Multilayer evaporation of MAFAPbI_3−xCl_x for the fabrication of efficient and large-scale device perovskite solar cells

Hybrid organic–inorganic perovskites have attracted much interest due to their excellent properties such as high absorption, high carrier mobilities and versatile band gap tuneability. Thus, perovskites solar cells (PSCs) emerged as ideal candidates for high performance solar cells with a certified efficiency up to 23.3% [1–6]. In addition, PSCs can be fabricated with relative ease using evaporation and/or solution-based methods. These methods have the potential for upscalable and cost-effective photovoltaic modules.

Currently, the two dominant fabrication methods are vacuum techniques and processing from solution [7–12]. Thermal evaporation with one or two step process has the advantage of producing high quality, pinhole-free and uniform perovskite films that are solvent-free, substrate-independent, reproducible, scalable and use more purified precursors during the evaporation process (due to the heating of the precursor before the deposition) [13–18]. However, as often reported in literature, thermal evaporation of organic compounds such as the volatile methylammonium iodide (MAI) is challenging and the precise thickness-control is frequently an obstacle [19–25]. Previously, we proposed a new evaporation method using a layer-by-layer deposition technique for the fabrication of efficient PSC devices with highly crystalline perovskite films [26].

Formamidinium lead triiodide (FAPbI₃) has advantages for achieving highly efficient PSCs because FA is thermally more stable than MA and also has a more red-shifted band gap towards the Shockley–Queisser limit [27, 28]. However, this perovskite suffers from phase instability at room temperature showing a photoinactive 'yellow phase' instead of the desired 'black phase'. Compositional engineering is a key solution to tackle this problem. Adding either A-cations, such as MA and Cs, or X-halides, such as Cl and Br, the black phase of FAPbI₃ can be stabilized at room temperature by increasing the entropy of the FAPbI₃ system [29–31]. In this regard, Isikgor et al [32] employed PbCl₂ as a source of chlorine and fabricated mixed MA/FA cation perovskite using a solution method. Based on this composition, they improved the perovskite crystallinity, resulting in a PCE of 18.14%. As shown by Yu et al [33], the presence of extra ${\rm C}{{{\rm H}}_{3}}{\rm NH}_{3}^{+}$ in the composition can slow down the perovskite formation and improve the crystallinity during annealing and thus the role of Cl is to facilitate removing the extra ${\rm C}{{{\rm H}}_{3}}{\rm NH}_{3}^{+}$ at low annealing temperature. In other related works, Jiang et al [34, 35] reported a two-step solution process for fabrication PSCs and they added MACl to the perovskite composition in the second step and improved the crystallinity as well as PCE of device over 20%.

In this work, we demonstrate a layer-by-layer thermal evaporation technique for the deposition of FAPbI₃ perovskite. We show a stabilized black phase at room temperature by inserting a thin layer of MACl. We optimized the thickness of MACl and the annealing temperature for the perovskite film which has a grain size of up to 2 µm and very little surface roughness of 14 nm. Based on this modification, PSCs with PCEs of 17.7% and 15.9% were achieved for active areas of 0.1 and 0.8 cm², respectively.

Figure 1 shows the schematic of our fabrication process using the layer-by-layer thermal evaporation technique to produce phase-stable FAPbI₃ perovskite by inserting a thin MACl layer (figure 1(a)). We deposited each precursor, i.e. PbI₂ and FAI/MACl in ten sequential steps. Note that MACl was deposited in the middle of the multilayers. This technique enables the reaction of the individual precursors to form a high-quality perovskite film and also gives us more freedom to tune the perovskite composition by adjusting the amount and thickness of the interlayers [26]. The photograph of perovskite films on TiO₂-coated FTO glass with different scales is shown in figure S1 (stacks.iop.org/JPhysD/52/034005/mmedia). Figure 2 shows the characterization results of the optimized perovskite film deposited by this technique. The top-view scanning electron microscopy (SEM) image shows high quality and a large grain size up to 2 µm (figure 2(a)). Energy dispersive x-ray spectroscopy (EDAX) elemental mapping reveals a uniform distribution of all the chemical elements as well as the presence of trace amounts of Cl after annealing at the optimal 130 °C (figures S2 and S3). In order to further study the role of MACl, the SEM images of perovskite film fabricated by MACl on PbI₂ and FAI on PbI₂ after annealing are shown in figure S4. As seen, the perovskite with MACl shows larger grain size as compared to its counterpart. Figure 2(b) shows a top-view atomic force microscopy (AFM) image of the optimized perovskite film. The film has a roughness of 14  ±  5 nm which could stem from the presence of MACl during annealing. A 3D AFM image of the corresponding film depicted in figure S5 clearly confirms the surface roughness. Figure 2(c) shows the powder x-ray diffraction (pXRD) pattern with typical peaks at 14.2°, 28.5°, and 43.2° corresponding to (1 1 0), (2 2 0), and (3 3 0) lattice planes, respectively [36]. This indicates that the film has a preferential orientation along with the (1 1 0) direction. The optical UV absorption and photoluminescence (PL) spectra of the perovskite film are shown in figure 2(d) showing a band gap of 1.57 eV. In addition, the UV-visible spectra of the perovskite film without and with MACl were measured as shown in figure S6. The results indicate that there is a slight blue-shift in the spectrum of the perovskite film by adding MACl due to the presence of chlorine atoms in the perovskite composition.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The main optimization parameter in this study was the thickness of the MACl layer and the annealing temperature. In a first step, the MACl thickness was tuned from 40, 50, 60, 70 to 80 nm. Figure 3(a) shows the device efficiency as a function of the MACl thickness showing optimal device performance for a 60 nm-thick MACl layer. Higher thicknesses than 60 nm reduces the PCE gradually coinciding with the increased blue-shift in the absorption (that accordingly lowers the current). Additionally, we have optimized the amount of chlorine inside the composition by controlling the annealing temperature. As reported in the literature [12, 37], by increasing the annealing temperature beyond the typical 100 °C, the chlorine starts leaving the perovskite film in the form of MACl. Thus, by controlling the annealing temperature, we can tune the amount of the remaining Cl inside the perovskite composition. In this case, we changed the annealing temperature from 100, 110, 120, 130, 140 to 150 °C. Figure 3(b) shows the variation of PCE versus the annealing temperature. Our results show that 130 °C is the optimal annealing condition. Above 130 °C, the PCE is reduced slowly due to less amount of Cl in the lattice and poorer phase stability.

Zoom In Zoom Out Reset image size

In order to further study the quality of the perovskite films, PSCs with a small (0.1 cm²) and a large (0.8 cm²) area were fabricated. Figure 4(a) illustrates the cross-sectional view SEM image of the PSC revealing the FTO-glass, TiO₂ electron transporting layer (ETL), perovskite absorber, spiro-OMeTAD as a hole transporting layer (HTL), and the gold electrode. The current density–voltage (J–V) curves of the best performing PSCs with small and large active areas are shown in figure 4(b). The figure of merits for these devices under forward and reverse scans are summarized in table 1. The optimized small area PSC resulted in a short circuit current density of (J_sc) of 22.7 mA cm⁻², a V_oc of 1040 mV, a fill factor (FF) of 75%, and a PCE of 17.7% at reverse bias. The device with large area showed slightly lower photovoltaic parameters as compared to small area one. As seen, the device parameters were: J_sc of 21.94 mA cm⁻², a V_oc of 1020 mV, a FF of 71% and a PCE of 15.9% under reverse scan. This indicates the potential of the multilayer vacuum technique for large-scale device. In contrast, figure S7 shows the J–V curves of a MAPbI₃ PSC measured under reverse and forward scans. As can be seen, the PCE is 16.2% which is lower than the MAFAPbI_3−xCl_x perovskite device due to lower V_oc and FF. Moreover, the hysteresis indexes of small area and large area devices were calculated by the following formula: hysteresis index  =  (PCE_backward  −  PCE_forward)/PCE_forward)  ×  100 and obtained to be 2.31% and 2.92%, respectively. These values are smaller than the hysteresis index of the MAPbI₃ device with 4.51% (figure S7), maybe due to the larger grain size and better crystallinity. The statistical photovoltaic data for the corresponding devices are shown in figure S8. It can be observed that the average values of the photovoltaic parameters for large scale devices are lower than that found in small area devices. This shortening is not related to the quality of evaporated perovskite film because the performance of the large-scale device in our work has been limited by spin-coating process of TiO₂ ETL and spiro HTL.

Zoom In Zoom Out Reset image size

In addition, the external quantum efficiency (EQE) of the corresponding devices has been measured to further confirm the J_sc values as demonstrated in figure 4(c). The EQE response for large area PSC is slightly reduced compared to the small area PSC resulting overall in a lower J_sc, which is still in good agreement with the J–V measurements. The integrated current densities calculated from EQE curves were 21.75 mA cm⁻² and 20.91 mA cm⁻² for small and large area devices, respectively. Since the stability of PSCs is a big challenge for commercialization, we have tested the shelf life stability of our device as well. Figure 4(d) demonstrates the PCE, measured over 60 d, kept in dry box with 30% relative humidity condition. The result shows that the device retains ~97% of its initial PCE value after 60 d.

In summary, FAPbI₃ perovskite films have been fabricated through a layer-by-layer thermal evaporation technique and stabilized by inserting an optimum thickness of an MACl layer. After optimization of the annealing temperature, we obtained a perovskite film with high quality, large grain size up to 2 µm, and low roughness. Based on this fabrication technique and optimization processes, devices with PCEs of 17.7% and 15.9% were achieved for 0.1 and 0.8 cm² active areas, respectively. Our results indicate the potential of our fabrication process and optimized composition as an alternative for commercialization of PSCs.

Device fabrication

Film characterization

Device measurement

M M T wants to thank school of engineering at Hong Kong University of Science and Technology for their support.