Introduction

Pure ZnO is a diamagnetic-semiconductor wide bandgap (∼3.37 eV) material with high exciton binding energy (≈60 meV), enables efficient excitonic emission at room temperature that makes it most suitable material for optoelectronic applications1,2,3. ZnO can be prepared in different forms, viz. single crystals, epitaxial films, along with various nanostructures, to meet different applications4. When ZnO is substituted with transition metals (TM: Fe, Co, Ni, etc.)5,6,7 and/or rare earth elements (RE: Er, Eu, Yb, etc.)8,9,10,11, then ZnO can act as a ‘diluted magnetic semiconductor’ (DMS) materials. It is noted that the 4f- RE-metals substituted ZnO have higher magnetic moments, compared with 3d TMs. The electrons could facilitate the ferromagnetic (FM) interaction among the RE ions as a result of the interaction between f electrons and host s electrons. Therefore, comprehending the electronic structure and magnetic characteristics of RE-substituted ZnO is crucial because of its prospective use in electronic/photo-electronic and spintronic devices. Especially, ZnO-substituted with RE erbium (ZnO: Er) can exhibit FM behaviors. The FM behaviours of ZnO: Er are often attributed to defects, like oxygen vacancies within the ZnO lattice, which can interact with the Er+-ions and facilitate spin alignment. Some studies report the observation of FM behaviors in ZnO: Er at room temperature. However, the exact magnetic mechanism and extent of this FM can depend on the synthesis methods, temperature, and Er+-concentration. In case of thin film, the magnetic behaviors are dependent not only on these parameters but also on the nature of the substrate used for the deposition of ZnO: Er thin films, as the substrate’s properties can affect the film’s structure, strain, defects, and ultimately, its magnetic properties. The substrate can also influence the magnetic anisotropy of the film, making it easier for the magnetization to align in a particular direction. In ZnO: Er thin films, the Er-dopant introduces magnetic moments that can interact with the surrounding ZnO matrix. The substrate can influence the interaction between Er-ions and the ZnO matrix, as whole affecting the magnetic properties of the film. The study of magnetic properties in ZnO: Er is an active research field for potential applications in electronic and spintronics due to its ability to manipulate both the charge and spin of electrons. The magnetic moment arises in Er-substituted ZnO from the localized unpaired 4f electrons of Er+-ions substituting Zn-sites in the lattice structure of ZnO. These Er+-ions are localized after substitution and can align with an applied magnetic field. However, the exact mechanism behind the ferromagnetism in Er-substituted ZnO is still under investigation, and achieving stable ferromagnetism at room temperature remains challenging. Among the various deposition techniques, sputtering is a potential technique, which can form thin film with strong adhesion layer12. The quality of thin films is affected by a number of sputtering parameters, including sputtering power, substrate temperature, chamber pressure, and sputtering time13. Because the nature of the substrate has a significant impact on the properties of thin films, selecting the appropriate substrate is crucial in many applications because the quality of the thin film nucleation/growth behavior depends on the substrate’s orientation, roughness, cleanliness, and temperature12,13,14,15,16,17,18. The thermal expansion coefficient mismatch between a ZnO thin film and its substrate significantly impacts the thin film’s quality14. Therefore, it affects the film’s structural, optical, electrical and magnetic properties.

In this present study, we have grown RE element Er-substituted polycrystalline-ZnO (ZnO: Er) thin films on two different substrates, viz. sapphire and silicon [(ZnO: Er)sapphire/silicon], using the DC sputtering technique. Sapphire and silicon substrates were chosen to see the crystal strain effect on deposited materials as there is a large lattice mismatch between ZnO and (ZnO: Er)sapphire/silicon. These choice of substrates are also used to study the grown of the thin films on a particular orientation. Therefore, the substrate effects of sapphire and silicon on the structural, electronic/atomic, and magnetic properties of RE element Er-substituted polycrystalline-ZnO [(ZnO: Er)sapphire/silicon], thin films have been studied. We have discussed several effects associated with the ZnO materials for the development of ZnO-based dilute magnetic semiconductors (DMSs) and their device-based applications. A comparative electronic/atomic/bonding-structure of (ZnO: Er)sapphire/silicon thin films has been studied using X-ray absorption near edge structure (XANES) and X-ray photoelectron spectroscopy (XPS) measurements, whereas magnetic behaviors are studied using SQUID-measured magnetic hysteresis M-H loops. XANES spectroscopy is an ideal technique because it is an element-specific probe that is highly sensitive to the local structure around the absorbing crystalline and non-crystalline atoms19,20. In addition, crystalline surface morphology is studied using scanning electron microscopy (SEM) images. The evolution of local electronic structures is proposed, where the Er in (ZnO: Er)sapphire/silicon thin film could behave as DMSs and could be used in electronic/spintronics applications.

Results and discussion

Figure 1a,b shows the SEM micrograph of (ZnO: Er)sapphire/silicon thin films. Differences in micrograph size/textures are due to the different substrate’s mismatch in lattice, stress, Er-content present (at%) in the ZnO matrix, and the thermal expansion coefficient between the ZnO: Er thin film and the substrates (sapphire/silicon). From the inset Fig. 1b, a branch of leaf-shaped cone arm structure with non-uniform diameter along its axes is observed in the (ZnO: Er)silicon thin film matrix, while no such structure is observed in (ZnO: Er)sapphire. Zhuo et al.21. also observed similar four legs regular-shaped hexagonal cone arm and defined as the tetrahedral structure in nanostructures-ZnO materials. The XRD spectra of (ZnO: Er)sapphire/silicon is shown in Fig. 1c,d. Different (ZnO: Er)sapphire/silicon diffraction peaks and their orientations are identified/indicated in the XRD spectra22,23,24,25,26, as shown in Fig. 1c,d. It is observed that both the films (ZnO: Er)sapphire/silicon are partially oriented epitaxial polycrystalline hexagonal wurtzite structure, exhibiting (002), (111), (102), (110), (103), (004) orientations. Moreover, the XRD patterns of (ZnO: Er)sapphire is grown along (111) plane; whereas (ZnO: Er)silicon is grown along (002) and (111) planes. In addition, some other additional minor peaks of ZnO and the substrate (sapphire/silicon)-peaks are also observed in the respective (ZnO: Er)sapphire/silicon thin films. The (ZnO: Er)sapphire thin films provide better crystalline quality and surface morphology compared to (ZnO: Er)silicon thin films. The silicon-substrate is a cost-effective and readily available option, especially for certain applications where UV performance is not the primary concern; whereas sapphire has a wurtzite structure, which aligns well with the wurtzite structure of ZnO film that facilitates better growth with smoother surfaces, leading better crystalline quality with reduced threading dislocations, improved crystalline orientation and larger grain sizes compared to those grown on silicon. In this context the lattice constants, crystallite size (from Scherrer’s formula) and the strains are obtained from XRD analysis and tabulated in Table 1 to support crystalline quality of the (ZnO: Er)sapphire/silicon thin films. It is clearly observed that the micro-strain is induced in (ZnO: Er)Sapphire thin film is approximately two times larger than (ZnO: Er)Silicon thin film which are the responsible for growing most preferentially orientation along (111) plane in case of sapphire substrate and (002), (111) plane in Si-substrate. From the obtained c/a values (see Table 1), it is observed that the ZnO structure is distorted in (ZnO: Er)sapphire thin film compared to (ZnO: Er)silicon thin film. It is well known that the oxide of Er i.e. Er2O3 exhibit anti-ferromagnetic at low temperature (~ 3.3 K) and paramagnetic at room temperature or above of 3.3 K27, as a result the magnetic contribution from Er2O3 in this present work is ruled out. Again, the Er-cluster or Er-metal exhibits both ferromagnetic as well as paramagnetic at low temperature28. As, we have not observed Er-clusters or Er-atomic cluster, so it is also ruled out in the present study. The Raman spectra of (ZnO: Er)sapphire/silicon thin films are shown in Fig. 1e. The spectral features shows that the (ZnO: Er)sapphire/silicon thin films are highly photoluminescence (PL) behaviours as inset Fig. 1e, while (ZnO: Er)sapphire is higher PL than (ZnO: Er)silicon indicating the formation of highly defective (ZnO: Er)sapphire thin film. In a (ZnO: Er)sapphire thin film, Raman mode peaks are difficult to observe due to the dominance of PL intensity over the Raman scattering signal. The presence of this high PL behaviour, particularly in (ZnO: Er)sapphire thin film is due to formation higher defective thin films compare to (ZnO: Er)silicon and responsible for the higher ferromagnetic behaviors29,30,31,32,33,34. All the observed spectroscopic Raman peaks are identified in Fig. 1e can be assigned to a wurzite ZnO structure35,36,37. However, in the Raman spectra the peak at 99 cm−1 is associated to the E2 (low) dominated by the vibrations of the heavy Zn-sub-lattice and the band at around 438 cm−1 attributed to the E2 (high) mode, mostly involves the oxygen atoms of (ZnO: Er)sapphire/silicon thin film35,36,37. A significant shift 483 cm−1 for the (ZnO: Er)sapphire is observed due to the formation of different phase of ZnO: Er films. The peak 620 cm−1 is ascribed as the combination of optical-acoustic mode38. The peak at 200 cm−1 is the overtone of the E2 (low) and the peak at 336 cm−1 is due to the combination E2 (high) and E2 (low) modes37,38. The E2 (high) mode centered at 438 cm−1 has a stronger intensity and narrower line-width in (ZnO: Er)silicon film, which indicates that the grown of ZnO: Er is a hexagonal wurtzite structure37,38.

Fig. 1
Fig. 1
Full size image

(a) Scanning electron microscopy (SEM) of ZnO: Er deposited on (a) Sapphire substrate and (b) Silicon substrate. XRD patterns of ZnO: Er (c) On sapphire and (d) On silicon. Raman spectra of ZnO: Er deposited on sapphire and silicon substrate.

Table 1 The values of cell parameters from XRD, crystallite size, strain for ZnO:Er deposited on sapphire and silicon substrate.
Full size table

XPS has been used to analyse the chemical composition (Zn, O, Er), quantification (at%), electronic structure and their defects. The quantificational analysis of XPS spectra shows ≈ 2.5 at% and ≈ 1.0 at% Er is substituted in (ZnO: Er)sapphire and (ZnO: Er)silicon thin films, respectively, although the deposition parameters/conditions remained unchanged during the deposition of (ZnO: Er)silicon/sapphire thin films. It is also observed that the Er-content in the film structure is not only different at%, but their peak positions in Zn 3p and O 1s XPS spectra of (ZnO: Er)sapphire/silicon as shown in Fig. 2a-d, e-h) respectively, are shifted at higher energy levels compared to pure ZnO, indicating that the deposition of ZnO: Er thin films are substrate-dependent. The corresponding Zn 3p and O 1s XPS spectra and their convolutions are shown in Fig. 2a–d and e–h, respectively. The Zn 3p spectrum for pure ZnO consists of two strong peaks at ≈88.4 eV and ≈91.5 eV, correspond to Zn 3p3/2 and Zn 3p1/2 states, respectively. The difference between the Zn 3p3/2 and Zn 3p1/2 binding energies found to be ≈ 3.1 eV, which closely matches with the value reported earlier39,40. Oxygen in the ZnO matrix is coupled to Zn+-ions in the + 2-oxidation state, as indicated by the locations of both peaks and the difference in their binding energies39,40. The Zn 3p spectral features of (ZnO: Er)sapphire/silicon thin films closely resemble each other with red-shift ≈ 0.8 eV [3p3/2: ZnO (pure), 88.4 eV →(ZnO: Er)sapphire/(ZnO: Er)silicon, 89.5 eV/90.3 eV; 3p1/2: ZnO (pure), 91.5 eV →(ZnO: Er)sapphire/(ZnO: Er)silicon, 92.5 eV/93.8 eV] and increase their peak widths [ZnO→(ZnO: Er)sapphire→(ZnO: Er)silicon] with variations of their intensities. This suggests that the introduction of Er influences the electronic states of Zn due to the existence of Zn interstitials/vacancies. The O 1s peak at 531.4 (± 0.1) eV of pure ZnO is believed to be due to the wurtzite structure of hexagonally arranged Zn2+ ions in the metal oxide as confirmed from XRD results. Different components, such as oxygen vacancies and absorbed oxygen/hydroxyl groups41,42,43, could be the reason for the other peak at higher binding energy, 533.0 (±0.5) eV. Based on existing literature, it is inferred that the peak is associated with oxygen vacancies in the system42,43. The incorporation of Er-ions into the ZnO lattice results in a heightened intensity of the peak at elevated binding energy, signifying an increase in oxygen vacancies. This increase in oxygen vacancies is, in turn, responsible for the improved saturation magnetization associated with Er. This implies higher room temperature ferromagnetism in (ZnO: Er)sapphire/silicon thin films compared to diamagnetic ZnO. The room temperature (RT) magnetic behavioral M-H hysteresis loops of ZnO (ZnO: Er) are shown in Fig. 2i and the Er 4d XPS spectrum of (ZnO: Er)sapphire/silicon thin films are shown in Fig. 2j inset in Fig. 2i. In the Er 4d XPS spectrum, the binding energy of ≈ 169.4 eV is attributed to the 4d levels of Er3+ ions are in a trivalent (Er3+) state in (ZnO: Er)sapphire/silicon thin films via an interaction with an unfilled shell. Thus, we can deduce that the RT-FM is a fundamental characteristic of the materials involved by Er³⁺ ions and may be in a high-spin configuration (4f116s0). While Er³⁺ ions contribute to magnetism, the mechanism is complex and often involves other factors like substitutions, carrier mediation and unpaired electron, rather than simply a high-spin configuration. In the RT magnetic M-H hysteresis loops, the contribution of the diamagnetic signals of substrates was adequately corrected. It is observed that the magnetic moment (MS =2.3 emu/g) and coercivity (HC =63 Oe) of (ZnO: Er)sapphire thin film is higher than (ZnO: Er)silicon thin film (MS =0.5 emu/g, HC =15 Oe) that confirms the magnetism is also substrate dependent. This observation is consistent with Er content (at %) obtained from XPS elemental/quantitative analysis, which indicates that the Er at% is substrate dependent. The (ZnO: Er)sapphire has a higher Er (~ 2.5 at%), and the magnetization is also higher than the films deposited on a silicon substrate with an Er concentration of ~ 1.0 at%. In the context of Er-substitution, the predominant defects in ZnO are probably oxygen vacancies (V0) and zinc vacancies (VZn). Specifically, V0 is expected to be more prevalent in Zn-rich environments due to its lower formation energy compared to zinc interstitials, while VZn is likely to be more prominent in oxygen-rich conditions. Thus, it is plausible to infer that there may be a significant presence of zinc and oxygen vacancies in (ZnO: Er)sapphire/silicon thin films. Our experimental findings may align with the bound-magnetic-polaron (BMP) model suggested by Coey et al.44, as the presence of both zinc and oxygen vacancies promotes the development of ferromagnetism in diluted magnetic semiconductors (DMSs)45,46. It is noted that the magnetic contribution is not due to Er-oxide and/or Er-metal. As Er has more affinity towards oxygen, so oxygen deficiency occurred in parent ZnO. Higher Er-content leads to higher oxygen deficiency and forms oxygen vacancy clusters which are responsible for observation of ferro-magnetic behavior. This is the reason why we have given much emphasis on vacancy mediated ferromagnetism. According to the BMP model, magnetic exchange can be mediated by shallow donors (Zni) electrons that form bound magnetic polarons, and the presence of dopant, i.e., the magnetic exchange interaction between Zn or O vacancies of Er³⁺ ions occupying the same space, is aligned with Er spins3+, forming BMPs. With the reduction of V0 and VZn by Er-substitution in ZnO, neighboring Er-ions coupled via a V0 and VZn (ferromagnetic exchange) are connected by a zinc bond (no exchange interaction) or oxygen bond (super-exchange interaction), which are responsible for the alteration of MS values. It is noted that the exchange interaction has not arisen due to Er. Due to high oxidation state of Er, the oxygen deficiency has been observed in the ZnO structural matrix. The exchange interaction is Er coupled and/or mediated which is responsible for the observation of ferromagnetism. The substrate can also influence the formation and characteristics of these defects, thus affecting the BMP formation and the overall magnetic behavior. In the case of (ZnO: Er)silicon, the ferromagnetic behavior is weakened, possibly due to Er ions3+ near the surface sites with their ionic radii mismatch47.

Fig. 2
Fig. 2
Full size image

(a) X-ray photoelectron spectroscopy (XPS) of (a) Zn 3p of ZnO, ZnO: Er; Zn 3p decomposed into Gaussian features (b) pure-ZnO, (c) ZnO: Er (Sapphire), (d) ZnO: Er (silicon). XPS of (e) O 1s of ZnO, ZnO: Er; O 1s decomposed into Gaussian features (f) pure-ZnO, (g) ZnO: Er (Sapphire), (h) ZnO: Er (silicon). (i) Magnetic M-H hysteresis loops of ZnO, ZnO: Er (Sapphire), and ZnO: Er (silicon). (j) Inset XPS of Er 4d.

To investigate the electronic/atomic structures and gain insight into their microstructure, we have used the Zn and O K-edges, Zn and Er L3-edge XANES spectra, along with Zn K-edge and Er L3-edge EXAFS spectra of ZnO (ZnO: Er)sapphire/silicon thin films. The Zn K- and O K-edges XANES spectra of ZnO (ZnO: Er)sapphire/silicon thin films are shown in Fig. 3a,b. As a reference, the Zn K-edge XANES spectrum of Zn-foil (Zn0) is also shown in Fig. 3a. To identify the proper absorption edge, first-order differentiation of Zn K-edge XANES spectra is shown in the lower inset of Fig. 3a. The transition of Zn 1s →Zn 4p dominates the Zn K-edge absorption spectra, which show the excited 1s core electrons of the zinc atoms. The Zn 1s→ 4p transitions are the distinctive peaks that are seen on the leading absorption edge. ༴he fully populated 3d levels of metal Zn-atoms (Zn0) are pushed below the Fermi level. Nevertheless, the zinc atoms’ oxides produce empty states in the 3d levels, Zn2+, because their electrons are required for bonding. Therefore, one might assume that an electron excited from the core levels would have a d-level state. However, quantum selection criteria, which prohibit s electrons from moving in d shells (i.e., Δl = ± 1) prevent the 1s → 3d transition. The Zn K-edge absorption edge of Zn-foil (Zn0) and pure-ZnO are observed at ≈ 9659.1 eV and ≈ 9665.5 eV respectively, along with a weak pre-edge at ≈9657.0 eV (marked with a black line), which are the hybridization of O-2p and Zn2+ 3d orbitals48. The ZnO exhibits molecular bands characterized by narrower 4p levels, with the closest neighbouring atoms being oxygen. The outer-shell electrons of the oxygen atoms are confined to the 2p levels, while the unoccupied levels on the zinc atoms correspond to the 4p levels. Quantum selection rules arise due to an interaction between the p-levels, which results in the narrowing of the 4p levels in ZnO materials (i.e. Zn2+). The absorption edges of Er-substituted ZnO, (ZnO: Er)sapphire/silicon are observed at ≈ 9663.0 eV/9662.5 eV, in between Zn-foil and pure ZnO as shown in the upper inset of Fig. 3a. The edge shifts from ≈ 9665.5 eV (ZnO) → ≈9663.0 eV/9662.5 eV (ZnO: Er)silicon/sapphire indicate the oxidation states of ZnO: Er (between Zn2+ and Zn0) are decreased on (ZnO: Er)silicon/sapphire thin films compared to pure ZnO. The oxidation states of (ZnO: Er)sapphire thin film is nearer to Zn-foil (Zn0) than (ZnO: Er)silicon, due to the edge shift being ≈ 0.5 eV lower than (ZnO: Er)silicon. It is also noticed that the pre-edge of Er-substituted ZnO is merged with the absorption edge in both (ZnO: Er)sapphire/silicon films. Similar edge shifts are also observed in O K-edge XANES spectra, illustrate in the upper inset Fig. 3b, provide a magnified perspectives of the pre-edge characteristics following the subtraction of the background as lower inset Fig. 3b, which has been accomplished using a best-fitted Gaussian curve as denoted by the dotted line. This variation in the general line shapes, the absorption intensity, and the positions of the peaks of the Zn and O K-edge XANES spectra, as shown in Fig. 3a, b of the pure-ZnO and (ZnO: Er)sapphire/silicon, has been regarded as evidence of substrate-dependent in O/Zn-p states. According to the dipole-transition selection rule, features A, B, C, and D, as shown in Fig. 3a, can be attributed to the Zn 4p-derived state. In Zn K-edge, the main feature at ≈ 9668.0 eV (marked as B) is attributed to the Zn-atoms in tetrahedral site symmetry of ZnO49. The upper inset of Fig. 3a shows a comparison of Zn K-edge spectra of ZnO and (ZnO: Er)sapphire/silicon thin films. A detailed view of the near-edge characteristics following the subtraction of the background, which has been accomplished using a best-fitted Gaussian curve as represented by the dotted line. No pre-edge feature is observed in (ZnO: Er)sapphire/silicon, but peak shifts occur on the (ZnO: Er)sapphire/silicon thin films. The peaks at ≈ 9683/9712 eV (marked as C/D) of ZnO/(ZnO: Er)silicon films because of the diverse scattering effects of the photoelectrons in the neighboring atoms environment that shifted to ≈ 9681/9714 eV for the (ZnO: Er)sapphire films. We have obtained the intensities by integrating the area within the range 9648.5–9595.5 eV for ZnO, 9653.0–9697.0 eV for (ZnO: Er)sapphire/silicon thin film and found that the intensity of ZnO: Er is reduces from ZnO (29 unit) → ZnO: Er (sapphire/silicon: 25.5/25.3 unit). Again, the intensity of (ZnO: Er)sapphire films is higher than the films (ZnO: Er)silicon, which implies an increase of the Zn 4p-orbital occupation or the positive effective charge on the Zn-ions that supports the formation of higher ferromagnetic behavior in ZnO: Er thin films. Figure 3b shows O K-edge XANES spectra of ZnO/(ZnO: Er)sapphire/silicon. In accordance with the dipole-transition selection rule, the distinct peaks observed in the energy range 531–565 eV originate from O 1s-derived states transitioning to 2pσ-derived (along the bilayer) and O 2pπ-derived (along the c axis) states50,51,52. The contributions in the absorption edge mainly arise from O 2p - Zn 4s hybridized states53. The strong features A is due to the asymmetry of the wurtzite structure. The absorption edges are shifted to lower energy, when Er is substituted with ZnO, which can be seen in an enlarged depiction of the pre-edge (feature-A′) after the background has been eliminated, utilizing a best-fitted Gaussian curve as indicated by the dotted line as shown in the upper inset of Fig. 3b. The fact that the intensity of (ZnO: Er)sapphire/silicon thin films (feature-A′) was found to be lower than pure ZnO (upper inset Fig. 3b) indicating the total number of unoccupied O 2p-derived states is decreased, which can be seen as the transfer of electrons from Er → O 2p states due to O 2p-Er 5d hybridizations in (ZnO: Er)sapphire/silicon thin films50,51. The lower inset Fig. 3b shows the differences in the O K-edge near edge features of [(ZnO: Er)sapphire/silicon - (ZnO)] thin films. The overall spectral intensity of (ZnO: Er)sapphire/silicon has been reduced and the near-edge feature has been broadened in comparison to those of ZnO, indicating a reduction in the number of unoccupied O 2p states because of the Er-diffusion and heightened negative effective charge of the O ion. The rise in the effective charge of the O ions is explained by the observation that Pauling’s electronegativity of Er (1.24) is considerably less than that of Zn (1.65)54. The enhancement of these characteristic features (A′, B′, C′, D′, E′, F′, G′) signifies the elevated local density of states (DOS) that originate from the defects and/or dangling bonds present in ZnO/(ZnO: Er)sapphire/silicon thin films55. Again, intensities of the features B′, C′, G′ is higher in (ZnO: Er)sapphire thin films compared to (ZnO: Er)silicon films; whereas the features A′, D′, E′, F′ of the (ZnO: Er)silicon films is higher than (ZnO: Er)sapphire films, indicating the population of defects and/or vacancies are substrate dependent. The overall intensities of (ZnO: Er)sapphire film is lower than (ZnO: Er)silicon suggesting a rise in the occupancy of the O 2p orbitals and an increase in the negative effective charge of the O-ion that supports the higher ferromagnetism in the (ZnO: Er)sapphire films.

Fig. 3
Fig. 3
Full size image

(a) Zn K-edge X-ray absorption near edge structure (XANES) spectra of ZnO and ZnO: Er deposited on sapphire/silicon substrate. The lower inset displays the 1 st order differentiation of the Zn K-edge XANES spectra and upper inset display a magnified view of the pre-edge features after the background has been subtracted using a best-fitted Gaussian line. (b) O K-edge XANES spectra of ZnO and ZnO: Er deposited on sapphire/silicon substrate. The upper inset displays a magnified view of the pre-edge features after the background has been subtracted using a best-fitted Gaussian line, whereas the lower inset plots differences of O K-edge between ZnO: Er and ZnO (ZnO: Er-ZnO).

Figure 4a,b shows Zn L3-edge and Er L3-edge XANES spectra of ZnO and (ZnO: Er)sapphire/silicon thin films. XANES measurements examine the Zn L-edge, providing knowledge of the unoccupied Zn d and Zn s-states. The Zn L3 and L2 regions, correspond to the Zn 2p → Zn 4s and Zn 2p → Zn 3danti-bonding states, respectively56. The spectral characteristics observed in the L2 region indicate the electron transition from Zn 2p state to Zn 3d state. No pre-edge feature is observed in ZnO/(ZnO: Er)sapphire/silicon implying no Zn 4s band transition occurs. The Zn L3-edge absorption peaks arise due to the transition of the 2p core electron to the unoccupied 4s/4d levels as final states. In accordance with the dipole-transition selection rule, Zn L3-edge XANES investigates the unoccupied states derived from Zn s- and d-orbitals. The Zn-3d orbital is completely filled, making the lowest unoccupied orbital of the Zn-ion is Zn-4s, succeeded by Zn-4p and Zn-4d57. The degenerate states 2p3/2 and 2p1/2 emerge due to spin-orbit coupling, leading to multiplets centered at ≈1027.8 eV and ≈1032.0 eV. The crystal field (octahedral) raises the degeneracy of 2p3/2 and 2p1/2 levels, creating of t2g and eg sub-band symmetries58. The fine structure multiplets arise from two effects: (i) interaction between the 3d (electron) and 2p core hole, and (ii) crystal field created by neighboring ions at a Zn2+ site59. The inset at the bottom of Fig. 4a is a magnified view of the near-edge features after subtracting the background using a best-fitted Gaussian curve indicated by the dotted line. It is found that the absorption features A/B(wide range) are observed at ≈1027.8/1031.9 eV in ZnO, which is shifted towards lower energy level (A′/B′) at ≈1027.3/1028.9 eV for the (ZnO: Er)sapphire film and ≈1027.3/1029.0 eV in (ZnO: Er)silicon film, respectively. The overall peak width and intensities of (ZnO: Er)sapphire/silicon thin films are decreased compared to pure-ZnO, indicating that the quantity of unoccupied Zn 4sd states in proximity to the conduction-band minimum is lowered as the Er is substituted with ZnO. We have integrated the Zn L3-edge region and found that the intensity of pure ZnO ≈ 0.77 (a.u.) is higher relative to (ZnO: Er)sapphire/silicon ≈0.49/0.47 (a.u.). The intensity of (ZnO: Er)sapphire films higher than the (ZnO: Er)silicon, indicating that the unoccupied Zn 4sd states are higher in (ZnO: Er)sapphire than (ZnO: Er)silicon. The augmentation of the defining features denotes the heightened local density of states (DOS) that originate from defects or dangling bonds within various ZnO nanostructures51. The (ZnO: Er)sapphire thin film shows higher absorption intensity concerning the (ZnO: Er)silicon thin film, signifying a greater prevalence of defects and/or vacancies at O sites in (ZnO: Er)sapphire thin film. As a result, the ferromagnetic nature is stronger in the (ZnO: Er)sapphire thin film than (ZnO: Er)silicon. To investigate the local electronic structure of Er in ZnO, we used the XANES spectrum of Er-substituted ZnO at Er L3-edge. The variation of white line intensity of (ZnO: Er)sapphire/silicon thin films is at ≈8363.3 eV and ≈8398.0 eV respectively, as shown in Fig. 4b. The peaks at ≈8363.3 eV (R) and ≈8398.0 eV (T) arise due to the 2p−5d transition of Er upon x-ray absorption60,61,62,63,64,65,66. The magnitude of the white line peak signifies the occupancy of 5d valence electronic states, which is contingent upon the symmetry of the absorbing atom. A high-order symmetry, such as that of the ErO6 octahedron (Oh point group), results in a concentrated density of states (DOS) of the 5d states, attributed to the equivalent coordination of six O atoms around Er. This concentrated DOS leads to a significant pre-edge peak; nevertheless, the peak height of the (ZnO: Er)silicon thin film is reduced as a result symmetry degradation59. The observed decrease in intensity to a reduced DOS concentration, which can be ascribed to the less symmetric geometry in the coordination53,58. Thus, the symmetry around Er increases for the (ZnO: Er)sapphire thin film as shown in the inset lower of Fig. 4b, implying the higher DOS. This higher DOS (2p−5d) of (ZnO: Er)sapphire thin film implies a higher ferromagnetic nature compared to the (ZnO: Er)silicon film.

Fig. 4
Fig. 4
Full size image

(a) Zn L3-edge and (b) Er L3-edge X-ray absorption near edge structure (XANES) spectra of ZnO and ZnO: Er deposited on sapphire/silicon substrate. Inset (a, b) display a magnified view of the Zn L3 and Er L3 pre-edge features after the background has been subtracted using a best-fitted Gaussian line.

The best tool for further study of the local structure around Zn atoms in Er-substituted ZnO is extended x-ray absorption fine structure (EXAFS) spectroscopy. The oscillations observed in EXAFS arise from the scattering and interference of photoelectron waves from source atoms with elevated kinetic energies. Due to their kinetic energy, photoelectrons can move between adjacent atoms. The outer shell electrons of these neighboring atoms exert a coulombic force on the incoming photoelectrons, which then follow a path characterized as either single or multiple scattering. The fluctuation EXAFS spectra can be extracted by using the ATHENA and ARTEMIS programs.67 In this case, Fourier transform of k3χ was performed, where k is the photoelectron wave vector, defined as k = (4π m (E − E0)/h)1/2, where m is the electron mass, E is the photon energy, E0 is the energy threshold of the absorption edge, and h is the Planck constant. This Fourier transform gives the radial structural function (RSF) around Er i.e., the atomic distribution. This involves the reduction of data from the absorption spectra and the use of Fourier transform to obtain the χ(R) versus R plots, the development of theoretical EXAFS spectra starting from a hypothesized crystallographic structure, and finally the fitting of the experimental χ(R) versus R data with the theoretical data using the ARTEMIS programs67. The fitting parameters consist of the coordination number (N), bond distances, and disorder (Debye-Waller) factors (σ2), which indicate the mean square fluctuations in the distances. To analyze the local structure around Zn and Er in detail, χ(R) versus R plots generated for ZnO and (ZnO: Er)sapphire/silicon thin films from the µ(E) versus E spectra through the Fourier transform of the χ (k) versus k plots with appropriate k(k3) weighting are presented in Fig. 5, measured at the Zn K-edge and Er L3-edge. Figures 5a,b illustrate the amplitude of the Fourier transform (FT) obtained from the EXAFS k3χ data for ZnO and (ZnO: Er)sapphire/silicon films, along with Zn-Foil as reference, and their respective oscillations at the Zn K-edge. The peak observed in Zn-foil at ≈2.58 Å is described as a Zn-Zn bond, and the peak at ≈1.84 Å observed in pure ZnO is defined as Zn-O coordination bond. It shows that the line shapes and radial distribution of the FT spectra at Zn K-edge of the Er-substituted ZnO are not identical to ZnO, but with a slight change in their intensities and peak positions. The Zn-O bond lengths of ZnO: Er films are reduced to ≈1.63 Å/1.87Å (sapphire/silicon-substrate) compared to pure ZnO films (≈1.84 Å), but the Zn-Zn bond length increased to 3.09 Å/3.16 Å [(ZnO: Er)sapphire/(ZnO: Er)silicon] compared to Zn-foil (≈2.58 Å)48. In addition, one extra peak at ≈1.35 Å is observed in (ZnO: Er)siliconfilms, equivalent to the weak peak at ≈1.15 Å observed in pure ZnO films and also defined as the Zn-O bond50. These results suggest that the Er-atoms take the place of the host Zn ions in the core of ZnO: Er; the local atomic configuration at the Zn sites is markedly distorted, owing to the substantial difference in ionic radii of Er3+ (0.89 Å) and Zn2+ (0.74 Å). It is considered that the Er atoms substituted at the Zn sites on the surface layer and formed dangling bonds/defects with a high degree of disorder. The EXAFS k3χ data as shown in Fig. 5b indicates the spectral features of (ZnO: Er)silicon film is identical with pure ZnO but non-identical with (ZnO: Er)sapphire film indicates along with a slight phase shift in scattering data beyond k = 6 Å−1. This indicates a minor disturbance on the formations at closer distance to the source Zn atoms in (ZnO: Er)sapphire/silicon thin films. An observable shift at the Zn-O bond peak position points out a disturbance in the atomic locations. It is highlighted a displacement of defects in (ZnO: Er)sapphire/silicon compared to the host elements of the ZnO materials in their wurtzite crystal structure with Er-substitution. These disorders and/or defects are higher in (ZnO: Er)sapphire thin film, implying higher ferromagnetic behaviors. Figure 5c shows the radial structure function of (ZnO: Er)sapphire/silicon and Fig. 5d is the k3-weighted EXAFS spectra of (ZnO: Er)sapphire/silicon. In EXAFS, the peak between 1.4 Å and 2.2 Å is the contribution of oxygen atoms in the ZnO: Er system53. The nearest-neighbor atom from Er is found at ≈ 1.7 Å/1.84 Å [(ZnO: Er)sapphire/(ZnO: Er)silicon] consistent with the observations reported by other groups61, where they found that the first neighbor atom should be O. In addition, a shoulder peak at ∼2.4 Å was observed near the first nearest-neighbor peak in (ZnO: Er)sapphire films, suggesting the structural distortion of pseudo-octahedral55,60, implying the formation of higher ferromagnetic nature. The Fourier transformed (FT)χ(k) with k weight as shown in Fig. 5d, gives a pseudo-radial structure function around Er59. In this, a slight phase shift in scattering data beyond k = 6 Å−1 indicates a slight disturbance of the atoms closer to the source Zn atoms in (ZnO: Er)sapphire/silicon thin film.

Fig. 5
Fig. 5
Full size image

Magnitude of Fourier transform (FT) of EXAFS k3χ data at the (a) Zn K-edge and (b) Er L3-edge and (c, d) corresponding EXAFS data within the range k = 0 to 8 Å, weighted by ZnO and ZnO: Er deposited on sapphire/silicon substrate along with Zn-Foil.

In conclusion, the DC sputtering method successfully synthesized the pure ZnO and (ZnO: Er)sapphire/silicon thin films. ZnO was crystallized in the hexagonal wurtzite structure of ZnO without any impurity phases. The most common defects in ZnO/(ZnO: Er)sapphire/silicon thin films are VO/VZn vacancies, which are confirmed from the transitions of Zn 1s→Zn 4p, Zn 2p → Zn 4s, Zn 2p → Zn 3d, and O 1s →2p in XANES study and peak shifts of Zn 2p/O 1s at higher binding energy level in XPS study. In XANES study, transitions occurred with the hybridization of O 2pσ and O 2pπ-derived states. These transitions/hybridizations in (ZnO: Er)sapphire/silicon thin films are lower than pure ZnO due to Er-substitution with ZnO, which implies the existence of zinc/oxygen vacancies in the (ZnO: Er)sapphire/silicon thin films. The Er-ions are in trivalent states in ZnO: Er thin films, and the ferromagnetism is caused by the substitution of Zn2+ in the ZnO lattice by Er3+-ions. Results shows that the magnetic moment in (ZnO: Er)sapphire is higher than (ZnO: Er)silicon thin film due to more Er3+-ions occupying adjacent cation lattice positions. With the reduction of oxygen vacancies by Er-substitution in ZnO, neighboring Er-ions coupled via oxygen vacancy/bonds (ferromagnetic exchange/super-exchange interaction) are coupled in (ZnO: Er)sapphire, which are responsible for the higher value of MS. The overall occupying of Er3+-ions due to higher vacancies/hybridization in (ZnO: Er)sapphire thin film, provides the higher ferromagnetic signals at room temperature as observed in magnetic M-H hysteresis loops, where the magnetic parameters MS/HC in (ZnO: Er)sapphire films is greater than the (ZnO: Er)silicon films. These (ZnO: Er)sapphire/silicon thin film DMSs could be employed for optoelectronic as well as spintronic applications.

Methods

Preparation of ZnO and Er-substituted ZnO

ZnO and Er-ZnO thin films were grown using the DC sputtering technique using sapphire and silicon substrates, respectively. Highly dense ZnO and Er-substituted ZnO (Zn0.95Er0.05O) targets of size 2.5 cm diameter and 0.5 cm thickness were prepared by solid state reaction technique. The targets were sintered at 650 °C for 5 h to have higher densification. These targets were used in DC sputtering unit. The sputtering chamber was evacuated with high purity Ar and then O2 separately to remove the unwanted volatile impurities present within the chamber. The sputtering was carried out using a plasmagen gas mixture of argon and oxygen with a percentage ratio of 70:30 and the chamber pressure was kept at ~ 10−6 Torr. The sputtering power was fixed at maximum of 100 W. The silicon and sapphire substrates were used for depositing ZnO and Er-substituted ZnO over it. Initially the substrates were ultrasonically cleaned using acetone, isopropanol, HF and mixture of HNO3 & H2SO4, stepwise. Finally, the substrates were cleaned using potassium dichromate and isopropanol. Once the substrate was cleaned, it was fixed in the substrate holder present in the sputtering unit. The substrate temperature was kept at 700 °C during deposition and later annealed at the same temperature to improve the oxygen content with a heating and cooling rate of 5 °C/min.

Characterization

The surface morphology and microstructure were investigated by scanning electron microscopy (SEM) (Philips, XL30). The crystal structure was analyzed with Rigaku Miniflex XRD in the Bragg diffraction angle of 10° to 80° using CuKα radiation source (λ = 1.5404 Å) at 40 KV and 15 mA. Raman spectra were measured using HORIBA scientific XploRA at 532 nm (2.41 eV) LASER light excitation energy. Magnetic properties of ZnO: Er are typically measured using techniques like the superconducting quantum interference device (SQUID) to analyses magnetization curves. The Zn K-edge and Er L3-edge XANES measurements were taken at the Wiggler-17C1 beamline; whereas Zn L3 edge and O K-edge XANES measurements were taken at the 20A1 beamlines, and XPS measurements were taken from TLS-09A1 beamlines of Taiwan Light Source at the National Synchrotron Radiation Research Center (NSRRC) in Hsinchu, Taiwan. The Zn K-edge EXAFS spectra were obtained in fluorescence mode, while the O K-edge and Zr/Zn L3-edge spectra were obtained in surface-sensitive electron-yield mode. The resolution was set to 0.1 eV (0.2 eV) at a photon energy of 530 eV (1020 eV) for the O K-edge (Zn L3-edge) XANES measurements. All XANES measurements were carried out to study the electronic structures of ZnO: Er thin films deposited on sapphire and silicon substrates. Zn K-edge and Er L3-edge EXAFS were performed to investigate the local structure of Er in ZnO: Er thin films.