Low-Temperature Atomic Layer Deposition of Metal Oxide Layers for

Subscriber access provided by University of South Dakota

Energy, Environmental, and Catalysis Applications

Low Temperature Atomic Layer Deposition of Metal Oxides Layers for Perovskite Solar Cells with High Efficiency and Stability under Harsh Environmental Conditions Yifan Lv, Piaohan Xu, Guoqi Ren, Fei Chen, Huirong Nan, Ruqing Liu, Dong Wang, Xiao Tan, Xiaoyuan Liu, Hui Zhang, and Zhi-Kuan Chen ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.8b07346 • Publication Date (Web): 28 Jun 2018 Downloaded from http://pubs.acs.org on June 29, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Low Temperature Atomic Layer Deposition of Metal Oxides Layers for Perovskite Solar Cells with High Efficiency and Stability under Harsh Environmental Conditions Yifan Lv, Piaohan Xu, Guoqi Ren, Fei Chen*, Huirong Nan, Ruqing Liu, Dong Wang, Xiao Tan, Xiaoyuan Liu, Hui Zhang*, Zhi-Kuan Chen* Key Laboratory of Flexible Electronics (KLOFE) and Institute of Advanced Materials (IAM) Jiangsu National Synergetic Innovation Center for Advanced Materials (SICAM) Nanjing Tech University (NanjingTech), 30 South Puzhu Road, Nanjing 211816, P. R. China

Corresponding

e-mails:

[email protected];

[email protected];

[email protected]

Abstract Rapid progress achieved on perovskite solar cells raises the expectation for their further development towards practical applications. Moisture sensitivity of perovskite materials is one of the major obstacles which limits the long-term durability of the perovskite solar cells, especially in outdoor operation where rainfall and water accumulation on the solar panels often occur. Micro/nano pinholes within the functional layers of the devices usually lead to water vapor penetration, thus subsequent decomposition of perovskites, and finally poor device performance and shortened operational lifetime. In this work, low temperature atomic layer deposition (ALD) technique was utilized to incorporate pinhole-free metal oxide layers (TiO2& Al2O3) into an inverted perovskite solar cell consisting of ITO/NiO/Perovskite/PC61BM/TiO2/Ag. The interface properties between the inserted TiO2 layer and the perovskite layer were investigated by XPS study. The results showed that TiO2 ALD fabrication process had made negligible degradation to the perovskite layer. The TiO2 layer can significantly reduce interfacial charge recombination loss, improve interfacial contact, and enhance water resistance. Maximum power conversion efficiency (PCE) of 18.3% was achieved for devices with TiO2 interface layers. A stacked Al2O3 encapsulation layer was designed and deposited on top of the devices to further improve device stability under harsh environmental conditions. The encapsulated devices with the best performance retained 97% of the initial PCE after being stored in ambient for a thousand hours. They also showed great water resistance, no significant degradation in terms of PCE and photocurrent of the devices was observed after they were being immersed in deionized water for as long as two hours. Our approach offers a promising way of developing highly

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

efficient and stable perovskite solar cells under real world operational conditions. Keywords: perovskite solar cells, atomic layer deposition, interface, stability, encapsulation.

Introduction Solar cells, which convert ecologically friendly solar energy into electrical power using the photovoltaic effect, are expected to replace traditional fossil fuels and meet global energy needs. To effectively harvest the sunlight for power generation, many types of absorbing semiconductors have been invented, produced and commercialized. Amongst of them, organo-metal hybrid perovskites, such as CH3NH3PbI3, are one of the most promising photovoltaic materials due to their superior photophysical and optoelectronic properties, i.e., direct and tunable optical band-gaps, high hole/electron transport mobilities, cost-effective solution processability, high crystallinity and certain flexibility.1-6 Since their first application in photovoltaics in 2009,7 perovskite based solar cells (PSCs) have attracted intensive research attention owing to their rapid progress on PCE.8,9 which has rocketed up from 3.8%7 to 22.7%10 in a few years. Despite the great achievement in PCE of PSC devices, sustained operational stability of the devices under normal conditions still remains an obstacle to their further development towards practical applications. Perovskite materials are sensitive to water and oxygen in ambient environment and are readily to be decomposed in the presence of moisture.11,12 In a sandwich-like PSC device architecture, neighboring functional layers, i.e., charge transport layers and interface layers, can double their positions and act as effective barriers to prevent perovskites from contacting with water moisture permeated from ambient environment. However, the widely used charge transport layers in PSCs, i.e., PC61BM and Spiro-MeOTAD, normally show a high density of small micro/nano sized pinholes within the films formed during the spin-coating procedure.13,14 The pinholes provide pathways for moisture penetration into the perovskite layer, causing its decomposition; also, they act as channels for the immigration of iodine containing volatile species from perovskite to top electrode, causing corrosion of electrodes, thus device degradation.15,16 Thus, pin-holes free, charge transport layers are desired for realizing high moisture resistance, thus high stability of PSCs. Great research efforts have been devoted to address the stability issue of PSCs, from aspects of improving the intrinsic moisture resistance of perovskite materials and implementing continuous and compact pin-hole free functional layers, especially charge transport layers. Huang et al. addressed the underlining mechanism for morphology-dependent degradation of perovskite films and demonstrated a way of increasing the intrinsic stability of perovskites through scaling up its


Page 2 of 20

Page 3 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


grain size.17 Grätzel et al. introduced a phosphonic acid ammonium crosslinker additive into perovskite films, which facilities the growth of perovskite crystals and connects the neighboring grains into a mesoporous network through strong hydrogen bonding, resulting in an increase of both device performance and moisture stability.18 Wang et al. developed a lead-free Copper (II) based perovskite absorber with extraordinary hydrophobic behavior.19 Smith et al. used two dimensional hybrid perovskites (PEA)2(MA)2[Pb3I10] as absorbers in PSCs and found that they are much more moisture-resistant than 3D perovskite.20 Research achievements on charge transport layers mainly focused on the development of metal oxides charge extraction materials, as metal oxides present higher carrier mobility and superior stability than organic materials, such as solution processed inorganic NiO and ZnO implemented as hole/electron transport layers reported by Yang et al.;21 heavily p-doped NixMg1-xO and n-doped TiOx charge transport layers fabricated by spray pyrolysis, which facilities rapid carrier extraction and large area stable PSCs with >90% of the initial PCE remaining after 1000 hours of light soaking.22 Besides of inorganic metal oxides, functional organic transport layers were also developed. Huang et al., achieved highly efficient and stable PSCs with a cross-linkable silane-functionalized and doped fullerene layer.23 Atomic layer deposition is a thin film deposition technique widely employed to produce homogeneous, compact and pinhole-free films with finely controlled morphology and crystallinity.24 Recently, ultra-thin metal oxides deposited by ALD, i.e., NiO,25 TiO2,26,27 SnO2,28-31 Al2O332,33 and ZnO,34,35 have been explored as effective charge transport and encapsulation layers in PSCs, some works implemented direct deposition of metal oxides films onto perovskite layer successfully by low temperature (< 100 oC) ALD process.27,29-35 The metal oxide functional incorporated PSCs achieved both enhanced device efficiency and sustained stability against moisture. In this study, low temperature ALD technique was utilized to implement pinhole-free metal oxide layers (TiO2 & Al2O3) in an inverted perovskite solar cell device architecture consisting of ITO/NiO/CH3NH3PbI3/PC61BM/TiO2/Ag/Al2O3. The inserted ALD TiO2 layer between PC61BM/Ag together with PC61BM layer acted as a composite electron transport layer for reduced interfacial recombination loss. The interface properties, i.e., the elemental compositions and chemical states of elements and their variations during the TiO2 ALD process were studied by XPS measurement in detail. We found that our ALD process did not causing any damage to the perovskite layer through deliberately pre-depositing excessive Ti precursor at the beginning. A newly designed composite ALD Al2O3 layer in a stacked structure of “Al2O3 (10nm)/Al2O3 with intermediates of Al precursor (20nm)/Al2O3 (30nm)” was employed as an encapsulation layer to enhance the device resistance against water. We optimized the thickness of the ALD TiO2 layer in the devices in order to achieve desired water resistance and improved interfacial contact without affecting the electrical conductivity ACS Paragon Plus Environment


and light transmission. We compared performance of devices with TiO2 layer with the control device, and found the ALD TiO2 incorporated devices achieved an improved Voc of 1.04 V and an enhanced PCE of 18.3 %. We investigated performance and stability of devices incorporated with ALD TiO2 layer and Al2O3 encapsulation layer in both ambient and harsh environment. We found that 97% of the original PCE of the device remained in the most stable device after a thousand hours exposure to ambient air, and no significant degradation was observed even after the devices was immersed in deionized (DI) water for two hours. Our results show their significance of addressing the practical application of PSCs in outdoor environment.

Results and discussion Organic-inorganic hybrid perovskite materials CH3NH3PbI3 were employed as the photosensitive layers to fabricate solar cells in our study. Perovskite solar cells with an inverted device configuration of ITO/NiO/CH3NH3PbI3/PC61BM/Ag were employed as control devices to investigate the effects of functional ALD metal oxide layers on device performance, where solution processed NiO and PC61BM layers act as the hole and electron transport layers, respectively.22 The band alignments for this structure is shown in Figure S1. The HOMO and LOMO energy levels of the ALD TiO2 interface layer was measured and deduced by ionization potential spectroscopy (IPS) (Figure S2) combined with UV-vis absorption spectroscopy (Figure S3). The results are given in Table S1.An effective charge carrier cascade was finally set up in the device. On exposure to light, charge carriers are generated in the perovskite layer and electrons and holes are subsequently collected by their contact layers, NiO and PC61BM. The metal oxide functional layers TiO2 and Al2O3 were fabricated by employing low temperature ALD technique using tetrakis (dimethylamido) titanium (TDMAT), trimethylaluminum (TMA) and water vapour precursors, which present high reactivity and high vapour pressure.36,37 The schematic diagram of ALD deposition process is illustrated in Figure S4. In order to investigate the feasibility of depositing metal oxides onto perovskites by low temperature ALD technique, and to investigate whether the ALD TiO2 process causes damage to the underlying perovskite layer. ~60 nm thick ALD TiO2 layers were firstly attempted to deposit directly on CH3NH3PbI3 films (TiO2/CH3NH3PbI3/glass) without covering with PC61BM layer. CH3NH3PbI3 films on glass substrates were prepared with one-step synthesis method using lead iodide and CH3NH3I precursors at a molar ratio of 1:1 beforehand.38,39 The detailed ALD process is described in the experimental section. Thin-layer X-ray diffraction (XRD) measurements were conducted on perovskites films without/with ALD TiO2 layer (Figure S5). Perovskites films with TiO2 layer showed identical perovskite phase as the pristine perovskite film, with all crystallographic signatures matching well with the tetragonal perovskite crystal phase. The sharp


Page 4 of 20

Page 5 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


diffraction peaks at 14°, 28.3°, and 31.8° arise from the (110), (220) and (312) lattice planes and the peak intensities kept constant, confirming that the TiO2 layer could not shield the XRD signals. In addition, no characteristic XRD peaks from ALD TiO2 layer can be distinguished in the XRD pattern, which suggests that low temperature ALD TiO2 films were amorphous that is in agreement with previous studies.40-42 UV-vis absorption spectra were measured and presented a same profile with the absorption edge at ~800 nm (Figure S6), indicating that the optoelectronic properties of perovskite films were not influenced during the ALD process. Therefore, low temperature ALD had allowed deposition of TiO2 films on the CH3NH3PbI3 perovskite surface successfully. Scanning emission microscopy (SEM) measurement was carried out to examine the surface morphology of ALD TiO2 films on silicon substrate. Compact, uniform and continuous TiO2 films are shown in Figure S7. X-ray photoelectron spectroscopy (XPS) measurements were further carried out to systemically investigate the perovskite/TiO2 interface properties during the ALD TiO2 process, the XPS data analysis on Ti2p, O1s, Pb4f and Pb4f7/2 are summarized in Figure 1. The core energy level spectrum of Ti2p from XPS analysis is shown in Figure 1a. As expected, no XPS signal from Ti was observed in the Ti2p region over the surface of CH3NH3PbI3 film before ALD TiO2 deposition. As the number of ALD cycles increased, the thickness of TiO2 layer gradually increased. The XPS peaks for Ti2p1/2 and Ti2p3/2 began appearing and the peak intensity increased as well. It demonstrated the presence of TiO2 on the surface of CH3NH3PbI3layer, and the increased Ti content with increased number of ALD cycles. The binding energies for Ti2p1/2 and Ti2p3/2 were observed at 464.4 eV and 458.5 eV with the spin-orbital splitting of them of 5.9 eV, as reported.43 The core energy level spectrum of O1s from XPS analysis after is shown in Figure 1b. O signal can be observed over the surface of CH3NH3PbI3layer before the TiO2 ALD process, indicated by XPS peaks at 531.3 eV and 532.1 eV correlated to O-H bonds and O-Pb bonds, respectively, which originated from the unavoidable contact of the CH3NH3PbI3 surface with O2 and H2O, thus partial surface oxidation in ambient air.44,45 3 cycles of Ti precursor (TDMAT) was then deposited over the CH3NH3PbI3 layer. Consequently, a new XPS peak at 529.7 eV appeared owing to the formation of O-Ti bond and it increased with increasing the ALD cycles.44 It illustrated that the deposited TDMAT had reacted with H2O and other oxygen-containing groups adsorbed onto the perovskite layer and formed a TiO2 thin layer, also TiO2 is the main product after 40 ALD cycles. The chemical composition of titanium oxide was further analyzed by XPS spectra of Ti 2p3/2 (Figure S8). The spectrum was decomposed using a standard nonlinear-least-squares fitting procedure with the Voigt function. The Ti 2p3/2spectrum were decomposed into two components originating from Ti4+and Ti3+ with the binding energy difference of ~1.1 eV. The Ti4+ state was associated with the stoichiometric TiO2, showed by an XPS peak with much higher intensity; while the Ti3+ was an intermediate oxidation state (Ti2O3), ACS Paragon Plus Environment


showed by a weak XPS peak, which suggested an almost pure TiO2 layer was fabricated in our experiments.44 The core energy level spectrum of Pb4fare shown in Figure 1c. A prominent Pb4f7/2 XPS peak was observed at 138.6 eV correlated to the binding energy of Pb in the perovskite layer before TiO2 ALD process.46 Then 3 cycles of Ti precursor (TDMAT) was deposited over the CH3NH3PbI3layer. We found that the Pb4f7/2 XPS peaks showed an obvious decrease owing to the formation of a thin TiO2 layer over the CH3NH3PbI3 surface. In succession, 10 cycles of H2O precursor were deposited on top of the CH3NH3PbI3 layer with the thin TiO2 layer. The Pb4f7/2 XPS peak was then analyzed and decomposed (Figure 1d). A very weak XPS peak correlated to Pb-O bond at 138.0 eV and a strong peak at 138.6 eV correlated to the Pb in the perovskite lattice were observed. No signal for metallic Pb and Pb in degraded perovskite which are at 136.9 eV and 139.5 eV, can be seen.46,47 The XPS results demonstrated that the possible degradation on the perovskite layer can be avoided efficiently by pre-depositing excessive Ti precursor (TDMAT) at the beginning of TiO2 ALD process. The Pb4f5/2 and Pb4f7/2 peaks decreased with increasing number of ALD cycles due to the gradually increased thickness of TiO2 layer (Figure 1c). The core energy level spectrum of I3d (3d3/2 and 3d5/2) were also analyzed. It showed a similar trend of variation as that of Pb4f. The I3d5/2 peak at 619.2 eV was decomposed and was found to correlate only with I in the perovskite lattice. No signal for extra halides can be seen (Figure S9).48

Figure 1. XPS spectra of a) Ti 2p peaks including Ti 2p1/2 and Ti 2p3/2 b) O 1s peaks and c) Pb 4f peaks


Page 6 of 20

Page 7 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


including Pb 2p5/2 and Pb 2p7/2 of the CH3NH3PbI3 layer covered with different number of TiO2 ALD cycles. d) XPS spectra of Pb2p7/2 of the CH3NH3PbI3 layer and the CH3NH3PbI3 layer deposited with 3 cycles of Ti precursor followed by 10 cycles of H2O precursor.

A thin ALD TiO2 layer of ~2 nm was then deposited on top of PC61BM layer by 20 ALD cycles to form a composite electron transport layer in our devices. Atomic force microscopy (AFM) was

used

to

image

the

surfaces

of

CH3NH3PbI3,

CH3NH3PbI3/PC61BM

and

CH3NH3PbI3/PC61BM/ALD TiO2 in sequence. As seen from Figure 2a, b & c, the pristine CH3NH3PbI3 films achieved a surface roughness of 16 nm in RMS. The average crystal grain size was measured to be ~500 nm (Figure S10). CH3NH3PbI3 films coated with a 40 nm thick PC61BM layer obtained a reduced surface roughness of 4.7 nm. Pinholes existed in these films as shown in the AFM images. Through depositing a thin ALD TiO2 layer of less than 2 nm by 20 ALD cycles on top of PC61BM layer, the films were further flattened, and the surface roughness reduced to 2.7 nm in RMS. It is worth mentioning that the morphology and surface roughness of CH3NH3PbI3/PC61BM layer largely determined the growth direction of TiO2 nanoparticles, thus the surface roughness of TiO2 layer during ALD process, since ALD deposition is basically “conformal” and CH3NH3PbI3/PC61BM layer with spikes of nanometer size facilitates “sideway” growth of TiO2, which results in much smoother surface after deposition of a sufficiently thick TiO2.49 The deposition of ALD TiO2 layer could fill partial pinholes within the CH3NH3PbI3/PC61BM films, leading to a more compact contact with back electrode. The stability of CH3NH3PbI3 films, films coated with PC61BM and films deposited with ALD TiO2 were examined by X-ray diffraction (XRD) measurement. XRD patterns were achieved from fresh sample and samples kept in a chamber (BPS50CL) at constant temperature (20 °C) and humidity (50%) for five days (Figure S11). The X ray diffraction peaks’ intensity for pristine CH3NH3PbI3 films had reduced significantly after being stored for five days in the chamber. The diffraction peak observed at 12.6o arises from PbI2, which indicates CH3NH3PbI3 films had decomposed into PbI2 and CH3NH3I due to permeated moisture. Visually, the color of the perovskite films turned form dark brown to yellow as seen from the inserted picture in Figure S11. Solution processed PC61BM layer coated perovskite films also degraded after being stored for five days due to the intrinsic pin-holes in the PC61BM layer, as evidenced by the Xray diffraction peak from PbI2. Contact angle measurement was carried out to investigate the permeability of the PC61BM layer. As shown in Figure S12, a water droplet on PC61BM surface showed a contact angle of around 100 degrees at the beginning and decreased gradually. After two minutes, the droplet suddenly immersed into the film. Presumably, the water droplet penetrated through the PC61BM layers via pinholes, which were formed when being prepared via a simple spincoating process on the rough perovskite surface. Depositing an ALD TiO2 layer on the PC61BM layer



can effectively fill the pinholes in the PC61BM layer. ~2 nm and ~20 nm thick ALD TiO2 layers were deposited onto the PC61BM layer to test their capability of enhancing the resistance of perovskites against moisture. As shown in Figure S11c & d, thin TiO2 layer deposited CH3NH3PbI3 films also decomposed after five days, but thick TiO2 layer protected CH3NH3PbI3 films showed a greatly enhanced stability, evidenced by the XRD patterns with little change. However, to deposit a thick ALD TiO2 layer between the electrode and PC61BM layer is detrimental to the device efficiency owing to the lower charge transport mobility of the ALD TiO2.50 A TiO2 layer with an optimized thickness of ~ 2nm was inserted into our solar cells to form a composite electron transport layer combined with the PC61BM layer.

Figure 2. AFM images (5×5 µm) of a) perovskite film, b) PC61BM film covered in perovskite and c) 20 cycle ALD TiO2 layer on perovskite/PC61BM surface.

Devices consisting of ITO/NiO(30nm)/CH3NH3PbI3(320nm)/PC61BM(40nm)/ALD TiO2/Ag were then fabricated to investigate the effect of ALD TiO2 layer on the device performance, especially the operational stability. Performance of control devices were also investigated for comparison. Photoluminescence quenching experiments and time-resolved photoluminescence (TRPL) measurements were firstly carried out to investigate the recombination mechanism at the contact interface between the perovskite layer and the composite electron transport layer. On contact with charge transport layers, perovskites films normally show strong photoluminescence quenching as evidence of efficient charge transfer from the photoactive layer to the transport layer. Figure 3a shows the steady-state photoluminescence of glass/perovskite, glass/CH3NH3PbI3/PC61BM, glass/CH3NH3PbI3/PC61BM/TiO2. A clear quenching of more than 90% by 40 nm thick PC61BM was observed. Through modifying PC61BM layer with ALD deposited TiO2, a nearly 100% quenching efficiency was achieved due to reduced defect states in PCBM layer. The TRPL measurements of the CH3NH3PbI3 shows a decrease in the photoluminescence lifetime from 134.4 ns to 4.7 ns and 2.6 ns in the presence of PC61BM and PC61BM/TiO2, respectively (Figure 3b), indicating that charge carriers within the perovskite layer can be extracted more efficiently by PC61BM/TiO2 than a sole


Page 8 of 20

Page 9 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


PC61BM layer.

Figure 3. a) PL and b) TRPL spectra of MAPbI3 (black), MAPbI3/PC61BM (red) and MAPbI3/PC61BM/TiO2 (blue).

The J-V characterization curves of the control device is shown in Figure 4a, the short circuit current density (Jsc), open-circuit voltage (Voc) and fill factor (FF) were measured to be 22.2 mA/cm2, 1.01 V, 73.2% and 16.4%, respectively, which are comparable with previously reported results.51 To improve device’s resistance to humidity, an ALD TiO2 layer was deposited between PC61BM and Ag electrode of the control device by ALD process. The thickness of TiO2 layer was optimized in order to achieve an optimized device performance. Performance of devices with TiO2 layer of various thicknesses is summarized in Table S2. An optimum TiO2 layer thickness was found to be around 2 nm, which was deposited through 20 ALD reaction cycles. Any increase in the thickness was detrimental to the device performance due to the low conductivity of the amorphous ALD TiO2.33 The best performance with a Voc of 1.04 V, a Jsc of 22.8 mA/cm2, an FF of 76.9% and a PCE of 18.3% was achieved as shown in Figure 4a. The performance comparison of devices in Voc, Jsc and PCE with/without TiO2 layer is summarized in Figure 4b, c & d. In order to examine the consistency of device performance in Voc and PCE, 40 devices for each type were fabricated and the characterization results are summarized in Figure 4c & d. Compared with the control device, Voc showed an increment from 1.01V to 1.04 V, resulting from more favorable band alignment and hole blocking effect by the ALD TiO2 layer for spatial separation of electrons and holes, thus reduced recombination loss at the interface.13,31,35 No apparent improvement in Jsc and FF was observed (Figure S13). A short stability test under continuous light irradiation was carried out as well. Figure 4b shows the evolution of device performance parameters, i.e., Jsc and PCE for sixty seconds during stability test. The control devices achieved a stable photocurrent density of 20.5 mA/cm2 and stabilized efficiency of 16.4% at an applied bias of 0.8 V (black curves); while the ALD TiO2 modified devices showed a photocurrent density of 20.7 mA/cm2 at an applied bias of 0.84 V,



resulting in a stabilized efficiency of 17.4% (red curves).

Figure 4. a) J-V curves, b) Steady-state photocurrent and efficiency and plot of c) Voc and d) PCE of the devices with (red) or without (black) TiO2 interfacial layer, the inset in a) is the device configuration. In c) and d), X mark indicates the maximum and minimum values, small check inside the marks the average values.

In order to further improve devices’ stability (resistance to water) and retard degradation of the devices, an additional ALD Al2O3 layer was then deposited on top of the entire device to act as an encapsulation layer. Al2O3 is the most frequently-used permeation barrier against water and oxygen, and has been employed in various optoelectronic devices as encapsulation layers owing to its superior properties, i.e., compactness, pin-hole free, continuous and high chemical resistence.16,32,33 A SEM image of ALD deposited Al2O3 layers on silicon substrates is shown in Figure S14. The ALD process was completed at 60 oC. We monitored the stability of devices with and without an ALD Al2O3 encapsulation layer under ambient environment (25 oC, 40-60% in humidity). As seen from Figure 5a, the unencapsulated devices degraded quickly and achieved a T80 of 16 hours due to the rapid permeation of moisture into the devices and decomposition of perovskite layer. T80 is the period for devices to degrade 20% in PCE. The devices with TiO2 interface layers achieved a T80 of 40 hours due to the poor coverage of TiO2. When protected by a 60 nm compact ALD Al2O3 layer, the device operational stability was significantly improved and the T80 increased to six hundred hours. However, six hundred hours operation lifetime of the devices is still unfavorable for practical applications in real environment. We attributed the limited operational lifetime of the devices to the limited


Page 10 of 20

Page 11 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


resistance against water due to the intrinsic hydrophilicity of Al2O3 film, as proved by contact angle measurement of ALD Al2O3 (Figure S15). Water vapor can slowly diffuse through Al2O3 layer and destroy the devices. The mechanism of penetration of water through Al2O3 layer is proposed and illustrated in Figure 5b. Water molecules form H-bonds with oxygen on the surface and accumulate. The attached water molecules then propagate slowly from the surface into the film with a very low diffusion rate and form more H-bonds with the intrinsic oxygen inside the film. Finally, it moves towards and reaches the perovskite layer and decomposes it.

Figure 5. a) Evolution of normalized PCE of the control devices (black), devices with TiO2 interface layer (red), devices with TiO2 interface layer and Al2O3 encapsulation layer (blue) and devices with TiO2 interface layer and stacked Al2O3 encapsulation layer (green) stored in an ambient environment. b) Mechanism of water permeation and water blocking process by the stacked Al2O3 encapsulation layer. c) Evolution of normalized PCE for the Al2O3 protected devices immersed in water. The inset picture shows photographs of the devices with Al2O3 (right) and without Al2O3 (left) layers after being kept in water for an hour. d) The steady-state photocurrent of the devices with (red) or without Al2O3 protection layer (black) vs time while devices were immersed in DI water.

To slow down the penetration rate of water across the Al2O3 layer and further prolong the operational lifetime of our devices, we designed and fabricated novel structurally sandwich-like Al2O3 encapsulation layer, which consists of three layers, i.e., a 10 nm compact Al2O3 bottom layer touching with the devices, a 20 nm intermediate Al2O3 layer containing extra reaction intermediates



Page 12 of 20

O-Al-(CH3)3-x53 and a 30 nm compact Al2O3 layer at the most surface. The top 30 nm Al2O3 layer provides the first line of defense against water penetration when it attacks this stacked encapsulation layer, the 20 nm intermediate Al2O3 layer would react with water molecules and generate more compact Al2O3 very slowly when they reach it due to the extra intermediates (Figure 5b).37 The detailed reaction mechanism of the formation of the intermediates can be illustrated by the following reactions53: xOH(s)+Al(CH3)3(g) → xO-Al-(CH3)3-x(s)+xCH4(g)

(1)

xO-Al-(CH3)3-x(s)+(3-x)H2O(g) → xO-Al-(OH)3-x(s)+(3-x)CH4(g) (2) xO-Al-(OH)3-x(s)+ 2H2O(g) → O-Al-(OH)2(s)+ (3-x)CH4(g)

(3)

Al-O-Al(s)+H2O(g) ↔ -Al-OH(s)+-Al-OH(s)

(4)

The reactions consume and eliminate water as well as prolong the diffusion time of water due to more newly produced Al2O3. The whole process is presumably very slow as the water penetration is slow. The bottom 10 nm Al2O3 layer prevents devices from chemical corrosion by the reactive TMA in the intermediate layer. XPS measurement was carried out to investigate this deduction. The core energy level XPS spectrum of Al and C were analyzed (Figure S16). The signals of Al2p were collected from the ALD Al2O3 layer and the Al2O3 intermediates, and the signal of C1s was collected from the Al2O3 intermediates. Compared with the Al2p spectrum of the normal ALD Al2O3 layer (Peaking at 74.3 eV),46,47 an additional shoulder peak at 73.6 eV was observed from the Al2O3 intermediates, which is correlated to Al-C bond. The signal of C-Al bond was also found in the C1s spectrum of the Al2O3 intermediates at 283.6 eV.54 The C1s spectrum can be decomposed into three components, including signals of C-H bond, C-O bond and C-Al bond. The results demonstrated that O-Al-(CH3)3-x should be the main component of the Al2O3 intermediates. The cross-sectional SEM image for the stacked Al2O3 layer encapsulated device is shown in Figure S17, it can be seen that the composite Al2O3 layer was compact and had a full coverage over the device. Our CH3NH3PbI3 solar cells protected with this novel encapsulation layer exhibited excellent stability, 93% of the initial PCE remained after being stored in ambient for a thousand hours (Figure 5a) and the encapsulated device with the best performance can remain 97% of the initial PCE (Figure S18). Additionally, we also investigated devices’ resistance against water. Both control devices and devices with TiO2 layer and stacked Al2O3 layer were immersed in DI water for one hour. During the process, the color of control devices visually turned into yellow within one minute, indicating the decomposition of CH3NH3PbI3 into PbI2. No visible change was observed from the devices


Page 13 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


encapsulated with Al2O3 (Figure 5c). More visual details can be found in a video record in supporting information. Figure 5c also shows the PCEs of both control devices and encapsulated devices after they were immersed in DI water for various durations. The encapsulated devices can retain more than 95% of its initial efficiency after two hours of soaking in DI water. The steady-state performance of the device in DI water environment was also monitored. As shown in Figure 5d, the photocurrent of the control devices decreased by 80% in ten seconds and reduced to zero within one minute. The encapsulated devices showed much better water resistance, and the photocurrent remained relatively stable within two minutes.

Conclusion In conclusion, we have fabricated organic-inorganic CH3NH3PbI3 solar cells with an inverted device configuration of ITO/NiO/CH3NH3PbI3/PC61BM/Ag. Through low temperature ALD process, an interface TiO2 layer was deposited onto the PC61BM layer in the solar cells to form a composite electron transport layer and negligible degradation to the perovskite layer during ALD process can be seen, evidenced by a detailed XPS study on the interface. The compact TiO2 layer could fill up partial pinholes in the PC61BM layer, thus improve the resistance of underlying perovskite layer against water vapor. The TiO2 layer facilities more favorable band alignment and hole blocking effect for spatial separation of electrons and holes, thus reduced recombination loss at the interface. An improved Voc of 1.04 V and an enhanced efficiency with a PCE of 18.3 % were achieved from the ALD TiO2 incorporated devices. Through depositing a 60 nm ALD Al2O3 layer onto the devices to act as an encapsulation layer, the operational lifetime of the devices achieved an increment by a factor of ~38. Furthermore, a composite ALD Al2O3 layer in a stacked structure of “Al2O3 (10nm)/Al2O3 with intermediates of Al precursor (20nm)/Al2O3 (30nm)” was designed to be employed as an encapsulation layer. Devices protected with this novel encapsulation layer exhibited excellent stability, the encapsulated devices of the best performance retained 97% of the initial PCE after being stored in ambient for a thousand hours. Devices encapsulated with composite Al2O3 layers also showed great water resistance, more than 95% of the initial efficiency can be retained after being immersed in DI water for two hours. Our work shows that incorporation of low temperature ALD metal oxides functional layers into the perovskite photovoltaic devices can be quite promising to improve the device performance, enhance the resistance of devices to water vapor and prolong the device operational lifetime simultaneously.

Experimental Section Materials: Lead iodide(PbI2) and methylammonium iodide (MAI, CH3NH3I) were purchased



Page 14 of 20

from Xi’an Polymer Light Technology Corporation. Dimethyl sulfoxide, chlorobenzene, dimethylformamide, Ni(NO3)2·6H2O, 1,2-ethanediamine and ethylene glycol were purchased from Sigma-Aldrich. PC61BM was purchased from Borun New Material Technology Co., Ltd. TDMAT and TMA were purchased from MNT® Micro and Nanotech Co., Ltd.

Device fabrication: The ITO glasses were cleaned in cleaning agent, deionized water, acetone, and isopropanol under sonication for 10 min in sequence. Then they were dried by N2 and treated by UV-ozone plasma for 20 min. The procedure for preparing NiO hole transport layer was the same as previously reported method.21 1.25 mol PbI2 and MAI were dissolved in 1mL mixed solvent consisting of DMF and DMSO (4:1 in volume) and stirred at 60 Z overnight. The solution was dropped onto NiO coated ITO substrates at a spin speed of 4000 r.p.m. for 40 s in a glove box. After spin-coating for 8 s, 0.3 mL of PC61BM in chlorobenzene at 6 mg/mL was quickly dropped onto the sample surface to form the CH3NH3PbI3 layer,52 followed by solvent annealing with DMF at 100 Z for 10 min.39 15 mg/mL PC61BM in chlorobenzene was then spin-coated onto the CH3NH3PbI3 layer at 3000 r.p.m. Then the TiO2 interface layer was deposited by ALD process. 120 nm thick Ag electrodes were deposited under a high vacuum (

Low-Temperature Atomic Layer Deposition of Metal Oxide Layers for

Recommend Documents