Enhancing Surface Modification and Carrier Extraction in Inverted Perovskite Solar Cells via Self-Assembled Monolayers

1. Introduction

The perovskite solar cells (PSCs) have experienced a significant increase in efficiency, from 3.8 % in 2009 to 26.1 % in the present day.[1,2] These solar cells are typically composed of a perovskite light-absorbing layer sandwiched between an n-type electron transport layer (ETL) and a p-type hole transport layer (HTL).[3,4,5,6] Upon the incidence of light on the solar cell, an exciton is generated within the active layer, which subsequently separates into electrons and holes. These free carriers are then transported through the ETL and HTL, respectively, to the external electrode.[7] As for HTLs, molecular layers based on organic or conductive polymers have been widely studied to enhance the efficiency of PSCs.[8] However, these types of HTL layers have notable limitations in terms of reliability due to their environmental sensitivity.[9,10,11] As a result, there is an urgent need to explore alternative HTL materials that are not only efficient and stable, but also offer potential scalability and a reduction in overall manufacturing costs. Inorganic compounds with energy levels suitable for PSCs have been investigated. Nickel oxide (NiO), an inorganic HTL, exhibits semiconductor properties, including excellent stability, high hole mobility, a wide band gap (~3.5 eV), high optical transparency, and a solution process coating method, demonstrating its potential for use in PSCs.[12,13,14] However, a reduction in efficiency due to surface recombination, traps, and defects remains a challenge.[15] To address these issues, research on passivation[16,17], additives[18,19,20], and buffer layers[21,22] is being undertaken. Recently, several efforts have been made to reduce defects and enhance the performance of devices using self-assembled monolayers (SAM).[23] A SAM molecule is made up of three parts: a head group, a linkage tail and a functional group.[24] The head group attaches directly to the substrate and helps stabilize the surface, using various substances such as silane, carbonyl acid, phosphoric acid, and thiol.[25] The linkage tail connects the head group and functional group, and it affects the optical properties and performance, depending on its molecular size and tilt angle.[26] The functional group, which is located on the surface that is exposed, can alter the properties of the surface based on the material used.

Herein, we fabricated inverted p-i-n PSCs using HC-A1 and HC-A4 molecules. These multi-functional molecules, previously employed in dye-sensitized solar cells, exhibited excellent stability and offered the potential to control charge recombination.[27,28] The impact of introducing an additional layer of HC-A1 and HC-A4 between the perovskite and the transporting layer was investigated, both structurally and optically, using various techniques. The application of HC-A1 and HC-A4 resulted in an improvement in the grain morphology and uniformity of the perovskite active layer. In particular, HC-A4 led to significant surface modification and efficient carrier extraction with its pronounced hydrophilic properties. As a result, HC-A4-applied solar cells exhibited the highest power conversion efficiency (PCE) of 20%.

2. Materials and Methods

2.1. Materials

Indium-doped tin oxide (ITO) substrates (15Ω/sq) purchased from Nippon Sheet Glass. N,N-dimethylformamide (DMF; 99.8%), dimethyl sulfoxide (DMSO; ≥99.9%), and lead(II) bromide (PbBr; 99.9%) purchased from Sigma-Aldrich. Lead(II) iodide (PbI; >98.0%) and cesium iodide (CsI; >99.0%) purchased from TCI. Formamidinium iodide (FAI), methylammonium bromide (MABr) purchased from GreatCell Solar. Bathocuproine (BCP; 99.5+%) and fullerene (C₆₀) purchased from Nano-C, and nanoparticles (NiCT-7, 2.5wt% NiO in Ethanol) purchased from Nano Clean Tech. SAM materials (HC-A1 and HC-A4) were synthesized following previous work [27,28], and ethanol (>99.9%) and tetrahydrofuran (THF; 99%) purchased from EMSURE® (Merck) and JUNSEI, respectively.

2.2. Device Fabrication

ITO/glass substrates were sequentially cleaned with deionized (DI) water, Hellmanex, acetone, ethyl alcohol, and isopropanol for 20 min and then dried using N₂ gas. Any remaining organic residuals were then eliminated through a UV ozone surface treatment for 30 min. The NiO nanoparticles was stirred for 30 min, and 50 μl of the filtered NiO nanoparticle solution was spin-coated onto the ITO substrate at 4000 rpm for 40 s, and then annealed at 350 ℃ for 30 min. The CsFAMAPb(IBr)₃ precursor solution was then prepared by mixing 568 mg of PbI₂, 79.8 mg of PbBr₂, 187 mg of FAI, 12.2 mg of MABr, 66 mg of CsI, and 220 mg of DMSO with 755 mg of DMF solution and stirring for 1 h. The precursor solution was then spin-coated in two steps: 500 rpm for 5 s and 5000 rpm for 45 s, with 300 μl of chlorobenzene dropped as an anti-solvent after 15s into the second step. The resulting perovskite film was annealed at 100 ℃ for 45 min. All spin-coating processes were conducted within a glove box. Subsequently, a C₆₀ layer, electron transporting layer (ETL), was deposited by thermal evaporation under high vacuum (10^-7 Torr) conditions, achieving a thickness of 15 nm at a deposition rate of 0.5 Å/s. A BCP solution containing 6 mg of BCP in ethyl alcohol was then spin-coated at 4000 rpm for 30 s and dried at room temperature in an N₂ atmosphere. Finally, 100 nm Ag electrodes were deposited by thermal evaporation under high vacuum (10^-7 Torr) conditions, with thicknesses of 10 nm and 90 nm sequentially deposited at rates of 0.3 Å/sec and 1.0 Å/sec, respectively.

2.3. Device Characterization

The perovskite structure was analyzed using a high-resolution X-ray diffractrometer (HR-XRD; Empyrean) The absorption and transmittance spectra were obtained using a spectrophotometer (UV-Vis; Hitachi, UH4150). The top-view and cross-sectional scanning electron microscopy (SEM) images were measured by a field-emission scanning electron microscope (FE-SEM; Hitachi, SU8220). Photoluminescence (PL) spectroscopy was performed using a confocal microscope (NANOBASE) equipped with a 532 nm laser (100 nW), and PL spectra and mapping images were collected using a CCD detector. Time-resolved PL spectra were obtained using a 405 nm pulsed laser with a repetition rate of 32 MHz with a 10 s integration time, and an excitation power of 21.3 nW was maintained to prevent light-induced damage to the perovskite sample. Atomic force microscope (AFM) images were measured using an AFM system (NX-10; Park systems) with a non-contact mode cantilever (PPP-NCHR). The scan rate was controlled as 0.5 Hz. The current-voltage (J-V) characteristics were measured under an AM 1.5G solar illumination (100 mW/cm²) using a 450 W Xenon arc lamp and a class AAA solar simulator (Newport) equipped with Keithley 4200S-SCS. Additionally, external quantum efficiency (EQE) measurements were conducted with a 150 W Xenon arc lamp and IQE200 (Newport) across various wavelengths.

3. Results and Discussion

3.1. Fabrication and Characterization of SAM

Figure 1 illustrates the molecular structure of HC-A1 and HC-A4, along with a detailed fabrication process. The Y-shaped structure consists of a carboxylic acid acceptor linked to porphyrin dye (HC-A1) and naphthalene (HC-A4), respectively. The various functional groups of SAM molecules can facilitate charge extraction characteristics between the perovskite and the transporting layer. Solutions of 1 mmol HC-A1 (6 ml ethanol, 4 ml THF) and 1 mmol HC-A4 (10 ml ethanol) were stirred for 1 h, and the NiO nanoparticle-coated substrate was immersed in the SAM solution for 2 h. The SAM-modified substrate was then rinsed in ethanol for 10 s to remove any residual unbounded molecules. Following drying with N₂ gas, the final substrate was annealed at 100 ℃ for 10 min.

Figure 2a,b show the transmittance and absorption spectra of ITO/NiO/SAM according to the application of either HC-A1 or HC-A4. Since the additional SAM layer existed on the ITO/NiO substrate, a high transmittance of 60~80% was obtained in the visible range. The optical band gap was determined to be 4.04 eV, both with and without the SAM layer. Subsequently, we explored the characteristics of the perovskite layer with SAM modifications. Figure 2c,d present the XRD and absorption spectra of ITO/NiO/CFM, modified with the inclusion of either HC-A1 or HC-A4 between NiO and CFM perovskite. In the absence of a SAM layer, the perovskite primarily exhibited diffraction peaks at 14.1° and 20.0°, corresponding to the (001) and (011) planes of CFM. The introduction of SAM layers resulted in a slight shift of the entire diffraction peaks towards lower angles. Although HC-A1 and HC-A4 molecules are not integrated into the perovskite structure, the diffraction changes can be understood as strain changes, suggesting a slight expansion of the perovskite unit cell due to the addition of the SAM layer below.[29] Consistent with this observation, the optical band gap of CFM perovskite was confirmed at 790 nm (1.57 eV), without noticeable changes following SAM modification.

3.2. Surface Modification via SAM Molecules

The CFM core structures remained unaffected by the presence of HC-A1 and HC-A4 molecules; however, their grain morphology improved with the addition of SAM layers. Figure 3 presents the top-view and cross-sectional SEM images of ITO/NiO/CFM structures, both with and without SAM layer. The thickness of CFM perovskite was 500 nm across all films. Interestingly, both HC-A1 and HC-A4 applied perovskite films exhibited significant improvement in the vertical CFM grain structure. Additionally, when comparing the control and the HC-A4 samples, an improvement in the grain size on the surface was also observed. We hypothesize that SAM layers, consisting of different functional groups, may alter the surface properties of NiO.

Figure 4a,b show the surface difference between HC-A1 and HC-A4 molecules, as evidenced by the contact angle and AFM images. The SAM molecules were applied on the top of the ITO/NiO structure. The contact angles for HC-A1 and HC-A4 were confirmed to be 87.5° and 50.3°, respectively, placing both molecules within the hydrophilic region. However, HC-A4 demonstrated more pronounced hydrophilic properties, attributable to its functional groups. The root-mean-square (RMS) values for the ITO/NiO/SAM structure were also measured, revealing a more uniform surface for HC-A4 compared to HC-A1. The RMS values were determined to be 2.235 nm for HC-A1 and 1.155 nm for HC-A4. Furthermore, we observed a hollow-like CFM surface on the HC-A1 layer through top-view SEM images with low magnification. This observation implies that the perovskite film does not form completely due to the near hydrophobic-like properties of HC-A1. The uniformity of the perovskite film is a critical factor in determining optical properties and device performance; thus, we anticipated that HC-A4 could outperform HC-A1 in all aspects.

3.3. Carrier Extraction via SAM Molecules

We measured PL and TRPL spectra of the ITO/NiO/SAM/CFM structures as shown in Figure 5a,b. The PL intensity was remarkably increased with SAM molecules attributable to the improved surface and vertical grain morphology. Consistent with this observation, the PL peak position shifted from 761.66 nm to 755.76 nm following SAM modification, suggesting enhanced crystallization and passivated defect level. The TRPL spectra, presented in Figure 5b, were fitted using a tri-exponential function, and the average lifetimes (τ_ave) were calculated and listed on Table 1. In comparison to the control sample (73.81 ns), both the HC-A1 and HC-A4 samples demonstrated improved lifetimes of 201.11 ns and 117.33 ns, respectively. This improvement is associated with the quality of perovskite layer in conjunction with the SAM layer. However, it is worth noting that the HC-A4 sample exhibited a faster lifetime than the HC-A1 sample. A comparison of the TRPL properties with the HTL-including structure revealed the hole extraction ability of each sample. In other words, HC-A4 transported holes from the perovskite to the NiO layer more efficiently than HC-A1. Moreover, a missing part of PL was observed for the HC-A1 sample in Figure 5c, which is attributed to the hollow-like CFM surface depicted in Figure 4c. Otherwise, the HC-A4 modified film displayed more uniform PL properties than the other samples at the same scale.

3.4. Device Characterization with SAM Molecules

Finally, we fabricated a full p-i-n device structure of ITO/NiO/SAM/CFM/C₆₀/BCP/Ag, as depicted in Figure 6a. The energy level diagram for each layer, as illustrated in Figure 6b, shows that the functional group of SAM molecules can facilitate its energy level. The J–V and EQE curves were measured in Figure 6c,d, and the device parameters were summarized in Table 2. The control device without SAM molecules exhibited a V_OC of 1.02 V, J_SC of 23.13 mA/cm², FF of 79.81%, and PCE of 18.81%. Due to its hollow-surface, HC-A1 showed the lowest efficiency of 13.04% with poor device J_SC and FF values. Conversely, HC-A4 device demonstrated enhanced device parameters of 1.05 V, 23.34 mA/cm², 81.80%, and 20.00%, along with the highest integrated J_SC of 19.88 mA/cm². The uniform surface modification and the rapid charge extraction by the additional HC-A4 layer contribute positively to the device performance. Furthermore, the favorable band alignment of HC-A4 between CFM and NiO, as shown in Figure 6b, can promote hole extraction, thereby enhancing the overall efficiency of the device.

4. Conclusions

In this study, we successfully fabricated a p-i-n perovskite device structure utilizing HC-A1 and HC-A4 molecules. The application of these SAM molecules improved the grain morphology and uniformity of the perovskite film, which are critical factors in determining optical properties and device performance. Notably, HC-A4 outperformed HC-A1 due to its pronounced hydrophilic properties with a contact angle of 50.3°, attributable to its distinct functional groups. Overall, the HC-A4 applied film showed efficient carrier extraction properties, achieving a carrier lifetime of 117.33 ns, and recorded the highest PCE of 20%. Our study presented HC-A4 as a promising candidate for enhancing device performance, thereby expanding the scope of research on the utilizing of SAM molecules in PSCs.