Recent Advances in Fabrication Techniques of Perovskite Solar Cells: A Review

Corresponding Author: Mukter Zaman Faculty of Engineering (FOE), Multimedia University, 63100 Cyberjaya, Malaysia Tell: +6010-5391581 Email: mukter@mmu.edu.my Abstract: In this study, thorough review is performed on various reported fabrication techniques of the Perovskite Solar Cell (PSC) and the Hole Transport Materials (HTM) layer, in order to provide a critical insight on the direction for further advancement in related fields as well as a better understanding on the device physics. So far, the best short-circuit current density of 22.9 mA/cm 2 has been reported in CH3NH3PbI3-xClx scaffold by TiCl4 nanorods treated by NRS and the best open circuit voltage of 0.763V has been reported in TiO2/perovskite/spiro-OMeTAD/Au. On the other hand, so far the highest PCE of the PSCs reported in the literature is 19.3%. Interestingly, PSC with 200nm mesoporous TiO2 has been reported to achieve PCE value of 16.46%. A 5-layer TiO2/perovskite/spiroOMeTAD/Au integrated with 40 nm ZnO thin film has been reported to achieve Fill Factor (FF) of 76%. Although promising data has been obtained through numerous researches, there are still many other optimizations yet to be done before mass commercialization of PSC is possible. Optimizing the fabrication process of carrier transport material as well as the incorporation of sensitizer into the device structure are the key aspects that should be focused on in future researches.


Benchmarking of PSC Research and Development
Perovskite is a special type of solar cell material, named after the Russian mineralogist, L. A. Perovski (1792A. Perovski ( -1856 and is often considered as one of the most promising types of next-generation solar cell material. PSC has the perovskite crystal structure, which is shown in Fig. 1a and generally consists of organic and inorganic materials. The organic material converts solar energy to electrical energy, while the inorganic material provides the crystal structure. Generally, the PSC device is fabricated on top of a conductive substrate. The device structure consists of a layer of Electron Transport Material (ETM), followed by a perovskite sensitizer which converts photons into electrons and holes, a layer of Hole Transport Material (HTM) and finally the electrode layers to complete the circuit externally with the conductive substrate. A typical device structure of a PSC is shown in Fig. 1b. Despite the considerable amount of speculations that arose when PSC was first introduced in 2009, PSC is now considered to have a promising future because of its high energy conversion efficiency at lower cost than conventional solar cells, as well as its high versatility which allows it to be coated on various types of substrate materials. Although the commercialization progress of PSC is still currently in its infancy stage, its fabrication methods are found to be amazingly versatile, as there are many proven alternatives to fabricate PSC successfully.
In terms of energy conversion efficiency, PSC has been improving dramatically in the last few years. Researchers have demonstrated that the energy conversion efficiency of PSCs could go up to 20%. Nevertheless, stability of PSC remains as a critical concern. While conventional silicon-based solar cells can last for up to a quarter century without significant degradation in its performance, PSC with the best performance reported so far is only capable to maintain optimal performance for up to 100 h (Alharbi et al., 2014;Peplow, 2014). The relatively poor stability of PSC is because perovskite compound, which is water soluble, would break down easily in the presence of sunlight. Therefore, the current challenge in the development of PSCs is the stability issue rather than performance issue.  (Green et al., 2014) In this review, the fundamental working principle of PSC will be discussed along with all the reported possible combinations of PSC components, followed by different fabrication techniques of the PSC and HTM layers, as well as the synthesis processes of the PSC components reported in the literature. A critical review on different reported conditions such as but not limited to various atmosphere condition, thin film properties, HTM, electron transport, external electric field, blocking layer, transport layer, CBO and VBO absorber layer, heat treated ion assisted film, thermal annealed effect, reduced graphene oxide, blade coating and low temperature solution will be provided systematically. This review would be useful to benchmark all the research activities for PSC in order to obtain an insight for future studies on related fields.

Overview of Synthesis Methods of PSC
Different types of perovskite can be fabricated in different methods, among which fabrication methods of CH 3 NH 3 PbI 3 (iodine based), CH 3 NH 3 Sn 0.5 Pb 0.5 I 3 , CH 3 NH 3 SnI 3 , CH 3 NH 3 PbBr 3 (bromine based) , K 2 La 2 Ti 3 O 10 , La 0.7 Sr 0.3 MnO 3 , hybrid PSC with a planar heterojunction structure have been reported as the most promising methods and will therefore be explained in this section. This section will end with a summary of the different fabrication methods discussed.

Synthesis of CH 3 NH 3 PbI 3 Sensitizer (Iodine-based)
The most well-known perovskite compound is CH 3 NH 3 PbI 3 , which can be fabricated by mixing PbI 2 and CH 3 NH 3 I. In order to produce the CH 3 NH 3 I compound, methylamine (30%) is mixed with HI (45%) in ice-bath for few hours so that the chemicals can react effectively (Jeon et al., 2014). The solution also can be synthesized under nitrogen atmosphere at 0°C . After that, the evaporation of the methylamine and HI solvent will be conducted at high temperature. The remaining solid is dissolved in ethanol and recrystallized in diethyl ether. From this process, CH 3 NH 3 I compound which is pure-white in color (Xin et al., 2014;Sheikh et al., 2015) would be obtained. It can also be synthesized by adding 24 mL of methylamine and 10 mL of hydroiodic acid with 100 mL of ethanol with continuous stirring for 2 h in ice bath. With a help of rotary evaporator at 50°C to exclude the solvent, the precipitate is collected. Later on the product is recrystallized in ethanol and followed by the filtering and washing the end product with diethyl ether. Next step is drying for 24 h at 60°C in vacuum oven (Dkhissi et al., 2015;Li et al., 2015). CH 3 NH 3 PbI 3 can be fabricated by dissolving CH 3 NH 3 I and PbI 2 in gamma-butyrolactone and maintaining uniform stirring for 24 h. Next, the solution is deposited on top of the substrate by spin coating and heated up for 30 min at 100°C. Another method to synthesize CH 3 NH 3 PbI 3 is by dissolving PbI 2 in N, Ndimethylformamide at 70°C  and then spin coating it on top of the substrate. The layer is then dried by heating the substrate at 70°C for 15 min. Then, the substrate is dipped into 2-propanol solution that contains CH 3 NH 3 I to form CH 3 NH 3 PbI 3 layer (Della Gaspera et al., 2015). Another method to fabricate perovskite CH 3 NH 3 PbI 3 is by adding 30 mL of hydrochloric acid with methylamine and stirring the solution without any interruption for 2 h. Then, the precipitates of CH 3 NH 3 I are collected by distillation at 50°C under reduced pressure and washed with diethyl ether. Next, the precipitates are recrystallized using ethanol and diethyl ether and dried in a vacuum oven at 70°C for 24 h. The purified CH 3 NH 3 I powder is dissolved in gamma-butyrolactone with PbI 2 overnight at 60°C with continuous stirring to form the precursor (Xiao et al., 2014). In addition to the perovskite structure, another compound, known as CH 3 NH 3 PbI 3- x Cl x , has been reported in similar context. It can be fabricated with Methylamine Iodide (MAI) to PbCl 2 molar ratio of 3:1, where the mixed solution is stirred for 24 h at 60°C and then added into the prepared CH 3 NH 3 PbI 3 Sheikh et al., 2015).
Several performance curves of the CH 3 NH 3 PbI 3 PSC with varying film thicknesses, when illuminated from the FTO side, are shown in Fig. 2. From the I-V curve CH 3 NH 3 PbI 3 PSC with 290 nm film thickness is reported to have the highest current density compared to PSC with other film thicknesses. At the same time, the 290 nm PSC also showed highest IPCE and lowest transmittance at wavelength ranging from 550 to 800 nm (Della Gaspera et al., 2015).

Synthesis of Cocktail PSC
The substrate material used to fabricate cocktail perovskite is FTO glass/compact. First, the patterning and cleaning of the substrate is carried out. Then, by using spray pyrolysis method titanium(IV) cetylacetonate is coated on top of the substrate at 470°C, followed by dipping the substrate in TiCl 4 aqueous solution at 70°C for 30 min, before heating the substrate in air at 500°C. The deposition of TiO 2 nanoparticles on the substrate can be done by spin coating method, followed by 30 min of heating in air at 500°C. The next step is conducted under N 2 atmosphere to prepare the perovskite layer. Mixture of SnI 2 /PbI 2 and CH 3 NH 3 I (40 wt%, 1:1/mol) in DMF is spin coated and heated at 60°C for 30 min on a porous titania layer. Polymers are dissipated by spin coating on the substrate to employ the hole transporting layer P3HT. The flow diagram of the whole experiment procedure is shown in Fig. 3a (Ogomi et al., 2014;Bradley, 2015). Figure 3b shows the stability of this perovskite in the air with and without PbI 2 . Without PbI 2 , a rapid decrement in the 900 nm absorption in air can be observed. Eventually, the absorption drops from 1.0 to 0 in 60 min. Meanwhile, when PbI 2 is incorporated in the compound, the 900 nm absorption is relatively stable in air. The efficiency of this PSC is reported to be 14.4% (Ogomi et al., 2014).

Synthesis of CH 3 NH 3 PbBr 3 (Bromine-based)
The synthesis of CH 3 NH 3 PbBr 3 starts by mixing 48 mL of methylamine with 57 mL of hydrobromic acid for 2 h with constant stirring. After that, by evaporating the solution for 1 h at 50°C, the precipitate can be recovered. The arrows in Fig. 4a indicates the transfer of charge carriers, namely electrons and holes, inside the compound structure. The inset images in Fig. 4b and 4c are showing the absorption and room temperature photoluminescence (excited at 325 nm) spectra of the CH 3 NH 3 PbBr 3 films with and without thermal annealing at 80°C. The product of the chemical mixing is recrystallized by dipping into tetrahydrofuran twice and followed by drying process in vacuum oven at 60°C overnight. The dried CH 3 NH 3 Br is then mixed with PbBr 2 in N, N-dimethylformamide with constant stirring for 12 h at 60°C to form the perovskite precursor .

Synthesis of K 2 La 2 Ti 3 O 10 Sensitizer
PSC can be fabricated from other compound such as K 2 La 2 Ti 3 O 10 . It is prepared by conventional solid state reactions. First, stoichiometric amounts of K 2 CO 3 , La 2 O 3 and TiO 2 powders are mixed. In order to compensate the volatilization of the mixture, there is an excess amount of carbonate added. The mixture is calcinated in platinum crucible at 1000°C in air for 24 h. In order to remove the excess alkali in the substance, the product is washed by distilled water and dried at 500°C to obtain anhydrous oxides. Next step is to protonate the surface. The powder form product is immersed in aqueous HNO 3 solution at room temperature for 15 h. Then, centrifuging process is carried out to obtain white precipitates. Later, the precipitate is immersed in aqueous ammonia solution for 15 h in nitrogen environment. Lastly, the precipitated powder is dried and annealed with the presence of nitrogen gas for 1 h at 400°C. The end product is a light-yellowish powder. The absorbance and transmittance spectra of this type of PSC are shown in Fig. 5a and 5b (Huang et al., 2010).

Synthesis of La 0.7 Sr 0.3 MnO 3 Nano-Perovskite
La 0.7 Sr 0.3 MnO 3 is prepared using the citrate method. The chemicals used in the synthesis process are 0.0031 mol of La(NO 3 ) 3 ·6H 2 O, 0.0043 mol of Mn(NO 3 ) 2 ·4H 2 O and 0.0013 mol of Sr(NO3) 2 precursors in the presence of 0.0205 mol of citric acid. All the chemicals with respective number of moles are added together and left to evaporate overnight at 60°C. Then, the products are dried subsequently at 80°C and 150°C for 24 h. To get the end product, the mixture is calcinated for 9 h at 900°C. The resulting material is powdered. The absorbance of this type of PSC can be seen in Fig. 5c (Ghiasi and Malekzadeh, 2014).

Synthesis of Hybrid PSC with Planar Heterojunction
Structure (ITO/ZnO/CH 3 NH 3 PbI 3 /spiro-OMeTAD/Ag) An organic/inorganic hybrid solar cell is the combination of a mesoporous scaffold, an organic hole transporter and a perovskite light absorber. They have emerged at the forefront of solution process to enable the operation of photovoltaic devices. The sintering process for mesoporous metal-oxide support requires temperature up to 500°C for 8 h. The hybrid PSC device architecture, photograph of a bending test on the hybrid PSC fabricated on a flexible PET substrate, J-V characteristic curve, IPCE spectrum and the efficiency curve of the hybrid PSC are shown in Fig. 6a-f. Thin film of ZnO nanoparticles can be used as an electron transport layer in CH 3 NH 3 PbI 3 based solar cells. ZnO layer does not require sintering and it is thinner than TiO 2 . Based on these advantages a flexible solar cell with power efficiency of 10% can be fabricated. By using ZnO thin film, the performance of the solar cells prepared on rigid substrates is improved also. These cells can achieve a power conversion efficiencies as high as 15.7% under measurement of AM1.5 G illumination. First step is the synthesis of ZnO nanoparticles. ZnO nanoparticles are prepared using a precipitation method with zinc acetate as precursor material and KOH as precipitating agent. The ZnO nanoparticles solution is filtered by going through a 0.45 mm PVDF syringe filter. This solar is fabricated on ITO coated glass substrate that has the 20Ωsq −1 sheet resistance. Then, a thin layer of ZnO nanoparticle is spin coated on the substrate at 3,000 RPM for 30 sec. Following that, a solution of PbI 2 (dissolved in N, N-dimethylformamide with concentration of 460 mg mL −1 ) is spin coated on the top of the ZnO layer at 3,000 RPM for 15 sec. After several minutes of drying in the air, the substrate is dipped into 10 mg mL −1 solution of CH 3 NH 3 I for 40 sec before being dried under a flow of clean air. The spiro-OMeTAD base hole transfer layer (80 mg spiro-OMeTAD, 28.5 mL 4-tertbutylpyridine and 17.5 mL lithium-bis(trifluoromethanesulfonyl)imide (Li-TFSI) solution (520 mg Li-TFSI in 1 mL acetonitrile) all dissolved in 1 mL chlorobenzene) is deposited by 4,000 RPM for 30 sec. Finally, at base pressure of 2×10 −6 mbar, a 150 nm-thick silver layer is deposited by thermal evaporation. All the fabrication steps are carried out under ambient conditions.
The photovoltaic parameter graphs of this type of PSC are shown in Fig. 6. Figure 6d shows the J-V characteristic curve measurements of the highestperforming ITO/ZnO/CH 3 NH 3 PbI 3 /spiro-OMeTAD/Ag flexible device under 100 mWcm -2 AM1.5G illumination and in the dark. Figure 6e represents the IPCE spectrum of the highest-performing ITO/ZnO/CH 3 NH 3 PbI 3 /spiro-OMeTAD/Ag flexible device. The integrated product of the IPCE spectrum with the AM1.5G photon flux is also shown (black line). All measurements in Fig. 6d-f were performed on a single device and were performed from the highest radius of curvature to the lowest (Liu and Kelly, 2014).

Synthesis of Inorganic-Organic Hybrid PSC
The fabrication methodology for a bilayer architecture has the key features of full solution-based processes to synthesize mesoscopic and planar structures. In the fabrication process, poly(triarylamine) acts as the HTM layer and CH 3 NH 3 Pb(I 1-x Br x ) 3 , x = 0.1-0.15 acts as the absorber layer. First, with the spray pyrolysis method a thick layer of TiO 2 (70 nm) is deposited on top of a F-doped SnO 2 (FTO, Pilkington, TEC8) substrate. In order to avoid direct contact between hole conducting layer and the FTO, a 20 mM titanium diisopropoxide bis(acetylacetonate) solution is used at 450°C. Next, a 200-300 nm thick mesoporous TiO 2 film is spin coated on top of the bl-TiO 2 /FTO substrate. Then, in order for the organic components to be removed, the film is calcinated at 500°C for 1 h in air. After that, 27.86 mL of CH 3 NH 2 (40% in methanol) and 30 mL of HI (57 wt% in water) or 44 mL HBr (48 wt% in water are mixed in a 250 mL round-bottom flask at 0°C for 4 h with continuous stirring. By evaporation at 55°C for 1 hour, the precipitate is recovered. MAI and MABr are dissolved in ethanol before being recrystallized from diethyl ether and dried at 60°C in a vacuum oven for 24 h. The prepared powders of PbI 2 and PbBr 2 for 0.8 M MAPb (I 1-x Br x ) 3 (x = 0.1-0.15) solution are stirred in a mixture of GBL and DMSO (7:3 v/v) at 60°C for 12 h. Using a continuous two-step spin coating process at 1,000 and 5,000 RPM for 10 and 20 sec respectively, the resulting solution is coated on top of the mp-TiO 2 /bl-TiO 2 /FTO substrate. In the second spin coating step, the substrate is treated with toluene drop casting (around 1×1 cm). A solution of poly (triarylamine) (15 mg, PTAA, EM Index, M w = 17,500 gmol −1 ) in toluene (1.5 mL) is mixed with 15 µL of a solution of lithium bistrifluoromethanesulphonimidate (170 mg) in acetonitrile (1 mL) and 7.5 µL 4-tert-butylpyridine and spin-coated on the MAPb(I 1-x Br x ) 3 (x = 0.1-0.15)/mp-TiO 2 /bl-TiO 2 /FTO substrate at 3,000 RPM for 30 sec. Finally, an Au counterelectrode is deposited by thermal evaporation. The active area of this electrode is fixed at 0.16 cm 2 .
The photovoltaic parameter result of this PSC is shown in Fig. 7 and Table 1. The J SC , V OC and FF values obtained from the I-V curve of the reverse scan are 19.2 mAcm −2 , 1.09 V and 0.69 respectively, yielding a PCE of 14.4% under standard AM 1.5 conditions. In contrast, the corresponding values from the I-V curve of the forward scan are 19.2 mAcm −2 , 1.07 V and 0.44 respectively, showing a pronounced discrepancy of 9.1% in overall efficiency. Certified power conversion efficiency up to 16.2% is achieved in this PSC (Jeon et al., 2014).

Synthesis of HTM
The type of layers on top of the substrates, namely the Electron Transport Layer (ETL) and Hole Transporting Layer (HTL), is an important consideration in the fabrication of solar cell. These layers will facilitate the movement of holes and electrons around the device, thus a good choice in the fabrication of these layers will increase the efficiency of the PSC. From Table 2, one can observe that most of the fabrication methods of HTM layers are almost similar. Besides that, this table is useful to find the suitable method for the fabrication of HTM layer (either polymer or spiro-OTAD).

Summary of Fabrication Methods
Fabrication plays an important role in development of PSC and The summary of fabrication methods and chemicals are shown in Table 3. Based on Table 3, one can deduce that spin coating plays a major role in the fabrication of PSC.

Impact of Various Factors on PSC Performance
Despite the high efficiency of PSC, it still faces several challenges such as instability and some other issues, thus researches are going on to explore the effect of those issues on the performance of PSC. In this section, a review on the different impacts of various factors such as atmosphere condition, film thickness and properties, HTM, electron transport, external electric field, blocking layer, transport layer, CBO and VBO absorber layer, heat treated ion assisted film, thermal annealed effect, reduced graphene oxide, blade coating and low temperature solution on the performance of PSC.

Impact of Storage Conditions on PSC Performance
In this section, the impact of varying storage conditions on the photovoltaic performance of TiO 2 CH 3 NH 3 PbI 3x Cl x spiro-OMeTAD solar cells will be analyzed. The results of the PSC performance under different dark storage conditions are summarized in Table  4. Based on the results shown in Table 4, it is apparent that the dry air is the most affordable storage condition to accomplish high energy conversion efficiency during the first few hours. Precursor composition, perovskite morphology, pore filling and processing temperature are significant factors in evaluating the solar cell efficiency. During the initial couple of hours, dry air is the most suitable storage condition for high energy conversion efficiency.
Ambient atmospheric condition, proper oxidization and p-doping spiro-OMeTAD are important elements that need to be considered during the drying process as it is tightly related to efficiency. To achieve efficiency of 15%, TiO 2 and spiro-bi fluorine have to be used as the electron collector and hole conductor respectively. Absence of oxygen during the drying process can lead to decrement and increment in the electron conductivity and the recombination of charge carrier respectively. Electron conductivity of the LiTFSI-doped spiro-OMeTAD increases by two orders of magnitude when it is exposed to dry air. Some of the other factors that have been improved relatively to dry N 2 and vacuum conditions are photo-current density, FF and PCE.    The performance graphs of three solar cells treated under different dark storage conditions are shown in Fig.  8. The charge carrier density is enhanced with high resistivity and p doping. As mentioned before, long term stability is a big issue in PSC and this concern can only be minimized by keeping the device away from detrimental moisture and by encapsulation process after proper oxidation of HTM. Fermi level of the spiro-OMeTAD becomes lower due to the light illumination and this will result in higher V OC . The highest efficiency of 11.2% is achieved under dry air condition, whereas the efficiency is only 5.4 and 7.2% under vacuum and dry N 2 conditions respectively. Furthermore, PSC stored under dry air condition has the highest open circuit voltage of 0.89 V, highest photo-current density of 18.4 mA/cm 2 and largest fill factor of 68.9%. Although the PSC stored in dry air exhibits a marginally better performance than PSCs stored under dry N 2 and vacuum conditions, the atmospheric gas storage condition does not have much effect on the final efficiency. Besides, the light harvesting efficiency remains generally the same under different storage conditions (Sheikh et al., 2015).

Impact of Film Thickness on PSC Performance
Even though a considerable amount of literature on the on performance of perovskite has been published, the correlation between the film thickness of CH 3 NH 3 PbI 3- x Cl x (hybrid) perovskite layers and its performance are rarely addressed. As shown in Fig. 9a and 9b, the effects of CH 3 NH 3 PbI 3-x Cl x film thickness on both J SC and PCE are moderate when the thickness of CH 3 NH 3 PbI 3-x Cl x film is larger than 400 nm. When the film thickness is decreased to around 200 nm, both J SC and PCE are improved by 15 and 37%, respectively. Photo-responses below 500 nm are moderately enhanced in comparison with photo-responses in the wavelength region from 550 to 800 nm, which are dramatically enhanced. Minimization of the surface charge carrier recombination due to the enlarged surface coverage leads to the reduction of energy loss. Efficiencies of PHJ pero-HSCs are heavily depending on the film morphology of CH 3 NH 3 PbI 3-x Cl x layer when the film thickness is larger than 400 nm. Based on Table 5, due to the large surface roughness in the films thicker than 600 nm, both J SC and PCEs will decrease. From Table 5, the relationship between the efficiencies of "planar heterojunction" and the thin film thicknesses, as well as the morphologies of solution-processed CH 3 NH 3 PbI 3-x Cl x perovskite layers can be observed .
In Table 6, the performance results are based on TiO 2 thin films with different film thickness for quantum dot sensitized PSCs. As long as the absorption coefficient of quantum dots perovskite is high, the photo-current density of thinner films will be relatively higher compared to thicker films. By controlling the concentration of coating solution, post-annealing condition and TiO 2 surface modification on the quantum dot PSC, the efficiency of 6.54% can be achieved at AM 1.5G 1 sun illumination, as presented in Table 6. Because quantum dot tends to dissolve constantly into the redox electrolyte, one big issue of (CH 3 NH 3 ) PbI 3 quantum dot sensitized solar cell is its stability. It is stable for approximately 10 min (about 80% degradation) under continuous irradiation (Im et al., 2011). As shown in Table 7, the performance of the device is extremely well. This device has an eminently reproducible V OC and it is in the excess of 1.0 V. The average current density of this short circuit (J SC ) is nearly 20 mA cm 2 , with an average PCE of 13.7% and a very good fill factor of 70% (Liu and Kelly, 2014).

Impact of Superior Grain Growth and Light Harvesting Properties on PSC Performance
Another approach to improve the efficiency of PSC is by supervising the grain growth and light harvesting properties. The performance of PSC cells made of nanoassembled TiO 2 structure dispensed on TiO 2 compact layer, integrated with PbI 2 or CH 3 NH 3 PbI 3 layers, are shown in Fig. 10. Because of the relatively smoother, nano-grained nature of the surface of PbI 2 , perovskite crystal size is smaller in case of TiO 2 NP. Larger perovskite crystallites also can enhance light scattering and reduce recombination at the TiO 2 /CH 3 NH 3 PbI 3 /HTM interface. Higher amount of PbI 2 retained on the TiO 2 NB film can lead to enhanced grain growth of perovskite crystallites in the NB case and result in poor performance. Efficiency improvement using NB cell can be observed in the increment in both J SC and V OC . Performance of the PSC can be enhanced (based on Table 8, efficiency increases about 17%) by using NBs, in addition of a striking improvement of the grain size. Larger V OC in NB cell is attributed to the larger CH 3 NH 3 PbI 3 crystal size, which leads to the reduction in the grain boundary density (Tathavadekar et al., 2015).  . 10. UV-vis absorbance spectra of (a) PbI 2 -loaded and (b) CH 3 NH 3 PbI 3 -loaded TiO 2 (NP and NB) films on FTO substrate respectively (c) J-V characteristic curves for NP cell and NB cell (Tathavadekar et al., 2015) Fig. 11. Parameters of PSCs with and without HTM as a function of E v_absorber -ϕ M (Minemoto and Murata, 2014)

Impact of Hole Transport Material (HTM) on PSC Performance
Although HTM can contribute to the efficiency, it cannot resolve the problem of ϕ M mismatching. Presence of the HTM will increase the amount of V BI and thus the elimination of HTM is required to preserve the stability of the device. Nonetheless, to achieve a high efficiency, elimination of mesoporous layer is necessary. Improvement of the absorber quality, the matching of ϕ M and back contact material are essential for further improvement of device performance. In open circuit, the monotonical decrease in V BI leads to the decrease in the circuit voltage and reduction of electric field across the absorbance. By moving ϕ M upward to vacuum level, both V OC and FF will decrease, as shown in Fig. 11. Exclusive of the presence of HTM, the choice of the back contact material is critical in achieving high efficiency. Indicating the ϕ M matching is one of the requirements for achieving high efficiency up to 17.9% (Minemoto and Murata, 2014).

Impact of Compact Layer (CL) Thickness on PSC Performance
In this section, the influence of the CL thickness on disparate electron transport path in perovskite is investigated. Generally, increase in the thickness of the CL will improve the performance of the device. Decrease in the area of contact is the reason for the suppression of the electron transport from FTO to HTM and HTM and FTO electrodes form ohmic contacts, leading to the recombination of charge carriers. When the thickness of CL is increased, J SC , V OC and η of device will be increased also. The graph in Fig. 12 indicates that the optimal perovskite performance can be achieved at CL thickness of 90 nm. However, there is a reduction in transparency at 90 nm CL and reduction in transmittance at the thickness of 180 nm. Any further increment in the CL thickness will reduce the recombination retardation and lead to poor performance. Higher optical transmittance allows improvement in absorbance of light, which in turn increases the number of excited electrons. As shown in Table 9, cells with thicker CL have lower current and cells without CL have low energy conversion efficiency.
The CL thickness dependency varies in three sets of fabricated solar cells: FTO/CL/nc-TiO 2 /Perovskite/HTM/Carbon (1; classical cell), FTO/CL/ncTiO 2 /Perovskite/Carbon (2; omitting HTM) and FTO/CL/Perovskite/HTM/Carbon (3; omitting nc-TiO 2 ). Figure 12a-d present the photovoltaic characteristics of these three sets. The smoothest TiO 2 layer is at the thickness of 90 nm and at the thickness of 180 nm there will be a decrement in the transmittance, while at 390 nm there will be 20% decrement in the transmittance. To achieve a high performance for solar cell, the thickness should be less than 180 nm. A continuous layer can be formed by tightly coating a CL on the FTO (Wang et al., 2014a).

Impact of External Electric Field on Planar Heterojunction PSC Performance
By operating at a low temperature, planner heterojunction PSC can be fabricated. Figure 13 and 14 illustrate the performance of PSCs under different bias scanning range and different external bias respectively. Performance of this solar cells is strongly reliant on the external bias, while the light soaking effect is inconsequential.
By the appropriate optimization of the bias conditions, power conversion efficiency can be increased by 8-fold. Besides, PCE can increase significantly from 1% to over 8% by applying an external bias prior to the device measurement. Larger scan bias range means better performance (From 0.13% at -1 to 1 V to 4.29% at -1 to 6 V). But when the range is too high (for instance 1 to 8 V), the device may get destroyed due to the large injection of current density and eventually, as a result the performance will degrade. Another factor that can lead to decrement in performance is negative bias. On the other hand, positive bias leads to the increase in the performance of PSC. The reproducibility and stability of the device will be increased if ionic transport decreases. There are three possible reasons that can explain the dependency performance on external bias, which are the trap states in perovskite, ferroelectric nature of perovskite and ionic transport in perovskite (Li et al., 2015).

Impact of Blocking Layer on Organic Metal Halide PSC Performance
To form the blocking layer, titanium precursor solution, which is fabricated either by spin coating, spray pyrolysis or atomic layer deposition, is used. Addition of a compact electron blocking layer will prevent charge recombination originated in FTO and the hole conducting layer, thus ensuring an effective charge. Compact and appropriate thickness of the electron BL can lead to increase in the conductivity and photovoltaic performance. In addition, it can ensure the film uniformity of PSC. For high energy conversion efficiency, a highly compact BL film is very critical. Spin coating the precursor solution of TTIP and then PTA will increase the performance of the solar cell, nevertheless the pinholes within the compact layer will pose negative effect on the performance. As shown in Fig. 15a, the diffraction intensity will be affected if the thickness of the film that is generated by PSP is thicker than that of SP and PS. The device performance will have much improvement (based on Table 10, J SC of 12.96 mA cm −2 and V OC of 0.738 V together with FF of 48.77%), after performing spin coating of the precursor solution of PTA on the BL film of PS. This will reveal the importance of the BL in the device. Blocking layer can ensure that an effective charge can be extracted by preventing the hole conducting layer (Jianping et al., 2014).

Impact of Hole Transparent Layer on PSC Performance
In this section, the influence of hole transparent layer in CH 3 NH 3 PbI 3 PSC will be discussed. Generally, hole transparent layer is valuable because it exhibits better charge carrier collection efficiency. However, it also increases and decreases V OC and recombination respectively. Selectivity of contact in the layer increases due to the recombination at the hole collection electrodes. As shown in Fig. 15b and 15c, at the band edge of perovskite (wavelength 750 to 800 nm) absorption will drop strongly, but for the small wavelength (less than 500 nm), absorption will decrease gradually. At wavelengths below 500 nm, the absorption is unity for the both PSCs with and without HTL and because of the constant ratio on the EQEs around this wavelength range, the internal quantum efficiency will have to be changed, i.e., charge carrier collection efficiency. For the device without HTL it is decreased by 18% (3.5 mA/cm 2 ) (Tress et al., 2014).

Impact of Conduction Band Offset (CBO) and Valence Band Offset (VBO) on PSC Performance
Interface recombination will become prominent due to the high reduction of activation energy when the valence band energy of absorber layer is higher than the conduction band energy of buffer layer. As a result, the spike formed by the CBO will serve as a fence for photogenerated carrier flow. Figure 16a-d show the performance parameters of PSC with varying CBO, VBO and interface defect density values. High value of CBO induces double-diode-like curvature and leads to low FF and stable V OC . When CBO decreases from 0.0 to -0.4 eV, fill factor will be decreased while J SC remains almost unchanged. Negative CBO reduces V OC and V OC corresponding to E a, however it will increase the interface recombination. From here, open circuit states that the conduction band of the buffer is at least equal or higher than the amount of the absorber.
Low and high VBO have the same hole concentrations, but high VBO means poor FF. Negative VBO decreases both V OC and FF, but J SC will stay almost unchanged. Increase in VBO will decrease FF and electric field across the absorber. In the similar situation, prominent increase in the interface recombination will happen due to the higher position of the valence band of the HTM in relative to the absorber. Too low position for the valence band of HTM leads to the incomplete depletion of the absorber. Based on the results, the optimum conduction and valence band energy of buffer and HTM are 0.003 eV higher and 0.002 eV lower than the corresponding bands of absorber respectively. These results are essential guidelines in designing new buffer, absorber and HTM (Minemoto and Murata, 2015).

Impact of Heat Treated Films, Ion-Assisted Films and Thermal Annealing on PSC Performance
Heat treated films and ion assisted films can be used to accelerate the reaction between CH 3 NH 3 I and PbI 2 . By annealing a better crystallinity and high phase purity can be obtained, however with ion assisted greater amorphous structure will be obtained, also different morphologies of grown thin films will be created. Table  11 presents the performance data of PSC with varying annealing time. After heat treating, optical properties, crystallinity and phase purity of the solar cell will increase in more stable film. Phase purity of perovskite can be achieved by only use of ion assisted deposition, without the help of post deposition heat treatment. There is a relation between beam energy and crystallinity of the film, without the ion beam irradiation the crystallinity of film decreases, however the decrease in the beam energy increases the crystallinity of the resultant film. Besides that, by using heat a solar cell with better crystallinity and high phase purity can be achieved, while greater amorphous structure will be achieved with ion assisted process. Different morphologies of grown thin films can also be fabricated (Cimaroli and Yanfa, 2014). In order to improve the device performance, the CH 3 NH 3 PbBr 3 /ZC-NA composite is annealed at 80°C. After the thermal annealing, while the V OC remains unchanged, there is a slight increase in J SC and fill factor and that leads to improvement in averaged η (increases from 3.34% up to 4.31% under AM 1.5 illumination (100 mW/cm 2 )). As shown in Fig. 17a, the drop in transmittance at 800 nm is indicative of the direct band gap of the perovskite. Some of the other effects of thermal annealing are such as stabling the presence of CH 3 NH 3 PbBr 3 crystals and improvement of device performance. Enhancing the crystallinity and reducing the charge recombination will increase the charge transport properties of the perovskite (Boix et al., 2014).
The performance data shown in Table 12 is obtained under one sun illumination (100 mWcm −2 ) at AM1.5 G condition from 40.26 wt%, using equimolar mixture of CH 3 NH 3 I and PbI 2 in g-butyrolactone. The electrolyte composition is 0.9 MLiI, 0.45 MI 2 , 0.5 MtBP and 0.05 M urea in ethyl acetate. Photovoltaic property can be adjusted to certain extent by manipulating the annealing temperature. Photovoltaic performance of 5.5 mm-thick TiO 2 layer is slightly improved from voltage of 0.6046 V, photo-current of 13.31 mA/cm 2 and fill factor of 0.5013 to 0.6172 V, 13.76 mA/cm 2 and 0.557 respectively, as presented in Table 12. This improvement is due to the increase of the annealing temperature to 100°C (Im et al., 2011).

Impact of Buffer Layer on PSC Performance
The fundamental purpose of the buffer layer is to provide a layer between aqueous solutions for depositing HTM during solar cell preparation and also to absorb photons under photo-illumination. Buffer layer leads to weak light absorbance at the wavelength ranging from 550 to 600 nm and it will increase J SC due to the ignorable attenuation effect. As a result, it can protect the solar cell from contacting with the solution .

Impact of Air Mass on PSC Performance
Generally, optical air mass is defined as the optical path length scale for the solar beam over the atmosphere to the optical path over a standard atmosphere at sea level, when sun is at the zenith position. Dry air molecules reduce this ratio to a purely geometrical value. Variation in air mass will lead to amplification or attenuation of solar irradiance, photon density distribution and strong absorption in the visible region. It should be noted that a considerable loss of air mass ratio will occur at AM 1.5. However, high AM will shift the photon distribution into shorter wavelength regions. The blue spectral region is where attenuation of AM is higher and where majority of absorption bands occur. This demonstrates the strict condition imposed by air mass on the performance of PSC (Senthilarasu et al., 2015).

Impact of Aerosol Optical Depth (AOD) on PSC Performance
The small particles that are suspended in the air are called aerosol. The parameter used in measuring the atmospheric attenuation of solar radiation by aerosol is called the aerosol optical depth. Increase in the AOD will change the incoming irradiance. Due to improvement of turbidity due to increment in AOD, there will be more attenuation and scattering of solar irradiance in the visible UV spectra. The narrow absorption band in the blue wavelength region of the spectrum will be more affected under the turbidity condition. If the aerosol concentration in the atmosphere is low, the spectral losses will be lower than the air mass condition and there will be no effect on the PSC performance (Senthilarasu et al., 2015).

Impact of Perceptible Water on PSC Performance
The integrated amount of water vapor in zenith direction between the surface and the top of the atmosphere is called perceptible water. To be more precise, the thickness of the liquid water that would be formed if all the vapor in the zenith direction were condensed at the surface of a unit area is also described as the thickness of the liquid water. The improvement in PSC performance by increasing perceptible water, which results in the enhancement of absorption of solar radiation near the infrared spectra, can be observed (Senthilarasu et al., 2015).

Impact of MoO 3 /NPB as Hole Extraction Layers on Fully Vacuum Processed PSC Performance
Fully vacuum processed PSC will yield better reproducibility. Acceleration in the advancement of the PSCs, high V OC and free from the restriction of their solvents are some other benefits of this process. Employing MoO 3 /NPB layer as the hole extraction layer is the main reason which is responsible for the high V OC . From Fig. 17, the best results that can be achieved under the aforesaid conditions are 13.7% for PCE, 18.1 mA/cm 2 for JSC, 1.12 V for V OC and 0.67 for FF. In Fig. 17c, the scan step of J-V characteristics is 0.02 V with 0.3 sec of interval time for each step and the scan direction is from negative to positive voltage. Vacuum processed PSC is a promising tool to apply the various charge transport materials in the layered structures (Kim et al., 2015).

Impact of Poly (3-Hexylthiophene) Hole Transport Layer on High-Efficiency CHNHPbI (3-x) Cl x PSC Performance
One of the suitable HTM choices for high-efficiency PSC is doped P3HT. The performance of the fabricated device can achieve up to 9.3% PCE and high V OC . As shown in Fig 18a, the black-colored line represents the absorbance of a dehydrated TiO 2 sample, while the darkgray-colored line and the light-gray-colored line represent the absorbance of a UV-treated sample and a sample without any treatment respectively. The TiO 2 thickness is 700 nm and a 40% w/w perovskite precursor solution is spin coated at 2000 RPM and annealed at 120°C for 1 h. All measurements are carried out using an integrating sphere. Absorbance of FTO/glass has been subtracted from the absorbance spectra of all the samples. On the other hand, in Fig. 18b, the current density over voltage curves of the best FTO/c-TiO 2 /nc-TiO 2 /CH 3 NH 3 PbI (3-x) Cl x /P3HT/Au (red-colored line and symbols) and FTO/c-TiO 2 /nc-TiO2/CH 3 NH 3 PbI (3- x) Cl x /spiro-OMeTAD/Au (black-colored line and symbol) solar cells are presented. The PSC efficiency can be further improved by replacing Al 2 O 3 with TiO 2 . Adjusting the perovskite composition through addition of Br or other elements also serves the same objective (Di Giacomo et al., 2014).

Impact of Reduced Graphene Oxide (RGO) on Highly-Efficient and Stable Planar PSC Performance
RGO nano-sheets are used as electrode interlayer. A highly-efficient and stable CH 3 NH 3 PbI 3 PSC, high RGO conductivity, better aligned energy levels and better growth of the crystalline CH 3 NH 3 PbI 3 can be expected through the addition of the novel RGO HTM. PSCs with RGO have an excellent reproducible efficiency up to 10.8%. Researches have revealed that facilitated charge collection with retarded charge recombination, high RGO conductivity, better aligned energy level and better growth of crystalline CH 3 NH 3 PbI 3 on RGO HTMs are the reasons for improvement in the performance of PSCs with RGO HTM.  For example, Table 13 shows a comparison of PSC performance parameters between PSCs with HTM made of PEDOT: PSS,  show the performance curves of the three PSCs with varying HTM choices. This approach could motivate future researches into the design and development of HTMs and promise the improvement of PSC technology in terms of having lower cost with higher performance along the excellent stativity (Yeo et al., 2015).

Impact of PbI 2 Blade Coating as Air-Flow Assistance on PSC Performance
Blade coating can be performed in replacement of conventional spin coating to increase the deposition area. Blade coating can even reduce the use of material due to the peculiar compact structure of the PbI 2 obtained from blade-coating. To improve the efficiency, a long dipping time (30 min) into the CH 3 NH 3 I solution is required. By legitimate boost of the blade deposition and the air flow dynamics over the very large areas, most of the issues related with limited efficiency can be resolved. Figure 18c represents the current density over voltage characteristic curve of an 8 cm 2 module with 6.7% efficiency and large areas solar cells connected in series, prepared by PbI 2 blade coating deposition and 30 min of dipping in CH 3 NH 3 I solution. The V OC of the complete module is the sum of the V OC of each cell, indicating a good connection between the cells. The efficiency based on this PSC is reported to be up to 10.3% (Razza et al., 2015).

Impact of ZnO Nanoparticles on TiO 2 Thin mp-Electrodes for PSC
By manipulating the mixing ratio of ZnO nanoparticles, the thin TiO 2 electrodes for two-step process sensitize for PSC can be modified. Table 14 illustrates the effect of varying amounts of ZnO nanoparticles in the electrodes on the performance parameters of PSC. The best efficiency result is obtained for the cell with electrodes containing 30 wt.% ZnO, as depicted by the performance curves of PSCs with nonmodified electrodes and 30 wt% ZnO-modified electrodes shown in Fig. 20b and 20c. On the other hand, as shown in the two-step process in Fig. 20a, after the removal of the ZnO nanoparticles from the electrode structures, the increased current density is attributed to the facilitated conversion of molecular precursors to perovskite (Lee et al., 2014).
To ensure that the performance results obtained are solely dependent on the modification of the mesoporous electrode, it is necessary to hold all other parameters such as film thickness and post processing steps constant. This method has proven itself as a viable alternative for better improvement of the efficiency vin mesoporous and nanostructured solar cells (Lee et al., 2014).

Impact of Low-Temperature Solution on High-Efficiency and High-Flexibility PSC Performance
The goal of thin film photovoltaics is to reduce the energy cost in solar panels compared to their silicon model. By using an electron collection layer, one of the highest efficiencies for the solution-processed PSC can be achieved. However, this electron collector layer needs to be sintered at 500°C. As shown in Table 15, the photovoltaic parameters of a low cost, solution-based disposition procedure nano-composites of TiO 2 nanoparticles and graphene as the electron collection layer in mesosuperstruct PSC, is reported. Figure 21a and b, on the other hand, show the performance curves of the samples compared in Table 15 and the best-performing solar cells in bright and dark conditions respectively. The entire device can be fabricated with temperature less than 150°C due to the graphene nano-flakes that will provide superior charge collection in nano-composites. Based on Table 15, the graphene/metal oxide nano-composite has the potential as a low-cost solar cell. A relatively good absorbance for PSC can be achieved even with relatively easy solutionbased process at low temperature using low-cost materials and on top of that the process does not rely on the use of elements in limited supply. Electron collection layers have a critical role in the most efficient thin film of PSC, as they facilitate selective charge collection. CdS, the most common commercial thin film of solar cell, has been in operation for this purpose also. They devote to parasitic absorption in the solar cell. It is necessary to mention that they can prevent leakage and shunting for the currents under reverse bias (Zhong et al., 2015;You et al., 2014).

Impact of Cesium-doped Methylammonium on Light Absorber Hybrid PSC Performance
This goal of Cesium doping is to improve the performance of inverted-type perovskite/fullerene planar heterojunction hybrid solar cells. Base on Table 16, high performance is achieved at 10% Cs doping in the MAPbI 3 perovskite structure, represented by enhancements in the morphology, device efficiency in light absorption and energy difference between the valance band. Because of the increment in the in short-circuit current density and open-circuit voltage) with light absorption in optimal device thickness, shown in Fig. 21c, this device efficiency (with 10% Cs doped) is improved from 5.51 to 7.68%. Such approach to perovskite design offers the possibility to enhance the performance of conventional perovskite solar cells with n-type titania or alumina layers and to achieve even higher open-circuit voltages by employing tandem device structures (Choi et al., 2014).

Summary of Various Impacts on PSC Performance
Besides the sensitizer itself, there are many other factors that influence the PSC efficiency. The sensitizer is only responsible in gaining high efficiency as it only converts the photons into electron-hole pairs. Certain sensitizers can carry out the conversion at certain wavelengths and different sensitizers can have very different recombination rates of electron-hole pairs. To complete the circuit however, the perovskite sensitizer will need to have the electrons and holes to be conducted in opposite directions and into the external circuit. The hole transport material, or HTM, is where the holes will be collected and the electron transport material, or ETM, is where the electrons will be collected. This collection process is crucial to prevent the recombination of electron-hole pairs within the sensitizer itself.
In order to optimize the efficiency through improvement of sensitizers, the focus is mainly on the structure of the sensitizer. The sensitizer is mostly made out of CH 3 NH 3 PbI 3 or mixed-halide CH 3 NH 3 PbI 3-x Cl x material, fabricated in all-solid-state or combined with Li salt, or both. Moreover, the sensitizer can also be fabricated alongside with poly-(vinylpyrolidone) (PVP) to improve the efficiency. Table 17 summarizes the performance parameters of PSCs with several different sensitizer materials.
As one can observe in Table 15, HTM, which is to collect the holes generated by the sensitizer, have been shown to have a large impact on the PSC efficiency. Spiro-OMeTAD in PSCs is the most widely used HTM, while P3HT comes second. There are cost concerns when it comes to HTM materials, as spiro-OMeTAD is comparatively costly. Table 18 summarizes the performance parameters of PSCs with several different HTMs.
ETM is used to collect the electrons generated by the sensitizer. As ETM is generally made out of metal oxides, there are many different types of approaches in optimizing the ETM, including annealing and combining the ETM with other materials, or even embedding the sensitizer into the metal oxide itself. Some of the experiments have also shown that coating the sensitizers within the ETM will increase the efficiency and some have taken a different approach, that is to coat the sensitizer on top of a scaffolding TiO 2 layer. Table 19 summarizes the performance parameters of PSCs with several different ETMs. By changing some of the interface materials within the PSC device, the efficiency can change drastically. There are also some other experiments where the substrate has been changed from an ITO substrate/rigid glass to a flexible PET/ITO substrate with a decrease in efficiency. When the flexible substrates are bent or rolled over for multiple times, the efficiency drops. Overall in terms of efficiency, there are a multitude of variations in attempting higher efficiency, ranging from optimization of HTM and ETM, fabrication of sensitizer, crystal structure of ETM and how the sensitizer is embedded into the ETM itself. There are even hybrid materials where a combination of materials is used to fabricate the sensitizer.

Conclusion
Throughout the entire review paper, it can be noted that there are many different influences that will play a part in how perovskite can be manifested into commercial products without losing its properties. There are also many fabrication methods, ranging from commercial-friendly methods to highly sophisticated and costly methods. Some of the low-cost techniques have been tried and tested to fabricate PSC devices, such as spray pyrolysis technique (Razza et al., 2015) and also spin coating of dye-sol paste (Ramos et al., 2014).
With different tables of efficiency based on the variation of how the PSC device is made, it can be foreseen that a combination of said variations can be used to maximize the performance of future PSC. Flexible PSCs have also been fabricated, with a very good efficiency of 6.9% for a relatively new field (You et al., 2014;Roldan-Carmona et al., 2014). As the PSCs are made out of multiple layers of different materials, these layers, have shown its importance in the device.
We can see that there are certain optimizations done to the PSCs through different trials such as the data shown by Nagarjuna et al. (2015) where only the hole transport material layer is varied and it showed that the layer does indeed play a role in maximizing efficiency. In the sensitizer layer however, it has been said that the liquid electrolytes will pose a problem in the long run as it can potentially leak out or dried over time and thus Xiao et al. (2014) showed that with power conversion of 7.34%, an all-solid-state PSC can be done. The electron transport materials are mostly made out of metal oxides and is shown that the lattice structure of said metal oxide materials can also affect the efficiency drastically, as shown by Yamamoto et al. (2014). However, the fabrication of the ETM alone can incur a high cost onto the device, as sintering is required. Some of the experiments done had the sensitizer embedded into the ETM layer, resulting in an efficiency of 12.7% (Yamamoto et al., 2014) and some have varied the number of layers of ETM to be fabricated and the resulting efficiency of double-layered ETM reached 5.28% (Yamamoto et al., 2014).
Perovskite parade attractive and unusual physical properties which has persuaded the world to study it terms of the theoretical modeling and practical applications. Studies on perovskite were able to open the doors to the many revolutionary discoveries, which allowed scientists to design devices with new concepts. Perovskite has a promising future due to low price, adaptability and thermal stability, which often depend on bulk and surface characteristics. The possibility of being coated on flexible or semi-transparent substrate, light weight thin-films are some advantages of PSC. Moreover, having high and rapidly improved efficiencies can also be characterized as advantages of this method. Even though there were many accomplishments on improvement of PSCs, but they are still on early stages of commercialization in compared to their more mature rivals. The main concerns with PSCs are still their steadiness and toxicity. By overcoming these two issues, PSCs can enter the commercial world and even mount among its rivals and become the main commercialized PSCs.