A simple synthesis of transparent and highly conducting p-type CuxAl1−xSy nanocomposite thin films as the hole transporting layer for organic solar cells

Inorganic p-type films with high mobility are very important for opto-electronic applications. It is very difficult to synthesize p-type films with a wider, tunable band gap energy and suitable band energy levels. In this research, p-type copper aluminum sulfide (CuxAl1−xSy) films with tunable optical band gap, carrier density, hole mobility and conductivity were first synthesized using a simple, low cost and low temperature chemical bath deposition method. These in situ fabricated CuxAl1−xSy films were deposited at 60 °C using an aqueous solution of copper(ii) chloride dihydrate (CuCl2·2H2O), aluminium nitrate nonohydrate [Al(NO3)3·9H2O], thiourea [(NH2)2CS], and ammonium hydroxide, with citric acid as the complexing agent. Upon varying the ratio of the precursor, the band gap of the CuxAl1−xSy films can be tuned from 2.63 eV to 4.01 eV. The highest hole mobility obtained was 1.52 cm2 V−1 s−1 and the best conductivity obtained was 546 S cm−1. The CuxAl1−xSy films were used as a hole transporting layer (HTL) in organic solar cells (OSCs), and a good performance of the OSCs was demonstrated using the CuxAl1−xSy films as the HTL. These results demonstrate the remarkable potential of CuxAl1−xSy as hole transport material for opto-electronic devices.


Introduction
Industrialization in the past years has yielded an increasing energy demand, which was resolved using non-renewable resources. However, utilization of these resources results in serious environmental pollution and depletion of the fossil fuel energy. To avoid this green renewable energy must be developed and solar cells are one of the most important ways to overcome this problem. [1][2][3] Scientists have devoted much time to developing better carrier transporting layers with excellent properties. [4][5][6] At present, organic solar cells (OSCs) have attracted much attention because of their light weight, exibility, ease of production and high efficiency. [7][8][9][10][11][12] Poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT: PSS) 13,14 is a widely used hole transporting layer (HTL) which is used to modify the anode interface to improve the hole collection ability. However, because of its high hygroscopicity and acidity, 15 PEDOT:PSS has an adverse impact on the device stability. To solve this problem, several materials with high transmissivity within the range of visible light have been used as alternatives, and these are materials such as molybdenum oxide (MoOx), [16][17][18] nickel oxide (NiOx) 19,20 and so on. However, these materials are either toxic or in short supply. In addition, some of them oen require complex vacuum systems for deposition. Therefore, it is highly desirable to develop a highly transparent, earth-abundant, low-cost, non-toxic and non-corrosive HTL for highly efficient OSCs and other optoelectronic applications. Inorganic p-type lms with high mobility are very promising for opto-electronic applications. Nevertheless, it is difficult to synthesize p-type lms with wider, tunable band gaps and suitable band energy levels.
Chalcogenide (CuAlS 2 ) which is known to be a p-type semiconductor, is a promising material because of its high band gap energy (E g ) and hole conductivity. 21,22 There are various methods used to prepare CuAlS 2 , such as chemical vapor transport deposition, solid phase reaction, solvothermal deposition, chemical spray pyrolysis, hydrothermal methods and chemical bath deposition (CBD). [23][24][25][26][27][28] However, these methods involve either high temperature or high pressure and the prepared CuAlS 2 always has a low conductivity, whereas the lm prepared using CBD is not very uniform and lacks measureable electrical properties.
In this research, a CBD technique was used to grow Cu x -Al 1Àx S y thin lm with a tunable band gap in situ. Compared with other deposition methods, [23][24][25][26][27] this method has several advantages such as: (1) the precursors are dissolved in distilled water, which is non-toxic, low cost and environmentally friendly, (2) the chemicals are commercially available and inexpensive, the deposition temperature is 60 C which is really low and very safe, and (3) the process is simple. With this method Cu x Al 1Àx S y lms with a large area can easily be grown without using sophisticated instruments. In addition, citric acid and ammonium hydroxide are used to adjust the speed of the reaction to obtain the required lm.
In comparison with copper(II) sulde (CuS) 29-31 lm, the Cu x Al 1Àx S y lm is easy to grow without needing any sophisticated instruments. Most importantly, the variable energy levels of Cu x Al 1Àx S y lms with a tunable band gap are very attractive. In this research, non-toxic and earth-abundant Cu x Al 1Àx S y thin lm was used as HTL for OSCs with a blend of poly(3hexylthiophene) (P3HT) and 6,6-phenyl C 61 butyric acid methyl ester (PCBM) was used as the active layer. Photovoltaic devices were made with the structure of uorine doped tin oxide (FTO)/Cu x Al 1Àx S y /P3HT:PCBM/Al. The optimized OSCs illuminated under simulated AM1.5G, 100 mW cm À2 white light yielded a power conversion efficiency (PCE) of 2.67% with an open-circuit voltage (V oc ) of 0.596 V, a short-circuit current density (J sc ) of 9.21 mA cm À2 , and a ll factor (FF) of 48.7%, which was comparable to that obtained with reference organic photovoltaics with PEDOT:PSS as HTL.

Material
The polymer donor (P3HT) was purchased from Rieke Metals and acceptor PC 61 BM was obtained from Nano-C. Copper(II) chloride dihydrate (CuCl 2 $2H 2 O), aluminium nitrate nonohydrate [Al(NO 3 ) 3 $9H 2 O], thiourea [(NH 2 ) 2 CS], citric acid (all analytically pure reagents) and ammonium hydroxide were purchased from traditional Chinese chemical companies. All the materials were used without further purication. The solvent used was deionized water.

Preparation of Cu x Al 1Àx S y lms on FTO glass
The FTO substrate was cleaned using ultrasonic cleaning in deionized water, followed by acetone and alcohol for 10 min each. Finally, the substrates were dried in a temperature controlled drying oven.
Aqueous solutions of 0.01 M CuCl 2 $2H 2 O, 0.01 M Al(NO 3 ) 3 -$9H 2 O, 0.04 M (NH 2 ) 2 CS, 0.02 M citric acid and pH adjuster (ammonium hydroxide) was used to prepare a CuAlS 2 thin lm. Firstly, CuCl 2 $2H 2 O, Al(NO 3 ) 3 $9H 2 O and citric acid were placed into a beaker using deionized water as the solvent and then stirred continuously for a few minutes until the solution became homogenous. Then, ammonium hydroxide was added dropwise into the solution until it became a bluish violet colour, and solution was denoted as solution A. The (NH 2 ) 2 CS was placed into another beaker and dissolved in deionized water, which was denoted as solution B. Solution A and solution B were mixed together to obtain the nal solution, and ammonium hydroxide was added dropwise to adjust the pH to 8.8. Subsequently, the FTO substrates were placed into the nal solution at 60 C to obtain the required CuAlS 2 lm. To prepare the Cu x Al 1Àx S y lm (x ¼ 0.2, 0.3, 0.5, 0.7, 0.8, 0.9, 1), the molar ratio of CuCl 2 $H 2 O and Al(NO 3 ) 3 $9H 2 O was varied, while maintaining the total concentration at 0.02 M, as the remaining steps were processed as described previously. Then the Cu x -Al 1Àx S y lms were annealed at 150 C for 10 min and cooled down to room temperature (RT) prior to use.

Fabrication of solar cells
The organic solution was prepared by dissolving 10 mg of P3HT and 10 mg of PCBM in 0.5 ml of chlorobenzene with vigorous magnetic stirring for 24 h before use. The organic solution was used for the active layer deposition. It was spin-coated onto Cu x Al 1Àx S y lms at 500 rpm for 6 s and then at 1000 rpm for 20 s, to give a thickness of about 200 nm. The 100 nm thick Al top electrodes were thermally evaporated through a shadow mask under a pressure of about 10 À4 Pa. Finally, the fabricated devices were thermally annealed on a hot plate at 150 C for 10 min in an argon lled glovebox. The active area of the device was 0.04 cm 2 as dened by the shadow mask.

Films and device characterization
The transmittance of the lms was measured with an ultraviolet-visible -near infrared (UV-Vis-NIR) spectrophotometer (Cary 5000, Varian) in the 300-800 nm wavelength range at RT. The lm thickness was measured using ellipsometry. Field- Fig. 2 The transmission spectra of Cu x Al 1Àx S y films deposited on glass after annealing at 150 C for 10 min (a), and the diagram of (ahn) 2 against hn calculated from the transmission spectra (b). Table 1 Summary of optical and electrical properties of the Cu x Al 1Àx S y film annealed at 150 C for 10 min Hall mobility (cm 2 V À1 s À1 ) This journal is © The Royal Society of Chemistry 2018 emission scanning electron microscopy (SEM; FEI XL-30) was used to observe the morphology of the samples. Transmission electron microscopy (TEM; Jeol JEM-2010) was used for the observation of the ultrastructure. Energy dispersive spectrometry (EDS; FEI XL-30) was used to determine the components of the samples. X-ray photoelectron spectroscopy (XPS) and ultraviolet photoelectron spectroscopy (UPS) were performed using a XPS/UPS system (ThermoScientic, ESCLAB 250Xi, USA). The compositions and chemical states of the CuAlS 2 lms were examined using XPS. Before being tested, the samples were sputter cleaned, to remove atmospheric contamination in the XPS chamber for approximately 30 s, using the lower energy of Ar + , and the Ar + gun was operated at 0.5 kV under a pressure of 1 Â 10 À7 Pa. The vacuum pressure of the analysis chamber was greater than 1 Â 10 À8 Pa. A whole survey scan to identify the overall surface composition and chemical states was performed, using a monochrome Al Ka X-ray source (1486.68 eV), with detection of photoelectrons at a 150 eV energy pass and a channel width of 500 meV. The surface carbon signal at 284.6 eV was used as an internal standard. The work function and band energy levels were measured using UPS. UPS was carried out using helium I a radiation from a discharge lamp operated at 90 W, a pass energy of 10 eV, and a channel width of 25 meV. A À9 V bias was applied to the samples, in order to separate the sample and determine the low kinetic energy cutoffs. The conductivity, carrier concentration and mobility were measured using a Hall effect measurement system (Lake Shore Cryotronics, 7704A). The current-voltage (J-V) curves of the devices were obtained using a computer controlled Source Measure Unit (Keithley 2400) and the device test was carried out in a glove box under illumination of AM1.5G, 100 mW cm À2 (the light intensity was calibrated using a silicon photodiode) at RT using a solar simulator.

Results and discussion
The SEM images of Cu x Al 1Àx S y thin lms annealed at 150 C for 10 min with various composition are shown in Fig. 1. The ratio shown in each SEM image is the ratio of Cu:Al which refers to x  ¼ 0.2, 0.3, 0.5, 0.7, 0.8, 0.9, or 1. The lms deposited at 60 C for 10 min were very uniform. The particle size was dened by testing several small particles using a scaleplate when magnifying the picture. The average size of the particles was about 50-100 nm. As observed, both a cluster-by-cluster process and an ion-by-ion process occurred in the reaction, and the ion-by-ion process was dominant in the formation which could be seen from the small crystal domains in the lms. The CuAlS 2 (5 : 5) and CuS (10 : 0) lms show a similar, smoother lm because of their structure, whereas the Cu x Al 1Àx S y (8 : 2) lm has similar round domains. The transmission spectra of Cu x Al 1Àx S y thin lms are shown in Fig. 2(a) (x ¼ 0.2, 0.3, 0.5, 0.7, 0.8, 0.9, 1). The lm transmittance increased with Al content when the thickness of the lms was approximately 200 nm. It should be noted that in this research, the lm thicknesses were all tested using an ellipsometer.
The energy gap E g was calculated from the equation: where a is the absorption coefficient, which can be calculated from the equation: B is a constant and r is an index, which could have values of 1/2, 3/2, 2 and 3, depending on the nature of the electronic transition. The exponent r ¼ 1/2 is for allowed direct transition, r ¼ 3/ 2 is for forbidden direct transition, r ¼ 2 is for forbidden indirect transition, and r ¼ 3 is for allowed indirect transition. The CuAlS 2 lm exhibited allowed direct transition. The optical band gap of Cu x Al 1Àx S y lm according to the dependence of (ahn) 2 on hn was conrmed. The estimated band gaps of the Cu x Al 1Àx S y lms at different Cu concentrations are illustrated in Fig. 2(b). Using Fig. 2(b), the CuAlS 2 lm (x ¼ 0.5) was calculated to have a band gap of 3.60 eV which matched well with those gures reported in the literature. Also, as expected, CuS lm (x ¼ 1) had an optical gap of 2.63 eV, which was very close to the reported values (2.4 eV). As can be seen in Fig. 2(b), starting from CuS, the band gap of the lms increased from 2.63 eV to 4.01 eV with the decrease of x (which means that the ratios of Cu to Al became smaller) and the transmittance increased with the decreasing Cu content. Thus, metal chalcogenides with a tunable optical band gap could be obtained by varying the precursor ratio of Cu and Al, and this is promising for applications in optoelectronic devices.
The electrical properties were characterized using Hall effect measurements. As seen in Table 1, there was a variation of optical and electrical parameters of the Cu x Al 1Àx S y thin lms which were annealed at 150 C for 10 min. For each composition with a particular molar ratio, more than three samples were made and tested to ensure the reproducibility, and all the lms displayed p-type conductivity. The data listed in Table 1 for the reference was the one which was closer to the average result. Overall, the resistivity shows a tendency to decrease with increasing Cu concentration, as shown in Fig. 3(b). A maximum conductivity of 546 S cm À1 in lms with x ¼ 0.7 was achieved, which was much higher than the values reported for the p-type HTLs. 32,33 It is interesting that the resistivity of the lms between x ¼ 0.7 and x ¼ 0.8 has a sudden rise, the reason for this is not known at the moment. Hole concentration and mobility were measured using a 7704A Hall system (Lake Shore Cryotronics). As shown in Fig. 3(c), hole mobility appears to increase gradually as the Cu concentration increased, within the range of 0.2 < x < 0.5, and then gently decreased and this takes no account of the CuS lm. Compared with other samples, the high conductivity of the lm with x ¼ 0.7 originates from the relatively higher carrier concentration and mobility. In Fig. 3(d), hole concentration varies from (1-4) Â 10 21 cm À3 , which is in the range of a highly doped degenerate semiconductor. Fig. 3 reveals that the band gaps of 2.63-4.01 eV were comparable to p-type transparent materials such as aluminium copper dioxide (CuAlO 2 ; 3.6 eV), and the hole conductivity was relatively high, which was ascribed to the considerable mobility of 0.4-1.5 cm 2 V À1 s À1 and a hole concentration of 1À4 Â 10 21 cm À3 . CuS is well known as a p-type conductor. Thus, the hole conduction in the Cu x Al 1Àx S y lms was attributed to the CuS phase. The conducting network formed by CuS (even in Al rich samples) leads to the high p-type conductivity in the lms.
As discussed previously, transparency is more dependent on Cu and Al contents, whereas the Cu content plays an important role in the hole conductivity as well. Of the lms with <1000 U sq À1 sheet resistance, except for x ¼ 1, the highest transparency was found at x ¼ 0.2 because it had the highest Al content whereas the lowest sheet resistance and the highest hole mobility was found at x ¼ 0.5. The EDS measurements were used to determine the components of the Cu x Al 1Àx S y lms. As shown in Table 2, the atom percentage of Al decreased from 76.02% to 0% when x increased, which may be the main reason for the reduction of the optical band gap. In addition, the increasing atom percentage of Cu and S when x increases might be responsible for the increasing hole mobility. For the CuS lm (x ¼ 1), the mole ratio of Cu and S revealed the existence of copper sulde (Cu 2 S) and CuS. The possible existence of the O element in the lm may result in the higher energy gap (2.63 eV).
The XPS spectra of CuAlS 2 lms (x ¼ 0.5, annealed at 150 C) deposited on glass are shown in Fig. 4. Fig. 4(a) shows a full scale scan of the results of the XPS which found the peaks of Cu, Al, O, S and C. The magnied peaks of the Al 2p, Cu 2p and S 2p scan are shown in Fig. 4(b), (c) and (d), respectively. Fig. 4(b) shows the peak tting of the Al 2p spectra. The peak tting for the Al 2p line was divided into two peaks, the peaks at 77 eV and 74 eV revealed the presence of Cu 3p 3/2 and Al 2p. The peaks of Cu 2p were at 952.9 eV and 932.9 eV. The core levels of Cu 2p 1/2 indicated that there was a divalent Cu ion in the product whereas the core level of Cu 2p 3/2 refers to the Cu + . 34 The peaks at 162.0 eV belonged to S 2p. These results were consistent with Fig. 6 The UPS spectrum of Cu x Al 1Àx S y film on FTO glass. (a) This journal is © The Royal Society of Chemistry 2018 RSC Adv., 2018, 8, 16887-16896 | 16893 Paper the results found in the literature and have proved that the atomic ratio of elemental Cu and Al was approximately 2 : 1. For further analysis of the CuAlS 2 lms, TEM measurements were made to observe the ultrastructure. As seen in Fig. 5(a), the particles seem very large which may be because of the thick lm. Fig. 5(b) is a typical high resolution TEM (HRTEM) image of the lm of CuAlS 2 , where 10 crystal planes were tested to get an average interplanar distance for each crystal orientation. It was concluded that there is CuAlS 2 (112) in the CuAlS 2 lm with a corresponding interplanar distance of 3.04Å. The selected area electron diffraction (SAED) pattern of the CuAlS 2 lm in Fig. 5(c) shows clearly that the CuAlS 2 lm is a polycrystalline compound.
The UPS measurement was carried out using helium I a as the UV source. It can be concluded from Fig. 6 that the binding energies of the Cu x Al 1Àx S y lms (x ¼ 0.2, 0.3, 0.5, 0.7, 1) were 16.07 eV, 15.89 eV, 16.08 eV, 15.92 eV, 15.89 eV, respectively. Thus, the work function of the Cu x Al 1Àx S y lms (x ¼ 0.2, 0.3, 0.5, 0.7, 1) annealed at 150 C for 10 min was 5.15 eV, 5.33 eV, 5.14 eV, 5.30 eV, 5.33 eV, respectively. Fig. 6 also shows the energy difference between the top of valence band (E V ) and the Fermi level, and it was concluded that the E V of the Cu x Al 1Àx S y lms (x ¼ 0.2, 0.3, 0.5, 0.7, 1) were 0.50 eV, 0.25 eV, 0.21 eV, 0.08 eV, 0.13 eV below the Fermi level, respectively, which were 5.65 eV, 5.58 eV, 5.35 eV, 5.38 eV, 5.46 eV, respectively. Take account of the optical band gap of Cu x Al 1Àx S y lms mentioned previously in Table 1, the bottom of the conduction band (E A ) is set at 1.64 eV, 1.82 eV, 1.75, eV 2.13 eV, 2.83 eV for x ¼ 0.2, 0.3, 0.5, 0.7, 1, respectively. The device structure and energy level alignment are shown in Fig. 7. From the band alignment, the energy band structure of the Cu x Al 1Àx S y lms (x ¼ 0.2, 0.3, 0.5, 0.7, 1) was determined, and the energy band levels obtained for the Cu x Al 1Àx S y lms with excellent p-type conductivity could be suitable for many opto-electronic devices in the future.
To demonstrate the application of p-Cu x Al 1Àx S y lms in photovoltaic devices, several heterojunction OSCs were fabricated with the Cu x Al 1Àx S y lm used as hole transporting layers. The J-V characteristics, which were measured under standard test conditions (1000 W m À2 , air mass 1.5 global (AM1.5G) spectrum and 25 C) for the 0.04 cm 2 device, are presented in Fig. 8(a). Notably, it was found that a 40 nm CuAlS 2 lm annealed at 150 C provided superior performance. As summarized in Table 3, for the FTO/Cu x Al 1Àx S y/ P3HT:PCBM/Al devices, the PCE of the OSCs increased from 1.22% to 2.45% when x increased from 0.2 to 0.5 because of the increasing hole mobility and the better matched band between Cu x Al 1Àx S y and P3HT. However, the PCE decreased while x increased from 0.5  to 1, and the reasons for this phenomenon might be the decreasing band gap and the decline of the valence band edge. Compared with CuS lm and other Cu x Al 1Àx S y lms (x ¼ 0.7, 0.9) in Fig. 2(a), more light was transmitted by the CuAlS 2 lm with a wider band gap, which allows more use of the incident light under the same conditions. Furthermore, the variation of E V is in agreement with the changes in PCE, the E V of CuAlS 2 lm is closer to the highest occupied molecular orbital (HOMO) of P3HT compared with other lms as shown in Fig. 7(b). Thus, the solar cells using CuAlS 2 lm as HTL produced the highest PCE. Although the energy levels between Cu x Al 1Àx S y and P3HT were not sufficiently matched to achieve good hole collection, the variable energy levels of the Cu x Al 1Àx S y lms with tunable band gap were still attractive, and there will be more research with better matched bands in future. Finally, the best PCE was achieved using CuAlS 2 lm as HTL, and a J sc of 9.21 mA cm À2 and a V oc of 596 mV were observed, delivering a PCE of 2.67%, which is shown in Fig. 8(b). Further comparison with other HTL can be found in Table S1 (ESI †). Therefore, the CuAlS 2 lm is a good material for solar cell applications.

Conclusion
In summary, the optical and electrical characteristics of the Cu x Al 1Àx S y thin lms prepared on glass sheets and FTO glass using CBD were studied. Upon varying the precursor ratio of the CBD solution, p-type Cu x Al 1Àx S y lms with different optical and electrical properties were obtained. The band gap can be adjusted from 2.63 eV to 4.01 eV. In particular, the CuAlS 2 lm had a band gap of 3.60 eV and a hole mobility of 1.00 cm 2 V À1 s À1 .
Consequently, the lms could be used as a hole transporting layer for photocells. The addition of a CuAlS 2 layer between the anode and the active layer in OSCs can signicantly improve the device performance, leading to a 2.67% power efficiency with a device structure of FTO/CuAlS 2 /P3HT:PCBM/Al under simulated AM1.5G 100 mW cm À2 illumination. This indicates that CuAlS 2 is a very promising alternative HTL for OSCs and other opto-electronic devices.

Conflicts of interest
There are no conicts to declare.