Improvement of the Dynamic Spin-Washing Effect with an Optimized

Subscriber access provided by GRIFFITH UNIVERSITY

Article

Improvement of Dynamic Spin-Washing Effect with Optimized Process of Perovskite Solar Cell in Ambient Air by Kriging Method Atthaporn Ariyarit, Ryohei Yoshikawa, Issei Takenaka, Frédéric Gillot, and Seimei Shiratori Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.7b02515 • Publication Date (Web): 08 Sep 2017 Downloaded from http://pubs.acs.org on September 9, 2017

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 free 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 accessible to all readers and 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.

Industrial & Engineering Chemistry Research 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 23

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

Industrial & Engineering Chemistry Research

Improvement of Dynamic Spin-Washing Effect with Optimized Process of Perovskite Solar Cell in Ambient Air by Kriging Method Atthaporn Ariyarit1, Ryohei Yoshikawa1, Issei Takenaka1, Frédéric Gillot2 and Seimei Shiratori1,* 1

School of Integrated Design Engineering, Keio University, 3-14-1 Hiyoshi, Kohoku-ku, Yokohama, Kanagawa-ken 223-8522, Japan Tel: + 81 - 45 - 566 - 1602, Fax: +81- 45 -566 - 1602 [email protected] 2 Laboratoire de Tribologie et Dynamique des Systèmes, Ecole Centrale de Lyon-36, Avenue Guy de Collongues, 69130 Ecully, France

ACS Paragon Plus Environment


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

Abstract Lead halide perovskite solar cells (PVSCs) exhibit high efficiencies and are based on abundant materials; further, they can be deposited on flexible substrates using easy fabrication processes. Here, we report a way to fabricate a 1-cm2 perovskite solar cell outside a glove box (humidity below 40%) via a modified two-step method with a FTO /TiO2/perovskite/Spiro-OMeTAD /Au structure. This modified two-step method is based on the "dynamic spin-washing method". After dipping the perovskite films into a methylammonium iodide solution, they were washed with isopropyl alcohol (IPA) at different spin speeds. The surface morphology and crystallinity of the perovskite layer was drastically improved using this method. We also investigated and conducted an experiment to explain the effect of the dynamic spin-washing on the surface morphology and crystallinity. Lastly, we optimized the thickness of the TiO2 layer and observed a maximum power conversion efficiency of 13.75% and 11.78% for small- and large-area devices, respectively. Keyword: Perovskite solar cell, Washing process, two-step method, Kriging Graphic abstract


Page 2 of 23

Page 3 of 23

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


Introduction Various material combinations have been used in the study of lead halide perovskite solar cells (PVSCs), which could potentially be fabricated as flexible devices and also fabricated with large areas as a commercial product.1-3 Introducing the perovskite crystal for solar cells has many benefits, such as a high absorption coefficient, long carrier lifetime, and low recombination rate with a tunable bandgap.4-5 In 2016, Sabela et al. achieved a maximum efficiency of PVSCs over 22%.6 The efficiency of the PVSCs has been improved by testing different fabrication techniques, device structures, and optimizing the experimental parameters.7-9 Most of the reports on improving the fabrication methods of PVSCs are focused on small device active areas (i.e. an active area is smaller than 0.2 cm2). There have only been a few investigations into cells fabricated with areas of several centimeters squared via methods such as doctor-blading, inkjet printing, spray deposition, and spin coating.10-13 Le et al. fabricated large-scale PVSCs using a vacuum flash–assisted solution process.14 By using such processes, the crystalline quality of perovskites has improved and power conversion efficiencies (PCE) above 20% have been achieved. Gouda et al. developed a vapor treatment method for the surface of perovskite films and thus they achieved a PCE of around 15%15; their findings demonstrated that improving the crystalline quality and homogeneity of perovskite films are beneficial for scaling up PVSCs products. However, the studies were all conducted with fabrication processes requiring a glove box with an inert atmosphere to prevent the perovskite layer from being damaged by the vapor. The difficulty of controlling crystalline quality and homogeneity of perovskite layer is a bottleneck for the reproducibility in fabricating large-scale PVSCs under ambient air.16-18 To date, there have only been a few reports on the fabrication of large-area PVSCs in ambient air despite many research groups successfully developing PVSC fabrication processes in ambient air for small device areas (1 cm2. The difficult point to solve these problems for centimeter area is the environment effect making the perovskite surface non-uniformity that leading to hard for re-processing and high error rate of PCE. Therefore, to understand the effect of vapor during the fabrication process is necessary to the future development of fabrication process of perovskite thin film.



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

Here we designed a fabrication process for a perovskite layer that is referred to as the "dynamic spin-washing method", which involves a conventional mesoporous TiO2-based device to produce cells with areas >1 cm2 at atmospheric pressure (for a humidity level below 40%). The key component of this two-step method is the "dynamic spin-washing" process conducted after dipping the film into a Methyl Ammonium Iodide(MAI) solution. This experiment demonstrates that our washing process has the potential to improve the uniformity and quality of perovskite crystalline thin films. We also present the important mechanism for the improvement of the spin-washing process in ambient air condition. It has strong effect for making the surface of perovskite more uniform nd lower amount of pin-hole. Moreover, we designed the experiments so as to optimize the deposition conditions for the electron transporting TiO2 layer. After optimizing our process, we achieved a PCE of 13.75% for small-area devices and 11.8 % for large-area devices.

Experiment Synthesis of Methyl Ammonium Iodide (MAI)22 In this experiment, we synthesized MAI powder by mixing 5.8 mL of hydroiodic acid (57 wt% in water, Sigma-Aldrich), 5 mL of methylamine (40 wt% in methanol, Tokyo Chemical Industry Co. Ltd.), and 20 mL of ethanol in a round bottom flask at 0 °C (ice bath) and sintering for 2 h under vacuum. Then the MAI was crystallized out of solution via rotary evaporation of the solvent at 50 °C. The MAI powder was washed three times with diethyl ether (99%, Wako) in an ultrasonic bath and dried under vacuum at 50 °C for 8 h.

Device Fabrication FTO glass substrates (2 cm × 2.5 cm) were etched with HCl and zinc power. After that, the pattern-etched FTO substrates were cleaned by ultrasonication in deionized water, acetone, and ethanol (20 min in each solvent) and dried by blowing air onto them. Then the substrates were heated at 60 °C for 30 mins to remove the remnant solvent. The dense-TiO2 layer was fabricated by spin coating a solution composed of 0.1 M titanium diisopropoxide bis(acetylacetonate) (Sigma-Aldrich) dissolved in n-butanol (99%, Wako) at 3000 rpm. Then, the samples were annealed at 150 °C for 5 min. These processes were repeated three times to deposit a dense TiO2 precursor film. After that, the films were annealed at 500 °C for 60 min. After cooling, the mesoporous-TiO2 layer was deposited by spin coating 0.4 g of TiO2 nanoparticles diluted in ethanol (1.4 g ) at 3000 rpm for 30 s; the samples were then postannealed at 500 °C for 60 min. To prepare the perovskite layer with the two-step method, 1 M of PbI2 was dissolved in a mixed solvent that contained dimethylformamide (DMF, 99%, Wako) and dimethyl sulfoxide (DMSO 99%, Wako). The volume ratio of the solvents was 9:1. The samples and precursor solutions were preheated at 70 °C before deposition. The precursor solution was coated onto the substrate by spin coating (typically at 500 rpm for 3 s and subsequently at 5000 rpm for 30 s). The PbI2 films were annealed at 70 °C for 15 min to remove the remnant solvent in the film. After the surface temperature of the PbI2 thin films decreased to room temperature, the samples were typically dipped into the MAI solution (10 mg mL-1 in IPA) for 90 s. Then the perovskite thin films were transferred to the spin coater to


Page 4 of 23

Page 5 of 23

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


remove the remaining solution on the surface by spinning at 4000 rpm with IPA for 30 s. After that, the films were annealed at 90 °C for 30 min. To fabricate the hole transport layers, the samples were spin coated at 4000 rpm with a solution of the hole transport material. This solution was prepared by mixing 68 mM of spiro-OMeTAD (Sigma-Aldrich), 26 mM LiTFSI (Sigma-Aldrich), and 55 mM 4-tert-butylpyridine (TBP, Sigma-Aldrich) in acetonitrile and chlorobenzene (V/V = 1:10). Finally, 100-nm-thick gold electrodes were deposited by magnetron sputtering. The active device area was 0.09 cm2 for the small surface area and 1 cm2 for large surface area devices. The overall process for this experiment is shown in Figure 1.

Figure 1 Schematic illustration of the fabrication of perovskite solar cells based on the modified two-step spin method proposed in this study.

Characterization Surface topographies of perovskite layer were observed by field emission scanning electron microscope (FE-SEM, S-4700, Hitachi, Japan) with an accelerating voltage of 5 kV. The thicknesses of each layer were measured by stylus profilometry (Dektak 3030, ULVAC, Japan). Light transmittance over the range of 200 to 1000 nm was measured using spectrophotometry (UVmini-1240, Shimadzu, Japan). Crystallization of the perovskite layer was analyzed by X-ray diffraction (XRD, D8, DISCOVER, Bruker, Japan) using Cu Kα radiation (40 kV, 40 mA), with a scan rate and a step size of 2 °/min and 0.01 °/min, respectively. The surface elemental composition of the samples was identified by XPS (JEOL JPS-9000MX). The current density-voltage (J-V) characteristics of the solar cells were measured using an AM 1.5 solar simulator (100 mW cm−2) and an automatic polarization system (Hokuto Denko HSV-100) that used a 500 W Xe lamp (UXL-500SX, Ushio) as a light source.



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

Kriging Model Method Kriging model23,24 is an expression of unknown function yˆ ( x) as yˆ ( x ) = µ ( x ) + ε ( x )

(1)

where µ ( x ) is a global model and ε ( x) is a local model. Sample points (x) are interpolated with Gaussian random function. The correlation between Z ( xi ) and Z( x j ) is related to the distance between the two corresponding points x i and x j . The distance function between the point xi and x j is expressed as m

d ( xi , x j ) = ∑θ k | xik − xkj |2

(2)

k =1

where θ k (0 ≤ θ k ≤ ∞) is the k th element of the correlation vector parameter, θ . The correlation between the points, x i and x j , is defined as Corr[ Z ( xi ), Z ( x j )] = exp[ − d ( xi , x j )]

(3)

Kriging prediction can be expressed as,

yˆ ( x) = µˆ ( x) + rT R −1 (F − µˆ )

(4)

where F = [ f ( x1 ), f ( x2 ), f ( x3 ),..., f ( xN )] is the value of the evaluated function value at

Xn = {x1 , x2 , x3 ,..., xn } , R denotes the n× n matrix whose (i, j ) entry is Corr[ Z ( xi ), Z ( x j )] , r is the vector whose i th element is ri ( x ) = Corr[ Z ( x ), Z ( x i )]

(5)

The unknown parameter, θ for the Kriging model can be estimated by finding the maximum likelihood estimation (MLE): n 1 Ln ( µ , σ 2 , θ ) = − ln (σ 2 ) − ln (| R |) 2 2


(6)

Page 6 of 23

Page 7 of 23

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


MLE is an m -dimensional unconstrained non-linear optimization problem. In this paper, a genetic algorithm (NSGA-II)23 is used to solve this problem. For a given θ , σ 2 can be defined as

σ2 =

(F − µˆ ) ' R −1 (F − µˆ ) n

(7)

Then, µ ( x ) is assumed to be a constant value in the original Kriging as

µ ( x) =

1′R −1F 1′R −1 1

(8)

The mean square error s 2 ( x ) at point x of this function can calculate by the following equation: s 2 ( x ) = σ [1 − r ′R −1r +

(1 − 1R −1r ) 2 ] 1′R −1 1

(9)

where 1 denotes an n -dimensional unit vector.

Result and Discussion In this experiment, we fabricated a perovskite solar cell by modifying a two-step spin method for a mesoporous TiO2 based structure, namely for a FTO/dense-TiO2/mesoporous TiO2/perovskite/spiro-OMeTAD/Au stack. This structure has already been improved to achieve a very high PCE (over 20%) in many reports.25,26 However, only a few studies succeeded to fabricate this structure at atmospheric pressure and for a large surface area; this is because of the difficulties in controlling film uniformity so as to achieve fewer pin-holes and a high reproducibility of the film quality.27 To modify the two-step spin method, we changed the washing process for the perovskite film by dropping IPA onto the samples during the spin coating of the film. Figure2 shows the perovskite surface characteristics via SEM and AFM images. From this figure, we can observe that the surface morphology changes depend on the spin speed of the IPA during the washing of the perovskite thin film. In case of 2000 rpm for spin-speed washing, the grain size is larger than 3000,4000 rpm; cross section image also confirms the larger grain size at lower washing spin-speed. The reason of larger grain size at lower washing spin-speed is the re-crystalline of PbI2 crystal with vapor around the interface between perovskite film surface and vapor in the air.28 Therefore, to study the effect of vapor around surface during film forming stage is necessary that explain in the next section. The surface of the perovskite thin films appear smoother for higher washing spin speeds; the lower the surface roughness of the perovskite films is one of the important factors to improve



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

the power efficiency. Moreover, we observed that the surface roughness of the perovskite films dropped from 74 nm when washing at 2000 rpm to 35 nm when washing perovskite thin film at 4000 rpm. The surface roughness of the perovskite films was measured by AFM. Not only did the smoothness of the perovskite film surface increase, the number of pin-holes also decreased when the washing spin speed was increased.

Figure 2 Surface morphology analysis and cross section image of perovskite layer when washing the surface with different spin speeds The crystalline quality of the perovskite films is an important parameter to optimize the PCE of PVSCs. One way to investigate the quality of a perovskite layer is via XRD. Figure 3 shows the XRD pattern of a perovskite thin film washed with IPA at different spin speed. In a previous report on assessing the crystalline quality of a perovskite thin film, Hsieh et al. defined the conversion ratio of PbI2 into methylammonium lead iodide (MAPbI3) with the following equation;29

.

. .


(10)

Page 8 of 23

Page 9 of 23

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


where I12.7 is the intensity of the PbI2 peak at 12.7° and I14.2 is the intensity of the MAPbI3 peak at 14.2°. The values of the conversion ratio when changing the washing spin speed are shown in the Table S1. We found that a large amount of PbI2 remained at low washing spinspeeds. A higher conversion ratio of MAPbI3 was observed when we increased the spinwashing speed. To compare the conversion ratio of MAPbI3, the maximum value of is required. Therefore, we confirm the maximum by fabricating perovskite layer inside the glovebox without IPA washing. The value was calculated as 0.94. This value is a little bit higher than washing with IPA at 4000 rpm(0.92), so it can confirm the benefit of washing with IPA at high spin-speed can prevent the converting of MAPbI3 phase to PbI2 phase.

Figure 3 XRD pattern of perovskite films after washing them with IPA at different spin speeds We designed an experiment to investigate the phenomena that cause the formation of high roughness films and different conversion ratios for different spin speeds. First, a PbI2 thin film was deposited via spin coating on top of the TiO2 layer and dried to evaporate the solvent in the PbI2 thin film; after that, the film was dipped into the MAI solution. Then, the sample was mounted onto the holder of the spin coater and we waited until some areas of the film were dried (as shown in a photograph in Figure 4); subsequently, we conducted the dynamic spin-washing step. Figure 4 shows a picture of the sample before and after using the dynamic spin-washing method, highlighting the surface morphology of the dry and the wet



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

area. The dry area of the sample reveals a non-uniform perovskite film with a rough surface area. Further, Figure 4 demonstrates that the surface of the perovskite film was directly affected by the IPA evaporation: this is because we observed a smooth and uniform surface in the wet area of the film before the dynamic-spin drying process. Not only was the film's surface affected by the presence or absence of the solvent, but the crystalline quality of the PbI2 appeared also to be much higher in the wet area than in the dry area; the XRD result of the dry surface area and wet surface area are shown in the Supporting Information. To explain the reason for the differences in the structures and the quality of PbI2 when washing and drying the film with difference spin-speeds, we needed to model the results of Koo et al.30 In that report, they investigated the effects of water and IPA content in the air in a single wafer cleaning system. Their simulation showed that the content of water and IPA in the air and the spin-speed rate are related to the evaporation rate of the IPA and water; further, a higher evaporation rate of both water and IPA was confirmed when the spin-speed was increased. Additionally, the evaporation rate of water is lower than that of IPA owing to the higher boiling point of water. In humid conditions, it is necessary to consider the influence of water condensation. This phenomenon is harmful to the surface of perovskite films because the perovskite surface readily reacts with water, converting the surface back to PbI2.31

Figure 4 a) Photograph of the sample before and after using the dynamic-spin washing method with the SEM surface images of b) wet area and c) dry area Figure 5 shows the schematic of the film formation process for different film surfaces. For the dry part of the film at atmospheric pressure, the amount of condensed water (from water vapor in the ambient air) is higher than the amount of water evaporated from the perovskite


Page 10 of 23

Page 11 of 23

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


surface; thus, the presence of ambient vapor that condenses on the surface converts an additional portion of the perovskite film into PbI2. Moreover, the evaporation rate of IPA is non-uniform owing to the amount of IPA on the surface not being the same across the entire area of the film; for example, areas at the edge of the film have a lower IPA content than central areas. For this reason, the surface of a normal dry process results in a non-uniform film with a high surface roughness. Moreover, it can be seen from a cross sectional image (shown in Figure 6) that the wet and dry areas have different thicknesses; different surface morphologies of the perovskite film were formed during the waiting step while the films dried.

Figure 5 Schematic of explanation of the origin of the different surface structures when washing the surface of the perovskite films with IPA at different spin speeds. The MAI solution did not dry uniformly and some areas retained a small amount of solvent. In the completely dried area, the film thickness gradually decreased as the size of the crystallites from the re-converted perovskites to PbI2 grew. Schlipf el al. reported that the size of a perovskite crystal is larger than that of PbI2.32 At a low washing spin-speed, the water evaporation rate is still higher than the water condensation rate, and thus the surface of the film also has a high surface roughness. However, the conversion of perovskite crystals into PbI2 is lower than for the normal dry fabrication case because of the reduced amount of water on the surface owing to a higher rate of water evaporation. At high washing spin-speeds, the water evaporation rate is lower than the rate of water condensation. Thus, most of the perovskite surface has a low surface roughness with a low amount of residual PbI2, because the surface of the film is not converted into PbI2 owing to the low amount of water in the vicinity of the film surface.



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

Page 12 of 23

Figure 6 Cross-sectional schematic of the area with the residual solution and the dry area at the boundary between difference perovskite surface morphologies. Table 1 shows the J–V parameters for the small area PVSCs fabricated with different IPA spin-washing speeds. These results show a correlation between the spin speeds and the surface morphology and crystalline quality of the perovskite films. The PCE of the PVSCs gradually increased as we increased the spin-washing speed. We observed a significant increase in the fill factor (FF) and short-circuit current density (Jsc); the FF increased from 0.47 to 0.59 and Jsc increased from 16.5 to 22.5 mA/cm2. The J–V curves also confirmed that the dynamic spin-washing method can improve the PCE of PVSC devices. In particular, in the case of the 1-cm2-area device, the PCE of the devices when washing at 4000 rpm was almost double than those at 2000 rpm. Table 1 The J–V parameters of the small area of PVSCs fabricated at different IPA spinwashing speeds Washing spin speed

Jsc

Voc

(rpm)

(mA/cm2)

(V)

2000

16.5

0.91

0.47

7.2

3000

18.2

0.97

0.58

10.3

4000

22.5

0.95

0.59

12.7

FF

PCE (%)


Page 13 of 23

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


The J–V parameters of large area PVSCs fabricated with different washing IPA spin speeds are shown in Table 2. When we fabricated perovskites with a large surface area, the benefit of our method could be more clearly observed because of the higher crystalline uniformity and area of the perovskite coverage required for the large-scale device.33,34 The FF increased from 0.32 to 0.53 when we increased the washing speed from 2000 to 4000 rpm and the Jsc increased from 15.3 to 19.7 mA/cm2. Table 2 The J–V parameters of the 1 cm2 area of PVSCs fabricated at different IPA spinwashing speeds Washing spin speed

Jsc

Voc

FF

PCE

(rpm)

(mA/cm2)

(V)

2000

15.3

0.95

0.32

4.79

3000

17.3

0.98

0.41

6.95

4000

19.7

0.98

0.53

10.3

(%)

To optimize the efficiency of the PVSC devices in this experiment we chose the experimental methods using the kriging model to optimize the TiO2 layer; this method was already used in our previous studies.35 In this study, we optimized the TiO2 layer, as this layer has a large impact on increasing the Voc and Jsc by reducing the recombination at the interface between the perovskite layer and the FTO substrate36. In a device with a mesoporous structure, there are two kinds of TiO2 layers, the first of which is a dense TiO2 layer and the second of which is a mesoporous TiO2 layer. Therefore, we selected the input parameters to be the spin-speed for fabricating a dense TiO2 layer (Spin Speed 1) and the spin-speed for fabricating a mesoporous TiO2 layer (Spin Speed 2). To optimize the real effect of the TiO2 layer and to reduce the cost of optimization, we designed a half-cell that has no hole transporting layer. Before starting to design the experiment, we set the parameter values of the dense-TiO2 layer by fabricating devices with only a dense-TiO2 layer and a perovskite layer (FTO/denseTiO2/perovskite/Au). By changing the spin-speed for fabricating the dense TiO2 layer we estimated the range of input parameters. The detail of estimated the range of input parameters is shown in supporting information. We started the design of the experiment by choosing the parameters by Latin hypercube sampling (LHS). Table S4 shows the input parameters for fabricating the dense and mesoporous TiO2 layers assigned by LHS37. The benefit of random data sampling by LHS is



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

that we can decrease the number of experiments while maintaining uniformity throughout the experimental space. The optimization problem can be written as follows: Maximize: f1 = Jsc Maximize: f2 = Voc Maximize: f3 = FF Maximize: f4 = PCE Subject to: 2000 rpm ≤ Spin Speed 1 ≤ 6000 rpm 2000 rpm ≤ Spin Speed 2 ≤ 6000 rpm After we randomly selected the input parameters, we ran the experiment and got the output parameters. The output parameters were the J–V parameters (Jsc, Voc, FF, and PCE). The output parameters found based on the input parameters gathered using the LHS method are shown in Table S5. The relationship of output parameter when fabricated the PVSCs by using LHS condition are shown in Supporting Information Figure S3, Figure S4. We observed a large difference in the efficiency when we changed the dense and mesoporous TiO2 layer; thus, these input parameters are effective for determining the optimum parameters for the experiment. The relationship between output parameters (a) short circuit current, (b) open circuit voltage, (c) fill factor and (d) PCE and input condition random by LHS method. Applying the output parameters in the kriging model yielded the computational optimization results shown in Figure 7.23,24 This figure shows the simulation of the output surface distribution for the Jsc, Voc, FF, and PCE. From these surface plots, we can approximate a reasonable region for the input parameters to achieve a high output value.


Page 14 of 23

Page 15 of 23

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


Figure 7 Surface plot of the relationship between the experimental parameters (spin speed for the dense and mesoporous TiO2 and the outputs for a) Jsc, b) Voc, c) FF, and d) PCE. Figure 7 (a) and (b) show the effect of the input parameters on Jsc and Voc. We can observe that Spin Speed 1 (dense-TiO2) and Spin Speed 2 (mesoporous-TiO2) have a larger effect on Jsc than Voc. Figure 7(c) shows the effect of the input parameters on FF. With regards to choosing the range of parameters for the experiment (Supporting information Table S3) and also the data set from LHS experiment (Table S5), the input parameter FF is related only to Spin Speed 1 (dense-TiO2) with a nearly linear relationship, which is given in Figure 8. By combine 2 data sets, we can confirm the best condition to fabricate dense TiO2 at spin speed around 5000-6000 rpm. Based on the data in Figure 7(d) we can approximate the best conditions for fabricating the TiO2 layer; the best conditions for fabricating dense TiO2 are a spin speed of around 5600–5800 rpm and for mesoporous TiO2 a spin speed of around 2400– 2600 rpm is ideal. Therefore, we chose 5700 rpm for the dense TiO2 layer and 2500 rpm for the mesoporous TiO2 layer. By using this technique, we can observe the thickness of denseTiO2 has strongly effect to change the FF value.



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

Figure 8 The relationship between FF and spin speed for fabricated dense TiO2 layer in data set from estimate input parameter experiment and LHS After we got the optimized value, we made 6 small-area devices(half cell) under these conditions and evaluated the device performances. The average efficiency was 5.81% with a standard deviation of 0.16 and an average Jsc of 12.11 mA cm-2, Voc of 0.81 V, and FF of 0.591. The detail of Current density-voltage parameters of perovskite solar cell(half cell) when fabricated with the optimal condition is shown in Supporting information(Table S6). Table 3 The J–V parameters of the perovskite solar cells when fabricated with small and large device areas of 0.09 and 1 cm2, respectively Condition

Jsc

Voc

FF

PCE

(mA/cm2)

(V)

Best cell(0.09cm2)

21.51

0.99

0.646

13.75

Average(0.09 cm2)

20.13

0.93

0.597

11.53

Best cell(1 cm2)

19.83

0.99

0.599

11.78

Average(1 cm2)

18.78

0.91

0.563

9.54

(%)


Page 16 of 23

Page 17 of 23

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


Figure 9 The J–V characteristics of the highest performance of (a) small-area devices and (b) large-area devices when fabricated under the optimized conditions (the surface area of the small and large devices was 0.09 and 1 cm2, respectively). Moreover, we fabricated the PVSCs by using optimized condition with 18 cells for small-area devices and 10 cells for 1 cm2 area devices. The maximum efficiency in this experiment under optimized conditions for fabricating TiO2 with a full cell is given in Table 3 along with the results of the best cells and the average values. The best value for a small-area device was 13.75% with the Jsc, Voc, and FF values were 21.51 mA cm-2, 0.99 V, and 0.64, respectively. The best performance for the large-area devices was 11.78%, and the Jsc, Voc, and FF values were 19.53 mA cm-2, 0.99 V, and 0.59, respectively. Figure 9 shows the J–V characteristics of the highest performance cell of a) the small-area devices and b) the large-area devices when fabricated under optimized conditions. Figure 10 shows the cell performance distribution of these devices for small-area devices and 1 cm2 area devices. The result is small-area devices have a lower error range of PCE value than 1 cm2 area devices.



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

Figure 10 cell performance distribution of small-area devices and 1 cm2 area devices.

Conclusion In conclusion, we fabricated PVSCs in ambient air with a ITO/dense-TiO2/mesoporousTiO2/perovskite/Spiro-OMeTAD/Au structure with a large surface area by using a modified dynamic spin-washing method. We found that a high spin-washing speed had a significant effect on increasing the PCE by preventing vapor reactions. Vapor reactions strongly effect re-conversion of perovskites into PbI2 and cause the surfaces of perovskite films to be rough. From this experiment, we conclude that not only is the growth of the crystalline phase of the perovskite film important, but the washing step of the perovskite film is also important when it comes to scaling up the area of PVSCs. Moreover, by using Kriging model for predict the maximum efficiency by changing the TiO2 thickness. We can observe the changing of parameters FF is related only the thickness of the dense TiO2 layer. Lastly, we succeeded in achieving a PCE of 13.75% for a device area of 0.09 cm2 and 11.78% for a device area of 1 cm2 by using the optimized process. This modified two-step method will allow researchers to develop PVSCs for practical devices or commercial products in the near future.

ASSOCIATED CONTENT *S Supporting Information The Supporting Information is available free of charge on the ACS Publications website at DOI:xxx-xxxxxxxxxxxxxx This supporting information includes:


Page 18 of 23

Page 19 of 23

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


The value of conversion ratio of PbI2 into MAPbI3. XRD data of wet area and dry area. Detail and result of estimated the range of input parameters. Detail and result of optimization data.

ACKNOWLEDGMENTS The authors thank to recived the financial support from Marubun research promotion foundation 2017. We are deeply grateful to Takeo Moriya for taking the AFM result.

Reference (1) Chen, W.; Wu, Y; Yue, Y.; Liu, J.; Zhang, W.; Yang, X.; Chen, H.; Bi, E. & Han, L., “Efficient and stable large-area perovskite solar cells with inorganic charge extraction layers”, Science. 2015, 350, 944. (2) You, J.; Hong, Z.; Yang, Y.M.; Chen, Q.; Cai, M.; Song, T.B.; Chen, C.C.; Lu, S.; Liu, T.; Zhou, H. & Yang, Y., “Low-Temperature Solution-Processed Perovskite Solar Cells with High Efficiency and Flexibility”, ACS Nano. 2014, 8, 1674. (3) Nie, W.; Tsai, H.; Asadpour, R.; Blancon, J.C.; Neukirch, A.J.; Gupta, G.; Crochet, J.J.; Chhowalla, M.; Tretiak, S.; Alam, M.A.; Wang, H.L. & Mohite, A.D.,” High-efficiency solution-processed perovskite solar cells with millimeter-scale grains”, Science. 2015, 347, 522. (4) Kim, H.A.; Lee, J.W.; Yantara, N.; Boix, P.P.; Kulkarni, S.A.; Mhaisalkar, S.; Grätzel, M. & Park, N.G., “High Efficiency Solid-State Sensitized Solar Cell-Based on Submicrometer Rutile TiO2 Nanorod and CH3NH3PbI3 Perovskite Sensitizer”, Nano letters. 2013, 13, 2412. (5) Ponseca Jr, C.S.; Savenije, T.J.; Abdellah, M.; Zheng, K.; Yartsev, A.; Pascher, T.; Harlang, T.; Chabera, P.; Pullerits, T.; Stepanov, A., Wolf, J. P. & Sundström, V., “ Organometal Halide Perovskite Solar Cell Materials Rationalized: Ultrafast Charge Generation, High and Microsecond-Long Balanced Mobilities, and Slow Recombination”, J. Am. Chem. 2014, 136, 5189. (6) Saliba, M.; Matsui, T.; Seo, J.Y.; Domanski, K.; Correa-Baena, J.P.; Nazeeruddin, M.K.; Zakeeruddin, S.M.; Tress, W.; Abate, A.; Hagfeldt, A. & Grätzel, M., ”Cesium-containing triple cation perovskite solar cells: improved stability, reproducibility and high efficiency”, Energy Environ Sci. 2016, 9, 1989.



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

(7) Sudchanham, J.; Batmunkh, M.; Reutrakul, V.; Shapter, J.G.; Raston, C.L. & Pakawatpanurut, P., “Vortex Fluidics Improved Morphology of CH3NH3PbI3-xClx Films for Perovskite Solar Cells”, ChemistrySelect. 2017, 2, 369. (8) Eperon, G.E.; Burlakov, V.M.; Docampo, P.; Goriely, A. & Snaith, H.J., ” Morphological Control for High Performance, Solution-Processed Planar Heterojunction Perovskite Solar Cells”, Adv Funct Mater. 2014, 24, 151. (9) Quarti, C.; Mosconi, E.; Ball, J.M.; D'Innocenzo, V.; Tao, C.; Pathak, S.; Snaith, H.J.; Petrozz, A. & De Angelis, F.,” Structural and optical properties of methylammonium lead iodide across the tetragonal to cubic phase transition: implications for perovskite solar cells”, Energy Environ Sci. 2016, 9, 155. (10) Kutes, Y.; Ye, L.; Zhou, Y.; Pang, S.; Huey, B.D. & Padture, N.P., “Direct Observation of Ferroelectric Domains in Solution-Processed CH3NH3PbI3 Perovskite Thin Films”, J. Phys. Chem. Lett. 2014, 5, 3335. (11) Razza, S.; Giacomo, F.D.; Matteocci, F.; Cina, L.; Palma, A.L.; Casaluci, S.; Cameron, P.; D'Epifanio, A.; Licoccia, S. & Brown, T.M., “Perovskite solar cells and large area modules (100 cm2) based on an air flow-assisted PbI2 blade coating deposition process”, J. Power Sources. 2015, 277, 286. (12) Wu, C. G.; Chiang, C. H.; Tseng, Z. L.; Nazeeruddin, M. K.; Hagfeldt, A. & Grätzel, M., “High efficiency stable inverted perovskite solar cells without current hysteresis”, Energy Environ Sci. 2015, 8, 2725. (13) Huang, H.; Shi, J.; Zhu, L.; Li, D.; Luo, Y. & Meng, Q., “Two-step ultrasonic spray deposition of CH 3NH3PbI3 for efficient and large-area perovskite solar cell.”, Nano Energy 2016, 27, 352. (14) Li, X.; Bi, D.; Yi, C.; Décoppet, J.D.; Luo, J.; Zakeeruddin, S.M.; Hagfeldt, A.& Grätzel, M.,” A vacuum flash–assisted solution process for high-efficiency large-area perovskite solar cells”, Science 2016, 353, 58. (15) Gouda, L.; Gottesman, R.; Tirosh, S.; Haltzi, E.; Hu, J.; Ginsburg, A.; Keller, D. A.; Bouhadana, Y. & Zaban, A, ” Vapor and healing treatment for CH3NH3PbI3-xClx films toward large-area perovskite solar cells”, Nanoscale. 2016, 8, 6386. (16) Holliman, P. J. ; Jones, E. W.; Connell, A.; Ghosh, S.; Furnell, L. & Hobbs, R.J., “Solvent issues during processing and device lifetime for perovskite solar cells”, Mater. Res. Innovations 2015, 19, 508. (17) Zhao, Y., & Zhu, K., “Three-step sequential solution deposition of PbI2 -free CH3NH3PbI3 perovskite.”, J. Mater. Chem. A 2015, 3, 9086. (18) Shen, H.; Wu, Y.; Peng, J.; Duong, T.; Fu, X.; Barugkin, C.; White, T.P.; Weber, K. & Catchpole, K. R. “Improved Reproducibility for Perovskite Solar Cells with 1 cm2 Active Area by a Modified Two-Step Process.” ACS Appl. Mater. Interfaces 2017, 9, 5974.


Page 20 of 23

Page 21 of 23

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


(19) Carlo, A.D.; Matteocci, F.; Razza, S.; Mincuzzi, M.; Giacomo, F.D.; Casaluci, S.; Gentilini, D.; Brown, T.M.; Reale, A.; Brunetti, F.; Licoccia, S. & D'Epifanio, A., “Mesoscopic perovskite solar cells and modules”. In Title of Published Proceedings: Nanotechnology (IEEE-NANO), 2014 IEEE 14th International Conference, Toronto, ON, Canada, 18-21 Aug. 2014, 70-74 (20) Tai, Q.; You, P.; Sang, H.; Liu, Z.; Hu, C.; Chan, H.L. & Yan, F, ”Efficient and stable perovskite solar cells prepared in ambient air irrespective of the humidity”, Nat. Commun. 2016, 7, 11105. (21) Sveinbjörnsson, K.; Aitola, K.; Zhang, J.; Johansson, M.B.; Zhang, X.; Correa-Baena, J.P.; Hagfeldt, A.; Boschlo, G. & Johansson, E.M., ” Ambient air-processed mixed-ion perovskites for high-efficiency solar cells”, J. Mater. Chem. A. 2016, 4, 16536. (22) Ahn, N.; Son, D.Y.; Jang, I.H.; Kang, S.M.; Choi, M. & Park, N.G., “Highly Reproducible Perovskite Solar Cells with Average Efficiency of 18.3% and Best Efficiency of 19.7% Fabricated via Lewis Base Adduct of Lead(II) Iodide”, J. Am. Chem. 2015, 137, 8696. (23) Gao, Y. & Wang, X., ”An effective warpage optimization method in injection molding based on the Kriging model”, Int J Adv Manuf Technol. 2008, 37, 953. (24) Lee, K. H., & Park, G. J., ”A global robust optimization using Kriging based approximation model.”, JSME Int J., Ser. C. 2006, 49, 779. (25) Giordano, F.; Abate, A.; Baena, J.P.C., Saliba, M.; Matsui, T.; Im, S.H.; Zakeeruddin, S.M.; Nazeeruddin, M.K.; Hagfeldt, A. & Graetzel, M.,” Enhanced electronic properties in mesoporous TiO2 via lithium doping for high-efficiency perovskite solar cells”, Nat. Commun. 2016, 7, 10379. (26) Choi, J.; Song, S.; Hörantner, M.T.; Snaith, H.J. & Park, T., ” Well-Defined Nanostructured, Single Crystalline TiO2 Electron Transport Layer for Efficient Planar Perovskite Solar Cells”, ACS nano. 2016, 10, 6029. (27) Gao, L. L.; Li, C.X.; Li, C.J. & Yang, G.J.,“Large-area high-efficiency perovskite solar cells based on perovskite films dried by the multi-flow air knife method in air”, J. Mater. Chem. A. 2017, 5, 1548. (28) Cheng, Y.; Xu, X.; Xie, Y.; Li, H. W.; Qing, J.; Ma, C.; Lee, C.S.; So, F. & Tsang, S. W. “18% High ‐ Efficiency Air ‐ Processed Perovskite Solar Cells Made in a Humid Atmosphere of 70% RH.” Solar RRL. Early View, DOI: 10.1002/solr.201700097 (29) Hsieh, T.Y.; Wei, T.C.; Wu, K.L.; Ikegami, M. & Miyasaka, T.,“Efficient perovskite solar cells fabricated using an aqueous lead nitrate precursor”, Chem. Commun. 2015, 51, 13294.



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

(30) Koo, J.; Kim, T.; Jung, C. & Lee, J., “Effects of water and iso-propyl alcohol relative humidities on single wafer cleaning system performance”, Int. J .Heat Mass Transf. 2007, 50, 4275. (31) Chen, Q.; Zhou, H.; Hong, Z.; Luo, S.; Duan, H.S.; Wang, H.H.; Liu, Y.; Li, G. & Yang, Y., ”Planar heterojunction perovskite solar cells via vapor-assisted solution process”, J. Am. Chem. 2013, 136, 622. (32) Schlipf, P. J.; Docampo, P.; Schaffer, C.J.; Körstgens, V.; Bießmann, L.; Hanusch, F.; Giesbrecht, N.; Bernstorff, S.; Bein, T. and Müller-Buschbaum, P., ”A Closer Look into Two-Step Perovskite Conversion with X-ray Scattering”, Phys. Chem. Lett. 2015, 6, 1265. (33) Chang, C.Y.; Huang, W.K.; Chang, Y. C.; Lee, K. T. & Chen, C.T., ”A solutionprocessed n-doped fullerene cathode interfacial layer for efficient and stable large-area perovskite solar cells”, J. Mater. Chem. A. 2016, 4, 640. (34) Lang, F.; Gluba, M.A.; Albrecht, S.; Rappich, J.; Korte, L.; Rech, B. & Nickel, N.H.,“Perovskite solar cells with large-area CVD-graphene for tandem solar cells”, Phys. Chem. Lett. 2015, 6, 2745. (35) Ariyarit, A.; Takenaka, I.; Yoshikawa, R.; Gillot, F. & Shiratori, S.,” Boosting the optimization process of perovskite solar cells by partial sampling and kriging method”, RSC Adv. 2016, 6, 98052. (36) Kim, H. S., & Park, N. G., ”Parameters affecting I–V hysteresis of CH3NH3PbI3 perovskite solar cells: effects of perovskite crystal size and mesoporous TiO2 layer”, Phys. Chem. Lett. 2014, 5, 2927. (37) Stein, M., “Large Sample Properties of Simulations Using Latin Hypercube Sampling”, Technometrics 1887, 29, 143.


Page 22 of 23

Page 23 of 23

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


Schematic of explanation of the origin of the different surface structures when washing the surface of the perovskite films with IPA at different spin speeds for cm2 perovskite solar cell. 338x190mm (96 x 96 DPI)


Improvement of the Dynamic Spin-Washing Effect with an Optimized

Recommend Documents