Accepted: May 27, 2022 | Published Online: May 29, 2022

Citation: Bayahia D, Shirbeeny W, Madkhli AY (2022) Synthesis and Characterization of Sm and Ti-doped CuO Thin Films. Int J Opt Photonic Eng 7:048.

Copyright: © 2022 Bayahia D, et al. This This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Abstract

Thin films of CuO, Ti- and Sm-doped CuO were deposited on silicon substrates using the DC and RF for metal and metal-oxide deposition, respectively. Intentionally relatively high deposition power was applied to the metal (Ti or Sm). This guarantees a wider bandgap of CuO for possible high emission intensity by Ti or rare Earths. First, the crystal structures of the thinfilms were investigated via the X-ray diffraction method. The variation in crystal lattice dimensions responded to the Sm and Ti hosting, indicated by the redshift in the diffraction angle. Next, the optical properties of the thin films were studied by Spectroscopic Ellipsometry (SE). The effective medium approximation (EMA) method and the Tauc-Lorentz oscillator model fit ellipsometric parameters (Ψ and Δ). EMA model enabled the extraction of the optical parameters of the dopant metal and the metal-doped CuO. As a result, the doped metal mass thickness could be determined. The ellipsometry also helped estimate the thin film thickness and bandgap energy. Moreover, photoluminescence spectra verified the bandgap structure of the thin films. Furthermore, the PL results showed that the inclusion of Sm in CuO modified the host material's bandgap energy and elevated the CuO's emission intensity compared to Ti-doped CuO. Due to its high emission intensity and low absorption coefficient, Sm-doped CuO could be elected in LED and Laser diode potential applications.

Keywords

EMA, Sputtering, emission spectrum, photoluminescence

Introduction

Cupric oxide (CuO) is a p-type semiconductor with essential beneficial properties like low fabrication cost, non-toxicity, and richness of the crucial materials [1]. These advantages have triggered its usage for several device applications [2-9]. Properties of CuO are significantly controlled by its size, morphology, and bandgap of the thin films [10], which are considerably dependent on the deposition processes. Various reports of different CuO nanostructures synthesis have been revealed [4-6]. Metal-doped CuO has been studied due to its narrow bandgap. Then, Undoped and Yb-doped CuO thin films with various dopant concentrations and their performances in heterojunction solar cells were investigated [11]. Moreover, S-doped CuO nanoclusters were prepared to convert CO2 into valuable chemicals. Promptly, the catalytic activity and stability of CO2 have been considerably improved after sulfur substitution in CuO [12].

On the other hand, catalytic degradation of rhodamine B was examined by shape control of the Sn-doped CuO nanoparticles [13]. Photoelectrochemical properties of Ni-doped CuO nanorods were investigated, then the maximum photocurrent density was studied vs. the dopant concentration. Finally, a route for improving the photocurrent density was proposed [14]. Zr-doped copper oxide nanoparticles were positively fabricated via the Pechini method; their physical and antibacterial properties were inspected [15].

Though Samarium-doped CuO has rarely been studied [16], this work explores the optical properties of the Ti and Sm-doped CuO thin films deposited by DC and radio frequency magnetron sputtering. The dopants were successfully incorporated in CuO. The dopant's percentage was estimated using the effective medium approximation method [17,18]. Moreover, the variable angle spectroscopic ellipsometry (VASE) was employed to extract the refractive index, the extinction coefficient, the thin film thickness, roughness, and the bandgap energy [17,19-22].

Experimental

Thin-film synthesization

The CuO, Ti, and Sm targets were obtained from VIN KAROLA INSTRUMENTS with a purity of 99.99%, 3-inch diameter, and 0.125-inch thickness. The deposition was carried out on the glass and silicon substrates; the glass substrate is required for the XRD and transmission study. The silicon was chosen for the ellipsometry measurements and AFM investigation. The deposition process was conducted to simultaneously deposit Ti (or Sm) and CuO on the glass and silicon substrates. The CuO target was fixed to the radiofrequency (RF) probe. The Ti (or Sm) target was attached to the DC side in the sputtering chamber. The silicon and glass substrates were set on a rotating holder 10 cm from the target, formerly adjusted to a 40 turn/min rate. The starting presser of the chamber was set to reach 1.5 × 10-6 mbar by employing UNIVEX 350 deposition and Leybold Vacuum system. Ar gas was then allowed to flow at a constant rate of 90 sccm. An applied power of 60W was used on the DC probe, on which the dopant metal target (Ti or Sm) was attached, and 120 W power was applied to the CuO target. The duration of the deposition was fixed at 10 minutes. The undoped CuO thin film was deposited on silicon and glass substrates by applying the same conditions but applying no power to the metallic target (Ti or Sm) and keeping it covered.

Thin-film characterization

The crystal structure and lattice parameters of the CuO and metal-doped CuO thin films were investigated using a high-resolution X-ray diffractometer (Ultima-IV; Rigaku, Japan), Cu K-line of wavelength 1.54060 Å, and operating conditions 40 kV/20 mA. The surface roughness and thin film thickness were studied by AFM imaging. Optical characteristics and bandgap of the metal-doped CuO thin films were studied by employing M-2000 Variable Angle Spectroscopic Ellipsometry (VASE) by J. A. Woollam Co., backed by Complete EASE software. The dopant content mass thickness was estimated using the effective medium approximation (EMA).

The effective complex dielectric function (pseudo-dielectric function) $\tilde{\epsilon }$ is given by

$$\tilde{\epsilon }=f_A{\tilde{\epsilon }}_A+f_B{\tilde{\epsilon }}_B+f_C{\tilde{\epsilon }}_C+\mathrm{...}\text{ (1)}$$

Where fA, fB, and fC and are the volume fractions of the constituent materials with dielectric functions ${\tilde{\epsilon }}_A$, ${\tilde{\epsilon }}_B$ and ${\tilde{\epsilon }}_C$. The sum of the volume fractions of all constituents must equal 1.

A more accurate model is the Maxwell-Garnett EMA [23], which is derived assuming spherical inclusions of materials B and C in a host matrix of material A, and for three constituents. It is written as

$$\frac{\tilde{\epsilon }-{\tilde{\epsilon }}_A}{\tilde{\epsilon }+2{\tilde{\epsilon }}_A}=f_B\frac{{\tilde{\epsilon }}_B-{\tilde{\epsilon }}_A}{{\tilde{\epsilon }}_B+2{\tilde{\epsilon }}_B}+f_C\frac{{\tilde{\epsilon }}_C-{\tilde{\epsilon }}_A}{{\tilde{\epsilon }}_C+2{\tilde{\epsilon }}_A}\text{ (2)}$$

an extension of the model (for two constituents only) using a parameter called a depolarization factorp. The depolarization factor weights the contribution of material B. A depolarization factor of 1/3 assumes that material B is spherical; p of 1 represents flat disks or a laminar microstructure, while a p of 0 assumes material B is composed of needle-like or columnar inclusions.

For two inclusions, CuO and Sm, (or Ti), f_C = 0, and Eq. (2) becomes

$$\frac{\epsilon -{\epsilon }_{Sm}}{p\epsilon +\left(1-p\right){\epsilon }_{Sm}}=f_{Sm}\frac{{\epsilon }_{CuO}-{\epsilon }_{Sm}}{p{\epsilon }_{CuO}+\left(1-p\right){\epsilon }_{Sm}}\text{ (3)}$$

and f_sm is the volume fraction of the Sm (or Ti). The mass thickness MT of the dopant is calculated using:

$$M{T}_{sm}=f_{Sm}xthinfilmthickness\text{ (4)}$$

The ellipsometry measurements revealed two quantities, namely Ψ and Δ. The ellipsometry process is illustrated in Figure 1, and the ellipsometry parameters Ψ and Δ are defined as follows

$$\rho \text{ = }\frac{r_p}{r_s}\text{ = }tan\left(\Psi \right)e^{i\Delta }$$

Where, rp and rs are the reflection amplitudes of the p and s polarized lights.

The Complete EASE software uses various oscillator models. However, the Tauc-Lorentz model was the best fit Ψ and Δ for undoped CuO. Then, the optical parameters and the thin film thickness were extracted. To obtain the optical parameters of the doped CuO and the incorporated metal, Ψ and Δ were measured for Ti- and Sm-doped CuO thin films. In this step, the CuO was considered the first material, while the Ti (or Sm) was set as the second material. The parameters of the Tauc-Lorentz model predetermined in the case of undoped CuO were used in this case. Still, the model was converted into EMA to allow the insertion of the metal (2^nd material) parameters.

Results and Discussion

XRD investigation

A comparative study of the XRD spectra of Sm- and Ti-doped CuO thin films with dopants sputtered at 60 W is shown in Figure 2. Additionally, the XRD data of as-deposited CuO is also included. According to the XRD database card number (01-072-0629), the crystal unit cell is monoclinic. The crystallite size D and the unit cell dimensions of the undoped CuO, Ti-, and Sm-doped CuO thin films are listed in Table 1, Table 2, and Table 3, respectively. Doping CuO by Sm or Ti effectively influenced the crystal lattice parameters and grain size; this is ascribed to the variation in ionic radii of Sm (0.958 pm) and Ti (0.860 pm) ions compared to the Cuionic radius (0.73 pm) [24]. In addition, a redshift (a shift toward a larger diffraction angle) was noticed as Ti and Sm were doped in the CuO. This shift was attributed to the occupation of Ti or Sm interstitial sites in the CuO lattice. Accordingly, the Ti or Sm may occupy energy levels in the conduction band leading to the enlargement of the energy bandgap of the metal-doped CuO. This behavior also influenced the grain size of the Ti-doped and Sm-doped CuO compared to undoped CuO.

Surface morphology characterization

The surface morphologies of the undoped CuO, Ti-, and Sm-doped CuO thin films are shown in Figure 3a, Figure 3b, and Figure 3c, respectively. It is seen in Figure 3c that there are more surface granules than in Figure 3b. Additionally, the granules of the undoped CuO are of the smallest size among the doped CuO thin films, which may be correlated to the grains size determined by XRD study. In Table 4, the mass thickness extracted from EMA model (next section) of the Sm is considerably greater than that of the Ti. Since its melting point is significantly lower than Ti's, which is expected to have more Sm deposited on the substrate than Ti, this behavior is consistent with the grain size correlation of the Sm and Ti-doped CuO. Additionally, since the same sputtering power was applied to CuO in all experiments, the thickness of the CuO conserved its value but the Ti- and Sm-doped CuO thin films thicknesses increased by the addition of the Ti or Sm, as concluded from Table 4. The surface of the metal-doped CuO thin film was about 2 nm roughness; however, in the undoped CuO thin film, the roughness was about 5 nm. Therefore, the granules distribution over the surfaces was almost uniform.

Optical characterization of thin films

Using the Complete EASE - J.A. Woollam, the spectroscopic ellipsometry parameters Ψ(λ) and Δ(λ) were fitted to the Tauc-Lorentz oscillator model by considering the CuO as a host medium and Sm (or Ti) as the guest element. In this step, the thin films deposited on silicon substrates were used. Using this routine, we could extract the metal-doped CuO's optical parameters and optical parameters of the Ti (or Sm) hosted by CuO. The estimation of the optical parameters was done by obeying the following steps: 1st plotting the undoped CuO ellipsometry parameters Ψ(λ) and Δ(λ) and fitting them to a suitable oscillator model, Tauc-Lorentz in this work. 2^nd plotting the Sm- or Ti-doped CuO ellipsometry parameters Ψ(λ) and Δ(λ) and carrying fitting process using the Tauc-Lorentz model for the CuO and EMA for the dopant (Ti or Sm). The roughness, the total thin film thickness, and the volume fraction f_Ti (or f_Sm) of the dopantcould be estimated. The mass thickness of the dopant was calculated using equation (4). The influence of Sm and Ti doping on the CuO bandgap is shown in Figure 4. The bandgap energies of the undoped CuO, Ti-doped CuO, and Sm-doped CuO were extracted from the Tauc plot and had the values 2.89 eV, 2.92 eV, 3.08 eV, respectively. The bandgap broadening in the case of Ti or Sm may be attributed to the formation of additional energy levels of the dopant in the conduction band. Doping CuO with Ti or Sm influenced the thin film total thickness; however, the doped ratio varied according to the element, despite the same applied power.

PL characterization

The emission spectra of undoped CuO, Ti-doped CuO, and Sm-doped CuO are displayed in Figure 5. The excitation wavelength was set at 363 nm and was included in the emission spectra. Five emission lines were uncovered after performing deconvolution. The emission lines are denoted by λ_n, n = 2 to 6 and are listed in Table 5. The photoluminescence emission mainly succeeded in interpreting the existence of defects like vacancies or impurities of materials. As seen in Figure 6, vacancies may belong to copper or oxygen for the undoped CuO thin film, while the impurities may be interstitial copper or anti-oxygen sites O_Cu [25]. These defects stimulate the formation of new energy levels in the bandgap, and as an effect, emissions will occur from these trapped levels during the excitation of the sample. Emission occurs due to the exciton's radiative formation, and the emission bands are generally caused by deep level or trap site emissions due to oxygen vacancies [25]. Specifically, for the undoped CuO Figure 6, the emissions at 380.3 nm and 438.8 nm belong to Cu2+ transitions [26]. The lines 410.4 nm and 465.6 nm are O2- vacancies [27], while the emission at 488.5 nm is trap center. The deconvoluted emission lines of Sm-doped CuO are shown in Figure 7. The emission lines corresponding to C²⁺ are 380.3 nm and 451.4 nm. It is worthy to mention here that the 1^st Cu²⁺ emission line did not affect by the Sm doping. However, the line 451.4 was redshifted due to Sm inclusion. This is attributed to the formation of energy level in the bandgap. In addition, oxygen vacancies appeared at 407.4 nm and 470.0 nm, where they shifted their positions due to the slight stress added by the inclusion of Sm²⁺. The trap level, however, appeared at 491.9 nm. All shifts occurred in the case Sm-doped CuO emissions were produced due to the large ionic radius of Sm²⁺ compared to the Cu2+. The Ti-doped CuO photoluminescence emission lines are depicted in Figure 8. It is well noticed in Table 5 that there were no considerable shifts due to the inclusion of Ti in CuO, this was attributed to the fact that the ionic radii of the Ti and Cu are close to each other, therefore no observable stress occurred in the lattice. The insertion of Ti or Sm in the CuO formed additional energy levels in the conduction band leading to the Moss-Burstein shift [28]. The strongest emission lines of Ti²⁺ and Sm²⁺ lie beyond the excitation energy of the instrument, so the PL spectra in Figure 8 did not show these emission lines. The Ti²⁺ emission line is 251.6 nm (4.9 eV) and is produced by the transition 3D3 to 3F°4.

Conclusion

In conclusion, the potential of Sm or Ti to serve as useful dopants in CuO thin films was investigated. It was shown that the incorporation of Ti or Sm in CuO influenced the structure of the CuO thin film. However, the optical and emission properties of the CuO were significantly affected by the Sm inclusion. The Sm-doped CuO showed high photoluminescence intensity along the visible region. Additionally, Sm-doped CuO thin film revealed broader optical bandgap energy and low absorption. This work encourages the incorporation of Sm in CuO as a possible solution for CuO usage a LED and diode laser applications.

---