Khalid
Mahmood
*a,
Saad
Sarwar
b and
Muhammad Taqi
Mehran
b
aDepartment of Chemical & Polymer Engineering, University of Engineering & Technology Lahore, Faisalabad Campus, 3½ Km. Khurrianwala – Makkuana By-Pass, Faisalabad, Pakistan. E-mail: khalid@kaist.ac.kr
bUniversity of Science and Technology (UST), 217-Gajeong-ro, Yuseong-gu, Daejeon, 34113, Republic of Korea
First published on 17th March 2017
Methyl ammonium lead halide-based hybrid perovskite solar cells (PSCs) have been intensively studied in recent years because of their high efficiency and low processing costs. Although there are limited constraints for choosing the planar electron transport layer (ETL) or mesoscale electron transporting material (ETM), a great deal of effort is required in designing complex nanostructures which are effective as ETL/ETM to achieve high open circuit voltage (Voc) and high fill factor (FF) in PSCs. In this review, various inorganic and organic ETLs, as well as inorganic ETM systems, used for PSCs are summarized. The transport mechanism of electrons in these different ETL/ETM materials is discussed along with their effect on Voc on the basis of energy band diagrams with respect to the perovskite absorber. The authors also discuss the microstructure/nanostructure aspect of mesoscopic ETMs, doping and surface functionalization, and the influence of these parameters on solar cell behaviour, performance, and hysteresis effects. The authors also discussed the microstructure/nanostructure aspect of ETL on shape of current density vs. voltage (J–V) hysteresis in this review. Technical issues and recent progress of ETL to improve device efficiency and stability in terms of materials, process, and characterization are summarized.
To date, several electron transport layers (ETLs) have been reported, including metal oxides (mainly transition metal oxides), metal oxide composites, and metal oxide heterostructure nanoparticles.
The crystal structure of hybrid perovskite materials and hybrid PSCs is well known, and the crystal structure analysis has been discussed thoroughly by different groups.1 There have been more than 500 articles published within a 5 year span with different device geometry from the concept of dye sensitized solar cells (DSSC)2 to p–n/p–i–n geometry,33 bulk hetero-junction solar cells,34–36 and inverted planar structure.37–40 Efficiency increased in order from the initial PSC invented by Miyasaka's group.2 Later work by different groups improved the efficiency by using different materials from the bottom to the top electrode. This was found to be the easiest way to improve the power conversion efficiency (PCE) by increasing the open circuit voltage (Voc) and short circuit current (Jsc) in the early stages of device research.
The ETL, sometimes called the electron extraction layer or electron collection layer, is where electrons are injected from the absorber layer, transported through the electron transporting materials (ETM), and finally collected by the electrode. Sometimes the ETL is replaced by a composite photoelectrode to improve the charge collection efficiency by lowering the series resistance.34–36 Moreover, a one-dimensional nanorod ETL provides fast charge transport in a photovoltaic cell. Excellent pore filling achieved by manipulating the gaps between the nanorods and nanosheets was found to be beneficial for enhancing the fill factor (FF) and PCE.41–52 The gaps between the nanorods can be controlled by the concentration of the solution and seed nanoparticle films. The efficient charge collection leads to ideal current–voltage behaviour. This is a method by which to improve the external quantum efficiency (EQE), giving maximum efficiency without thermalization.
A schematic diagram of device architecture of the PSC is presented in Fig. 1. The simple PSC consists of an electron transport layer/perovskite/hole transport layer on the FTO/glass substrate. Low and high work function electrodes can be used at the ETL and hole transport layer (HTL) sides, respectively. The balanced diffusion length of electron and hole is more than 500 nm in a perovskite absorber, which inhibits charge recombination in the perovskite absorber.25,26 Recently, Dong et al. reported electron–hole diffusion length of more than 175 microns in CH3NH3PbI3.27 Therefore, a smooth electron injection requires after charge separation for further process. Sometimes a compact hole blocking layer is used before the ETL,53,54 and an electron blocking layer55,56 is used before the HTL to block the excess hole and electron at either side, respectively. Some earlier studies focused on compact hole blocking layer (HBL) being beneficial for PCE by blocking the hole from the ETL side.53,54 The electron and hole clouding in the device architecture leads to a high recombination rate. Using a planar hetero junction PSC, the PCE can be significantly improved by provision of a proper path for both electron and hole.
In this review, we discuss the material and electrical aspects of the importance of the ETL along with their device physics. There is evidence that the exciton might dissociate at the ETL/perovskite and perovskite/HTL interfaces. The other factor is exciton diffusion and separation, which is not clearly understood in PSCs. Low exciton energy in perovskite appears to have benefit for high exciton diffusion length, which is an advantage over polymer/organic solar cells. Another important factor is that PSCs fit to inorganic as well as polymer/organic ETL and HTL. Inorganic ETL/HTL has a longer electron/hole diffusion length compared with organic/polymer ETL/HTL. It has been reported that hybrid solar cells do not necessarily need a HTL but do need an ETL/scaffold.57 So choice of the proper ETL for PSC is vital. Several reviews have been reported within a short period of time on PSC research.58–66
In this review, we summarize different ETL used in hybrid PSC and explore how design of ETL helps to improve the overall device performance. We also discuss how to improve solar cell efficiency in terms of improving charge injection, charge collection, charge separation, and charge transport in ETM. These parameters are interlinking to achieve the overall performance. We also address the effect of microstructure and morphology of ETL materials and how to achieve better pore filling, and composite structure to achieve high electron injection and transport on open circuit voltage. Finally, we review the different processing methods of ETL and their advantages and disadvantages. Herein, the most recent results of this particular optimization process such as annealing temperature and annealing time are reviewed, and how certain ETM retard the charge recombination.
A complete list of electron, hole injection and exciton annihilation processes is described by Luo et al.66 The large variety of ETL/ETM makes it hard to define the characteristics of the ideal ETL. In terms of considering the band offset between the ETL/perovskite, Minemoto et al. theoretically calculated that the band offset between the ETL/absorber and HTL/absorber should be 0.0–0.3 eV, as shown in Fig. 2.67 Fig. 2 shows simulated J–V curves of PSC with different CBO values for the buffer/absorber interface.
Fig. 2 J–V curves of perovskite solar cells with different CBO values of buffer/absorber interface. Here, the VBO of the absorber/HTM interface was set to 0 and the interface defect density of IDL 1 was set to 1016 cm−3.67 |
To date, TiO2 is the ETL material most widely used by different research groups. But, Snaith et al. pointed out that a TiO2-based PSC had lower stability than a TiO2-free PSC.75 They studied the stability of ss-DSSC and PSC using transient absorption spectroscopy (Fig. 3). Snaith's group initiated the idea that the absorption tails of mesoporous TiO2 extend into the visible region, which may degrade the stability over a longer timescale.76 They suggested that a low cost UV filter could efficiently block the UV and enhance the stability for a longer period of time.
Ultra-thin amorphous TiOx provides uniformity and homogeneity in short range order, giving more highly stable electronic properties than mesoporous TiO2. A 12.2% efficient PSC with high bending and durability was achieved using 20 nm of atomic layer deposited (ALD) TiOx by Kim et al.18
Cadmium selenide (CdSe) has a band gap of 1.8 eV, with HOMO 0.5 eV lower than the perovskite active layer and LUMO 0.6 eV higher than the perovskite active layer. CdSe nanocrystal used as an electron transport/extraction layer for PSC shows PCE up to 11.7%.78 Strong PL quenching has been observed in CdSe/CH3NH3PbI3/Spiro-OMeTAD. The PL quenching effect is necessary in PV devices for better charge separation and recombination resistance. SnO2 shows better environmental stability than TiO2, as reported by Song et al.79 Using the concept of light scattering efficiency, Park et al. used TiO2 nanohelix ETL deposited by oblique-angle e-beam evaporation to produce hysteresis-free PSC.80 The drawback of using TiO2 in PSCs is that it must be annealed at high temperature, which can be a hindrance in commercialization as well as use in flexible devices as the substrate cannot withstand such high temperatures. While mesoporous TiO2 incorporated PSCs have shown good efficiency, their counterpart, the planar type PSC typically suffers from hysteresis issues. Recently several groups have addressed this issue by replacing the TiO2 with low temperature processed SnO2 in planar structures. SnO2 has advantages over TiO2 in terms of its high electron mobility, anti-reflection property, and deeper conduction band, which could increase charge transfer from perovskite to ETL and reduce charge accumulation at the interface. Weijun Ke et al. deposited SnO2 by spin-coating precursor solutions of SnCl2·H2O and heating in the air at 180 °C.147 A high efficiency of 17.21% was achieved using this method. Although less than in the planar TiO2 structure, hysteresis was still present in PSCs made using this method. Qi Jiang et al. obtained a certified high efficiency of 19.9% by employing solution-processed SnO2 at a low temperature of 150 °C, which was almost free of hysteresis.148 However, passivation of the PbI2 phase in the perovskite layer was introduced in this study, which might have played a role in such a high efficiency of the device. Researchers have also employed atomic layer deposited (ALD) SnO2 in a PSC, which served as the top cell in a tandem device. An efficiency of 20% was obtained with this configuration. However, as ADL is not cost-effective commercially, the researchers employed a combination of spin-coating and chemical bath deposition of SnO2 to form the ETL.149 Devices made using this method showed better performance than devices made by ALD, mainly because of an improved FF. The higher roughness of the SnO2 achieved from the new technique results in better contact of the ETL and perovskite layer, thus increasing the FF. The devices showed a mild hysteresis, although much better than that of the planar TiO2 devices, and showed a highest efficiency of 21% (planar structure). The devices retained 82% of efficiency after ageing for 60 h under illumination of 100 mW cm−2. This is similar to the ageing behaviour of high temperature mesoporous films. A steady-state efficiency of 17.54% was also obtained using an effective interface engineering technique for planar PSCs employing SnO2 as an electron selective layer.150
Fig. 4 TR-PL decays of bare CH3NH3PbI3 deposited on Pyrex glass and the FTO/TiO2/CH3NH3PbI3 devices employing NP30, NP40, NP50, and NP65. It can be seen clearly that the TR-PL decay is accelerated with decrease in the size of TiO2 NPs. The time resolution of the TR-PL measurement was ∼190 ps.77 |
Fig. 5 Current extraction and transient absorption spectroscopy of TiO2-based solar cells. In the top panel (a), the current collected from full perovskite-sensitized TiO2 solar cells with semi-transparent electrodes (15 nm gold) after a 10 ns laser pulse (550 nm) is plotted against time after the excitation. Non-encapsulated (blue), encapsulated after ambient light exposure (grey), and encapsulated after UV aging (red) (5 min at 40 mW cm−2, 365 nm) are compared. The area underneath the curves represents the total amount of collected charge. In the bottom panel (b), the transient absorption trace at short circuit of the hole on Spiro-OMeTAD is probed at 510 nm, performed simultaneously with the current extraction. The lines are merely guides to the eye. The table in the figure shows the charge generated (G, estimated from the early time transient absorption signal), collected (C, by integrating the current decay), and the ratio between the two as a percentage (C/G). The laser intensity was 80 mJ cm−2, giving charge densities similar to those found in devices under operating conditions.75 |
Fig. 6 UV-Vis spectra of sensitized films. (a) Film of TiO2/CH3NH3PbI3 before and after degradation, (b) Film of TiO2/CH3NH3PbI3/Al2O3 before and after degradation. The degradation time is 18 h.83 |
Fig. 7 (a) Electron lifetime as a function of open circuit voltage in Spiro-OMeTAD/CH3NH3PbI3/ZrO2 (or TiO2) solar cells. (b)84 Photocurrent transient decay occurs as a function of short-circuit current in Spiro-OMeTAD/CH3NH3PbI3/ZrO2 (or TiO2) solar cells.84 |
Using an ultra-thin Al2O3 layer, excellent moisture and humidity stability was demonstrated by Dong et al.83 Using computational analysis along with experimental results, they proved that thin layer Al2O3 acts an excellent absorber of H2O from the perovskite layer.
The ETL provides a scaffold material for the perovskite layer to crystallize well to form pinhole-free compact films. As is known, the crystallization and nucleation behaviour of the perovskite layer depends on the nucleation site of the substrate, so careful treatment of ETM is crucial to obtain a uniform, compact, and smooth perovskite layer with lesser voids to avoid unnecessary leakage of current. In addition to use of different ETM discussed above, Ke et al. claimed that PSC do not necessarily need ETM. Ultraviolet–ozone treatment of ITO substrates, with a planar junction architecture processed by a solution method, has achieved a PCE of over 14% and an open-circuit voltage of 1.06 V (Fig. 8).96
Fig. 8 The PL spectra of CH3NH3PbI3, CdSe/CH3NH3PbI3, CdSe/CH3NH3PbI3/Spiro-OMeTAD on glass substrates. The excited wavelength is 600 nm.78 |
With an ETL, a high conversion efficiency of ∼8.39% along with a high short circuit current (Jsc) of ∼15.24 mA cm−2, an open circuit voltage (Voc) of ∼0.830 V, and a high fill factor (FF) of ∼0.66 was obtained using Ti-coated ITO/PET substrate. Intensity-modulated photocurrent (IMPS)/photovoltage spectroscopy (IMVS) studies demonstrated that the fabricated flexible PSC had considerably reduced recombination rate and higher charge transfer rate.97
Fig. 9 Block diagrams of mesoscopic photoanodes with thinner (left) and thicker (right) ALD TiO2 overlayer (red) deposited on the mesoporous nanoparticle TiO2 (light blue) films. The np-TiO2-ALD TiO2 film is infiltrated with CH3NH3PbI3 absorber (green). For thicker ALD TiO2 overlayers, the pore diameter is reduced and improper pore filling of the perovskite absorber is highlighted (right).98 |
Without an ETL, electrons suffer a sudden voltage drop which may degrade the device performance. A novel ITO/TiO2 core–shell nanostructured photoanode was fabricated by Han et al.99 using VLS grown ITO nanowire with ALD deposited TiO2, which gave better electron injection and charge collection efficiency compared with TiO2 ETL. At the same time, this kind of nanostructure gives lower series resistance to the devices, with a lesser diffusion path for charge carrier, and acts as both photoanode and ETL for DSSC as well as hybrid PSC. The same concept was implemented using Al2O3/ZnO nanoparticle to improve the efficiency in PSC.35 Such nanoarchitectures could be implemented for ITO/ZnO, ITO/Al2O3 and ITO/V2O5, etc., because these metal oxides have already been tested as efficient ETL materials and can be deposited by the ALD process.
WO3 is not considered to be an efficient ETM for DSSC.100,101 In comparison with conventional TiO2 ETL, WOx exhibited comparable light transmittance but had higher electrical conductivity. PSCs using WO3 ETL exhibited comparable photoelectric conversion efficiency with TiO2-based PSCs, larger short-circuit current density, and high lower open-circuit voltage.102 However, if a thin layer of TiO2 is coated on the WO3, better electron injection can be observed in comparison with the WO3 alone. This leads to high charge collection efficiency and retards the recombination in PSC. Mahmood et al. reported a mesoscopic WO3/TiO2 nanostructure to get a smooth electron injection to get high Voc and PCE.47 As the FF depends on the series resistance, shunt resistance, and diode characteristics, careful choice of nanostructure, thickness, and pore filling could provide improved FF. A recorded Voc (approx ∼1.3 V) was achieved by tailoring both a low-lying HOMO level as well as the appropriate electronic and optical properties by Edri et al.57 using a TiO2/alumina/CH3NH3PbBr3/PDI structure. Their efficiency was limited, however, to 0.56%.
A high band gap material on m-TiO2 ETL would have reduced the charge recombination rate. PSCs are sensitive to moisture and humidity. It is believed that H2O can decompose CH3NH3PbI3 into PbI2 and CH3NH3I, which can degrade performance of PSC.82,103 The surface of m-TiO2 is prone to H2O/hydroxyl groups, which interact with the NH4 and CH3 in the perovskite layer, and degrade the charge transport.104 Long exposure of m-TiO2 to air and pre heating or UV treatment before fabrication of the perovskite layer can result in the above interface problem. Hence, some previous reports advised TiCl4 treatment, which can prevent such effects in PSC.53 Using a thin layer of MgO on m-TiO2, the recombination resistance can be reduced (Fig. 10).53
Fig. 10 EIS of PSC with and without MgO coating under (a) light illumination and (b) dark conditions.104 |
Charge retarding behaviour using heterostructure core–shell nanoparticles is well known in DSCC. Using MgO/TiO2, Al2O3/TiO2, ZnO/TiO2 etc., electron–hole recombination can be suppressed.72,73 MgO-coated TiO2 nanoparticle (NP)-based electron collecting layers were fabricated by Han et al., to prevent charge recombination at the perovskite/TiO2 interface in PSC.73 Open circuit voltage (Voc) and fill factor (FF) with MgO/TiO2 ETL are 4.7% and 6.1% higher, respectively, than with mesoporous TiO2 ETL, and the recombination times for MgO/TiO2 NP-based PSC are about three times longer than that of TiO2 NP-based solar cells. Longer recombination time was achieved by adding an MgO thin layer which retards the electron–hole recombination.71 The detailed mechanism is shown in Fig. 11.
Fig. 11 (a) Schematic representation of the charge transfer mechanism in TiO2-based perovskite solar cell with a MgO layer, band energy diagram of the MgO/TiO2 core–shell NP-based perovskite solar cell, (b) electron lifetime of MgO/TiO2 core–shell NP-based perovskite solar cells compared with TiO2 NP-based perovskite solar cells and (c) open circuit voltage (Voc) and fill factor (FF) of the MgO/TiO2 core–shell NP-based perovskite solar cell with varying magnesium methoxide concentrations (0.1 wt%, 0.5 wt%, and 2.0 wt%).73 |
N-type doping leads to an increase in the conduction band edge which can efficiently decrease the charge recombination process. Several n-type dopings have been reported in TiO2 as well as ZnO. Pure anatase mesoporous TiO2 was used as ETL material in the initial stages of PSC research. The band gap of TiO2 is about 3.2 eV with HOMO and LUMO situated at 3.2 eV and 5.4 eV, respectively, which is best fit to the energy level diagram. However, doping to the TiO2 improves the conduction band edge as well as the band gap TiO2. For this reason, n-type doping in ETL gives a smooth electron injection from the perovskite layer to ETL (Fig. 12).
Fig. 12 Incident photon-to-electron conversion efficiency spectra of devices based on 0.5% Y-TiO2 (black) and TiO2 (red). The inset shows the UV-Vis absorbance spectra of 0.5% Y-TiO2/CH3NH3PbI3 and TiO2/CH3NH3PbI3 electrodes.34 |
In this way, the Voc can be improved to shorten level, which would raise the FF, and consequently the PCE. n-type doping in TiO2 can be achieved using a number of elements. Using Sn doped TiO2, Zhang et al.51 tuned the band gap of TiO2 and improved the efficiency up to 67%. By n-type doping, the conduction band edge can be increased, which can suppress recombination of electron and hole and provide a smooth electron injection. Using this concept, many n-type doped TiO2 have been used to get higher Voc. Manseki et al. used Mg doped TiO2 as ETL to increase the band gap of TiO2 and CB edge, leading to improved Voc.52 Their results showed that the band gap increased up to 1 eV which affects the Voc, increasing it from 587 meV to 802 meV. Electron-rich, nitrogen doped ZnO (N:ZnO) NR-based electron transporting materials (ETMs) with enhanced electron mobility produced using ammonium acetate show consistently higher efficiencies by one to three power points than undoped ZnO NRs.107 Using Pb doped TiO2 nanofibre, Xiao et al. reported 9.03% efficient PSC with an HTL.108 Using Al-doped ZnO (AZO) modified ZnO nanorods as ETL, enhanced Voc and PCE were achieved (Fig. 13).109
Fig. 13 (a) Nyquist plots of the device treated with AZO at different Al doping levels at applied bias (600 mV) under dark conditions. (b) Equivalent circuit employed to fit the IS data. Plots of (c) the series resistance (RS) and (d) the charge transport resistance (Rrec) obtained from IS results as a function of the applied bias.109 |
For flexible PSCs, ALD deposited TiO2, ZnO, SiO2, HfO2 and ZrO2 are better options for growth on flexible substrate. However, other methods such as hydrothermal, spin coating, electrospray method spray deposition are good for PSC fabricated on rigid substrates. Using a hydrothermal method, the microstructure can be easily tuned to tune the optical properties and device architecture. The electrospray method is a better choice for growth of many novel nanostructures with multilayer structure for sensitized and solid state PSC.110 Electrospray has many process parameters, and can adjust the thickness as well as morphologies. Magnetron sputtering grown compact ZnO as ETL was deposited by Liang et al.48 They observed 13.4% and 8.03% PCE on ITO/glass and ITO/PET substrate. The sheet resistance of ZnO/ITO/glass was limited to 8.0 ohm per cm2, which gave a low series resistance to the device. At the same time, compact ZnO films by magnetron sputtering mean that the shunt resistance would be high enough to give such high efficiency. A complete list of fabrication methods and their electronic properties is given in Table 1. The HOMO, LUMO, and band gap are also listed and the band energy diagram with respect to perovskite is tabulated. As per theoretical study by Minemoto et al.,67 the band offset between the ETL and perovskite layer should be in the range of 0.0–0.3 eV, to give an optimum Voc, which regulates the FF in PSC. This information is useful for fabricating highly efficient PSC by choosing an optimum thickness, microstructure, doping, and device architecture. Band gap tailing by Nb doping in TiO2 and faster electron injection and electron lifetime was reported by Kim et al.111 They showed light Nd doped TiO2 to have the highest efficiency of 13.4% (Fig. 14).
Solar cell type | V oc (mV) | J sc (mA cm−2) | FF (%) | PCE (%) |
---|---|---|---|---|
PSC-30 | 995 | 21.61 | 66.59 | 14.32 |
PSC-40 | 1029 | 21.45 | 69.74 | 15.24 |
PSC-50 | 1049 | 21.58 | 75.94 | 17.19 |
PSC-65 | 1005 | 21.27 | 72.35 | 15.47 |
Fig. 14 (a) Nyquist plot of the PSCs measured at open-circuit conditions. The impedance spectra measured in the frequency range of 10 mHz to 2 MHz at open circuit potential under 1 sun illumination. (b) tR of the PSCs as a function of Voc. The detailed measurement conditions of tR are the same as those of the tc measurements (Fig. 6).111 |
Polymer | Device structure | J sc (mA cm−2) | V oc (V) | FF | PCE (%)Ref |
---|---|---|---|---|---|
PEI/Y-TiO2 | ITO/PEIE/CH3NH3PbICl/Spiro-OMeTAD/Au | 22.75 | 1.13 | 0.750 | 19.03 (ref. 71) |
PEI/N:ZnO | ITO/ZnO/PEI/CH3NH3PbI3/SpiroOMeTAD/Au | 21.5 | 9.645 | 0.700 | 16.1 (ref. 107) |
ZnO/PCBM | ITO/PEDOT:PSS/CH3NH3PbI3−xClx/PCBM/ZnO/Ag | 16.0 | 0.95 | 0.745 | 11.3 (ref. 145) |
Perylene-diimide/PCBM | ITO/PEDOT:PSS/CH3NH3PbI3−xClx/PCBM/PDINO/Ag | 18.8 | 0.95 | 0.785 | 14.0 (ref. 145) |
PEI/PCBM | FTO/PEI/PCBM/MAPbI3/PTAA/Au | 21.8 | 0.98 | 0.720 | 15.3 (ref. 138) |
Rhodamine 101/LiF | ITO/PEDOT:PSS/CH3NH3PbI3−xClx/rhodamine 101/LiF/Ag | 17.72 | 1.01 | 0.730 | 13.2 (ref. 134) |
PCBM:polystyrene | FTO/NiO/CH3NH3PbI3−xClx/PCBM:PS)/Al | 15.62 | 1.07 | 0.640 | 10.68 (ref. 39) |
PN4N | ITO/PEDOT:PSS/CH3NH3PbI3−xClx/PCBM/PN4N/Al | 20.61 | 1.0 | 0.725 | 15 (ref. 146) |
PFN | ITO/PEDOT:PSS/CH3NH3PbI3−xClx/PCBM/PFN/Al | 20.47 | 0.98 | 0.552 | 8.7 (ref. 146) |
Fig. 15 Schematics of device structure based on (a) conventional low aspect ratio (LAR) NR arrays, and (b) high aspect ratio (HAR) NR arrays with additional PEI coating (the inset shows the schematic representation of PEI as a capping agent to control nanorod growth). Energy diagrams of the devices based on (c) LAR or HAR NRs without PEI coating and (d) LAR or HAR NRs with PEI coating. The conduction band energy of ZnO was lowered by 0.07 eV by nitrogen doping and by 0.37 eV by dipole formation by the PEI layer.107 |
Generally, the PSC needs a mesoporous ETL, compact ETL, or a scaffold material for growth of the perovskite absorber. The most common practice is to use a commercially available mesoporous TiO2 layer, which has superior electrical properties because the mesoporous TiO2 is deposited easily by spin-coating, but the thickness and roughness cannot be easily controlled. Also, perovskite is sometimes inserted into the TiO2 layer and this cannot be distinguished using a cross-sectional scanning electron microscope. A number of reports describe such intermixed TiO2 and perovskite layers using the 1S and 2S methods. The microstructure of CH3NH3PbI3 has particle size in the range of 100–500 nm with large pore size. When the HTM is deposited on the perovskite layer, the HTL may insert into the gaps, touching the ETL and providing a shunt path. Consequently, the efficiency of the solar cell is degraded. The effects of morphology and pore filling of mesoporous TiO2 layer and perovskite layer in terms of device performance were studied by Nanova et al.69 TEM-EELS (electron energy loss spectroscopy) measurements were used to study the mechanism and optimum pore structure of TiO2/perovskite. A finer grain of TiO2 efficiently accompanies the perovskite, resulting in an interlinked structure and better PCE. Hence, an over-layer perovskite absorber is needed to overcome the problems described above (Fig. 16).
Fig. 16 Correlation between morphology and device characteristics. (a) Device 2 exhibits a compact and well-connected TiO2 network. (b) The TiO2 in device 3 is a much finer grained structure (very small and unconnected red domains). (c) I–V characteristics of the devices. The higher fill factors of devices 1 and 2 can be correlated to the well-connected TiO2 network and hence better electron transport properties.66 |
Fast electron transport and charge collection is possible using nanostructured ETL materials such as nanowire, nanosheets, and nanoneedles. These NSs provide better charge injection and unidirectional charge transport. Smooth injection is the key point at the ETL and perovskite interface, which must satisfy the condition that the LUMO of the ETL is lower than that of the perovskite. However, the CH3NH3PbI3 and (Cl, Br) doped CH3NH3PbI3 (CH3NH3PbClx, CH3NH3PbClxBryI1−y) were found to be beneficial because of the optimum band gap giving broad absorption in the UV-Vis region and high extinction coefficient.113–115 A schematic band energy level diagram of hybrid perovskite materials is presented in Fig. 1. To provide a smoother path for the electron the LUMO level of the ETL should be lower than the LUMO level of the intrinsic layer, and the HOMO level should be higher than that of the intrinsic layer.
Nanostructure ETL such as nanowire and nanorods were found to be beneficial for charge collection. A better surface morphology, roughness, and nanoarchitecture give efficient electron injection and a short carrier path for the electron. Transition metal oxides (TiO2, ZnO) can have a variety of nanostructures, and can be easily doped by a number of elements to increase/decrease the Fermi level, which affects the electron injection efficiency and consequently leads to enhanced transport properties and better charge collection efficiency. Park et al. used ZnO nanorods to efficiently collect the electron from the lower wavelength region.46
Using a dense compact TiO2 layer, the series resistance of photovoltaics (PV) was improved by Ke et al.116 Their approach was to minimize the series resistance, giving a higher recombination time of the PSC. They achieved about 15.07% efficiency with Voc up to 1.04 V. Using electrospun deposited TiO2 nanofibre of different TiO2 thicknesses, Dharani et al. reported Voc = 844 mV with PCE = 9.8%.50 They predicted that Voc decreased with increasing fibre thickness.
Hydrothermally deposited one-dimensional TiO2 array rather than a mesoporous TiO2 layer is a better ETL for PSC. A schematic of a 1D TiO2/perovskite solar cell is presented in Fig. 17.50 The schematic gives misleading information that the HTL is in contact with the TiO2 ETL, which might have resulted in high shunt resistance and low recombination time, so degrading the solar cell performance. Hence, the J–V curves did not maintain the current with increasing Voc. At the same time the series resistance is comparatively high using long TiO2 nanorod. A highly efficient 1D TiO2 can be fabricated by proper choice of thickness of TiO2 and doped TiO2 nanorod with excellent pore filling and over layer perovskite material with controlled pore size. Large pore size and particle size of perovskite films allow insertion of the HTL into the TiO2 ETL, giving low shunt resistance which gives low FF.
Control over nanostructure ETL with different morphology and size shows better electron injection, electron transfer, and charge collection efficiency.36,37,100 The thickness and gap between nanorods should be controlled to decrease the series resistance and increase the shunt resistance. A non-porous ETL needs a thin compact hole blocking layer for better film growth. Increasing the nanowire length provides higher series resistance and also gives a large carrier path to extract the electron. If the nanowire size can be controlled, the perovskite can be infiltrated into the nanowire gaps, which would compensate the series resistance by several orders. At the same time, a proper morphology must be considered to obtain high efficiency. The microstructure cannot be easily controlled using the spin coating process. Extra additives or doping is required to initiate a large number of nucleation sites which can grow nanocrystalline perovskite nanoparticles with high surface coverage. This method allows high shunt resistance to get high fill factor with high external quantum efficiency (EQE).
By controlling the photoelectrode thickness and perovskite thickness, Dhar et al. efficiently controlled the charge injection to the photoelectrode. A PSC of 12% efficiency was fabricated using novel (001)-oriented nanoplatelets of anatase thin film as the photoanode.117 A thicker ETL material provides a longer diffusion path for the electron to reach the electrode, which limits the charge collection efficiency and also transmittance of the incident photon decreases with increasing thickness. So when designing an ETL, the optimum thickness should be considered. Oh et al. reported ternary Zn2SnO4 (ZSO) as the electron-transporting electrode of a CH3NH3PbI3 PSC as an alternative to the conventional TiO2 electrode.118–120 They obtained different thicknesses of ZSO compact layer or ZSO mesoscopic layer using the spin coating method.118 They observed that transmittance of ZSO films had negligible effect on increasing thickness. In addition, time-resolved and frequency-resolved measurements revealed that the ZSO-based PSC exhibits faster electron transport (∼10 times) and superior charge collection capability compared with its TiO2-based counterpart with similar thickness and conversion efficiency. Using Zn2SnO4/PVP nanofibres, Mali et al. reported 7.3% efficiency (Fig. 18).
Fig. 18 Electron diffusion coefficients of the ZSO-based perovskite solar cell with the highest conversion efficiency. Electron diffusion coefficients of a TiO2-based perovskite solar cell with similar thickness (∼300 nm) and conversion efficiency (∼7%) are also presented for a comparison.118 |
Using ZnO/PCBM ETL in a PHJ PSC, trap-assisted recombination was lowered by several orders.123 Interface modification plays a crucial factor by passivation or deposition of polymer on inorganic ETL, found to enhance the efficiency of PSC by improving the electron transport/extraction. The aim of using PCBM on ZnO123 and TiO2 (ref. 124 and 125) is to adjust the quasi Fermi level of ETL and HTL, as the Voc of the PHJ PSC is decided by the energy-level offset between the quasi Fermi levels of the ETM and the HTL adjacent to the perovskite active layer.126,127 In addition, efficiency can also be enhanced by use of an induced dipole moment to enhance charge collection of electron and reduce charge recombination (Fig. 19).
Fig. 19 (a and c) Photocurrent and photovoltage decay of devices employing the TiO2 ETL incorporated with different acetylacetonate-based additives and (b and d) the related carrier transport lifetime and electron lifetime. Intensity is shown in log scale and 1 sun condition is equivalent to 1000 W.121 |
The role of the ETM/ETL is to collect electrons efficiently and retard the hole toward the HTL. Efficient charge injection of electron or hole can be analysed by time-resolved photoluminescence decays and the photoluminescence quantum efficiencies of the perovskite films on ETL. With a BHJ PSC architecture, using PC61BM/CH3NH3PbI3:PC61BM composite thin films, the device performance was improved by means of excellent charge collection. Fig. 21 shows the photoluminescence (PL) spectra of the pristine CH3NH3PbI3, CH3NH3PbI3:PC61BM, and PC61BM/CH3NH3PbI3:PC61BM thin films. PL quenching was observed in CH3NH3PbI3:PC61BM and PC61BM/CH3NH3PbI3:PC61BM composites compared with CH3NH3PbI3, which confirms efficient exciton dissociation at “donor/acceptor” interfaces and subsequently electron extraction. With this concept, reduced electron–hole pair recombination and faster charge extraction/collection was obtained in the conventional BHJ PSCs.124 At the same time, this composite film shows higher absorbance than the CH3NH3PbI3 (Fig. 20–22).
Fig. 20 (a) Light-intensity dependence of the Voc of the perovskite solar cells with ZnO and ZnO/PCBM. The lines represent the best fits to the data, for which the slope is indicated. (b) Recombination resistance at different applied biases extracted from IS characterization of the perovskite solar cells.128 |
Fig. 21 Photoluminescence (PL) spectra of the pristine CH3NH3PbI3, CH3NH3PbI3:PC61BM, and PC61BM/CH3NH3PbI3:PC61BM thin films.124 |
Fig. 22 Absorption spectroscopy of CH3NH3PbI3, CH3NH3PbI3:PC61BM, and PC61BM/CH3NH3PbI3:PC61BM thin films.124 |
Fig. 23 (a) Photovoltaic performance of 25 individual devices (red: devices with popcorn NPs; black: control devices). (b) J–V curves of the best-performance perovskite device with and without popcorn NPs, respectively.130 |
In a similar way, a DMAPA-C60 layer on PCBM lowers the work function (WF) by inducing an interface dipole moment, creating a selective quasi-ohmic contact at the interface with Ag (Fig. 24).131 This formation of a dipole layer by either PEI or DMAPA-C60 directly impacts on the WF of electrode and suppresses the energy barrier between the ETL and electrode.104
A remarkable FF of 86.7% with enhanced electron extraction efficiency was achieved by bulk heterojunction (BHJ) pero-HSCs fabricated by blending perovskite materials with fullerene derivatives A10C60 (Fig. 25).132
Fig. 24 Different polymer additives used in organic electron transport layers.131 |
Fig. 25 (a) PL spectra of CH3NH3PbI3, CH3NH3PbI3:A10C60, CH3NH3PbI3/PC61BM and CH3NH3PbI3:A10C60/PC61BM thin films; insets, enlarged PL spectra of CH3NH3PbI3/PC61BM and CH3NH3PbI3:A10C60/PC61BM thin films. (b) Nyquist plots of PHJ and BHJ pero-HSCs measured in the dark and at an applied voltage close to the open-circuit voltage of the pero-HSCs. (c) Recombination resistance resulting from impedance spectrometry measurements fitted for different applied biases at 1 sun illumination.132 |
Limitation of the band gap and absorbance in the UV-Vis region and their photoresponses limits the device performance. It has been confirmed that the band gap of MAMeX3 can be tuned by changing M, A, Me, X, or MA; however, devices suffer in terms of Jsc, Voc, and FF. By implementing another photoactive material in a PSC, which could behave as both ETL and active material, better performance would be achieved than using a material with single active layer. Poly(diketopyrrolopyrrole-terthiophene) (PDPP3T) is a typical low-band gap polymer which is used as an active material in PSCs. It also has lower HOMO level than that of hybrid perovskite material, enabling use as the ETL in PHJ or BHJ solar cell.133 By implanting PDPP3T–PC61BM BHJ into PSCs, the absorbance of a cell was successfully enhanced in the long-wavelength region. Fig. 26 shows a comparison of absorbance of CH3NH3PbI3/(PDPP3T–PC61BM) blend films (Fig. 26).
Fig. 26 Comparison of absorption spectra for PDPP3T–PC61BM (1:2) blend film, CH3NH3PbI3 film, and CH3NH3PbI3/(PDPP3T–PC61BM) (1:2) bilayer film.133 |
Even when the band energy level of PCBM matches with cathode, some interfacial issues can arise such as interfacial trap and defect states, which significantly affect the charge collection. Interlayer polymers can overcome these difficulties and help to achieve smooth energy level alignment. These polymers can help electrons to efficiently collect at the cathode of PSCs in cases of energy mismatch between PCBM and cathode. In addition, interlayers can provide better surface roughness for electrode deposition, which ultimately affects the charge collection. It is believed that polymer ETL passivates defects of the perovskite layer, which might not occur when using inorganic ETL/scaffold. Another advantage of polymer ETL is that it does not require high temperature deposition.
In general, a compact TiO2 layer is used as the hole blocking layer (HBL) in m-TiO2 based PSCs. The BCP layer on PCBM ETL acts in the same way as the TiO2 compact layer does in inorganic ETL. By employing double interfacial layers (rhodamine101/LiF), Sun et al. achieved better efficiency than that of PCBM/LiF and PCBM/rhodamine 101.134 Moreover, zwitterions with positive and negative charges on the same molecule can lower the work function of electrodes by formation of interface dipoles.135,136 In the same way, this was achieved using interlayer PEI on TiO2 [Yang] or ZnO [Mahmood].
The addition of PS aids formation of a highly smooth and uniform PCBM on the perovskite active layer that is more effective in preventing undesirable electron–hole recombination between the perovskite layer and the top electrode.39
Fig. 27 J–V curves of the best perovskite solar cells with bare TiO2 (dark gray curve) and glycine-treated TiO2 (red curve).137 |
Fig. 28 Nyquist plots of perovskite solar cells based on bare TiO2 (symbol B) and glycine-treated TiO2 (symbol). Symbols are experimental data and solid lines correspond to the fitting results using the equivalent circuit shown in the inset figure. EIS spectra were recorded using a potentiostat/galvanostat instrument (PGSTAT 302N, Autolab) equipped with a FRA2 module under constant light illumination of 100 mW cm2. A DC applied bias voltage was set at open circuit voltage of the perovskite solar cells with a sinusoidal ac potential perturbation of 15 mV over a frequency range from 1 MHz to 0.01 Hz.137 |
Conjugated polymer electrolytes (CPEs) are another class of efficient interface materials that form the electron dipole moment at the interface of either metal/ETL or ETL/perovskite. These concepts of interface engineering have been regarded as beneficial for improving the device performance by charge collection, extraction/injection. Ryu et al. achieved PCE of about 15% in FTO/PEI/PCBM/MAPbI3/PTAA/Au PHJ PSC.138 To improve efficiency, well-organized mesoporous TiO2 photoelectrodes were fabricated by block copolymer-induced sol–gel assembly by Sarkar et al.139
V oc can be increased by reducing electron–hole recombination occurring at the HTL/ETL inside a perovskite pinhole. Functionalizing the ETL/scaffold by organic/polymer molecules would provide better recombination resistance. Different concentrations of 3-aminopropyl (3-oxobutanoic acid) functionalized silica nanoparticles were used to improve Voc and PCE by a one-step solution process.140 The advantage of functionalized SiO2 NPs over compact TiO2, is that hole transfer is hindered from the perovskite as well as the HTL. The functionalized SiO2 lowers the conduction band edge to give a smooth electron injection to the metal electrode. Absorption of perovskite films was improved in a broad range of light spectrum by employing 3-aminopropyl (3-oxobutanoic acid) functionalized silica nanoparticles at different concentrations, as shown in Fig. 29.
Fig. 29 (a) I–V curves of best performing devices for each f-SiO2 wt%; (b) UV-Vis reflectance spectra of perovskite films with varying f-SiO2 wt% loadings.140 |
C60SAM functionalized mesoporous TiO2 as ETL has shown recorded efficiency of ∼11.7% in a perovskite-sensitized solar cell using Spiro-OMeTAD as a transparent hole transporter.141 In this approach, C60SAM functionalized mesoporous TiO2 enhances both the absorption and the electron injection. It was observed that the contact between the perovskite and the planar n-type TiO2 charge collection layer was not effective for electron transfer. Modifying this contact with an organic electron-accepting fullerene self-assembled monolayer (C60-SAM) modified TiO2, significantly enhanced the electron transfer.141 Most of the interlayer organic/polymer molecules passivate or inhibit the formation of trap states at the interface, on the TiO2 surface through the anchoring group. This results in reduction of the non-radiative recombination channels at this interface.142 Through electroluminescence (EL) studies, Wojciechowski et al. showed a 5/10-fold increase in the electroluminescence signal for a fullerene-modified sample compared with a standard cell, indicating that the non-radiative recombination at the heterojunction was reduced with the C60-SAM treatment. The observed increase in EL intensity also suggests that the functionalization led to a reduction of interface defect sites, either on the TiO2 or in the perovskite, which would otherwise cause non-radiative recombination of injected carriers. This is consistent with recent results by Q. Wang et al., who suggested that fullerene can passivate trap states in the perovskite itself. Using triple cathode buffer layers consisting of PCBM/C60/LiF, a PCE of up to 14.24% was achieved with excellent stability (Fig. 30).143
Fig. 30 (a) Absorption spectra of P3HT and perovskite films with and without C60SAM fullerene functionalization. The materials were cast on glass substrates which had been coated with a 600 nm thick m-TiO2 layer. Inset: energy levels of the system components relative to vacuum. (b) Schematic of the device 4d. (c) Microstructure and morphology effect (c) J–V characteristics taken under AM1.5 simulated sunlight (100 mW cm−2 irradiance) and in the dark for devices with and without C60SAM fullerene functionalization. Inset: external quantum efficiencies (EQE) for the same devices.141 |
Growing interest in PSCs and increased knowledge on how to improve their efficiency by manipulating the microstructure, morphologies, doping, and device architecture has resulted in development of several techniques to fulfil these requirements. For example, several methods have been suggested to deposit ETM/ETL. The most common technique is spin coating of m-TiO2. The thickness of m-TiO2 can be controlled by the concentration of m-TiO2 in ethanol solution in iso-propyl alcohols. This method has been found to be more reliable and easier than other techniques. Nanostructured ETM/ETL can be prepared by a hydrothermal method. Kim et al. observed PCE of more than 13% using 20 nm of ALD-deposited TiO2 on an ITO/PEN substrate. They reported that 20 nm ALD-deposited a-TiOx effectively collected the electron from the perovskite layer without mesoporous TiO2.18 Wang et al. used nanocomposites of graphene and TiO2 nanoparticles to produce highly efficient ETL in PSC. They postulated that most of the electron density resides on the graphene and ensuing “depleted” TiO2 is contacting the perovskite will have a lower recombination rate with the holes in the perovskite because of reduced electron density in the TiO2, thus contributing to the increase in the photocurrent.144
This journal is © The Royal Society of Chemistry 2017 |