Hamza Shaili*a,
Elmehdi Salmanib,
Mustapha Beraichc,
Abderrahim Elhata,
Mustapha Rouchdia,
Mhamed Taibid,
Hamid Ez-Zahraouyb,
Najem Hassanaina and
Ahmed Mzerda
aGroup of STCE-Energy Research Center (ERC), Faculty of Sciences, Mohammed V University, B. P. 1014, Rabat, Morocco. E-mail: hamza.shaili13@gmail.com; Tel: +21 2604710051
bLaboratory of Condensed Matter and Interdisciplinary Sciences Department of Physics, Faculty of Sciences, Mohammed V University, Rabat, Morocco
cLaboratory of Physics of Condensed Matter, Department of Physics, Ibn Tofail University, Kenitra, Morocco
dCSM, LPCMIO, Ecole Normale Supérieure, Mohammed V University, Rabat, Morocco
First published on 26th July 2021
Delafossite materials are considered to be a promising range of transparent conductive oxides for optoelectronic applications. The complications that have held back their implementation in practical devices lie in the complex growth methods that are required and in the formation of undesirable secondary phases. Herein, a fast, simple, and low-cost deposition method allowing the deposition of high-quality 2H-CuFeO2 nanostructured thin films is employed. The effect of Sr doping on the properties of spray-coated CuFeO2 thin film annealed at 850 °C is reported. X-ray diffraction (XRD) analysis revealed the delafossite structures of all the samples corresponding to the 2H-CuFeO2 phase. The lattice spacing decreased with increasing substitution of Sr at the Cu site. Raman analysis further authenticated the structural results collected via XRD analysis. Surface scanning using field-emission scanning electron microscopy revealed the formation of nanostructured CuFeO2 thin film possessing high crystalline quality, with the nanocrystal size increasing as the dopant content was increased. Energy-dispersive X-ray analysis allowed the quantification of the elements content via determining the ratios of the main elements as well as the dopant content in each sample. The optical properties of the samples showed strong light absorption in the visible region with a decrease in the band gap values with Sr insertion. First-principles calculations using density functional theory (DFT) were conducted to strengthen the experimental findings regarding the nature of the bonds in the hexagonal lattice of the CuFeO2 compound and the effect of Sr doping on its characteristics. The electrical properties measured at room temperature revealed p-type conductivity with tunable resistivity, while the samples displayed increased electron mobility as a function of the dopant content. Consequently, our work introduced an efficient and cost-effective synthesis route for the preparation of high-quality nanostructured 2H-CuFeO2 thin films, paving the way to facilitate further device applications.
One of the most promising classes of TCO candidates that has been studied and developed is the copper-based delafossite oxides CuMO2 (M = Fe, Cr, Al, Ga, Mn, Co…), which exhibit interesting optical and electrical properties compared to conventional TCOs. This interesting type of semiconducting oxides has been considered for several applications, including photocatalysis for solar energy conversion,5 luminescent materials,6 thermoelectrics,7 gas sensors8, batteries9, and photocathodes to produce hydrogen via water splitting.10 Herein, the emphasis will be on the CuFeO2 compound, which is composed of earth-abundant elements and has stable structural properties at room temperature and interesting electrical properties.7 CuFeO2 also offers good optical properties, a tunable band gap, and controllable conductivity, making it suitable for several applications such as anodes in lithium-ion batteries11 and in photovoltaics.12
According to the literature, CuFeO2 is a p-type semiconductor that crystallizes in two different structures, the rhombohedral (3R)-type13 and the hexagonal (2H)-type structure.14 The majority of papers that have reported the effective synthesis of CuFeO2, whether in nanoparticle or thin-film form, indicated a structure corresponding to the rhombohedral 3R-CuFeO2 phase or a mixture of the two phases. Only a limited number of papers have reported the successful synthesis of the 2H-CuFeO2 phase, and several difficulties were associated with the process.12,15 Furthermore, numerous attempts to engineer the physical and chemical properties of the CuFeO2 compound through a doping process using a variety of dopant elements have been conducted.16–18 Considerable attention has been paid to the use of alkali elements to adjust the characteristics of CuFeO2 for various fields of application. Anna Wuttig et al. reported Mg-doped CuFeO2 and tested its compatibility as a photocathode to optimize CO2 reduction.19 Qiang Gao et al. reported Ca-doped CuFeO2 and the modification of its initial properties.20 To the best of our knowledge, there are no reports describing tuning the properties of the material CuFeO2 using strontium either experimentally or through first-principles calculations.
3R-CuFeO2 thin films have been synthesized via several physical and chemical methods including pulsed laser deposition,21 radiofrequency magnetron sputtering,22 aerosol-assisted chemical vapor deposition,23 electrodeposition,24 spray pyrolysis,25 and spin-coating.26 In most previous attempts to prepare CuFeO2 in thin-film form, complex and expensive routes were utilized, which poses a major drawback for practical applications. So far, only a few attempts have been made to grow a CuFeO2 thin film corresponding to the hexagonal phase, especially through chemical deposition methods. Therefore, in the present paper, we aimed to develop a much simpler and faster chemical synthesis approach to obtain this material in thin-film form and to study its properties in detail. Another important objective of this work was to identify the impact of strontium doping on the characteristics of CuFeO2, with particular interest in the optical and electrical properties. To strengthen the quality of the present work, a theoretical part employing density functional theory (DFT) calculations was included to examine the effect of Sr doping on the electronic and optical properties of the pristine CuFeO2 lattice. This will allow a better understanding of the bonds formed between the atoms and their interactions, as well as the effects of the defects induced by doping on the electronic properties of the initial lattice.
Fig. 1 A schematic diagram of the ultrasonic spray deposition method and the annealing process in the furnace. |
(1) |
The calculated lattice parameters are illustrated in Fig. 3a. The values of the lattice constants (a and c) for the pristine sample are in agreement with previously reported results for the 2H-CuFeO2 structure.12,14 The lattice parameters show similar behavior in which both a and c decrease upon Sr addition. This decrease may be attributed to several factors, including the concentration of the dopant, defects (vacancies, interstitial, and dislocation), external strain developed due to the calcination temperature, and the difference between the ionic radii of Sr2+ (1.18 Å) and Cu+ (0.77 Å), which was observed earlier via the increasing intensities. The structure remains hexagonal and is not altered by the addition of different amounts of strontium into the CuFeO2 lattice.
Furthermore, the effect of strontium incorporation on the crystallite size of the nanostructured CuFeO2 thin film was investigated by employing the Debye–Scherrer formula:
(2) |
Fig. 3 Variations of the calculated CuFeO2 (a) lattice constants and (b) crystallite size and FWHM values as a function of the Sr content. |
Typical Raman analysis was performed at room temperature to further analyze the purity of the as-synthesized films, as well as to examine the homogeneity and crystallinity of the samples and understand the change in cation ordering. Raman analysis allows the determination of local structures and identification of changes in cation ordering of the samples. Fig. 4 presents the Raman modes obtained for the pristine and Sr-doped CuFeO2 samples. Three major vibrational modes were detected at 351, 375 and 690 cm−1, all of which were assigned to the 2H-CuFeO2 structure.21 The peaks at 351 and 690 cm−1 are assigned to Eg and A1g vibrational modes, respectively. The additional peak located at 495 cm−1 is generally attributed to non-zero wavevector phonons and is generated by the presence of crystalline defects.21 So far, only a few papers have reported the Raman analysis of CuFeO2 in the hexagonal structure. Y. Jin et al. reported the first Raman analysis of hexagonal CuFeO2 nanoplates and recorded almost the same major modes.12 It can be noted that after the first dopant fraction was introduced, a small decrease in the intensities of the bands was detected with no change in the modes as illustrated in Fig. S2.† The two main modes show a slight shift toward higher wavenumber in comparison to those in the undoped sample, indicating the effective insertion of Sr2+ cations in the place of Cu+, which consequently impacts the Cu–O bonds. The decrease in the intensities became more notable as the Sr content increased, but the modes were maintained. Indeed, the results suggest the high crystallinity of the deposited films with a consistent structure throughout the doping process. Hence, the Raman analysis further confirms the obtained phase and indicates the high quality of the prepared samples.
It is well known that morphological properties strongly impact the efficiency and construction of optoelectronic devices. Therefore, to be considered for further optoelectronic applications, a full and thorough assessment of the surface quality must take place. The morphology of the synthesized films was examined by means of field emission scanning electron microscopy (FE-SEM) and energy dispersive spectroscopy (EDX) to observe and evaluate the modifications induced by doping on the quality, grain size, and composition of the films, as shown in Fig. 5. It can be seen from Fig. 5a that the deposited samples exhibit a textured, uniform and homogeneous surface with an even distribution over the substrate. In addition, the micrographs in Fig. 5b–d show the formation of high-density quasi-hexagon-shaped grains with different sizes. An increase in the grain size was noted when the lower concentration of the dopant was introduced. At 5% Sr content, the grains became bigger and denser due to the expansion in the CuFeO2 lattice. This increase was observed earlier from the XRD data as a result of the difference in ionic radii, indicating agreement between the obtained results. It can be observed that the doping process enhances the quality of the obtained nanocrystals, which can have an immense effect on the transport properties of the film, particularly in terms of increasing its carrier mobility.
Moreover, EDX analysis was conducted to inspect the elemental composition of the prepared films, as presented in Fig. 6. The data illustrate the obtained percentages, with the presence of sufficient quantities of the main elements (Cu, Fe, and O). The ratios of the elements indicate the high purity and near-stoichiometric nature of the samples with rich oxygen content, as illustrated in Table S2.† In addition, the Si signal is detected from the quartz substrate as a result of the difference between the thickness of the film and the high EDX interaction volume and penetration depth of the electron beam. This may slightly reduce the accuracy of the quantitative analysis. The presence of the dopant Sr was confirmed in the doped films in varying quantities, depending on the percentage added. It can be noted that the ratios became slightly imbalanced as the doping element was introduced. Additionally, elemental atomic mapping was included to visualize the distribution of the elements on the resulting surfaces, allowing the quantification of each component with high precision. The mapping images in Fig. 6 once again demonstrate the presence of the main elements with a near-stoichiometric distribution. In addition, the images exhibit a uniform distribution of the elements with a higher concentration for the oxygen element and a small portion of the Sr element.
The UV-vis spectra of the pure and Sr-doped CuFeO2 thin films are presented in Fig. 7a. All the films exhibit strong absorption behavior in the UV light region (200–400 nm), with a notable increase in the visible region, which continues to increase exponentially as the wavelength increases into the near-infrared (NIR) region. It can be noted that the absorption edge of the doped films is shifted toward higher wavelength, especially for the 5% Sr-doped CuFeO2 film. Furthermore, when the dopant fractions were introduced, the absorption increased slightly, demonstrating that strontium increased the absorption of the metal oxide, as has been reported previously.33 To summarize, strontium doping effectively enhanced the quality of the films, particularly for the 3 at% Sr film, which directly impacted their photon absorption ability. The band gaps of the pristine and Sr-doped CuFeO2 were calculated using the Tauc plot according to the equation:
(3) |
Fig. 7 (a) Recorded optical absorbance as a function of wavelength. (b) Plots of (αhν)1/2 versus hν and (c) the band gap variations of pristine and Sr-doped CuFeO2 nanostructured thin films. |
Ab initio calculations were performed to uncover the nature of the bonds formed between the different elements and to investigate the electronic properties and the effect of strontium incorporation on the pristine CuFeO2 lattice. In order to conduct a comparative study, the doping concentrations selected for DFT calculations were the same as those used in the experiment. Fig. 8 shows the unit cell of the pristine CuFeO2 compound with the positioning of each atom. In general, an idealized solid exhibits perfect periodicity, which allows the theoretical calculations to be restricted to just one unit cell as long as the bulk properties are considered, which is validated in the calculation method employed in the present work. To the best of our knowledge, this is the first report of DFT calculations on the properties of CuFeO2 in the hexagonal structure. Doping a system with non-magnetic impurities commonly leads to a non-magnetic system. Therefore, in this work, we chose to perform doping with a non-magnetic element (Sr2+ ions) to avoid any augmentation of the magnetic behavior of the system and improve the properties of the material under investigation.
Fig. 8 The crystal structure of CuFeO2 crystallizing in the hexagonal structure. Color code: Cu = blue, Fe = brown, O = red. |
Fig. 9 presents the spin-up and spin-down band structures calculated for the pristine CuFeO2 using both the GGA and GGA-SIC approximations. Defect-free CuFeO2 is a semiconductor with the valence-band maximum (VBM) and the conduction-band minimum (CBM) positioned at the Γ(0,0,0) point at the center of the Brillouin zone, indicating a direct-band-gap semiconductor nature. The band gaps were calculated to be 1.08 and 1.67 eV using the GGA and GGA-SIC approximations, respectively. The obtained band gap for GGA-SIC is in close agreement with previous experimental values.12 Furthermore, semi-conducting behavior can clearly be seen for the spin-up band structure, while the spin-down band structure is metallic in both approximations. A comparison of the calculated value with the band gap (1.75 eV) observed experimentally using the UV-vis spectrometer indicates the presence of point defects.
Fig. 9 The band structure of 2H-CuFeO2 using the GGA-PBE ((a) spin-up and (b) spin-down) and GGA-SIC approximations ((c) spin-up and (d) spin-down). |
Usually, the regular GGA-PBE tends to underestimate the actual properties of the material under investigation, meaning that the exchange and the correlation effects in transition metal oxides are not considered. This leads to important self-interaction errors32. This can be clearly observed in the density of states calculated using the GGA-PBE approximation, while the GGA-SIC provided a value that was relatively closer to that collected from experimental measurements. This band gap increase is due to the delocalization of electrons in such materials and nonzero self-interaction errors of DFT. When a semiconductor is doped, the Fermi level tends to shift towards higher energy values, which can be explained by the fact that the doping process enhances the stability of the compound corresponding to the low energies. Fig. 10 displays the total and partial DOS (spin-up and -down) of the pristine and strontium-doped CuFeO2 using the GGA and GGA-SIC approximations. It can clearly be noted from Fig. 10a that for the GGA-PBE approximation, the spin-up and spin-down channels do not overlap, indicating the half-metallic behavior of the pristine structure. For that reason, the self-interaction corrected (SIC) was implemented to reduce this behavior to the greatest extent possible. The DOS in Fig. 10b also reveals that CuFeO2 acts as a half-metal, and the same behavior was noted for the spin-up channel. However, for the spin-down channel, the metallic behavior is greatly reduced due to the fact that the Fe band gap opened up with SIC correction. The conduction band minima (CB) of CuFeO2 are composed essentially of Fe-3d states with some overlap between Cu-3d and O-2p, while the valence band maxima (VB) are largely dominated by the Cu-3d states with a minor contribution from O-2p states. However, the contribution of the Cu-d states becomes prominent in the −0.5 to −0.3 eV zone. An important contribution was observed from the Sr-3d states as the dopant was induced, with increased domination of the Fe-3d states for both of the doping percentages. The band gap became narrower relative to that of pristine CuFeO2 and settled at 1.22 eV for 5 at% strontium content. In summary, first-principles calculations using DFT permitted an in-depth investigation of the properties of the hexagonal CuFeO2, as well as of the effect of Sr2+ incorporation in tuning the initial properties; the results were consistent with the experimental findings.
Fig. 10 Calculated partial and total DOS for pure and Sr-doped CuFeO2: (a) 0% with GGA, (b) 0% with GGA-SIC, (c) 3% Sr content, and (d) 5% Sr content. |
Evaluation of the potentials of certain materials for optoelectronic applications relies heavily on their electrical properties. Thus, it is essential to determine the electrical properties of the prepared samples and the changes resulting from the alkali-element doping of the CuFeO2 films in order to enhance their properties and thereby evaluate their potential for further applications. For this purpose, the electrical properties of pure and Sr-doped CuFeO2 thin films were measured at room temperature to determine their semiconducting nature and to investigate the tunability of the conductivity and carrier mobility due to the strontium doping of the pristine CuFeO2 nanostructured films. Fig. 11 shows the electrical properties measured using the van der Pauw configuration and their variation as a function of the dopant content. The Hall coefficient (RH) measurements indicated the p-type nature of the CuFeO2 compound, as reported previously.20,22 The resistivity of the undoped CuFeO2 thin film was measured to be 4.95 Ω cm. An intense decrease in the electrical resistivity was observed at the lower doping percentage, manifesting the effective capability of Sr-doping to increase the conductivity of metal oxides. The resistivity value further decreased to settle at 3.5 Ω cm for 5 at% Sr content. The same tendency was observed for the carrier concentration, which decreased greatly due to the doping. Furthermore, the increased carrier mobility due to doping was suggested above from the SEM images, which showed an augmentation in grain size leading to an enhancement of the carrier mobility from 7.9 cm2 V−1 s−1 for the undoped sample to 12.1 cm2 V−1 s−1 for the 5 at% Sr sample. Hence, the prepared CuFeO2 thin film possessed superior electrical properties compared to other delafossite materials, with these enhanced characteristics being generated by the doping process.
Fig. 11 Resistivity, carrier concentration, and mobility of pure and Sr-doped CuFeO2 films as a function of the Sr content, measured at room temperature. |
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1ra03734j |
This journal is © The Royal Society of Chemistry 2021 |