Buried-in interface with two-terminal functional groups for perovskite-based photovoltaic solar cells

Interface plays an important role in perovskite solar cells. Herein, a functional molecular with two-terminal donor groups was deposited between the tin oxide (SnO2) electron transport layer and halide perovskite to induce the perovskite crystal growth and passivate defects at the interface. It is found that isonicotinohydrazide (INHA) can anchor Pb2+ cluster in precursor and promote uniform nucleation, which helps to adjust the crystal growth in perovskite films. As well, more analysis shows that interfacial modification can greatly reduce trap defects and therefore facilitate photogenerated carrier-transferring. The efficient electron transfer and reduced interface traps correlate well with the corresponding fill factors and open-circuit voltages (VOC) of working devices. The resulting perovskite solar cell exhibits striking improvements to reach the champion efficiency of 21.12%. The long-term stability is also significantly enhanced by comparing to the pristine devices. This work highlights the effect of INHA/perovskite interfacial interaction and provides a multi-functional passivation strategy for further perfecting perovskite films.


Introduction
Halide lead perovskites can be seen as some of the most promising photovoltaic materials owing to their enormous potential for solar-to-electricity conversion [1]. The performance of perovskite photovoltaic device has achieved a striking power conversion efficiency (PCE) of 25.7% within a decade [2], comparable to silicon solar cell (~ 26.7%) and CuInGaSn (CIGS, ~ 23.4%) thin-film photovoltaic devices [3]. The excellent PCE is brought about by the merits of perovskite physique per se, such as high symmetry crystalline, large absorption coefficient across the visible spectrum, long carrier lifetime, and high tolerance to defects [4][5][6][7][8]. The researchers make sustained effort aim to attain the highest efficiency close to theory limitation [9] and finally realize the commercialization of perovskite solar cells (PSC).
The perovskites are represented as a prototype of ABX 3 , where A cation is a methylammonium (MA + ), formamidinium (FA + ), or Cs + ; B denotes as metal cations (Ge 2+ , Sn 2+ , Pb 2+ , etc.), and X denotes as halide (Br -, I -, or Cl -) [10]. One of the most intrinsic features distinguishing the perovskite from other photovoltaic materials is its ioniccovalent bonding in perovskite crystal, where the A cations fill-in the void between [PbX 6 ] 4octahedra to build a three-dimensional network [11,12]. It is well known that the optoelectronic quality of photovoltaic semiconductors is affected by lattice relaxation and disorder caused by crystal growth or additives [13][14][15]. A large number of studies have been reported on improving the growth of perovskite crystals, such as component engineering [16], solvent engineering [17], interface passivation engineering [18], and more. Most researchers addressed that the ionic nature of perovskite crystal makes it vulnerable to interface defects in terms of chemical interactions with functional groups. Therefore, the interfacial chemical interactions are crucial factors in the regulation of perovskite optoelectronic performance [19]. For instance, Zhang group reported self-assembled monolayers (amphiphilic molecule) between electron transport layer (ETL) and perovskite absorber, evidencing energy level alignment and reduction of trap density due to interfacial reaction [20]. Levine et al. advanced the carrier propagating process at buried interfaces of perovskite and selective contacts layers [21]. They conformed the interfacial chemical structure dominated the carrier transfer and density of interface traps, providing the electronic kinetic model to understand the energy loss in perovskite solar device. Additionally, Wang et al. proceeded with a mixture of a functional group (acceptor group) into perovskite precursor to prepare an interface passivation on top of perovskite surface [22]. Dai's group also demonstrated the "molecule glue" with an I-rich donor group to attain the high interfacial toughness [23]. Jiang, Xiong, and Shin groups summarized multiple moleculars with donor or acceptor groups between the charge transport layer and perovskite, guiding us for further optimizing performances of perovskite solar devices [24][25][26]. The previous studies discussed various functional molecules on perovskite interfacial passivation. [27][28][29][30][31][32][33]. However, the key chain of evidence is equivocal-the origin of molecular/perovskite interface formation and how the thin interfacial layer affect the crystal growth and passivation of perovskite traps. SnO 2 is acknowledged to be a functional semiconductor material with excellent performance [34,35]. In this paper, we demonstrated the main process of perovskite crystal growth based on SnO 2 /INHA substrate and discuss the influence of INHA/perovskite interface on perovskite growth and photovoltaic performance of devices. Here, a functional molecular (isonicotinohydrazide) with twoterminal donor-groups was planted on the SnO 2 substrate to form a thin passivated layer. On the one hand, INHA molecules are anchored on the surface of SnO 2 through donor groups, reducing the oxygen vacancy defects of SnO 2 , and increasing the wettability of perovskite precursor on SnO 2 . [36][37][38] On the other hand, the donor group on the other side forms a coordination bond with Pb 2+ in perovskite, which improves the crystallization process of perovskite. These coordinated bonds promoted a uniform crystal growth in the whole perovskite film. The functional INHA works efficiently for perovskite solar device, improving the PCE from 18.17% to 21.12% significantly. We further provide the defect state based on DFT calculation, confirming the reduction of traps at the INHA/ perovskite interface. This work provides a mentality for further advancing the development of interface passivation and should trigger interest in searching for further functional molecular leading to photovoltaic device with precisely controlled crystal growth.

Results and discussion
First of all, we deposited INHA thin-film on SnO 2 electron transport layer. As seen in Fig. 1a, the two donor groups (= N − and − CO −) are located at either end due to the steric effect of pyridine group, offering a typical interfacial framework to passivate the perovskite interface. With INHAbased SnO 2 at the interface, the donor group is attached with SnO 2 on top of substrate surface. As well, the other donor strongly coordinates with Pb 2+ cluster in perovskite precursor, constructing an adhesive interface for perovskite casting in Fig. 1c and e. As the FTIR showed in Figs. 1b and S1, the functional groups are identified as N-N-C with vibration peaks at 1260.5 cm -1 , C − N with stretching vibration peak at 1289.8 cm −1 , and − C = O with peaks at 1607.1 cm -1 , respectively. The aggregates of the INHA-based film are subjected to intermolecular hydrogen bonding and function of steric effect. We propose the intermolecular hydrogen bonding (− CO … HN −) aligns the thin-film in order, combining with the steric effect of large pyridyl group to assemble the interfacial frame on the SnO 2 surface. On the other hand, the two-terminal donor groups are random planted on the SnO 2 substrate in terms of similar coordinated ability thereof. As seen in Fig. 1d, a smooth and even SnO 2 film is presented in atomic force microscopy (AFM) images. In contrast, with INHA on top of SnO 2 , there lay a map of undulating surface in Fig. 1e. The morphology with high roughness conveys that the INHA is successfully located on top of the surface. As confirmed by the XPS spectra in Fig. S2b, the noticeable N 1p peak on the SnO 2 film also demonstrates that the INHA is deposited on the surfaces of SnO 2 substrate. To evaluate the water-resistivity change of the INHA covered films, we monitored the contact angles of the INHA layers on SnO 2 films. The control SnO 2 film is visualized with a contact angle of 24.4° in Fig. 1c. By comparison, the contact angle of the INHA covered film is as low as 15.7° (Fig. 1c, below), which elicits the increment of wettability and the coordination potency of INHA upon perovskite film. Considering the dielectric nature of the functional molecular, there is a concern about the conductivity of electric dipole INHA/ perovskite interface [39][40][41]. In Fig. 2f, the conductivity of INHA modified SnO 2 layer is negligible loss, because the amount of INHA on surface is marginal and thus offers a better ohmic contact at the interface. These results suggest that the INHA has a considerable effect on the SnO 2 substrate, which is expected to promote the crystallization of the perovskite films.
FA-based perovskite films (FA 0.8 MA 0.15 Cs 0.05 Pb(I 0.87 Br 0.13 ) 3 ) were subsequently deposited by antisolventassisted spin coating approach. The impact of INHA on the morphological properties of the perovskite films was examined in this section. Figure 2a shows the scanning electron microscope (SEM) images of the prepared perovskite films from pristine and INHA modified SnO 2 . The INHA modified perovskite film presents a dense and large crystal with average grain size of ~ 565 nm, which is more homogeneous than the pristine film with average size of ~ 375 nm (Figs. 2a and S3). The perovskite crystal is enlarged with INHA incorporation, which can be owned to the coordination interaction at the INHA/perovskite interface to enable the regulation of crystal growth. The distinctly different crystallization behavior of the perovskite films processed from INHA passivation was identified by atomic force microscopy (AFM) images in Fig. S4. Apparently, the perovskite film based on INHA-modified SnO 2 exhibits a uniform and smooth surface (RMS = 7.63 nm), whereas the film deposited on pristine SnO 2 yields a rough surface (RMS = 17.14 nm) with many protuberances of small grains. We further calculated Fermi level and valence band of SnO 2 and SnO 2 / INHA ETLs extracted from ultraviolet photoelectron spectroscopy (UPS) measurement. The pristine SnO 2 (~ 40 nm) was deposited by spin coating SnO 2 colloid at low temperature following previous reports. The cooled ITO/SnO 2 films were soaked into 0.1 M INHA aqueous solution for 10 min followed by washing with deionized water and blow-drying.
As shown in Fig. 2b and c, the secondary electron cutoffs for the INHA modified SnO 2 is 16.91 eV. The corresponding work functions calculated by subtracting secondary electron cutoffs from the excitation energy for INHA modified SnO 2 is 4.35 eV. The Fermi level of INHA modified SnO 2 ETL is 4.35 eV from the onsets of secondary electron cutoff region, which yields a higher bandgap of 4.26 eV. With INHA modification, the conduction band level of SnO 2 is optimally matched well with band level of perovskite absorber layer, which can facilitate charge transfer at SnO 2 / perovskite interface in solar device. Figure S5 shows the UV-vis absorption spectra of perovskite films with and without INHA modification. There presents a similar absorption slope at ~ 780 nm in both perovskite films. To distinguish the interfacial passivation effect on perovskite photoelectric property, more measurements were applied to monitor the subtle differentiation of perovskite films on INHA-modified substrate. As seen in Fig. 2d, the steady-state photoluminescence (PL) shows a PL peak at 760 nm, which presents a stronger PL intensity on INHA modified film compared to the control. The improved emission intensity suggests the reduction of the interfacial recombination owned to interface defects [42]. We further activate the perovskite film from back side (INHA/perovskite side) in Fig. 2e. It is noticed that there is no significant difference  Fig. 3b. The external quantum efficiency (EQE) spectra give a J int of 22.05 mA/cm 2 , which is in agree with the J SC value from J-V curves (Fig. 3c). In addition, the INHA-modified device experiences a negligible current decline during continuous illumination over 100 s (AM 1.5G), suggesting good light and thermal stability of perovskite working device. The steady-state output efficiency of the champion perovskite solar device shows a non-degraded PCE of ~ 20.61% under maximum power points (MPP) under a bias value of 0.97 V in Fig. 3d. Furthermore, the average PCEs of the INHAmodified cells are boosted significantly from 17.51% to 19.68%. The average PCEs of the FA-based solar cells is also addressed by analyzing the reproducibility of J-V values from 30 perovskite devices each in Fig. S7.
Additionally, the optoelectronic properties of FA-based perovskite film were measured to address how the INHApassivation works on the enhanced photovoltaic performance. We characterized light intensity-dependent V OC to evaluate the charge recombination of perovskite devices. Generally, the trap-assisted non-radiative recombination is critical factor to reduce the performance of solar devices [43,44]. Figure 3e Equation S1). As seen in Fig. 3e, the V OC dependent on the light intensity displaying a slope with 1.30 K b T/q lower than the pristine of 1.70 K b T/q, which can suppress nonradiative recombination at the interface significantly. We also evaluated the Nyquist plots collected from EIS values in Fig. 3f. The series resistance (R S ) in optimized device decreases from 40.6 Ω to 21.7 Ω, which promote the photon-generated carriers transfer. Taken together, these results explain the INHA-passivation toward the enhancement of optoelectronic qualities, consistent with its enhanced crystal quality by XRD results (Fig. S5), accounting for the improved device performance.
It should be noted that there is no consensus to explain the perovskite growth induced crystal mismatch until to now [45]. In this regard, we offer a proposal of INHA-assisted perovskite growth depending on our experimental results. In our work, one-step spin-coating approach was employed to deposit a perovskite thin-film in Fig. 4a. The spin-coating provides a high uniformity and relatively thin-films required for effective photovoltaic devices. Once the perovskite precursor was casted on the substrate, the absorption occurs at the substrate/solution interface [46]. Subsequently, the rotation of the substrate at a certain speed lay the centrifugal force combined with surface tension of the solution, resulting in an even perovskite liquid film with a thickness of ~ 1 µm (Fig. 4a, process II). In practical perovskite-film deposition, we employ antisolvent process to extract the solvent (DMF, DMSO, etc.) to manipulate the growth of perovskite crystal (process-III). Under antisolvent approaching, the solution oversaturation at the perovskite/air surface produces the nuclei and triggers the perovskite crystal growth in Fig. 4b. The effect of antisolvent is of significance, with a consequence of the crystal growth orientation from top to bottom due to the nuclei distribution. The top-to-bottom growth orientation causes a depletion region beneath the perovskite/air surface, resulting in mismatched crystalline at the SnO 2 /perovskite interface (Figs. 4b and 2f). This behavior is of importance for thinfilm deposition, especially for 2D-perovskite with phase segregation in vertical direction [47,48]. In this present work, the INHA/perovskite interface alleviates the uneven distribution of nuclei by anchoring Pb 2+ cluster at the bottom (SnO 2 /perovskite side), which is expected to promote the perovskite crystal growth and passivate the interface charge traps (Fig. 4c).
As aforementioned, the nuclei are the sine qua non of perovskite crystal; therefore, their distribution dominates the growth of perovskite films and interfacial trap state [49,50]. where N is number of nuclei, t denotes as time, A is preexponential factor, k B is Boltzmann's constant, and T is temperature. It can be found that the best way to increase the crystal growth and quality of perovskite films is to manipulate the nuclei (N) under a given ΔG. Figure 5a shows a vivid illustration of the substrate interface with absorption of perovskite cluster. Once the nuclei are populated at the perovskite/air interface due to antisolvent treatment, the primary nuclei spontaneously exhaust the surrounding compound, leading to the formation of depletion region beneath the perovskite/air interface. This process occurs less than one second and determines the final perovskite's performance. The presence of INHA is critical that their terminal-groups anchor perovskite cluster to form the absorption layer with a certain concertation of perovskite nuclei. The absorbed layer counters against ingredient deficient within depletion region (top-crust effect), which offers a buffer layer for perovskite crystal growth. The INHA/perovskite interface is expected to heal the crystalline mismatch for efficient charge separation at the interface [51,52]. (1) In interpreting the reduction of trap density at the INHA/ perovskite interface, we further characterized the charge defect density state and carrier transfer behavior by TRPL, DFT, and quantitative trap state. The time-resolved PL (TRPL) in Fig. 5b shows that the INHA passivated film displays a longer average carrier lifetime (ave ~ 57.13 µs) than the control INHAple (ave ~ 23.86 µs). The high carrier lifetime at the INHA/perovskite interface is in favor of the charge transfer and effective suppression of defectsinduced recombination at the interface, contributing to higher efficiency and V OC values. Density functional theory (DFT) calculations were also performed to study the coordination provided by the interfacial bonding between the perovskite surface and INHA substrate, where alpha-FAPbI 3 represents FA-based perovskite in Fig. 5c. The interfacial bonding with INHA has the characteristics of the halogen bond, where a high electron density region on a donor group (mainly − CO) forms attractive interaction with uncoordinated Pb 2+ cluster in perovskite. The INHA-Pb likewise experiences orbital overlap that analogous to Pb-X coupling. The strong coordinate enhances the bonding with the perovskite surface, which can upshift the conduction band (CBM) and broaden the bandgap. The upshifted CBM along with passivated defects increase photogenerated charge density at the interface, attributing to the V OC enhancement. We further extract the trap density by the log plots of the dark J-V curves. Figure 5d shows the dark J-V curves of the films in where V TFL is voltage kink point, N t is defect numbers, e is the elementary charge of the electron, L denotes the thickness, ε is the perovskite dielectric constant and ε 0 represents the vacuum permittivity. As showed in Fig. 5d and Eq. (2), the control perovskite film presents a trap density of 5.23 × 10 16 cm -3 and is consistent with previous publications [10,53]. However, the INHA/perovskite is calculated to be value of the 2.33 × 10 16 cm -3 , which suggests that the incorporation of INHA can significantly reduce the trap defects at the interface, yielding a high photovoltaic performance with improved V OC and FF. In addition, the strong bonding between INHA and perovskite is thermodynamically favorable and presents excellent stability at the interface for perovskite working devices. We performed the stability of the devices that were stored at atmosphere without any encapsulation (Fig. 6a). The control devices undergo a gradual efficiency decline during one month's tracking, retaining 70% of their initial efficiency. In comparison, the device processed from the INHA passivation preserves its 90% initial efficiency for the entire one month's testing. We measured the shelf-stability of encapsulated devices under ambient conditions in terms of XRD peaks shifting. It is seen that the INHA based devices does not undergo any peaks shifting for 5 weeks, which is related to the improved contacts between the perovskite and interface layers. Overall, these results verify significantly enhanced stability of the INHA-modified device as compared to the control devices.

Conclusions
In summary, we analyzed the influence of INHA on the interface of ETL/perovskite and found that INHA can adjust the interface defects of SnO 2 /perovskite and promote the crystallization of perovskite. Because INHA is anchored on the surface of SnO 2 , the oxygen vacancy defect of SnO 2 electron transport layer is effectively suppressed. At the same time, the coordination effect of INHA and Pb 2+ improves the crystallization process of perovskite and promotes the formation of high-quality perovskite film. These comprehensive factors enhance the separation and transmission of carriers at the interface. Equally important, the influence of INHA passivation layer on the photoelectric performance of perovskite solar cells was discussed. INHA act efficiently for perovskite solar device, improving the PCE from 18.17% to 21.12%, and enhanced the device stability. To the best of our knowledge, the effect of INHA on perovskite growth is first expounded in this work. The INHA-assisted crystal growth is introduced for deposition of perovskite thin-films, which is expected to promote the preparation of large-area films and devices.