Molecularly Engineered Thienyl-Triphenylamine Substituted Zinc Phthalocyanine as Dopant Free Hole Transporting Materials in Perovskite Solar Cells

To endure the success of perovskite solar cells (PSCs), developing dopant-free hole transporting materials is of paramount importance to induce long-term stability. Phthalocyanine has emerged as a viable alternative for common Spiro-OMeTAD, due to their excellent optoelectrical properties and chemical stability. We report the design and semiconducting properties of a thienyl-triphenylamine tetrasubstituted Zinc (II) phthalocyanine, and its application into PSCs. The PSCs fabricated with Zinc phthalocyanine without the use of any dopant and additives gave a competitive power conversion efficiency in excess of 14.5%, along with long-term stability. the green solid obtained washed with MeOH at reflux, and finally purified by preparative TLC (DCM:MeOH, 96:4) affording ZnTPPc (32 mg, 52 %) as a green powder. UV-vis: λ max (DMF)/nm (log Ԑ): 390 (5.01), 647 (4.57), 719 (5.17). HRMS (MALDI-TOF-MS): For C 128 H 92 N 12 O 8 S 4 Zn (M) + calc. 2116.533, found 2116.581. solution diluted by anhydrous ethanol. The substrates kept 500 30 min the Mesoporous TiO 2 films were obtained by spin-coating a mesoporous dispersion of TiO 2 nanoparticles (30 NRD from Dyesol) and heated at 500 °C for 30 min to produce the mesoporous structure. The substrates were transferred to an argon-filled glovebox for depositing perovskite films. The two-step deposition method was adopted to prepare perovskite . In short, 1.3M PbI 2 precursor solution dissolving in a mixed solvent (volume ratio of DMF/DMSO=9.5/0.5) was spin coating at 2,000 rpm for 20 s to prepare PbI 2 layer. The mixed organic precursor solution containing formamidinium iodide: methylammonium bromide: methylammonium chloride= 60: 6: 6 in 1 mL isopropanol was dynamically added to the PbI 2 layer and was spin-coated at 2,000 rpm for 20 s. The as-prepared film was annealed at 150 °C for 15 min in the glovebox. The ZnTPPc film layer was deposited atop perovskite films at a spin rate of 4,000 rpm for 20 s with 10 mM (~ 21 mg/mL) solution with chlorobenzene as the solvent. The Spio-OMeTAD hole transporting layer was prepared by spin-coated the Spiro-OMeTAD solution prepared by dissolving 72.3 mg material in 1 mL chlorobenzene. Finally, a ~ 70 nm gold electrode was deposited on top of the hole-transporting layer by thermal evaporation. J–V curves were recorded with a under AM 1.5 G, 100 mW cm 2 illumination from a 450 W 3A solar simulator (ORIEL, 94023 A). This was calibrated using an NREL certified calibrated monocrystalline silicon solar cell. parameters: active area defined by black cm scan pre-sweep Surface and cross-sectional SEM images recorded by using a The EQE spectra measured using a W Xenon lamp (Newport) attached to IQE200B (Oriel) motorized 1/4m monochromator as the light source. The water contact angle was tested by contact angle goniometer (Ossila). Absorption spectra collected Cary


Introduction
The last decade has witnessed an enormous development in the investigation of organic-inorganic hybrid perovskite solar cells (PSCs) due to their power conversion efficiency (PCE) skyrocketing from 3.8% to 25.2%. [1][2][3][4][5] This rapid development is owing to the outstanding properties of organic-inorganic halide perovskites: panchromatic light absorption and long-range ambipolar carrier diffusion as well as the choice of solution-processability at low temperature. [6][7][8] In the device configuration of PSCs, the light harvester perovskites layers are sandwiched between an electron and hole transporting material (ETM, HTM). This can be categorized into normal (n-i-p) and inverted (p-i-n) structures according to whether ETM or HTM is in contact with the bottom transparent conductive electrode (TCO).
In a standard normal device structure (TCO/ETM/perovskite/HTM/metal electrode), HTMs atop of perovskite layer play pivotal roles in achieving high efficiency and long-term stability in PSCs. The hole transporting layers extract and transport photogenerated holes from perovskite to the metal electrode rapidly, avoiding undesired charge recombination losses at the interfaces and also inhibits the degradation process induced by oxygen and moisture. [9][10][11] An effective HTM must fulfill the following requirements: simplified synthesis procedure, solution-processability, uniform film formation (compact morphology) without any pinhole, the relatively high charge transporting abilities, the well aligned highest occupied molecular orbital (HOMO) /valence band with valence band of perovskite, superior stability under ambient and thermal condition. [12][13][14][15] To date, 2,2´,7,7´-tetrakis-(N,N-di-p, methoxyphenylamine)-9,9´-spirobifluorene (Spiro-OMeTAD) is most investigated HTM in n-i-p structure type devices because of their simple orthogonal solution-processability (to common perovskites) and high performance of the devices. [16][17][18][19][20][21] However, Spiro-OMeTAD possesses low conductivity and mobility in its pristine form, and introducing a hygroscopic p-type dopant and additive such as lithium bis(trifluoromethanesulfonyl)imide (Li-TFSI) and 4-tert-Butylpyridine (t-BP) is prerequisite to enhance the charge transporting ability of Spiro-OMeTAD, but on the contrary, this can also accelerate the degradation of perovskite and lowers the stability of the fabricated devices. The tedious multi-step synthesis and costly purification process of Spiro-OMeTAD along with the use of indispensable hydrophilic dopants and additives limits its further extensive application in PSCs. 22,23 Recently, among alternatives that meet HTM's criteria, including small molecules, conjugated polymers, and inorganic p-type semiconductors , etc., have been specially designed and synthesized to supersede Spiro-OMeTAD. 12,14,15,[24][25][26] Among them, one such promising category of attractive HTM is metal phthalocyanine (Pcs), analogs which own organic macrocycles with an extended π-conjugated system and planar structure. [27][28][29][30] They are cost-effective and attractive due to their facile synthesis and purification, high hydrophobicity, as well as excellent stability under thermal and photo conditions. Moreover, they exhibit a unique combination of chemical, optical and electronic properties that can be tailored by molecular engineering. The peripheral and nonperipheral substitution of Pcs is an effective approach to regulate their chemical and electrical properties. 29,31,32 The efficiency of Pcs-based PSC was significantly improved to nearly 20% since the first paper reported on vacuum-deposited CuPc in PSCs in 2015. [33][34][35] Catchpole et al. applied solution-processed copper (II)  2,9,16,23-tetra-tert-butyl-29H,31H-phthalocyanine as a doped  HTM for PSCs and reported the highest efficiency of the device  with Pcs. 35 Recently, Pc derivatives have emerged as an  appealing dopant-free HTMs. 27,36-39 Dopant-free Zinc (II)  2,9,16,23-tetra-tert-butyl-29H,31H-phthalocyanine dimers with  3,6-bisthienylldiketopyrrolopyrrole (ZnPc-DPP-ZnPc) as dopantfree HTM into PSCs yielded 16.8% PCE. 40 Similarly, Wang et al. successfully developed tetra-methoxyl triphenylamine substituted zinc phthalocyanine (OTPA-ZnPc) as a dopant-free HTM, and corresponding device gave 16.23% efficiency. 41 Xu et al. investigated properties and application of two CuPcs: methoxydiphenylamine / methoxy triphenyl amine-substituted copper Pcs, named as OMe-DPA-CuPc and OMe-TPA-CuPc, respectively. It showed that OMe-TPA-CuPc exhibit better π-π interaction and more parallel orientations of Pcs cores in bulk aggregates compared to OMe-DPA-CuPc, and the corresponding dopant-free HTM based PSCs gave PCE of 19.7%. 37 Substitution with methoxy triphenylamine moieties highlights the strong potential to future design Pc derivatives as dopant-free HTM. Thiophenes or its derivatives as the π-linker are widely investigated as organic semiconducting materials in thin-film devices. 42,43 Particularly, thienyl groups are employed to synthesize dopant-free small molecules HTMs, [43][44][45] but reports dealing with Pcs analogs with thiophene or its derivatives are in scarce. 29,46,47 In this work, guided by rational design, we synthesize and characterize Zinc (II) 2,9,16,23-tetra-thienyl-methoxytriphenylamine-29H,31H-phthalocyanine, named as ZnTPPc (Fig. 1), and its integration as a dopant-free HTM in PSCs is investigated. Notably, the PSC based on dopant-free ZnTPPc delivered a PCE of 14.5%, which is very competitive to that of the device employing the wellknown undoped HTMs. Particularly, the facile synthesis of ZnTPPc based on commercial precursor materials makes it a promising new class of HTMs with high stability.

Results and discussion
Synthesis and characterization of ZnTPPc Our objective was to synthesize a tetrasubstituted ZnTPPc with triphenylamine-thienyl conjugated aromatic electron-rich substituents. These substituents were selected in order to improve the solubility and hole mobility of Pc as a HTM. ZnTPPc was synthesized from autocondensation of phthalonitrile 2 in the presence of Zn(OAc)2 and a drop of DBN as catalyst with 52% of yield (Scheme 1). The phthalonitrile 2 was obtained by Suzuki coupling of arylboronate and bromo-thienyl-phthalonitrile 3, which in turn can be easily prepared from 4-(thiophen-2-yl)phthalonitrile 4 by bromination with NBS. Phthalonitrile 4 was synthesized by a Suzuki reaction of 4-iodophthalonitrile and thiophene-2-boronic acid pinacol ester. Palladium acetate was the best catalyst for this reaction, better than the usual Pd2(dba)3. The phthalonitriles 2-4 were characterized by 1 H-NMR and 13 C-NMR spectroscopy (Fig. S1-S6, ESI †). The final product ZnTPPc was systematically characterized by 1 H-NMR spectroscopy, HR-MALDI-TOF mass spectrometry UV-vis absorption and fluorescence spectra. The 1 H-NMR spectra in DMSO-d6 of ZnTPPc showed broad aromatic signals meanwhile the aliphatic ones are well defined. This may be due to aggregation by stacking and to the presence of four different region isomers (Fig. S7, ESI †). The absorption spectra of ZnTPPc in DMF (Fig. 2a) presented the Soret bands located in the ultraviolet region around 390 nm and the Q band in the region between 600 nm and 780 nm. Intensive absorptions at 647 nm and 719 nm, respectively, could be ascribed to strong π-π* transitions of Q-band. 37 The ZnTPPc show strong, narrow steady-state fluorescence with peak position at 768 nm (Fig. 2a). The optical bandgap (Eg) of ZnTPPc was estimated from the corresponding intersection of absorption and fluorescence spectra, thus, resulting in 1.65 eV.

ZnTPPc
The electrochemical characterization was performed to evaluate the energy level alignment of the perovskite/ZnTPPc interface via using cyclic voltammetry as displayed in Fig. 2b. The ferrocene redox couple was used as the external standard and ZnTPPc showed two oxidation peaks at 0.35 V and 0.79 V (vs Fc/Fc + ), respectively. The highest occupied molecular orbital energy level (EHOMO) of ZnTPPc was calculated according to Equation: EHOMO=-4.8-Eox, where Eox is the first onset oxidation potential. The EHOMO value was estimated to be -5.15 eV and located in between valence band of perovskite (-5.9 eV) and work function of gold (-5.1 eV), which could facilitate the hole carrier extraction from perovskite to metal electrode. 48 While the lowest unoccupied molecular orbital energy level (ELUMO) of ZnTPPc was -3.44 eV, determined by adding the previously obtained Eg to EHOMO value. ELUMO lies higher than the conduction band of perovskite (-4.4 eV), which can efficiently block the photo-generated electron and avoid recombination at the interface of perovskite/HTM.
Particularly, the charge transporting abilities of HTMs are crucial to fabricate high-performance PSCs and the conductivity and hole mobility of ZnTPPc were firstly studied. The electrical conductivity was determined by measuring the current-voltage (I-V) curve of the device in a structure (FTO/HTM/Ag) following the equation of σ = IA -1 V −1 d, where A is the active area of the device, d is the thickness of HTM.
The calculated conductivity of ZnTPPc was found to be 2.9 × 10 -6 S cm −1 , and for pristine Spiro-OMeTAD was 6.9 × 10 -7 S cm −1 , showing the similar values reported elsewhere (10 -7 -10 -8 S cm -1 ). 10,49 The conductivity value of ZnTPPc is notably one magnitude higher than the value of pristine Spiro-OMeTAD. Besides, the hole mobility of the ZnTPPc and pristine Spiro, was evaluated by space charge limited current (SCLC) method using a hole-only device with the architecture of FTO/PEDOT:PSS/HTM/Ag. The current density-voltage (J-V) curve was tested under the dark and ambient condition and the hole mobility was calculated using the Mott-Gurney law (J=9εε0μVapp 2 /8d 3 ). The obtained hole mobility of ZnTPPc was 4.1 × 10 -5 cm 2 V -1 s -1 , which is higher than that of pristine Spiro-OMeTAD (1.7 × 10 -5 cm 2 V -1 s -1 ). 50,51 The relatively higher conductivity and mobility of ZnTPPc will allow improvement of in charge transporting ability and fill factor thus device performance.
The film-forming ability of ZnTPPc was investigated with the help of microscopy techniques such as scanning electron microscopy (SEM) and atomic force microscopy (AFM). The Fig. 3a & 3b showed the top-view SEM of bare perovskite and perovskite covered with ZnTPPc (perovskite/ZnTPPc), respectively. Compared to bare perovskite, the surface microstructure of perovskite/ZnTPPc film shows full-coverage without any voids or defects. Fig. 3c & 3d, suggests similar observation with AFM topographical images and bare perovskite film displayed a root mean square (RMS) roughness value of 29.1 nm. The surface of perovskite/ZnTPPc becomes smoother and decreased RMS value of 19.6 nm was noted. The uniform, smooth HTM film can extract photo-generated holes effectively and avoid interfacial recombination loss, and improve the device performance. In all, the ZnTPPc armed with such excellent properties such as matched energy level with perovskite and improve electrical properties, film-forming ability demonstrates great potential towards fabricating efficient PSCs.

Photovoltaic performance of PSCs with ZnTPPc as HTM
To validate the potential of ZnTPPc as an efficient HTM in PSCs, we fabricated devices with an architect of FTO/b&mp-TiO2/perovskite/ZnTPPc/Au (b&mp-TiO2 presents blocking and mesoporous TiO2 layer, Fig 4a) and for comparative purpose, pristine and doped Spiro-OMeTAD were used as reference HTM. Fig 4b shows a representative cross-sectional SEM image of PSC device with ZnTPPc as an HTM, where the mixed perovskite atop of mp-TiO2 was prepared by a two-step deposition method and the thickness was ~ 450 nm. 17,18,25 A thin layer of ZnTPPc covered on perovskite film allows around 60 nm thick film. The energy level diagram (Fig. 3c) of ZnTPPc based device shows a well matched HOMO-LUMO energy level with respect the perovskite and thus exhibit the stronger possibility of transporting the holes efficiently. The J-V curves of the champion device with different HTMs were measured under reverse scan and corresponding photovoltaic parameters are shown in Fig. 4d and Table 1. The PSC with ZnTPPc yielded an overall PCE of 14.5%, with an open-circuit voltage (Voc) of 994.3 mV, a short circuit current density (Jsc) of 21.0 mA cm -2 , and a fill factor (FF) of 69.42%. A reference device with conventional Spiro-OMeTAD doped with lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) and 4-tert-Butylpyridine (t-BP) was fabricated with the same condition, and exhibited a PCE of 18.47% with a Voc of 1075.9 mV, a Jsc of 23.62 mA cm -2 , and a FF of 72.69%, which was slightly higher than that of the device with pristine ZnTPPc. However, when using pristine Spiro-OMeTAD as a HTM, the device showed inferior PV parameters and registered a PCE of 5.14%, with a Voc of 993.9 mV, Jsc of 15.25 mA cm -2 , FF of 33.92%. Notably, low Jsc and FF was found due to the poor electrical properties of Spiro-OMeTAD in its pristine form. The series resistance (Rs) of different HTMs are shown in Table  1. The ZnTPPc-based device displayed the Rs of 58 Ω, which is comparable with the value of doped Spiro-OMeTAD device (44 Ω) and lower than the value of pristine Spiro-OMeTAD device (424 Ω). Subsequently, the relative low Rs of ZnTPPc-based device allowed us to achieve improved FF and subsequently PCE. [52][53][54] Fig. 4e presents the external quantum efficiency (EQE) spectra and the integrated current density (integrated Jsc). The integrated Jsc values of PSCs with ZnTPPc, pristine and doped Spiro-OMeTAD are 19.41, 19.06, and 20.86 mA cm -2 , respectively, close to the Jsc value attained from J-V curves under the standard solar AM 1.5G. The lower Jsc value of the ZnTPPc-based device obtained from EQE spectra compared with doped Spiro-OMeTAD is in agreement with the value from the J-V curves. The device with ZnTPPc and Spiro-OMeTAD own the same EQE value at wavelength ranging from 300 nm to 580 nm, while the device with ZnTPPc showed lower EQE value than that of the device with Spiro-OMeTAD at wavelength ranging from 580 nm to 800 nm. Considering the strong absorbance of Pcs at wavelength ranging from 580 nm to 800 nm (Fig. 1a & 4f), one possible reason of decreased EQE and Jsc of ZnTPPc-based device is that unabsorbed long-wavelength light passing through the perovskite active layer was absorbed by ZnTPPc instead of being reflected by the electrode and reabsorbed by the perovskite layer, leading to the decrease of photocurrent and efficiency. Photoluminescence (PL) characterization for perovskite with different HTMs on the quartz substrate was used to analyse the charge carrier separation and transport process. [55][56][57][58] In order to observe the hole extraction ability of different HTMs, steady state photoluminescence (PL) measurement was made with the structure of quartz/perovskite/HTM. Fig. 4g. shows the steady state PL spectra for pristine perovskite and with different HTM excited from the quartz side at 600 nm. The strong PL emission peak centred at 798 nm was observed for pristine perovskite while a decrease in PL intensity was noticed with all the HTM samples indicating the efficient hole extraction. Although, HTMs can quench the PL of perovskite, however, doped Spiro-OMeTAD exhibit the maximum PL quenching compared to undoped Spiro and ZnTPPc.
Interestingly, in case of ZnTPPc , blue shifted PL peak (from 798 to 787 nm at the Pero/ZnTPPC was observed, which can be ascribed to a decrease in the surface trap states 59 on the surface and grain boundaries of perovskite layer, since surface roughness and film morphology was improved after ZnTPPc deposition as shown in Figure 3b-d. To evaluate the reliability of the device, the stabilized power output was measured (Fig. 4h). The device with ZnTPPc showed 15.7 mA cm -2 under the constant bias of near the maximum power point (0.78 V), and yielded stable PCE of 12.2%, in good agreement with the average PCE obtained from forward and reverse scan (12.92%). Besides, to check the reproducibility, 20 independent devices were statistically analysed, and the histograms of PCE (Fig. 4i) as well as Voc, Jsc and FF (Fig. S9 ESI †) are summarized in Table 1. The average PCE was 11.80±1.51%. Our results suggest reproducibility of PSCs with ZnTPPc as an HTM possibly due to its core structure, which has high tolerance power and nature.  Apart from the performance in terms of efficiency, the long-term stability of corresponding devices is another important parameter to further assess the potential of HTMs. These un-encapsulated devices were stored in a dry box under 40 -50% relative humidity and room temperature condition. The ZnTPPc-based PSCs present average PCE of 12.75% after aging for 432 h, while the initial average values of the device were 12.31%, showing slight increases after aging (Fig. 5 and Table S2). Such phenomena typically originated from the optimization of interface contact by solvent drying with storage time. In contrast, the average PCE of the device with doped Spiro-OMeTAD dropped from 18.28% to 16.91% after aging 120 h under similar conditions (Fig. S10 and Table S2, ESI †). Importantly, the average PCE of the ZnTPPc-based device remains 88% of initial efficiency after aging for over 3000 h (Fig. 5e and Table S3, ESI †). This point towards ZnTPPc based PSCs superior stability. The plausible reasons for improved stability are as follows: the perovskite/ZnTPPc show a higher water contact angle of 98° compared to that of Spiro-OMeTAD (76°) (Fig. 5f), the excellent hydrophobic nature of ZnTPPc can effectively block the water penetration into perovskite layers.
To conclude, through molecular engineering, we have designed a tetrasubstituted ZnTPPc with electron-rich thienyl-triphenylamine conjugated aromatic substituents. We found that ZnTPPc possesses superior electrical properties such as relative high conductivity and mobility compared to pristine Spiro-OMeTAD, suitable energetic level alignment and can supress the trap state on the surface of perovskite. The perovskite solar cells fabricated thereof employing ZnTPPc as a hole selective layer gave a highly competitive power conversion efficiency over 14.5%, without the use of any dopant and additives. Furthermore the ultrathin layer will cut down materials consumption and will further lower the cost. ZnTPPc maintained the initial efficiency after the aging measurement.

Measurements for materials
NMR spectra were measured with a Bruker AC 300. UV-vis were recorded with a Helios Gamma spectrophotometer. Fluorescence spectra were recorded with a Perkin Elmer LS 55 Luminescence spectrophotometer. High-resolution mass spectra were obtained from a Bruker Microflex LRF20 matrix assisted laser desorption/ionization time of flight (MALDI-TOF) using dithranol as matrix.
Cyclic voltammetry was measured in a conventional three-electrode cell using a µ-AUTOLAB type III potentiostat/galvanostat at 298 K over benzonitrile and deaerated sample solutions containing 0.10 M tetrabutylammonium hexafluorophosphate (TBAPF6) as supporting electrolyte. Platinum as working electrode, Ag/AgNO3 as the reference electrode, and platinum as wire counter electrode were employed. Ferrocene/Ferrocenium was used as an internal standard for all measurements.

Device fabrication and characterization
The pre-cleaned FTO (NSG10) substrates were treated with UV-ozone machine for 60 min each time and were heated to 500 °C. A compact blocking layer of TiO2 were deposited by spray pyrolysis using a 5% titanium (IV) diisopropoxide bis(acetylacetonate) solution diluted by anhydrous ethanol. The substrates were kept at 500 °C for 30 min after the spraying. Mesoporous TiO2 films were obtained by spin-coating a mesoporous dispersion of TiO2 nanoparticles (30 NRD from Dyesol) and heated at 500 °C for 30 min to produce the mesoporous structure. The substrates were transferred to an argon-filled glovebox for depositing perovskite films. The two-step deposition method was adopted to prepare perovskite . In short, 1.3M PbI2 precursor solution dissolving in a mixed solvent (volume ratio of DMF/DMSO=9.5/0.5) was spin coating at 2,000 rpm for 20 s to prepare PbI2 layer. The mixed organic precursor solution containing formamidinium iodide: methylammonium bromide: methylammonium chloride= 60: 6: 6 in 1 mL isopropanol was dynamically added to the PbI2 layer and was spin-coated at 2,000 rpm for 20 s. The as-prepared film was annealed at 150 °C for 15 min in the glovebox. The ZnTPPc film layer was deposited atop perovskite films at a spin rate of 4,000 rpm for 20 s with 10 mM (~ 21 mg/mL) solution with chlorobenzene as the solvent. The Spio-OMeTAD hole transporting layer was prepared by spin-coated the Spiro-OMeTAD solution prepared by dissolving 72.3 mg material in 1 mL chlorobenzene. Finally, a ~ 70 nm gold electrode was deposited on top of the hole-transporting layer by thermal evaporation. J-V curves were recorded with a Keithley 2400 source-measurement-unit under AM 1.5 G, 100 mW cm 2 illumination from a 450 W 3A solar simulator (ORIEL, 94023 A). This was calibrated using an NREL certified calibrated monocrystalline silicon solar cell. Testing parameters: active area defined by black masks: 0.09 cm 2 , scan rate: 100 mV s -1 , pre-sweep delay: 10 s). Surface and cross-sectional SEM images were recorded by using a Hitachi S-4800. The EQE spectra was measured using a 150 W Xenon lamp (Newport) attached to IQE200B (Oriel) motorized 1/4m monochromator as the light source. The water contact angle was tested by contact angle goniometer (Ossila). Absorption spectra was collected by Cary 60. Atomic Force Microscopy (AFM) images were acquired with CSI Nano observer AFM.