Enhancing the efficiency of planar heterojunction perovskite solar cells via interfacial engineering with 3-aminopropyl trimethoxy silane hydrolysate

The interfacial compatibility between compact TiO2 and perovskite layers is critical for the performance of planar heterojunction perovskite solar cells (PSCs). A compact TiO2 film employed as an electron-transport layer (ETL) was modified using 3-aminopropyl trimethoxy silane (APMS) hydrolysate. The power conversion efficiency (PCE) of PSCs composed of an APMS-hydrolysate-modified TiO2 layer increased from 13.45 to 15.79%, which was associated with a significant enhancement in the fill factor (FF) from 62.23 to 68.04%. The results indicate that APMS hydrolysate can enhance the wettability of γ-butyrolactone (GBL) on the TiO2 surface, form a perfect CH3NH3PbI3 film, and increase the recombination resistance at the interface. This work demonstrates a simple but efficient method to improve the TiO2/perovskite interface that can be greatly beneficial for developing high-performance PSCs.


Introduction
Since the first report on perovskite solar cells (PSCs) in 2009 by Miyasaka et al., the power conversion efficiencies (PCEs) of PSCs have rapidly increased from approximately 3.8 to over 20% [1,2]. Two dominant device structures, known as mesostructured and planar heterojunction (PHJ) cells, have been employed to fabricate

2.
3. Modification of the TiO 2 layer by 3-aminopropyl trimethoxy silane hydrolysate APMS was hydrolysed in an 80 : 20 w/w mixture of ethanol and deionized water at a concentration of 10% w/w and with acetic acid added at 5% w/w relative to the solvent. The FTO substrate with the compact TiO 2 layer was immersed into the hydrolysate, and the grafting reaction was sustained for 2 h [14]. After that, the TiO 2 layer with APMS hydrolysate was washed with water and then immersed into 0.5 mol l −1 HI solution for 2 h. Subsequently, the treated compact TiO 2 layer was washed by deionized water and dried in a N 2 stream.

Fabrication of PSCs
MAI and PbI 2 were added into a mixture of GBL and DMSO (7 : 3 v/v) at 60°C for 12 h. The precursor solution was coated onto the TiO 2 layer with and without APMS modification at 5000 r.p.m. for 55 s. During the spin-coating process, the substrate was treated by chlorobenzene drop-casting [15]. The substrate was then dried on a hot plate at 100°C for 10 min. A spiro-MeOTAD solution was prepared by dissolving 72.3 mg of spiro-MeOTAD in 1.0 ml of chlorobenzene, into which 28.8 µl of tBP and 17.5 µl of a Li-TFSI solution (520 mg Li-TFSI in 1 ml acetonitrile, Sigma-Aldrich, 99.8%) were added. The spiro-MeOTAD solution was spin-coated onto the perovskite films at 5000 r.p.m. for 30 s. Finally, an Au electrode with a thickness of 80 nm was thermally evaporated onto the spiro-MeOTAD-coated substrates.

Characterization
The surface of bare compact TiO 2 (c-TiO 2 ) and APMS-hydrolysate-treated c-TiO 2 was investigated using X-ray photoelectron spectroscopy (XPS, ESCALAB-210) with Al Kα radiation (1486.6 eV). The current density (J)-voltage (V) characteristics were measured with a computer-controlled Keithley 2400 under AM 1.5 illumination (100 mw cm −2 ) from a Newport solar simulator. The incident photon-to-electron conversion efficiency (IPCE) was measured by a QEXL Solar Cell Spectral Response instrument. Contact angles were measured by a KRUSS DSA30. The morphology of c-TiO 2 /perovskite and c-TiO 2 /APMShydrolysate/perovskite was characterized using field emission scanning electron microscopy (FE-SEM, Hitachi S-4800). X-ray diffraction (XRD) patterns of the CH 3 NH 3 PbI 3 films were recorded on a HaoYuan DX-700 diffractometer. Capacitance-voltage (C-V) measurements and electrochemical impedance spectroscopy (EIS) were performed on an Electrochemical Workstation (VMP3, Bio-Logic, France).

Results and discussion
The n-i-p planar PSC structure with APMS hydrolysate between TiO 2 and CH 3 NH 3 PbI 3 layer is shown in figure 1a. Cross-sectional SEM image exhibits each functional layer in figure 1b. (HO) 3 -Si-R-NH 3 I was expected to grow on the surface of the compact TiO 2 layer, in which NH + 3 and I − moieties were incorporated into the surface of the CH 3 NH 3 PbI 3 layer, thus anchoring it to the perovskite layer [7,10].
As seen in figure 1b and electronic supplementary material, figure S1, no obvious modified layer can be observed. The APMS hydrolysate layer was considered as self-assembly monolayer or extremely thin layer, covering the TiO 2 surface [16]. To investigate the hydrolytic degree of APMS and the grafting onto the surface of the compact TiO 2 layer, X-ray photoelectron spectroscopy (XPS) was used to analyse the TiO 2 film before and after surface modification [17]. The XPS wide-scan survey spectra are shown in figure 2a. All of the peaks were calibrated using the C 1s peak (284.8 eV) as a reference [18]. Peaks of Si 2p, N 1s and I 3d can be clearly observed in the XPS spectra of TiO 2 before and after modification with APMS (indicated by rectangles in figure 2a). A detailed analysis of these XPS spectra provides clear evidence that the films were chemically modified, which was confirmed by the Si 2p, N 1s and I 3d spectra from fitted curves obtained using XPSPEAK software. As illustrated in figure 2b, the Si 2p peak was located at 102.07 eV following modification, while N 1s was at 401.25 eV (figure 2c), and the peak for I was located at 618.41 eV (figure 2d). Thus, the APMS hydrolysis product was clearly grafted onto the hydroxyl-rich TiO 2 layer.
The current density (J)-voltage (V) characteristics of the PSCs both without and with APMS hydrolysate modification, measured under reverse scan are shown in figure 3a.       fitted curve shifted right to a higher efficiency. This shift indicates that the average efficiency of the PSCs with modified compact TiO 2 layers is higher than that of the cells with unmodified layers. Using another analysis method, the PCEs of PSCs modified with APMS hydrolysate exhibit an efficiency range of 10-15%, which is higher than that of unmodified cells, 8-13%. In order to further demonstrate the statistics, a PC 60 BM layer was introduced as an interface layer as reported [19][20][21]. As a control group, the only difference in the experiment was to replace APMS hydrolysate with PC 60 BM. PCBM solution was spin-coated onto the clean substrates at 6000 r.p.m. for 40 s. The cross-sectional SEM image is shown in electronic supplementary material, figure S1b. The best device obtained a PCE of 14.82% with a V oc of 1.0 V, a FF of 64.73% and a J sc of 22.75 mA cm −2 . Figure 3c presents the incident photon-to-electron conversion efficiency (IPCE) spectra of the PSCs. The curves of all of the PSCs display a wide photoresponse from 350 to 800 nm, which is consistent with the absorption spectrum of CH 3 NH 3 PbI 3 . Photocurrent generation was initiated at 1.55 eV, which is in good agreement with the bandgap of CH 3 NH 3 PbI 3 [22]. The integrated photocurrents estimated from the IPCE curve were 21.1 and 20.0 mA cm −2 for the two devices with and without APMS hydrolysate, respectively, which agree well with the values obtained from the J-V measurements. Moreover, the device with APMS hydrolysate obviously demonstrates higher IPCE values, especially in the range of 450-750 nm, as evidenced by the higher IPCE value exhibited at 500 nm (77.6%) related to its bare compact-TiO 2 ETL-based counterpart (74.0%). This higher IPCE may benefit from more efficient electron collection and less charge recombination due to the insertion of APMS hydrolysate, thereby increasing the J sc .
The contact angles of droplets of GBL on the modified and unmodified TiO 2 films were measured to investigate the mechanism through which APMS hydrolysate improves the efficiency of the PSCs [23]. As shown in figure 4a,b, the contact angle of GBL on the untreated TiO 2 film is 23.83°, while the GBL spreads out on the APMS-hydrolysate-modified TiO 2 film. These results suggest that the GBL wettability on TiO 2 was enhanced.
SEM images of the perovskite on both the bare and modified TiO 2 layers are shown in figure 4c,d. The perovskite crystals on the modified TiO 2 layer are more uniform than those on the bare TiO 2 layer, and the crystals grown directly on the bare TiO 2 surface exhibit a slightly smaller grain size than those grown  on the modified surface. The morphological evolution of the perovskite film with APMS hydrolysate could be attributed to the improved miscibility of the substrate with the perovskite, wherein the amino group is expected to become ammonium and incorporate into the crystalline structure of the perovskite, as shown in figure 1. Larger grain sizes result in fewer grain boundaries for the photogenerated charges to traverse, thereby decreasing charge losses due to recombination at grain boundaries [24]. Figure 4e illustrates the X-ray diffraction (XRD) patterns of the CH 3 NH 3 PbI 3 films grown on TiO 2 both with and without APMS hydrolysate modification. The peaks at 14.1°, 28.4°and 42.1°can be attributed to the (110), (220) and (330) faces of the CH 3 NH 3 PbI 3 crystalline structure, respectively [7]. The characteristic peaks appear at the same angles, indicating the pure perovskite phase on both surfaces without a change in the crystal orientation [25]. In addition, the diffraction peaks at 14.1°and 28.4°w ere significantly enhanced by APMS hydrolysate modification, indicating an improvement in the crystallinity of the CH 3 NH 3 PbI 3 film.
The steady-state PL spectra of the CH 3 NH 3 PbI 3 films are shown in figure 4f to illustrate the charge transport and dynamics of the corresponding devices fabricated on the c-TiO 2 layer without and with APMS hydrolysate treatment. The samples were excited at 440 nm, and both perovskite films exhibited an emissive band with a maximum at approximately 770 nm and a broad emission band ranging from 720 to 850 nm. With the introduction of APMS hydrolysate, a significant fluorescence quenching of perovskite is exhibited, which indicates enhanced electron transport from the perovskite to the APMS-hydrolysatetreated c-TiO 2 layer. This phenomenon demonstrates that the introduction of APMS hydrolysate could facilitate charge transfer between the perovskite and TiO 2 layer. Owing to the shorter diffusion length of electrons than that of holes in perovskite materials, efficient electron transfer and extraction balance the electron and hole transport in PSC devices, resulting in a significant enhancement of the FF [17].
The electronic trap states are able to delocalize charge carriers and induce high capacitance at the interface, which can be readily detected using impedance spectroscopy [25]. To further investigate the trap states on the compact TiO 2 surface both before and after modification, capacitance-voltage measurements were performed on the PSCs fabricated with and without APMS hydrolysate treatment at 1 kHz (figure 5a). At this frequency, the capacitance clearly varies with increasing bias voltage, which is indicative of charge accumulation at the compact layer and can thus reflect its capacitance [26]. As shown in figure 5a, the capacitances of devices treated with APMS hydrolysate are lower, confirming the passivation effect of APMS hydrolysate. The change in capacitance supports the speculation that fewer carriers gather in the traps and that a longer lifetime results in higher performing PSCs fabricated with APMS-hydrolysate-treated compact TiO 2 .
To further investigate the effect of the APMS hydrolysate at the TiO 2 /CH 3 NH 3 PbI 3 interface on the photovoltaic performance, electrochemical impedance spectroscopy (EIS) was employed to characterize the charge-transfer dynamics of PSCs by analysing the variation in the impedance related to the different . Two semicircles are revealed, one in the high-frequency range and another in the low-frequency range, which were measured from 0.1 Hz to 1 MHz. In such PSCs, the resistance at the TiO 2 /CH 3 NH 3 PbI 3 /HTM interface can be determined from the high-frequency (10-100 kHz) semicircle in the Nyquist plots [18,27]. As reflected by the Nyquist plots, the recombination resistance is higher after APMS hydrolysate is grafted onto the TiO 2 layer, which indicates higher current losses via recombination and an increase in the FF. This increased resistance indicates the retardation of electron back-flow from the TiO 2 layer into the perovskite [28][29][30].

Conclusion
Modifying the interface between TiO 2 and perovskite layers by inserting APMS hydrolysate was demonstrated to enhance the photovoltaic performance of solution-processed PHJ PSCs. The PCE improved from 13.45 to 15.79% (representing the best solar cells), and the average PCE increased from 12.01 to 14.20%, thus confirming the desired effect of APMS hydrolysate. First, the wettability of GBL on the TiO 2 surface was enhanced, and as a result, the CH 3 NH 3 PbI 3 precursor solution spread out over the compact TiO 2 layer. Second, perovskite crystals were larger and more uniform on the modified layer. The surface traps of TiO 2 could be passivated by the APMS hydrolysate, and a portion of the molecule was believed to have incorporated into the perovskite crystal. Third, EIS tests revealed that the recombination resistance of the TiO 2 /perovskite interface increased. Our work highlights the effects of APMS hydrolysate on the performance of the TiO 2 /perovskite heterojunction in PHJ PSCs.
Data accessibility. The datasets supporting this article have been uploaded as the electronic supplementary material.