Accepted Manuscript Feasibility study of atmospheric-pressure dielectric barrier discharge treatment on CH3NH3PbI3 films for inverted planar perovskite solar cells
Jui-Hsuan Tsai, I-Chun Cheng, Cheng-Che Hsu, Chu-Chen Chueh, Jian-Zhang Chen PII:
S0013-4686(18)32213-8
DOI:
10.1016/j.electacta.2018.09.203
Reference:
EA 32794
To appear in:
Electrochimica Acta
Received Date:
27 May 2018
Accepted Date:
30 September 2018
Please cite this article as: Jui-Hsuan Tsai, I-Chun Cheng, Cheng-Che Hsu, Chu-Chen Chueh, JianZhang Chen, Feasibility study of atmospheric-pressure dielectric barrier discharge treatment on CH NH3PbI3 films for inverted planar perovskite solar cells, Electrochimica Acta (2018), doi: 10.1016/j. 3 electacta.2018.09.203
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Feasibility study of atmospheric-pressure dielectric barrier discharge treatment on CH3NH3PbI3 films for inverted planar perovskite solar cells Jui-Hsuan Tsai1,4,^, I-Chun Cheng2,#, Cheng-Che Hsu3,$, Chu-Chen Chueh3,4,%, JianZhang Chen1,4,* 1Graduate
Institute of Applied Mechanics, National Taiwan University, Taipei City 10617, Taiwan 2Department of Electrical Engineering & Graduate Institute of Photonics and Optoelectronics, National Taiwan University, Taipei City 10617, Taiwan 3Department of Chemical Engineering, National Taiwan University, Taipei City 10617, Taiwan 4Advanced Research Center for Green Materials Science and Technology, National Taiwan University, Taipei City 10617, Taiwan Corresponding Authors *Jian-Zhang Chen Email:
[email protected] TEL: +886-2-33665694 #I-Chun Cheng Email:
[email protected] TEL: +886-2-33669648 $Cheng-Che
Hsu Email:
[email protected] TEL: +886-2-33663034 %Chu-Chen Chueh Email:
[email protected] TEL: +886-2-33664947 ^ Jui-Hsuan Tsai Email:
[email protected] TEL: +886-2-33669423
Abstract A facile surface treatment of the CH3NH3PbI3 film is conducted using a portable
1
ACCEPTED MANUSCRIPT
dielectric barrier discharge device inside a nitrogen-filled glove box. DBD treatment can promote the surface crystallization of the CH3NH3PbI3 film, increase the conductivity of CH3NH3PbI3 film, and manifest efficient defect passivation. In addition, DBD treatment can remove surface N-containing fragments of perovskite film to produce a lead-rich surface, which not only affords self-passivation effect but also increases the interactions at the perovskite/fullerene interface. Benefitting from the reduced defect density at the perovskite interface, the DBD-treated perovskite film can enable a ~12.5% enhancement in power conversion efficiency of the derived doped PC61BM-based perovskite solar cells. Such enhancement is mainly contributed from the increased open circuit voltage (from 0.885 to 0.922 V) and fill factor (0.698 to 0.759). Detailed optimization of the DBD treatment is carefully investigated in this study. Keywords: Atmospheric pressure plasma; dielectric barrier discharge, perovskite solar cell; passivation; grain growth
2
ACCEPTED MANUSCRIPT
1. Introduction Low-pressure plasma is a mature technology that has been extensively used in the integrated circuit (IC) and display industry. Its constituent components like vacuum chambers and pumping systems generally demand routine maintenance to ensure it stable operation. Therefore, many studies are focusing on regular-pressure plasma technology as an alternative. Atmospheric-pressure plasma (APP) can be operated at regular pressure without requiring vacuum chambers and pumps. By using different electrode configurations and excitation methods, APPs with various heavy particle temperatures and charge densities can be generated [1, 2]. Typical APPs include dielectric barrier discharge (DBD), corona discharge, transferred arc, microplasma, and atmospheric pressure plasma jet (APPJ). Various types of APPs have diverse properties, thus enabling applications in different fields such as thin-film deposition [3, 4], nanoparticle synthesis [5], liquid precursor reduction [6], surface modification and cleaning [7, 8], sintering of reduced graphene oxides and carbon nanotubes [9, 10], and materials etching [11]. It is worth noting that the sheath structure of APPs is much thinner than that of low-pressure plasma (sometimes it even collapses). In addition, the mean free path of atoms or molecules in APPs is much shorter than that of species in low-pressure plasma. Therefore, compared with low pressure plasma, ion bombardment or sputtering effect of APPs is very limited because ions will lose energy upon collisions with higher probability. The organic-inorganic hybrid perovskite solar cells (PSCs), especially for the inverted p-i-n device configuration have attracted worldwide attention owing to their superiority in power conversion efficiency (PCE) and low temperature processing procedure [12, 13]. One important research subject in this field is to reduce the defect
3
ACCEPTED MANUSCRIPT
density at the relevant perovskite interface to further optimize the resultant performance [11, 14-17]. To this end, several methods including vapor-assisted solution, gas-assisted solution, and vacuum flash-assisted solution processes [15, 18, 19] have been reported and shown promising effects on enhanced crystallization. On the other hand, adequate amount of PbI2 at the perovskite grain boundaries upon releasing organic species via annealing has been proven as an effective approach to reduce the carrier recombination, which is described as the self-passivation of the perovskite layer [20]. Given the powerful surface treatment capability, the plasma processes have been introduced in PSC fabrication to amend the associated interfaces. For example, lowpressure Ar plasma treatment was introduced to modify the surface composition and defect types of the deposited perovskite film; this can efficiently enhance charge collection across the perovskite–electrode interface and suppress charge recombination [21]. APP-enhanced CVD was applied to roll-to-roll processes of large-area functional titania used as hole-blocking electron-transport layers in PSCs [22]. DBD and APPJ were used to decontaminate and activate FTO glass substrates for PSCs [23]. Atmospheric Ar/O2 plasma was used as a key energy source for the activation of sol–gel film to formulate dense SnO2 from a SnCl2 precursor with record rate of ≤5 min [24]. Atmospheric-pressure hydrogen plasma was also used for dry-etching perovskite films [25]. A low-pressure oxygen plasma treatment has been lightly used as an acid-free approach
to
enhance
the
conductivity
of
poly(3,4-ethylenedioxythiophene)-
poly(styrenesulfonate) (PEDOT:PSS) layers for PSCs [26]. In this study, a portable atmospheric-pressure DBD device was used to treat the perovskite films inside a nitrogen-filled glove box. In this home-made portable DBD
4
ACCEPTED MANUSCRIPT
device, the plasma species are generated and diffused to the treated material surface; it is a diffusion process that has almost no ion bombardment or sputtering effects. This is different from jet-type APPs in which the high-rate flows may accelerate ions to bombard material surface. As such, we do not have to consider ion bombardment or sputtering effect on the treated material in our process. DBD treatment can enlarge the grain size by two times, remove surface N-containing fragments, and manifest efficient defect passivation capability. As a result, the DBD-treated perovskite film can enable a ~12.5% enhancement in PCE of the derived doped PC61BM-based PSCs as compared to that of the untreated device. Such enhanced PCE was mainly contributed from the increased open circuit voltage (VOC) from 0.885 to 0.922 and the fill factor (FF) from 0.698 to 0.759, highlighting its efficacy in surface passivation.
2. Experimental details 2.1 Film deposition and device fabrication Patterned indium tin oxide (ITO)-coated glass was sequentially cleaned by sonication with deionized water, acetone, and isopropanol for 15 min and treated using a UV-ozone cleaner for 15 min. PEDOT:PSS (Clevios P VP AI 4083, Heraeus) was spincoated on the patterned ITO glass at 4000 rpm for 40 s, followed by annealing at 140°C for 20 min. Then, the samples were transferred into a nitrogen-filled glove box. CH3NH3PbI3 film was deposited by a one-step process, as described in our previous study [23]. CH3NH3PbI3 film was annealed at 100°C for 10 min, following which it was treated by the portable DBD for 0, 10, 20, 40, or 60 s inside a glove box. Figure 1 shows the DBD setup. The DBD was operated with an AC voltage (~20 kV, 0.3~2 kHz) excitation.
5
ACCEPTED MANUSCRIPT
The plasma was produced between the dielectric and the ground mesh; the generated plasma reactive species diffused to the surface of the target sample being treated. After DBD treatment, PC61BM (nano-C) (20 mg/ml in chlorobenzene) was spin-coated on the CH3NH3PbI3 film at 2000 rpm for 30 s for one testing structure. For the other structure with higher PCE, PC61BM was doped with 2 μl of DMOAP (Sigma-Aldrich, 42 wt% in methanol). DMOPA is mixed with PC61BM before spin-coating. Then, bathocuproine (BCP, Alfa Aesar) (0.5 mg/ml in 2-propanol) was spin-coated on the PC61BM film at 6000 rpm for 20 s. Finally, a 100-nm-thick Ag was deposited using an e-beam evaporator; the area was defined as 0.08 cm2.
(a)
(b) Figure 1. (a) Experimental setup of DBD system. (b) Structures of inverted planar PSCs. 2.2 Film and device characterization The optical emission spectra (OES) of DBD were monitored using spectrometers (Ocean FLAME-S-XR1-ES) with attached Wi-Fi module. The surface morphology of the CH3NH3PbI3 film was observed by scanning electron microscopy (SEM, JOEL JSM7800 Prime). The crystallization of the CH3NH3PbI3 film was analyzed by grazing 6
ACCEPTED MANUSCRIPT
incidence X-ray diffractometry (GIXRD, BRUKER D8 DISCOVER SSS multi-function high power X-Ray diffractometer). The surface chemical bonding status of the CH3NH3PbI3 film was inspected by X-ray photoelectron spectrometry (XPS, VG ESCA Scientific Theta Probe). The conductivity of the CH3NH3PbI3 film was measured by using a digital sourcemeter (Keithley 2636A). For conductivity measurement, it was measured through plane. The perovskite film was deposited on strip-patterned ITO glass, the width of ITO is 4 mm. Gold strip with a width of 2 mm, perpendicular to the ITO strip, was deposited on the perovskite layer. Therefore, the defined area for conductivity measurement is 8 mm2. The current density-voltage (J-V) curve was measured using a precision source/measure unit (Agilent B2902A) under simulated solar AM1.5 light (ABET Sun 2000 Solar Simulator).
3. Results and discussion Figure 2(a) shows the OES during the DBD processing of CH3NH3PbI3 with different processing times. The plasma intensity increased over time and reached its maximum at ~30 s. Figure 2(b) shows the OES of DBD at 60 s. The N2 2nd positive system C3u-g peaks are in the range of 300‒410 nm, and the NO system A2+-X2 peaks are in the range of 220‒280 nm. The DBD instrument was operated in a nitrogenfilled glove box with low oxygen level; therefore, the intensities of the NO system peaks are fairly low compared with those of the N2 emission system.
7
ACCEPTED MANUSCRIPT
(a)
(b) Figure 2. (a) OES evolution during DBD processing of CH3NH3PbI3 layer. (b) OES of DBD at 60 s.
Figure 3 shows surface SEM images of CH3NH3PbI3 with different DBD treatment durations. As the DBD treatment time increased, the grain size of CH3NH3PbI3 increased. As the DBD treatment time reached 40 s, some precipitates were found on the edge of the grains. As the DBD treatment time reached 60 s, the grains became less dense. In DBD plasma, electrons and heavy particles may react with the surface material of the 8
ACCEPTED MANUSCRIPT
perovskite layer. The heavy particles could react with organic portion of perovskite film and make CH3NH3- detached to form free radicals. In grain boundary regions, two free radicals could bond together and eliminate the boundary. Once the content of CH3NH3- is low, the perovskite may decompose and precipitate PbI2. In addition, our DBD device slightly heated up the sample, this may also cause grain growth. On the surface, plasma reactive species react with the surface material. Because DBD also slightly heats up samples; therefore, there is also some thermal effect. In the bulk, the main effect results from the heating of the DBD treatment. On the surface, it is a synergetic effect of the plasma reactive species and heat. As mentioned in the Introduction section, APPs have thin sheath and very short mean free path compared with low-pressure plasmas, and our home-made DBD device is a diffuse-type without jet flows. Therefore, almost no ion bombardment or sputtering is expected in our DBD material treatment process. In principle, charge recombination can be categorized into radiative and nonradiative processes. For perovskite films, non-radiative recombination has been generally cited to arise from the grain boundaries, impurities, and the interface with neighboring layers [27]. The larger grain sizes (less grain boundaries) can suppress recombination to benefit the PSC performance [28]. Therefore, by modulating the morphology of the perovskite layer, the density of grain boundaries and traps can be reduced, and the potential loss in the device can be minimized. High density grain boundaries can hinder carrier transport, thereby increasing the series resistance and resulting in lower FF [29]. After DBD treatment, the grain size of CH3NH3PbI3 increased. The overall growth of small precipitates in the grain boundaries might lower the absorption efficacy of the CH3NH3PbI3 layer and engender additional charge trapping sites. It is noted that
9
ACCEPTED MANUSCRIPT
CH3NH3PbI3 with 60-s DBD treatment showed less dense crystals accompanied with more impurities in the CH3NH3PbI3 layer, which is unfavorable for realizing highefficiency PSCs.
(a)
(b)
(c)
(d)
(e) Figure 3. SEM top view images of CH3NH3PbI3 films with DBD treatments for (a) 0 s, (b) 10 s, (c) 20 s, (d) 40 s, and (e) 60 s. Figure 4(a) shows XRD patterns of the CH3NH3PbI3 films under various DBD treatment durations. No significant difference was seen in the diffraction peaks of CH3NH3PbI3 for all DBD processing times. To measure the conductivity of the CH3NH3PbI3 films, gold pads were deposited on the CH3NH3PbI3 films on ITO glass substrates. Figure 4(b) shows the current-voltage curves of CH3NH3PbI3 films treated by DBD for various durations. Table 1 shows measured conductivity values. After DBD 10
ACCEPTED MANUSCRIPT
treatment, the conductivity of CH3NH3PbI3 films increased from 1.39 mS·m-1 (0 s) to 1.96 mS·m-1 (20 s) and then decreased to 1.43 mS·m-1 (60 s). The improvement can be attributed to the grain size enlargement. Larger grains result in lower density carrier trap and grain boundaries, thereby increasing the conductivity. However, with treatment durations above 40 s, the small precipitates that formed in the grain boundaries increased the barrier for carrier transports, thereby decreasing the conductivity. Therefore, the DBD treatment of perovskite film for 20 s might be the best condition for device fabrication.
(a)
(b) Figure 4. (a) XRD and (b) conductivity of CH3NH3PbI3 films with various DBD treatment durations. Table 1. Conductivity of CH3NH3PbI3 films with various DBD treatment durations.
11
ACCEPTED MANUSCRIPT
DBD treatment duration (s)
Conductivity (mS·m-1)
0
1.39±0.33
10
1.62±0.50
20
1.96±0.38
40
1.60±0.24
60
1.43±0.29
Figure 5(a)‒(d) shows the XPS C1s, N1s, Pb4f, and I3d spectra of the CH3NH3PbI3 films treated by DBD for various durations. C1s and N1s binding energy peaks show more obvious change with DBD treatment for various durations. Figure 6(a)‒(e) shows the detailed C1s spectra fitting results. Table 2 shows the atomic ratios. An apparent peak reduction is noted in the N1s peak. The N1s peak nearly disappeared with 60-s DBD treatment. The nitrogen content decreased from ~0.1 (10%) to 0.037 (3.7%) as the DBD treatment time increased to 60 s. The iodine content decreased only slightly within the error margin with DBD treatment. DBD treatment is more reactive with N-containing residue. The Pb content increased monotonically from 0.186 to 0.263 as the DBD treatment time increased from 0 to 60 s. An abrupt increase was seen as the DBD treatment time increased from 40 to 60 s. After DBD treatment, N on the surface of the CH3NH3PbI3 film were released to form a lead-rich surface, enabling the selfpassivation effect as discussed earlier [21]. Adequate amount of PbI2 precipitates around the grain boundaries could also exert self-passivation effect to reduce carrier
12
ACCEPTED MANUSCRIPT
recombination [20].
(a)
(b)
(c)
(d) 13
ACCEPTED MANUSCRIPT
Figure 5. XPS (a) C1s, (b) N1s, (c) Pb4f, and (d) I3d spectra of CH3NH3PbI3 films with various DBD treatment durations. Table 2. Atomic ratios of CH3NH3PbI3 films with various DBD treatment durations. DBD treatment
C
N
Pb
I
0
0.234
0.097
0.186
0.483
10
0.217
0.104
0.188
0.491
20
0.250
0.087
0.194
0.469
40
0.262
0.074
0.207
0.457
60
0.248
0.037
0.263
0.452
duration (s)
Figure 6(a)‒(e) shows the XPS C1s fine scan results of the CH3NH3PbI3 films with DBD treatments for various durations. The peaks at 284.8 eV and 285.9 eV correspond to C-C and C-N bonds, respectively [30]. The C-N bond can be attributed to CH3NH3I in CH3NH3PbI3 or solvent DMF, and the C-C bond can be attributed to the normal contaminations in the environment [31]. After DBD treatment, the C-N bond content decreased significantly; this is consistent with the XPS N1s spectra.
(a)
(b)
14
ACCEPTED MANUSCRIPT
(c)
(d)
(e) Figure 6. XPS C1s spectra of CH3NH3PbI3 film with DBD treatment durations of (a) 0 s, (b) 10 s, (c) 20 s, (d) 40 s and (e) 60 s. To verify whether DBD treatment is beneficial for PSCs, two types of p-i-n structure PSCs were fabricated using DBD-treated CH3NH3PbI3, as shown in Figure 1(b). As shown in Figure 7, the regular ITO/PEDOT:PSS/CH3NH3PbI3/PC61BM/BCP/Ag structure
showed
poorer
performance,
whereas
the
ITO/PEDOT:PSS/CH3NH3PbI3/DMOAP-doped PC61BM/BCP/Ag yielded higher PCE owing to enhanced electron-collecting efficiency enabled by DMOAP doping in the PC61BM layer. Figure 7(a) and (b) show the corresponding J-V curves for PSCs with and without DMOAP doping in the PC61BM electron-transporting layer (ETL), respectively. Table 3 summarizes the photovoltaic (PV) parameters of PSCs. With DBD treatment, PCE improved from 3.814% to 5.816%; VOC, from 0.827 to 0.875 eV; and FF, from
15
ACCEPTED MANUSCRIPT
37.60% to 46.84% (forward scan) for PSCs without DMOAP doping in the PC61BM layer. For PSCs with DMOAP doping in the PC61BM layer, PCE improved from 8.755% to 9.846%; VOC, from 0.885 to 0.922 eV; and FF, from 69.84% to 75.91% (reverse scan). The enhancements of the VOC and FF could be attributed to the enlargement of the grain size; this reduces the densities of traps and grain boundaries and thus improves the PCE. In addition, the PCE improvement rate is higher in PSCs without DMOAP doping in the PC61BM layer (52.5%) than in those with DMOAP doping in the PC61BM layer (12.5%); this suggests the unsatisfactory passivating effectiveness of bare PC61BM layer, revealing the deficiency at this corresponding interface. Because DBD treatment can transform the CH3NH3PbI3 surface to a lead-rich surface, it is thus suspected that this conversion can promote the associated interaction between the perovskite and fullerene to reduce charge recombination. The PCE improvement (12.5%) observed in the DMOAP-doped PC61BM-based device is smaller, which suggests the DMOAP-doped PC61BM possesses a better passivation capability than pristine PC61BM.
(a)
(b)
Figure 7. J-V characteristics of PSC (a) without DMOAP and (b) with DMAOP doping in the PC61BM layer. Table 3. PV parameters of PSCs. Voc (V) Jsc (mA/cm2) FF (%) PCE (%) 16
ACCEPTED MANUSCRIPT
DBD 0 forward
0.827
12.264
37.595
3.814
s
reverse
0.793
12.187
37.070
3.583
DBD
forward
0.875
14.190
46.835
5.816
20 s
reverse
0.807
15.299
40.299
4.973
DBD 0 forward
0.887
13.469
60.147
7.189
without DMOAP
with DMOAP
s
reverse
0.885
14.166
69.843
8.755
DBD
forward
0.930
13.563
58.105
7.331
20 s
reverse
0.922
14.071
75.910
9.846
4. Summary A portable atmospheric-pressure DBD device is used for CH3NH3PbI3 treatment inside a nitrogen-filled glove box.
DBD treatment increases the grain size and
conductivity of the CH3NH3PbI3 film through a reduction in defect (or carrier recombination center) density. XPS results show a clear reduction of nitrogen and carbon contents owing to DBD treatment, suggesting the conversion of a lead-rich surface with shows self-induced passivation effect. DBD treatment enable a 52.5% improved PCE in a PC61BM-based device, whereas the improvement becomes less significant (12.5%) in a DMOAP-doped PC61BM-based device. This study provides a new method to improve the quality of perovskite films
Acknowledgments
17
ACCEPTED MANUSCRIPT
This work is financially supported by the “Advanced Research Center for Green Materials Science and Technology” from The Featured Area Research Center Program of the Higher Education Sprout Project by the Ministry of Education (107L9006) and the Ministry of Science and Technology in Taiwan (MOST 106-2221-E-002-193-MY2 & MOST 107-3017-F-002-001). We would like to thank Ms. Yuan-Tze Lee for her assistance with the SEM operation. The facility support from the Nano-ElectroMechanical-Systems (NEMS) Research Center at National Taiwan University, Taiwan is also gratefully acknowledged.
References [1] M. Laroussi, T. Akan, Plasma Processes and Polymers, 4 (2007) 777-788. [2] A. Schutze, J.Y. Jeong, S.E. Babayan, J. Park, G.S. Selwyn, R.F. Hicks, IEEE transactions on plasma science, 26 (1998) 1685-1694. [3] S. Babayan, J. Jeong, A. Schütze, V. Tu, M. Moravej, G. Selwyn, R. Hicks, Plasma Sources Science and Technology, 10 (2001) 573. [4] A.C. Bose, Y. Shimizu, D. Mariotti, T. Sasaki, K. Terashima, N. Koshizaki, Nanotechnology, 17 (2006) 5976-5982. [5] J. Patel, L. Nemcova, P. Maguire, W.G. Graham, D. Mariotti, Nanotechnology, 24 (2013). [6] J.Z. Chen, C.C. Hsu, C. Wang, W.Y. Liao, C.H. Wu, T.J. Wu, H.W. Liu, H.M. Chang, S.T. Lien, H.C. Li, C.M. Hsu, P.K. Kao, Y.J. Yang, I.C. Cheng, Coatings, 5 (2015) 26-38. [7] S.L. Chen, S. Wang, Y.B. Wang, B.H. Guo, G.Q. Li, Z.S. Chang, G.J. Zhang, Applied Surface Science, 414 (2017) 107-113. [8] J. Munoz, J.A. Bravo, M.D. Calzada, Applied Surface Science, 407 (2017) 72-81. [9] F.H. Kuok, C.Y. Liao, T.H. Wan, P.W. Yeh, I.C. Cheng, J.Z. Chen, Journal of Alloys and Compounds, 692 (2017) 558-562. [10] F.-H. Kuok, K.-Y. Kan, I.-S. Yu, C.-W. Chen, C.-C. Hsu, I.-C. Cheng, Applied Surface Science, 425 (2017) 321-328. [11] L. Liu, D. Ye, Y. Yu, L. Liu, Y. Wu, Carbon, 111 (2017) 121-127. [12] T. Liu, K. Chen, Q. Hu, R. Zhu, Q. Gong, Advanced Energy Materials, 6 (2016). [13] L. Meng, J. You, T.-F. Guo, Y. Yang, Accounts of chemical research, 49 (2015) 155165. [14] W. Nie, H. Tsai, R. Asadpour, J.-C. Blancon, A.J. Neukirch, G. Gupta, J.J. Crochet, M. Chhowalla, S. Tretiak, M.A. Alam, Science, 347 (2015) 522-525. [15] Y. Shao, Z. Xiao, C. Bi, Y. Yuan, J. Huang, Nature communications, 5 (2014) 5784. [16] N.K. Noel, A. Abate, S.D. Stranks, E.S. Parrott, V.M. Burlakov, A. Goriely, H.J. Snaith, 18
ACCEPTED MANUSCRIPT
ACS Nano, 8 (2014) 9815-9821. [17] D. Shi, V. Adinolfi, R. Comin, M. Yuan, E. Alarousu, A. Buin, Y. Chen, S. Hoogland, A. Rothenberger, K. Katsiev, Science, 347 (2015) 519-522. [18] F. Huang, Y. Dkhissi, W. Huang, M. Xiao, I. Benesperi, S. Rubanov, Y. Zhu, X. Lin, L. Jiang, Y. Zhou, Nano Energy, 10 (2014) 10-18. [19] X. Li, D. Bi, C. Yi, J.-D. Décoppet, J. Luo, S.M. Zakeeruddin, A. Hagfeldt, M. Grätzel, Science, (2016) aaf8060. [20] Q. Chen, H. Zhou, T.-B. Song, S. Luo, Z. Hong, H.-S. Duan, L. Dou, Y. Liu, Y. Yang, Nano Letters, 14 (2014) 4158-4163. [21] X. Xiao, C. Bao, Y. Fang, J. Dai, B.R. Ecker, C. Wang, Y. Lin, S. Tang, Y. Liu, Y. Deng, Advanced Materials, 30 (2018) 1705176. [22] J.L. Hodgkinson, H.M. Yates, A. Walter, D. Sacchetto, S.J. Moon, S. Nicolay, Journal of Materials Chemistry C, 6 (2018) 1988-1995. [23] J.-H. Tsai, I.-C. Cheng, C.-C. Hsu, J.-Z. Chen, Journal of Physics D: Applied Physics, 51 (2018) 025502. [24] H. Yu, H.I. Yeom, J.W. Lee, K. Lee, D. Hwang, J. Yun, J. Ryu, J. Lee, S. Bae, S.K. Kim, J. Jang, Advanced Materials, 30 (2018) 1704825. [25] H. Khachatryan, H.P. Kim, S.N. Lee, H.K. Kim, M. Kim, K.B. Kim, J. Jang, Materials Science Semiconductor Processing, 75 (2018) 1-9. [26] B. Vaagensmith, K.M. Reza, M.D.N. Hasan, H. Elbohy, N. Adhikari, A. Dubey, N. Kantack, E. Gaml, Q.Q. Qiao, ACS Applied Materials & Interfaces, 9 (2017) 35861-35870. [27] W. Yan, Y. Li, Y. Li, S. Ye, Z. Liu, S. Wang, Z. Bian, C. Huang, Nano Energy, 16 (2015) 428-437. [28] D. Huang, T. Goh, J. Kong, Y. Zheng, S. Zhao, Z. Xu, A.D. Taylor, Nanoscale, 9 (2017) 4236-4243. [29] J.-H. Im, I.-H. Jang, N. Pellet, M. Grätzel, N.-G. Park, Nature Nanotechnology, 9 (2014) 927-932. [30] Y.-L. Huang, A. Baji, H.-W. Tien, Y.-K. Yang, S.-Y. Yang, C.-C.M. Ma, H.-Y. Liu, Y.-W. Mai, N.-H. Wang, Nanotechnology, 22 (2011) 475603. [31] S.M. Jain, B. Philippe, E.M. Johansson, B.-w. Park, H. Rensmo, T. Edvinsson, G. Boschloo, Journal of Materials Chemistry A, 4 (2016) 2630-2642.
19