Hybrid CuI@g-C₃N₄/MoS₂ Thin Films for Energy Conversion Applications: A Photoelectrochemical Characterization

Abstract

In this work, we report the fabrication of CuI@g-C₃N₄/MoS₂ thin films by the thermal evaporation of Cu films and their conversion into hybrid films by a simple wet chemical method. Compared to pure CuI, CuI@g-C₃N₄/MoS₂ shows enhanced absorption near the UV region, which improves its DC photoconductivity. The conductivity of the films is enhanced by the addition of g-C₃N₄/MoS₂, which is distributed on the surface of the CuI film. The band gap of the films red-shifts upon adding g-C₃N₄/MoS₂. We evaluate this material’s potential application as a photodetector and in photocatalysis by evaluating its photoelectrochemical properties using impedance spectroscopy measurements, cyclic voltammetry, and DC photoresponse measurements. We find that upon the addition of g-C₃N₄/MoS₂, the conductivity of the films is increased, as evidenced by the time-dependent photo amperometry measurements. Also, a higher DC photoresponse is observed upon increasing the concentration of MoS₂. This work marks the first time a hybrid CuI@g-C₃N₄/MoS₂ film and its photoelectrochemical characteristics have ever been reported.

1. Introduction

Transparent semiconductor thin films are promising due to their applications in transparent electronics, such as OLED and AMOLED displays. Projections are such that this kind of technology is predicted to dominate an important portion of the near-future market [1]. Usually, semiconducting oxide thin films possess n-type conductivity, and their wide band gaps lead to efficient electron injectionn due to their bonding properties, so, from a technological point of view, it is important to produce p-type transparent semiconductors. The first instance of a p-type transparent semiconductor was proposed by Kawazoe et al., who reported a Delafossite CuAlO₂ [2]. Since then, many researchers have explored different materials, such as NiO [3] and Cu-Cr-O Delafossites [4], perovskites such as SrCuO₃, oxichalcogenides, and Cu_xO [5,6]. Among the available p-type transparent semiconductors, CuI is an important material with a direct band gap of 3.1 eV, relatively high hole mobility of 44 cm²/V-s, high p-type conductivity, and high carrier concentrations of up to 10²⁰ cm⁻³ [7]. The conductivity of the films can be modified by tuning the charge carrier concentrations by doping [8]. The material can be easily synthesized by many methods, such as thermal evaporation [9] and sputtering. Many devices based on transparent p-type CuI have been reported in the literature. One example is a transparent self-powered photodiode based on γ-CuI/β-Ga₂O₃ deposited using the thermal evaporation of the starting salts. This work reported a high rectification ratio and UV photovoltaic action [10]. Kim and co-workers obtained an n-ZnO/p-CuI with a passivation film made of ZnS. In their work, the rectification ratio of the n-ZnO/p-CuI was improved to 10⁷ by employing a ZnS film in between the ZnO and the CuI; this helped to passivate the oxygen vacancies, which improved the rectification ratio [11]. Some other materials employed for transparent diodes based on CuI are BaSnO₃ [12], SiZnSnO [13], AgI [14], and NiI₂ [15]. These materials provide a high rectification ratio and, thus, high-quality junction and electrical properties of the diode; however, their application is limited by the complicated nature and high cost of the experimental techniques used to obtain the materials and their desired properties. A way to tune and optimize material properties is by blending materials with other organic or inorganic materials to form hybrid films. Among the materials used to make hybrid films, the n-type material g-C₃N₄ is a relatively new and unexplored option for optoelectronic applications coupled with CuI. Usually, carbon nitride thin films find extensive applications in photocatalysis [16], anticorrosion coatings [17], perovskite solar cells [18], and solar fuel generation [19,20,21], among others. One example is carbon nitride thin films with a graphene-like structure in which an infinite extension of triazine rings or tri-s-triazine rings forms the basic structural unit. This kind of structure allows a unique electronic structure and crystal sites that allow the material to have a large surface area and, hence, more active sites for charge separations. The covalent bonding from the structural unit allows the material to have excellent chemical and thermal stability up to 600 °C. Its conduction and valence band positions and moderate bang gap of 2.7 eV make it an attractive material for photodiodes and solar cells [22]. The material has some drawbacks, such as its high recombination rate, low conductivity, and low specific surface area, which are easily improved by various methods, most of which are compatible with low-cost solution methods [23]. Depending on the application, the properties of carbon nitride can be tuned by controlling the defects on the material, or by passivation effects from the carbon nitride itself, acting as an additive [24]. For example, the band gap of g-C₃N₄ can be tuned with MoS₂ to form a hybrid material with an extended absorption edge near the visible spectrum so the material can utilize more photons of the visible spectrum and tune the optoelectronic properties of the material [25]. MoS₂ possesses an indirect band gap of 1.8 eV, which changes to a direct one when it is in the form of thin sheets. This material has a 2H hexagonal crystalline structure, which makes it compatible with carbon nitride. Its optical and electrical properties allow it to be used in different applications, such as sensors, photodetectors, photocatalysis, etc. [26]. Quantum dots of MoS₂ were embedded in g-C₃N₄ by mixing both powders appropriately. The resulting hybrid materials modified the absorption edge of the carbon nitride from 2.62 eV to 2.52 eV [27]. The simple mechanical milling of the exfoliated powders of both MoS₂ and C₃N₄ leads to a hybrid material with an absorption edge red-shifted to 500 nm, compared to the fundamental edge of carbon nitride near 400 nm [28]. Films of g-C₃N₄/GO/MoS₂ were prepared using a simple hydrothermal method. Incorporating the hybrid nanocomposite improved the photocurrent of the material by four times compared to the pristine carbon nitride material [29]. A heterojunction thin film of MoS₂/g-C₃N₄ was employed to increase the photocurrent of pristine carbon nitride. Aside from the shift to 1.69 eV, there was an improvement in the charge transfer due to the heterojunction [30]. Control of charge transfer is important for electronic applications, as it influences the rectification ratio and, thus, the quality of the diode. An appropriate band positioning of the respective n- and p-type semiconductor materials can modify charge transfer. The modification of band positioning can be achieved by the hybridization of the materials and doping. Carbon nitride offers a simple way of controlling band positions using chemical engineering and non-chemical methods [31]. Theoretical studies show that nanosheets of g-C₃N₄ can enhance the material’s conductivity [32]. So far, the only works reported on the heterojunction of CuI with g-C₃N₄ are for photocatalysis, supercapacitors, and hydrogen production [33,34,35]. Based on the lack of studies on CuI modified with MoS₂/g-C₃N₄, we report for the first time a hybrid film based on CuI with a surface modification of particles of g-C₃N₄/MoS₂. We establish this composite’s optical and electrical properties by employing DC photoresponse and impedance spectroscopy. We find an increase in conductivity and charge transfer upon adding g-C₃N₄/MoS₂ into the CuI film.

2. Materials and Methods

2.1. Synthesis of CuI@g-C₃N₄/MoS₂ Thin Films

2.1.1. Chemicals

This study used Cu beads (Cu, 99.99% purity), iodine powder (Fermont, 99.999% purity), melamine (Sigma, St. Louis, MO, USA, 99%), glycine (Sigma, purity > 98%), and 2-propanol. The reagents were used as received without further purification.

2.1.2. Synthesis of CuI Thin Films

FTO-coated glass (Qunguan 7–8 Ω/m²) and soda–lime glass (Corning) were cut to form a glass 25 × 25 × 1.2 mm in size in the case of FTO and 25 × 25 × 0.7 m in size in the case of soda–lime glass. Subsequently, these substrates were cleaned in an ultrasonic bath, employing a soap solution, water, and ethanol for 10 min each. Afterward, the substrates were cleaned with dry air. The cleaned substrates were placed in a thermal evaporation system (Torr International (Marlboro, NY, USA), Model No: THE2-2.5 kW-TP) at a constant rate of 2 Å/s under a high vacuum of 10-6 Torr. The thickness of the Cu layer was kept at 50 nm, as measured with a quartz detector and profilometry. The as-grown Cu layers were removed from the system and dipped in an I₂ solution at a fixed concentration of 0.15 M for 1 min. The films were readily converted into CuI, as evidenced by the color change from metallic and shiny brown to transparent reflective film. Afterward, they were washed with pure 2-propanol and dried with a gentle airflow.

2.1.3. Synthesis of g-C₃N₄/MoS₂ Composite

To synthesize the composite, we followed the subsequent methodology: First, we synthesized g-C₃N₄ via thermal decomposition of melamine on a conventional muffle furnace at 550 °C for 4 h, using a ramp of 2 °C/min. The as-prepared powder was pulverized to homogenize the particle size. Next, we prepared a colloidal suspension of the powders in 2-propanol at a concentration of 3 mg/mL and sonicated it for 24 h to obtain the graphitic form of the carbon nitride. To obtain the g-C₃N₄/MoS₂ composite, we added 1%, 3%, and 5% powder of MoS₂ to the previously mentioned colloidal suspension. The sonication process was repeated for these colloidal suspensions to obtain the composites.

2.1.4. Synthesis of CuI@g-C₃N₄/MoS₂ Hybrid Thin Films

The composite CuI@g-C₃N₄/MoS₂ was obtained using the previously prepared colloidal suspension to dissolve the iodine powder at a concentration of 0.15 M. The Cu films were dipped in these solutions (concentration of g-C₃N₄ 3mg/mL; concentration of MoS₂ varied based with respect to the concentration of g-C₃N₃ as, 1%, 3%, and 5% of weight) for 1 min, rinsed with 2-propanol, and dried with a gentle flow of warm air. These samples were labeled as CNM-1, 2, and 3, respectively.

2.2. Characterization

Structural characterization was conducted using a PANalytical (Malvern, UK) diffractometer in the 2θ range of 10–70° using a step size of 0.01° with a time per step of 10 s. Vibrational spectra were obtained with a DXR 2 Raman Microscope (Thermo, Waltham, MA, USA) with a 532 nm wavelength laser. The optical absorption, specular transmission, and reflection of the films were measured with a Uv-Vis-NIR Jasco V-770 Spectrophotometer (Thermo). The morphology of the films was studied with a Scanning Electron Microscope (SEM), Hitachi SU 8020 (Tokyo, Japan). A Transmission Electron Microscope JEOL 2010 (USA, MA)was employed to observe the interfacial characteristics of the composites. The chemical composition of the samples was determined by an X-ray photoelectron Spectrometer (XPS), Thermo K-Alpha, using monochromatized Al-Kα radiation (hv = 1486.68 eV). DC photocurrent measurements were performed with a pico ammeter/voltage source, Keithley 6487 (Cleveland, OH, USA). For this purpose, silver electrodes 5 × 5 mm in size were painted on the surface of the samples deposited in bare glass using Ag colloidal paste (SPI Supplies, Nieuw-Vennep, the Netherlands). The samples were illuminated with a blue LED with 30 W power (power density of 4.8 W/cm²) and a wavelength of 460–465 nm. Electrochemical studies were conducted using an electrochemical workstation (CS350 EIS potentiostat/galvanostat, Corrtest Instruments, Wuhan, China). Cyclic voltammetry (CV), impedance spectroscopy (EIS), and Chronoamperometry (CA) analyses were performed with the CNM-1, CNM-2, and CNM-3 electrodes as photoanodes. Six films were synthesized using the specific method described in Section 2.1, ‘Synthesis of CuI@g-C₃N₄/MoS₂ Thin Films’. Each film was deposited onto an FTO substrate, ensuring an exposed material area of precisely 1 cm². The synthesized films included CuI, CNM-1, CNM-2, CNM-3, and the reference films CuI@g-C₃N₄ and CuI@MoS₂. Platinum foil was used as a counter electrode and Saturated Calomel Electrode (SCE) as a reference electrode. As the electrolyte, 0.5 M potassium chloride (KCl) solution was chosen. CV profiles were registered from −0.6 V to 0.6 V vs. SCE at scan rates of 100 mVs⁻¹. EIS was performed in open circuit potential (OCP) conditions with a frequency range of 1–10⁵ Hz and an amplitude of 10 mV. Before electrochemical studies were conducted, the solution was kept under an argon atmosphere for 30 min.

3. Results

3.1. Structural Characterization

Figure 1 shows the X-ray diffraction pattern of the CNM-1, CNM-2, and CNM-3 samples. A sample of a pure CuI film was also analyzed for comparison. It is observed that diffraction peaks occur at the 2θ angles of 25.57°, 29.51°, 42.306°, 49.89°, and 52.306°, which are due to the reflections of the (111), (200), (220), (311), and (222) planes of the zinc-blende structure of CuI (pdf number 060246) 29. From the diffraction pattern, the pure CuI film is observed to have a preferential orientation to the (111) family of planes. When the g-C₃N₄/MoS₂ composite is added into the CuI thin films, additional diffraction peaks are observed at 2θ values of 14.53° and 27.48°, which are due to the reflections of the (002) plane of the 2H hexagonal crystal of MoS₂ (pdf number 0771716) [36] and (002) plane of the tri-s-triazine building block of g-C₃N₄ [37]. These peaks are marked with a # and a *, respectively, in the diffractogram. They can be explained by the successful incorporation of the g-C₃N₄/MoS₂ composite into the CuI thin films. The intensity of the MoS₂ peak increases as the quantity of the powder increases from 1% wt to 5% wt, which is in agreement with the experiment. The concentration of g-C₃N₄ is kept constant in all experiments, so the peak intensity is always the same. Since the carbon nitride is in the monolayer form, the respective reflection is expected to have a low peak intensity due to the fewer planes contributing to this peak intensity. No other impurity phases, such as metallic Cu or CuO, are observed.

The average crystallite size of the samples was calculated using Scherrer’s equation [38]:

$$D={\displaystyle \frac{0.9\lambda }{\beta cos\theta }}$$

(1)

where D is the average crystallite size, λ is the diffractometer wavelength, β is the full-width at half maxima of the peak in radians, and θ is the Bragg’s angle in radians.

To conduct the analysis, we selected the (111) reflection of all the samples under consideration. The results showed that the as-prepared sample of CuI thin films shows an average crystallite size of 58 nm. The crystallite size increases slightly in the range from 60 to 65 nm for samples CNM-1, CNM-2, and CNM-3, respectively.

Figure 2 depicts the Raman spectra of all the samples. We observe Raman peaks for all the samples at 121.91 cm⁻¹ due to the transverse optical vibrational mode of the Cu-I bond. Peaks appear at 383 and 407 cm⁻¹ in sample CNM-3, which are due to the vibrational modes of MoS₂, the E_2g¹ longitudinal in-plane vibration of the S atom plane with adjacent Mo atoms, and the optical mode A_1g due to the out-of-plane S vibrations that occur because of optical excitation [39]. The bulk sample of MoS₂ is shown in Figure 3. It can be observed that both Raman modes are also present at 378 and 404 cm⁻¹, respectively. We observe a red-shift in the A_1g mode of the CNM-3 sample. The shift is 4 cm⁻¹, which could be due to decreased interlayer interaction [40]. A red-shift in E_2g¹ is also observed in this sample. This can be explained by the electrical interaction nature of the S-Mo-S layers, since exfoliation plays no significant role in the elastic constant of the S-S vibrational mode. Exfoliation usually tends to decrease the layers of bulk MoS₂ to the order of a few layers or even monolayer thickness. Since we observe a red-shift in the mode, we can assume that ionic chalcogen–metal interaction results from this red-shift with decreased layers [41]. Peak intensity also increases when the sample is in the thin film rather than in the bulk phase. This effect is due to the optical interference of the laser radiation and the emitted Raman scattering radiation [41]. The higher background and noise present in this sample could be due to the inelastic scattering produced by the powder sample. All these results confirm that the phases present are crystalline and consist of CuI and the composite material.

The complete conversion of metallic Cu thin films into CuI follows the reaction

2Cu + I₂ → 2CuI

(2)

in which the iodine from the solution readily reacts with the Cu atoms on the surface of thin films to form solid CuI. The electronegativity value of Cu is listed as 1.9 on the Pauling Scale, and the I electronegativity value is listed as 2.26 on the same scale; this difference explains why the reaction proceeds in 1 min [42]. Other sources of Cu have different reaction kinetics, such as Cu₃N, in which the reaction rate is slower than the metallic Cu precursor, due to the differences in electronegativity values [43]. Oxides, such as CuO, can be found after the application of specific deposition methods. For example, Li and co-workers reported that the synthesis of CuI from metallic Cu using a hydrothermal method at 200 °C with a stainless steel autoclave produced CuO on the resulting films [38]. In our case, no oxides can be detected by XRD or Raman. Also, no metallic Cu can be observed from these measurements, so we can conclude that a complete conversion occurred. An increase in thickness from 50 nm to 150 nm is observed in the samples, independent of the concentration of MoS₂ employed. This resulting increase in thickness is due to the volume expansion of the lattice upon the addition of iodine. The preferential orientation in the (111) direction is due to the lower surface energy of the respective plane [44].

3.2. Chemical Composition

Figure 4 shows the survey spectra and high-resolution spectra of the MoS₂ and g-C₃N₄ powders. Figure 4a shows that only elements corresponding to the starting materials are present in the samples. The small oxygen signals present in the samples might be due to adventitious carbon. No other impurities are observed. Figure 4b shows the high-resolution spectra of the C 1s and N 1s levels for carbon nitride and Mo 3d and S 2p for MoS₂ powders. The carbon nitride sample from the melamine precursor shows two distinctive photoelectron peaks for the C 1s core level. The peak at 284.8 eV is due to C=C coordinated carbon, while the peak at 287.98 eV corresponds to the carbon bonded with sp³ hybridization in the C-N=C backbone. For the N 1s core level, the peak at 398.48 eV corresponds to the sp²-bond in N in the triazine rings, and the peak at 399.78 eV is due to three carbons coordinated with a N, labeled as N-(C)₃ in the figure. The peak at 400.88 eV is due to the amino functional group present in the carbon nitride [45]. In the case of MoS₂, the Mo 3d core level shows a doublet peak, which is centered at 229.68 and 232.78 eV, and these peaks are due to the Mo⁴⁺ 3d_5/2 and Mo⁴⁺ 3d_3/2 components present in the MoS₂ with 2H symmetry. Also, a small peak is observed for this core level centered at 226.98 eV, which is due to the 2s component of the material. For the S 2p core level, a doublet photoelectron peak is centered at 162.48 eV and 163.68 eV due to the S 2p_3/2 and S 2p_1/2 components of MoS₂. These values agree with the reported values for bulk MoS₂ [46]. The Mo 3d level has a spin–orbit coupling component of 3.1 eV and an intensity ratio of 3:2 for Mo 3d_5/2 and Mo 3d_3/2, respectively. The S 2p level has a spin–orbit coupling component of 1.2 eV and an intensity ratio of 2:1 for the S 2p_3/2 and S 2p_1/2 components, respectively. These values agree with the reported values found in the literature [47]. The atomic percentage for g-C₃N₄ is 49.78% for C and 48.89% for N, an almost stoichiometric ratio. The atomic percentage for MoS₂ is 13.59% for Mo and 29.12% for S. The atomic ratio for MoS₂ is 2.1:1, which is in good agreement with the chemical formula of this compound.

Figure 5 shows the survey spectra for samples CNM-1, CNM-2, and CNM-3. The spectra indicate the presence of Cu, I, C, N, Mo, and S due to the CuI thin films and the composite g-C₃N₄/MoS₂. All the samples present an O 1s signal, which could be due to contamination from adventitious carbon and oxygen.

To gain further insight into the chemical states and chemical interactions of CuI and the composite, we analyze the high-resolution spectra for the core levels C 1s, N 1s, Mo 3d, S 2p, Cu 2p, and I 3d for samples CNM-1, CNM-2, and CNM-3, respectively. The results are presented in Figure 6. It can be observed that for the C 1s core level, there are three different photoelectron peaks centered at 284.6 eV, 286.28 eV, and 288.28 eV, which are due to C-C from graphitic and adventitious carbon on the surface, C-O from adventitious carbon, and the C-N=C backbone, respectively. For the N 1s core level, photoelectron peaks centered at 399.2 eV, 400.6 eV, and 402.3 eV are due to sp²-bonded N in triazine rings, the N-(C)₃ coordinated N and the amino functional groups, respectively [45]. An additional signal centered at 395.2 eV is present because of the Mo 3p component due to the Mo-S bond [46]. The Mo 3d core level shows a doublet peak, with the peaks centered at 229.4 eV and 232.5 eV, which are the Mo 3d_5/2 and Mo 3d_3/2 components of the Mo-S bond, having a spin–orbit component of 3.1 eV. Also, a small peak centered at 226.8 eV is present due to the S 2s photoelectron peak from the Mo-S bond [46]. The S 2p core level has a doublet photoelectron peak, with binding energies of 162.1 and 163.e eV due to the S 2p_3/2 and 2p_1/2 doublets for the Mo-S bond. They have a spin–orbit coupling component of 1.16 eV [46]. The Cu 2p core level has a doublet peak with binding energies of 932.4 eV and 952.2 eV due to the Cu 2p_3/2 and Cu 2p_1/2 components of the Cu-I bond. A minor feature at 931.1 eV is due to the iodine photoelectron peak. The I 3d core level has a doublet photoelectron peak with binding energies at 619.8 and 631.3 eV due to the I 3d_5/2 and I 3d_3/2 components of the Cu-I bond [48]. These results indicate that the Cu film is readily converted into CuI upon reacting with the composite solution of iodine, which confirms our XRD studies. Compared with Figure 4, the intensity of the C-C photoelectron peak greatly intensifies on the thin film. The N-C=N signal is shifted to higher binding energies than the bulk sample. Also, the peak intensity for the C=C peak is stronger than the bulk phase. This shows that adventitious carbon is highly present in thin films. Compared with the bulk phase, in which a higher number of s-triazine units are present, the exfoliated phase can contain fewer units in the bulk, which explains the lower intensity of the N-C=N signal. The signals of the N 1s components are all in the same ratios compared to the bulk sample, which is consistent with previous reports [49]. No changes in the binding energies for the C and N levels are expected upon exfoliation, as previously reported by Gowri and co-workers [50].

The valence band spectra of the samples were analyzed. Figure 7 presents the valence band spectra of the bulk samples and samples CNM-1, CNM-2, and CNM-3, respectively. The valence band edge was extracted from the XPS valence spectrum using a method similar to that in [51]; a straight line is extrapolated from the valence band edge to the onset of the band. The intersection of these two straight lines is the valence band edge. The calculated band edges for the pure samples are 2.06 eV, 1.07 eV, and 0.33 eV for carbon nitride, MoS₂, and pure CuI, respectively. For the CuI@g-C₃N₄/MoS₂ thin films, the extrapolated valence band edges are 0.02 eV, 0.97 eV, and 0.7 eV for samples CNM-1, CNM-2, and CNM-3. The valence band for the CNM-1 sample is lowered compared to pure CuI, while the samples CNM-2 and CNM-3 are moved to higher energy levels compared to CuI. This shift in energy comes from the interaction of the electron cloud with π bonding in carbon nitride and from the higher interaction of I species with MoS₂ and carbon nitride.

3.3. Morphology

Figure 8 shows the sample’s SEM and high-angle annular dark-field TEM images. Figure 8a shows the SEM micrograph of a pure CuI thin film. The surface morphology of CuI consists of a rough surface with large crystals. Usually, the surface morphology of CuI consists of large and irregularly shaped crystals due to the fast kinetics of the reaction of Cu with the iodine solution. For example, different kinds of morphologies are obtained by the spin coating of CuI powders in acetonitrile solvent, in which conditions such as concentration and doping atoms alter the surface roughness of the film [52]. Figure 8b, Figure 8c, and Figure 8d are the SEM micrographs of samples CNM-1, CNM-2, and CNM-3, respectively. It can be observed that some islands are present on the surface of the film, which are the composite material g-C₃N₄/MoS₂. The film morphology of CNM samples consists of smaller grains than the pure CuI thin film. This is because of a stabilization effect due to the solution of I_2, which coordinates with the suspended g-C₃N₄ to generate a stable suspension [53]. This stable suspension yields higher surface energy to the Cu thin film, which, in return, leads to more uniform film morphology. Different methodologies produce different crystal orientations and morphologies [54]. As some theoretical research has proven, the observation of the agglomerates of g-C₃N₄/MoS₂ leads to a more uneven film surface but can lead to more active sites for photocatalytic reactions [55]. TEM analysis is used to study the g-C₃N₄/MoS₂ agglomerates. Figure 8e is the transmission image of the sample. It can be observed that the largest particle consists of the MoS₂ nanoparticles supported on g-C₃N₄. The carbon nitride consists of only a few layers, as expected from the mechanical exfoliation process. To address the bulk CuI@g-C₃N₄/MoS₂ heterojunction, we performed a dipping process on the Cu grid used to support the composite, resulting in the material being formed on the grid and, simultaneously, supported by it. The high-angle annular dark-field image in Figure 8f confirms this. The thickest material is the formed CuI, while the layered semiconductor is g-C₃N₄. The corresponding HRTEM micrograph of the interface of MoS₂ and g-C₃N₄ and SAED analysis of the CuI@g-C₃N₄/MoS₂ of sample CNM-3 are shown in Figure S1.

3.4. Optical Properties

Figure 9 shows the absorbance spectra of all the films. It can be observed that all the films possess a strong absorption onset at 400 nm, which corresponds to the fundamental absorption edge. Another strong absorption peak appears at 337 nm, which is due to the transition from the split-off band [56]. The fundamental edge of CuI is of a direct semiconductor, as already reported in the literature [36]. The absorption intensity increases with the increase in the concentration of MoS₂, as can be seen in the figure. This increase is due to the absorption of the composite g-C₃N₄/MoS₂ [57]. Absorption decreases upon increasing the concentration from 3% to 5%. This is explained by the SEM micrographs, in which larger particles are observed on the surface of the samples, and thus, this enhances the scattering of light. Absorption onsets from g-C₃N₄/MoS₂ nanomaterials are difficult to observe in the figure. To confirm the presence of these absorption onsets, we calculate the first derivative of the absorption with respect to the wavelength to determine the transitions present in our samples. This method was reviewed by Ojeda and Rojas [58] and implemented by Trenczek-Zajac and co-workers [59]. The derivate equals zero at the wavelength at which the maximum absorption occurs. In the figure, three transitions appear with wavelengths of 450 nm, 510 nm, and 620 nm, corresponding to the direct transition from the band gap of g-C₃N₄ and transitions from MoS₂, respectively. The same plot for pure CuI is also presented for comparison. No transitions appear for the same wavelength, thus confirming the composite’s presence on the matrix films’ surface. The plots are given as a Supplementary File (Figure S2).

Figure 10 shows the Tauc Plots of the samples. The plots are obtained with the following equation [36]:

$${\left(\alpha h\nu \right)}^n=A(h\nu -E_g)$$

(3)

where α is the absorption coefficient of the thin film, hυ is the photon energy, E_g is the optical band gap of the thin film, A is a constant, and the value of n is equal to 2 for a direct allowed transition [60]. To obtain the Tauc Plots, the plot of (αhv)ⁿ vs. photon energy is obtained. We extrapolate a tangent from the linear part of this plot into the photon energy axis. The calculated band gap for the pure CuI sample is 3.01 eV. The same value is obtained for samples CNM-1 and CNM-2. A slightly lower value of 2.99 eV is obtained for sample CNM-3, possibly due to a higher concentration of MoS₂ in the g-C₃N₄/MoS₂ composite. There is evidence that the addition of MoS₂ into carbon nitride allows the absorption range of the base material to be extended [61]; however, in our case, such band gap modification is negligible. The addition of the composite into the CuI film does not change the shape of the CuI film or its crystal size, so it is expected that the band gap of the film does not change.

3.5. Photoelectrochemical Characterization

Figure 11 shows the DC photoresponse and the DC I-V curves of samples CNM-1, CNM-2, and CNM-3 and pure CuI. The photoresponse was obtained using a 30 W blue LED with an applied bias of 1 V. In the I-V curves, the conductivity of the samples is increased from 21 (Ω·cm)⁻¹ to 58 (Ω·cm)⁻¹. The value of the conductivity of base CuI is in the range of a few tens to hundreds of (Ω·cm)⁻¹ [62]. This value can be increased to 200 by doping with Se [63]. Our method allows for an increase in conductivity without doping, which can lead to a detrimental effect when doping with high concentrations of atoms. The photoresponse of the films is evaluated at a 1 V bias voltage. It can be observed that upon illumination with the blue LED, the current increases from 0.9 μA to 1.2 μA. There is an evident increase in photocurrent when adding the composite material to CuI, to a highest value of 19.8 μA. Wang and co-workers demonstrated a p-CuI/n-Si heterojunction photodetector doped with Zn. Their results show that upon illumination with a UV light source of 325 nm, there is was increase in photocurrent due to the built-in heterojunction and rectifying behavior. However, some pin-holes were still present on the surface of the samples. Also, the single-step deposition of their CuI thin films allowed for faster crystallization, which led to inhomogeneities and hindered electrical performance [52]. Our work allows for a UV photodetector without the need for a Si substrate, which offers an evident advantage over what is normally reported in the literature, for example, a heterojunction based on TiO₂ and Si [54,64]. There is a considerable increase in current for sample CNM-3 due to the increase in the conductivity value, which saturates the photogenerated charge carriers. During our experiments, the conductivity of CuI was enhanced due to the presence of MoS₂ in the film. The composite allows for a heterojunction between g-C₃N₄ and MoS₂. When more MoS₂ is added to the thin film, charge extraction is enhanced due to the contact between CuI and MoS₂ [65]. Increasing the concentration of MoS₂ yields bigger crystals on top of the surface of CuI, which can contribute more charge carriers that can be extracted upon applying the bias. For comparison, samples were prepared with g-C₃N₄ and MoS₂. It is observed that adding the materials separately does not enhance the photoconductivity of the base CuI film, which demonstrates that the composite enhances electrical behavior. These results can be found in the Supplementary File (Figure S3).

Figure 12 shows the photoelectrochemical characterization of all the samples. Figure 12a presents the cyclic voltammogram of the CuI film in a 0.5 M KCl electrolyte, revealing four characteristic electrochemical features. Upon immersion in the aqueous electrolyte, the surface oxidation of the CuI film may occur, potentially leading to the formation of Cu(II) species. In the cathodic scan, two peaks are observed at approximately +0.2 V and −0.2 V vs. SCE. The peak at +0.2 V could correspond to the reduction of Cu(II) to Cu(I), suggesting a residual oxygen reduction or adsorption process [66]. The peak at −0.2 V is consistent with the reduction of Cu(I) to Cu(0). In the subsequent anodic scan, oxidation processes occur, including the oxidation of Cu(0) to Cu(I) and the oxidation of I⁻ to I² [67,68]. These processes result in quasi-reversible electrochemical behavior. The oxidation states observed in the CV are consistent with the results obtained from XPS analysis. Figure 12b, Figure 12c and Figure 12d correspond to the cyclic voltammogram of the CNM-1, CNM-2, and CNM-3 films, respectively. A higher current was obtained in CNM-1, CNM2, and CNM-3 than in CuI, suggesting an improved charge transfer process. In the anodic direction, peaks corresponding to Cu(0)/Cu(I) and I⁻/I² were observed, consistent with the cyclic voltammogram of pure CuI. Notably, a broad peak was observed at approximately −0.2 V/SCE in the CNM samples. To elucidate the involved process, cyclic voltammetry was carried out on CuI@g-C₃N₄ and CuI@MoS reference films (Figure S4a,b), revealing that the CuI profile is more pronounced in CuI@g-C₃N₄ than CuI@MoS₂. This possibly suggests a stronger interaction between CuI and g-C₃N₄ owing to the presence of triazine rings and π-π interactions from g-C₃N₄, which contribute to a redistribution of charge density at the Cu sites, requiring higher potentials for oxidation. Conversely, the Mo-S bond in CuI@MoS₂, as revealed by Mo 3d and S 2p spectra in XPS analysis, may facilitate charge transfer and lower oxidation potentials in CuI@MoS₂. Electrochemical impedance spectroscopy (EIS) was performed under dark and light conditions to investigate charge transfer resistance and separation. Nyquist plots, depicting the real impedance (Z′) on the x-axis and the imaginary impedance (Z″) on the y-axis, are presented in Figure 12e–h, as well as Figure S4c,d. The reference films, CuI@g-C₃N₄ and CuI@MoS₂ (Figure S4c,d), displayed resistive behavior. Notably, CuI@g-C₃N₄ exhibited an increased impedance at low frequencies under illumination compared to dark conditions. This suggests a reduced charge transfer rate, potentially due to enhanced electron–hole recombination before charge transport. In contrast, CuI@MoS₂, with its ~1.8 eV band gap and semi-metallic character, showed minimal changes between dark and light conditions, likely due to its high charge carrier density. The CuI and CNM-1 films (Figure 12e) displayed a Nyquist profile similar to g-C₃N₄, indicating rapid charge recombination. However, CNM-1 exhibited a lower intrinsic resistance, suggesting that the composite enhanced charge transport. Interestingly, CNM-2 films, composed of CuI@g-C₃N₄/MoS₂ with 3% MoS₂, demonstrated improved charge separation, as evidenced by a reduced arc under illumination, indicating less charge recombination. Conversely, CNM-3, with 5% MoS₂, showed no significant change in its Nyquist profile, potentially due to the increased MoS₂ content imparting more prominent semi-metallic characteristics.

4. Discussion

As already mentioned, adding g-C₃N₄/MoS₂ composites enhances the conductivity of CuI. Similar results can be obtained by doping the thin films with elements like Se, in which case only DC conductivity is increased to almost 200 (Ω·cm)⁻¹. In terms of the application of these films, they can be used as counter electrodes or transparent p-type semiconductors. In our case, the moderate conductivity values permit us to apply this material as a hole transport material in perovskite solar cells and as an efficient photocatalyst since the optical absorption is increased. Compared to the pure CuI thin films, which rely only on Cu vacancies for this purpose, the composite provides active sites for the photocatalytic process. Also, as the impedance spectroscopy measurements revealed, the heterojunction between g-C₃N₄/MoS₂ and CuI enhances electron transfer upon illumination. Photoconductivity can be modulated by varying the concentration of MoS₂.

5. Conclusions

In conclusion, CuI@g-C₃N₄/MoS₂ hybrid thin films have been synthesized for the first time. A combination of thermal evaporation and a simple chemical method was used for synthesis. The crystalline structure of the CuI thin film is not changed upon the addition of the composite. g-C₃N₄/MoS₂ is distributed on the surface of the films in the form of agglomerated particles, which allows for enhanced optical absorption. The material’s band gap is extended into the visible region when the concentration of MoS₂ in the composite colloidal suspension is modulated. Therefore, the electrical and electrochemical properties of the films can be easily modulated by varying the concentration of this material. The addition of the composite enhances electron transport and injection from CuI, which allows it to be applied as a photocatalyst and as a hole transport material in perovskite solar cells.