Investigation of Fe doped ZnO thin films by X-ray absorption spectroscopy

Abstract

Fe doped ZnO thin films with varying Fe doping concentrations have been deposited by rf magnetron sputtering technique on c-Si substrates. X-ray Diffraction (XRD) measurements show the pure wurtzite structure of all the samples. X-ray near edge spectroscopy (XANES) measurements confirm that Fe is present in the samples in Fe²⁺ and Fe³⁺ oxidation states. Extended X-ray Absorption Fine Structure (EXAFS) measurements at Zn and Fe K edges prove that Zn atoms are being substituted by Fe dopants, Zn vacancies are created in the samples near the dopant sites and also support the presence of mixed oxidation states obtained by XANES measurements. Oxygen K-edge XANES results support the conjecture that Fe goes to Zn sites in tetrahedral symmetry. Magnetization measurements show the presence of room temperature ferromagnetism (RTFM) in the samples and saturation magnetization is found to increase as Fe doping concentration increases. The observed ferromagnetism seems to be intrinsic in nature and it is explained in the light of XANES and EXAFS observations.

---

Introduction

Dilute magnetic semiconductors (DMSs) with ferromagnetic (FM) ordering above room temperature have attracted great interest for their promising applications in the emerging field of spintronics.^1–5 Ferromagnetic ordering above room temperature could be achieved in a DMS material by doping the semiconductor with a very small quantity of a transition metal (TM) element. Hence both the charge and spin of the electron can be utilized for device operation such as spin-valve transistors, spin light-emitting diodes, non-volatile memory, logic devices, optical isolators and ultrafast optical switches, etc.^6,7 The search for room temperature ferromagnetic DMS in doped ZnO gained pace after theoretical predictions of this by Dietl et al.² in 5% Mn doped ZnO, due to hole doping. The theoretical demonstration by first-principle electronic structure calculations by Sato and Katayama-Yoshida⁸ suggested that transition-metal (TM = Ti, V, Cr, Mn, Fe, Co, Ni) doped ZnO compounds are ferromagnetic provided that the carriers produced by the TM doping formed a partially filled spin-split impurity band. Giuli et al.⁹ very recently reported that transition metal-doped zinc oxide can be used as a new lithium-ion anode material to increase the energy density of high power lithium-ion batteries.

Several methods have been used to produce ZnO based DMS nanocrystalline powder and thin films, such as sol gel,^10,11 solid state reaction,¹² molecular beam epitaxy,¹³ laser ablation¹⁴ and radio frequency (rf) sputtering¹⁵ etc. Despite the successful preparation of TM doped ZnO, the origin of ferromagnetism in these DMS is still an open question. Different groups have reported different reasons for the origin of ferromagnetism in these DMS systems, particularly whether FM arises due to the precipitation of metallic cluster in a homogenous ZnO host matrix¹⁶ or due to the uniformly distributed TM cations in the semiconductor lattice or due to the presence of an entirely separate magnetic phase^12,17 or due to the defects.¹⁸ Thus the above issue is still a fairly unresolved question and further experimental and theoretical studies are required particularly to explore the local environment around the host and the dopant cations carefully to obtain unambiguous results regarding the origin of FM in these systems.

X-ray Absorption Spectroscopy (XAS), which consists of two techniques, X-ray near edge spectroscopy (XANES) and Extended X-ray Absorption Fine Structure (EXAFS), yields element-specific local structure information (like bond length, coordination number, disorder, coordination geometry and oxidation state) around the host and dopant cations separately and hence is very useful in studying such TM doped DMS systems. In a recent communication, on Mn doped ZnO thin films, by employing the local structure investigation tool of XANES and EXAFS at Zn, Mn and O K edges, we have shown that oxygen vacancy plays a crucial role in inducing ferromagnetism in the samples, while the presence of Mn clusters destroy room temperature ferromagnetism (RTFM) in the samples.¹⁹ In the present paper, we have carried out similar investigations by XAS measurements on the origin of RTFM in Fe doped ZnO thin films prepared by rf sputtering technique.

Experimental details

Preparation of samples

Pure and Fe (1, 2, 4, 6 and 10%) doped ZnO thin films have been deposited using an in-house built magnetron sputtering system.²⁰ Sputtering is a physical vapor deposition (PVD) process in which a plasma is created between two electrodes in an inert gas atmosphere (generally argon) using an electric field. In case of dc sputtering, positively charged ions (Ar⁺) from the plasma are accelerated towards the negatively charged electrode where the material to be deposited or the “target” is kept.²¹ The positive ions are accelerated by potentials ranging from a few hundred to a few thousand electron volts and strike the negative electrode with sufficient force to dislodge and eject atoms from the target. These atoms (cluster of atoms or adatoms) condense on substrate surfaces that are placed in proximity to the magnetron sputtering cathode and is generally kept grounded. In case of rf sputtering technique, which is used to sputter insulators like in this case, sputtering conditions prevail at both target and substrate regions, however, target sputtering is made much more significant by reducing its effective area compared to the substrate. Magnetron sputtering technique, additionally uses a closed magnetic field around the target material so that electrons undergo helical movement near the target, enhancing the efficiency of the ionization process. This allows creation and sustention of plasma at relatively lower pressures which reduces both contamination in the growing film and energy loses of the sputtered atoms through collision with gas molecules.

In the present case, the target materials were high purity Fe doped ZnO (1, 2, 4, 6 and 10%) discs of 75 mm diameter. Substrates used for all the depositions are crystalline silicon (111) having dimension of 25 mm × 10 mm and the depositions have been carried out under high purity argon ambient, the Ar flow rate being controlled by using a mass flow controller. The substrates were not intentionally heated during deposition; however it has been observed that during deposition, the substrate temperature reaches up to 50–80 °C due to the presence of the magnetron plasma. A quartz crystal thickness monitor has been used in the deposition system near the target for in situ monitoring of the rate of deposition and total thickness deposited. The crystal monitor is calibrated by ex situ thickness measurement of a undoped ZnO thin film by spectroscopic ellipsometry. The other deposition parameters used are sputtering pressure of 4 × 10⁻³ mbar, substrate to target distance of 4.5 cm, RF power of 100 W and all the films were deposited for 20 minutes.

Characterization

Preliminary structural characterisations of the films were carried out by Grazing Incidence X-ray Diffraction (GIXRD) measurements using a Rigaku SmartLab diffractometer with monochromatic Cu Kα X-rays at a grazing angle of incidence and with a wavelength of 1.54 Å. The XANES and EXAFS measurements have been carried out at the Energy-Scanning EXAFS beamline (BL-09) at the INDUS-2 Synchrotron Source (2.5 GeV, 200 mA) at Raja Ramanna Centre for Advanced Technology (RRCAT), Indore, India.^22,23 The beamline optics consists of a Rh/Pt coated collimating meridional cylindrical mirror and the collimated beam reflected by the mirror is monochromatized by a Si(111) (2d = 6.2709) based double crystal monochromator (DCM). The second crystal of the DCM is a sagittal cylindrical crystal which is used for horizontal focusing of the beam while another Rh/Pt coated bendable post mirror facing down is used for vertical focusing of the beam at the sample position. In case of the present thin film samples, the measurement is done in fluorescence mode using a Si drift detector (Vortex detector) in 45° geometry. Rejection of the higher harmonics content in the X-ray beam is performed by detuning the second crystal of the DCM.

XANES measurements at the O K edges of the Fe doped ZnO samples were performed at room temperature in the total electron yield (TEY) mode at the soft X-ray absorption spectroscopy (SXAS) beamline (BL-01) of the Indus-2 synchrotron radiation source at RRCAT, Indore, India.²⁴ M–H measurements at room temperature (RT) were performed with a Superconducting Quantum Interference Device Vibrating Sample Magnetometer (SQUID-VSM-050 MPMS), Quantum Design, USA.

Results and discussions

The phase purity of Fe doped ZnO films have been checked by GIXRD measurement which (Fig. 1) shows no secondary phase other than wurtzite hexagonal ZnO phase (space group P6₃mc; JCPDS card no. 36-1451) in the samples. It has been observed from Fig. 1 that with the increase in Fe doping concentration, hexagonal crystal structure of ZnO remains preserved and there is no appearance of other impurity peaks however the diffraction peaks are shifted to lower angles and it is consistent with the other reports on Fe doped ZnO films.^25–28 The shift in the (100) peak has been shown in enlarged scale in the inset of Fig. 1. Similar shift of the XRD peaks towards lower angles with increase in Fe doping concentration has also been observed by Gautam et al.²⁹ for Fe doped nanorod samples which has been attributed to slight increase in Zn–O and Zn–Zn bond lengths with increase in Fe doping concentration in the ZnO samples. In wurtzite structure of ZnO unit cell, one oxygen atom neighboring four Zn atoms form the geometry of a triangular pyramid with oxygen located at the pyramid centre and substitution of Zn²⁺ ions (ionic radii 0.60 Å in tetrahedral coordination³⁰) by smaller Fe³⁺ ions (ionic radii 0.49 Å in tetrahedral coordination³⁰) of smaller radii results in a small increase in the Zn–O and Zn–Zn bond distances. We have used the position of (100) diffraction peak to calculate the lattice constant (a) using the formula given by Wang et al.²⁷ The calculated lattice parameter and Full Width Half Maxima (FWHM) of the (100) diffraction peak are plotted with Fe doping concentration in Fig. 2, which shows that both the above quantities increase with Fe doping concentration. Chen et al.²⁶ also observed similar broadening in the FWHM and explained it on the basis of strain and disorder induced by difference in the ionic radii of Fe and Zn.

Fig. 1 GIXRD plots of undoped ZnO and Fe doped ZnO thin films (1%, 2%, 4%, 6% and 10% doping). Inset: variation of the (100) peak position with increase in Fe doping concentration.

Fig. 2 Variation of lattice constant (a) and FWHM of the diffraction peak (100) with Fe doping concentration.

XANES spectra measured at Fe K edge for Fe doped ZnO films are shown in Fig. 3. Wei et al.²⁸ had calculated the full XANES spectra of Fe doped ZnO films using full multiple scattering ab initio calculations for two different configurations viz., Fe at the Zn site and Fe at the tetrahedral site in Fe₃O₄. The shallow shoulder feature obtained in case of our samples denoted by A in Fig. 3, is quite similar to the XANES spectra calculated with Fe at Zn site in ZnO as reported by Wei et al.²⁸ which confirms substitution of Zn atoms by Fe dopants in our samples.

Fig. 3 Normalised XANES spectra of undoped ZnO and Fe doped ZnO thin films (1%, 2%, 4%, 6% and 10% doping) at Fe K-edge. The reference spectra of Fe metal, FeO, Fe₃O₄ and Fe₂O₃ are given for comparison. The inset figure shows the expanded part of the pre-edge region.

The XANES spectra of four standard samples with different oxidation states of Fe viz., Fe metal foil (+0), FeO (+2), Fe₃O₄ (+2/+3) and Fe₂O₃ (+3) have also been shown in Fig. 3 along with the XANES spectra of the Fe doped ZnO films. It is well known that transition metals show the pre-edge features prior to the main absorption edge.³¹ While the main absorption edge manifests transition of the core level electron to continuum, the pre-edge peak is mainly due to electric dipole transition of the core shell (1s) electron to the 3d–2p hybridized orbital, the pre-edge centroid position depends on the oxidation state of absorbing atom and the intensity of the pre-edge depends on the coordination geometry. In general, the pre-edge intensity of the tetrahedral geometry (non centro-symmetry) is higher than the intensity of the octahedral geometry (centro-symmetry). The intense pre-edge peak in the present Fe doped ZnO samples, as shown in Fig. 3, indicates the presence of tetrahedral coordination. Similar pre-edge peaks in the Fe K-edge XANES spectra have also been observed by Gautam et al.²⁹ for their Fe doped ZnO nanorods. The XANES spectra also indicate that Fe K-edges of the samples fall in-between that of FeO and Fe₂O₃ standards, with more proximity to Fe₂O₃, indicating the presence of Fe in mixed oxidation states of +2 and +3 in our samples, with higher probability of +3 state. To know the oxidation state of Fe quantitatively, we have performed linear combination fit (LCF) using the XANES spectra of all four standards. The best fitting can only be obtained after excluding the Fe metal standard which indicates that there is no metal cluster present in our samples. This is in contrast to our earlier result¹⁹ on Mn doped ZnO thin films which show signature of Mn clusters at higher doping concentrations. Signature of Fe clustering have not been observed for our sol–gel derived Fe doped ZnO nanoparticles also reported earlier.¹¹ The LCF results (Table 1) shows that Fe is present in 30 : 70 ratio in +2 and +3 oxidation state in the samples and the ratio does not vary significantly with the change in Fe doping concentration. Mixed oxidation state of Fe has been observed by Gautam et al.²⁹ also in polycrystalline Fe doped ZnO nanorods and they have also obtained similar ratio using LCF calculations. Seo et al.³² have also observed a mixed oxidation state of Fe ions in ZnO thin films prepared by rf magnetron sputtering on sapphire substrate. However, many researchers^8,11,33 have also found that Fe exists in purely trivalent state in their Fe doped ZnO samples. The expected oxidation state of Fe is +2 if it replaces Zn atom in ZnO while the presence of Fe in +3 state may be due to the presence of nearby cationic vacancy (Zn vacancy) at the substitutional sites. The XANES spectra of the samples at Zn K edge are shown in Fig. 4 which completely resemble that of pure ZnO sample over the whole Fe doping concentration range. This indicates that Zn related impurity or defects around Zn sites are negligible in the films.

Table 1 Contribution of Fe²⁺ and Fe³⁺ oxidation states obtained from linear combination fitting of XANES spectra

Sample	Fe^2+	Fe^3+	Rfactor
Zn0.99Fe0.01O	29%	71%	0.0015
Zn0.98Fe0.02O	29%	71%	0.0020
Zn0.96Fe0.04O	32%	68%	0.0017
Zn0.94Fe0.06O	30%	70%	0.002
Zn0.90Fe0.10O	29%	71%	0.002

---

Fig. 4 Normalised XANES spectra of undoped ZnO and Fe doped ZnO thin films (1%, 2%, 4%, 6% and 10% doping) measured at Zn K-edge along with Zn metal foil.

The normalized EXAFS spectra of pure and Fe doped ZnO samples are shown in Fig. 5 and 6 at Zn and Fe K edges respectively. In order to take care of the oscillations in the absorption spectra, μ(E) has been converted to absorption function χ(E) defined as follows:³⁴

where, E₀ is the absorption edge energy, μ₀(E₀) is the bare atom background and Δμ₀(E₀) is the step in μ(E) value at the absorption edge. The energy dependent absorption function χ(E) has been converted to the wave number dependent absorption function χ(k) using the relation,where, m is the electron mass. χ(k) is weighted by k² to amplify the oscillation at high k and the χ(k)k² functions are fourier transformed in R space to generate the χ(R) versus R spectra [or the Fourier transformed (FT)-EXAFS spectra] in terms of the real distances from the center of the absorbing atom. The set of EXAFS data analysis software available within IFEFFIT package have been used for EXAFS data analysis.³⁵ This includes background reduction and Fourier transform to derive the χ(R) versus R spectra from the absorption spectra (using ATHENA software), generation of the theoretical EXAFS spectra starting from an assumed crystallographic structure and finally fitting of experimental data with the theoretical spectra using ARTEMIS software.

Fig. 5 Normalised EXAFS spectra of undoped ZnO and Fe doped ZnO (1%, 2%, 4%, 6% and 10% doping) thin films measured at Zn K-edge.

Fig. 6 Normalised EXAFS spectra of Fe doped ZnO (1%, 2%, 4%, 6% and 10% doping) thin films measured at Fe K-edge.

The Fourier transform EXAFS spectra or χ(R) versus R plots have been generated for all the samples from the μ(E) versus E spectra following the methodology described above and are shown in Fig. 7 for the data measured at Zn K-edge. The structural parameters (atomic coordination and lattice parameters) of ZnO used for simulation of the theoretical EXAFS spectra of the samples have been obtained from ref. 36 and the best fit χ(R) versus R plots (fitting range R = 1.0–3.5 Å) of the samples have also been shown in Fig. 7 along with the experimental data. The bond distances, co-ordination numbers (including scattering amplitudes) and disorder (Debye–Waller) factors (σ²), which give the mean square fluctuations in the distances, have been used as fitting parameters and the best fit results are summarized in Table 2 for Zn K-edge measurements. It should be mentioned here that the amplitude reduction factors (S₀²) for the different paths have been obtained from fitting of FT-EXAFS spectrum of a commercial ZnO powder sample and have been kept constant during the fitting of the thin film spectra.

Fig. 7 Fourier transformed EXAFS spectra of (a) ZnO (b) 1% Fe doped ZnO (c) 2% Fe doped ZnO (d) 4% Fe doped ZnO (e) 6% Fe doped ZnO and (f) 10% Fe doped ZnO (scatter points) at Zn K edge along with best fit theoretical plots (solid line).

Table 2 Bond length, coordination number and disorder factor obtained by fitting of FT-EXAFS data measured at Zn K edge

Path	Parameter	ZnO:pure	ZnO:1% Fe	ZnO:2% Fe	ZnO:4% Fe	ZnO:6% Fe	ZnO:10% Fe
Zn–O	R (Å)	1.94 ± 0.01	1.95 ± 0.01	1.94 ± 0.01	1.94 ± 0.02	1.96 ± 0.01	1.96 ± 0.01
	N	4	4	4	4	4	4
	σ^2	0.0060 ± 0.0009	0.0063 ± 0.0009	0.0060 ± 0.0009	0.0074 ± 0.0009	0.0073 ± 0.0008	0.0072 ± 0.009
Zn–Zn	R (Å)	3.19 ± 0.01	3.18 ± 0.01	3.19 ± 0.01	3.18 ± 0.01	3.21 ± 0.02	3.19 ± 0.01
	N	12	12	12	12	12	12
	σ^2	0.0097 ± 0.0007	0.0099 ± 0.0008	0.0117 ± 0.0009	0.0129 ± 0.009	0.0123 ± 0.0007	0.0146 ± 0.008

---

The goodness of the fit in the above process is generally expressed by the R_factor which is defined as:³⁷

where, χ_dat and χ_th refer to the experimental and theoretical χ(R) values respectively and Im and Re refer to the imaginary and real parts of the respective quantities.

The first peak at 1.4 Å (phase uncorrected spectrum) in Zn K-edge FT-EXAFS spectrum (Fig. 7) of pure ZnO has contributions from four oxygen atoms at a bond length of 1.94 Å. The bond length obtained is in good agreement with other studies.³⁸ The second peak at 2.8 Å (phase uncorrected spectrum) is fitted with single Zn–Zn shell with coordination number of 12 and at a Zn–Zn distance of 3.19 Å. It can be seen from Table 2 that the Zn–O and Zn–Zn bond lengths do not change with increase in Fe doping concentration in the samples, however the Debye–Waller factor (σ²) increases which is due to the increase in disorder in ZnO lattice due to substitution of Zn atoms by Fe dopants. It should be noted here that in accordance to the shift of the XRD peaks of ZnO samples towards lower angles with Fe doping as discussed above, we expected to observe marginal increase in the Zn–O and Zn–Zn bond lengths by EXAFS measurements as has been observed by Gautam et al.²⁹ However, we have not observed any such change possibly due to fact that it is beyond the detection limit of our measurement set-up. It should also be noted from Table 2 that there is no change in oxygen or Zn coordination around Zn sites with increase in Fe doping concentration in the films. This result is quite different from what we have observed in case of Mn doped ZnO thin films,¹⁹ which showed the presence of oxygen vacancy around Zn sites that increases with increase in Mn doping concentration.

The Fourier transform EXAFS spectra at Fe K edge are shown in Fig. 8. The first peak at 1.3 Å in the phase uncorrected spectra is a contribution of the nearest Fe–O shell and second peak at 2.7 Å is a contribution of next nearest Fe–Zn shell. To carry out the fitting theoretical FT-EXAFS spectrum has been generated by substituting Zn atoms in ZnO structure by Fe atoms. The bond distances, co-ordination numbers and disorder factors have been used as fitting parameters and best fit results of Fe K edge FT-EXAFS data are shown in Table 3. The first peak in 1% Fe doped ZnO is a contribution of first oxygen shell at distance of 1.92 Å. This bond length is found to be slightly lower than Zn–O bond lengths obtained in the samples from Zn K-edge EXAFS measurements. This decrease in the bond length is expected since the ionic radii of tetrahedral coordinated Fe³⁺ (0.49 Å) is lower than the ionic radii of tetrahedral coordinated Zn²⁺ (0.60 Å).³⁰ This result corroborates the presence of Fe primarily in Fe³⁺ state in the samples as obtained from the Fe K-edge XANES results described earlier. Similar decrease in the bond length is also observed by other researchers also.⁸ The Fe–O bond length further decreases for 2% Fe doped sample and remains same for all the doping concentration. The second peak in the Fe K-edge FT-EXAFS spectra is a contribution of next nearest Fe–Zn shell at a distance of ∼3.20 Å. The Fe–Zn bond length is found to be comparable to the Zn–Zn bond length which suggests that Fe goes to Zn sites in the ZnO lattice. Zn coordination in this shell is however lower than the ideal value of 12 indicating the presence of Zn vacancies in the samples near the dopant sites. The presence of Zn vacancy and mixed oxidation state of the dopants, in case of Fe doped ZnO have been observed by many other researchers also.^39–42 Karmakar et al.³⁹ have found that Fe is present in the +2 and +3 oxidation state and suggested that presence of the Fe in +3 state is due to the hole doping in the system by Zn vacancy, however no experimental evidence is provided on the presence of Zn vacancy. Our experimental results support the theoretical prediction of Debernardi et al.⁴² by density functional theory which suggests that presence of Zn vacancies at the next nearest positions of Fe atoms is energetically more favorable than isolated Zn vacancies. In case of our samples also, the presence of Zn vacancies are not observed from Zn K-edge EXAFS measurements, however they are found in the neighborhood of Fe atoms from the Fe K-edge EXAFS measurements.

Fig. 8 Fourier transformed EXAFS spectra of (a) 1% Fe doped ZnO (b) 2% Fe doped ZnO (c) 4% Fe doped ZnO (d) 6% Fe doped ZnO and (e) 10% Fe doped ZnO (scatter points) at Fe K edge along with best fit theoretical plots (solid line).

Table 3 Bond length, coordination number and disorder factor obtained by fitting of FT-EXAFS data measured at Fe K edge

Path	Parameter	ZnO:1% Fe	ZnO:2% Fe	ZnO:4% Fe	ZnO:6% Fe	ZnO:10% Fe
Fe–O	R (Å)	1.92 ± 0.02	1.88 ± 0.01	1.87 ± 0.01	1.87 ± 0.01	1.88 ± 0.01
	N	3.57 ± 0.10	3.57 ± 0.14	3.57 ± 0.10	3.81 ± 0.14	3.91 ± 0.13
	σ^2	0.0047 ± 0.0009	0.0047 ± 0.0009	0.0063 ± 0.0008	0.0090 ± 0.0009	0.0099 ± 0.0009
Fe–Zn	R (Å)	3.21 ± 0.02	3.20 ± 0.01	3.19 ± 0.01	3.20 ± 0.02	3.20 ± 0.02
	N	9.46 ± 0.28	10.73 ± 0.32	10.73 ± 0.32	11.44 ± 0.46	11.72 ± 0.47
	σ^2	0.0196 ± 0.002	0.0197 ± 0.002	0.0194 ± 0.001	0.0233 ± 0.002	0.0256 ± 0.002

---

As can be seen from Table 3, the variation in the bond length with Fe doping concentration is not significant, however the disorder factor (σ²) is found to increase with increase in Fe doping concentration manifesting again substitution of Zn atoms by Fe dopants. These results also corroborate with the results obtained from Zn K-edge EXAFS measurement and with the fact that FWHM of XRD peaks of the samples increase with increase in Fe doping concentration as discussed earlier. Normalized XANES spectra at O K-edge for samples with different Fe concentrations are shown in Fig. 9. The spectral features are denoted by A, B, C and D in the figure. The main peak denoted by B at ∼536 eV is due to the transition of 1s electron to O 2p–Zn 4sp hybridized state as stated by other workers also.^43,44 The peak A at 530 eV, which is like a shallow shoulder for the undoped sample, shows significant enhancement with increase in Fe doping concentration. This feature at A is attributed to transitions from 1s core state of oxygen to 2p oxygen state hybridized with 3d state of Fe.⁴⁵ In case of undoped ZnO, the transition to oxygen 2p-Zn 3d hybridized state is not possible since the d orbital of Zn is fully occupied. As Fe atoms replace Zn atoms in ZnO lattice, the empty orbital of Fe 3d state are contributing to this transition and enhances this peak at position A. The peak at D is due to the O 2p hybridization with extended Fe 4sp orbital.⁴⁰ Thus O K-edge XANES measurements also support the above conjecture that Zn atoms are being substituted by Fe dopants in the samples.

Fig. 9 O K edge XANES spectra of undoped and Fe doped ZnO samples (1%, 2%, 4%, 6% and 10% doping).

The magnetization (M) versus field (H) curves for 2%, 4% and 10% Fe doped ZnO samples measured at RT are shown in Fig. 10. The diamagnetic contribution from the substrate as well as the host ZnO has been subtracted from each plot. The inset shows the enlarged view of low field region. All three samples exhibit RTFM ordering with significant and increasing saturation magnetization (M_S) (0.14 emu g⁻¹ for 2%, 0.18 emu g⁻¹ for 4% and 0.63 emu g⁻¹ for 10% Fe doping at 5 kOe) with increase in Fe dopant concentration. However, the coercive field (H_C) is low (∼51 Oe) for all samples and it indicates that all Fe doped ZnO thin film samples are ferromagnetically soft. The magnetization measurements by Kumar et al.⁴⁵ on 1% Fe doped ZnO samples shows soft ferromagnetic behavior at room temperature with 27 Oe coercive field and ∼0.125 emu g⁻¹ saturation magnetization. The saturation magnetization is found to increase with Fe doping concentration in the samples. Magnetic measurements on undoped ZnO and Fe doped ZnO nanocrystals by Inamdar et al.³³ showed similar coercivity and saturation magnetization.

Fig. 10 Room temperature M–H curve of Fe doped ZnO thin films (2%, 4% and 10% doping).

The origin of room temperature ferromagnetism in TM doped DMS material is generally attributed to two major causes, intrinsic or extrinsic. The extrinsic origin is due to the possible presence of any secondary magnetic phase viz., metallic Fe cluster or Fe based oxide phases in case of the present set of samples. The intrinsic origin, on the other hand, is due to the substitution of host atoms by dopants or presence of defects and interstitials. For the present set of Fe doped ZnO thin film samples, we have observed from GIXRD, XANES and EXAFS measurements that Fe gets incorporated into the Zn sites of ZnO lattice and no evidence of Fe metallic cluster or extra Fe related secondary oxide phases like FeO, Fe₂O₃, ZnFe₂O₄ and Fe₃O₄ is found upto the maximum Fe doping concentration. Thus it can be concluded that the origin of ferromagnetism in the present set of samples is intrinsic in nature and due to Fe substitution at Zn sites and subsequent creation of Zn vacancies near Fe sites. Wang et al.⁴¹ have also suggested that p-type defects such as Zn vacancies play a crucial role in tuning and stabilization of ferromagnetism in Fe doped ZnO thin films.

Summary and conclusions

Fe doped ZnO thin films with varying Fe doping concentrations have been deposited with rf magnetron sputtering technique on c-Si substrates. GIXRD measurements show pure wurtzite structure of the all samples. XANES measurements show that Fe is present in the samples with mixed oxidation states of Fe²⁺ and Fe³⁺ and quantitative ratio of the two contributions have been obtained using linear combination fitting of the XANES spectra. EXAFS measurements at Zn and Fe K edges show that Zn atoms are being substituted by Fe dopants and Zn vacancies are created near to Fe substituting sites. EXAFS results support the presence of mixed oxidation states of Fe²⁺ and Fe³⁺ in the samples as observed also by XANES measurements. O K edge XANES results also support that Fe is going to Zn sites in the ZnO lattice. Magnetization measurement shows the room temperature ferromagnetic behaviour of the samples with increasing saturation magnetization as Fe doping concentration increases in the samples. Magnetic behaviour of the Fe doped ZnO thin film samples is thus due to substitution of host Zn atoms by Fe dopants and subsequent creation of Zn vacancies near Fe sites and not due to any extrinsic effect such as presence of Fe metallic cluster or Fe related secondary oxide phases in the samples.

References

1. H. Ohno, Science, 1998, 281, 951–955.
2. T. Dietl, H. Ohno, F. Matsukura, J. Cibert and D. Ferrand, Science, 2000, 287, 1019–1022.
3. K. Ando, H. Saito, Z. Jin, T. Fukumura, M. Kawasaki, Y. Matsumoto and H. Koinuma, J. Appl. Phys., 2001, 89, 7284–7286.
4. K. Takamura, F. Matsukura, D. Chiba and H. Ohno, Appl. Phys. Lett., 2002, 81, 2590–2592.
5. Y. Matsumoto, M. Murakami, T. Shono, T. Hasegawa, T. Fukumura, M. Kawasaki, P. Ahmet, T. Chikyow, S. Y. Koshihara and H. Koinuma, Science, 2001, 291, 854–856.
6. S. A. Chambers, Mater. Today, 2002, 34–39.
7. H. Ohno, F. Matsukura and Y. Ohno, JSAP Int., 2002, 5, 4–13.
8. K. Sato and H. Katayama-Yoshida, Phys. E, 2001, 10, 251–255.
9. G. Giuli, A. Trapananti, F. Mueller, D. Bresser, F. d'Acapito and S. Passerini, Inorg. Chem., 2015, 54, 9393–9400.
10. H. Liu, J. Yang, Y. Zhang, L. Yang, M. Wei and X. Ding, J. Phys.: Condens. Matter, 2009, 21, 145803.
11. S. Kumar, N. Tiwari, S. N. Jha, S. Chatterjee, D. Bhattacharyya, N. K. Sahoo and A. K. Ghosh, RSC Adv., 2015, 5, 94658–94669.
12. D. Wang, Z. Q. Chen, D. D. Wang, J. Gong, C. Y. Cao, Z. Tang and L. R. Huang, J. Magn. Magn. Mater., 2010, 322, 3642–3647.
13. C. Tie-Xin, C. Liang, Z. Wang, H. Yu-Yan, Z. Zhi-Yuan, X. Fa-Qiang, I. Kurash, Q. Hai-Jie and W. Jia-Ou, Chin. Phys. B, 2013, 22, 026101.
14. S. D. Yoon, Y. Chen, D. Heiman, A. Yang, N. Sun, C. Vittoria and V. G. Harris, J. Appl. Phys., 2006, 99, 08M109.
15. A. J. Chen, X. M. Wu, Z. D. Sha, L. J. Zhuge and Y. D. Meng, J. Phys. D: Appl. Phys., 2006, 39, 4762–4765.
16. J. H. Park, M. G. Kim, H. M. Jang and S. Ryu, Appl. Phys. Lett., 2004, 84, 1338–1340.
17. X. Z. Li, J. Zhang and D. J. Sellmyer, Solid State Commun., 2007, 141, 398–401.
18. W. Yan, Z. Sun, Q. Liu, Z. Li, Z. Pan, J. Wang, S. Weia, D. Wang, Y. Zhou and X. Zhang, Appl. Phys. Lett., 2007, 91, 062113.
19. A. K. Yadav, S. M. Haque, D. Shukla, R. J. Choudhary, S. N. Jha and D. Bhattacharyya, AIP Adv., 2015, 5, 117138.
20. A. Biswas, R. Sampathkumar, A. Kumar, D. Bhattacharyya, N. K. Sahoo, K. D. Lagoo, R. D. Veerapur, M. Padmanabhan, R. K. Puri, D. Bhattacharya, S. Singh and S. Basu, Rev. Sci. Instrum., 2014, 85, 123103.
21. J. George, Preparation of Thin Films, Marcel Dekker, NY, 1992.
22. A. K. Poswal, A. Agrawal, A. K. Yadav, C. Nayak, S. Basu, S. R. Kane, C. K. Garg, D. Bhattacharyya, S. N. Jha and N. K. Sahoo, AIP Conf. Proc., 2014, 1591, 649.
23. S. Basu, C. Nayak, A. K. Yadav, A. Agrawal, A. K. Poswal, D. Bhattacharyya, S. N. Jha and N. K. Sahoo, J. Phys.: Conf. Ser., 2014, 493, 012032.
24. http://www.cat.ernet.in/technology/accel/srul/beamlines/mcd_pes.html.
25. Z. C. Chen, L. J. Zhug, X. M. Wu and Y. D. Meng, Thin Solid Films, 2007, 515, 5462–5465.
26. A. J. Chen, X. M. Wu, Z. D. Sha, L. J. Zhuge and Y. D. Meng, J. Phys. D: Appl. Phys., 2006, 39, 4762–4765.
27. C. Z. Wang, Z. Chen, Y. He, L. Li and D. Zhang, Appl. Surf. Sci., 2009, 255, 6881–6887.
28. X. X. Wei, C. Song, K. W. Geng, F. Zeng, B. He and F. Pan, J. Phys.: Condens. Matter, 2006, 18, 7471–7479.
29. S. Gautam, S. Kumar, P. Thakur, K. H. Chae, R. Kumar, B. H. Koo and C. G. Lee, J. Phys. D: Appl. Phys., 2009, 42, 175406.
30. R. D. Shannon, Acta Crystallogr., 1976, 32, 751–767.
31. T. Yamamoto, X-Ray Spectrom., 2008, 37, 572–584.
32. S.-Y. Seo, C.-H. Kwak, S.-H. Kim, B.-H. Kim, C.-I. Park, S.-H. Park and S.-W. Han, J. Cryst. Growth, 2010, 312, 2093–2097.
33. D. Y. Inamdar, A. K. Pathak, I. Dubenko, N. Ali and S. Mahamuni, J. Phys. Chem. C, 2011, 115, 23671–23676.
34. X-Ray Absorption: Principles, Applications, Techniques of EXAFS, SEXAFS and XANES, ed. D. C. Konigsberger and R. Prince, Wiley, New York, 1988.
35. M. Newville, B. Ravel, D. Haskel, J. J. Rehr, E. A. Stern and Y. Yacoby, Phys. B, 1995, 154, 208–209.
36. E. H. Kisi and M. M. Elcombe, Acta Crystallogr., 1989, 45, 1867–1870.
37. R. Bhunia, A. K. Yadav, S. N. Jha, D. Bhattacharyya, S. Hussain, R. Bhar and A. K. Pal, Polymer, 2015, 78, 1–12.
38. S. Basu, D. Y. Inamdar, S. Mahamuni, A. Chakrabarti, C. Kamal, G. R. Kumar, S. N. Jha and D. Bhattacharyya, J. Phys. Chem. C, 2014, 118, 9154–9164.
39. D. Karmakar, S. K. Mandal, R. M. Kadam, P. L. Paulose, A. K. Rajarajan, T. K. Nath, A. K. Das, I. Dasgupta and G. P. Das, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 75, 144404.
40. W.-G. Zhang, B. Lu, L.-Q. Zhang, J.-G. Lu, M. Fang, K.-W. Wu, B.-H. Zhao and Z.-Z. Ye, Thin Solid Films, 2011, 519, 6624–6628.
41. Q. Wang, Q. Sun, P. Jena and Y. Kawazoe, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 79, 115407.
42. A. Debernardi and M. Fanciulli, Appl. Phys. Lett., 2007, 90, 212510.
43. V. Vaithianathan, B. Lee, C. Chang, K. Asokan and S. S. Kim, Appl. Phys. Lett., 2006, 88, 112103.
44. J.-H. Guo, L. Vayssieres, C. Persson, R. Ahuja, B. Johansson and J. Nordgren, J. Phys.: Condens. Matter, 2002, 14, 6969–6974.
45. S. Kumar, Y. J. Kim, B. H. Koo, S. K. Sharma, J. M. Vargas, M. Knobel, S. Gautam, K. H. Chae, D. K. Kim, Y. K. Kim and C. G. Lee, J. Appl. Phys., 2009, 105, 07C520.

---