Soluble hexamethyl-substituted subphthalocyanine as a dopant-free hole transport material for planar perovskite solar cells

Boron subphthalocyanine (SubPc) has special physical and chemical properties, originating from its non-centrosymmetric, near-planar taper structure and large conjugated system; it can act as an alternative to the small molecule hole-transporting material 2,2′,7,7′-tetrakis-(N,N-di-p-methoxyphenylamine)-9,9′-spirobifluorene in perovskite solar cells (PSCs). To achieve a higher solubility in common organic solvents and a more suitable highest occupied molecular orbital energy level that aligns with the valence band of the perovskite material, a SubPc molecule with a hexamethyl substitution at its peripheral position (Me6-SubPc) was successfully designed and synthesized in a one-step method. Completely solution processed PSCs were fabricated with only a small hysteresis, a power conversion efficiency of 6.96% and Voc of 0.986 V.


Introduction
Owing to their low cost, high open circuit voltage and simple structure, organic-inorganic lead halide perovskite solar cells (PSCs) have attracted significant attention since first being reported by the Miyasaka group. The power conversion efficiency (PCE) of PSCs has increased from 3.8% in 2009 to 22.1% in 2016 [1][2][3][4][5].
These highly efficient PSCs always use light-absorbing sensitizers, electron transporting materials (ETMs) and holetransporting materials (HTMs) [6][7][8]. HTMs serve as hole transport channels, which facilitates hole extraction and retards  1 H-NMR spectra were obtained from a CDCl 3 solution on a Bruker Ascend 400 NMR spectrometer. Highresolution mass spectrometry was recorded on a Q-Exactive mass spectrometer. The UV-Vis spectrum of the solution (in CH 2 Cl 2 ) was recorded on a PerkinElmer Lambda750S spectrophotometer. The HOMO level was measured directly using an ionization energy measurement system (IPS-4).

Fabrication and characterization of perovskite solar cells
The structure of the PSCs fabricated in this study was FTO/SnO 2 /PCBM/perovskite/Me 6 -SubPc/Au on patterned FTO glass. The patterned FTO glass was washed with detergent, ethanol, chlorobenzene and acetone in an ultrasonic bath, subsequently dried with nitrogen stream, and finally treated with ultraviolet-ozone for 20 min. A SnO 2 electron transport layer (ETL) was prepared according to a method we reported on previously [28]. A 0.1 M precursor solution of SnO 2 ·2H 2 O in ethanol was spin coated on the surface of the cleaned FTO substrates with a spin speed of 2000 revolutions per minute (r.p.m.) for 45 s to obtain a compact and uniform SnO 2 ETL. Then, the FTO substrates coated with the SnO 2 ETL were annealed on a hotplate in air at 180°C for 1 h; 15 mg PCBM was dissolved in 1 ml chlorobenzene, and the solution was spin coated onto the surface of the FTO /SnO 2 ETL samples with a spin speed of 2000 r.p.m. and heated on a hotplate at 100°C for 10 min. The perovskite layer (thickness of about 500 nm) was prepared by solvent engineering. The thickness of the perovskite layer was gauged using an Ambios Technology (Santa Cruz, USA) XP-2 profilometer. The Me 6 -SubPc solutions were prepared by dissolving 15 mg Me 6 -SubPc in 1 ml chlorobenzene. The Me 6 -SubPc solution was spin coated onto the surface of the perovskite layer with a spin speed of 4000 r.p.m. for 30 s. The Au electrodes were subsequently deposited through thermal evaporation under a vacuum of approximately 1 × 10 −6 T; the electrode thickness was monitored in situ using quartz crystal monitors during the evaporation.
The current density-voltage (J-V) curves were measured with a computer-controlled Keithley 236 source measurement unit coupled with a Zolix ss150 solar simulator. A xenon lamp coupled with AM1.5 solar spectrum filters was used as the light source, and the optical power at the sample was 100 mW cm −2 . The intensity of the solar simulator's light was calibrated using a standard silicon solar cell. The external quantum efficiency spectrum was measured using a solar cell quantum efficiency/external quantum efficiency measurement system (Zolix Solar cell scan 100) with a model SR830 DSP lockin amplifier coupled with a WDG3 monochromator and a 500 W xenon lamp. Thin film stacks with SnO 2 /PCBM/perovskite and SnO 2 /PCBM/perovskite/Me 6 -SubPc structures on FTO patterned glass were fabricated following the same procedure as for PSC fabrication. Grazing incidence X-ray diffraction patterns of the samples were recorded on a Smartlab 9 kW diffractometer with a Göbel mirror attachment. The irradiation occurred with a parallel CuK α 1,2 X-ray beam (grazing incident angle of 2.000°(θ)); the detector was set to collect the diffraction data in a 2θ range of 3-30°with a step-size of 0.02°(2θ) at a fixed speed of 1 s step −1 . The UV-Vis spectrum of a Me 6 -SubPc film was recorded on a PerkinElmer Lambda750S spectrophotometer. The Me 6 -SubPc film was prepared by spin-coating a solution of 15 mg Me 6 -SubPc in chlorobenzene onto a cleaned FTO substrate. The morphologies of the thin film samples were probed using atomic force microscopy (AFM) with a Keysight Technologies (5500AFM/STM) scanning probe in tapping mode. Steady-state photoluminescence (PL) spectra were measured by FLS980 Spectrometer-Edinburgh Instruments. The wavelength of excitation light was 470 nm and the samples were excited from the perovskite or Me 6 -SubPc layer. Impedance spectroscopy measurements were carried out by CHI 760E under illumination conditions with different direct current bias potentials from 0.3 to 0.8 V and frequency from 100 kHz to 0.01 Hz. The carrier mobility was extracted by fitting the J-V curves according to the modified Mott-Gurney equation [43]: where J is the current density, ε 0 is the permittivity of free space, ε r is the relative permittivity, µ is the zero-field mobility, V is the applied voltage and d is the thickness of active layer. 3. Results and discussion 3.1. Synthesis and characterization of hexamethyl-substituted subphthalocyanine Me 6 -SubPc was synthesized using a modified version of a previously reported one-step method (synthetic route used in this study shown in scheme 1). The product was recrystallized three times for application in PSCs. The X-ray data for Me 6 -SubPc was collected on a Bruker ApexII diffractometer equipped with a charge-coupled device-area detector using graphite-monochromated Cu Kα radiation (k = 1.54178 Å) generated from a sealed tube source. Data were collected and reduced using the SMART and SAINT software [44] in the Bruker package. The structure was solved by means of direct methods using SHELXT-2014 [45,46] and refined via full-matrix least-squares methods based on Fo [45,46] with SHELXL-2014 in the framework of OLEX [47] software. For disordered solvent molecules, as observed on the Fourier density map, a standard SQUEEZE protocol implemented into PLATON software was used to account for areas of unassigned electron density [48]. The non-H atoms found in the electron density map were refined anisotropically. All non-H atoms were placed in calculated positions with the 'riding-model technique' with displacement parameters bonded to a parent atom. A summary of the crystallographic data and of the refinement parameters for Me 6 -SubPc is provided in table 1.
A perspective view of the Me 6 -SubPc molecule with accompanying atomic labelling scheme is shown in figure 1a. Selected bond lengths and angles are presented in table 2. The compound crystallized in the space group I4/m. In the taper structure of Me 6 -SubPc, a cupped geometry of the delocalized ring system can be observed; it is forced by the tetrahedral boron centre. In this molecule, the boron atom is located nearly 0.6 Å out of the plane of its three bonded nitrogen atoms. The thus created cavity of each molecule is occupied by one CH 2 Cl 2 molecule and its neighbour Me 6 -SubPc molecules. Sphere-like structures are formed via the contribution of four Me 6 -SubPc molecules along with four CH 2 Cl 2 molecules with the help of weak CH-N hydrogen bonding interactions (figure 1b). The packing arrangements of the crystal lattice result from π-π stacking interactions between sphere-like structures, as well as weak hydrogen bonding. As shown in figure 1c,d, two different π-π stacking interactions can be observed in the crystal structure. A parallel face-to-face perfect interaction is visible in figure 1c, with an interplanar distance of 3.5 Å and a small shift distance of 0.3 Å. Offset or slipped packing can also be observed between the phenyl rings in the neighbouring Me 6 -SubPc molecules with an interplanar distance of 3.5 Å and a shift distance of 1.7 Å (figure 1d). Various weak hydrogen bonding interactions, such as between C(1)H-Cl(1) and C(1)H-N(3) as well as between C(10)H-Cl(1) and C(10)H-N(3), also contribute to the highly efficient, closely packed crystal structure, which should be favourable for carrier transport. Figure 2a shows the solution UV-Vis absorption spectra of different concentrations of Me 6 -SubPc. Me 6 -SubPc has a good solubility in DCM, chlorobenzene, dichlorobenzene and other organic solvents. The UV-Vis absorption spectra of a solution and film are shown in figure 2b. The Me 6 -SubPc solution in DCM has strong absorption peaks at 573 nm (Q-band) and 313 nm (Soret band). There is only one strong absorption peak at about 573 nm, which is attributed to the transition to an orbitally degenerate state fitted with the Faraday A term of the Me 6 -SubPc ring monomer. The Soret band is attributed to three transitions to non-degenerate states (Faraday B terms) [34]. Compared with phthalocyanine molecules, the Soret band and Q-band of Me 6 -SubPc appear at shorter wavelengths because of the smaller π-conjugation system. The differences in the absorption spectra between Me 6 -SubPc in solution and a film of SubPc were smaller than for a solution and thin film of phthalocyanine. A 13 nm red shift in conjunction with a broader absorption for each transition was observed for both the Soret band      The optical bandgap of Me 6 -SubPc is 2.07 eV, as estimated from the absorption edge of the solution. The HOMO level of Me 6 -SubPc was measured by IPS and was found to be 5.3 eV (figure 3a). The lowest unoccupied molecular orbital level of Me 6 -SubPc was calculated using the Me 6 -SubPc optical bandgap and its HOMO level; it was found to be 3.23 eV. The energy levels of the different layers of the PSCs were obtained from the literature and are shown in figure 3b [30,31]. As shown in figure 3b, the HOMO level of Me 6 -SubPc is more compatible with perovskites than the HOMO level of SubPc, and thus Me 6 -SubPc can efficiently transfer holes generated in the perovskites to the Au electrodes.

Solar cell performance
The PSCs using Me 6 -SubPc as HTMs with a conventional planar device architecture consisting of FTO/SnO 2 /PCBM/perovskite/Me 6 -SubPc/Au were fabricated and characterized. Compared with TiO 2 requiring high temperature treatment [13,49] and unstable ZnO [50,51], we chose SnO 2 with ideal electron mobility and stability as ETL, and simultaneously used PCBM to modify the interface between perovskite and ETL to improve the electron extraction efficiency [30,31,52]. A cross-sectional scanning   The photovoltaic performances of PSCs using Me 6 -SubPc as HTMs were measured under a xenon lamp with air mass (AM) 1.5 solar spectrum filters and an optical power of 100 mW cm −2 . Herein, the SnO 2 layer (with its high electron mobility) and the effective fullerene passivation layer worked together as an electron selection layer.
We tested 10 PSCs in which Me 6 -SubPc was the hole transport material. Compared with previously reported results in the literature [42], devices using Me 6  6% [42]. The device with the Me 6 -SubPc HTL had a higher V oc of 0.986 V, reduced J sc of 17.21 mA cm −2 , and FF of 0.41, with an increased PCE of 6.96%. The photovoltaic parameters of the devices with Me 6 -SubPc or dopant-free spiro-OMeTAD as their HTL are shown in table 3. As shown in figure 5a,b and table 3, the performance of the device with Me 6 -SubPc is also better than the device performance of dopant-free spiro-OMeTAD. It has been previously reported in the literature that hole transport materials with deeper HOMO levels yield PSCs with higher V oc values [49]. In our work, the HOMO level of SubPc was −5.6 eV, which is deeper than the valence band edge and is detrimental to hole injection from the perovskites to SubPc. It is worth noting that Me 6 -SubPc has a HOMO level of −5.3 eV, which is higher than the valence band edge of perovskites and favourable for hole transport from the perovskites to Me 6 -SubPc. This reduces charge recombination at the interface and thus reduces voltage loss, which was also proved by PL spectra. As shown in the electronic supplementary material, figure S1a, there was a strong PL peak at 778 nm for the reference perovskite layer. After a spin coated Me 6 -SubPc layer on the top of perovskite, the PL spectra of perovskite films can be effectively quenched, indicating that photo-induced excitons can be separated and transferred effectively. The timeresolved photoluminescence (TRPL) spectroscopy was tested to further understand the charge transfer mechanism in PSCs with Me 6 -SubPc HTL. Electronic supplementary material, figure S1b shows the TRPL spectra of films of perovskite and perovskite/Me 6 -SubPc. It is clearly seen that the PL decay lifetime for the perovskite film is shorter than the perovskite/Me 6 -SubPc film, which also proved an efficient hole extraction from perovskite to Me 6 -SubPc. The hole mobility of Me 6 -SubPc was determined by means of hole-only devices (electronic supplementary material, figure S2) using the space charge limited current and found to be 1.4 × 10−5 cm 2 V −1 s −1 . The low hole mobilities were consistent with the reduced J sc and FF of the devices with Me 6 -SubPc HTL. Impedance data of device was recorded and characterized by a single semicircle (electronic supplementary material, figure S3), whereby the capacitance relates to the device properties. This was also a factor affecting the J sc and FF.  Table 3. Photovoltaic parameters of the best device for a reverse and forward scan, averaged photovoltaic parameters of 10 devices with Me 6 -SubPc as HTL and photovoltaic parameters of the best device with dopant-free spiro-OMeTAD as HTL. voltage at 0.1 V s −1 in both the reverse and forward scan direction under simulated solar illumination (AM 1.5, 100 mW cm −2 ). The results are depicted in figure 5b, which demonstrate that the hysteresis in the Me 6 -SubPc-based devices is small. To study the stability of PSCs with Me 6 -SubPc as their HTL, a stable bias of 0.6 V was applied to the device near maximum power, and the stabilized efficiency was recorded (figure 5c). With Me 6 -SubPc as a HTM, the PSC achieved a stable PCE of 6.51%. To further study the performance of PSCs with Me 6 -SubPc as the HTM, the incident photon-to-electron conversion efficiencies of perovskite devices were tested to confirm the photovoltaic performance. As shown in figure 5d, the Me 6 -SubPc-based device exhibited enhanced IPCE over the whole visible wavelength range from 300 to 800 nm, which was in good agreement with electrical characteristics.

Conclusion
We designed and synthesized a Me 6 -SubPc using a one-step method. Compared with unmodified subphthalocyanine, Me 6 -SubPc has a higher solubility in organic solvents, as well as a frontier orbital energy level that is well aligned with the valence band of the perovskite material. PSCs employing Me 6 -SubPc as a hole transport material were prepared using a completely solution-phase process, achieving a V oc of 0.986 V and a PCE of 6.96%. We believe that Me 6 -SubPc molecules can be further modified such as by changing the chlorine substituent on the B central atom to help improve the as-fabricated device performance.
Data accessibility. Our supplementary data for the steady-state PL, time-resolved PL, experimental and calculated J-V characteristics and impedance spectroscopy characterization are available in the electronic supplementary material; the single crystal data file for hexamethyl-substituted subphthalocyanine has been deposited in Dryad and available at: http://dx.doi.org/10.5061/dryad.3d7c900 [53].