Efﬁcient and Stable Perovskite Large Area Cells by Low-Cost Fluorene-Xantene-Based Hole Transporting Layer

: Among the new generation photovoltaics, perovskite solar cell (PSC) technology reached top efﬁciencies in a few years. Currently, the main objective to further develop PSCs is related to the fabrication of stable devices with cost-effective materials and reliable fabrication processes to achieve a possible industrialization pathway. In the n-i-p device conﬁguration, the hole transporting material (HTM) used most is the highly doped organic spiro-ﬂuorene-based material (Spiro-OMeTAD). In addition to the high cost related to its complex synthesis, this material has different issues such as poor photo, thermal and moisture stability. Here, we test on small and large area PSCs a commer-cially available HTM (X55, Dyenamo) with a new core made by low-cost ﬂuorene–xantene units. The one-pot synthesis of this compound reduces 30 times its cost with respect to Spiro-OMeTAD. The optoelectronic performances and properties are characterized through JV measurement, IPCE (incident photon to current efﬁciency), steady-state photoluminescence and ISOS stability test. SEM (scanning electron microscope) images reveal a uniform and pinhole free coverage of the X55 HTM surface, which reduces the charge recombination losses and improves the device performance relative to Spiro-OMeTAD from 16% to 17%. The ISOS-D-1 stability test on large area cells without any encapsulation reports an efﬁciency drop of about 15% after 1000 h compared to 30% for the reference case.


Introduction
In the last decade, halide perovskite (PVSK) photovoltaic (PV) technology showed an efficiency improvement from 9.7% to 25.5%, approaching the Si record of 26.7% [1,2]. The reasons behind such rapid development are related to the high absorption coefficient, the ambipolar charge transport, the high charge carriers lifetime and diffusion length, the flexible bandgap tuning, the low exciton binding energy and the defect tolerance of the PVSK absorber [3][4][5]. The affinity with fabrication procedures based on solution process deposition techniques from organic electronics and the experience acquired from previous PV technologies (dye-sensitized and copper indium gallium selenide solar cells) further contribute to the success of PVSK PV in terms of performance [6][7][8][9][10][11][12]. The PVSK absorber is formed by an organic/inorganic cation (e.g., methylammonium CH 3 NH 3 + , formamidinium CH(NH 2 ) 2 + , Cs or their mixture), a divalent IV-A group metal (lead, tin or germanium) and an inorganic halide anion (I − , Br − or Cl − ). The device configurations of a perovskite solar cell (PSC) include mesoscopic/planar direct (n-i-p) and inverted (p-i-n) structures and the full printable mesoscopic structures [13]. In the mesoscopic devices, the mesoporous layer (typically TiO 2 ) acts as a scaffold by supporting the perovskite crystals. The TiO 2 layer is also used as an electron transport layer (ETL) that delivers electrons to the electrode. Holes are extracted by a p-type material and transported to the back electrode.
that delivers electrons to the electrode. Holes are extracted by a p-type material and transported to the back electrode. The typical n-i-p mesoporous structure (Figure 1a) is composed of a TCO (tin conductive oxide) cathode, a thin TiO2 compact layer to suppress the shunting paths, a mesoporous oxide layer filled with a highly crystalline perovskite absorber, a hole transporting layer (HTL) and a top metal/carbon electrode [14,15]. The HTL transports the photogenerated holes to the counter-electrode and stops the photogenerated electrons by lowering the recombinations. Organic hole transport materials (HTMs) are considered the ideal p-type semiconductor materials because of the solution process deposition by coating techniques, good hole mobility, tunable energy level and the adaptable chemical synthesis [16][17][18]. Different organic HTMs such as Spiro-OMeTAD, PTAA, PEDOT:PSS and P3HT are reported in literature [19][20][21][22][23]. Spiro-OMeTAD has a good solubility in organic solvents such as toluene and chlorobenzene to guarantee the deposition of the HTL on the perovskite material, and to reduce pinholes by improving the morphology [17,24]. The complex and expensive synthesis to realize the three-dimensional Spiro-OMeTAD core unit makes its application difficult on large area devices [25]. Moreover, Spiro-OMeTAD suffers in the presence of high temperature and moisture [16]. Recently, Bo et al. synthesized a new Spiro-compound formed by lower cost fluorene/xantene units with respect to the bi-fluorene one [17]. They applied the one-pot Buchwald-Hartwig cross-coupling synthesis to further reduce the material cost. This HTM (named X55) exhibits a high hole mobility and conductivity, an excellent 3D structure and a good film-forming feature. In the present work, we compare X55 HTM and the reference Spiro-OMeTAD on small and large area cells: X55-based devices have higher efficiency and, as we show for the first time, improved stability (ISOS-D-1) with respect to the reference case without any encapsulation.

Materials and Methods
The experimentation was conducted on small (0.09 cm 2 ) and large area (1.01 cm 2 ) cells (Figure 1b,c). In both cases we used the n-i-p structure with a mesoporous TiO2 (mp-TiO2) scaffold layer (Figure 1a).
The process is in a glove box environment except for compact TiO2 (c-TiO2) and mp-TiO2 deposition; 2.5 × 2.5 cm 2 FTO (fluorinated tin oxide)-covered glasses (Pilkington, 7 Ω/sq.) are cleaned with a soap/water solution and then in an ultrasonic bath with acetone, ethanol and 2-propanol for 10   The HTL transports the photogenerated holes to the counter-electrode and stops the photogenerated electrons by lowering the recombinations. Organic hole transport materials (HTMs) are considered the ideal p-type semiconductor materials because of the solution process deposition by coating techniques, good hole mobility, tunable energy level and the adaptable chemical synthesis [16][17][18]. Different organic HTMs such as Spiro-OMeTAD, PTAA, PEDOT:PSS and P3HT are reported in literature [19][20][21][22][23]. Spiro-OMeTAD has a good solubility in organic solvents such as toluene and chlorobenzene to guarantee the deposition of the HTL on the perovskite material, and to reduce pinholes by improving the morphology [17,24]. The complex and expensive synthesis to realize the three-dimensional Spiro-OMeTAD core unit makes its application difficult on large area devices [25]. Moreover, Spiro-OMeTAD suffers in the presence of high temperature and moisture [16]. Recently, Bo et al. synthesized a new Spiro-compound formed by lower cost fluorene/xantene units with respect to the bi-fluorene one [17]. They applied the one-pot Buchwald-Hartwig cross-coupling synthesis to further reduce the material cost. This HTM (named X55) exhibits a high hole mobility and conductivity, an excellent 3D structure and a good film-forming feature. In the present work, we compare X55 HTM and the reference Spiro-OMeTAD on small and large area cells: X55-based devices have higher efficiency and, as we show for the first time, improved stability (ISOS-D-1) with respect to the reference case without any encapsulation.

Materials and Methods
The experimentation was conducted on small (0.09 cm 2 ) and large area (1.01 cm 2 ) cells (Figure 1b,c). In both cases we used the n-i-p structure with a mesoporous TiO 2 (mp-TiO 2 ) scaffold layer (Figure 1a).
The process is in a glove box environment except for compact TiO 2 (c-TiO 2 ) and mp-TiO 2 deposition; 2.5 × 2. ending for both small and large area devices. After spin coating, the perovskite is annealed at 100 • C for 60 min. Spiro-OMeTAD (Borum, Hadsten, Denmark) and X55 (Dyenamo, Stockholm, Sweden) solutions are in chlorobenzene (73.52 and 70 mg/mL) doped with TBP (26.77 µL for both), Co salt (7.2 µL from a stock solution of 376 mg/mL in ACN) and Li-TFSI (16.66 and 16.61 µL from a stock solution of 520 mg/mL in ACN). Both the HTMs are spin coated (3000 rpm, 2 s ramp up, 30 s) in the glove box environment atop the perovskite forming the HTL. Following this, 80 nm-thick gold counter-electrode is thermally evaporated. The morphological analysis is realized with a field emission scanning electron microscope (Zeiss Auriga, Oberkochen, Germany). PL and IPCE measurement systems (Arkeo) are from Cicci Research. The photovoltaic characteristics and the MPP tracking are measured with a class A sun simulator (Sun 2000, Abet) at AM 1.5 1000 W/m 2 calibrated with an SKS 1110 sensor (Skye Instruments Ltd., Llandrindod Wells, UK); the system is equipped with a 2612 source meter (Keithley Instruments Inc., Cleveland, OH, USA) and a LabVIEW interface.

Results and Discussion
In Table 1 and Figure 2a, we show the statistics of 16 small area cells for each HTM and the J-V curves from the best-performing cells, respectively.

Results and Discussion
In Table 1 and Figure 2a, we show the statistics of 16 small area cells for each HTM and the J-V curves from the best-performing cells, respectively.  Both HTMs devices (Figure 2b) show low hysteresis index (HI = 1.02, defined as the relationship between the reverse and forward scan efficiencies [26]); however, X55-based cells have a higher Jsc with respect to Spiro-OMeTAD (21.46 vs. 20.18 mA/cm 2 ) and consequently a higher efficiency (17% vs. 16.2%). The measured Jsc is in accordance with the integrated Jsc from the IPCE (incident photon to current efficiency) spectra (Figure 2c). The performance discrepancy between X55 and Spiro-OMeTAD HTM is mainly related to the better uniform surface coverage of the perovskite layer by X55 with respect to the Spiro-OMeTAD, where pinholes are clearly observed (Figure 3). The reduction of pinholes can limit the charge recombination losses at the perovskite/HTM interface. Both HTMs devices (Figure 2b) show low hysteresis index (HI = 1.02, defined as the relationship between the reverse and forward scan efficiencies [26]); however, X55-based cells have a higher Jsc with respect to Spiro-OMeTAD (21.46 vs. 20.18 mA/cm 2 ) and consequently a higher efficiency (17% vs. 16.2%). The measured Jsc is in accordance with the integrated Jsc from the IPCE (incident photon to current efficiency) spectra (Figure 2c). The performance discrepancy between X55 and Spiro-OMeTAD HTM is mainly related to the better uniform surface coverage of the perovskite layer by X55 with respect to the Spiro-OMeTAD, where pinholes are clearly observed (Figure 3). The reduction of pinholes can limit the charge recombination losses at the perovskite/HTM interface. The effective perovskite coverage by X55 is related to the good film-forming ability because of the presence of three SFX (spiro[fluorene-9,90-xanthene]) units in the molecular structure, which grants an optimal 3D geometry of the molecule that is able to make a uniform film above the absorber material. Concerning this, Bo et al. reported the absorption and photoluminescence (PL) spectra of X55 and Spiro-OMeTAD in toluene [17]. The three SFX units involve a larger conjugated system of X55 with respect to Spiro-OMeTAD, which causes the red shift of the X55 emission band. X55 and Spiro-OMeTAD have an absorption band in the visible (peak at 406 nm) and UV (peak at 388 nm) light region, respectively. In both cases the emission bands are red shifted. As reported in literature, the charge recombination loss reduction improves the charge collection efficiency and device performance ( Figure 2) [17]. The steady state PL helps explain the hole transfer process at the perovskite-HTM interface and the higher current density with X55 HTM with respect to Spiro-OMeTAD HTM. HTM/PVSK combination lowers PL intensity with respect to the PVSK layer [27]. The interfacial hole transfer between X55 and PVSK is higher with respect to Spiro-OMeTAD and PVSK, because of the PL quenching (i.e., the PL intensity is lower than in the second case) (Figure 4).  The effective perovskite coverage by X55 is related to the good film-forming ability because of the presence of three SFX (spiro[fluorene-9,90-xanthene]) units in the molecular structure, which grants an optimal 3D geometry of the molecule that is able to make a uniform film above the absorber material. Concerning this, Bo et al. reported the absorption and photoluminescence (PL) spectra of X55 and Spiro-OMeTAD in toluene [17]. The three SFX units involve a larger conjugated system of X55 with respect to Spiro-OMeTAD, which causes the red shift of the X55 emission band. X55 and Spiro-OMeTAD have an absorption band in the visible (peak at 406 nm) and UV (peak at 388 nm) light region, respectively. In both cases the emission bands are red shifted. As reported in literature, the charge recombination loss reduction improves the charge collection efficiency and device performance ( Figure 2) [17]. The steady state PL helps explain the hole transfer process at the perovskite-HTM interface and the higher current density with X55 HTM with respect to Spiro-OMeTAD HTM. HTM/PVSK combination lowers PL intensity with respect to the PVSK layer [27]. The interfacial hole transfer between X55 and PVSK is higher with respect to Spiro-OMeTAD and PVSK, because of the PL quenching (i.e., the PL intensity is lower than in the second case) (Figure 4). Both HTMs devices (Figure 2b) show low hysteresis index (HI = 1.02, defined as the relationship between the reverse and forward scan efficiencies [26]); however, X55-based cells have a higher Jsc with respect to Spiro-OMeTAD (21.46 vs. 20.18 mA/cm 2 ) and consequently a higher efficiency (17% vs. 16.2%). The measured Jsc is in accordance with the integrated Jsc from the IPCE (incident photon to current efficiency) spectra (Figure 2c). The performance discrepancy between X55 and Spiro-OMeTAD HTM is mainly related to the better uniform surface coverage of the perovskite layer by X55 with respect to the Spiro-OMeTAD, where pinholes are clearly observed (Figure 3). The reduction of pinholes can limit the charge recombination losses at the perovskite/HTM interface. The effective perovskite coverage by X55 is related to the good film-forming ability because of the presence of three SFX (spiro[fluorene-9,90-xanthene]) units in the molecular structure, which grants an optimal 3D geometry of the molecule that is able to make a uniform film above the absorber material. Concerning this, Bo et al. reported the absorption and photoluminescence (PL) spectra of X55 and Spiro-OMeTAD in toluene [17]. The three SFX units involve a larger conjugated system of X55 with respect to Spiro-OMeTAD, which causes the red shift of the X55 emission band. X55 and Spiro-OMeTAD have an absorption band in the visible (peak at 406 nm) and UV (peak at 388 nm) light region, respectively. In both cases the emission bands are red shifted. As reported in literature, the charge recombination loss reduction improves the charge collection efficiency and device performance ( Figure 2) [17]. The steady state PL helps explain the hole transfer process at the perovskite-HTM interface and the higher current density with X55 HTM with respect to Spiro-OMeTAD HTM. HTM/PVSK combination lowers PL intensity with respect to the PVSK layer [27]. The interfacial hole transfer between X55 and PVSK is higher with respect to Spiro-OMeTAD and PVSK, because of the PL quenching (i.e., the PL intensity is lower than in the second case) (Figure 4).  The recombinations reduction can be well explained by analyzing the energy level alignment between the perovskite and the HTL. The HOMO energy levels of X55 and Spiro-OMeTAD are 5.23 and 5.13 eV, respectively [17,28]. In this condition and considering the valence band of the triple-cation perovskite used [29,30], both HTMs can extract the charge from the perovskite material, but the hole injection is favored in the X55 case because the hole contact is related to the HOMO level-X55-based devices exhibit higher Voc with respect to Spiro-OMeTAD-based cells (clearly visible in the following large area devices) [17,25,31].
One of the key factors for further development of PVSK solar technology is the upscaling from a small area cell to module size [10,32]. The first step in this direction is the fabrication of a large area cell with a size greater than 1 cm 2 [33]. We found an efficiency loss of about 5% by scaling to a large area size cell, mainly because of the fill factor (FF) ( Table 2 and Figure 5a). This drop is mainly related to the front contact sheet resistance loss due to the increased area and diminished layer heterogeneity [32,34,35]. The recombinations reduction can be well explained by analyzing the energy level alignment between the perovskite and the HTL. The HOMO energy levels of X55 and Spiro-OMeTAD are 5.23 and 5.13 eV, respectively [17,28]. In this condition and considering the valence band of the triple-cation perovskite used [29,30], both HTMs can extract the charge from the perovskite material, but the hole injection is favored in the X55 case because the hole contact is related to the HOMO level-X55-based devices exhibit higher Voc with respect to Spiro-OMeTAD-based cells (clearly visible in the following large area devices) [17,25,31].
One of the key factors for further development of PVSK solar technology is the upscaling from a small area cell to module size [10,32]. The first step in this direction is the fabrication of a large area cell with a size greater than 1 cm 2 [33]. We found an efficiency loss of about 5% by scaling to a large area size cell, mainly because of the fill factor (FF) ( Table 2 and Figure 5a). This drop is mainly related to the front contact sheet resistance loss due to the increased area and diminished layer heterogeneity [32,34,35].  The current density is about 20 mA/cm 2 , as reported in the IPCE spectra (Figure 5b). The best efficiencies are 16.05% (15% stabilized at MPP) and 14.9% (13.9% stabilized at MPP) for the X55-and Spiro-OMeTAD-based devices, respectively (Figure 5a,c). As can be noticed, the performance gap between the two HTMs is higher in the large area (8%) The current density is about 20 mA/cm 2 , as reported in the IPCE spectra (Figure 5b). The best efficiencies are 16.05% (15% stabilized at MPP) and 14.9% (13.9% stabilized at MPP) for the X55-and Spiro-OMeTAD-based devices, respectively (Figure 5a,c). As can be noticed, the performance gap between the two HTMs is higher in the large area (8%) than in the small area cells (5%). The better film-forming properties of X55 devices also show a higher V OC because of low defects and pin holes.
The commercialization of a photovoltaic technology is possible only if it is stable during the lifetime [9,36]. In literature, small area devices with X55 HTM were tested at room temperature (RT) and 20% humidity. Here, we checked the stability (Figure 6) of Energies 2021, 14, 6081 6 of 8 nonencapsulated large area devices for more than 1000 h in an ambient condition (ISOS-D-1, relative humidity of around 50%) [37].
Energies 2021, 14, x FOR PEER REVIEW 6 of 8 than in the small area cells (5%). The better film-forming properties of X55 devices also show a higher VOC because of low defects and pin holes. The commercialization of a photovoltaic technology is possible only if it is stable during the lifetime [9,36]. In literature, small area devices with X55 HTM were tested at room temperature (RT) and 20% humidity. Here, we checked the stability (Figure 6) of nonencapsulated large area devices for more than 1000 h in an ambient condition (ISOS-D-1, relative humidity of around 50%) [37]. After 1000 h, the efficiency drop is less than 15% for the X55 HTM and more than 30% for the Spiro-OMeTAD HTM. Moreover, the X55-based cell has no HI change during the test (from 1.02 to 1.05), while cells with Spiro-OMeTAD experience a large variation of the hysteresis cell between the beginning (HI = 1.02) and the end (HI = 1.3) of the test. Analyzing the shape of the curves [38], we estimate a T80 of 497 h for the reference case (PCE = 10.6%); X55 HTM is still above 80% of PCE (13.35%) after 1000 h. The high uniformity and homogeneity and the low recombination rate of the X55 HTL are the main causes of the better stability with respect to the reference case [17]. Furthermore, the two methoxy groups in the X55 structure have a higher hydrophobicity feature relative to the Spiro-OMeTAD HTM. Consequently, the moisture presence does not have a detrimental effect on the device performance.

Conclusions
The HTL plays a crucial role in terms of efficiency, stability and cost of PSCs. Devices based on the expensive Spiro-OMeTAD HTM have very high efficiency but poor stability in presence of humidity. In this work, we experimented with the cost-effective X55 HTM, which has a much deeper HOMO level and a higher hole mobility and conductivity than Spiro-OMeTAD. The results show a higher efficiency (17% vs. 16%), film homogeneity and better hole transport for the X55 with respect to the reference Spiro-OMeTAD. Then, we tested and publish for the first time the promising X55 HTM on large area devices during MPP tracking and under ISOS-D-1 stability test for more than 1000 h. On large area cells, the best efficiency is 16.05% and 14.9% for X55 and Spiro-OMeTAD HTM, respectively. We then found a stabilized efficiency at MPP of 15% for X55 and 13.9% for Spiro-OMeTAD. The efficiency drop was about 15% and 30% for the X55-and Spiro-OMeTAD-based device after 1000 h at ISOS-D-1 condition, respectively. The sturdiness relative to the moisture and the low upscaling losses make the X55 HTM a potential candidate for stable large area modules. After 1000 h, the efficiency drop is less than 15% for the X55 HTM and more than 30% for the Spiro-OMeTAD HTM. Moreover, the X55-based cell has no HI change during the test (from 1.02 to 1.05), while cells with Spiro-OMeTAD experience a large variation of the hysteresis cell between the beginning (HI = 1.02) and the end (HI = 1.3) of the test. Analyzing the shape of the curves [38], we estimate a T 80 of 497 h for the reference case (PCE = 10.6%); X55 HTM is still above 80% of PCE (13.35%) after 1000 h. The high uniformity and homogeneity and the low recombination rate of the X55 HTL are the main causes of the better stability with respect to the reference case [17]. Furthermore, the two methoxy groups in the X55 structure have a higher hydrophobicity feature relative to the Spiro-OMeTAD HTM. Consequently, the moisture presence does not have a detrimental effect on the device performance.

Conclusions
The HTL plays a crucial role in terms of efficiency, stability and cost of PSCs. Devices based on the expensive Spiro-OMeTAD HTM have very high efficiency but poor stability in presence of humidity. In this work, we experimented with the cost-effective X55 HTM, which has a much deeper HOMO level and a higher hole mobility and conductivity than Spiro-OMeTAD. The results show a higher efficiency (17% vs. 16%), film homogeneity and better hole transport for the X55 with respect to the reference Spiro-OMeTAD. Then, we tested and publish for the first time the promising X55 HTM on large area devices during MPP tracking and under ISOS-D-1 stability test for more than 1000 h. On large area cells, the best efficiency is 16.05% and 14.9% for X55 and Spiro-OMeTAD HTM, respectively. We then found a stabilized efficiency at MPP of 15% for X55 and 13.9% for Spiro-OMeTAD. The efficiency drop was about 15% and 30% for the X55-and Spiro-OMeTAD-based device after 1000 h at ISOS-D-1 condition, respectively. The sturdiness relative to the moisture and the low upscaling losses make the X55 HTM a potential candidate for stable large area modules.

Data Availability Statement:
The data presented in this study are available on request from the corresponding authors.