Effect of Electrochemically Deposited MgO Coating on Printable Perovskite Solar Cell Performance

Herein, we studied the effect of MgO coating thickness on the performance of printable perovskite solar cells (PSCs) by varying the electrodeposition time of Mg(OH)2 on the fluorine-doped tin oxide (FTO)/TiO2 electrode. Electrodeposited Mg(OH)2 in the electrode was confirmed by energy dispersive X-ray (EDX) analysis and scanning electron microscopic (SEM) images. The performance of printable PSC structures on different deposition times of Mg(OH)2 was evaluated on the basis of their photocurrent density-voltage characteristics. The overall results confirmed that the insulating MgO coating has an adverse effect on the photovoltaic performance of the solid state printable PSCs. However, a marginal improvement in the device efficiency was obtained for the device made with the 30 s electrodeposited TiO2 electrode. We believe that this undesirable effect on the photovoltaic performance of the printable PSCs is due to the higher coverage of TiO2 by the insulating MgO layer attained by the electrodeposition technique.


Introduction
Perovskite solar cell (PSC) technology has made tremendous progress over the last few years, with a significant increase in power conversion efficiency with recent devices reaching over 22% [1].Initially, perovskite (CH 3 NH 3 PbI 3 ) was used as a sensitizer to replace organic dye molecules in dye-sensitized solar cells by Miyasaka et al. [2].However, corrosion of the CH 3 NH 3 PbI 3 by the I − /I 3 − electrolyte hindered the interest of this new sensitizer until realizing the possibility of replacing the electrolyte with a solid organic hole transport material (HTM), spiro MeOTAD [3].Since then, both mesoscopic and planar heterojunction PSCs have been fabricated with different architectures and preparation methods [4][5][6][7].Recently, huge interest was given to printable PSCs with carbon counter electrodes as they demonstrate enormous potential for achieving high efficiency, long lifetime and low manufacturing costs which may lead to future commercialization [7][8][9].
In PSCs, often a thin compact TiO 2 layer (~50 nm) is formed in order to prevent back electron transport from fluorine-doped tin oxide (FTO) to either perovskite or HTM.However, unevenness, surface defects and the presence of pin holes in this layer are often responsible for reducing the cell performance.Therefore, over the last years, many attempts have been made to enhance the overall cell performance of PSCs by retarding the back transfer of photo-generated electrons through the FTO/TiO 2 interface by surface modification of TiO 2 using insulating metal oxides [10][11][12] and hydroxides [13,14] or high band gap semiconductors [15,16] that form a blocking layer between the perovskite sensitizer and TiO 2 layer to block the back electron flow towards the HTM.For instance, TiO 2 modification with a monolayer of silane [17] and ZnO with 3-aminopropanioc acid self-assembled monolayer [18] have been found to enhance the device efficiency by retarding the back electron transfer processes of PSCs.Wang et al. [19] found that magnesium oxide and magnesium hydroxide, formed at the surface of TiO 2 , suppress the recombination, achieving an improvement of V oc and hence photo-conversion efficiency (PCE).Similarly, Jung et al. [10] coat an ultrathin MgO layer on TiO 2 and found an improvement in the fill factor and V oc of the device, which they believe to be due to the retarded charge recombination at the interface between MgO and CH 3 NH 3 PbI 3 .Conversely, Ke et al. [20] and Liu et al. [21] have shown that a higher V oc can be obtained for planar PSCs without a compact TiO 2 layer, suggesting that the recombination pathways in PSCs are still unclear and more investigation with different device configurations is required for better understanding.
Compared to other widely used coating techniques such as spin coating and screen printing, the electrodeposition is considered to be a versatile technique for producing surface coatings, owing to its precise controllability, better adherence to substrate, rapid deposition rate with a higher uniformity, room temperature operation and relatively low cost [22][23][24].In this study, we grew a conformal Mg(OH) 2 coating by the electrodeposition method on the surface of FTO/TiO 2 and investigated the effect of this insulating oxide on their photovoltaic device performance.Our results confirmed that there is an adverse effect on the device performance with the MgO coating obtained by electrodeposition of Mg(OH) 2 on the printable PSCs.

Materials and Methods
Fluorine-doped tin oxide (FTO) glass substrates were etched with a laser before being cleaned ultrasonically with detergent, deionized water and ethanol successively.Then, some of the substrates were coated with a TiO 2 compact layer by aerosol spray pyrolysis at 500 • C using a precursor containing 300 µL of titanium diisopropoxide bis (acetylacetonate).The treated film was annealed at 500 • C for 30 min inside an oven.Then, the mesoporous TiO 2 layer was deposited by screen printing using a TiO 2 paste [3 g of F-6 powder (Showatitanium, Toyama, Japan) with 0.5 mL of acetic acid, 15 mL of ethyl cellulose (45-55 mPa•s, TCI, 10 wt % in EtOH) and 50 g of α-terpineol].After the coating, the film was dried at 125 • C for 5 min and sintered at 500 • C for 30 min using an oven.The Mg(OH) 2 coatings were electrodeposited on the FTO/TiO 2 or FTO substrates in an aqueous electrolyte solution composed of Mg(CH 3 COO) 2 •4H 2 O having a concentration of 0.01 M. The electrodeposition was carried out in the three-electrode configuration using the TiO 2 -coated FTO or FTO substrate as the working electrode with the cathode area of 1 cm 2 , Ag/AgCl electrode, and Pt as the reference and counter electrodes, respectively.The electrodeposition was conducted at a constant current of 0.6 mA (Chronopotentiometry) using a Potentiostat/Galvanostat.After the deposition, the films were removed from the electrolyte solution, washed with distilled water and allowed to dry at room temperature.The electrodeposition time was varied for 10 s, 30 s, 1 min, 2 min, 4 min, 6 min, 10 min and 20 min and the deposited Mg(OH) 2 is converted to MgO during the post-annealing of successive layers.In devices A and B, a ZrO 2 space layer was printed on the film using a ZrO 2 paste [3 g of ZrO 2 powder (40-50 nm, Alfa Aesar, Lancashire, UK) with 0.5 mL of acetic acid, 15 mL of ethyl cellulose (45-55 mPa•s, TCI, 10 wt % in EtOH) and 50 g of α-terpineol].The film was annealed at 400 • C for 30 min inside an oven after drying at 125 • C for 5 min on a hot plate.Then, a NiO mesoporous layer was coated on respective devices using a NiO paste consisting of 3 g of NiO powder (20 nm, Iolitec Ionic Liquids Technologies GmbH, Heilbronn, Germany) with 0.5 mL of acetic acid, 15 mL of ethyl cellulose (45-55 mPa•s, TCI, 10 wt % in EtOH) and 50 g of α-terpineol.After the pre-drying at 125 • C, the NiO layer was sintered at 500 • C for 30 min using an oven.Finally, a carbon black/graphite layer was coated on the top of the ZrO 2 or NiO layer by the screen printing method using a pot milled carbon black/graphite paste [1 g of Printex L6 carbon (Evonik Industries, Frankfurt, Germany), 1 g of graphite, 4 g of graphite flakes (~325 mesh, Alfa Aeser, Haverhill, MA, USA), 5 g of TiO 2 (P25), 20 g of α-terpineol and 30 g of ethyl cellulose (10 wt % in EtOH) and sintered at 400 • C for 30 min inside an oven.The synthesis of CH 3 NH 3 PbI 3 (MAPbI 3 ) and deposition on the devices was carried out by the two-step deposition method onto the carbon black/graphite layer (using 1.2 M PbI 2 in DMF solution and CH 3 NH 3 I solution (10 mg/mL)].Upon drying at 70 • C for 30 min, the films darkened in colour, indicating the formation of MAPbI 3 in the solar cell.All of the processes were performed under ambient conditions.
Photovoltaic measurements were conducted using an AM 1.5 solar simulator equipped with a xenon lamp (Yamashita Denso, Tokyo, Japan).The power of the simulated light was calibrated to 100 mW•cm −2 by a reference Si photodiode (Bunkou Keiki, Tokyo, Japan).J-V curves were obtained by applying an external bias to the cell and the generated photocurrent was measured with a B2901A, Agilent voltage current source (Santa Clara, CA, USA).The active area of the cells was fixed at 0.04 cm 2 .The SEM images were obtained by JOEL-JSM-6510 scanning electron microscopes (JEOL, Tokyo, Japan) and EDX measured using a TE3030, Hitachi machine (Tokyo, Japan).

Results and Discussion
In order to investigate the effect of MgO on the photovoltaic performance of printable PSCs, we have deposited a MgO layer on FTO/TiO 2 (devices B and C) or FTO (device D) by varying the Mg(OH) 2 electrodeposition time on different mesoscopic structures (Figure 1).The deposited Mg(OH) 2 is converted to MgO upon post-annealing of successive layers.The devices made with a standard four-layer structure of TiO 2 /ZrO 2 /NiO/Carbon (MAPbI 3 ) are depicted as A and the devices with the same architecture with MgO coating on TiO 2 are illustrated as device B. Device C is employed with the intention of fabricating a thin insulation layer of MgO on TiO 2 and FTO (with the deposition time) before coating the hole-transporting NiO layer.Device D is designed according to the meso-super-structured solar cell structure where the MgO layer is acting as a scaffold for MAPbI 3 in the device.
3 cellulose (45-55 mPa•s, TCI, 10 wt % in EtOH) and 50 g of α-terpineol.After the pre-drying at 125 °C, the NiO layer was sintered at 500 °C for 30 min using an oven.Finally, a carbon black/graphite layer was coated on the top of the ZrO2 or NiO layer by the screen printing method using a pot milled carbon black/graphite paste (1 g of Printex L6 carbon (Evonik Industries, Frankfurt, Germany), 1 g of graphite, 4 g of graphite flakes (~325 mesh, Alfa Aeser, MA, USA), 5 g of TiO2 (P25), 20 g of α-terpineol and 30 g of ethyl cellulose (10 wt % in EtOH) and sintered at 400 °C for 30 min inside an oven.The synthesis of CH3NH3PbI3 (MAPbI3) and deposition on the devices was carried out by the two-step deposition method onto the carbon black/graphite layer (using 1.2M PbI2 in DMF solution and CH3NH3I solution (10 mg/mL)).Upon drying at 70 °C for 30 min, the films darkened in colour, indicating the formation of MAPbI3 in the solar cell.All of the processes were performed under ambient conditions.
Photovoltaic measurements were conducted using an AM 1.5 solar simulator equipped with a xenon lamp (Yamashita Denso, Tokyo, Japan).The power of the simulated light was calibrated to 100 mW•cm −2 by a reference Si photodiode (Bunkou Keiki, Tokyo, Japan).J-V curves were obtained by applying an external bias to the cell and the generated photocurrent was measured with a B2901A, Agilent voltage current source (CA, USA).The active area of the cells was fixed at 0.04 cm 2 .The SEM images were obtained by JOEL-JSM-6510 scanning electron microscopes (JEOL, Tokyo, Japan) and EDX measured using a TE3030, Hitachi machine (Tokyo, Japan).

Results and Discussion
In order to investigate the effect of MgO on the photovoltaic performance of printable PSCs, we have deposited a MgO layer on FTO/TiO2 (devices B and C) or FTO (device D) by varying the Mg(OH)2 electrodeposition time on different mesoscopic structures (Figure 1).The deposited Mg(OH)2 is converted to MgO upon post-annealing of successive layers.The devices made with a standard four-layer structure of TiO2/ZrO2/NiO/Carbon (MAPbI3) are depicted as A and the devices with the same architecture with MgO coating on TiO2are illustrated as device B. Device C is employed with the intention of fabricating a thin insulation layer of MgO on TiO2 and FTO (with the deposition time) before coating the hole-transporting NiO layer.Device D is designed according to the mesosuper-structured solar cell structure where the MgO layer is acting as a scaffold for MAPbI3 in the device.During the electrodeposition of Mg(OH)2 in the devices B, C and D, the below electrochemical reactions could take place at the cathode surface [13], During the electrodeposition of Mg(OH) 2 in the devices B, C and D, the below electrochemical reactions could take place at the cathode surface [13], As a result of these reactions (i.e., with the formation of OH − ), a steep increase of local pH (~9) in the electrodeposition solution near to the cathode could occur, leading to the formation of Mg(OH) 2 due to the poor solubility of Mg(OH) 2 (K sp of Mg(OH) 2 is 1.2 × 10 −11 mol 3 •dm −9 at 25 • C) [25,26].This flocculated Mg(OH) 2 in the solution can be hetero-coagulated on TiO 2 and FTO electrostatically due to their opposite surface charges.(Isoelectric points of TiO 2 ~6.2 FTO ~6 and Mg(OH) 2 ~12) [13].
Once the electrodeposition is started, it is more likely that the low resistive exposed FTO surface (i.e., pinholes) and the TiO 2 surface in the vicinity are coated with Mg(OH) 2 .However, with the increasing deposition time, the alkaline pH boundary extends further away from the interior surface which could lead to the TiO 2 nanoparticles being covered by Mg(OH) 2 .
The surface topographic FEG-SEM images of electrodeposited Mg(OH) 2 for 2 min and 10 min deposition times on FTO glass substrate are shown in Figure 2a,b respectively.The images showed that the substrates are completely covered by electrodeposited flower-like Mg(OH) 2 spheres [13,27].Figure 2c,d shows the cross-sectional images of devices C and D respectively.As shown in Figure 2c, the thickness of the TiO 2 <MgO>/ZrO 2 /NiO (MAPbI 3 ) is around 2 µm whereas the carbon counter electrode is estimated to be around 25 µm.The MgO layer in the device cannot be seen clearly as it is too thin to be visible in the cross-sectional images.However, EDX (Figure 2e) mapping of the cross-sectional image of device D, confirms the presence of Mg, Ni and carbon in the electrode.
Once the electrodeposition is started, it is more likely that the low resistive exposed FTO surface (i.e., pinholes) and the TiO2 surface in the vicinity are coated with Mg(OH)2.However, with the increasing deposition time, the alkaline pH boundary extends further away from the interior surface which could lead to the TiO2 nanoparticles being covered by Mg(OH)2.
The surface topographic FEG-SEM images of electrodeposited Mg(OH)2 for 2 min and 10 min deposition times on FTO glass substrate are shown in Figure 2a,b respectively.The images showed that the substrates are completely covered by electrodeposited flower-like Mg(OH)2 spheres [13,27].Figure 2c,d shows the cross-sectional images of devices C and D respectively.As shown in Figure 2c, the thickness of the TiO2<MgO>/ZrO2/NiO(MAPbI3) is around 2 µ m whereas the carbon counter electrode is estimated to be around 25 µ m.The MgO layer in the device cannot be seen clearly as it is too thin to be visible in the cross-sectional images.However, EDX (Figure 2e) mapping of the crosssectional image of device D, confirms the presence of Mg, Ni and carbon in the electrode.The influence of the MgO layer and its thickness (by varying the Mg(OH)2 electrodeposition time) on the photovoltaic performance of these different printable PSC structures was evaluated by their I-V characteristics.Figure 3 illustrates the variations observed in the I-V characteristics of the The influence of the MgO layer and its thickness (by varying the Mg(OH) 2 electrodeposition time) on the photovoltaic performance of these different printable PSC structures was evaluated by their I-V characteristics.Figure 3 illustrates the variations observed in the I-V characteristics of the TiO 2 /ZrO 2 /NiO/Carbon(MAPbI 3 ) structured device (device A) against the Mg(OH) 2 deposition time (device B).The device A yielded a V oc of 0.78 V, a short-circuit current density (J sc ) of 9.30 mA•cm −2 and a fill factor of 0.44, which correspond to a PCE of 3.19%.The I-V plots and the average photovoltaic parameters in Table 1 of devices show that the device's performance is clearly influenced by the systematic growth of Mg(OH) 2 on TiO 2 .The cell prepared with the 30 s Mg(OH) 2 electrodeposited electrode showed a V oc of 0.75 V, J sc of 8.48 mA•cm −2 , fill factor of 0.52 and PCE of 3.26%.It is noticeable that the J sc has systematically decreased as the Mg(OH) 2 coating time varied from 1 to 10 min, suggesting the blocking of porous TiO 2 .It is more likely that the coverage of Mg(OH) 2 on TiO 2 prevents MAPbI 3 on TiO 2 which was evident by the small reduction of photocurrent density up to 30 s.The trend was continued as the electrodeposition time further increased.The device with 10 min of Mg(OH) 2 coating showed the lowest J sc of 0.61 mA•cm −2 , with a trivial improvement in the V oc (0.79 V) and fill factor (0.53) compared to device A.
Coatings 2017, 7, 36 5 TiO2/ZrO2/NiO/Carbon(MAPbI3) structured device (device A) against the Mg(OH)2 deposition time (device B).The device A yielded a Voc of 0.78 V, a short-circuit current density (Jsc) of 9.30 mA•cm −2 and a fill factor of 0.44, which correspond to a PCE of 3.19%.The I-V plots and the average photovoltaic parameters in Table 1 of devices show that the device's performance is clearly influenced by the systematic growth of Mg(OH)2 on TiO2.The cell prepared with the 30 s Mg(OH)2 electrodeposited electrode showed a Voc of 0.75 V, Jsc of 8.48 mA•cm −2 , fill factor of 0.52 and PCE of 3.26%.It is noticeable that the Jsc has systematically decreased as the Mg(OH)2 coating time varied from 1 to 10 min, suggesting the blocking of porous TiO2.It is more likely that the coverage of Mg(OH)2 on TiO2 prevents MAPbI3 on TiO2 which was evident by the small reduction of photocurrent density up to 30 s.The trend was continued as the electrodeposition time further increased.The device with 10 min of Mg(OH)2 coating showed the lowest Jsc of 0.61 mA•cm −2 , with a trivial improvement in the Voc (0.79 V) and fill factor (0.53) compared to device A.  In device C, Mg(OH)2 was electrodeposited after coating the mesoporous TiO2 layer without a compact TiO2 layer and the MgO layer was replaced with TiO2 and ZrO2 layers in the device D. As shown in Figure 4a and Table 2, device C demonstrated poor device performance compared to the device A. This is probably due to the higher recombination rate at the TiO2/NiO interface despite the MgO layer on TiO2.The bare device demonstrated a Jsc of 4.99 mA•cm −2 , a Voc of 0.10 V and a FF of 0.26, leading to a PCE of 0.14%.The device made with a coating of Mg(OH)2 for 30 s on the FTO/TiO2 electrode showed a PCE of 0.05%, with a significant decrease in both the Jsc (from 4.99 to 3.75 mA•cm −2 ) and Voc (from 0.10 to 0.06 V).Although an improvement in PCE was observed when the electrodeposition time increases from 2 to 4 min, the device efficiency is still inferior to the performance of the bare solar cell (Table 2).In device C, Mg(OH) 2 was electrodeposited after coating the mesoporous TiO 2 layer without a compact TiO 2 layer and the MgO layer was replaced with TiO 2 and ZrO 2 layers in the device D. As shown in Figure 4a and Table 2, device C demonstrated poor device performance compared to the device A. This is probably due to the higher recombination rate at the TiO 2 /NiO interface despite the MgO layer on TiO 2 .The bare device demonstrated a J sc of 4.99 mA•cm −2 , a V oc of 0.10 V and a FF of 0.26, leading to a PCE of 0.14%.The device made with a coating of Mg(OH) 2 for 30 s on the FTO/TiO 2 electrode showed a PCE of 0.05%, with a significant decrease in both the J sc (from 4.99 to 3.75 mA•cm −2 ) and V oc (from 0.10 to 0.06 V).Although an improvement in PCE was observed when the electrodeposition time increases from 2 to 4 min, the device efficiency is still inferior to the performance of the bare solar cell (Table 2).
Coatings 2017, 7, 36 Following the meso-super-structured solar cell proposed by Snaith et al. [3], we fabricated device D with FTO/Mg(OH)2/NiO/Carbon(MAPbI3) configuration.In this structure, MgO acts as a scaffold for MAPbI3.The I-V characteristics of the devices with different electrodeposition times of Mg(OH)2 are shown in Figure 4b.It was clear that up to 10 min electrodeposition time of Mg(OH)2, the diode characteristic did not appear due to the shorting resulted by the low thickness of the MgO layer.However, the diode characteristic emerged above the 10 min electrodeposition time of Mg(OH)2.The device made with the 20 min deposited MgO layer showed a Jsc of 0.32 mA•cm −2 , a Voc of 0.12 V, a fill factor of 0.27 and a PCE of 0.01% (Table 2).However, the higher thickness of MgO does not improve the photocurrent density of the device as expected, which could be due to the higher sheet resistance that results from the thick insulating MgO layer at the higher deposition times.
Overall, the results confirmed that the MgO coating on FTO/TiO2 (device B and C) or FTO (device D) in printable PSCs does not show much benefit in improving the PCE compared to the bare devices, which could be due to the higher coverage of TiO2 by the insulating MgO layer attained by the electrodeposition of Mg(OH)2.

Conclusions
In summary, we studied the effect of MgO coating on the photovoltaic performance of printable PSCs.Electrodeposition of Mg(OH)2 was conducted on the surface of mesoporous TiO2 on FTO in devices A, B and C and FTO in device D. The effect of electrodeposition time on the performance of printable PSCs was evaluated on the basis of their key cell parameters.The overall results confirmed  The device made with the 20 min deposited MgO layer showed a J sc of 0.32 mA•cm −2 , a V oc of 0.12 V, a fill factor of 0.27 and a PCE of 0.01% (Table 2).However, the higher thickness of MgO does not improve the photocurrent density of the device as expected, which could be due to the higher sheet resistance that results from the thick insulating MgO layer at the higher deposition times.
Overall, the results confirmed that the MgO coating on FTO/TiO 2 (devices B and C) or FTO (device D) in printable PSCs does not show much benefit in improving the PCE compared to the bare devices, which could be due to the higher coverage of TiO 2 by the insulating MgO layer attained by the electrodeposition of Mg(OH) 2 .

Conclusions
In summary, we studied the effect of MgO coating on the photovoltaic performance of printable PSCs.Electrodeposition of Mg(OH) 2 was conducted on the surface of mesoporous TiO 2 on FTO in devices A, B and C and FTO in device D. The effect of electrodeposition time on the performance of printable PSCs was evaluated on the basis of their key cell parameters.The overall results confirmed that the insulating MgO coating has an adverse effect on the photovoltaic performance of the solid state printable PSCs.We believe that this adverse effect on the photovoltaic performance of the printable PSCs is due to the higher coverage of TiO 2 by the insulating MgO layer attained by the electrodeposition of Mg(OH) 2 .

Figure 2 .
Figure 2. Surface topographical images of electrodeposited Mg(OH)2on the FTO substrate for (a) 2 min and (b) 10 min (inset shows the higher magnification), cross-sectional images of PSCs device B (c) and device D (d) and EDX mapping images of the MgO layer (e) NiO layer (f) and Carbon (g) in the device D.

Figure 2 .
Figure 2. Surface topographical images of electrodeposited Mg(OH) 2 on the FTO substrate for (a) 2 min and (b) 10 min (inset shows the higher magnification), cross-sectional images of PSCs device B (c) and device D (d) and EDX mapping images of the MgO layer (e), NiO layer (f) and Carbon (g) in the device D.

Figure 3 .
Figure 3. J-V characteristics of printable PSCs (device A) with different electrodeposition times of Mg(OH)2 (device B).

Figure 3 .
Figure 3. J-V characteristics of printable PSCs (device A) with different electrodeposition times of Mg(OH) 2 (device B).

Figure 4 .
Figure 4. J-V characteristics of printable PSCs of device C (a) and device D (b) with different electrodeposition times of Mg(OH)2.

Figure 4 .
Figure 4. J-V characteristics of printable PSCs of device C (a) and device D (b) with different electrodeposition times of Mg(OH) 2 .

Table 1 .
Summary of the average solar cell parameters of device A and B against the electrodeposition times of Mg(OH)2.

Table 1 .
Summary of the average solar cell parameters of devices A and B against the electrodeposition times of Mg(OH) 2 .

Table 2 .
Summary of the average solar cell parameters of device C and D against the electrodeposition time of Mg(OH)2.

Table 2 .
Summary of the average solar cell parameters of devices C and D against the electrodeposition time of Mg(OH) 2 .