, H., Yan, W., Li, Y., Sun, W., Peng, H., Liu, Z., Bian, Z., Li, [40] Liu, L., Zhou, B., Deng, L., Fu, W., Zhang, J., Wu, M., Zhang, W., Zou,
Y., Huang, C., Adv. Mater. 2016, 28, 9648. B., Zhong, H., J. Phys. Chem. C 2014, 118, 26964.
[31] Ye, S., Sun, W., Li, Y., Yan, W., Peng, H., Bian, Z., Liu, Z., Huang, C., [41] Bagul, S. V., Chavhan, S.D., Sharma, R., J. Phys. Chem. Solids 2007, 68, 1623.
Nano Lett. 2015, 15, 3723. [42] Kundu, M., Hasegawa, T., Terabe, K., Yamamoto, K., Aono, M., Sci.
[32] Madhavan, V. E., Zimmermann, I., Roldan-Carmona, C., Grancini, Technol. Adv. Mater. 2016, 9, 035011.
G., Buffiere, M., Belaidi, A., Nazeeruddin, M. K., ACS Energy Lett. [43] Pattrick, R. A. D., Mosselmans, J. F. W., Charnock, J. M., England,
2016, 1, 1112. K. E. R., Helz, G. R., Garner, C. D., Vaughan, D. J., Geochim.
[33] Yang, G., Wang, Y.-L., Xu, J.-J., Lei, H.-W., Chen, C., Shan, H.-Q., Liu, Cosmochim. Ac. 1997, 10, 2023.
X.-Y., Xu, Z.-X., Fang, G.-J., Nano Energy 2016, 31, 322. [44] Cui, J., Jiang, R., Lu, W., Xu, S., Wang, L., Small 2017, 13, 1602235.
[34] Zhu, Q., Bao, X., Yu, J., Zhu, D., Qiu, M., Yang, R., Dong, L., ACS [45] Koo, B., Jung, H., Park, M., Kim, J.-Y., Son, H. J., Cho, J., Ko, M. J., Adv.
Appl. Mater. Inter. 2016, 8, 2652. Funct. Mater. 2016, 30, 5400.

2

Uploaded by

Copyright:

Available Formats

2

Uploaded by

Copyright:

Available Formats

www.advancedsciencenews.com www.solar-rrl.

You might also like