Surface Characterization of Zinc Selenide Thin Films Obtained by RF co-sputtering

Abstract

In this work, radio frequency magnetron sputtering was used to deposit zinc selenide thin films on p-type silicon (100) wafers and glass substrates in a high vacuum chamber. Several surface characterization instruments were implemented to study the thin films. X-ray photoelectron spectroscopy results revealed that oxidized Zn bound to Se (Zn-Se) at 1022.7 ± 0.1 eV becomes the dominant oxidized species when Se concentration exceeds 70%. Scanning electron microscopy coupled with energy dispersive spectroscopy showed that incorporating Se in Zn thin films will lead to formation of ZnSe grains on the surface. Contact angle measurements indicated that ZnSe-60 exhibited the lowest total surface free energy value of 24.94 mN/m. Lastly, ultraviolet-visible spectrophotometry and ultraviolet photoelectron spectroscopy data evinced that the energy band gap gradually increases with increasing Se concentration with ZnSe-70 having the highest work function value of 4.91 eV.

INTRODUCTION

Binary metal chalcogenide-based semiconductors, specifically from the group II-VI elements, have been stimulating profound interest from the materials science community. These set of materials exhibit unique properties such as having exceptional chemical stability, non-linear optical properties, and a wide band gap.¹⁻⁴ Owing to their favorable characteristics, II-VI semiconductors have been applied in the fields of solar cells, optoelectronic devices, IR detectors, in-vivo imaging agents, and anti-bacterial fabrics.⁵⁻¹⁰ Among these group of materials, zinc selenide (ZnSe) has been of particular interest due to its remarkable photosensitivity, a direct and wide band gap (2.67 eV), and optoelectronic tunability.^11,12 Thus, ZnSe has been implemented in applications such as photovoltaic devices,¹³ lasers,¹⁴ sensors,¹⁵ and light-emitting diodes;¹⁶ instilling its prospects as a versatile material to be widely utilized in the semiconductor industry.

Solid-state fabrication has become an integral component that has major influence over many different industrial sectors including semiconductor fabrication. To consolidate an effective solid-state fabrication process, utilizing materials with a high surface to volume ratio is of utmost importance.^12,17 One type of material that has high surface to volume ratio are thin films. Thin films have been fabricated through various means such as thermal evaporation,^1,17−18 vacuum evaporation,¹⁹ sol-gel method,²⁰ chemical vapor deposition,²¹ and radio frequency (RF) magnetron sputtering.^22,23 Out of these numerous methods, RF magnetron sputtering is one technique that can effectively fabricate ZnSe thin films as it carries many advantages such as facile composition tunability, strong adhesion, high deposition rate, and uniform deposition.^3,24 Additionally, RF magnetron sputtering can fabricate thin films regardless of the electronic properties of the target, which contrasts with direct current sputtering. Therefore, RF magnetron sputtering can be used to simultaneously co-sputter both zinc (Zn) and selenium (Se) to fabricate ZnSe thin films.

Despite ZnSe thin films being extensively studied for the aforementioned reasons, specific studies of RF magnetron sputtered ZnSe thin films are still rare. Moreover, detailed investigations on the chemical environment and how compositional ratios of Zn to Se affects the resulting ZnSe thin films are even more in rarity. In this study, we investigate ZnSe thin films with various Zn to Se ratios fabricated by RF magnetron sputtering. Numerous characterization tools, such as X-ray photoelectron spectroscopy (XPS), ultraviolet photoelectron spectroscopy (UPS), scanning electron microscopy (SEM) coupled with energy dispersive spectroscopy, contact angle measurement setups, and UV-vis spectroscopy, are employed to study the physical and chemical properties of the thin films.

EXPERIMENTAL

ZnSe thin films were fabricated by a RF magnetron cosputtering system housed in an ultra-high vacuum chamber (UHV). The temperature and the base pressure in the UHV system were maintained at 295 K and 2.7×10^-5 Pa by using an ion pump and a turbo molecular pump backed with a rotary vane pump. Thin films were deposited on either glass or p-type Si(100) wafer as a substrate. The substrate was cleaned with acetone and then dried with nitrogen gas before it was mounted in the UHV system. After the substrate was inserted in the UHV system, the substrate was rotated with a speed of 5 rpm. Pure Ar gas was injected as a sputter gas with a constant flow rate of 20 standard cubic centimeter per minute (sccm). Zn and Se targets (99.99%, Vacuum Thin Film Materials, Incheon, Korea) were used as the sputtering targets. At a working pressure of 5.33 Pa, the pre-sputtering process was carried out by applying a RF sputter power of 20 W on each target for 5 minutes. The shields between the targets and substrate were closed during this process.

The deposition rates were calculated with the measured thickness of thin films by using a surface profiler (Alphastep 500, Tencor, California, USA) to adjust the RF power on each target during the sputtering process. It was determined that the rate of the deposition rates increasing with increasing RF sputter power for Zn and Se was similar. It is worthy to note that the deposition rate of the insulating Se thin films was roughly 11 nm/min, which is slower than the conducting Zn thin films. Based on these results, we controlled the RF sputter power on each target to fabricate thin films with the following relative Se percentages in the thin films: 0, 20, 40, 60, 70, and 100%. By controlling the sputtering time, the thickness of the thin films was fixed at 100 nm. In this paper, the notation of the thin films will be designated as ZnSe-X, where X represents the relative ratio of Se.

To analyze the chemical environment and atomic ratio of the elements in the thin films, XPS (ESCALab MKII, VG, East Sussex, UK) with a Mg Ka X-ray source (1253.6 eV) was used. The survey XPS spectra were taken in constant analyzer energy mode with an energy step size of 0.5 eV and a pass energy of 50 eV. The high resolution XPS spectra for Zn 2p and Se 3d were taken by applying am energy step size of 0.05 eV and a pass energy of 20 eV. To obtain detailed chemical information, the high resolution XPS spectra with a FWHM of 2.0 ± 0.2 eV were deconvoluted by using the XPSPEAK 4.1 software in a 30/70 of Gaussian/Lorentzian ratio. And the background was subtracted using Shirley mode. The micrographs containing the morphology of the surface of the thin films were taken using SEM (MIRA3, TESCAN, Brno, Czech Republic). EDS (X-max^N 50, Oxford instruments, Abingdon, UK) was conducted to measure the compositional ratio of elements in smaller areas than XPS. Distilled water (DW) and ethylene glycol (EG) were used to measure the contact angle between these two liquids and the surface of the thin films from a homemade contact angle measurement setup. The roughness of the surface of the thin films were obtained by AFM (Dimension Icon, Bruker, Massachusetts, USA). The transmittance was measured at the wavelength range from 350 to 900 nm by using a UV-vis spectrophotometer (UV-2600, Shimadzu, Kyoto, Japan). The band gap energy was calculated from the UV-vis spectroscopy results. To align the energy level for the thin films, the work function (WF) and the valence band maximum (VBM) were calculated from the results obtained by UPS (ESCALab MKII, VG, East Sussex, UK), which was operated with a filament current of 46 mA, a filament voltage of 0.66 kV, and a He I source (21.2 eV).

RESULTS AND DISCUSSION

XPS spectroscopy was performed to examine the chemical environment and relative ratio of Zn to Se in the ZnSe thin films. Fig. 1 displays the XPS survey spectra of Zn, Se and ZnSe thin films. Based on Fig. 1, peaks corresponding to Zn 2p (located between binding energies:

Figure 1. XPS survey spectra of thin films with various relative atomic ratios for Se.

1014 and 1052 eV) and Se 3d (located between binding energies: 50 and 60 eV) can be observed. Therefore, we were able to confirm that Zn, Se, and ZnSe thin films were successfully deposited onto the substrates. In addition, the relative ratio values of Zn to Se with different chemical environments in each thin film are listed in Table 1. Interestingly, the peak corresponding to Se is weakly present in the pure Zn spectra. This might be attributed to the sublimation of the Se target in the UHV system. The peak corresponding to O 1s is also present in the survey spectra because surface oxidation of the thin films is inevitable when they are exposed to air. However, the O 1s peak was hardly detected in the pure Se thin film since Se does not oxidize readily.

Table 1. The relative atomic ratio of Zn to Se with different chemical environments in each thin film

The deconvoluted high resolution XPS spectra for Zn 2p_3/2, Se 3d, and O 1s are shown in Figs. 2(a), 2(b), and 2(c) respectively. In Fig. 2(a), Zn 2p_3/2 spectra were deconvoluted into three peaks: metallic Zn (Zn⁰) at 1021.9 ± 0.1 eV, oxidized Zn bound to oxygen (Zn-O) at 1022.9 ± 0.1 eV, and oxidized Zn bound to Se (Zn-Se) at 1022.7 ± 0.1 eV.²⁵⁻²⁷ The Se 3d peaks have minimal spin orbit splitting (0.86 eV), which results in Se 3d_3/2 and Se 3d_5/2 peaks overlapping and appearing as a singlet peak. In Fig. 2(b), the Se 3d spectra were deconvoluted into two peaks: neutral Se (Se⁰) at 54.8 ± 0.1 eV and reduced Se bound to Zn (Se-Zn) at 54.0 ± 0.1 eV.^28,29 In Fig. 2(c), O 1s spectra were deconvoluted into two peaks: reduced O bound to Zn (O-Zn) at 530.2 ± 0.1 eV and oxygen contaminants (surf.-O) at 531.1 ± 0.1 eV.³⁰ Table 1 lists the relative ratio values of Zn and Se species in different chemical environments. Table 1 lists the relative ratio values of Zn and Se species in different chemical environments. Based on Fig. 2. and Table 1, we can infer how the relative atomic ratios of Zn to Se affects the oxidation states of each element. In terms of Zn, it can be clearly seen that Zn was present as metallic Zn or ZnO from the high resolution XPS spectra of pure Zn thin films. As Se is incorporated in the Zn thin films, both metallic Zn and ZnO diminished and ZnSe gradually became the dominant species with increasing Se. Evidently in ZnSe-70, Zn only existed in the form bounded to Se. In the case of Se, Se⁰ or Se-Zn were the prevalent species. To be specific, when the relative ratio of Se in the ZnSe thin films exceeded 40%, the relative ratio of Zn bonded to O and neutral Se decreased while the amount of Zn bonded to Se increased. In terms of O, the peak corresponding to oxygen contaminants may be due to moisture in the atmosphere trapped on the surface of the thin films. This trend in O 1s peaks was similar to that of oxidized Zn combined with O in the Zn 2p_3/2 peak.

Figure 2. The deconvoluted XPS high resolution spectra for (a) Zn 2p_3/2, (b) Se 3d, and (c) O 1s of thin films with various relative atomic ratios.

The SEM micrographs pertaining to the surface of the ZnSe thin films and EDS data for ZnSe-40, ZnSe-60, and ZnSe-70 are shown in Fig. 3. The relative weight and atomic ratios of Zn to Se from both the basal surface and grain are listed in Table 2. The two monoatomic films: pure Zn and pure Se showed a relatively uniform and homogenous surface. For ZnSe-20, sharp grains of irregular morphology formed and were scattered on the surface. The other three thin films: ZnSe-40, ZnSe-60 and ZnSe-70 showed circular grains on the surface. These grains can be viewed as aggregates of ZnSe and the detailed formation mechanism has been previously reported by H. Gong et al.³¹ According to their studies, Zn acts as a seed and ZnSe grows around it, which continuously aggregates to form grains. This would explain the existence of the sparsely scattered small grains in ZnSe-40 and higher amounts of Se being detected in the grains than the basal surface. On the other hand, ZnSe-60 and ZnSe-70 contained more Se than ZnSe-40, which in turn correlates to lower amounts of Zn seeds being present in these two samples. Hence, larger grains form due to a higher amount of ZnSe aggregating around the lower amount of Zn seeds. In addition, grains form along the Zn seeds, which results in higher Zn concentration in the grains than the basal surface. It is also noteworthy that there are fewer grains in ZnSe-70 than ZnSe-60, which can be ascribed to the lower amount of Zn seeds.

Figure 3. SEM micrographs of (a) ZnSe-40, (b) ZnSe-60, and (c) ZnSe-70 thin films along with the EDS spectra; spectra 1, 3, 5 were obtained from the basal surface and spectra 2, 4, 6 were obtained from a grain on each thin film.

Table 2. The relative weight and atomic ratio of Zn to Se in the ZnSe-40, ZnSe-60, and ZnSe-70 thin films. Each ratio was measured at basal surfaces and grains

Fig. 4(a) summarizes the correlation between the measured contact angle of DW and EG on the thin films according to the Se concentration along with roughness of surface. For all samples, the measured contact angle with DW was larger than EG. Distinct grains were observed in ZnSe-40, ZnSe-60, and ZnSe-70, which resulted in the contact angle of DW on these films being less than 90°. On the contrary, densely packed particles on the surface of Zn, Se, and ZnSe-20 thin films were observed and the contact angle of DW on these films were above 90°. According to the lotus effect, the contact angle increases when the surface has a regular structure.³² Irregular grain protrusions formed on the surface of ZnSe-40, ZnSe-60, and ZnSe-70 improve the cohesion of DW droplets, resulting in a low contact angle. Furthermore, the contact angle of ZnSe-60 thin film was higher than ZnSe-40 and ZnSe-70 due to the presence of more grains on the surface. Since this grain affects the roughness value of the surface, the contact angle and roughness tend to be similar. The surface free energy (SFE) of the thin films were calculated using the Young’s equation:³³

Figure 4. (a) The measured roughness and contact angle of the thin films with DW and EG; inset images display a droplet of DW on the surface of ZnSe-20 and ZnSe-70 thin films. (b) The surface free energy of thin films calculated by contact angle measurements.

γ_L cosθ = γ_S - γ_SL ,       (1)

where γ_L is the surface tension of the used liquid, θ is the contact angle between the liquid and surface, γ_S is the SFE of the thin films, and γ_SL is the SFE between the thin films and liquid. The surface tension of DW at temperatures 20 ℃ to 95 ℃ was determined using the following equation:³⁴

\(\begin{aligned}\gamma_{\mathrm{L}}=235.8\left(\frac{374-\mathrm{T}}{647.15}\right)^{1.256}\left[1-0.625\left(\frac{374-\mathrm{T}}{647.15}\right)\right],\end{aligned}\)       (2)

where T is the DW temperature in degrees Celsius. The surface tension of EG at temperatures 20 ℃ to 170 ℃ was calculated using the following equation:³⁵

\(\begin{aligned}\gamma_{\mathrm{L}}=48.97-\frac{\mathrm{T}}{15}, \end{aligned}\)       (3)

where T is the EG temperature in degrees Celsius. According to Fowkes, the total SFE can be categorized into two components: the dispersive (non-polar) and non-dispersive (polar). Owens and Wendt modified the Fowkes hypothesis by incorporating the hydrogen bonding term. The modified equation is given as:³⁶

\(\begin{aligned}\gamma_{\mathrm{SL}}=\gamma_{\mathrm{S}}+\gamma_{\mathrm{L}}-2 \sqrt{\gamma_{\mathrm{S}}^{\mathrm{d}} \gamma_{\mathrm{L}}^{\mathrm{d}}}-2 \sqrt{\gamma_{\mathrm{S}}^{\mathrm{p}} \gamma_{\mathrm{L}}^{\mathrm{p}}}, \end{aligned}\)       (4)

where superscripts d and p denote the dispersive and polar components of the SFE. Merging equations (1) and (4), we obtain:

\(\begin{aligned}\gamma_{\mathrm{L}}(1+\cos \theta)=2 \sqrt{\gamma_{\mathrm{s}}^{\mathrm{d}} \gamma_{\mathrm{L}}^{\mathrm{d}}}+2 \sqrt{\gamma_{\mathrm{s}}^{\mathrm{p}} \gamma_{\mathrm{L}}^{\mathrm{p}}}, \end{aligned}\)       (5)

Utilizing equation (5), the dispersive, polar, and total SFE can be calculated, and the obtained values are displayed in Fig. 4(b) accordingly. The dispersive component had stronger influence over the total SFE for the two monoatomic and ZnSe-20 thin films, whereas the polar component had stronger influence for ZnSe-40, ZnSe-60, and ZnSe-70 thin films. In other words, thin films that contained grains: ZnSe-40, ZnSe-60, and ZnSe-70 were polar and hydrophilic with semi-wettability. Among the three, ZnSe-60 had the lowest calculated total SFE value of 24.94 mN/m.

Transmittance data for the thin films are displayed in Fig. 5(a). For the pristine Se, ZnSe-60 and ZnSe-70 thin films, the transmittance was higher than 60% in the long wavelength region of the visible light spectrum. In other words, the transmittance of ZnSe was higher than pure Se. On the contrary, thin films with higher Zn content: pure Zn, ZnSe-20, and ZnSe-40 exhibited very low transmittance. This observation indicates that lower content of Se results in lower transmittance, which might be attributed to the optically opaque nature of Zn metal. To further understand the optical properties, the absorption coefficient (α) was calculated using the following equations:³⁷

Figure 5. (a) The transmittance of thin films with various relative atomic ratios. (b) (αhν)² vs. hν plots.​​​​​​​

\(\begin{aligned}\alpha=\frac{1}{\mathrm{~d}} \ln \left(\frac{1}{\mathrm{t}}\right),\end{aligned}\)       (6)

where d and t represent the thickness and the transmittance of thin films, respectively. The correlation between the calculated absorption coefficient and the incident photon energy (hν) can be characterized using the Tauc plot, which is given as:³⁷

αhv = A(hv - E_g)ⁿ,       (7)

where A is a constant, E_g is the energy band gap of the thin films, and n is a constant that is determined by the type of interband transition. To be specific, n may pertain to the following values: 1/2, 3/2, 2, and 3; corresponding to direct allowed, direct forbidden, indirect allowed, and indirect forbidden transition, respectively. (αhν)² versus hν plots for direct allowed transition are shown in Fig. 5(b). The energy band gap was estimated by extrapolating it to the photon energy axis. In terms of our results, the energy band gap was determined to be 2.49 eV for the non-metallic amorphous phase Se thin film. The ZnSe-20 thin film, which has low Zn content, had a similar energy band gap (2.48 eV) to the Se thin film. Out of the fabricated thin films, ZnSe-60 had the highest energy band gap of 2.68 eV. In essence, the energy band gap decreased as the Zn content increased for thin films with higher Zn compositional ratios.

UPS data was used to calculate the VBM and WF. A representative UPS spectrum of ZnSe-60 is displayed in Fig. 6(a), which can be classified into two regions. Region 1 indicates the low kinetic energy region where the secondary cutoff can be determined and region 2 indicates the high kinetic energy region where the Fermi level can be determined. A least square extrapolation method was used to obtain the secondary cutoff and Fermi level.³⁸ Subsequently, the WF was determined using the following equation³⁹:

Figure 6. (a) The UPS spectra of ZnSe-60 thin film; Region 1 and 2 pertain to the secondary cutoff and Fermi level, respectively. (b) The aligned energy level with band gap and work function of thin films with various relative atomic ratios.​​​​​​​

WF = hν – (Fermi Level – Secondary Cutoff),       (8)

where hν is the radiation energy of the He I source (21.2 eV). Setting the Fermi level to be 0 eV, the VBM can be calculated from the tangent of the edge near the Fermi level.

The pristine Se thin exhibited the highest work function and this value decreased with increasing Zn concentration. This propensity implies that when the Zn content is higher than Se in the thin film, free electrons can be released with more ease. Fig. 6(b) presents the aligned energy level for the different thin films with respect to the vacuum level. EV denotes the vacuum level, CBM denotes the conduction band minimum, and EF denotes the Fermi level. The energy levels were determined from the VBM and WF values solved from the energy band gap acquired through the transmittance data and UPS results. The Fermi level was located closer to the VBM than the CBM at the energy level, providing evidence for ZnSe thin films exhibiting p-type semiconducting properties. The p-type behavior could be ascribed to the presence of oxygen, which acts as an impurity. As already confirmed by XPS results, the impurity O was present on the thin films and it has been previously reported that O can trap electrons and form an acceptor level.⁴⁰ Hence, the newly formed acceptor level was within the energy band gap and the Fermi level was located closer to the VBM, ultimately giving rise to p-type semiconducting behavior.

CONCLUSION

In conclusion, zinc selenide thin films were successfully deposited on p-type silicon (100) wafers and glass substrates using radio frequency magnetron co-sputtering in a high vacuum chamber. The deposited thin films were analyzed using the following characterization techniques: X-ray photoelectron spectroscopy, scanning electron microscopy coupled with energy dispersive spectroscopy, contact angle measurements, ultraviolet-visible spectrophotometry, and ultraviolet photoelectron spectroscopy. Our findings revealed that with increasing Se content: Zn-Se became the dominant oxidized Zn species, larger but fewer ZnSe grains formed at higher Se concentration, ZnSe thin films became more hydrophilic, and the band gap gradually increased.