Effect of Codoping Cl Anion and 5-AVA Cation on Performance of Large-Area Perovskite Solar Cells with Double-Mesoporous Layers

For the perovskite solar cells (PSCs), the performance of the PSCs has become the focus of the research by improving the quality of the perovskite absorption layer. So far, the performance of the large-area PSCs is lower than that of small-area PSCs. In the paper, the experiments were designed to improve the photovoltaic performance of the large-area PSCs by improved processing technique. Here we investigated the optoelectronic properties of the prototypical CH 3 NH 3 PbI 3 (MAPbI 3 ) further modulated by introducing other extrinsic ions (specifically codoped Cl and 5-AVA). Moreover, we used inorganic electron extraction layer to achieve very rapid photogenerated carrier extraction eliminating local structural defects over large areas. Ultimately, we fabricated a best-performing perovskite solar cell based on codoping Cl anion and 5-AVA cation which uses a double layer of mesoporous TiO 2 and ZrO 2 as a scaffold infiltrated with perovskite and does not require a hole-conducting layer. The experiment results indicated that an average efficiency of double-mesoporous layer-based devices with codoping Cl anion and 5-AVA cation was obtained with exceeding 50% enhancement, compared to that of pure single-mesoporous layer-based device.


Introduction
Organic-inorganic metal halide perovskite solar cells, particularly the MAPbX 3 (where MA denotes methylammonium cations and X = I, Cl, and Br) perovskite as a light harvesting material, have recently attracted tremendous attention due to their low cost and outstanding desirable properties [1].Currently, continuous efforts will be focused on the investigation of the high efficiency and stability of the CH 3 NH 3 PbX 3 PSCs with electron or hole transport layer.However, employing perovskites with mixed cations and halides has also become significant because the pure perovskite compounds suitable for photovoltaic applications, which are mainly MAPbX 3 , FAPbX 3 (where FA denotes formamidinium cations), and CsPbX 3 (where Cs refers to inorganic cesium), possess numerous disadvantages.Cl incorporation into the perovskite crystal structure was initially investigated by Colella et al. [2].They found that the Cl doping dramatically improves the charge transport within the perovskite layer.Moreover, it was reported that adding PbCl 2 to the precursor solution could influence the morphology of perovskite, resulting in an increase of the power conversion efficiency (PCE) [3][4][5].In 2015, Chen et al. found that the chlorine incorporation mainly improved the carrier transport across the heterojunction interfaces, rather than within the perovskite crystals, leading to a PCE of 17.91% [6].In addition, Suzuki and coworkers investigated the role of halogen doping using iodine, bromine, and chlorine compounds as dopant on the photovoltaic performance and microstructures of CH 3 NH 3 PbI 3−− Br  Cl  perovskite solar cells [7].Materials property evolution has been recorded when replacing lead with tin, where the absorption edge can be successfully redshifted to over 1000 nm in CH 3 NH 3 Sn  Pb (1−) I 3 perovskites [8].Regarding the site of MA cation, 5-AVA replaced part of the MA cations in the cuboctahedral site of MAPbI 3 , forming the mixed cation (5-AVA)  (MA) 1− PbI 3 perovskite reported by Mei et al. [9].The FA 1− MA  Pb(I 1− Br  ) 3 perovskite was prepared to further improve the performance of perovskite based photovoltaics [10].In addition, the group of McMeekin fabricated PSCs that reached open-circuit voltages of 1.2 volts and power conversion efficiency of over 17% on small area [11].Bi et al. also produced the perovskite films in a single step from a solution containing a mixture of FAI, PbI 2 , MABr, and PbBr 2 [12].To date, Saliba et al. presented the notion that the use of all three cations (MA, FA, and Cs) offers additional versatility in tuning high quality perovskite films.The final result shows that the combination of different cations can combine the merits of the components while avoiding their flaws [13].In addition, a double layer of mesoporous TiO 2 and ZrO 2 as a scaffold infiltrated with perovskite was used [9].Liu et al. fabricated solar cells based on TiO 2 /ZrO 2 /carbon films with efficiency exceeding 13% [14].
Currently, PSCs typically require an expensive and airsensitive hole transporter (e.g., spiro-MeOTAD) and a noble metal electrode (Au or Ag).To tackle high manufacturing cost, Ku and coworkers proposed for the first time that carbon black/graphite can substitute noble metallic materials as an efficient counter electrode in the application of holeconductor-free CH 3 NH 3 PbI 3 perovskite/TiO 2 heterojunction solar cells [15].Wei et al. employed abundant candle soot as an efficient hole extractor and reported a new strategy of clamping solar cells by judiciously interfacing the candle soot with CH 3 NH 3 PbI 3 films [16].Subsequently, a planar perovskite solar cell that incorporates a nanocarbon holeextraction layer is fabricated for the first time by an inkjet printing technique with a precisely controlled pattern and interface [17].Among the various perovskite solar cells, the carbon-based hole-transporter-free type is the most promising for applications, owing to its low materials and manufacturing cost, high efficiency, and best long-term stability.However, the fill factor is typically below 0.75, limiting the cell performance.Therefore, Wei et al. highlighted two outstanding advantages for multiwall-carbon-nanotube-based devices: high FF and hysteresis-free performance [18].Thus far, for carbon-based HTM-free PSCs, the highest power conversion efficiency of 14.38% is obtained, which is a record value [19].In addition, Chan et al. proposed that the carbon layer as a p-type electrode was fabricated by screen-printing the carbon paste, which was prepared by mixing graphite, carbon black, Al 2 O 3 , ethyl cellulose, and alpha-terpineol [20].Here, we adopted facile method and low-cost screen-printing carbon paste to fabricate the carbon layer as a back electrode.
Although numerous efforts have been concentrated on the investigation of small-area PSCs, the investigation of PSCs based on doping and without hole transport layer at an aperture area >1 cm 2 is insufficient.Chen et al. fabricated the large-area PSCs with an aperture area >1 cm 2 showing the highest PCE > 15% [21].At present, utilizing films fabricated by vacuum-flash solution processing method, Li et al. demonstrated solar cells with an aperture area exceeding 1 cm 2 showing a maximum efficiency of 20.5% and a certified PCE of 19.6% [22].
Here we fabricate large-area PSCs with single-mesoporous layers using doping of the different elements to investigate the role of the extrinsic ion in the context of optoelectronic properties.The experiments indicate the performance of codoping perovskite is higher than that of singledoping perovskite.Furthermore, in order to compare with the performance of single-mesoporous-based PSCs with codoping, double-mesoporous-based PSCs with codoping are fabricated.Finally, an average efficiency of double-mesoporous layer-based devices with codoping Cl anion and 5-AVA cation was obtained with exceeding 50% enhancement, compared to that of pure single-mesoporous layer-based device.

Experiment and Methods
2.1.Materials.Unless stated otherwise, all materials were purchased from Sinopharm Chemical Reagent Beijing Co., Ltd., or Shanghai MaterWin New Materials Co., Ltd., and used as received.ZrO 2 printing paste was purchased from Wuhan National Laboratory for Optoelectronics.And carbon paste for screen-printing was purchased from Guangzhou Seaside Technology development Co., Ltd.

The Preparation of Perovskite Precursor Solution.
The Cl anion was introduced into the perovskite lattice with a mixture of MAI and MACl in the precursor solution with the mole ratio of 4 : 1; that is, MAI: MACl precursor was prepared by mixing MAI and MACl in 2-propanol with the mole ratio (4 : 1) described elsewhere [23].And the concentration of MA + in the precursor solution for doping Cl − and undoping was kept at a 1 : 1 molar ratio.Li et al. [24] and Mei et al. [9] fabricated the (5-AVA)  (MA) 1− PbI 3 precursor solution with a mixture of 5-AVAI and MAI of the mole ratio between 1 : 20 and 1 : 30.In the present work, the perovskite based on doping 5-AVA + precursor solution was fabricated in the same procedure, except that a mixture of MAI and HOOC(CH 2 ) 4 NH 3 I (5-AVAI) with the mole ratio of 1 : 19 was used instead of MAI solution to introduce 5-AVA during the film growth.And the concentration of iodide anions in the precursor solution with/without doping 5-AVA + was maintained at a 1 : 1 molar ratio.The (5-AVA)  (MA) 1− PbI 3− Cl  precursor solution was prepared in the same procedure, except that a mixture of 5-AVAI, MACl, and MAI with the mole ratio of 1 : 4 : 15 was used.The total molar concentration is equal to the concentration of MAI (10 mg/mL).

Device Fabrication.
The fluorine-doped tin oxide (FTO) conductive glasses were cleaned with detergent in an ultrasonic bath and rinsed with deionized water, followed by washing with the mixture of deionized water, ethanol, and acetone at 1 : 1 : 1 by volume, and they suffered from an ultraviolet treatment for one hour.The compact TiO 2 layer (cp-TiO 2 ) was spin-coated on the cleaned FTO conductive glasses at 2000 rpm for 60 s and annealed at 125 ∘ C for 5 min and then sintered at 500 ∘ C for 30 min.
In our trials, for the preparation of mesoporous layer, the following two steps were performed.(1) Single-mesoporous layer: TiO 2 paste was diluted by ethanol at 1 : 3.5 by weight and deposited on the FTO/cp-TiO 2 substrate by spin-coating method at 2000 rpm for 60 s and annealed at 125 ∘ C for 5 min and then sintered at 500 ∘ C for 30 min to form mesoporous TiO 2 (mp-TiO 2 ) layer.(2) Double-mesoporous layer: ZrO 2 paste was diluted by ethanol at 1 : 2 by weight and spin-coated on the FTO/cp-TiO 2 /mp-TiO 2 substrate at 2000 rpm for 60 s and annealed at 125 ∘ C for 5 min, and then sintered at 500 ∘ C for 30 min.
The substrates were preheated at 70 ∘ C on a hot plate for 15 min.Then, the PbI 2 solution (1 mol/L) keept at 70 ∘ C was spin-coated on the top of mesoporous layer at 5500 rpm for 60 s, followed by thermal annealing at 100 ∘ C for 30 min.Then, MAI (dissolved in 2-propanol) and a mixture of MAI/MACl or MAI/5-AVAI or MAI/MACl/5-AVAI were spin-coated on top of the dried PbI 2 layer in a two-step program at 0 rpm and 5500 rpm for 30 s and 60 s, respectively.Then, the films were dried at 100 ∘ C for 30 min.Finally, a carbon counter electrode was screen-printed on the top of perovskite layer, employing carbon slurry diluted with cyclohexanone to 85 wt% and annealed at 120 ∘ C for 15 min.Then, the solar cells were hotpressed at 120 ∘ C for 5 s and pressure was set to 0.2 MPa.The entire fabrication processes of PSCs were carried out under humidity of about 30% and outside the glove box.The active area is about 1.5 ± 0.2 cm 2 .

Characterization.
The current density-voltage (J-V) curves were characterized with a VersaSTAT3 source meter under standard air-mass 1.5 global (AM1.5G)illumination (100 mW/cm 2 ) provided by a solar simulator (Oriel Sol 3A).The scanning electron microscopy (SEM) images were performed with an S-4300 (CARL ZEISS, Germany) to observe the surface morphology of thin films.X-ray diffraction (XRD) was used to analyze the crystallinity and size of sample.The XRD measurement was obtained by Bruker D8 focus (Bruker Corporation, Germany).The analysis of element was detected by Energy Dispersive X-Ray Spectroscopy (EDS) (model SIGMA, CARL ZEISS, Germany).The ultravioletvisible (UV-vis) absorption spectra were obtained by the test system, which consisted of AvaSpec-ULS2048XL, test software, AvaLight-DH-S-BAL as a light source, AvaSpere-50-REFL, AvaSphere-IRRAD, and so on.The monochromatic incident photon-to-electron (IPCE) was measured with an IPCE measurement system (Institute of Physics, CAS, China) in our laboratory.

Device Performance of PSCs Based on Doping Perovskite
Layers.In this paper, we research the role of the different doping extrinsic ion (undoped, doped Cl − , doped 5-AVA + , and codoped Cl − and 5-AVA + ) on the perovskite film layers.Device performance is characterized by current densityvoltage (J-V) measurements under standard AM1.5G illumination (100 mW/cm 2 ) under atmospheric condition, with statistical distribution shown in Figure 1 and the mean values on a batch of seven photovoltaic devices were illustrated in Table 1.As shown in Figure 1, reference devices, that is, without Cl incorporation (w/o), obtain short-circuit current density ( SC ) in the range of 3.62-5.14mA/cm 2 , open-circuit voltages ( OC ) in the range of 563-808 mV, and fill factors (FF) in the range of 28%-33%, resulting in the efficiency ranging from 0.61% to 1.31%.By comparison, better performance The enhanced performance of devices using doping Cl − and codoping in terms of  SC ,  OC , and FF is possibly related to the reduction of both parasitic current loss and series resistance [6].The results indicate that doping is beneficial to improving the performance of the PSCs.

Effect of the Doped Cl
− on the Performance of the PSCs.The J-V curves for best-performing PSCs with/without Cl incorporation are shown in Figure 2. The doping Cl − based perovskite devices show greatly improved performance, with  SC ,  OC , and FF reaching values of 4.90 mA/cm 2 , 763 mV, and 40%, respectively, corresponding to a PCE of 1.43%.However, the optimal PSCs without doping produced  SC of 5.14 mA/cm 2 ,  OC of 808 mV, and FF of 33%, yielding an efficiency of 1.31%.The reason may be that the Cl incorporation improves the carrier transport across the heterojunction interfaces, so as to suppress the recombination of interfaces, where consequently the efficiency is improved; that is, the devices with Cl incorporation present the reduction of both parasitic current loss and series resistance, which is in good accordance with the relatively high FF obtained for this device [6].It is inferred that the improved average FF is associated with the Cl incorporation.SEM measurement was conducted to characterize the morphological evolution of the perovskite films (Figures 3(a) and 3(b)).Compared with the w/o doped perovskite, it is noted that the perovskite size based on doping Cl − becomes slightly larger.In order to gain insight into the effect of Cl incorporation, XRD analysis was performed to characterize the crystal structure of MAPbI 3 perovskites with Cl incorporation prepared on the FTO/cp-TiO 2 /mp-TiO 2 substrates (Figure 4).As shown in Figure 4, the diffraction peak of perovskite based on doping Cl − and w/o doping at 14.09 ∘ is assigned to the (110) lattice plane.As a comparison, International Journal of Photoenergy it is found that the full width at half maximum (FWHM) of perovskite based on doping Cl − with (110) diffraction peak becomes narrow; this indicates that the size of MAPbI 3 perovskite with doping Cl − increases.This is in agreement with the analysis of SEM measurement.In addition, reducing the peak intensity of residual PbI 2 of the perovskite based on doping Cl − is also observed.The result shows that the conversion of PbI 2 into the perovskite with doping Cl − increases.EDS measurement is also performed to further analyze the Cl element incorporation into the perovskite lattice (Figure 3(c)).The combination of XRD and EDS indicates that the Cl element is introduced into the MAPbI 3 perovskite lattice.
In addition, both the UV-vis spectra and IPCE curves of the devices with/without doping Cl − are presented in Figures 5(a) and 5(b).The corresponding integrated photocurrent calculated from the IPCE spectra with the AM1.5 solar emission are also shown in Figure 5(b).As shown in Figure 5(a), it is noted that the absorbance of the perovskite film with doping Cl − is slightly higher than that of the devices without doping Cl − in the range from 400 nm to 750 nm, which is consistent with the previous report [2].The quantum efficiency in the IPCE of the PSCs based on doping Cl − has a slight improvement, corresponding to the trend of UV-vis spectra.perovskite.Moreover, compared with the perovskite based on no doping, although the  SC and  OC of the device with doping 5-AVA + are lower, the efficiency is higher, which is related to the improvement of fill factor.The reason may be that the defect concentration of newly emerging (5-AVA)  (MA) 1− PbI 3 perovskite is lower than that of pure MAPbI 3 perovskite, leading to the reduction of internal recombination [9].The XRD patterns of pure MAPbI The absorption intensity of (5-AVA)  (MA) 1− PbI 3 perovskite increases between 380 nm and 630 nm.Furthermore, the IPCE curves of the device with/without doping 5-AVA + are plotted in Figure 8(b).As shown in Figure 8(b), the change of external quantum efficiency is in keeping with that of UV-vis spectra.And the integrated photocurrent from IPCE measurement based on doping 5-AVA + is higher than that of the device without doping.The result is in agreement with the J-V measurement.

Effect of the Codoped Cl Anion and 5-AVA Cation on
the Performance of the PSCs.The J-V curves measured at a simulated AM1.5 solar irradiation of 100 mW/cm 2 for the best-performing PSCs based on doping Cl − , 5-AVA + , and codoping are plotted in Figure 9(a).It is found that the photovoltaic performance of perovskite based on codoping is the best, indicating that the Cl − and 5-AVA + coincorporation has a positive effect.The small addition of chloride ions to the pure MAPbI 3 results in such a remarkable increase in the electron-hole diffusion length, predominantly arising from substantial inhibition of nonradiative electron-hole recombination; that is, the larger diffusion length in the mixed halide perovskite results from a much longer recombination lifetime [25].de Quilettes et al. have shown employing confocal fluorescence microscopy correlated with scanning electron microscopy [26] where the electron-hole recombination rate is strongly reduced in the presence of grain boundaries (GBs), while it is increased with Cl doping.This experiment clearly suggests that GBs are detrimental to perovskite solar cell performance due to accelerated electronic energy loss to heat.
International Journal of Photoenergy   To clarify the role of GBs and GB doping on the electronhole recombination within MAPbI 3 perovskites, a real-time atomistic simulation is needed to mimic the electronicvibrational dynamics directly as they occur in experiments.Long and coauthors presented the notion that GBs greatly accelerate the recombination, while introduction of substitutional Cl dopants at the boundary reduces the recombination, making it even slower than in the pristine MAPbI 3 [27].It is found that the carrier transport across interface between perovskites and carrier transport layers is improved through the Cl incorporation, rather than within the perovskite film [6].With regard to the influence of organic cation doping, Mercier [28] employed 4-ammonium-butyric acid (4-ABA) as a template for the engineering of (4-ABA) 2 PbI 4 and mixed-cation (4-ABA) 2 MAPb 2 I 7 perovskite crystals.The ABA molecules formed linear hydrogen-bonded chains that serve as supramolecular synthons for layered plumbo iodide perovskite structures [28].Mei et al. presented that the 5-AVA influences the crystal growth of (5-AVA)  (MA) 1− PbI 3 in a similar fashion through the formation of hydrogen bonds between its COOH and NH 3 + groups and I − ions from the PbI 6 octahedra [9].
In order to explore the role of Cl and 5-AVA in perovskite film, the IPCE curves of the devices based on doping Cl − , 5-AVA + , and codoping are shown in Figure 9(b).Compared with the devices of doping Cl − or 5-AVA + , the external quantum efficiency of the devices based on codoping Cl − and 5-AVA + is remarkably improved in the range of from 380 nm to 600 nm.The integrated photocurrent from IPCE measurement based on the codoping is higher than that of the devices based on single-doping.The results are in agreement with those of J-V measurements.scaffold than device with porous TiO 2 is beneficial for higher loading amounts of (5-AVA)  (MA) (1−) PbI (3−) Cl  perovskite [14].And the integrated photocurrent calculated from IPCE measurement of the (5-AVA)  (MA) (1−) PbI (3−) Cl based perovskite based on double-mesoporous layer is higher than that of single-mesoporous layer.The result is consistent with that of J-V measurement.Furthermore, the UVvis spectra of single-/double-mesoporous layer-based device with codoping Cl anion and 5-AVA cation are provided in Figure 12.As a comparison, the absorption intensity of the device based on double-mesoporous layer is lower than that of the device with single-mesoporous layer.This indicates that the light transmittance of the perovskite with doublemesoporous layer is weaker than that of single-mesoporousbased perovskite.The reason may be related to increasing the thickness of mesoporous layer, leading to increasing the internal recombination.However, the quantum efficiency in the IPCE of double-mesoporous-based device is higher than that of single-mesoporous-based device in the range of 380 nm to 550 nm.It is attributed to the insulation properties of ZrO 2 layer, resulting in avoiding the direct contact between carbon counter electrode and the cp-TiO 2 layer.

International Journal of Photoenergy
A hysteresis effect often occurs during measurement of the J-V curves of perovskite solar cells.The typical J-V characteristics of (5-AVA)  (MA) (1−) PbI (3−) Cl  PSCs with single-and double-mesoporous layer at different scanning directions with a scan rate of 0.1 V/s are obtained under AM1.5 illumination (Figure 13).In terms of the scan direction, by comparison, no significant J-V hysteresis is observed in the device with double-mesoporous layer, which may be associated with increasing the thickness of mesoporous layer [29].

Conclusions
In summary, in our experiment, we investigated the role of the doping and single-/double-mesoporous layer in modifying the perovskite layers.We fabricated the best-performing International Journal of Photoenergy perovskite solar cell based on codoping Cl anion and 5-AVA cation which uses a double layer of mesoporous TiO 2 and ZrO 2 as a scaffold infiltrated with perovskite and does not require a hole-conducting layer.An average efficiency of double-mesoporous layer-based devices with codoping Cl anion and 5-AVA cation was obtained with exceeding 50% enhancement, compared to that of pure single-mesoporous layer-based MAPbI 3 , due to the positive effect of codoping.Furthermore, the ZrO 2 with insulation property is blocking the flow of photogenerated electrons to the back contact, suppressing recombination with the holes from the perovskite at the back contact.In a word, the final experiment results are in agreement with the expectation, and it is confirmed that the performance of perovskite solar cell can be further improved by codoping.Continuous efforts will be concentrated on the investigation of charge transfer behaviors to further improve the device performance.For example, besides Cl and 5-AVA, other extrinsic ions for codoping can be incorporated into the perovskite solar cells by simply changing the reactants.

International 2 )Figure 2 :Figure 3 :
Figure 2: J-V characterization of the optimal PSCs with/without Cl incorporation.

Figure 4 : 2 )Figure 5 :
Figure 4: XRD patterns of perovskite films fabricated with different approaches (from top to bottom: doped Cl − and w/o doped).

Figure 11 :
Figure 11: The IPCE curves and integrated photocurrent curves of the best-performing codoping-based PSCs with single-and doublemesoporous layer.

Figure 12 :
Figure 12: The UV-vis spectra of the device based on single-and double-mesoporous layer.

Figure 13 :
Figure 13: Hysteresis behavior of perovskite solar cells.The typical J-V characteristics of (5-AVA)  (MA) (1−) PbI (3−) Cl  -based PSCs with (a) single-and (b) double-mesoporous layer at different scanning directions: from short circuit to forward bias (black) and from forward bias to short circuit (red).

Table 1 :
Experiment average photovoltaic parameters of singlemesoporous PSCs devices based on different doping.SC in the range of 3.53-4.9mA/cm 2 ,  OC in the range of 683-763 mV, and FF in the range of 33%-40%, yielding the PCE ranging from 0.93% to 1.43%.Compared with the reference devices, the devices based on doping 5-AVA + have the  SC value of 4.05-4.96mA/cm 2 ,  OC value of 656-761 mV, and FF value of 32%-38%, resulting in the PCE ranging from 0.92% to 1.42%.The best-performing PSCs for codoping Cl − and 5-AVA + devices produce the  SC value of 4.41-6.00mA/cm 2 ,  OC in the range of 661-767 mV, and FF in the range of 32%-41%, yielding the PCE ranging from 1.00% to 1.76%.As shown in Table1, the average performance for devices based on doping Cl − and codoping is superior to the devices based on no doping.