Perovskite Photo-Sensors with Solution-Processed TiO₂ under Low Temperature Process and Ultra-Thin Polyethylenimine Ethoxylated as Electron Injection Layer

Abstract

A perovskite photo-sensor is promising for a lightweight, thin, flexible, easy-to-coat fabrication process, and a higher incident photon-to-current conversion efficiency. We have investigated perovskite photo-sensors with a solution-processed compact TiO₂ under a low-temperature process and an ultra-thin polyethylenimine ethoxylated (PEIE) as an electron injection layer. The TiO₂ film is grown from an aqueous solution of titanium tetrachloride (TiCl₄) at 70 °C by a chemical bath deposition method. For an alternative process, the ultra-thin PEIE is spin coated on the TiO₂ film. Then, the perovskite layer is deposited on the substrate by the one- or two-step methods in the glovebox. Next, a hole transport layer of 2,2,7,7-tetrakis(N,N-di-p-methoxyphenylamine)-9, 9-spiro-bifluorene (Spiro-OMeTAD) solution is spin coated. The fabricated device structure is a photodiode structure of FTO/TiO₂/(without or with) PEIE/(one- or two-step) perovskite layer/Spiro-OMeTAD/Au. For the sensing characteristics, a ratio of photo-to-dark-current density was 2.88 × 10⁴ for the device with PEIE layer. In addition, a power-law relationship is discussed.

1. Introduction

Organic electronic devices, such as organic light-emitting diodes (OLEDs) [1,2,3], organic thin-film transistors (OTFTs) [4,5,6,7,8], organic sensing devices [9,10,11] and organic solar cells (OSCs) [12,13,14,15] are widely studied. Especially in OSCs, one of the most interesting topics is an organic–inorganic hybrid perovskite solar cell (PeSC) that has been considered an alternative photovoltaic technology due to its excellent photoelectric conversion efficiency (PCE) along with its low material costs, and its recorded power conversion efficiency of 25.5% [16]. A compact titanium dioxide (TiO₂) layer is the most common electron-transport layer for the PeSC [15]. For achieving the higher efficiency PeSCs with a suitable electronic property of TiO₂, several methods, such as sol−gel, spin coating, slot-die coating, blade coating, electrodeposition, atomic layer deposition, and electron beam (EB) evaporation have been used [17,18,19]. On the other hand, by employing the perovskite materials as an active layer of photo-sensing characteristics, a perovskite photo-sensors (PePS) using a photodiode (PD) structure [20,21,22,23,24], an X-ray detection using the PD structure [25], a photodetector with a photoconductive phenomenon [26], and a printable and flexible photodetector [27,28] were reported. This series of perovskite devices are effective for their higher carrier generation of electron and hole, and their higher carrier mobility during light irradiation. During this time, we studied the PePS with a solution-processed TiO₂ at a low temperature of 70 °C by the chemical bath deposition (CBD) method [29]. For improving device performance, an ultra-thin polyethylenimine ethoxylated (PEIE) [30,31,32,33] as an electron injection layer is fabricated and device characteristics are evaluated. One of the preliminary experimental results is presented at the international conference on the 26th and 27th International Workshop on Active-Matrix Flatpanel Displays and Devices [34,35].

2. Experimental Section

The typical device structure of the PePS under study is shown in Figure 1. Except for the solution-processed TiO₂ layer, the fabrication of the device structure and evaluation of the device characteristics are identical to that of our previous PeSC experiment as follows [29,34]. First, an FTO glass substrate was cleaned using the process of ultra-sonication dipped into organic-alkali and UV-irradiation in a UV-ozone chamber. Then, 20 nm thickness of the TiO₂ layer was grown onto the FTO substrates using a mixture of solution TiCl₄ and pure water (2000 μL/100 mL) at 70 °C for 50 min by the CBD method. Then, the substrate was rinsed into pure water. For the case of device fabrication with a very thin PEIE layer, 4 wt% solution of PEIE dissolved with ethanol was spin coated. Thereafter, spin-rinse using the ethanol solution was performed. After the UV-irradiation, the perovskite layer was deposited on theTiO₂ films by the one- or two-step method in the glovebox under a nitrogen environment. For fabricating the perovskite layer, two types of fabrication processes were tried and tested. For the case of the one-step method, perovskite precursor solutions were prepared by dissolving mixed powders of PbI₂ (Aldrich) and methyl-ammonium iodide (MAI, Aldrich) in the mixed solution of anhydrous N,N-dimethylformamide and dimethyl sulfoxide, where the ratio was 3:1. Then, the perovskite precursor solutions were spin coated on FTO/TiO₂ films under the nitrogen environment. During the spin-coating process, a single drop (180 μL) of chlorobenzene was added as antisolvent crystallization. Finally, the perovskite-coated FTO/TiO₂ electrode was heated at 100 °C for 10 min. For the case the of two-step method, the PbI₂ solution was prepared by dissolving PbI₃ powders into the mixture of N,N-dimethylformamide (DMF) and dimethylsulfoxide (DMSO), and a mixture of formamidinium iodine (FAI), methylammonium bromide (MABr), and methylammonium chloride (MACl) dissolved in isopropyl alcohol (IPA) as the second-step solution. Firstly, first-step PbI₂ solution was spin coated and annealed at 70 °C for 10 min. Additionally, sequentially, second-step solutions were spin coated and annealed at 150 °C for 30 min, and the perovskite film was formed. For finishing the coating of the perovskite layer, the 2,2,7,7-tetrakis(N,N-di-p-methoxyphenylamine)-9,9-spirobifluorene (Spiro-OMeTAD) solution was prepared by dissolving 80 mg of Spiro-OMeTAD (Aldrich), 22.4 μL of lithium bis(trifluoromethane-sulfonyl)imide (LiTSFI, Wako) solution (520 mg of LiTSFI in 1 mL acetonitrile, Aldrich), and 36 μL of 4-tert-butylpyridine (Aldrich) in 1 mL of chlorobenzene (Aldrich). This hole-transport layer was deposited on top of the FTO/TiO₂/perovskite substrate by the spin-coating method in air. Finally, 40 nm-thick gold was thermally evaporated with an evaporation rate of around 1.0 Å/s. Final fabricated device structures are FTO/TiO₂/(without or with PEIE)/(one- or two-step) perovskite/Spiro-OMeTAD/Au [29,34].

The surface morphologies were investigated using a field emission scanning electron microscope (FE-SEM, JEOL 6700F). The photo-sensing characteristics of the PePSs were measured using a semiconductor parameter analyzer (HP4145B) and solar simulator (Yamashita Denso, YSS E40) under Air Mass (AM) 1.5, 100 mW/cm². Photocurrent density J_p and dark-current density J_d are evaluated at the voltage of −0.2 V.

3. Results and Discussions

Figure 2 shows the FE-SEM observation on the TiO₂ layer fabricated by the CBD method [33]. The surface morphology was comparatively flat and grain size was observed between 0.1 and 1 μm. In our previous EB evaporation method [36], a flat surface was also observed; however, there are many raised portions with shadow regions on the TiO₂ surface. In the CBD method, there are few raised portions and many small structures of several 10 nm on the TiO₂ surface.

There is much literature on morphology-photovoltaic properties correlation [37], comparative study [38], and the effects of process parameters by the sequential-coating method [39] fabricated by one-step and two-step coating method. However, combined with the perovskite layer, the CBD method of TiO₂ is not confirmed in detail. Therefore, various combinations of perovskite layers using one-step and two-step methods on the TiO₂ using the CBD method are evaluated again. Figure 3 shows photographs of FE-SEM observation of perovskite structures. Figure 3a,b are the structure of FTO/TiO₂ (CBD method)/perovskite (one-step method), and Figure 3c,d are the structure of FTO/TiO₂ (CBD method)/perovskite (two-step method). Upper view photographs (Figure 3a,c) are lower magnification, and lower view photographs (Figure 3b,d) are higher magnification. The crystalline size of the one-step method was slightly larger than that of the two-step method. In Figure 3b, black opening points are holes without perovskite structure. In Figure 3d, the white particles are unreacted PbI₂.

Photo-sensing characteristics of current density (J) versus voltage (V) characteristics under dark and AM1.5 illumination for the case of the one-step method without PEIE are shown in Figure 4 [34] The J_p and J_d are 15.1 mA/cm² and 2.58 μA/cm², respectively. The ratio of photo- and dark-current density was 5.85 × 10³. This value was a little higher than that of the previous study with the EB-evaporated TiO₂ layer of 1.33 × 10³ [36]. One of the possible reasons why the ratio of photo- and dark-current density was larger compared to the device with the EB-evaporated TiO₂ is due to the smaller J_d for the case of a device with TiO₂ layer by the CBD method. In this method, the surface morphology of the TiO₂ layer was smoother with no obvious pin-hole than that of the EB evaporation method. This smoother surface reduces some defect points and resultant current flow due to the higher concentration of electric field at a convex and concave surface. By the way, we will discuss the polarity of the dark-current density changed at +0.57 V. This could be explained that some generation current will affect this offset voltage. As discussed later, in Figure 5, the current density below +0.57 V was smaller than in the case of the one-step method without PEIE. Thus, within the bias condition of the measurement under reversed bias condition, the current density measured by the dark condition is constant; in other words, there is no reverse bias dependence. Therefore, the generation current will occur at some generation point, for example, at the interfacial states between TiO₂ and the perovskite layer. Other remaining issues are a lower series resistance and an optimization of the thickness in the TiO₂ and the perovskite layer. Nevertheless, the photocurrent density of 15.1 mA/cm² could be achieved. For the evaluation of the solar cell characteristics, the PCE of 12.4%, the short-circuit current density of 15.1 mA/cm², the open-circuit voltage of 1.06 V, and the fill-factor of 0.77 were obtained. The fabrication condition of the perovskite layer for further lowering humidity is the still an issue for this lowness of device characteristics.

In order to obtain process controllability of the device fabrication process, the two-step method for the perovskite layer preparation is tested because the timing of the single-drop of the MAI solution is crucial and less reproducible than the single-step method. In the previous paper, the perovskite crystal formation on a modified self-assembled monolayer, 3-aminopropyl triethoxysilane (APTES), a protonated amine silane coupling agent (PASCA-Br), between TiO₂ and perovskite layers are discussed [29]. As in a similar matter to this situation, large polycrystalline domains will be grown on the surface and the electron injection condition will be also improved [32]. Figure 5a,b show the photo-sensing characteristics, without and with 4 wt% solutions-coated PEIE device, respectively. By inserting the PEIE layer, the photocurrent of 22.4 mA/cm² was a little decreased by comparing 27.6 mA/cm² for the device without PEIE layer. The ratio of photo to dark current was 2.23 × 10⁴ and 2.88 × 10⁴ for the case of without and with PEIE layer, respectively. At this moment, the photocurrent was decreased a little by inserting the PEIE layer. If the tunneling current between the perovskite layer and TiO₂ is dominant, the photocurrent will be increased regardless of the insertion of insulating PEIE layer. On the other hand, if a thicker PEIE layer is formed on the TiO₂ layer, the series resistance R_s will be increased. In addition, the dark-current will also be decreased. There is an optimum thickness of PEIE layer by changing the concentration and spin-rinse condition. As a result, the photodiode performance was slightly better for the device with the PEIE layer.

For this device fabrication condition, the series resistance R_s, estimated to be 466 Ωcm², was observed at the voltage of around +0.90 V, as shown in Figure 6. By inserting the PEIE layer, this series resistance was dramatically decreased to 20.5 Ωcm². As a result, the photo-sensing characteristics around the voltage of +0.9 V were different between Figure 5a,b. It is considered that a tunneling current was increased and better electron injection was carried out by employing the PEIE layer. Additionally, another interesting point for these experimental results is discussed. The photocurrent of 27.6 mA/cm² (−0.2 V) and 31.1 mA/cm² (−1.0 V) was larger than that of the short circuit in the common solar cells of 24.5 mA/cm² [40]. Therefore, in the common solar cells, there is still an improvement point for the perovskite crystal growth even in the PCE of over 20%. Finally, the dark current under the bias voltage is discussed. For the case of MAPbI₃ perovskite single-crystal system, a linear ohmic regime followed by the trap-filled regime with a steep increase in current was reported using a low trap-state density of the order of 10⁹ to 10¹⁰ cm⁻³ [41]. Figure 7 shows the logarithm plot of the current density vs. voltage characteristics under forward and reverse bias condition in the device with FTO/TiO₂/PEIE/two-step perovskite/Spiro-OMeTAD/Au. Under reverse bias condition, the dark-current density depends on the applied reverse voltage as $J~{\left|V\right|}^{1.6}$.

This index number could not explain the common sense of the diode characteristics. However, the dark-current density is changed by changing the reverse bias condition. Therefore, the dark current under reverse bias conditions may be generated at the depletion layer of the perovskite layer and/or bias-dependent carrier injection layer between the electrode and perovskite layer [42].

Many issues will be solved by reducing the humidity of the glovebox (up to now, 30 ppm at the present situation) in order to reduce pinholes in the perovskite films [42] and the physical properties of the photo-sensing device by varying the temperature and device structures.

4. Conclusions

We investigated the PePSs with the solution-processed compact TiO₂ under a low-temperature process. The fabricated device structure is a photodiode structure of FTO/TiO₂/(without or with) PEIE/(one- or two-step) perovskite layer/Spiro-OMeTAD/Au. For the sensing characteristics, a ratio of photo-to-dark-current density was 2.88 × 10⁴ for the device with PEIE layer. During the device fabrication process, the maximum temperature of the baking condition for the fabrication of perovskite layer is 150 °C. Therefore, this process condition possibly applies to a flexible substrate. In addition, we checked the validity of the two-step process and PEIE interfacial layer. Further investigations are necessary to realize practical PePS devices.