Long-term Stable 2D Dion-Jacobson Phase Perovskite Photodiode with Low Dark Current and High On/off Ratio①

GU Hao CHEN Shan-Ci ZHENG Qing-Dong

a (College of Chemistry, Fuzhou Uniνersity, Fuzhou 350000, China)

b (State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fuzhou 350002, China)

ABSTRACT A two-dimensional (2D) organic-inorganic hybrid perovskite (OIHP) material is considered as a promising candidate for a long-term stable photodetector owing to its outstanding phase and environmental stability. Herein, we demonstrate a perovskite photodiode based on the DJ phase 2D perovskite(PDA)(MA)n-1PbnI3n+1 (where PDA is 1,3-propylenediamine, MA is methylamine, nominal n = 4). The best-performance device exhibits a high detectivity of 4.57 × 1011 Jones, a responsivity of 0.25 A·W-1 at 480 nm,a low dark current density of 9.60 × 10-4 mA·cm-2, and a remarkable on/off ratio of 3.10 × 105. The unencapsulated device can maintain 95% of the initial photocurrent density after 90 days under an ambient atmosphere with relative humidity (RH) of 65%, demonstrating its improved stability than the 3D counterpart.The excellent stability of the photodiode based on 2D perovskite promises its bright commercial application future.

Keywords: two-dimensional perovskite, Dion-Jacobson, photodiode, on/off ratio, stability;

1 INTRODUCTION

As a kind of photoelectronic devices to convert optical signals into electrical signals, photodetectors are widely used in aerospace communication, environmental monitor,biological sensor, and other fields[1-5]. At present, commercial photodetectors are mostly based on III-V inorganic semiconductors such as silicon (Si) and germanium (Ge), but they have the drawbacks of high cost and complicated fabrication process[6-9]. Therefore, it is urgent to fabricate photodetectors with low cost and simple preparation process.In recent years, perovskite materials have attracted increasing attention due to their excellent merits including high carrier mobility, adjustable bandgap, long charge diffusion length, and simple fabrication process[10-13]. As a typical organic-inorganic hybrid perovskite (OIHP),CH3NH3PbI3(MAPbI3) has been used for various practical applications such as solar cells, light-emitting diodes (LED),photodetectors, lasers, etc.[14-18]. However, the poor moisture and heat stability is still a problem that needs to be solved.Recently, two-dimensional (2D) OIHP materials, which are fabricated by adding a long chain or large volume organic ammonium salt to the perovskite components[19,20], are emerging due to their great potential to solve the stability issue of the conventional three-dimensional (3D) perovskite materials.

The 2DOIHP material was first used to improve the performance of solar cells. Smith et al.used (PEA)2(MA)2-Pb3I10(where PEA+is phenylethylammonium) as the active layer to fabricate solar cells with improved stability[21].Some other organic ammonium salts have also been used to prepare 2DOIHP, such as 1,3-propylenedia- mine (PDA)[19],butylamine (BA)[22], and so on. The encouraging results of solar cells based on 2Dperovskites inspire researchers to further explore other applications of the 2Dperovskites. For example, Han et al. fabricated highly oriented thin films with a new 2Dperovskite (PA)2(FA)Pb2I7(where PA isn-pentylaminium and FA is formamidine) for photodetectors,which realized an ultrafast response time of 2.54 ns as well as a high responsivity of 1.73 × 1014Jones[23]. Tsai et al.used a 2Dperovskite (BA)2(MA)2Pb3I10for X-ray photodiodes that can work at a very low external bias[24]. Lim et al.developed a photodiode based on 2Dperovskite by incorporating PEAI into MAPbI3. The current density of the 2Dperovskite photodiode maintained 76% of the initial current density after 80 days in the ambient condition,compared to 15% for the control sample of 3Dperovskite photodiode[25].

According to the types of organic amines used in the 2Dperovskite, 2Dperovskite can be divided into Ruddlesden-Popper (RP) and Dion-Jacobson (DJ) phases,which are prepared by using monoamine and diamine,respectively[26]. The RP phase 2Dperovskite has a general formula of (RNH3)2An-1MnX3n+1, in which the monoammonium cations only interact with inorganic perovskite layers at one side, and van der Waals gap exists between the adjacent cations. In the DJ phase 2Dperovskite with a formula of(NH3RNH3)An-1MnX3n+1, hydrogen bonds are formed between the amino group and the inorganic component at both ends. Compared with the RP phase 2Dperovskites, the DJ ones can possess better stability owing to the elimination of the van der Waals gap (see illustration in Fig. 1)[19,27]. So far, there are only a few reports about the DJ phase perovskites for photodetectors[28,29]. And, in the reports, a photoconductor device configuration was adopted for the DJ phase 2Dperovskite photodetectors, which usually needs relatively high operating voltages. Meanwhile, they suffer from low on/off ratios (＜ 103) which limit their further applications. In contrast, photodiode-type photodetectors can work at zero bias and exhibit high on/off ratios owing to the rather low dark currents although their EQEs are normally less than 1. To the best of our knowledge, DJ phase 2Dperovskites have never been used for fabricating the photodiode-type photodetectors until now.

Fig. 1. Schematic illustration of RP and DJ phase 2D perovskites

Herein, we fabricated a high-performance stable photodetector based on a DJ phase 2Dperovskite derived from 1,3-propyleneamine (PDA). A high-quality film of(PDA)(MA)n-1PbnI3n+1(nominaln= 4) was prepared by one-step method. The resulting photodiodes exhibit lower dark current and better stability than the 3Dreference devices. The device based on DJ phase 2Dperovskite exhibits a responsivity of 0.25 A·W-1and a high detectivity of 4.57 × 1011Jones at 480 nm. In addition, the device has a dark current density as low as 9.60 × 10-4mA·cm-2and a maximum on/off ratio of 3.10 × 105, which are remarkable in the 2Dperovskite photodiodes. Furthermore, the devices based on the DJ phase 2Dperovskite layer without encapsulation can maintain 95% of the initial photocurrent after storage for 90 days in the air with a relative humidity(RH) of 65%, while the conventional 3Dperovskite photodiodes can only maintain 35% of the initial photocurrent after the same storage time.

2 EXPERIMENTAL

2. 1 Materials

1,3-Propylenediamine and hydroiodic acid were purchased from Aladdin. The SnO2colloid precursor (Tin (IV)oxide, 15% in H2O colloidal dispersion) was purchased from Alfa Aesar. Lead iodide (PbI2) and methylammonium iodide(MAI) were purchased from Xi’an Polymer Technology Crop..Dimethylformamide (DMF), dimethylsulfoxide(DMSO),γ-butyrolactone (GBL), and chlorobenzene (CB)were purchased from Sigma-Aldrich. 2,2′,7,7′-Tetrakis[N,Ndi(4-methoxyphenyl)amino]-9,9′-spirobifluorene (Spiro-OMeTAD) was purchased from Shenzhen Feiming Science and Technology Co., Ltd.

2. 2 Synthesis of (PDA)I2

1,3-Propylenediamine iodide ((PDA)I2) was prepared by adding dropwise hydroiodic acid (3.40 mL, 30 mmol) to a solution of 1,3-propylenediamine (0.74 g, 10 mmol) in ethanol at 0 ℃ and stirring for 2 h[19]. The white precipitate of the diammonium iodide was recovered by evaporating the solvent at 55 ℃ using a rotary evaporator. The precipitate was washed three times with ether and dried overnight in an oven at 60 ℃ under reduced pressure to obtain a white powder (1.70 g, 85% yield).

2. 3 Preparation of 2D perovskite film

The 2Dperovskite PDA(MA)3Pb4I13precursor solutions(the concentration of Pb2+is 0.8 M) were prepared by dissolving a stoichiometric ratio of 4:3:1 of PbI2, MAI, and(PDA)I2in the mixed solution of GBL:DMSO (7:3, volume ratio), and then stirred at 65 ℃ before use. The prepared solutions were filtered by PTFE syringe filter (0.2 μm) and coated onto the substrate with a consecutive spin-coating process at 6000 rpm for 35 s. Before the end of 15 s, 100 μL of chlorobenzene was dropped onto the perovskite film for better crystallization, followed by annealing at 120 ℃ for 10 min.

2. 4 Fabrication of devices

ITO glass substrates were cleanedνiasequential ultrasonic treatments with acetone, isopropyl alcohol, and deionized water for 30 min, respectively. Then, they are dried in the oven at 75 ℃. Before deposition of the SnO2electron transport layer (ETL), the ITO substrates were pretreated by UV-O3for 15 min. And the SnO2colloid solution with a concentration of 7.5 wt% was spin-coated on the ITO at 3000 rpm for 30 s, followed by annealing at 120 ℃ for 10 min and 150 ℃ for 20 min. Then, the coated substrates were transferred to a N2-filled glovebox to deposit the perovskite onto the ETL using the aforementioned method. Next, 100 μL of Spiro-OMeTAD in CB (72.3 mg·mL-1) containing additives of 17.5 μL of bis(trifluoromethane)-sulfonimide lithium salt (520 mg·mL-1in acetonitrile)and 28.8μL of 4-tert-butylpyridine was spin-coated at 4000 rpm for 30 s as a hole transport layer. Finally, 80 nm of Au was deposited as a top electrode using a thermal evaporator system under a vacuum of ～10-4Pa.

2. 5 Characterization of perovskite films and devices

The absorption spectra were obtained by using a UV-Vis spectrophotometer (Perkin-Elmer, Lambda 365). X-ray diffraction (XRD) spectra were measured by a Rigaku MiniFlex 600 diffractometer equipped with a CuKαradiation at room temperature. The ultraviolet photoelectron spectra(UPS) for perovskite films were carried out by using X-ray Photoelectron Spectroscopy (Thermo Fisher, ESCALAB 250Xi). The root-mean-square (RMS) roughness of perovskite films was investigated by Scanning Probe Microscope (Bruker, Dimension Icon). The scanning electron microscopy (SEM) images were obtained by using the Field Emission Scanning Electron Microscope(SU-8010). TheJ-Vcurves were measured by using an Oriel Sol3A simulator (Newport) under AM 1.5 irradiation (100 mW·cm-2) and recorded by a Keithley 2400 source measurement unit. The light intensity was calibrated by a standard Si reference solar cell that has been certified by the National Renewable Energy Laboratory (NREL). The external quantum efficiency (EQE) measurements were conducted by using a Newport EQE measuring system. The active areas of the devices were fixed at 6.25 mm2. The unencapsulated 10 devices were tested and stored in the ambient under 65% RH.

3 RESULTS AND DISCUSSION

The DJ phase 2Dperovskite (PDA)(MA)n-1PbnI3n+1(nominaln= 4) was prepared from a stoichiometric (4:3:1)reaction between PbI2, MAI, and (PDA)I2(see Experimental section for details). In order to identify the difference between 2Dand 3Dperovskite films, the MAPbI3films were also prepared for comparison. Fig. 2a shows the normalized absorption spectra for the 2Dand 3Dperovskite films. The optical bandgap (Eg) can be deduced from the absorption edge according toEg= 1240/λ, whereλis the absorption edge. The optical bandgaps of 2Dand 3Dperovskite films are 1.67 and 1.60 eV, respectively. X-ray diffraction (XRD)measurements were performed to study the crystallinity of 2Dand 3Dperovskite films. As shown in Fig. 2b, both of the perovskite films have two dominant peaks that appear around 14.2° and 28.4°, representing the diffraction from crystallographic planes (111) and (202), respectively[22,30].However, the peaks of the 3Dperovskite film slightly shift towards large angles, which indicates the extension in crystal lattice constant. In the region between 5° and 10°, the 2Dperovskite film shows the (040) and (060) peaks which characterize the 2Dperovskites[30]. Ultraviolet photoelectron spectroscopy (UPS) was also conducted to identify the difference between the 2Dand 3Dperovskite films on energy levels. The full UPS spectra are given in Fig. 2c. The magnified UPS spectra of the valence band edge (Eonset) and the secondary electron cutoff edge (Ecutoff) are depicted in Fig. 2d. As shown in Fig. 2d,Ecutoffvalues for the 2Dand 3Dperovskite films are 16.10 and 16.22 eV, respectively.Similarly, the 2Dand 3Dperovskite films have the sameEonsetvalue of 1.77 eV. The highest occupied molecular orbital (HOMO) energy level can be calculated according to HOMO = 21.22 eV -Ecutoff+Eonset, and the lowest unoccupied molecular orbital (LUMO) energy level can be calculated by the HOMO and optical bandgap[31]. Thus, the HOMO energy levels of 2Dand 3Dperovskite films are calculated to be -6.89 and -6.77 eV, respectively, with their corresponding LUMO energy levels to be -5.22 and -5.17 eV, respectively.

Fig. 2. (a) Absorption spectra of the 2D and 3D perovskite films. (b) X-ray diffraction (XRD) spectra of the 2D and 3D perovskite films.(c) Full UPS spectra of the 2D and 3D perovskite films. (d) UPS spectra for the valence band edge and secondary electron cutoff edge of the 2D and 3D perovskite films

We also studied the morphological evolution of 2Dand 3Dperovskite films by using scanning electron microscopy(SEM) and atomic force microscopy (AFM). The SEM and AFM images of the 2Dand 3Dfilms are shown in Fig. 3. As shown in Fig. 3a, the grain size of 2Dperovskite is about 50～200 nm, while the 3D perovskite has a larger grain size up to 300 nm (Fig. 3b). The root-mean-square (RMS)roughnesses of the 2Dand 3Dfilms (Fig. 3c and d) are determined to be 4.13 and 7.41 nm, respectively, suggesting the smoother morphology of the 2Dperovskite film. Besides,the 3Dperovskite exhibited large grain, which is consistent with the SEM results.

Fig. 3. SEM images of (a) 2D and (b) 3D perovskite films. The AFM images of (c) 2D and (d) 3D perovskite films

To investigate the photodetection performance of the(PDA)(MA)n-1PbnI3n+1(n= 4) 2Dperovskite material, photodiodes with a device architecture of ITO/SnO2/Perovskite/Spiro-OMeTAD/Au (as shown in Fig. 4a and Fig. S1)were fabricated. For comparison purposes, the photodiodes based on 3Dperovskite were also fabricated with the same device configuration. To prepare the perovskite photodiode,SnO2was spin-coated on top of the ITO substrate as an ETL,and then the perovskite film was deposited on the ETL. Later,Spiro-OMeTAD was spin-coated onto the perovskite layer as a hole transport layer (HTL) followed by the sequentially thermal evaporation of Au as a top electrode. To explore the difference between the devices based on 2Dand 3Dperovskite films, we tested the external quantum efficiencies(EQEs) of the perovskite photodiodes. The corresponding EQE spectra of the devices are shown in Fig. 4b. The photoresponse range of the devices based on the 2Dperovskite is 300～750 nm, which is narrower than that of the 3Ddevices (300～800 nm). The trend of EQE spectra is consistent with that of the absorption spectra. The maximum EQE for the devices based on the 3Dperovskite film is 77%at 470 nm, which is higher than that of the 2Ddevices (67%at 460 nm). The lower EQE for the devices based on 2Dperovskite film can be attributed to the existence of the insulating PDA layers which hinder the charge transport. As shown in Fig. 4c, the dark current densities of the photodiodes based on 2Dand 3Dperovskite films are 9.60 ×10-4and 3. 84 × 10-3mA·cm-2, respectively. The dark current of devices based on the 2Dperovskite film is much lower than that of the 3Dones, which is beneficial to improve the on/off ratio of photodiodes. The photocurrent of devices based on the 3Dperovskite is higher than that of the 2Ddevices (Fig. 4d), which is consistent with the results reported in the literature[28]. The on/off ratio determined byIphoto/Idarkis 3.10 × 105for the device based on the 2Dperovskite film. Owing to its low dark current, the on/off ratio for the 2Dperovskite-based device is much higher than that of the 3Ddevice (7.40 × 103).

Fig. 4. (a) Configuration of the photodiode. (b) EQE spectra, (c) dark current density, and (d) photocurrent density of the photodiodes based on 2D and 3D perovskite films measured under one solar (AM 1.5G, 100 mW·cm-2) illumination

Fig. 5. (a) Responsivity, and (b) detectivity of photodiodes based on the 2D and 3D perovskite filmsmeasured under one solar (AM 1.5G, 100 mW·cm-2) illumination

According to the definition, the responsibility is the ratio of photocurrent to the incident-light intensity, indicating the ability of a device in response to the optical signal.Thereafter, it can be given as follows[16,32]:

whereRis the spectral responsivity,Jphotothe photocurrent density andLlightthe incident light intensity. Since EQE can be calculated by the number of photo-generated electrons divided by the number of incident photons, equation (1) can also be described by the following equation (2):

whereqis the electron charge,λthe corresponding wavelength of light,hthe Planck's constant, andcthe light speed. Using the measured EQE results, we can calculate the wavelength-dependent responsibility of the perovskite photodiodes through equation (2). The highest responsivity data of the perovskite photodiodes were 0.25 A·W-1at 480 nm and 0.30 A·W-1at 530 nm for 2Dand 3Dperovskites,respectively (Fig. 5a). The high charge accumulation and increased recombination losses will hinder the photoelectric performance. Thus, the responsivity of the device based on 2Dperovskite is lower than that of the 3Dperovskite-based counterpart. Another parameter to evaluate the characteristics of the photodiode is detectivity (D). The detectivity characterizes how weak the light can be detected. It is determined by the responsivity and the dark current of a photodetector. In general, the detectivity can be calculated by equation (3):

whereqis the electron charge andJdthe dark current density.Considering all of the above equations, a high EQE along with a low dark current can promote the detectivity of a photodetector. The device based on 2Dperovskite films showed the highest detectivity of 4.57 × 1011Jones at 480 nm, which is higher than that of 3Dperovskite-based counterpart (2.76 × 1011Jones at 530 nm). Although the higher EQE and responsivity were obtained in the devices based on 3Dperovskite film, the devices based on 2Dperovskite film exhibited higher detectivity than the 3Dperovskite-based devices. It can be due to the fact that the organic cation (PDA+) in the perovskite crystalline structure reduces the dark current. However, the organic cation can also result in the reduction of photocurrent. Therefore, an appropriate amount of organic cations should be used in order to achieve the balance between the photo and dark current. The performance parameters of the devices based on 2Dand 3Dperovskite films are summarized in Table 1.Furthermore, we summarized the performance parameters of our and others’ work in Table 2. It is shown that the on/off ratio was ultrahigh for the devices based on 2Dperovskite film in this work, which is 2～3 orders of magnitude higher than the values of the other reported photodetectors based on the DJ phase 2Dperovskites[28,29]. Meanwhile, the responsivity and detectivity of 2Dperovskite-based devices are higher than the reports of DJ-type perovskites.

Table 1. Photodetection Performance Parameters for the Devices Based on 2D and 3D Perovskite Films

Table 2. Comparison of Performance Parameters for the Photodetectors Based on 2D DJ Perovskites

As can be seen from Table 1, the photodiodes based on 2Dperovskite displayed better performance with a lower dark current, higher detectivity and on/off ratio. Since stability is the most prominent advantage of 2Dperovskite materials compared to 3Dones, the stability of both unencapsulated devices was investigated under ambient condition (65% RH).The normalized photocurrent densities and the dark current densities of two types of photodiodes as a function of time are shown in Fig. 6. As shown in Fig. 6a, the normalized photocurrent density of the 3Dperovskite-photodiode showed a great decrease and only maintained 35% of the initial photocurrent value after 90 days of storage. In contrast,the 2Dperovskite-photodiode exhibited significantly improved stability with 95% of the initial photocurrent under the same storage condition. Fig. 6b shows the dark current densities of the two types of devices as a function of time.The dark current density of the devices based on 2Dperovskites maintained ～10-4mA·cm-2all along, while that of the 3Dones increased from 3.84 × 10-3mA·cm-2to 5.80× 10-3mA·cm-2after 90 days. It is well known that the increase of dark current is detrimental to the device performance. Therefore, the devices based on 2Dperovskites are more promising in practical applications.

Fig. 6. (a) Normalized photocurrent density, (b) the dark current density of the photodiodes based on 2D and 3D perovskite films as a function of the storage time

Fig. 7. Photoresponse performance of photodiodes based on 2D perovskite film measured under the repeated cycles of illumination of one solar light (100 mW·cm-2, AM 1.5G). (a) Time-dependent photocurrent response (I-t) curves. (b) Rise time. (c) Fall time.(d) The noise current measured under one solar (100 mW·cm-2, AM 1.5 G) illumination

We further investigated the light response speed of the photodiode based on 2Dperovskite film by testing the response time of the device under one solar light illumination (Figs. 7a and S2). Fig. 7a shows the timedependent photocurrent response curves within 10 s under switching state. Fig. 7b and c show the rise and fall time of the photodiode, respectively. The rise time (fall time) is defined as the time for the photocurrent to rise from 10% to 90% (fall from 90% to 10%) of the maximum value during the on and off cycles[31,33,34]. The calculated rise time for the

4 CONCLUSION

In summary, we incorporated a diamine (PDA) into the conventional 3Dperovskite component to form a DJ phase 2Dperovskite PDA(MA)n-1PbnI3n+1(nominaln= 4). The photodiode based on 2Dperovskite exhibited better performance with a high detectivity (4.57 × 1011Jones at 480 nm) and a low dark current density (9.60 × 10-4mA·cm-2) in comparison with the device based on 3Dperovskite. Also,2Dperovskite-based photodiode is 20 ms and the fall time is 45 ms, which exhibited comparable response time with the devices reported by others[25,28]. The fast rise and fall time suggested the fast response speed of the devices based on 2Dperovskite. Fig. 7d shows the level of the noise current of the 2Dperovskite-based photodiode. It is found that the overall noise current is stable at 10-9～10-10A·Hz-1/2. Moreover, the noise current stabilized quickly at low frequencies (within 5 Hz), which is also of great significance for the practical application of the photodiodes. The highest on/off ratio of the device based on 2Dperovskite is 3.10 × 105, which is impressive for 2Dperovskite-based photodiodes. The unencapsulated photodiodes based on 2Dperovskite can be stored under air with RH of 65% for over 90 days without significant degradation, showing their improved stability than the 3Dperovskite-based photodiodes. This long-term stable 2DDion-Jacobson phase perovskite photodiode with low dark current and high on/off ratio is promising for the practical switching applications.