Electrosprayed TiO₂ nanoporous hemispheres for enhanced electron transport and device performance of formamidinium based perovskite solar cells

Abstract

Titanium dioxide (TiO₂) nanoporous hemispheres (NHSs) with a radius of ∼200 nm are fabricated by electrospraying a hydrothermally synthesized TiO₂ nanoparticle (NP) suspension solution. The resulting TiO₂ NHSs are highly porous, which are beneficial to the infiltration of perovskites and provide a larger contact area, as building blocks to construct a mesoporous TiO₂ layer for FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ based perovskite solar cells (PSCs). By varying the TiO₂ NHS collecting period (15 s, 30 s, 60 s and 90 s) during the electrospraying process, the performance of PSCs changes with different TiO₂ NHS distribution densities. The optimized PSC employing TiO₂ NHSs (60 s) exhibits a photovoltaic conversion efficiency (PCE) as high as 19.3% with a J_sc of 23.8 mA cm⁻², a V_oc of 1.14 V and a FF of 0.71. Furthermore, the PSC possesses a reproducible PCE value with little hysteresis in its current density–voltage (J–V) curves. The small perturbation transient photovoltage (TPV) measurement reveals a longer free carrier lifetime within the TiO₂ NHS based PSC than that in the TiO₂ NP based PSC, and the time of flight (TOF) photoconductivity measurement shows that charge mobilities in this system are also enhanced. These characteristics make TiO₂ NHSs a promising electron transport material for efficient photovoltaic devices.

---

Introduction

Metal halide perovskites with the formula APbB₃ [A = Cs⁺, MA (CH₃NH₃⁺, methylammonium cation) or FA (NH = CHNH₃⁺, formamidinium cation); B = Cl⁻, Br⁻ and I⁻], as emerging light absorbing materials, offer a promising path to construct highly efficient photovoltaic devices, owing to their unique optoelectronic properties, including large optical tunability, high absorption coefficients and millimeter-scale carrier diffusion length. The efficiency of perovskite solar cells (PSCs) has kept growing in recent years and the highest efficiency reported is 22.1% up to now.^1–11 Among the recently proposed PSC architectures, the mesoporous TiO₂ layer functions as an “active” scaffold to enable electron transportation from the perovskite light-absorber to the fluorine doped tin oxide (FTO) layer through its conduction band (CB).^2–13 Therefore, the mesoporous TiO₂ layer is required to be: (1) unconsolidated to accept the perovskite, (2) nanoporous to enlarge the contact area with the perovskite to obtain highly efficient charge injection, and (3) defect-free to minimize grain boundaries for the purpose of suppressing charge recombination and trapping.¹³ As the thickness of the TiO₂ layer strongly affects the efficiency of mesoscopic PSCs, a submicron thick (200–300 nm) mesoporous TiO₂ layer made from TiO₂ nanoparticles (NPs) with a size of 20–50 nm is commonly adopted to enhance the electron transportation and minimize the loss of electrons caused by the charge recombination when fabricating PSCs.^14–19 While tiny TiO₂ NPs with a large surface area possess small pores that will block the effective infiltration of the perovskite, large TiO₂ NPs with a better infiltration ability of the perovskite normally generate a small surface area.¹³ For this reason, nanoporous TiO₂ spherical structures (SPs) synthesized by different methods,^20–37 which have been applied in dye sensitized solar cells^{20,21,26–31} and quantum dot solar cells,^32–36 are introduced to replace TiO₂ NPs in PSCs^13,37 and reveal superiorities over TiO₂ NPs in both providing a large surface area for the deposition of perovskites and generating efficient porous structures for electrolyte diffusion. For example, a notable photovoltaic conversion efficiency (PCE) improvement has been achieved from 14.73% of a 30 nm TiO₂ NP based PSC to 18.41% of a 100 nm TiO₂ SP based device.¹³ However, the large TiO₂ SP diameter significantly reduces the contact area between nanoporous TiO₂ SPs and the substrate, which further holds back the electron injection from the TiO₂ SP CB into the substrate. To solve this, mechanical compression treatment, which can increase the inter-particles connection as well as the contact area with the substrate, has been used in dye sensitized solar cells,^20,21 but this will cause counteracting of the enlarged pores brought about by SPs and potential risk to mechanically damage the compact TiO₂ layer or the FTO substrate.

Here, we establish a fabrication method which yields TiO₂ nanoporous hemispheres (NHSs) by electrospraying hydrothermally synthesized TiO₂ NP suspension solution in a controllable and scalable manner. The “soft” touch of the electrosprayed TiO₂ on the “hard” FTO/compact TiO₂ substrate automatically forms hemispherical structures. The resulting TiO₂ NHSs are monodispersed with ultralarge surface areas and a considerable number of pores to improve the infiltration of the perovskite, and the large surface area increases the contact area between the perovskite and TiO₂ NHSs, leading to efficient electron injection. All these properties make them preferable to replace the conventional mesoporous TiO₂ NP layer. Moreover, TiCl₄ treatment is applied to provide strong interconnectivity among the TiO₂ NPs within the TiO₂ NHS, making the charge transportation through the TiO₂ layer more efficient. We demonstrate that the fabricated TiO₂ layer formed by electrosprayed TiO₂ NHSs greatly enhances the photovoltaic properties of PSCs. Furthermore, we systematically study the origin of the enhanced photovoltaic properties of PSCs based on TiO₂ NHSs by performing small perturbation transient photovoltage (TPV) and time of flight (TOF) photoconductivity measurements to analyze the free carriers lifetime and mobilities.

Experimental section

Synthesis of MABr

MABr was synthesized by slowly dropping hydrobromic acid (48 wt% in water, Sigma-Aldrich) into a solution of CH₃NH₂ (33 wt% solution in absolute ethanol, Sigma-Aldrich) at 0 °C.¹¹ The solution was magnetically stirred for 4 hours at room temperature, followed by rotary evaporation at 50 °C until no obvious liquid remained. The crude solid was then dissolved in a small amount of ethanol (Sigma-Aldrich), and reprecipitated with diethyl ether (Sigma-Aldrich) three times. The resulting white solid was collected and dried at room temperature under vacuum for 2 days before usage.

Synthesis of TiO₂ nanoparticles

Titanium isopropoxide (Sigma-Aldrich) was used as a TiO₂ precursor. Acetic acid (Sigma-Aldrich) and nitric acid (Sigma-Aldrich) were used to control the pH (around 1 with roughly 0.1 M H⁺) of the solution, and the solution was transferred to Teflon cups in an autoclave for the following hydrothermal process (240 °C for 15 hours).³⁸ TiO₂ paste was prepared by adding 0.65 g TiO₂ nanoparticles and 0.5 g ethyl cellulose (Sigma-Aldrich) into 3.22 mL α-terpineol (Sigma-Aldrich) and 3.22 mL methoxyethanol (Sigma-Aldrich). The paste was then sonicated for 2 hours until a homogeneous suspension solution without aggregation was obtained. Next, the TiO₂ paste was diluted with ethanol in a weight ratio of 1 : 10. The diluted solution was loaded into a plastic syringe which was connected to a high voltage power supply, and then electrosprayed directly onto the compact TiO₂ layer.

Fabrication of PSCs

The patterned FTO-coated glass (NSG) was sequentially cleaned using detergent, DI water, acetone (Sigma-Aldrich), and isopropanol (Sigma-Aldrich) and finally treated under UV ozone for 10 minutes at 100 °C. A thin compact TiO₂ (C-TiO₂) layer was deposited on top of the cleaned FTO substrate by spin-coating the TiO₂ flat film precursor solution at 6000 rpm for 30 s and dried at 120 °C for 5 minutes in a nitrogen (N₂) glovebox. Then, it was gradually heated up to 450 °C for 30 minutes in air. To prepare the TiO₂ flat film precursor solution, titanium isopropoxide (1 mL, Sigma-Aldrich) and 12 M HCl solution (10 μL, Sigma-Aldrich) were diluted in ethanol (10 mL). Then the substrate was treated with a 50 mM TiCl₄ (Sigma-Aldrich) aqueous solution at 80 °C for 20 minutes followed by annealing at 450 °C for 20 minutes. To prepare the TiO₂ NHSs with a radius of about 200 nm, the electrospraying process was carried out with a commercial electrospinning setup (ET-2535, Ucalery). The electric field of 15 kV was applied between the metal orifice and the conducting collector, and the solution feeding rate was controlled by using a syringe pump at 0.3 mL h⁻¹. The electrospraying process generates droplets and simultaneously applies a charge by means of an electric field. During the electrospraying process, particle stacking proceeds under the forces of the electric field and liquid surface tension. The final shape of the TiO₂ NHSs was controlled by adjusting the parameters of the electrospraying process, including voltage (15 kV), collecting distance (10 cm), humidity (50%–70%) and feeding rate (0.3 mL h⁻¹), to ensure when the droplets reached the collector, the solvent was not evaporated completely. The shape of TiO₂ NHSs is therefore deformable, which further leads to a flat contact area with the substrate. The collecting periods are 15 s, 30 s, 60 s and 90 s, respectively. After the electrospraying process, the as-prepared TiO₂ meso-layer was placed on a hot plate and sintered at 500 °C for 20 minutes in air. Then, the substrate was treated with a 50 mM TiCl₄ (Sigma-Aldrich) aqueous solution at 80 °C for 20 minutes followed by annealing at 500 °C for 15 minutes. For the fabrication of the TiO₂ NP PSC, an ∼180 nm thick TiO₂ NP layer was deposited on top of the substrate by spin-coating the prepared TiO₂ paste (the other procedures are the same as the TiO₂ NHS PSC). Next, the perovskite film was deposited on the substrate by a spin-coating process in a glovebox: first, 2000 rpm for 10 s; second, 6000 rpm for 30 s, and then chlorobenzene (110 mL, Sigma-Aldrich) was dropped in 8–10 s during the second spin-coating process. The substrate was then heated at 100 °C for 60 minutes. The recipe of the precursor solution was the same as a recent study:¹¹ the precursor solution was prepared by dissolving 0.265 g PbI₂ (TCI), 0.037 g PbBr₂ (TCI), 0.011 g MABr and 0.094 g FAI (Dyesol) in 0.4 mL anhydrous N,N-dimethylformamide (Sigma-Aldrich) and 0.1 mL anhydrous dimethylsulfoxide (Sigma-Aldrich). Spiro-OMeTAD (Merck) was further deposited by spin-coating at 3000 rpm for 30 s. The spiro-OMeTAD solution was prepared by dissolving 74 mg spiro-OMeTAD, 28.5 μL 4-tert-butylpyridine (Sigma-Aldrich), 17.5 μL of a stock solution of 520 mg mL⁻¹ lithium bis (trifluoromethylsulphonyl) imide (Sigma-Aldrich) in acetonitrile (Sigma-Aldrich) and 29 μL of a stock solution of 100 mg mL⁻¹ tris(2-(1H-pyrazol-1-yl)-4-tert-butylpyridine)-cobalt(III) tris(bis(trifluoromethylsulfonyl) imide) (Sigma-Aldrich) in acetonitrile in 1 mL anhydrous chlorobenzene. Finally, 80 nm of gold was deposited as an electrode by thermal evaporation at an average deposition rate of 0.3 Å s⁻¹. Epoxy was used to protect the fabricated devices from coming into contact with moisture and the encapsulated devices were collected in a desiccator for the stability study.¹¹

Characterization

J–V characterization was performed in an inert environment (MBraun glovebox, N₂ atmosphere) with a solar simulator (SAN-EI Electric XES-301S 300 W Xe lamp JIS Class AAA) and a Keithley 2400 sourcemeter. The solar cells were masked with metal apertures to define the active areas which were typically 0.09 or 0.08 cm². The incident photon-to-current efficiency (IPCE) as a function of light wavelength was measured in the same glovebox with a customer-made setup consisting of an Oriel 300 W Xe lamp in combination with an Oriel Cornerstone 130 monochromator and a SRS 810 lock-in amplifier (Stanford Research Systems). A calibrated Si diode was used as the reference to calculate the number of photons incident on the device for each wavelength. The IPCE photocurrent was recorded by using a Keithley 2400 sourcemeter. The surface and cross section morphologies of perovskite films and the entire PSCs were investigated by using a scanning electron microscope (SEM, JEOL 6340F) at 5 kV. The morphologies of the prepared TiO₂ NPs and NHSs were studied by using a transmission electron microscope (TEM, JEOL 2100F) at 200 kV. X-ray diffraction (XRD) experiments were conducted by using a Shimadzu XRD-6000 X-ray diffractometer with Cu Kα radiation (λ = 1.54 Å). The simulations of optical power distribution within the functional layers were conducted by means of finite-difference time-domain methods (Lumerical FDTD Solutions). The absorption spectra were recorded on a PerkinElmer Lambda 950UV/VIS/NIR spectrometer. TOF photoconductivity measurements were carried out by generating the photocarriers at the FTO side of the device, and a pulsed Nd:YAG laser (wavelength of 532 nm, pulse width <4 ns, and pulse repetition rate 1 Hz) was used as the light source for the TOF experiment. Small perturbation TPV experiments were performed using the same laser, and the power intensity of the laser was adjusted by using neutral density filters to afford a small perturbation of the cell photovoltage (∼20 mV). A white light halogen lamp and neutral density filters were employed to vary the back illumination level. The photocurrents under different applied electric fields were documented by using an oscilloscope (Agilent 54845A).

Results and discussion

In order to study the photovoltaic properties of electrosprayed TiO₂ NHS based PSCs, TiO₂ NPs are first synthesized using a hydrothermal method, and then TiO₂ NHSs are fabricated by electrospraying this TiO₂ NP suspension solution. Fig. 1 schematically illustrates the fabrication process of both TiO₂ NHS based PSC and its control group based on conventional TiO₂ NPs.

Fig. 1 Fabrication process of PSCs.

The TiO₂ NPs are fabricated by a hydrothermal method with titanium isopropoxide as the precursor under acidic conditions (pH ∼ 1). The concentration of H⁺ in the solution is a critical factor to determine the synthesis product when preparing TiO₂ by hydrothermal reaction. To be specific, when under extreme acidic conditions (pH < 0), rutile TiO₂ is usually formed, while anatase TiO₂ is obtainable under significantly less acidic conditions (pH > 1).^12,38–40 So both acetic acid and nitric acid are used in our process to adjust the pH value of the aqueous solution. The SEM (Fig. 2a) and TEM images of different magnifications (Fig. 2b and c) present the mesoscopic and morphological structure of the synthesized TiO₂ NPs. Fig. 2b shows a group of TiO₂ NPs with an average size of 30 nm which is well controlled and uniformly distributed. High resolution TEM (Fig. 2c) illustrates the crystal structure of the synthesized TiO₂ NPs, which is oriented along the [101] direction, similar to some previously reported results.^39,40 To reveal the structure of the TiO₂ NPs in this study, the XRD pattern (Fig. 2d) is recorded, and the well-resolved diffraction peaks are indexed to (101), (004), (200), (105), (211), (204) and (215), indicating the anatase nature of the synthesized TiO₂ NPs.^{12,13,38–40}

Fig. 2 (a) SEM; (b) low-magnification TEM; (c) high-resolution TEM images, and (d) XRD pattern of the hydrothermal synthesized TiO₂ NPs. The SEM image (a) and XRD spectrum (d) were documented from a sample fabricated by spin-coating TiO₂ NPs on top of the FTO/C-TiO₂ substrate; to obtain the TEM images, the TiO₂ NPs were deposited on the copper mesh TEM grids for the measurement.

TiO₂ NHSs are fabricated by electrospraying the suspension solution of TiO₂ NPs and then form a dense structure due to the evaporation of the solvent. The SEM and TEM images of the electrosprayed TiO₂ NHSs are shown in Fig. 3. Compared to TiO₂ spherical structures formed with fully and ultrafast evaporation of the solvent during electrospraying,^20,21 we control both the size and shape of TiO₂ NHSs by adjusting various parameters during the electrospraying process. More precisely, by employing a relatively shorter collecting distance, a larger feeding rate and higher humidity, only a small amount of solvent remains within the suspension droplets when they reach the substrate. With the high momentum provided by the electric field, the droplets deform into hemispherical structures with flat surfaces contacting with the substrate. Fig. 3a and b clearly show the hemispherical structure of a single TiO₂ NHS and a group of TiO₂ NHSs on top of a FTO/compact TiO₂ substrate. TiO₂ NHSs are highly porous structures, as shown in Fig. 3c, which will improve the infiltration of the perovskite light-absorber into the mesoporous TiO₂ layer. However, the interconnectivity among the TiO₂ NPs within the electrosprayed TiO₂ NHS is weak, which may suppress the electron transport within the TiO₂ NHSs. So, we utilize diluted TiCl₄ aqueous solution to treat the as-electrosprayed TiO₂ NHSs for the purpose of improving the interconnectivity of the TiO₂ NPs.^20,21Fig. 3c and e show the TiO₂ NHSs before the TiCl₄ treatment, and Fig. 3d and f are the images of TiO₂ NHSs after the treatment. TiO₂ NPs within TiO₂ NHSs after being treated with TiCl₄ are connected more strongly than the case before the treatment, which provides a fast transporting channel for photogenerated electrons to transport inside the TiO₂ NHSs.

Fig. 3 Side view SEM images of (a) a single TiO₂ NHS and (b) a 60 s collecting time sample on the FTO/compact TiO₂ substrate. Plan view SEM images of TiO₂ NHSs (c) before and (d) after TiCl₄ treatment; TEM images of TiO₂ NHSs (e) before and (f) after TiCl₄ treatment.

The prepared TiO₂ NHSs are then employed to fabricate the mesoporous TiO₂ layer. For comparison, the same TiO₂ NPs synthesized by hydrothermal reaction are also applied using a spin-coating method as a control group. Then the FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ layer is deposited on the mesoporous TiO₂ layers by a two-step spin-coating process.¹¹Fig. 4a shows the surface morphology of the prepared FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ layer. Compared to the control group (Fig. S1†), the crystal size of the FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ film is more uniform on the electrosprayed TiO₂ NHS layer than that on the TiO₂ NP layer. In addition, the average ∼400 nm FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ crystal size is observed (with the range of 300–600 nm) in the TiO₂ NHS sample, while the average crystal size in the TiO₂ NP sample is ∼250 nm with the range of 100–450 nm. The large average crystal size and even size distribution of the deposited FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ on top of the TiO₂ NHS layer lead to less grain boundaries within the perovskite film, which reduces the recombination and trapping probability in the injection and transportation processes of photogenerated free electrons.¹³ XRD spectra (Fig. S2†) show that the intensities and band widths of FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ peaks are not varied due to the change of the mesoporous TiO₂ layer, and the characteristic XRD peak positions of the prepared FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ samples are identical to the previous results.¹¹

Fig. 4 (a) Surface SEM image of the FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ film based on TiO₂ NHSs; (b) cross-section SEM image of TiO₂ NHS based PSC. Photovoltaic performance: (c) J–V curves; and (d) IPCE of PSCs with different collecting time of TiO₂ NHSs.

Next, FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ based PSCs are fabricated as described in Fig. 1. Fig. 4b shows the cross-sectional SEM image of the fabricated PSC employing electrosprayed TiO₂ NHSs with the collecting period of 30 s, which gives the thickness and morphology for each functional layer of the device. As demonstrated, the functional layers within the PSC from bottom to top are FTO, a compact TiO₂ layer (∼40 nm), a mesoporous TiO₂ layer with FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ (∼600 nm), HTM (∼150 nm) and a Au electrode (∼80 nm), respectively. Each layer connects tightly to its adjacent layers, which enables the unimpeded interlayer transport of photogenerated free carriers. As can be seen from the image, a few TiO₂ NHSs located above the TiO₂ compact layer and the FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ film show high crystallinity.

PSCs with a different collecting period of TiO₂ NHSs are fabricated. Fig. 4c presents the typical current density–voltage (J–V) curves for these resulting devices. Generally, both the short circuit current density (J_sc) and the open circuit voltage (V_oc) remain at a similar level for PSCs with the TiO₂ NHSs collecting period of 15 s, 30 s and 60 s, but drop remarkably when the collecting period reaches 90 s. In addition, the PSC based on 60 s collecting of TiO₂ NHSs (PSC-60 s) offers the best fill factor (FF) of 0.63, while the FF of PSC-90 s is as low as 0.34. Overall, the 60 s TiO₂ NHSs collecting time yields the best PSC with a typical PCE of 16.01%. This is in good agreement with the IPCE measurement (Fig. 4d), which also indicates that PSC-60 s can convert photons into electricity more effectively. To clearly compare the photovoltaic performance of PSC with different collecting periods of TiO₂ NHSs, the main photovoltaic parameters (including J_sc, V_oc, FF, and PCE) for these devices are plotted in Fig. 5. It is obvious that the FF of PSC-60 s is superior to other resulting devices in this study, and the PCE indicates that 60 s is the optimum collecting time of efficient PSC based on the proposed electrosprayed TiO₂ NHSs.

Fig. 5 Main photovoltaic parameters. (a) J_sc, (b) V_oc, (c) FF and (d) PCE for PSCs based on different collecting periods of TiO₂ NHSs (15 s, 30 s, 60 s and 90 s). 20 devices are made and tested for each collecting period.

This is attributed to the different electron transporting mechanisms within TiO₂/perovskite heterojunctions, under different TiO₂ NHS distribution density conditions, as demonstrated in Fig. 6 (the NHS structure can be clearly seen in Fig. S3a†). To be more specific, for the collecting period of 15 s, 30 s and 60 s, the population of TiO₂ NHSs is small in relative terms of the substrate area, so most of the TiO₂ NHSs are evenly distributed and are in a monodisperse state as plotted in Fig. 6a–c (the SEM image of 30 s and 60 s collected TiO₂ NHSs is shown in Fig. S3b and S3c†). In this case, photoinduced electrons generated in the perovskite tend to inject into the CB of the nearest TiO₂ NHS, as illustrated in Fig. 6e. Electrons transport/inject more efficiently into the TiO₂ CB with more TiO₂ NHSs. Also, the hysteresis effect is significantly minimized within more mesoporous TiO₂ nanostructures.^11–13 Thus an enhanced J_sc and FF can be obtained in the PSC-60 s. But when the TiO₂ NHSs accumulate after a longer time (90 s in this case) when the density becomes relatively large, as illustrated in Fig. 6d and f (the SEM image of 90 s collected TiO₂ NHSs is shown in Fig. S3d†), TiO₂ NHSs pile up which lead to some non-direct contacts between TiO₂ NHSs and the FTO/TiO₂ compact layer substrate. In this case, some photogenerated electrons have to pass through the interface between upper and lower TiO₂ NHSs before they can finally reach the FTO. As a consequence, the distance, in which electrons need to transport to the substrate, is extended and more free electrons may be annihilated during this period,^9,12,13 which lead to a reduction in the number of electrons that finally reach FTO. Therefore, the performance of the PSC-90 s is worse than that of the PSC-60 s. And for the same reason, the IPCE curves of the PSCs shown in Fig. 4d remain at a similar level when the collecting time is below 60 s, and drop evidently when the collecting time increases to 90 s.

Fig. 6 Schematic drawing of the TiO₂ NHS dispersion with different collecting times. NHSs distribution status with a collecting time of (a) 15 s, (b) 30 s, (c) 60 s, and (d) 90 s, respectively, and cross-section schematics for the collecting time of (e) 60 s and (f) 90 s.

As a comparison, anatase TiO₂ NPs prepared by the same hydrothermal method are used in the PSCs with an identical device configuration and fabrication process to the TiO₂ NHSs, as shown in Fig. 1. Fig. 7a plots the J–V curves for the corresponding champion devices based on both TiO₂ NHSs and TiO₂ NPs. It is apparent that most of the photovoltaic parameters of the TiO₂ NHS based device, including V_oc (1.14 V), J_sc (23.8 mA cm⁻²) and FF (0.71), are superior to the corresponding data (V_oc of 1.12 V, J_sc of 22.2 mA cm⁻² and FF of 0.68) of the TiO₂ NP counterpart. Thus, the TiO₂ NHS based device generates a PCE of 19.3%, which is better than the 17.0% efficiency provided by the TiO₂ NP based device. For the IPCE curves as shown in Fig. 7b, there is an obvious enhancement response in the wavelength range from 500 nm to 760 nm. This enhancement may be related to an optical interference effect caused by TiO₂ NHS arrays. For λ > 500 nm the cavity of PSC operates in the low-finesse thin-film interference regime,⁴¹ so the TiO₂ NHS arrays with the dimension of sub-wavelength may cause the interference of input light and generate optical field enhancement within the perovskite layer, which will lead to more efficient light absorption. The FDTD⁴² simulations show the difference of optical power distribution at the TiO₂/perovskite interface between the NP based model and NHS (60 s) based model (Fig. S4a and S4b†), which indicates the effect of optical interference. Optical absorption spectra also show the enhanced light absorption in the FTO/C-TiO₂/TiO₂ NHSs 60 s/perovskite sample (Fig. S4c†). Furthermore, little hysteresis^11–13 is found in the J–V curve of the champion device of PSC-60 s (Fig. S5†). A preliminary stability investigation of these PSCs has been performed for 60 days, which shows that the devices stored in the dark at room temperature are relatively stable, with just a 0.7% PCE drop in the entire testing period (Table S1†).

Fig. 7 (a) J–V curves and (b) IPCE spectra of champion devices fabricated based on TiO₂ NHSs and the control group based on TiO₂ NPs.

To study the photogenerated free carrier behaviour in the PSC based on 60 s collected TiO₂ NHSs, small perturbation TPV is utilized to observe the fast decay component occurrence within the system with the time relaxation constant in the nanosecond (ns) scale.^43–45 Thus, we operate complementary small perturbation TPVs by using a white light source to vary the back illumination level and a ns-pulsed 532 nm Nd/YAG laser to induce the small voltage perturbations and the results are shown in Fig. 8a. The TPVs show one component of the photovoltage decays, a lifetime that spans from 102 to 174 μs in a wide back illumination level from 0.8 V to 0.2 V (as summarized in Table S2†), which is much longer than its mesoporous TiO₂ NP counterpart (only 4 μs in the back level of 1 V, shown in Fig. S6†). This is ascribed to the increase of the contact area between FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ and TiO₂. Also, the use of TiO₂ NHSs can further optimize the device structure to obtain an optimal combination of the series resistance and shunt resistance. Thus, the free carriers can transport a longer distance in this TiO₂ NHSs/FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ heterojunction and this TiO₂ NHS layer is indeed a good replacement to its predecessor-mesoporous TiO₂ NP layer in the application of efficient PSCs. Moreover, the obtained lifetime is much longer than that in the typical TiO₂ NP based CH₃NH₃PbI₃ solar cell (3 to 92 μs with the back illumination level 0.7 V to 0.1 V).⁴⁵ As a result, the free carriers are able to transport a longer distance in this FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ light absorbing layer, compared to the CH₃NH₃PbI₃ layer.

Fig. 8 (a) Normalized small perturbation TPVs of the PSC. (b) TOF hole transients and mobilities measured at different back voltages for the device. (c) TOF electron transients and mobilities measured at different back voltages for the device.

Charge mobility in this device is also studied with a TOF photoconductivity technique at room temperature.^46,47 By simply changing the applied background voltages of the device, the electron and hole transit time can be obtained separately. Furthermore, the electron and hole mobilities of the system can be easily calculated, given the heterojunction thickness, applied back voltage and free charge transit time. Fig. 8b shows the TOF hole transients and mobilities for laser generated holes obtained at different external applied voltages (the full spectra can be found in Fig. S7†). The transit time can be obtained from the peak of the TOF transients spectra and the charge mobility (μ) in the system can be further calculated with the relation μ = d²/Vt_T,^46,47 where d is the perovskite/HTM thickness (650 nm), V is the applied voltage, and t_T is the transit time. The calculated hole mobility in the FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃/spiro-OMeTAD heterojunction under ambient conditions is 4.0 × 10⁻³ cm² V⁻¹ s⁻¹, which is one order of magnitude higher than that in the CH₃NH₃PbI₃/spiro-OMeTAD heterojunction (∼10⁻⁴ cm² V⁻¹ s⁻¹)⁴⁸ and 4 times higher than that in the typical mesoporous TiO₂ NP based FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ PSC (1.0 × 10⁻³ cm² V⁻¹ s⁻¹, as shown in Fig. S8†), with an external applied electric field of 1.33 × 10⁴ V cm⁻¹. We found that the mobility values of free photogenerated holes are all above 10⁻³ cm² V⁻¹ s⁻¹ within the whole voltage range. The TOF transients and mobilities for photoinduced free electrons with different applied background voltages are shown in Fig. 8c. In this work, the electron mobility in the system is 2.4 × 10⁻³ cm² V⁻¹ s⁻¹ at an applied electric field of 1.54 × 10⁴ V cm⁻¹ for a 650 nm thick TiO₂/FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ heterojunction. The electron mobility values are around 10⁻³ cm² V⁻¹ s⁻¹ within the 1–5 V voltage range, which is nearly one order of magnitude higher than that in the typical mesoporous TiO₂ NP based FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ PSC (∼10⁻⁴ cm² V⁻¹ s⁻¹). The electron mobility in the device is slightly smaller than the hole mobility and this mobility mismatch of electrons and holes may influence the overall performance of the device, thus interface modifying such as metal ion doping and coating as well as the surface plasmonic effect of metal nanoparticles can be used to further increase the efficiency of this kind of PSC.^38,49–53

Conclusions

In summary, we have exploited an electrospray method using hydrothermally synthesized TiO₂ NP suspension solution to fabricate TiO₂ NHSs with an average radius of ∼200 nm. The TiO₂ NHSs have been further employed as a mesoporous layer of FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ based PSCs. Detailed electron transporting mechanisms are discussed based on the difference of TiO₂ NHS distribution densities with various electrospraying collecting periods. The PSC with the collecting period of 60 s demonstrates an optimized 19.3% PCE with a J_sc of 23.8 mA cm⁻², a V_oc of 1.14 mV and a FF of 0.71, which is superior to the TiO₂ NP based control group with the PCE of 17.0%. TPV measurement indicates that PSC-60 s, compared to the PSC adopting 30 nm TiO₂ NPs, exhibits notably efficient carrier injection/transport from FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ to both TiO₂ NHS and HTM layers. This is due to the increase of the contact area between FA_0.81MA_0.15Pb(I_0.836Br_0.15)₃ and TiO₂ caused by this unique TiO₂ NHS network and the optimized overall device structure. Moreover, the TOF results show that the photogenerated free carrier mobilities in the proposed system are higher than the TiO₂ NP based PSC.

Acknowledgements

This work is supported by the Singapore-Berkeley Energy Research Institute (SinBeRISE), Grantor Lloyd's Register Foundation (R-265-000-553-597), the Singapore Ministry of Education Academic Research Fund Tier 2 (MOE2015-T2-1-066, MOE2015-T2-2-010), and Nanyang Technological University (Start-up grant: Lei Wei). The authors thank Dr Xizu Wang, Dr Changyun Jiang, Dr Lim Siew Lay and Mr Goh Wei Peng from IMRE for fruitful discussion and access to the glovebox.

Notes and references

1. National Renewable Energy Laboratory (NREL), Research Cell Efficiency Records, http://www.nrel.gov/ncpv/images/efficiency_chart.jpg, accessed: November, 2016.
2. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Chem. Soc., 2009, 131, 6050.
3. J.-H. Im, C.-R. Lee, J.-W. Lee, S.-W. Park and N.-G. Park, Nanoscale, 2011, 3, 4088.
4. H. S. Kim, C. R. Lee, J. H. Im, K. B. Lee, T. Moehl, A. Marchioro, S. J. Moon, R. Humphry-Baker, J. H. Yum, J. E. Moser, M. Gratzel and N. G. Park, Sci. Rep., 2012, 2, 591.
5. J. H. Heo, S. H. Im, J. H. Noh, T. N. Mandal, C.-S. Lim, J. A. Chang, Y. H. Lee, H.-J. Kim, A. Sarkar, M. K. Nazeeruddin, M. Grätzel and S. I. Seok, Nat. Photonics, 2013, 7, 486.
6. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Science, 2012, 338, 643.
7. J. Burschka, N. Pellet, S. J. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin and M. Grätzel, Nature, 2013, 499, 316.
8. T. Leijtens, G. E. Eperon, S. Pathak, A. Abate, M. M. Lee and H. J. Snaith, Nat. Commun., 2013, 4, 2885.
9. H. Zhou, Q. Chen, G. Li, S. Luo, T. B. Song, H. S. Duan, Z. Hong, J. You, Y. Liu and Y. Yang, Science, 2014, 345, 542.
10. W. S. Yang, J. H. Noh, N. J. Jeon, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok, Science, 2015, 348, 1234.
11. D. Bi, W. Tress, M. I. Dar, P. Gao, J. Luo, C. Renevier, K. Schenk, A. Abate, F. Giordano, J. P. Baena, J. D. S. M. Zakeeruddin, M. K. Nazeeruddin, M. Grätzel and A. Hagfeldt, Sci. Adv., 2016, 2, e1501170.
12. D. Zhong, B. Cai, X. Wang, Z. Yang, S. Miao, W. H. Zhang and C. Li, Nano Energy, 2015, 11, 409.
13. I. S. Yang, J. S. You, S. D. Sung, C. W. Chung, J. Kim and W. I. Lee, Nano Energy, 2016, 20, 272.
14. N. J. Jeon, J. H. Noh, W. S. Yang, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok, Nature, 2015, 517, 476.
15. Z. Xiao, Y. Yuan, Y. Shao, Q. Wang, Q. Dong, C. Bi, P. Sharma, A. Gruverman and J. Huang, Nat. Mater., 2015, 14, 193.
16. H.-S. Kim, S. H. Im and N.-G. Park, J. Phys. Chem. C, 2014, 118, 5615.
17. G. Murugadoss, G. Mizuta, S. Tanaka, H. Nishino, T. Umeyama, H. Imahori and S. Ito, APL Mater., 2014, 2, 081511.
18. S. D. Sung, D. P. Ojha, J. S. You, J. Lee, J. Kim and W. I. Lee, Nanoscale, 2015, 7, 8898.
19. N. Y. Ahn, D. Y. Son, I. H. Jang, S. M. Kang, M. S. Choi and N. G. Park, J. Am. Chem. Soc., 2015, 137, 8696.
20. D. Hwang, H. Lee, S. Y. Jang, S. M. Jo, D. Kim, Y. Seo and D. Y. Kim, ACS Appl. Mater. Interfaces, 2011, 3, 2719.
21. H. Lee, D. Hwang, S. M. Jo, D. Kim, Y. Seo and D. Y. Kim, ACS Appl. Mater. Interfaces, 2012, 4, 3308.
22. A. Léaustic, F. Bahonneau and J. Livage, Chem. Mater., 1989, 1, 248.
23. X. Jiang, T. Herricks and Y. Xia, Adv. Mater., 2003, 15, 1205.
24. S. Eiden-Assmann, J. Widoniak and G. Maret, Chem. Mater., 2004, 16, 6.
25. J. H. Pan, Q. Wang and D. W. Bahnemann, Catal. Today, 2014, 230, 197.
26. D. Chen, F. Huang, Y.-B. Cheng and R. A. Caruso, Adv. Mater., 2009, 21, 2206.
27. Y. J. Kim, M. H. Lee, H. J. Kim, G. Lim, Y. S. Choi, N.-G. Park, K. Kim and W. I. Lee, Adv. Mater., 2009, 21, 3668.
28. I. G. Yu, Y. J. Kim, H. J. Kim, C. Lee and W. I. Lee, J. Mater. Chem., 2011, 21, 532.
29. S. H. Jang, Y. J. Kim, H. J. Kim and W. I. Lee, Electrochem. Commun., 2010, 12, 1283.
30. W.-Y. Cheng, J. R. Deka, Y.-C. Chiang, A. Rogeau and S.-Y. Lu, Chem. Mater., 2012, 24, 3255.
31. H.-E. Wang, L.-X. Zheng, C.-P. Liu, Y.-K. Liu, C.-Y. Luan, H. Cheng, Y. Y. Li, L. Martinu, J. A. Zapien and I. Bello, J. Phys. Chem. C, 2011, 115, 10419.
32. R. Zhou, Q. Zhang, E. Uchaker, J. Lan, M. Yin and G. Cao, J. Mater. Chem. A, 2014, 2, 2517.
33. X.-Y. Yu, J.-Y. Liao, K.-Q. Qiu, D.-B. Kuang and C.-Y. Su, ACS Nano, 2011, 5, 9494.
34. T. Shu, P. Xiang, Z.-M. Zhou, H. Wang, G.-H. Liu, H.-W. Han and Y. D. Zhao, Electrochim. Acta, 2012, 68, 166.
35. S. D. Sung, I. Lim, P. Kang, C. Lee and W. I. Lee, Chem. Commun., 2013, 49, 6054.
36. Y.-F. Xu, W.-Q. Wu, H.-S. Rao, H.-R. Chen, D.-B. Kuang and C.-Y. Su, Nano Energy, 2015, 11, 621.
37. T. Leijtens, B. Lauber, G. E. Eperon, S. D. Stranks and H. J. Snaith, J. Phys. Chem. Lett., 2014, 5, 1096.
38. J. T. Wang, J. M. Ball, E. M. Barea, A. Abate, J. A. Alexander-Webber, J. Huang, M. Saliba, I. Mora-Sero, J. Bisquert, H. J. Snaith and R. J. Nicholas, Nano Lett., 2013, 14, 724.
39. H. Cheng, J. Ma, Z. Zhao and L. Qi, Chem. Mater., 1995, 7, 663.
40. A. Testino, I. R. Bellobono, V. Buscaglia, C. Canevali, M. D'Arienzo, S. Polizzi, R. Scotti and F. Morazzoni, J. Am. Chem. Soc., 2007, 129, 3564.
41. Q. Lin, A. Armin, R. C. Nagiri, P. L. Burn and P. Meredith, Nat. Photonics, 2015, 9, 106.
42. K. S. Yee, IEEE Trans. Antennas Propag., 1966, 14, 302.
43. S. D. Stranks, V. M. Burlakov, T. Leijtens, J. M. Ball, A. Goriely and H. J. Snaith, Phys. Rev. Appl., 2014, 2, 034007.
44. V. Roiati, S. Colella, G. Lerario, L. De Marco, A. Rizzo, A. Listorti and G. Gigli, Energy Environ. Sci., 2014, 7, 1889.
45. L. Bertoluzzi, R. S. Sanchez, L. Liu, J. W. Lee, E. Mas-Marza, H. Han, N. G. Park, I. Mora-Sero and J. Bisquert, Energy Environ. Sci., 2015, 8, 910.
46. Q. Dong, Y. Fang, Y. Shao, P. Mulligan, J. Qiu, L. Cao and J. Huang, Science, 2015, 347, 967.
47. M. J. Tan, W. P. Goh, J. Li, G. Pundir, V. Chellappan and Z. K. Chen, ACS Appl. Mater. Interfaces, 2010, 2, 1414.
48. D. Bi, L. Yang, G. Boschloo, A. Hagfeldt and E. M. Johansson, J. Phys. Chem. Lett., 2013, 4, 1532.
49. W. Guo, C. Xu, X. Wang, S. Wang, C. Pan, C. Lin and Z. L. Wang, J. Am. Chem. Soc., 2012, 134, 4437.
50. H. Park, Y. Park, W. Kim and W. Choi, J. Photochem. Photobiol., C, 2013, 15, 1.
51. W. Zhang, M. Saliba, S. D. Stranks, Y. Sun, X. Shi, U. Wiesner and H. J. Snaith, Nano Lett., 2013, 13, 4505.
52. F. Giordano, A. Abate, J. P. Baena, M. Saliba, T. Matsui, S. H. Im, S. M. Zakeeruddin, M. K. Nazeeruddin, A. Hagfeldt and M. Graetzel, Nat. Commun., 2016, 9, 1989.
53. W. E. Zhang, M. Saliba, S. D. Stranks, Y. Sun, X. Shi, U. Wiesner and H. J. Snaith, Nano Lett., 2013, 13, 4505.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6nr07369g

‡ These authors contributed equally to this work.

---