2H-NbS2 film as a novel counter electrode for meso-structured perovskite solar cells

We report the use of 2H-NbS2 film as a novel counter electrode in perovskite solar cells fabricated with a cold isostatic pressing method. The 2H-NbS2 film, which was prepared through an exfoliation method followed by restacking from LixNbS2 powder, shows high electrical conductivity of 8.7 × 103 S cm−1 and work function of 5.20 eV. The two-dimensional transition metal dichalcogenide was used for the first time as a counter electrode in meso-structured perovskite solar cells. Through this process, we demonstrated a new alternative to noble metals. The perovskite solar cell base on the 2H-NbS2 counter electrode showed an open-circuit voltage of 1.046 V, comparable to that of gold, and a power conversion efficiency of 8.3%.

(e.g., Bi 2 Te 3 ), and thermoelectric (e.g., Bi 2 Te 3 ) phenomena 28 . The weakly bonded atomic layers in these materials facilitate the exfoliation from bulk to thin layers, thus providing a potential application in many fields such as lithium ion battery 29 . In this work, we used a cost-effective and flexibly metallic 2H-NbS 2 film as a counter electrode in perovskite solar cells. The cold isostatic pressing (CIP) method was successfully developed instead of the conventional thermal evaporation or solution based printing techniques 30 . The work function of the prepared NbS 2 film was measured to be 5.20 eV, matching the highest occupied molecular orbital energy level of the used hole transporting material 2,2′,7,7′-tetrakis(N,N-di-p-methoxyphenylamine)-9,9′-spirobifluorene (spiro-OMeTAD, 5.22 eV) 31,32 . This, together with a high electrical conductivity (8.7 × 10 3 S cm −1 ), indicated efficient charge extraction and collection at the counter electrode. The PSC with the configuration of FTO/TiO 2 / MAPbI 3 /spiro-OMeTAD/NbS 2 has a maximum power conversion efficiency of 8.3% from our measurements. We believe that with further optimization of the material and the fabrication technique, the performance of our devices can be further improved. Figure 1a shows the multi-step process for the preparation of NbS 2 thin film 33 . NbS 2 powder was firstly synthesized through the conventional solid-state reaction (SSR) from Nb and S 34 . The powder was then reheated with Li 2 S and Nb for the intercalation of Li cation within the NbS 2 layers. Exfoliation of the Li x NbS 2 in hydrochloric acid aqueous solution led to the exchange of Li + with H + , forming a homogeneous NbS 2 nanosheet colloid. From the atomic force microscopic (AFM) and high-resolution TEM (HRTEM) images, the prepared NbS 2 nanosheets were found to be just one to two layers in most cases with the thickness of around 1.1 nm in the hexagonal lattice structure ( Fig. S1a and b). A lattice spacing of 2.90 Å is assigned to the 2H-NbS 2 (100) plane. The final NbS 2 thin film was fabricated by restacking the prepared nanosheets through vacuum filtration, and the film thickness can be easily tuned by controlling the amount of the colloidal solution. The prepared NbS 2 film exhibits many wrinkles as shown in the scanning electron microscopy (SEM) image in Fig. 1b.

Results and Discussion
The X-ray diffraction (XRD) pattern of the final NbS 2 film was measured to check the influence of the Li exfoliation. As shown in Fig. 2a, the (001) plane is shifted from 13.7° to 14.9° due to the decreased inter-planar spacing after extracting the Li + . The only strongest peak at 14.9° confirms the 2 H phase and the good lattice orientation after restacking. Figure 2b shows the Hall-effect curve of the prepared NbS 2 film with a thickness of 600 nm. By fitting with a linear function with a slope of 5.5 × 10 −5 Ω T −1 , a carrier concentration of 1.9 × 10 21 cm −3 is extracted. To further investigate the charge transportation properties, a temperature-dependent resistance variation was measured using a four probe method (Fig. 2c). The prepared NbS 2 film exhibits a typical metallic feature with a gradual drop of resistance with the decrease of temperature. A clear superconducting transition is observed when the temperature moves down to 6 K, consistent with previous reports 35,36 . The conductivity of the NbS 2 film at 300 K is calculated to be 8.7 × 10 3 S cm −1 , which is lower than that of gold (4.6 × 10 5 S cm −1 ). However, it is about 15 times higher than that of the assembled graphene (550 S cm −1 ) 37 and 290 times higher than that of the carbon cloth based electrode (30 S cm −1 ) 26 , indicating that our NbS 2 films prepared by the SSR method is a good candidate No band gap in the bulk 2H-NbS 2 was observed, which further confirmed its metallic character (Fig. S2). A work function of 5.20 eV was obtained from the UPS spectrum, slightly higher than that of Au (Figs 2d and S4). Close work functions between NbS 2 and the spiro-OMeTAD hole transporting material (HTM) indicate that the NbS 2 electrode can provide efficient charge separation at the interface.
The exfoliation fabrication process of the device is illustrated in Fig. 3a. The FTO/TiO 2 /MAPbI 3 electrode was fabricated according to earlier publications 26 . The spiro-OMeTAD solution was then spin-coated on the top of the fresh prepared perovskite layer as a HTM. In the counter electrode part, the prepared NbS 2 film was transferred to the HTM surface, followed by sealing the whole device into a polyethylene envelope under vacuum. The envelope was put into the CIP chamber full of hydraulic oil, pressed with a pressure of 280 MPa at room temperature for a couple of minutes. After this high pressure treatment, the NbS 2 film was tightly plastered onto the spiro-OMeTAD surface (Figs 1c and S3a). No breakages or ruptures were observed under the mechanical pressure treatment from the SEM images. The NbS 2 film was even smoother and more compact comparing with the fresh one (Fig. 1c). Therefore, as a less energy consuming and easy preparation technique, the CIP could be a new method to be used on flexible films for device preparation. The whole device architecture and the energy level diagram are shown in Fig. 3b,c 31,32 . The TiO 2 and spiro-OMeTAD layers work as the electron and hole collector, respectively, with appropriate energy levels for efficient charge extraction.
The influence of NbS 2 thickness on the device performance was investigated. The NbS 2 films with the thickness of 300 nm, 600 nm and 1200 nm were prepared. The best photovoltaic performance was obtained while using the 600 nm film ( Fig. S5 and Table S1). It was found that all the performance parameters initially increase with the increase of the NbS 2 film thickness, followed by a drop at film thickness more than 600 nm. Considering the balance between resistance and conductivity, the 600 nm NbS 2 film is chosen in our further studies. For the CIP technique, the pressure is an important factor because the perovskite layer can crack during this process. To check this effect, traditional mesoporous based PSCs with or without spiro-OMeTAD were fabricated and the photovoltaic properties were tested before and after the CIP (280 MPa) treatment. As shown in Fig. S6a and b, no obvious changes were observed after the CIP treatment in both cases. It proves that 280 MPa CIP treatment does not have much effects on the perovskite/HTM layer. A lower pressure of 100 MPa was tested under the same conditions for comparison. As shown from the statistics data, the higher pressure (280 MPa) is slightly better (Fig. S6d). The sequential operation starting from 100 MPa, followed by 280 MPa for one more time did not make any difference (Fig. S6c). Figure 4a demonstrates the current-voltage characteristics of one of the best performing devices based on the NbS 2 and Au counter electrodes at a scan rate of 10 mV s −1 . The Au electrode was used as a reference for comparison. The photovoltaic parameters are summarized in Table 1. The reference device based on the Au counter and FF (0.53), which gave a PCE of 8.31%. The lower J sc and FF are mainly due to the large series resistance related to the NbS 2 counter electrode (302.8 Ω) comparing with that of Au (29.4 Ω) either from the NbS 2 /HTM interface or the NbS 2 itself, which then lead to an ineffective charge collection. A small hysteresis is observed in the J-V curves between the forward and reverse scans for NbS 2 . A slightly low FF leads to a 0.42% drop of PCE in the forward scan. From the incident photon-to-current conversion efficiency (IPCE) spectra, the integrated short-circuit current densities are in agreement with those obtained from the J-V curves (Fig. 4b). The device statistics based on two different counter electrodes over several batches are shown in Figs 4c and S7a. The average PCEs for PSCs based on NbS 2 and Au electrodes are 7.22 ± 1.02%, and 15.80 ± 0.45%, respectively (shown in the parenthesis in Table 1). To evaluate the stability of the devices, two unsealed best-performed solar cells based on NbS 2 and Au counter electrode were kept in ambient atmosphere (humidity > 50%) at room temperature (24 °C). The J-V curves were measured every 24 h (Figs 4d and S7b). The performance of both cells drops gradually with time, showing 50-60% of their initial value after a period of 7 days.
In summary, we took advantage of a general solid-state strategy to prepare the single-layered 2H-NbS 2 nanosheets. Through simple filtration of the nanosheet colloid, the flexible NbS 2 film with a high electrical conductivity of 8.7 × 10 3 S cm −1 and a work function of 5.20 eV was obtained. The cold isostatic pressing method was firstly used to combine the prepared NbS 2 film over the organic hole transporting layer for the preparation of perovskite solar cells. Based on this new counter electrode, the device showed a power conversion efficiency of 8.3% in our tests, and a comparable V oc (1046 mV) comparing with that of gold. The two-dimensional transition metal dichalcogenide film is used as a counter electrode in perovskite solar cells, avoiding the conventional use of expensive noble metals and the high-energy consuming thermal evaporation process. Although both the TMD materials and the isostatic pressing technique need to be further optimized to improve the performance, this method provides an alternative way for the selection of new type of electrodes for the photovoltaic devices.

Methods
NbS 2 film preparation. The NbS 2 powder was synthesized by heating the stoichiometric ratio of niobium (99.8%, Alfa Aesar) and sulfur (99.999%, Alfa Aesar) in an evacuated quartz tube at 900 °C for 20 h 33 . The prepared NbS 2 , together with Li 2 S and Nb in a molar ratio of 3:2:1, were reheated through solid state reaction at 800 °C for 10 h, for the preparation of Li x NbS 2. The NbS 2 nanosheets were obtained by chemical exfoliation of Li x NbS 2 in the hydrochloric acid (HCl) aqueous solution. In a typical exfoliation process, 30 mL of 1 M HCl was added to 200 mL of water. Then, 5 mg of Li x NbS 2 was added. After ultrasonication and stirring for 15 min, the homogeneous NbS 2 nanosheet colloid was obtained without any sediments inside, which was dialyzed several times to remove the impurities. The final NbS 2 film was obtained through vacuum filtration of the prepared NbS 2 nanosheets colloid using a PES (Sterlitech, Kent, WA) filter paper. In a typical experiment, 200 mL NbS 2 solution with the concentration same as above led to a membrane thickness of about 300 nm.
Fabrication of the Perovskite Solar Cells. The chemically etched (zinc powder and 4 M HCl) FTO conducting glass was sequentially cleaned by 2% Hellmanex water solution, deionized water, ethanol, and acetone. A 30-50 nm compact TiO 2 layer was then deposited by spray pyrolysis at 450 °C from a precursor solution of titanium diisopropoxidebis(acetylacetonate) in anhydrous ethanol, using oxygen as the carrier gas. The mes-oporousTiO 2 layer was prepared by spin coating the TiO 2 paste (Dyesol 30NRD, dilute 7 times in EtOH) at 4000 rpm for 20 s, followed by sintering at 500 °C for 30 min in air. The mixture of methylammonium iodide  Measurements and Characterization. The X-ray diffraction patterns were measured by the Bruker The work function of Au and NbS 2 film were determined by ultraviolet photoemission spectroscopy in Kratos Axis Ultra DLD multitechnique surface analysis system using He I (21.21 eV) photon lines from a discharge lamp with an error about 0.01 eV. The photocurrent density-voltage characteristics of the solar cells were measured using a 1000 W xenon solar simulator (Newport), and the light intensity was calibrated to 100 mW cm −2 by a NREL-calibrated Si cell (Oriel 91150). All devices were measured under the ambient air condition at room temperature (24 °C). The active area is 0.16 cm 2 defined by a black mask. The IPCE spectrum was measured by a Newport QE system equipped with a 300 W xenon lamp.