Dimethylammonium Cation-Induced 1D/3D Heterostructure for Efficient and Stable Perovskite Solar Cells

Mixed-dimensional perovskite engineering has been demonstrated as a simple and useful approach to achieving highly efficient and more-durable perovskite solar cells (PSCs), which have attracted increasing research interests worldwide. In this work, 1D/3D mixed-dimensional perovskite has been successfully obtained by introducing DMAI via a two-step deposition method. The additive DMA+ can facilitate the crystalline growth and form 1D DMAPbI3 at grain boundaries of 3D perovskite, leading to improved morphology, longer charge carrier lifetime, and remarkably reduced bulk trap density for perovskite films. Meanwhile, the presence of low-dimension perovskite is able to prevent the intrusion of moisture, resulting in enhanced long-term stability. As a result, the PSCs incorporated with 1D DMAPbI3 exhibited a first-class power conversion efficiency (PCE) of 21.43% and maintained 85% of their initial efficiency after storage under ambient conditions with ~45% RH for 1000 h.


Introduction
Halide perovskite solar cells (PSCs) have attracted significant attention due to their impressive potential for next-generation photovoltaic technology. They possess the merits of economical cost, capacity for large-scale fabrication, and remarkable power conversion efficiency (PCE) [1][2][3][4][5][6][7][8][9][10][11][12]. Since the first PSC fabricated by Kojima et al. in 2009 [1], numerous achievements have been accomplished to raise the PCE from 3.8% to 25.8% [5]. Despite inspiring development in past decades, the application of PSCs is severely restricted because of their poor stability against moisture, heat, and UV light, attributed to the inherent structural characteristics of conventional (three dimensional) perovskites [13][14][15][16]. To be specific, inevitable defects which form during fabrication process of PSCs and are located at grain boundaries and interfaces can not only serve as the center of trap-assisted non-radiative recombination, but also induce the decomposition of perovskite films in humid conditions and consequently decrease the efficiency and long-term stability of PSCs [17][18][19][20][21][22]. On the other hand, their severe toxicity is a potential threat to the environment and humankind [23].
To address these issues, various strategies have been attempted, such as additive engineering [23][24][25][26][27][28], interfacial modification [29][30][31][32][33][34], solvent engineering [2,35] and so on. Among them, the incorporation of one-dimensional (1D) perovskite has drawn great attention due to its unique characteristics in terms of stability [36][37][38][39]. The typical structural configuration of 1D perovskite can be described as the adjacent [PbX 6 ] 4− octahedra surrounded by A-site cations that are corner-sharing, edge-sharing, or face-sharing to form the "shoulder to shoulder" arrangement. The hydrophobic large-size organic spacer cations at the A-site can effectively prevent the destructive effect of moisture on the perovskite lattice [37,40]. On the other hand, this natural structure leads to intrinsic shortcomings such as ious aspects. However, the formation of low-dimension/3D perovskite has rarely been investigated, even though DMAPbI 3 has been proven to be 1D perovskite structure [56][57][58].
In this work, we have introduced DMAI into the precursor solution to prepare mixeddimensional perovskite and systematically investigate its specific effect on photovoltaic performance and long-term stability. All results revealed that DMAI could facilitate the crystallization process and coordinate with excessive PbI 2 to form 1D DMAPbI 3 , which has a face-sharing connection of [PbI 6 ] 4− octahedra. Moreover, the formation of 1D perovskite tends to locate at grain boundaries, leading to a reduced bulk trap density, efficient passivation of defects, and promotion of charge carrier transfer capability. Moreover, the presence of 1D perovskite significantly retarded the unfavorable intrusion of moisture, hence leading to the improved long-term stability of mixed-dimensional perovskites. As a result, the PSCs modified with 1D DMAPbI 3 achieved a first-rate efficiency of 21.43%, whereas the unencapsulated device retained 85% of its initial efficiency after aging under ambient condition with~45% RH for 1000 h.

Results and Discussion
We prepared pure 1D perovskite and mixed-dimensional perovskite films via a twostep method, where PbI 2 and organic salt precursor solutions were sequentially deposited on the substrates by spin-coating and annealed to form target perovskite films. Various concentrations of DMAI were dissolved in IPA with the typical organic ammonium halides. More experimental details are stated in the experimental section. The crystal structure and crystallographic parameters of 1D perovskite with the chemical formula of DMAPbI 3 (CCDC: 1497287) are shown in Figure 1a and Table S1, which consists of 1D chains of face-sharing [PbI 6 ] 4− octahedra with large DMA + cations [59]. X-ray diffraction (XRD) patterns were carried out to confirm the formation of DMAPbI 3 and reveal the effect of DMAI on the crystallization process of perovskite.
As shown in Figure 1b, the XRD pattern of the deposited 1D DMAPbI 3 perovskite film matches exactly with the calculated pattern from its single-crystal data; the strong peak located at 12.7 • can be assigned as the excess PbI 2 . Figure 1c shows the XRD patterns of mixed-dimensional perovskite films: five characteristic diffraction peaks at 14.1 • , 20.0 • , 24.4 • , 28.2 • , and 31.8 • are clearly observed, corresponding to the (100), (110), (111), (200), and (210) planes of 3D perovskite films [55]. For DMAI-treated perovskite films, no significant changes were reflected in the XRD patterns after adding small amounts of DMAI; as the concentration increased, another new diffraction peak appeared and gradually increased at 11.8 • , which was attributed to 1D DMAPbI 3 perovskites [57,58]. Meanwhile, the remnant PbI 2 peaks were almost eliminated when the concentration of DMAI was more than 5 mg/mL, suggesting that additional DMAI in the precursor solution can react with excess PbI 2 to form 1D/3D perovskite. Additionally, the maximum diffraction peaks attributed to (100) planes of 3D perovskite gradually shifted to a lower degree after addition of DMAI, as shown in Figure S1, indicating lattice expansion caused by substituting large cations into the perovskite A-site [60]. We carried out 2D grazing-incidence wide-angle Xray scattering (GIWAXS) measurements to investigate the crystal structure and orientation of mixed-dimensional perovskite films. Figure 1e and Figure S2 show the GIWAXS images and integrated intensity of corresponding samples; the 1D/3D samples exhibit a barely changed intensity of diffraction rings at q z = 1 Å −1 , which corresponds to the (100) lattice plane of 3D perovskite. In contrast, remarkably diminished intensity of PbI 2 diffraction spots at q z = 0.9 Å −1 were observed. In addition, a new diffraction ring at q z = 0.85 Å −1 that can be assigned as 1D DMAPbI 3 perovskite was observed in high-concentration samples. In the following discussion, the additional concentration of DMAI in modified samples was 0.5 mg/mL, unless otherwise noted.
The existence of 1D DMAPbI 3 in perovskite films and its influence on surface morphology are also characterized by top-view scanning electron microscopy (SEM) images, as shown in Figure 2. As shown in Figure 1b, the XRD pattern of the deposited 1D DMAPbI3 perovskite film matches exactly with the calculated pattern from its single-crystal data; the strong peak located at 12.7° can be assigned as the excess PbI2. Figure 1c shows the XRD patterns of mixed-dimensional perovskite films: five characteristic diffraction peaks at 14.1°, 20.0°, 24.4°, 28.2°, and 31.8° are clearly observed, corresponding to the (100), (110), (111), (200), and (210) planes of 3D perovskite films [55]. For DMAI-treated perovskite films, no significant changes were reflected in the XRD patterns after adding small amounts of DMAI; as the concentration increased, another new diffraction peak appeared and gradually increased at 11.8°, which was attributed to 1D DMAPbI3 perovskites [57,58]. Meanwhile, the remnant PbI2 peaks were almost eliminated when the concentration of DMAI was more than 5 mg/mL, suggesting that additional DMAI in the precursor solution can react with excess PbI2 to form 1D/3D perovskite. Additionally, the maximum diffraction peaks attributed to (100) planes of 3D perovskite gradually shifted to a lower degree after addition of DMAI, as shown in Figure S1, indicating lattice expansion caused by substituting large cations into the perovskite A-site [60]. We carried out 2D grazing-incidence wide-angle Xray scattering (GIWAXS) measurements to investigate the crystal structure and orientation of mixed-dimensional perovskite films. Figures 1e and S2 show the GIWAXS images and integrated intensity of corresponding samples; the 1D/3D samples exhibit a barely changed intensity of diffraction rings at qz = 1 Å −1 , which corresponds to the (100) lattice plane of 3D perovskite. In contrast, remarkably diminished intensity of PbI2 diffraction It is obvious that numerous lead iodide grains were distributed in the control sample, whereas fewer similar grains were observed in the 1D/3D modified samples as the content of DMAI continuously increased, and finally disappeared. As the concentration further increased, several fiber-shaped perovskite crystals formed (as shown in Figure S3), and the residual PbI 2 completely converted into 1D nanofiber perovskite crystals. When the addition was 10 mg/mL, we found more rod-shaped crystals form at the grain boundaries in morphology images. Therefore, it can be deduced that 1D perovskite was most likely located at the grain boundaries (GBs) for the samples with at low concentration addition of DMAI, which was beneficial to reducing non-radiative recombination through suppression and passivation of defects at GBs. To study the effect of DMAI on the crystallite size of perovskite, we measured the crystallite size, as seen in Figure 2a,b; the statistical results are illustrated in Figure 2c,d. The size ranged from 100 nm to 700 nm for the control sample while it increased to between 200 to 800 nm with the addition of DMAI. After fitting by the equation: we obtained the average crystallite size of 389 ± 111 nm and 503 ± 114 nm for the control and modified samples, respectively. These results indicate that DMAI could facilitate the crystallization process with the reduction of GBs.
X-ray photoelectron spectroscopy (XPS) was used to investigate the effect of DMAI on the surface chemistry properties of perovskite films. As shown in Figure 3a, two dominant peaks of Pb 2+ 4f 7/2 and Pb 2+ 4f 5/2 are located at 138.3 eV and 143.1 eV in the control film; both peaks shifted about 0.4 eV toward the lower-energy region when modified with DMAI, which was ascribed to interaction between uncoordinated Pb 2+ and DMA + cation ligands [32].
Molecules 2021, 26, x FOR PEER REVIEW 5 of 12 spots at qz = 0.9 Å −1 were observed. In addition, a new diffraction ring at qz = 0.85 Å −1 that can be assigned as 1D DMAPbI3 perovskite was observed in high-concentration samples. In the following discussion, the additional concentration of DMAI in modified samples was 0.5 mg/mL, unless otherwise noted. The existence of 1D DMAPbI3 in perovskite films and its influence on surface morphology are also characterized by top-view scanning electron microscopy (SEM) images, as shown in Figure 2. It is obvious that numerous lead iodide grains were distributed in the control sample, whereas fewer similar grains were observed in the 1D/3D modified samples as the content of DMAI continuously increased, and finally disappeared. As the concentration further increased, several fiber-shaped perovskite crystals formed (as shown in Figure S3), and the residual PbI2 completely converted into 1D nanofiber perovskite crystals. When the addition was 10 mg/mL, we found more rod-shaped crystals form at the grain boundaries in morphology images. Therefore, it can be deduced that 1D perovskite was most likely located at the grain boundaries (GBs) for the samples with at low concentration addition of DMAI, which was beneficial to reducing non-radiative recombination through suppression and passivation of defects at GBs. To study the effect of DMAI on the crystallite size of perovskite, we measured the crystallite size, as seen in Figure 2a  To further investigate the impact of 1D DMAPbI 3 on optical and electronic properties of mixed-dimensional perovskite films, ultraviolet-visible (UV-Vis) absorption, steady-state photoluminescence (SSPL), and time-resolved photoluminescence (TRPL) measurements were conducted. Figure 3b shows the UV-Vis absorption spectra of the perovskite films. All samples exhibited an absorption threshold around 800 nm and no apparent shifts occurred. Meanwhile, the Tauc-Plot spectra in Figure 3c reveal that the optical bandgap of control is estimated to be about 1.57 ± 0.02 eV, which is almost the same as the modified sample with incorporation of DMAI, suggesting that the addition of a small amount of DMAI has negligible effect on the optical bandgap of mixed 1D/3D perovskite, which has been proven in previous research [47]. Figure 3d shows that the addition of DMAI increases the SSPL intensity, which was attributed to the improved charge transport and suppressed nonradiative recombination in 1D/3D perovskite films with reduced trap density. Similarly, the bandgap calculated from PL was around 1.51 ± 0.05 eV for perovskite with and without the addition of DMAI ( Figure S5b), which was smaller than that of 1.57 ± 0.03 eV calculated from the UV-Vis absorption result. The decrease in the bandgap could be attributed to the existence of a Stokes shift in the perovskite materials. The charge carrier lifetime of control and modified films was characterized by TRPL spectra (Figure 3e); the excitation source wavelength was 440 nm, and the decay curves were fitted with an exponential function. The modified film showed a much longer lifetime (1993.7 ns) over the control film (657.4 ns).
The enhanced PL intensity and improved carrier lifetime imply the efficient passivation of defects in the mixed-dimensional perovskite films, which can improve device performance. We adopted the space charge limited current (SCLC) method to systematically quantify the trap density of perovskite films. The dark J-V characteristic curves of electron-only devices with a configuration of glass/ITO/SnO 2 /perovskite/PCBM/Au are shown in Figure 3f. The trap-state density can be calculated by the equation: Nt = 2ε 0 ε r V TFL /(qL 2 ), where ε 0 , ε r , and V TFL are the vacuum permittivity, the relative dielectric constant (ε r = 46.9), and the trap-filled limited voltage, respectively; q is the elementary charge, and L is the film thickness. The V TFL of modified film (0.18 V) was lower than that of the control sample (0.30 V), which corresponds to the reduced trap density from 4.04 × 10 15 cm −3 to 2.42 × 10 15 cm −3 , indicating that the incorporation of DMAI effectively reduced electron trap density and diminished trap-assisted non-radiative recombination [47].
To evaluate the photovoltaic (PV) performance of DMAI-modified PSCs, devices with a typical planar structure of glass/ITO/SnO 2 /perovskite/Spiro-OMeTAD/Au were fabricated and characterized. The optimum concentration of DMAI was 0.5 mg/mL; the current density voltage (J-V) curves of the control and modified devices are shown in Figure 4a and Figure S7, along with corresponding PV parameters listed in the insets and Table S2.
The best 1D/3D modified device exhibited a remarkably improved open-circuit voltage (V oc ) of 1.16 V, compared with that of 1.08 V for the control device. Meanwhile, the shortcircuit current density (J sc ) and fill factor (FF) were enhanced from 23.64 mA·cm −2 and 73.4% to 23.75 mA·cm −2 and 77.8%, respectively, resulting in a significant improvement of PCE from 18.74% to 21.43%. Furthermore, the 1D/3D device showed less hysteresis (2.1%) than the 5.4% of the control device. Figure 4b shows the external quantum efficiency (EQE) spectra and the integrated J sc of corresponding devices; the modified device exhibited a maximum EQE value of 93.1% within the wavelength range of 300 to 900 nm. Significantly, the incorporation of 1D perovskite enhanced the EQE value at 300-380 nm, similar phenomena to which have been reported in previous work [40,49,61]. The integrated J sc calculated from EQE spectra reached 22.55 mA·cm −2 and 22.64 mA·cm −2 for control and modified devices, respectively. To study the reproducibility of our devices, we followed the method mentioned in the previous report [62]; the statistical distribution of the PV parameters from 15 individual devices with different concentrations of DMAI are summarized in Figure S8, revealing that the introduction of 1D perovskite in PSCs improves reproducibility, and the increment of PCE is owed primarily to the enhanced V oc and FF.
To further investigate the mechanism of these improvements caused by the incorporation of 1D perovskite, we adopted light intensity-dependent Voc measurement, electrochemical impedance spectroscopy (EIS), and dark current measurement. As shown in Figure 4c, the modified PSC shows a smaller slope of 1.32 k B T/q compared to 1.72 k B T/q for the control device. Herein, the deviation between the calculated slope and unity k B T/q (where k B is the Boltzmann constant, T is the temperature, and q is the elementary charge) reflects the rate of trap-assisted Shockley -Read-Hall recombination; a smaller slope indicates a less trap-assisted non-radiative recombination with the addition of 1D perovskite [63][64][65]. Two semicircles can be found in each EIS curve; as shown in Figure 4d, each EIS curve can be classified to successively reflect charge transfer resistance (R tr ) and recombination resistance (R rec ) [66]. It is obvious that 1D/3D perovskite devices exhibited decreased R tr and increased R rec , indicating the promoted charge transfer and inhibited charge recombination, thus leading to improved V oc and J sc . Figure 4e shows the J-V curves in dark conditions, the lower leakage current of modified PSC compared to control sample indicating the notable suppressed charge recombination and promoted charge extraction attributed to the prominent reduction of trap density and efficient passivation of bulk defects, which simultaneously leads to the increase in FF [67,68]. All the results of the modified PSCs are in good agreement with the excellent properties of the corresponding films as stated above, indicating the incorporation of 1D perovskite significantly improves the PV performance of PSCs. on the surface chemistry properties of perovskite films. As shown in Figure 3a, two dominant peaks of Pb 2+ 4f7/2 and Pb 2+ 4f5/2 are located at 138.3 eV and 143.1 eV in the control film; both peaks shifted about 0.4 eV toward the lower-energy region when modified with DMAI, which was ascribed to interaction between uncoordinated Pb 2+ and DMA + cation ligands [32]. To further investigate the impact of 1D DMAPbI3 on optical and electronic properties of mixed-dimensional perovskite films, ultraviolet-visible (UV-Vis) absorption, steadystate photoluminescence (SSPL), and time-resolved photoluminescence (TRPL) measurements were conducted. Figure 3b shows the UV-Vis absorption spectra of the perovskite films. All samples exhibited an absorption threshold around 800 nm and no apparent shifts occurred. Meanwhile, the Tauc-Plot spectra in Figure 3c reveal that the optical bandgap of control is estimated to be about 1.57 ± 0.02 eV, which is almost the same as the modified sample with incorporation of DMAI, suggesting that the addition of a small amount of DMAI has negligible effect on the optical bandgap of mixed 1D/3D perovskite, which has been proven in previous research [47]. Figure 3d shows that the addition of DMAI increases the SSPL intensity, which was attributed to the improved charge transport and suppressed non-radiative recombination in 1D/3D perovskite films with reduced trap density. Similarly, the bandgap calculated from PL was around 1.51 ± 0.05 eV for perovskite with and without the addition of DMAI ( Figure S5b), which was smaller than that of 1.57 ± 0.03 eV calculated from the UV-Vis absorption result. The decrease in the bandgap could be attributed to the existence of a Stokes shift in the perovskite materials. The charge carrier lifetime of control and modified films was characterized by TRPL spectra (Figure 3e); the excitation source wavelength was 440 nm, and the decay curves were fitted with an exponential function. The modified film showed a much longer lifetime (1993.7 ns) over the control film (657.4 ns). The enhanced PL intensity and improved carrier lifetime imply the efficient passivation of defects in the mixed-dimensional perovskite films, which can improve device performance. We adopted the space charge limited current (SCLC) method to systematically quantify the trap density of perovskite films. The dark J-V characteristic curves of electron-only devices with a configuration of glass/ITO/SnO2/perovskite/PCBM/Au are shown in Figure  3f. The trap-state density can be calculated by the equation: Nt = 2ε0εr VTFL/(qL 2 ), where ε0, εr, and VTFL are the vacuum permittivity, the relative dielectric constant (εr = 46.9), and the trap-filled limited voltage, respectively; q is the elementary charge, and L is the film thickness. The VTFL of modified film (0.18 V) was lower than that of the control sample (0.30 V), which corresponds to the reduced trap density from 4.04 × 10 15 cm −3 to 2.42 × 10 15 cm −3 , indicating that the incorporation of DMAI effectively reduced electron trap density and diminished trap-assisted non-radiative recombination [47].
To evaluate the photovoltaic (PV) performance of DMAI-modified PSCs, devices with a typical planar structure of glass/ITO/SnO2/perovskite/Spiro-OMeTAD/Au were fabricated and characterized. The optimum concentration of DMAI was 0.5 mg/mL; the current density voltage (J-V) curves of the control and modified devices are shown in Figures 4a and S7, along with corresponding PV parameters listed in the insets and Table S2.  In addition, enhanced long-term stability of PSCs is expected to be achieved for the 1D/3D mixed perovskite device. The hydrophobicity of pristine and modified perovskite films is certified by water contact angle measurement, as shown in the insets in Figure 4f. A larger water contact angle was displayed for modified film than for the control film, indicating the existence of 1D perovskite, which is beneficial against moisture and thus inhibited destruction of the lattice. We have tracked the PV performance of unencapsulated devices under ambient conditions with~45% RH for 1000 h (Figure 4f). The modified device exhibited a prolonged long-term stability while maintaining 85% of its original PCE. However, the efficiency of the control sample dramatically decreased to about 50% of its initial value after aging for1000 h.

Conclusions
In summary, we have successfully synthesized 1D/3D mixed-dimensional perovskite via a two-step method by simply introducing DMAI to the precursor solution of organic cation halide. The addition of DMAI regulated crystalline growth and reduced residual PbI 2 simultaneously. Meanwhile, the incorporated 1D perovskite, which tended to locate at the grain boundaries, could significantly suppress the formation of bulk defects and retard the invasion of moisture. Consequently, the optimized film properties of inhibited nonradiative recombination, improved charge carrier transfer, and enhanced long-term stability has been achieved. The corresponding devices exhibited a first-class PCE of 21.43% and maintained 85% of initial efficiency after aging under ambient conditions with~45% RH for 1000 h.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules27217566/s1, Figure S1: XRD patterns zoomed in (100) planes of perovskite films treated with different concentration of DMAI; Figure S2: The radially integrated intensity from GIWAXS data of corresponding perovskite film samples; Figure S3: Top-view SEM images of perovskite films with 5 mg/mL, 10 mg/mL and 20 mg/mL DMAI; Figure S4: UV-vis absorption spectra (a) and corresponding Tauc-Plots (b) of perovskite films with different concentration of DMAI; Figure S5: (a) SSPL spectra of perovskite films with different concentration of DMAI. (b) The tauc plot calculated from the PL results with and without 0.5 mg/mL addition of DMAI; Figure S6: TRPL spectra of perovskite films with different concentration of DMAI; Figure S7: Forward and reverse scanning J-V curves of the champion control and modified devices; Figure S8: Statistical distribution of photovoltaic parameters for PSCs incorporated with different concentration of DMAI. Table S1: Crystallographic parameters for DMAPbI 3 ; Table S2: Photovoltaic parameters of the champion control and modified PSCs under forward and reverse scans.