Controlling growth of lead halide perovskites on organic semiconductor buffer layers

Organic–inorganic hybrid perovskite solar cells (PSCs) have attracted much attention because of their great power conversion efficiency (PCE). Perovskites are one of the desirable materials for the photoactive layer of solar cells in terms of bandgap, light-absorption coefficient, and carrier-diffusion length. ^1–3) The PCE reached 25% in 2021. ⁴⁾ The morphology of the perovskite can affect the charge collection efficiency that is related to PSC performance. ⁵⁾ In addition, its performance also depends on the orientation of the perovskite as previously reported. ⁶⁾ It is important to optimize the thin film growth of the perovskite to control the morphology and orientation. We have reported heteroepitaxial growth of CH₃NH₃PbI₃ (MAPbI₃) on a rubrene single crystal. ⁷⁾ However, previous reports used a stand-alone single crystal, which makes it difficult to fabricate PSC devices. A high-crystalline rubrene thin film is required to apply this crystallization method for device fabrication. There had been some technical difficulty in acquiring the high-crystalline rubrene thin film, ⁸⁾ but Haemori et al. reported the deposition of highly oriented rubrene on a pentacene buffer layer. ⁹⁾ Therefore, pentacene was used as the base layer material of rubrene in this study. Thus far, organic semiconductor buffer layers (OSBLs) have been incorporated in PSCs to improve their photovoltaic performance. ^10,11) OSBLs have various effects on PSCs. For example, they contribute to promoting good contact at the perovskite/electrode interface ¹²⁾ and realize good alignment of the HOMO level of OSBL to the VB edge of perovskite. ¹³⁾ In addition, Ryu et al. reported that a deep HOMO level of the hole transport layer improves the V_OC of PSC. ¹⁴⁾ Rubrene and pentacene can be used as hole transport layers of PSCs because of their high hole mobility (rubrene: 18 cm² V⁻¹ s⁻¹, pentacene: 1 cm² V⁻¹ s⁻¹). ^15,16) Yang et al. reported the PSC using pentacene as a buffer layer material. ¹⁷⁾ In their research, perovskite was deposited on a pentacene buffer layer by spin-coating, where the solvent (N,N-dimethylformamide, dimethyl sulfoxide) could give some damage to the pentacene underlayer, so we adopted a dry process as an alternative way to deposit MAPbI₃ on OSBL surfaces. In this paper, we used a rubrene/pentacene bilayer, pentacene layer, and rubrene layer as OSBLs in PSCs to compare the influences of crystallinity and morphology of perovskites grown on these OSBLs on the device performance. The fabrication of PSCs using rubrene as a buffer layer material was added to this paper newly from the SSDM2023 Extended Abstract. ¹⁸⁾

An ITO-coated glass substrate was treated in RF-generated oxygen plasma (3.5 × 10² Pa, 30 min). Poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS, Clevious, AI4083), which acts as a conductive organic polymer, ¹⁹⁾ was spin-coated on the substrate at 500 rpm for 5 s and 3000 rpm for 180 s. Then, the film was annealed at 135 °C for 10 min. The sample was transferred into the laser evaporation chamber for organic materials where an 808 nm continuous-wave (CW) semiconductor laser was used to evaporate the materials. A pentacene layer (35 nm) was deposited on the PEDOT:PSS by laser evaporation while controlling the substrate temperature at 55 °C. A rubrene layer (15 nm) was deposited on pentacene by laser evaporation with the same substrate temperature (55 °C). Then the sample was transferred into a different laser evaporation chamber (808 nm CW laser) that enables alternative evaporation of PbI₂ and CH₃NH₃I (MAI) (Fig. 1). A PbI₂ layer was initially deposited followed by a MAI layer and this process was repeated to form MAPbI₃ thin films. Silicon powder was mixed into the MAI and PbI₂ target to absorb the laser energy and promote heating of the source materials. ²⁰⁾ The laser for MAI was modulated as a 10 Hz square wave to control the deposition rate of MAI accurately by tuning the duty ratio of the square wave. ²⁰⁾ The sample was transferred into the laser evaporation chamber for organic materials again and a C₆₀ layer (7 nm) was deposited on the MAPbI₃ without heating the substrate. C₆₀ is one of the great electron transport materials because of its high electron mobility (1.6 cm² V⁻¹ s⁻¹). ²¹⁾ A bathocuproine (BCP) layer (3 nm) was deposited on the C₆₀ by laser evaporation without heating the substrate. Ag electrodes (100 nm) were deposited on the BCP by thermal evaporation without heating the substrate. BCP can suppress the quenching of exciton at the C₆₀/Ag interface. ²²⁾ The structures of PSCs prepared in this study are shown in Fig. 2. Crystallinity of the samples was analyzed by X-ray diffraction (XRD, RIGAKU Smart Lab, Cu Kα1 X-ray source). The surface morphologies of the thin films were observed by atomic force microscopy (AFM; Nanonavi/E-sweep, SII nanotechnology).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Figure 3 shows the out-of-plane XRD pattern of pentacene film deposited on amorphous PEDOT:PSS to confirm the orientation of pentacene film. The (001) diffraction peak of pentacene (thin film phase) ²³⁾ was observed. Out-of-plane XRD patterns of rubrene films deposited on pentacene and PEDOT:PSS are shown in Fig. 4 to demonstrate the importance of the pentacene underlayer. The (002) diffraction peak of rubrene (orthorhombic phase ²⁴⁾) was observed only for the film deposited on pentacene. The result of the previous research ⁹⁾ was confirmed, where the out-of-plane orientation of rubrene can be obtained by growing on pentacene. AFM images of rubrene films deposited on pentacene and PEDOT:PSS are shown in Figs. S1(a) and S1(b), respectively, where clear step and terrace structures are seen in the rubrene thin film deposited on pentacene. These results are consistent with Fig. 4. Figure 5 shows AFM images of MAPbI₃ thin films (∼30 nm) to investigate the effects of OSBL on MAPbI₃ film structure. MAPbI₃ fabricated on rubrene/pentacene [Fig. 5(a)] and pentacene [Fig. 5(b)] have a smoother surface than that grown on rubrene [Fig. 5(c)], where root mean square values are 1.23 nm, 0.94 nm, and 2.03 nm, respectively. We attributed these results to the difference in crystallinity. As seen in Figs. 3 and 4, the rubrene/pentacene bilayer and pentacene layer have out-of-plane orientation, while the rubrene layer is amorphous. Thus, highly oriented OSBLs are enhanced to grow smooth MAPbI₃ film. Figure S1(d) shows the AFM image of perovskite film on PEDOT:PSS. The perovskite film deposited on PEDOT:PSS contains particle-shaped morphology with a size of 100 nm. The perovskite film deposited on the rubrene/pentacene bilayer shows smoother morphology because of the highly crystalline characteristics of the underlayer. For comparing OSBLs in terms of device performance, J–V curves of PSCs and their performance parameters are shown in Fig. 6 and Table I, respectively. The device using a rubrene/pentacene bilayer as an OSBL exhibited the highest PCE (8.50%) among (a)–(c). The energy level diagrams of PSCs are shown in Fig. 7 following previous reports ^1,17,25,26) to discuss the effect of energy level on V_OC. The HOMO level of the HTL adjacent to the MAPbI₃ layer is deeper in the PSC with rubrene/pentacene than that with only pentacene. This difference can lead to the improvement of V_OC (0.61 V→0.89 V) comparing Figs. 6(a) and 6(b). The high crystallinity and reduced voids comparing Figs. 6(a) and 6(c) result in the improvement of J_SC (12.40 mA cm⁻²→16.64 mA cm⁻²). The voids can decline J_SC by decreasing the area that the photocurrent passes. The slight difference in J_SC between Figs. 6(a) and 6(b) appears for the same reason observed in Figs. 5(a) and S1(b). The EQE spectra shown in Fig. S3 also suggest that the PSC with only rubrene OSBL has the least J_SC in three types of devices. Furthermore, in Fig. 6(c), non-linear resistance appears near V_OC. It suggests that carrier transport at the interface between the perovskite and HTL is inhibited. Figure 8 shows the out-of-plane (a) and in-plane (b) XRD patterns of MAPbI₃ thin film (∼30 nm) deposited on the rubrene/pentacene bilayer to confirm the orientation of MAPbI₃ film. These patterns suggest that the MAPbI₃ film is a tetragonal structure. ^27,28) Only (002) diffraction peaks of MAPbI₃ were detected in Fig. 8(a). This result supports that the MAPbI₃ film is highly controlled to form an out-of-plane orientation. In Fig. 8(b), the peak marked with an asterisk was assigned to the (310) diffraction peak of MAPbI₃ according to previous research. ^29,30) Because the in-plane orientation of the rubrene thin film in this research is not controlled, that of the MAPbI₃ film is random. Incidentally, Fig. 9 shows the predicted structural model of MAPbI₃ film on rubrene and 〈110〉 and 〈002〉 plane directions are depicted by black arrows.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In this study, we used three types of OSBLs as underlayers of MAPbI₃ films, (a) rubrene/pentacene bilayer, (b) pentacene layer, and (c) rubrene layer and investigated the crystal orientation of each layer to compare their PSC performance. A MAPbI₃ thin film grown on (a) achieved pronounced out-of-plane orientation and smooth surface morphology. The difference in HOMO level between rubrene and pentacene led to higher V_OC of the PSC using (a) than (b). The PSC using (c) showed a lower J_SC than (a). because there are a lot of voids in the MAPbI₃ film grown on (c). Therefore, among (a)–(c), the greatest OSBL is (a), the rubrene/pentacene bilayer, which achieves improvement in crystallinity and highly efficient charge transport. The importance of a buffer layer in the PSC was verified in terms of both energy level alignment and crystallinity control of the perovskite film. We hope that this research will help to choose the new buffer layer material of the PSC that contributes to improving the crystallinity of the perovskite.

This work was supported by JST SPRING, Grant No. MJSP2123 and JSPS KAKENHI, Grant Nos. 20H02819, 20H02838, and 22H01946.