Improving the performance of perovskite solar cells by adding 1,8-diiodooctane in the CH3NH3PbI3 perovskite layer

Improving the performance of perovskite solar cells by adding 1,8-diiodooctane in the CH3NH3PbI3 perovskite layer

Solar Energy 176 (2018) 178–185 Contents lists available at ScienceDirect Solar Energy journal homepage: www.elsevier.com/locate/solener Improving ...

6MB Sizes 0 Downloads 138 Views

Solar Energy 176 (2018) 178–185

Contents lists available at ScienceDirect

Solar Energy journal homepage: www.elsevier.com/locate/solener

Improving the performance of perovskite solar cells by adding 1,8diiodooctane in the CH3NH3PbI3 perovskite layer

T

Chih-Hung Tsai , Chia-Ming Lin, Cheng-Hao Kuei ⁎

Department of Opto-Electronic Engineering, National Dong Hwa University, Hualien 97401, Taiwan

ARTICLE INFO

ABSTRACT

Keywords: Perovskite solar cells 1,8-diiodooctane Coordination Nucleation rate

In this study, different ratios of 1,8-diiodooctane (DIO) were added in the CH3NH3PbI3 perovskite layer of perovskite solar cells (PSCs). The effects of different ratios of DIO on the surface morphology of the perovskite layer were observed using an optical microscope (OM) and scanning electron microscope (SEM), whereas the surface roughness, crystal structure, and surface element bonding of the perovskite layer were observed using an atomic force microscope (AFM), X-ray diffractometer (XRD), and X-ray photoelectron spectroscope (XPS), respectively. The absorption spectrum of the perovskite layer was investigated using an ultraviolet–visible (UV/ VIS) spectrometer, and the band gap of the perovskite layer was calculated. Furthermore, the PSCs with various ratios of DIO in the perovskite layer were analyzed using current density–voltage (J–V), external quantum efficiency (EQE), and electrochemical impedance spectrum (EIS) measurements. The results showed that adding DIO to the perovskite layer can reduce the nucleation rate of the perovskite through coordination, thereby enhancing the crystallinity, uniformity, and coverage of the perovskite layer and reducing its surface roughness to yield PSC device performance that is superior to a device without DIO. The resulting power conversion efficiency (PCE) increased from 10.77% without DIO to 12.86% with 1% DIO, which was a 19.4% increase, thus verifying that adding 1% DIO to the perovskite layer facilitated increasing the device performance of the PSC.

1. Introduction

of eight calcium atoms that form a hexahedron, with the center of each side containing an oxygen atom, and a titanium atom is situated at the center of the hexahedron. The term “perovskite” is also a general term referring to materials with this crystal structure, and their general formula is ABX3, where A and B are two distinct cations and X is an anion. Along with dye-sensitized solar cells and quantum-dot-sensitized solar cells, thin-film solar cells using perovskite-type organic lead–iodide compounds as light-absorbing materials are another type of solar cell employing novel materials. Perovskite-type organic lead–iodide compounds (e.g., CH3NH3PbI3) possess unique photoelectric properties and exhibit excellent power conversion efficiency (PCE) in liquid-state sensitized cells and solid-state thin-film cells based on hole and electron transport materials. Perovskite crystal simultaneously possesses satisfactory electron and hole transport properties, which are the basis for successful preparation of perovskite planar heterostructure solar cells. In 2009, Miyasak et al. first applied the CH3NH3PbI3 perovskite material to a device structure of dye-sensitized solar cells, wherein the dye was replaced with perovskite, and together with the use of titanium dioxide and liquid electrolytes, an energy conversion efficiency of 3.8% was achieved (Kojima et al., 2009). This initiated research on PSCs. In 2011, Park et al. enhanced the efficiency of CH3NH3PbI3 PSCs to 6.54%

Because of recent energy shortages and environmental problems derived from global warming, countries worldwide have begun to attach great value to the development and use of renewable energy (Ott, 1995). Among the numerous types of renewable energy, solar energy is the pollution-free energy that is subject to the fewest environmental and geographical constraints. It does not require transportation and has abundant production resources, prompting governments in many countries to encourage the development of the solar energy industry and promote relevant regulations and commercial incentives. Scientists are also actively involved in relevant studies of solar energy (Kazmerski, 1997). Based on the materials used, solar cells can be divided into several types: silicon based (Goetzberger et al., 2003; Rovira et al., 2000), compound semiconductor (Hubbard et al., 2008; Kapur, 2003; Peharz et al., 2009), oxide materials (Zang, 2018); and organic semiconductor (Sariciftci et al., 1992; Tang, 1986; Nazeeruddin et al., 2001). The use of perovskite material structures to prepare perovskite solar cells (PSCs) has been the most prominent method in recent years. The term perovskite originated from an ore structure discovered in the 19th century. Its chemical formula is CaTiO3, and its structure consists ⁎

Corresponding author. E-mail address: [email protected] (C.-H. Tsai).

https://doi.org/10.1016/j.solener.2018.10.037 Received 27 July 2018; Received in revised form 3 October 2018; Accepted 10 October 2018 0038-092X/ © 2018 Elsevier Ltd. All rights reserved.

Solar Energy 176 (2018) 178–185

C.-H. Tsai et al.

(Im et al., 2011). During their collaboration with Grätzel in 2012, they replaced the previous liquid electrolyte with solid-state materials and achieved a device efficiency of up to 9.7% (Kim et al., 2012). In 2012; Snaith et al. published a study wherein an insulating material—aluminum oxide—was used as the electrode, and together with the perovskite material of CH3NH3PbI2Cl, the PCE was increased to 10.9%. The charge conduction capacity of the perovskite material was also discovered (Lee et al., 2012). In 2013; Grätzel et al. applied a two-step process for preparing all solid-state PSCs with a maximum efficiency of 15% (Burschka et al., 2013). After a few years of research and development, the current certified maximum PCE of PSCs has reached 22.1%. PSCs possess advantages of a high absorption coefficient, low nonradiative carrier recombination rate, high carrier mobility, and an adjustable band gap. The typical cell structure is composed of a conductive substrate, an electron transport layer (ETL), a perovskite layer, a hole transport layer (HTL), and metal electrodes. PSCs have a simple fabrication process, and the devices can be prepared in a wet process at a low temperature. Thus, they are advantageous for being applicable to flexible devices. Recently, various methods have been used to increase the efficiency of PSCs. For example, the performance of PSCs has been improved by employing a CdSe quantum dot/PCBM composite as a ETL (Zeng et al., 2017) and through enhancing the hole extraction by designing iodide concentration gradient in the perovskite layer (Wang et al., 2018). The perovskite layer is the core of a PSC, possessing the function of absorbing sunlight to generate free charge carriers, and its quality affects the performance of a PSC device. Studies on the optimization of the perovskite layer have been conducted in recent years. These included the modulation of the organic and halogen compositions in perovskite materials or of the perovskite layer thickness (Xi et al., 2015). In particular, the crystallinity, uniformity, and coverage of perovskite are the primary influences on PCE (Stranks et al., 2015); thus, optimizing these three features is the most effective approach for improving the PCE of PSCs. In previous literature it was shown that adjusting the concentration of the perovskite precursor solution and the spin-coating speed of the perovskite layer can improve the uniformity and coverage of the perovskite layer (Conings et al., 2014; Hu et al., 2015). In addition, preparing the perovskite layer through fast deposition-crystallization (FDC) and solvent–solvent extraction (SSE) can acquire a more satisfactory crystalline structure (Xiao et al., 2014; Zhou et al., 2015). Liang et al. noted that adding 1,8-diiodooctane (DIO) in perovskite precursor solutions can effectively enhance the crystallinity, uniformity, and coverage of perovskite, the efficiency of the PSCs with DIO additives reached 11.8% (Liang et al., 2014). In addition to its application in the perovskite layer, DIO has also been added to the ETL of PSCs to improve surface morphology. The device efficiency of PSCs was improved to 12.73% by using DIO as a solvent additive during the deposition of PCBM layers (Liu and Lee, 2015). These increases could be attributed to the improved morphological characteristics of the perovskite and PCBM layers prepared using DIO. DIO is even applied to organic and polymer solar cells, making it the most discussed supplementary material among PSCs, organic solar cells, and polymer solar cells in recent years (Zusan et al., 2015; Wang et al., 2014). In this study, a PSC was prepared through low-temperature spin coating with a structure of Glass/fluorine-doped tin oxide (FTO)/poly (3,4-ethylene dioxythiophene):poly(styrene sulfonate) (PEDOT:PSS)/ Perovskite active layer/[6,6]-phenyl-C61-butyric acid methyl ester (PCBM)/Ag. A complete analysis of the material properties and device properties was employed to investigate the effects of adding different proportions of DIO to the perovskite layer in the PSC.

Fig. 1. The perovskite precursor solutions based on various ratios of DIO.

of PbI2 powder and 0.1187 g of CH3NH3I (MAI) powder into a mixed solvent consisting of 600 μL of dimethyl sulfoxide (DMSO) and 400 μL of γ-butyrolactone (GBL). Subsequently, DIO was added to achieve different weight ratios: without DIO, with 1% DIO, and with 2% DIO. The solutions were then oscillated in an ultrasonic oscillator for 20 min and placed in a nitrogen atmosphere glove box and heated at 60 °C for 1 h to obtain the perovskite precursor solution (Fig. 1). The ETL solution was prepared by adding 0.002 g of PCBM powder to 90 μL of 1,2Dichlorobenzene (DCB) solvent. Subsequently, the ETL solution was oscillated in an ultrasonic oscillator for 5 min, thereby completing the preparation of the ETL solution. 2.2. Fabrication of PSCs Fig. 2 shows the process for producing the PSC. First, the substrate was etched and cleaned. A 0.6-cm-wide strip of Teflon tape was applied to one side of a 2 × 2 cm fluorine-doped tin oxide (FTO) glass substrate. Subsequently, 3 M tape was applied to the area of the substrate not covered by the Teflon tape. The Teflon tape was then removed, and zinc powder was applied to the surface of the substrate that was not covered by the 3 M tape. Hydrochloric acid was then dripped onto the zinc powder to etch away the FTO. Next, the substrate was rinsed with deionized water to remove the excessive zinc powder and hydrochloric acid. It was then placed in isopropyl alcohol and subjected to 5 min of ultrasonic oscillation. After oscillation, the substrate was dried with nitrogen gas. Teflon tape was then applied to the nonetched area on the opposite side of the substrate (0.6 × 2 cm) to serve as the measuring cathode, and the substrate was irradiated using a UV–Ozone cleaner for 15 min to remove organic substances and increase surface affinity. Next, the HTL was prepared by placing the cleaned substrate in a spin coater, and 50 μL of PEDOT:PSS was evenly applied to the substrate through 30-s spin coating at 4000 rpm, after which it was heated at 150 °C for 10 min. The perovskite layer was prepared by cooling the substrate to room temperature and placing it in a nitrogen atmosphere glove box (99.995%). The substrate coated with PEDOT:PSS was then placed in a spin coater, where 40 μL of perovskite precursor solution was evenly applied to its surface through two stages of spin coating: 10-s rotation at 1000 rpm during the first stage, and 22-s rotation at 7500 rpm during the second stage. Toluene was injected at 18 s during the second stage, and the substrate was subsequently moved onto a heating plate and heated at 100 °C for 5 min. After the substrate cooled down, the previously described spin coating of the perovskite layer was repeated again. The PCBM ETL was prepared by placing the substrate coated with the perovskite layer in a spin coater, and 20 μL of the ETL solution

2. Experiments 2.1. Preparation of perovskite precursor solution and ETL solution The perovskite precursor solution was prepared by adding 0.5762 g 179

Solar Energy 176 (2018) 178–185

C.-H. Tsai et al.

Fig. 2. The fabrication process of perovskite solar cell.

Fig. 3. SEM images at 10 K magnification for the perovskite layer with (a) 0%, (b) 1%, and (c) 2% DIO; and SEM images at 30 K magnification for the perovskite layer with (d) 0%, (e) 1%, and (f) 2% DIO.

was evenly applied onto its surface through a 30-s spin coating at 1750 rpm. After spin coating, it was removed from the glove box. Finally, the evaporation deposition of the Ag electrode was performed by adding a metal mask to the ETL-coated substrate, which was then placed in the vacuum chamber of a vapor deposition machine. A mechanical rough pump was used to pump the vacuum value to 2 × 10−2 Torr, and the turbine pump was then turned on and warmed up for 2 min. The fine pump valve was opened to pump the background vacuum value to approximately 2.2 × 10−5 Torr. The current supply was turned on to gradually increase the current value to 38 A, and the baffle was opened to heat and vaporize the Ag on the tungsten boat, after which it was dispersed onto the substrate. The vapor deposition lasted for 4 min, after which the baffle and current source for the evaporation deposition were turned off, and the vacuum of the chamber was broken to complete the PSC device. The active area of the PSC device in this work was 0.1 cm2.

2.3. Characterization of the perovskite layers and PSCs This study conducted a complete analysis of the perovskite layers and PSCs. The effects of various solvents on the surface morphology of the perovskite layer were observed using an optical microscope and scanning electron microscope (SEM), whereas the surface roughness, crystal structure, and surface element bonding of the perovskite layer were observed using an atomic force microscope (AFM), X-ray diffractometer (XRD), and X-ray photoelectron spectroscope (XPS), respectively. The absorption spectrum of the perovskite layer was investigated using an ultraviolet–visible (UV/VIS) spectrometer, and the band gap of the perovskite layer was calculated based on the measured absorption spectrum. This study also conducted a complete analysis of the PSCs, and the analysis methods included current density–voltage (J–V) curve, external quantum efficiency (EQE), and electrochemical impedance spectrum (EIS).

180

Solar Energy 176 (2018) 178–185

C.-H. Tsai et al.

3. Results and discussion

form a temporary metal complex and occupied two positions of I within the perovskite unit lattice. DIO was subsequently removed through heating, and the positions occupied by the original coordination were filled by the I atoms in the MAI. The bonding at the surrounding six vacancies indicated that the perovskite crystal was completed. Therefore, the brief coordination after the addition of DIO resulted in the perovskite being unable to complete crystallization within a short time. A heating procedure was required to remove the DIO and allow the MAI to completely react with the PbI2, because the purpose of adding DIO was to reduce the nucleation rate of the perovskite crystal. According to basic crystal growth theory, a slow nucleation rate forms more orderly and dense crystals; therefore, the slow crystallization of perovskite can improve the uniformity of the perovskite crystal, increase the charge transfer efficiency between the perovskite layer and charge transport layers, and enhance the efficiency of the PSC (Tsai et al., 2017). The SEM images showed that adding 1% DIO moderately reduced the nucleation rate of the perovskite crystallization and formed a uniform perovskite crystal. Because the evaporation of DIO was incomplete after adding 2% DIO, excessive DIO occupied too many positions in the perovskite unit lattice and resulted in the MAI being unable to bond effectively with PbI2. Therefore, the excessive amount of precipitated crystals observed in the SEM images is presumed to be PbI2 crystals. The perovskite layer without DIO exhibited poor crystallinity and coverage, and excessive pinholes resulted in current short-circuit, which is usually regarded as a defect of the device that reduces its efficiency. This study next analyzed the surface roughness of the perovskite layer using an AFM, and the results are shown in Fig. 5. The surface roughness was 19.0 nm when no DIO was added to the perovskite layer. Adding 1% DIO to the perovskite layer yielded the lowest surface roughness (14.9 nm), whereas adding 2% DIO produced a surface roughness of 19.6 nm, which was slightly higher than that without DIO. The AFM images showed that the perovskite layer without DIO had a poor crystallinity caused by disorderly perovskite nucleation, which yielded greater surface roughness. Adding 1% DIO to the perovskite layer reduced the nucleation rate of the perovskite, enabling it to undergo a homogenous and uniform crystallization and form a smoother perovskite layer. When 2% DIO was added, the excessive DIO resulted in its incomplete evaporation, which inhibited the combination of PbI2 and MAI, resulting in the precipitation of PbI2 crystals that increased the surface roughness. Adding 1% DIO to the perovskite layer yielded a lower surface roughness, which is advantageous to the performance of the planar PSC device. The crystal structure of the perovskite layer was analyzed using an XRD, and the analysis results are shown in Fig. 6. Fig. 6(a)–(c) show the results for the crystal structures with 0%, 1%, and 2% DIO, respectively. The experimental results indicated that the signals at 26.5°, 33.7°, 37.8°, and 51.6° were the characteristic peaks of FTO. The signals at 14.1°, 19.9°, 23.5°, 24.6°, 28.5°, 31.9°, 40.6°, and 43.2° were the characteristic peaks of the perovskite structure material, CH3NH3PbI3, and their corresponding lattice planes were (1 1 0), (1 1 2), (2 1 1), (2 0 2), (2 2 0), (3 1 0), (2 2 4), and (3 1 4) (Kim et al., 2017). In the crystal structure without DIO in Fig. 6(a), a signal belonging to the characteristic peak of PbI2 with a lattice plane of (0 0 1) was discovered at the angle of 12.7° (Wu et al., 2014). No PbI2 characteristic peak was discovered in the crystal structure with 1% DIO shown in Fig. 6(b), whereas a weak PbI2 characteristic peak was observed in the crystal structure with 2% DIO shown in Fig. 6(c). When the perovskite structure material (CH3NH3PbI3) left the nitrogen atmosphere glove box, it was prone to react with the moisture and oxygen in the atmosphere, causing some of the perovskite to decompose from the defective locations (Zhao and Park, 2015; Yamamoto et al., 2016). After adding 1% DIO to the perovskite layer, the dense and uniform crystallization and complete coverage reduced the decomposition of the perovskite layer by the moisture and oxygen. Because the DIO did not completely evaporate and occupied too many lattice sites when 2% DIO was added, the MAI

3.1. Characterization of the perovskite layers This study used an SEM and an AFM to investigate the effects of adding different ratios of DIO on the surface morphology of the perovskite layer, and the analyzed structure was Glass/FTO/PEDOT:PSS/ Perovskite active layer. In the SEM analysis, the effects on the surface morphology from adding distinct ratios of DIO to the active layer were observed. The results are shown in Fig. 3. Fig. 3(a)–(c) display SEM images at 10 K magnification for 0%, 1%, and 2% DIO, respectively. Fig. 3(d)–(f) display SEM images at 30 K magnification for 0%, 1%, and 2% DIO, respectively. The images at 10 K magnification revealed that the coverage of the perovskite layer with 1% DIO and 2% DIO was significantly greater than without DIO. Further increasing the magnification to 30 K times showed that the perovskite crystal with 1% DIO was extremely clear, with consistent crystal size and few pinholes. When the additive concentration was increased to 2% DIO, the perovskite crystal became indistinct, and precipitation of crystal particles, speculated to be PbI2 crystals, was observed. The reasons for this are discussed in a following section. Finally, the perovskite layer without DIO was nearly crystal-free and exhibited poor coverage; thus, numerous holes were observed through the SEM. DIO is a long-chain molecule with the chemical formula C8H16I2. Fig. 4(a) shows its three-dimensional structure, in which the front and end both connect to an I atom. The reaction principle of DIO added to perovskite is described by the hard and soft acid and base (HSAB) theory, which is widely used to describe organic and inorganic chemical reactions. It defines positive ions as acids and negative ions as alkalis and calculates chemical hardness based on atomic ionization energy and electron affinity: (ionization energy − electron affinity)/2. Acids and alkalis with equivalent chemical hardness are likely to form bonds. According to the HSAB theory, I and Pb can be regarded as alkali and acid, and their chemical hardnesses are 3.69 eV and 3.53 eV, respectively; thus, they are likely to react with each other (Pearson, 1988). When DIO was added to the perovskite, the I atoms connected to the head and tail of DIO were bonded to Pb as described by the HSAB theory. This process is also known as coordination reaction. Because I is considered a ligand, DIO possesses two ligands. The mechanism of the effects of DIO on perovskite is shown in Fig. 4(b). When DIO was added, it underwent coordination with Pb to

Fig. 4. The (a) chemical structure of DIO and (b) the mechanism of the effects of DIO on perovskite. 181

Solar Energy 176 (2018) 178–185

C.-H. Tsai et al.

could not completely react with the PbI2, and a weak PbI2 characteristic peak was observed. Nevertheless, the signal strength of the PbI2 with 2% DIO was lower than that of the perovskite layer without DIO. The strong PbI2 signal of the 0% DIO perovskite layer was caused by defects. The XRD analysis showed that adding 1% DIO to the perovskite layer yielded a denser and more uniform perovskite crystal with complete coverage. The defects caused by nucleation were very small; thus, it achieved satisfactory performance and was expected to exhibit favorable performance in the PSC device. This study then adopted XPS analysis to investigate the surface element bonding of the perovskite layer. The spectra of I3d and Pb4f were analyzed individually in the XPS analysis. Fig. 7(a) shows the analysis results of I3d. The orbital bonds of I3d3/2 and I3d5/2 in I- were measured. The binding energies of the I3d3/2 orbital bond were 631.28 eV (0% DIO), 630.98 eV (1% DIO), and 630.98 eV (2% DIO), whereas the binding energies of the I3d5/2 orbital bond were 619.78 eV (0% DIO), 619.48 eV (1% DIO), and 619.48 eV (2% DIO). The XRD analysis showed that the perovskite layer without DIO was decomposed by moisture and oxygen, thereby leading to increased PbI2 content. The XPS analysis identified an increase in binding energy in the I3d orbital bond of the perovskite layer without DIO. Fig. 7(b) displays the Pb4f analysis results; the orbital bonds of Pb4f5/2 and Pb4f7/2 in Pb2+ were measured. The binding energies of the Pb4f5/2 orbital bond were 143.58 eV (0% DIO), 142.78 eV (1% DIO), and 142.98 eV (2% DIO), whereas the binding energies of the Pb4f7/2 orbital bond were 138.78 eV (0% DIO), 137.93 eV (1% DIO), and 138.03 eV (2% DIO). A distinct increase in the binding energy of the Pb4f orbital bond was observed when no DIO was added. The XPS analysis demonstrated that the presence of PbI2 increased the binding energy of I and Pb, a phenomenon resulting from the stronger bonding strength of PbI2 that was noted in the literature (Li et al., 2015). XPS analysis showed that the perovskite layer without DIO had more residual PbI2 after crystallization and was unfavorable for device performance (Thakur et al., 2016). The effects of adding DIO to the perovskite layer on its optical properties were investigated using a UV/VIS spectrometer. Fig. 8 shows the results for the absorption spectroscopy analysis with a measured wavelength range of 300–800 nm. The results showed that the perovskite layer possessed favorable light absorption properties at 300–550 nm. At 300–400 nm, the perovskite layer with 1% DIO exhibited a higher absorption; at wavelengths of 420–540 nm, the perovskite layer without DIO possessed a higher absorption. This study further analyzed the effects of adding DIO to the perovskite layer on the energy band gap (Eg) of the perovskite. The Eg can be calculated from the known absorption spectrum through the Tauc–Sunds equation (Ogwu et al., 2005). First, the film thickness (t) of the perovskite layer was substituted into the equation I = I0 e−αt to calculate the α value, and the (αhυ)2 value and Eg were used as the ordinate and abscissa for the Tauc plot, as shown in Fig. 9. In the Tauc plot, the Eg value corresponding to the highest slope of the curve was the Eg of the perovskite layer. The experimental results indicated that the Eg values of the perovskite layer with different DIO ratios of 0%, 1%, and 2% were 1.587 eV, 1.589 eV, and 1.594 eV, respectively, with Eg exhibiting a slight increasing trend with the increasing ratio of DIO.

Fig. 5. The AFM images of perovskite layer based on (a) 0%, (b) 1%, and (c) 2% DIO in the perovskite layer.

3.2. Characterization of PSC devices The J–V curve, EQE, and EIS of the device were analyzed to investigate the effects of adding different ratios of DIO to the perovskite layer. The J–V analysis was first performed on the device to investigate the effects of adding distinct DIO ratios to the perovskite layer on the short-circuit density (Jsc), open-circuit voltage (Voc), fill factor (FF), and

Fig. 6. The XRD results of perovskite layer based on various ratios of DIO in the perovskite layer.

182

Solar Energy 176 (2018) 178–185

C.-H. Tsai et al.

Fig. 7. The XPS (a) I3d and (b) Pb4f analysis results of perovskite layer based on various ratios of DIO in the perovskite layer.

Fig. 8. The absorption spectra of the perovskite active layers based on various ratios of DIO in the perovskite layer.

PCE of the PSC device. Fig. 10 shows the J–V curves of the PSCs based on various ratios of DIO in the perovskite layer measured under forward and reverse scan directions. The J–V results showed that the PSCs exhibited negligible hysteresis. The reproducibility of the device performances based on various ratios of DIO in the perovskite layer was investigated. For each condition, five PSC devices were fabricated and analyzed. The values shown in Table 1 are averaged values and standard deviations of five devices made under identical conditions. The results illustrated that the ratios of DIO that yielded the lowest and highest Jsc were 2% DIO (19.93 mA/cm2) and 1% DIO (21.51 mA/cm2), respectively; the DIO ratios that produced the lowest and highest Voc were 0% DIO (0.88 V) and 2% DIO (0.97 V), respectively. The lowest and highest FF values (0.58 and 0.65) were yielded by 2% DIO and 1% DIO, respectively; the lowest and highest PCE values were yielded by 0% DIO (10.77%) and 1% DIO (12.86%). The FF of the PSC device varied because of the impedance between the interfaces of the device, the series resistance of the device, and the parallel resistance. The J–V analysis indicated that the perovskite layer possessed superior crystallinity, uniformity, coverage, and flatness when 1% DIO was added, and its Jsc increased by 0.76 mA/cm2 compared with the device without DIO. It also led to an increase in Voc by 0.04 V, and the 0.06 increase in FF was the most significant influence. Satisfactory crystallinity, uniformity, coverage, and flatness reduced the impedance between the interfaces and the series resistance of the device. Adding 1% DIO also improved the PCE by 19.4%, verifying that this particular DIO ratio was highly beneficial to device performance. Although adding 2% DIO

Fig. 9. The Tauc plots of the perovskite active layers based on various ratios of DIO in the perovskite layer.

183

Solar Energy 176 (2018) 178–185

C.-H. Tsai et al.

yielded superior device performance compared to the device without DIO, excessive DIO prevented the reaction between MAI and PbI2; thus, 1% DIO remains the optimal ratio. The J–V analysis also showed that the PSCs possessed higher Voc values when DIO was added in the perovskite layer. The increase in Voc indicates that the charge recombination (leakage current) is suppressed by the use of DIO in the perovskite layer (Liu and Lee, 2015). In addition, the experimental results indicated that the band gap of the perovskite layer exhibited a slight increasing trend with the increasing ratio of DIO. The higher Voc of PSCs with 2% DIO is attributed to the higher band gap of the perovskite layer (Yuan et al., 2016). This study then performed EQE analysis with a measurement wavelength range of 300–800 nm, and the analysis results are shown in Fig. 11(a). The PSC device showed favorable EQE at 380–750 nm. The highest peaks of EQE for the different ratios of DIO were 89.2% (1%), 86.2% (0%), and 85.8% (2%). Overall, the EQE values exhibited the same trend as the Jsc values (1% DIO = 21.51 mA/cm2, 0% DIO = 20.75 mA/cm2, and 2% DIO = 19.93 mA/cm2). The effects of adding DIO to the PSC device were verified again in the EQE analysis, with 1% being the optimal ratio for DIO addition and yielding the highest PCE and EQE. This study further performed EIS analysis of the device, and the results are shown in Fig. 11(b). The measured frequency ranged from 20 Hz to 1 MHz, and the AC amplitude was 0.01 V. A semicircle representing the impedance of the carrier that was transported between the ETL/perovskite layer/HTL interfaces was present in the EIS analysis of the PSC device, with a smaller impedance value being more favorable to the carrier transport (Liu et al., 2017; Tsai et al., 2017). The EIS results indicated that the PSC had the smallest interface impedance value when 1% DIO was added to the perovskite layer. Because the perovskite layer had a lower surface roughness and higher crystallinity, uniformity, and coverage when 1% DIO was added, the resulting interface impedance was lowest; conversely, the PSC device without DIO exhibited the highest internal impedance. The EIS analysis verified that a perovskite layer with high crystallinity, few defects, and low surface roughness can effectively reduce interface impedance. This study also compared the stability of the PSC devices with various ratios of DIO in the perovskite layer. The unencapsulated devices were kept in a dry box in the dark and their photovoltaic performance was measured at different time intervals under ambient air at room temperature. Fig. 12 shows that after 300 h, the devices with 1% and 2% DIO in the perovskite layer retained 75–80% of their efficiency

Fig. 10. The J–V curves of the PSCs based on various ratios of DIO in the perovskite layer under forward and reverse scan directions. Table 1 Photovoltaic characteristics of PSCs based on various ratios of DIO in the perovskite layer. Perovskite layer

Jsc (mA/cm2)

Voc (V)

FF

PCE (%)

W/o DIO With 1% DIO With 2% DIO

20.75 ± 0.09 21.51 ± 0.08 19.93 ± 0.08

0.88 ± 0.01 0.92 ± 0.01 0.97 ± 0.01

0.59 0.65 0.58

10.77 ± 0.17 12.86 ± 0.19 11.21 ± 0.16

Fig. 11. The (a) EQE spectra and (b) EIS Nyquist plots of the PSCs based on various ratios of DIO in the perovskite layer.

Fig. 12. The stability of the PSCs based on various ratios of DIO in the perovskite layer.

184

Solar Energy 176 (2018) 178–185

C.-H. Tsai et al.

compared with the original devices. By contrast, the devices without DIO in the perovskite layer merely retained 55% of their original efficiency. The results showed that after adding DIO to the perovskite layer, the dense and uniform crystallization and complete coverage reduced the decomposition of the perovskite layer by the moisture and oxygen, thereby extending the stability of such devices.

Ministry of Science and Technology (MOST) of Taiwan (Grant No. MOST 106-2221-E-259-014). References Ott, K.O., 1995. Prog. Nucl. Energy 29, 81. Kazmerski, L.L., 1997. Renew. Sustain. Energy Rev. 1, 71. Goetzberger, A., Hebling, C., Schock, H.W., 2003. Mater. Sci. Eng. R-Rep. 40, 1. Rovira, P.I., Ferlauto, A.S., Koh, Joohyun, Wronski, C.R., Collins, R.W., 2000. J. NonCryst. Solids. 266, 279. Hubbard, S.M., Cress, C.D., Bailey, C.G., Raffaelle, R.P., Bailey, S.G., Wilt, D.M., 2008. Appl. Phys. Lett. 92, 123512. Kapur, V., 2003. Thin Solid Films 53, 431. Peharz, G., Siefer, G., Bett, A.W., 2009. Sol. Energy 83, 1588. Zang, Z., 2018. Appl. Phys. Lett. 112, 042106. Sariciftci, N.S., Smilowitz, L., Heeger, A.J., Wudl, F., 1992. Science 258, 1474. Tang, C.W., 1986. Appl. Phys. Lett. 48, 183. Nazeeruddin, M., Pechy, P., Renouard, T., Zakeeruddin, S.M., Humphry-Baker, R., Comte, P., Liska, P., Cevey, L., Costa, E., Shklover, V., Spiccia, L., Deacon, G.B., Bignozzi, C.A., Grätzel, M., 2001. J. Am. Chem. Soc. 123, 1613. Kojima, A., Teshima, K., Shirai, Y., Miyasaka, T., 2009. J. Am. Chem. Soc. 131, 6050. Im, J.H., Lee, C.R., Lee, J.W., Park, S.W., Park, N.G., 2011. Nanoscale 3, 4088. Kim, H.S., Lee, C.R., Im, J.H., Lee, K.B., Moehl, T., Marchioro, A., Moon, S.J., HumphryBaker, R., Yum, J.H., Moser, J.E., Grätzel, M., 2012. Sci. Rep. 2, 591. Lee, M.M., Teuscher, J., Miyasaka, T., Murakami, T.N., Snaith, H.J., 2012. Science 338, 643. Burschka, J., Pellet, N., Moon, S.J., Humphry-Baker, R., Gao, P., Nazeeruddin, M.K., Grätzel, M., 2013. Nature 499, 316. Zeng, X., Zhou, T., Leng, C., Zang, Z., Wang, M., Hu, W., Tang, X., Lu, S., Fang, L., Zhou, Miao, 2017. J. Mater. Chem. A 5, 17499. Wang, M., Zang, Z., Yang, B., Hu, X., Sun, K., Sun, L., 2018. Sol. Energy Mater. Sol. Cells 185, 117. Xi, J., Wu, Z., Dong, H., Xia, B., Yuan, F., Jiao, B., Xiao, L., Gong, Q., Hou, X., 2015. Nanoscale 7, 10699. Stranks, S.D., Nayak, P.K., Zhang, W., Stergiopoulos, T., Snaith, H.J., 2015. Angew. Chem. Int. Ed. 54, 3240. Conings, B., Baeten, L., Dobbelaere, C.D., D’Haen, J., Manca, J., Boyen, H.G., 2014. Adv. Mater. 26, 2041. Hu, X., Du, P., Xu, W., Wang, K., Yi, C., Liu, C., Huang, F., Gong, X., Cao, Y., 2015. IEEE J. Photovolt. 5, 1402. Xiao, M., Huang, F., Huang, W., Dkhissi, Y., Zhu, Y., Etheridge, J., Gray-Weale, A., Bach, U., Cheng, Y.B., Spiccia, L., 2014. Angew. Chem. Int. Ed. 126, 10056. Zhou, Y., Yang, M., Wu, W., Vasiliev, A.L., Zhu, K., Padture, N.P., 2015. J. Mater. Chem. A 3, 8178. Liang, P.W., Liao, C.Y., Chueh, C.C., Zuo, F., Williams, S.T., Xin, X.K., Lin, J., Jen, A.K.Y., 2014. Adv. Mater. 26, 3748. Liu, Z., Lee, E.C., 2015. Org. Electron. 24, 101. Zusan, A., Gieseking, B., Zerson, M., Dyakonov, V., Magerle, R., Deibel, C., 2015. Sci. Rep. 5, 8286. Wang, Z., Zhang, F., Li, L., An, Q., Wang, J., Zhang, J., 2014. Appl. Surf. Sci. 305, 221. Pearson, R.G., 1988. Inorg. Chem. 27, 734. Tsai, C.M., Wu, G.W., Narra, S., Chang, H.M., Mohanta, N., Wu, H.P., Wang, C.L., Diau, E.W.G., 2017. J. Mater. Chem. A 5, 739. Kim, S., Chung, T., Bae, S., Lee, S.W., Lee, K.D., Kim, H., Lee, S., Kang, Y., Lee, H.S., Kim, D., 2017. Org. Electron. 41, 266. Wu, Y., Islam, A., Yang, X., Qin, C., Liu, J., Zhang, K., Peng, W., Han, L., 2014. Energy Environ. Sci. 7, 2934. Zhao, X., Park, N.G., 2015. Photonics 2, 1139. Yamamoto, K., Furumoto, Y., Shahiduzzaman, M., Kuwabara, T., Takahashi, K., Taima, T., 2016. Jpn. J. Appl. Phys. 55, 04ES07. Li, Y., Xu, X., Wang, C., Wang, C., Xie, F., Yang, J., Gao, Y., 2015. AIP Adv. 5, 09711. Thakur, U., Kwon, U., Hasan, M.M., Yin, W., Kim, D., Ha, N.Y., Lee, S., Ahn, T.K., Park, H.J., 2016. Sci. Rep. 6, 35994. Ogwu, A.A., Bouquerel, E., Ademosu, O., Moh, S., Crossan, E., Placido, F., 2005. J. Phys. D Appl. Phys. 38, 266. Yuan, M., Zhang, X., Kong, J., Zhou, W., Zhou, Z., Tian, Q., Meng, Y., Wu, S., Kou, D., 2016. Electrochim. Acta 215, 374. Liu, D., Li, Y., Yuan, J., Hong, Q., Shi, G., Yuan, D., Wei, J., Huang, C., Tang, J., Fung, M.K., 2017. J. Mater. Chem. A 5, 5701. Tsai, C.H., Huang, W.C., Wang, W.S., Shih, C.J., Chi, W.F., Hu, Y.C., Yu, Y.H., 2017. J. Colloid Interf. Sci. 495, 111.

4. Conclusions This study investigated the effects of adding different ratios of DIO to the CH3NH3PbI3 perovskite layer on the perovskite layer and PSC device properties through a complete material and device feature analysis. The structure of the PSC was Glass/FTO/PEDOT:PSS/ Perovskite layer/PCBM/Ag. In the surface morphology analysis, an SEM was employed to observe the changes in the surface morphology of the perovskite layer. The SEM analysis showed that the perovskite layer without DIO had poorer crystallinity, uniformity, and coverage, and numerous pinholes were observed. Adding 1% DIO increased the crystallinity and coverage of the perovskite layer and reduced the number of pinholes. When 2% DIO was added, the incomplete evaporation of DIO resulted in excess DIO that prevented the reaction between MAI and PbI2, causing the precipitation of PbI2 crystals on the surface. According to the AFM analysis, adding 1% DIO to the perovskite layer caused the DIO to undergo coordination with PbI2, after which the DIO was removed by heating, thereby reducing the nucleation rate of the perovskite and yielding the lowest surface roughness (14.9 nm). The XRD analysis indicated that when no DIO was added to the perovskite layer, the PbI2 characteristic peak with a lattice plane of (0 0 1) was observed at 12.7°, because perovskite was prone to react with the oxygen and moisture in the atmosphere and began decomposing at the defect locations (pinholes). Adding 1% DIO effectively reduced the number of pinholes; thus, no PbI2 characteristic peak was observed. This verified that the perovskite layer was more stable after the addition of DIO. The XPS analysis illustrated that the orbital bonds of I3d and Pb4f had higher binding energies when no DIO was added. The XPS results verified that perovskite with more defects is more easily decomposed into PbI2 in the atmosphere and is detrimental to device performance. Analysis of the optical properties indicated that adding DIO to the perovskite layer did not yield a substantial change in the absorption spectrum of the perovskite layer, and no significant change was observed in the energy band gap; only a slight upward trend was discovered. The J–V analysis results showed that the when 1% DIO was added to the perovskite layer, the Jsc, FF, and PCE of the PSC were the highest (21.51 mA/cm2, 0.65, and 12.86%, respectively). The EQE analysis results indicated that the PSC possessed the highest EQE peak (89.2%) when 1% DIO was added to the perovskite layer. Moreover, EIS analysis showed that the PSC also had the smallest interface impedance when 1% DIO was added. In summary, adding DIO to the perovskite layer can reduce the nucleation rate of the perovskite through coordination, thereby enhancing the crystallinity, uniformity, and coverage of the perovskite layer and reducing its surface roughness to yield PSC device performance that is superior to a device without DIO. The resulting PCE increased from 10.77% without DIO to 12.86% with 1% DIO, which was a 19.4% increase, thus verifying that adding 1% DIO to the perovskite layer facilitated increasing the device performance of the PSC. Acknowledgments The authors sincerely acknowledge the financial support from

185