Multifunctional role of dysprosium in HfO2

PCCP View Article Online

Published on 03 July 2017. Downloaded by Banaras Hindu University on 12/07/2017 10:14:55.

PAPER

Cite this: DOI: 10.1039/c7cp02800h

View Journal

Multifunctional role of dysprosium in HfO2: stabilization of the high temperature cubic phase, and magnetic and photoluminescence properties† Sandeep Kumar,a S. B. Raib and Chandana Rath

*a

Hafnium oxide (HfO2) can exist in different crystalline structures such as monoclinic at room temperature, tetragonal at 1700 1C and cubic at 2600 1C. In the present study, nanocrystalline powders of HfO2 synthesized by a Pechini type sol–gel technique show a monoclinic phase, P21/c, at room temperature. By incorporating Dy into the HfO2 lattice, the intensity of all diffraction peaks corresponding to P21/c decreases and at a concentration of 11 at% of Dy, the monoclinic phase transforms completely to % the cubic phase, Fm3m, showing a mixed phase of monoclinic and cubic at intermediate concentrations (5–9 at%) of Dy. For the first time, we have stabilized the high temperature cubic phase of HfO2 at room temperature by incorporating Dy. Selected area electron diffraction patterns confirm the monoclinic and the cubic phase as observed from the X-ray diffraction patterns. A mechanism for stabilization of the high temperature cubic phase in Hf1xDyxO2 has been analyzed based on the substitution of dysprosium for Received 29th April 2017, Accepted 3rd July 2017 DOI: 10.1039/c7cp02800h

hafnium ions and the formation of oxygen vacancies. While ferromagnetic ordering at room temperature observed in HfO2 nanoparticles is quenched after incorporating 1 at% of Dy, photoluminescence (PL) studies demonstrate excellent emissions in the blue and yellow region after exciting with UV light of wavelength 352 nm. Combining excitation and emission profiles, we have proposed a tentative energy band diagram

rsc.li/pccp

illustrating the energetic processes taking place in Hf1xDyxO2.

1 Introduction Lead-free binary oxides such as ZnO, TiO2, CeO2, SnO2, ZrO2 and HfO2 etc. are of special scientific interest because of their intriguing properties and numerous potential industrial applications such as high-k gate dielectric in MOSFETs, spintronics, scintillators and anti-reflective coatings, etc.1–5 More importantly, their non-toxic nature towards the environment is an added advantage over the lead (Pb) based materials. Among them, hafnium oxide (HfO2) has been extensively researched for many attractive applications and has emerged as the most promising candidate for an alternate high-k gate dielectric in silicon based technology.5 In bulk, HfO2 shows a polymorphism exhibiting the monoclinic phase (m-phase) at room temperature and a tetragonal structure (t-phase) between B1700 1C and B2600 1C which eventually transforms to the cubic phase (c-phase) above 2600 1C under ambient conditions.6 A systematic theoretical

a

School of Materials Science and Technology, Indian Institute of Technology, Banaras Hindu University, Varanasi, 221005, India. E-mail: [email protected], [email protected] b Department of Physics, Banaras Hindu University, Varanasi, 221005, India. E-mail: [email protected] † Electronic supplementary information (ESI) available. See DOI: 10.1039/c7cp02800h

This journal is © the Owner Societies 2017

study predicts that the high temperature tetragonal and cubic phases of HfO2 exhibit high dielectric constant values (k) such as B70 and B25–30, respectively.7,8 The stable high temperature phases of HfO2 at room temperature are technologically very important due to their appropriate high-k value. It has been reported that high-k and high temperature HfO2 phases can be stabilized by doping elements having a lower valency than the Hf4+ ion.9–11 Using density functional theory, Lee et al.12 predicted that while dopants such as Si, Ge, Sn, P, Al or Ti having ionic radii smaller than Hf favour the tetragonal phase, dopants like Y, Gd or Sc with larger ionic radii stabilize the cubic phase of HfO2 at room temperature. The substitution of Hf with larger cations thus obviously increases the lattice volume resulting in a more stable cubic phase. In this context, few experimental investigations on stabilization of the cubic phase at room temperature by doping different elements such as divalent or trivalent Mn, Ce and Y ions have been reported.13–15 Besides the stabilization of the high temperature cubic phase of HfO2, it is shown to possess rich magnetic and optical properties. HfO2 being a non-magnetic transition metal oxide with d0 cations, the recent discovery of unusual room temperature ferromagnetism (RTFM) in HfO2 thin films is fascinating and compelling for materials research communities.16 A wide group of researchers believe that the required defects such as

Phys. Chem. Chem. Phys.

View Article Online


Paper

oxygen vacancies (VO) can be held responsible for such type of ferromagnetic ordering present in non-magnetic oxides at room temperature.20–25 In this regard, Coey et al.17,18 report observation of RTFM in monoclinic HfO2 thin films originating due to unpaired electrons in extended molecular orbitals of the defects. N. H. Hong19 reports RTFM with a large magnetic moment in HfO2 thin films deposited by the pulsed laser deposition technique. Apart from experimental evidence, first principle calculations report the possibility of Hf vacancy induced short range magnetic interactions by forming high-spin defect states in the monoclinic HfO2.26,27 Apart from RTFM observed in thin films, ferromagnetic behavior has also been observed in HfO2 nanoparticles, nanoclusters and nanorods synthesized via various chemical routes at room temperature. Liu et al.28 have shown this ˇevic ´-Mitrovic´ et al.29 report the behavior in nanorods and Dohc same in nanopowders. Besides, RTFM has also been demonstrated in yttrium (Y) and nickel (Ni) doped HfO2 nanoparticles.29,30 However, the possible origin of such RTFM behavior observed in HfO2 and/or doped HfO2 is still confusing and debatable. Owing to the high transparency over most of the UV and visible region, HfO2 has been proven to be a strong phosphor material for scintillator applications. Interestingly, monoclinic (m) HfO2 based phosphors acting as hosts show remarkable photoluminescence (PL) properties when activated with an appropriate rare earth (RE) ion such as Eu, Ti, Tb, Sm and Ce etc.31–34 Most reports discuss the luminescent behavior of Eu doped HfO2 exhibiting mainly red emission.31,34,35,37 Lauria et al.35 report the modeling of the luminescence properties of Eu and Tb doped HfO2 nanoparticles by non-radiative trivalent Lu ions. Eu and Tb doped HfO2 luminescent nanocolloids have been used as a promising tool for in vivo diagnostics and therapy.36 The photoluminescent properties of Eu doped HfO2 powders have also shown to be intentionally altered by means of other codopants such as Li, Ta, Nb and V ions.37 Among lanthanides, dysprosium (Dy) as an efficient activator has been used in various hosts such as ZnO, MgO, and ZrO2 for desired optical applications.38–41 However, no studies have been reported on Dy doped HfO2 concerning either their structural or magnetic and/or optical properties when the size is reduced to the nanometer range. To study the structural evolution of HfO2 nanoparticles doped with Dy, we have synthesized HfO2 and Dy doped HfO2 nanoparticles by a versatile Pechini type sol–gel method. In HfO2 nanoparticles, the monoclinic phase of HfO2 transforms to the high temperature cubic phase when the Dy concentration reaches 11 at%. The cubic phase remains stable at room temperature. The intermediate concentration of Dy shows a mixed phase of monoclinic and cubic. To the best of our knowledge, no such structural transformation from monoclinic to cubic by doping Dy in HfO2 nanoparticles has been reported in the literature. Surprisingly, while the room temperature ferromagnetic behavior in HfO2 is quenched after doping 1 at% of Dy, the photoluminescence studies show emission of intense blue and yellow light. Systematic studies on the structural, microstructural, magnetic and photoluminescence properties of Dy doped HfO2 nanoparticles are discussed.


PCCP

2 Experimental details 2.1

Synthesis procedure

Pure and Dy doped hafnium oxide (HfO2) nanoparticles were synthesized with varying dysprosium (Dy) concentrations from 0 to 11 at% using a Pechini type sol–gel method. Dissolution of a stoichiometric amount of hafnium chloride (HfCl4, Alfa Aesar, 99.9%) and dysprosium nitrate (Dy(NO3)3, 99.9%, Sigma Aldrich) in 2 M citric acid (C6H8O7) solution was made under constant stirring with a Teflon coated magnetic bead. Ensuring the total solubility of chemical precursors, a clear sol was obtained. Ethylene glycol (EG) (C2H6O2) acting as a polymerizing agent was further added drop wise to the mixed solution. The sol was heat treated at 80 1C for 4 h to complete the condensation process and was dried at 120 1C for 20 h to form the precursor resin. It was light brown in color. Further, it was converted into a fine powder after grinding with an agate mortar and pestle. Finally, HfO2 and Dy doped HfO2 nanoparticles were calcined at 500, 700 and 900 1C for 5 h in air. The powders were white in color. 2.2

Characterization

The characterization of the prepared samples was carried out using EPMA (Electron probe micro analysis), XRD (X-ray diffraction), XPS (X-ray photoelectron spectroscopy), TEM (Transmission electron microscopy), magnetic measurements and PL (photoluminescence spectroscopy). EPMA, CAMECA, SX Five was utilized as an efficient tool for easy detection of Dy and Hf elements to investigate the composition of the prepared samples. Wavelength dispersive spectrometry (WDS) was employed for nondestructive elemental analysis. A focused electron beam obtained with 15 kV accelerating voltage having a beam current of 4 nA was ´rythriol) used for analysis. TAP (thallium acid phthalate), PET (pentae and LLIF (long lithium fluoride) crystals were further used to record the spectra. X-ray diffraction was carried out using an 18 kW rotating Cu anode based Rigaku Miniflex X-ray diffractometer in Bragg–Brentano configuration. X-ray photoelectron spectroscopy measurements were performed on an AMICUS X-ray photoelectron spectrometer using Mg Ka radiation having a pass energy of 75 eV. TEM images were taken on an FEI, Tecnai G2 20 Twin instrument. Magnetic measurements were carried out with an MPMS3 of Quantum Design. This instrument is able to measure very low values of moments within 7 T to +7 T magnetic field at temperatures down to 2 K. Room temperature excitation and photoluminescence spectra were measured with a Horiba, Jobin Yvon, Fluorolog using the excitation source of a 450 W UV lamp. The spectral resolution of the above spectrometer was 1 nm.

3 Results and discussion 3.1

EPMA analysis

The composition of the pure and Dy doped HfO2 samples has been examined using EPMA. In the EPMA technique, a high energy electron beam is incident on a sample surface that generates X-rays of different wavelengths. These X-rays generated


View Article Online


PCCP

Paper

Fig. 2 XRD patterns of Hf1xDyxO2 (0 r x r 0.11) samples calcined at 900 1C.

Fig. 1 (a) Back scattered electron (BSE) images for x = 0.05, 0.07, 0.09 and 0.11, respectively and (b) wavelength dispersive spectra (WDS) for x = 0.11.

through atomic transitions between different shells are then captured by a specific crystal monitor present in the instrument.42,43 Back scattered electron (BSE) imaging has been employed to investigate the distribution of constituting elements having high and low atomic numbers as depicted in Fig. 1(a) for x = 0.05, 0.07, 0.09 and 0.11, respectively. The contrast of different areas obtained in the BSE image is not similar everywhere since it is dependent on the atomic number of the constituting elements. Thus, BSE images contain reliable information regarding the distribution of the majority and trace elements in a sample. In these images, while portions of higher atomic number elements are brighter, areas having low atomic number elements show lower contrast.43 All BSE images explicitly show a uniform distribution of constituting elements present in Dy doped HfO2. Wavelength dispersive spectra (WDS) have also been taken for the prepared samples as shown in Fig. 1(b) for x = 0.11. WDS spectra confirm the presence of the majority species i.e. Hf and a trace of Dy in the Dy doped HfO2. The dominant peaks corresponding to Hf can be clearly seen and have been indexed as HfLa, HfLa2, HfLb, and HfLb2, respectively. The peaks associated with Dy have been labeled as DyLa and DyLb. 3.2

Microstructure and phase transformation

X-ray diffraction (XRD) patterns of HfO2 calcined at 500, 700 and 900 1C have been measured to examine the phase and crystallinity of the particles (see the ESI,† Fig. S1). While the sample calcined at 500 1C exhibits broad diffraction peaks indicating fine particles, with an increase in calcination temperature to 900 1C, diffraction


peaks are well distinguished and become sharper indicating a well crystalline nature. The observed diffraction peaks such as (011), (110), (1% 11), (111), (020), (200), (021), (2% 11) and (112) are well matched with the monoclinic phase (m-phase) of HfO2 with space group P21/c (JCPDS card no. 78-0049). No impurity phase has been identified within the detection limit of the instrument. As the HfO2 powders show a good crystalline nature, we have fixed the calcination temperature at 900 1C for HfO2 doped with the various concentrations of dysprosium (Dy). The XRD patterns of the Dy doped HfO2 are shown in Fig. 2. At lower Dy concentration (x = 0.01), the diffraction peaks are matched with the monoclinic phase as in pure HfO2 except for the decrease in peak intensity. However, at x = 0.05, in addition to the peaks corresponding to the monoclinic phase, a weak characteristic peak at 2y = 30.201 is observed, which matches with the (111) plane of the cubic phase (JCPDS card no. 53-0560). At x = 0.07, the diffraction peaks observed in the XRD pattern correspond to the mixed phase of monoclinic and cubic. At x = 0.09, the cubic phase becomes more pronounced at the expense of the monoclinic one. At the highest dopant concentration i.e. x = 0.11, characteristic diffraction peaks along the (111), (002), (022), (113) and (222) planes correspond to the pure cubic phase (c-phase) of HfO2 with the space group Fm3% m (JCPDS card no. 53-0560). No peak of the Dy2O3 phase is found in any of the Dy doped HfO2 samples. This suggests that Dy3+ ions substitute Hf4+ ions in both the monoclinic and cubic lattice. For the first time, we show that the monoclinic phase of HfO2 transforms to the high temperature cubic phase by doping Dy upto 11 at%. The high temperature cubic phase of HfO2 becomes stable at room temperature. Further, we have fitted the XRD patterns for x = 0 and 0.11 using Le-Bail profile fitting of FULLPROF program with pseudoVoigt function, shown in Fig. 3. Pure HfO2 is fitted with P21/c, the monoclinic phase, and x = 0.11 is fitted with Fm3% m corresponding to the cubic phase. Table 1 summarizes the


View Article Online


Paper

PCCP

Fig. 4 Variation of crystallite size (D) and lattice strain (e) in Hf1xDyxO2 (0 r x r 0.11) samples. Fig. 3 Le-Bail profile fitting of XRD patterns using the FULLPROF pro% gram: (a) x = 0 with P21/c and (b) x = 0.11 with Fm3m.

extracted cell volume and cell parameters from Le-Bail refinement. It is worth mentioning that incorporation of Dy in the HfO2 lattice increases the lattice parameter of the pure cubic phase of HfO2 from 5.095 to 5.119 Å for x = 0.11. It may be vividly observed that while no significant change in cell volume of the monoclinic phase is observed, the cubic phase shows an increase of 1.4%. The observed increase in the cell volume of the cubic phase further confirms the replacement of Hf4+ ions with Dy3+ which has a relatively larger ionic radius than the Hf4+ ion. This clearly indicates that a significant number of Dy3+ ions acquiring Hf4+ ion sites stabilize the high symmetry cubic phase even at room temperature. We have estimated the phase fraction by the integral peak area ratio of the cubic (111) peak to the monoclinic (111) or (1% 11) peak. Variation of ACP (111)/AMP (111) or AMP (111) with Dy concentration (figure not shown) shows % that the peak integral area ratio increases with increasing Dy concentration. While both the monoclinic and cubic phases coexist at an intermediate Dy concentration (x = 0.05–0.09), a complete cubic phase is eventually observed when x = 0.11. The Williamson–Hall (W–H) method is employed to separate the particle size and lattice strain of the XRD line broadening using the following eqn (1).44 b cos y 1 e sin y ¼ þ l D l

(1)

Table 1 Standard JCPDS data for the monoclinic and cubic phase of HfO2 and refined cell parameters for Hf1xDyxO2 (x = 0 and 0.11)

JCPDS card no./sample

Space group

78-0049 53-0560 Hf1xDyxO2 (x = 0) Hf1xDyxO2 (x = 0.11)

P21/c Fm3% m P21/c Fm3% m


Cell parameters a (Å)

b (Å)

c (Å)

Volume (Å3)

5.117 5.095 5.117 5.119

5.175 5.095 5.176 5.119

5.291 5.095 5.290 5.119

138.32 132.26 138.302 134.12

where b is the full width at half maximum (FWHM in radians), y is the Bragg angle, e is the lattice strain, D is the crystallite size and l is the wavelength of the X-rays. In general, a plot drawn between sin y and b cos y (straight line expected) allows one to extract the particle size and lattice strain from its intercept and slope, respectively. All samples show non-zero slopes indicating strain in both the monoclinic and cubic lattices. The variation in particle size and lattice strain with Dy concentration is plotted as shown in Fig. 4. In the pure monoclinic phase, while the particle size is B40 nm, the pure cubic phase shows particles of around 10 nm in size. Although the particle size is reduced by four times, the strain is doubled in the cubic phase showing a maximum at 9 at% of Dy. One may note that strain is found to be higher in the range of 5 to 9 at% when the mixed phase of monoclinic and cubic is observed. Transmission electron micrographs of the monoclinic (x = 0) and cubic phase (x = 0.11) of HfO2 show that the particles are slightly agglomerated and semi-spherical in shape (figure not shown). From the particle size distribution histogram shown in the insets of Fig. 5(a) and (c), the average particle size is found to be B36 and B10 nm for x = 0 and 0.11, respectively, which matches well with the estimated average particle size from the W–H plot. Fig. 5(a) and (c) give the distinguished diffraction rings observed from the selected area electron diffraction (SAED) patterns indexed as (1% 11), (1% 02), (1% 21) and (2% 13) of P21/c and (111), (200), (220) and (311) of Fm3% m. High resolution TEM shows the interplanar spacing (d) as B0.28 and B0.308 nm which corresponds to (200) and (111) of P21/c and Fm3% m, respectively as depicted in Fig. 5(b) and (d). It is known in the literature that bulk HfO2 exists in polymorphs such as the monoclinic phase at room temperature, and tetragonal and cubic phases at around B1700 1C and B2600 1C, respectively. It has been reported earlier that the high temperature phases of HfO2 can be stabilized at room temperature effectively by various dopants like Mn, Ce, Lu and Y etc. of different chemical valencies.13–15,35 For example, Dohcˇevic´-Mitrovic´ et al. demonstrated the complete stabilization of the pure cubic phase in HfO2 by doping 15% of trivalent yttrium by metathesis ´lvez-Barboza et al., through a simple modified synthesis.29 Ga


View Article Online


PCCP

Paper

Fig. 5 SAED pattern and high resolution TEM for x = 0 (a and b) and x = 0.11 (c and d), respectively. The insets of (a) and (c) show the particle size distribution histogram. The insets of (b) and (d) show the lattice planes.

sol–gel method, obtain the pure cubic phase of HfO2 nanoparticles after doping with 10% of trivalent Ce.45 Mendoza-Mendoza et al. also reported successful stabilization of the cubic phase by doping 10 mol% of Ce at a lower temperature through a molten LiCl/KCl chloride flux method.14 On the other hand, Gao et al. report a fully stabilized cubic phase after doping with multi-valent Mn in HfO2 nanoparticles by a conventional solid state reaction method.13 It has been shown that doping with 30% of Mn in HfO2 and sintering in an argon atmosphere, the transformation from Mn4+ to Mn2+ takes place, which stabilizes the cubic phase. However, sintering at 1000 1C in air, impurity phases of Mn2O3 and MnO are observed for more than 30% of Mn.13 Furthermore, such a high temperature cubic phase has not only been stabilized in HfO2 but also in a similar system like ZrO2. The key factors are the incorporation of dopants associated with the lattice expansion and consequent formation of oxygen vacancies. Therefore, the phase transformation from monoclinic to cubic in Dy doped HfO2 could be explained in the light of the stabilization process of the doped ZrO2 system.46 There is general agreement and it is well accepted in the literature that binary oxides are more prone to oxygen defects and to sample non-stoichiometry.47,48 Theoretically, Zhang et al., through first principles calculations predicted that a divalent or trivalent cation substituting Hf4+ ion induces oxygen vacancies (VO) in the host lattice to maintain the overall charge neutrality. These dopants lower the required formation energy to introduce an oxygen vacancy in the host lattice.49 The low temperature monoclinic phase possesses seven coordinated Hf4+ ions whereas the high temperature tetragonal or cubic phase is coordinated by eight Hf4+ ions. It is therefore expected in the present case that addition of Dy3+ ions in the HfO2 lattice may also generate oxygen vacancies which vary with an increase in Dy concentration.


Fig. 6

The XPS spectra of the O 1s region for x = 0 and x = 0.11.

The XPS spectra for the O 1s core level are shown in Fig. 6 for x = 0 and 0.11. All the peaks are calibrated with respect to the C 1s peak located at B284.6 eV. Due to the asymmetric shape of the spectra, we have deconvoluted the spectra into three peaks fitted with mixed Gaussian and Lorentzian functions using XPSPEAK version 4.1. The three curves centering at around 530.05, 531.1 and 533.0 eV are labeled as peak 1, 2 and 3, respectively. The peak with lower binding energy (BE) at 530.05 eV is associated with the oxygen atom in the O–Hf4+ bond.29 The higher BE peak at 531.1 eV belongs to O2 ions in oxygen deficient regions in HfO2 whereas the peak at 533.0 eV originates partly from the hydroxyl groups and from the chemisorbed oxygen lying on the surface of the samples.29,49,56 After deconvolution of each peak, due to the less pronounced peak corresponding to 533.0 eV in the spectra, we have ignored the area of the same peak. Comparing the peak area ratio of 2 and 1 (A2/A1), we observe that the A2/A1 ratio is almost double in the cubic phase than in the monoclinic one. We, therefore, confirm that due to higher oxygen vacancies in x = 0.11, the high temperature cubic phase of HfO2 is stabilized at room temperature. At the lowest concentration of Dy, HfO2 shows only the monoclinic phase due to the insufficient amount of oxygen vacancies. With further increase in the Dy concentration, the formation energy of oxygen vacancies is reduced significantly because of dipole formation between Dy and VO, introducing more oxygen vacancies in the lattice and hence, the monoclinic and cubic phases coexist for intermediate Dy concentrations


View Article Online


Paper

PCCP

(x = 0.05–0.07). An almost pure cubic phase is observed when the Dy concentration approaches 9 at%. Eventually, at x = 0.11, a nominal concentration of oxygen vacancies is present, which is sufficient to stabilize the cubic phase. Interestingly, one may notice that for dopants such as Y, Dy and Ce with ionic radii of 1.019 Å, 1.027 Å and 1.143 Å, respectively, i.e. more than Hf4+ and also more than 1.00 Å, a lower concentration of dopants is needed to stabilize the cubic phase at room temperature. On the other hand, a relatively much higher concentration of dopants is needed when the ionic radii of dopants like Mn2+ (0.96 Å) are slightly higher than Hf4+ but less than 1.00 Å. When the ionic radius (0.67 Å) of Mn4+ is quite less than the radius of Hf4+, no cubic phase is observed even upon increasing the sintering temperature upto 1400 1C in air.13 The above observations clearly indicate that the enhancement in lattice volume and nominal oxygen vacancies is needed to stabilize the cubic phase at room temperature with a low dopant concentration. 3.3

Magnetic properties

The magnetization (M) as a function of magnetic field (H) at 300 K for pure HfO2 (x = 0) calcined at 500, 700 and 900 1C is measured (figure not shown). The sample calcined at 500 1C exhibits a slim hysteresis loop at the lower magnetic field and at higher field, a linear increase in magnetization is observed. Upon increasing the calcination temperature to 900 1C, besides a linear increase in magnetization, the area under the hysteresis loop increases. The presence of a noticeable hysteresis loop demonstrates a weak ferromagnetic behavior at room temperature. While the coercivity (Hc) of B45 Oe remains the same in the sample calcined at 700 and 900 1C, the remanence (Mr) is enhanced by an order of magnitude in the sample calcined at 900 1C. The observed Hc is comparable to Hc in HfO2 reported by Coey et al.17 In addition, it is worth noting that the maximum magnetization (Mmax) decreases in sequence with increasing calcination temperature. The reduction in Mmax is attributed to a decrease in the surface to volume ratio. Due to an appreciable hysteresis loop observed in the sample calcined at 900 1C, we have measured M vs. H at 2 K and 35 K (figure not shown). It is evident that the area under the hysteresis loop dramatically increases with a decrease in the measurement temperature as expected. At 2 K, while Hc is B650 Oe, at room temperature, it drastically falls to B45 Oe which is an order magnitude less. This indicates that ferro or ferrimagnetic ordering exists even at room temperature, which becomes more prominent at low temperature. Such a behavior is consistent with earlier reports on the colloidal HfO2 nanorods, CaO nanopowders and CeO2 powders.51–53 While Mmax enhances by 30% at 2 K, Mr increases from B0.002 to B0.006 emu g1 with the temperature decreasing from 300 K to 2 K. Varying Dy concentration from 0 to 1 at% in HfO2, we have shown M vs. H at room temperature in Fig. 7(a). Magnetization increases linearly with increasing magnetic field indicating paramagnetic behavior irrespective of Dy concentration. A closer inspection at low magnetic field reveals that addition of even 1 at% of Dy leads to a collapse of the hysteresis loop. This indicates severe quenching of the weak ferro or ferrimagnetic


Fig. 7 M vs. H dependence measured at (a) 300 K and (b) 2 K for Hf1xDyxO2 (x = 0 and 0.1). The upper and lower inset in (b) shows a zoomed-in view at lower field for x = 0 and 0.01, respectively.

ordering observed in pure HfO2 by doping Dy. However, Mmax shows a proportional enhancement with increasing Dy concentration in HfO2. On the other hand, at 2 K we observe a wellsaturated hysteresis loop for Hf1xDyxO2 (x = 0 and 0.1) (Fig. 7(b)). It is interesting to note that Hc drastically reduces from B650 to B30 Oe by incorporating 1 at% of Dy and remains unchanged for the above concentration of Dy. An almost constant Hc value of B25 5 Oe is observed for all samples having more than 1 at% of Dy concentration. Saturation magnetization (Ms) increases from 0.14 to 16.2 emu g1 with increasing Dy concentration upto 11 at%. In order to confirm the ferro or ferrimagnetic nature in Dy doped HfO2, we have extrapolated the high temperature regime in the Curie–Weiss plot which gives a negative Curie– Weiss temperature. It suggests an antiferromagnetic (AFM) correlation at low temperature (see ESI,† Fig. S3). However, the presence of a hysteresis loop at low temperature reflects the finite magnetic moment. Combining both experimental observations, we conclude that non-compensated antiferromagnetically ordered spin results in a finite magnetic moment exhibiting a ferrimagnetic behaviour at low temperature. Further, susceptibility measurements with varying temperature shown in Fig. 8, demonstrate the transition from antiferro and/or ferrimagnetic to a paramagnetic phase for x = 0, 0.05, 0.07 and 0.11. We show that while the Neel temperature (TN) for x = 0 is found to be 45.5 K, for x = 0.05, 0.07 and 0.11, TN decreases to 4.6, 2.8 and 2.5 K, respectively. It has been shown that RTFM in non-magnetic oxides like TiO2, HfO2, ZnO, SnO2, Al2O3, MgO and CeO2 is an inherent


View Article Online


PCCP

Fig. 8 Variation of susceptibility with temperature in Hf1xDyxO2 (x = 0, 0.05, 0.07 and 0.11) nanoparticles.

characteristic property of the nanostructured materials. The origin of such ferromagnetic ordering is likely due to various defects such as oxygen vacancies, F-centers, defect complexes etc. and also depends on the crystallinity, shape and morphology of the particles.3,4,17,18,20–25,50,52,54–59,63 On the contrary to the above reports, Lin et al. report the absence of ferromagnetic ordering in ˇevic ´-Mitrovic´ et al. show HfO2 nanoclusters.60 In this context, Dohc that degradation of ferromagnetic ordering in Y doped HfO2 could be due to various (VO–YHf) defect complexes in different charge states such as (VO–YHf)0, (VO–YHf)+ and (VO–YHf)++. Among them, (VO–YHf)+ is the most stable. Since these defects are usually located very near to the valence band, electrons could not be moved to 5d states to form the spin-split defect band to establish ferromagnetism.29 On the other hand, in Pr doped CeO2 nanoparticles, Paunovic´ et al. believe that formation of Pr3+–VO–Pr3+ or Pr3+–VO–Ce3+ defect complexes are responsible for suppression of ferromagnetic ordering.61 According to predictions using the theoretical molecular mechanics method, generated oxygen vacancies mainly lie at the surface of nanoparticles that requires less energy to form an empty space corresponding to anions.62 Also, the oxygen vacancy at the surface is more efficient for mediating ferromagnetism than in the bulk.63 These oxygen vacancies can exist in different charge states by electron trapping such as VO, VO, VO, VO+ and VO++ forming a defect band near the conduction band.64 It is well known that two electrons trapped in an oxygen vacancy called F0 centers are able to develop antiferromagnetic ordering while singly occupied vacancies (F+ centers) establish long range ferromagnetic ordering that lies deep in the impurity band gap. Also, oxygen vacancies that are


Paper

unoccupied known as F++ centers do not contribute to ferromagnetic ordering.65,66 In the present case, XPS analysis revealed that pure HfO2 exhibiting monoclinic structure contains a considerable amount of oxygen vacancies at the surface of the nanoparticles. We believe that these induced oxygen vacancies mediate ferromagnetic ordering via an F+ center exchange mechanism. The addition of Dy3+ ions produces an oxygen vacancy (VO), which may bring two electrons in the lattice to compensate charge neutrality which increases with increase in Dy3+ concentration. In spite of a sufficient number of VO, it is surprising that a very small amount of substituted Dy3+ ions destroys the weak ferromagnetism observed in pure HfO2. In view of above reports and considering our experimental facts, it appears that quenching of ferromagnetic ordering in Dy doped HfO2 could be either due to partial compensation of F+ by F0 and/or F++ centers or because of defect states located far from the oxide conduction band minimum (CBM). However, it is obvious that substituted Dy3+ ions for Hf4+ ions in the lattice which are near to oxygen vacancies can create different kinds of unwanted complexes. In Dy doped HfO2, a similar defect state of Dy3+–VO–Dy3+ type can also be possibly formed which basically creates oxygen vacancies with no trapped electrons (F++ centers). The abundant formation of F++ centers negates long range ferromagnetic ordering and partially reduces the number of F+ centers, which hinders the appearance of ferromagnetism. On the other hand, defect complex formation can also displace the single electron energy levels of VO near or far from the valence band maximum that prevents spin-splitting of defect bands developing ferromagnetism.29,67 Therefore, the presence of Dy3+ ions in the HfO2 lattice can cause a reverse effect leading to the manifestation of the ferromagnetic ordering in Dy doped HfO2 systems. 3.4

Photoluminescence properties

Fig. 9 depicts the room temperature excitation spectra recorded under an emission wavelength of 490 nm for pure and Dy

Fig. 9 Room temperature excitation spectra of Hf1xDyxO2 (0 r x r 0.11) by monitoring the emission wavelength at 490 nm.


View Article Online


Paper

doped HfO2 nanoparticles. In sample x = 0, the spectrum consists of a broad band in the range of 240 to 340 nm centered at B296 nm and an another peak centered at B398 nm. The former peak is attributed to the host absorption band.35 The latter could originate from the most common defect state such as oxygen vacancies inherent in HfO2. In x = 0.01, along with a host absorption peak at B296 nm, the most intense peak is observed at 352 nm resulting from the electronic transition from the ground state, 6H15/2 to hypersensitive state, 6P7/2. In addition, sharper secondary peaks are observed in the range of 300–410 nm. These peaks detected at around 326, 338, 366, 383 and 393 nm are ascribed to different f–f transitions of Dy3+ ions from the ground state, such as 6H15/2 - 6P3/2, 6H15/2 - 4F5/2/4I9/2, 6 H15/2 - 6P5/2,6H15/2 - 4F7/2, and 6H15/2 - 4I13/2, respectively.68,69 The presence of the host absorption peak at B296 nm indicates a weak energy transfer from the host to Dy3+ activator ions. Further, the intensity of these excitation peaks decreases with increasing Dy concentration and finally reaches a minimum when x = 0.11. The appearance of sharp absorption peaks after doping 1 at% of Dy occurs at the expense of the host absorption band of the HfO2 lattice. The emission spectra of Hf1xDyxO2 (0 r x r 0.11) nanoparticles are collected after exciting with 352 nm which is the highest intensity excitation peak (Fig. 10). In x = 0, two emission peaks at 413 and 435 nm are superimposed and show a broad band in the range of 380–500 nm. These peaks are ascribed to the emissions of the host lattice. In addition to these peaks, for x = 0.01, two more emission peaks are observed at 490 and 577 nm falling in the blue and yellow regions, respectively. These peaks are the consequence of transitions from the metastable state to ground states such as 4F9/2 - 6H15/2 and 4 F9/2 - 6H13/2.69,70 With increasing Dy concentration, the emission peak intensity decreases gradually upto 7 at% and further at x = 0.09 and 0.11, the peak intensity is drastically reduced. Moreover, we observe the most efficient emission behavior when we dope only 1 at% of Dy in HfO2. It is well-known that emission from the electronic transition, 4F9/2 - 6H13/2 of the Dy3+ ion observed in the yellow (Y) region is hypersensitive to the crystal field environment. In a

Fig. 10 Room temperature emission spectra of Hf1xDyxO2 (0 r x r 0.11) at an excitation wavelength of 352 nm.


PCCP

site without inversion symmetry, usually, this transition becomes the dominant one.71 The other strongest emission peak corresponding to the magnetic transition, 4F9/2 - 6H15/2 detected in the blue (B) region bears the information of the presence of Dy3+ ions at a low symmetry site with inversion symmetry. Taking the intensity ratio of electronic to magnetic transition shown in eqn (2) gives an indication of the local symmetry of the Dy3+ ion site. YellowðYÞ I 4 F9=2 ! 6 H13=2 ¼ 4 BlueðBÞ I F9=2 ! 6 H15=2

(2)

The estimated Y–B ratio decreases with increasing Dy concentration as shown in Fig. 11(a) (blue line). One can notice that the Y–B ratio was found to be 0.95 0.03 at a lower Dy concentration and remains constant upto x = 0.07. However, at a higher Dy concentration, i.e. at x = 0.09 and 0.11, the Y–B ratio reduces more rapidly, suggesting that the intensity of the electronic transition decreases faster than that of the

Fig. 11 (a) Dependence of Y–B ratio and the relative intensity of the 490 nm emission peak in Hf1xDyxO2 (0 r x r 0.11) and (b) the CIE color space chromaticity diagram (a–f corresponds to x = 0–0.11) under an excitation wavelength of 352 nm.


View Article Online


PCCP

magnetic one.66 The reduction in the Y–B ratio above 7 at% of Dy concentration could be due to the changes in the local symmetry of Dy3+ ions in the HfO2 lattice. It is evident from the XRD analysis that while a high symmetry cubic phase obtained at x = 0.09 and 0.11, at x = 0.01–0.07, the low symmetry monoclinic phase remains present in the parent HfO2 lattice. The low symmetry in x = 0.01 and a high symmetry in x = 0.11 realized from XRD well supports the luminescence behavior as observed in the present case. Fig. 11(a) (red line) depicts the variation of the relative intensity of the emission peak at 490 nm as a function of Dy concentration in HfO2 nanoparticles. The emission intensity decreases rapidly with increasing Dy concentration in HfO2. Such a behavior clearly indicates the quenching of luminescence taking place due to the increase in Dy concentration. At a higher Dy concentration, this can be best explained in terms of the cross relaxation between two closely situated Dy3+–Dy3+ ions that includes mainly transition Dy3+ (4F9/2) + Dy3+ (6H15/2) - Dy3+ (6F3/2) + Dy3+ (6H11/2). Due to less separation between Dy3+ ions, the electrons move from a high energy level, 4F9/2 to 6F3/2 a lower energy level.70–72 The energy associated with such transition helps electrons sitting in the ground state of Dy3+ activator ions to get transferred to a higher energy level, 6H11/2. Such observation suggests that excellent photoluminescence properties can be perceived by doping a nominal amount of Dy (1 at%) in HfO2. Fig. 11(b) depicts the two dimensional (x, y) CIE color space chromaticity diagram of Dy doped HfO2 nanoparticles as a function of Dy concentration under an excitation wavelength of 352 nm. The CIE chromaticity coordinates are found to be (0.17, 0.12), (0.27, 0.27), (0.29, 0.29), (0.25, 0.24), (0.19, 0.13) and (0.19, 0.15) for x = 0, 0.01, 0.05, 0.07, 0.09 and 0.11, respectively. Among all samples, color coordinates of x = 0.01, 0.05 and 0.07 are close to that of cool white emission having the highest Y–B ratio. This indicates that by adjusting the Y–B ratio, one can also realize white light emission in the Dy doped HfO2 system for white LED (WLED) and display applications. We have proposed an energy band diagram to illustrate the different energetic transitions involved in the Dy doped HfO2 system as shown in Fig. 12. We have labeled the energy levels

Fig. 12 The proposed energy band diagram illustrating the charge transfer (CT) mechanism taking place in Hf1xDyxO2.


Paper

associated with the excitation and emission of Dy3+ ions (6H11/2, H13/2, 6H15/2, 4F9/2, 4I13/2, 4F7/2, 6P5/2, 6P7/2, 6P3/2 and 4I9/2) and the host lattice (VO, Oi0, Oi and Oi+). The host HfO2 (donor) partially absorbs the pumping energy and the electrons are promoted either to the host charge transfer band (CTB) and/or to the inherent defect states (VO) lying within the band gap of the host observed at B296 and 398 nm in the excitation spectrum. Consequently, the excited electrons relax from energy levels of the host CTB and from VO defect states to the ground state through deep defect levels lying near to the valence band of the host showing emissions at B413 and 435 nm and/or are partially transferred from the host CTB to different energy levels of the Dy3+ ion. The slight overlapping of the HfO2 broadband emission with the Dy doped HfO2 excitation band confirms the weak charge transfer from the host CTB to the Dy3+ activator (acceptor) ions. The similar emission profiles of the host suggest that the energy transfer process could be non-radiative.73 Such charge transfers from the host CTB to an activator ion (Sm3+ ions) have been shown by De la Rosa-Cruz et al. in Sm doped ZrO2.74 The electrons in the excited levels of the Dy3+ ion relax to different ground states such as 6H13/2 and 6H15/2 through a metastable state, 4F9/2 showing emissions at 490 and 577 nm. 6

4 Conclusions In the present work, HfO2 and Dy doped HfO2 nanoparticles were synthesized through a simple Pechini type sol–gel method followed by calcination at 900 1C in air. We have shown that the monoclinic phase of HfO2 nanoparticles transforms partially to the cubic phase after incorporating Dy of 5 at%. Further increasing the Dy (x = 0.07–0.11) concentration, the cubic phase is enhanced at the expense of the monoclinic one. At x = 0.11, the monoclinic phase completely transforms to the cubic phase. We have been able to stabilize the high temperature cubic phase of HfO2 at room temperature after incorporating Dy. Selected area electron diffraction patterns confirm the monoclinic and cubic phase in HfO2 and Hf0.89Dy0.11O2, respectively. High resolution TEM showed an interplanar spacing of B0.28 and B0.308 nm corresponding to (200) and (111) of P21/c and Fm3% m, respectively. A mechanism for stabilization of the high temperature cubic phase in Hf1xDyxO2 is proposed where the role of the substitution effect of Dy and the formation of oxygen vacancies was discussed. The magnetic properties of HfO2 nanoparticles revealed the presence of hysteresis at room temperature indicating ferromagnetism in contrast to the diamagnetic behavior in the bulk. Incorporation of even 1 at% of Dy quenched the ferromagnetic ordering. The disappearance of the ferromagnetic behavior in Dy doped HfO2 could be due to the formation of defect complex Dy3+–VO–Dy3+ which mainly creates F++ centers, highly unfavorable for long range ferromagnetic ordering. From the photoluminescence studies, we observed an excellent blue and yellow light emission by incorporating even 1 at% of Dy into the HfO2 lattice. The peak intensity of the blue and yellow emission decreases with increasing Dy concentration. The emissions of the Dy3+ ion are partially accompanied by a


View Article Online

Paper

non-radiative weak energy transfer from the host CTB to Dy3+ activator ions. Combining all the photoluminescence properties, we proposed an energy band diagram showing different transitions occurring in Hf1xDyxO2. Looking at the rich photoluminescence behavior of Hf0.99Dy0.01O2, it may be used as an efficient phosphor material for scintillator applications.


Acknowledgements Sandeep Kumar acknowledges UGC, New Delhi, India for providing a Rajiv Gandhi National Fellowship (RGNF). The authors also gratefully acknowledge Prof. N. V. Chalapathi Rao, Department of Geology, Banaras Hindu University, Varanasi, India for providing EPMA facility. Acknowledgement is also made to Central Instrument Facility Center (CIFC), Indian Institute of Technology (Banaras Hindu University), Varanasi, India for TEM and Magnetic measurements.

References 1 D. R. Uhlmann, T. Suratwala, K. Davidson, J. M. Boulton and G. Teowee, J. Non-Cryst. Solids, 1997, 218, 113–122. 2 P. A. Rodnyi and I. V. Khodyuk, Opt. Spectra, 2011, 111, 776–785. 3 P. Mohanty, D. Kabiraj, R. K. Mandal, P. K. Kulriya, A. S. K. Sinha and C. Rath, J. Magn. Magn. Mater., 2014, 355, 240–245. 4 P. Mohanty, V. P. Singh, N. C. Mishra, S. Ojha, D. Kanjilal and C. Rath, J. Phys. D: Appl. Phys., 2014, 47, 315001. 5 D. G. Wilk, R. M. Wallace and J. M. Anthony, J. Appl. Phys., 2001, 89, 5243–5275. 6 R. Ruh and V. A. Patel, J. Am. Ceram. Soc., 1973, 56, 606–607. 7 X. Zhao and D. Vanderbilt, Phys. Rev. B: Condens. Matter Mater. Phys., 2002, 65, 233106. 8 D. Fischer and A. Kersch, Appl. Phys. Lett., 2008, 92, 012908. 9 C. Adelmann, V. Sriramkumar, S. Van Elshocht, P. Lehnen, T. Conard and S. De Gendt, Appl. Phys. Lett., 2007, 91, 162902. 10 C. Wiemer, L. Lamagna, S. Baldovino, M. Perego, S. SchammChardon, P. E. Coulon, O. Salicio, G. Congedo, S. Spiga and M. Fanciulli, Appl. Phys. Lett., 2010, 96, 182901. 11 G. Dutta, Appl. Phys. Lett., 2009, 94, 012907. 12 C. K. Lee, E. Cho, H. S. Lee, C. S. Hwang and S. Han, Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 78, 012102. 13 L. Gao, L. Zhou, J. Feng, L. Bai, C. Li, Z. Liu, J. L. Soubeyroux and Y. Lu, Ceram. Int., 2012, 38, 2305–2311. 14 E. Mendoza-Mendoza, J. S. Quintero-Garcıá, B. A. Puenteńdez and L. A. Garcıá-Cerda, Urbina, O. S. Rodrı´guez-Ferna J. Alloys Compd., 2017, 692, 448–453. 15 B. Matovic´, D. Bucˇevac, M. Prekajski, V. Maksimovic´, D. Gautam, K. Yoshida and T. Yano, J. Eur. Ceram. Soc., 2012, 32, 1971–1976. 16 M. Venkatesan, C. B. Fitzgerald and J. M. Coey, Nature, 2004, 430, 630–633. 17 J. M. Coey, M. U. Venkatesan, P. Stamenov, C. B. Fitzgerald and L. S. Dorneles, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 72, 024450.


PCCP

18 J. M. Coey, Solid State Sci., 2005, 7, 660–667. 19 N. H. Hong, Phys. Status Solidi C, 2007, 4, 1270–1275. 20 M. D. Glinchuk, E. A. Eliseev, V. V. Khist and A. N. Morozovska, Thin Solid Films, 2013, 534, 685–692. ´, Phys. Rev. B: 21 N. H. Hong, J. Sakai, N. Poirot and V. Brize Condens. Matter Mater. Phys., 2006, 73, 132404. 22 R. Singh, J. Magn. Magn. Mater., 2013, 346, 58–73. 23 J. Ran and Z. Yan, J. Semicond., 2009, 30, 102002. 24 Q. Zhang, G. Chen and S. Yunoki, Solid State Sci., 2017, 252, 33–39. 25 H. Weng and J. Dong, Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 73, 132410. 26 C. D. Pemmaraju and S. Sanvito, Phys. Rev. Lett., 2005, 94, 217205. ń, M. C. Mun õz and J. Hafner, New J. Phys., 2008, 27 J. I. Beltra 10, 063031. 28 X. Liu, Y. Chen, L. Wang and D. L. Peng, J. Appl. Phys., 2013, 113, 0761021. 29 Z. D. Dohcˇevic´-Mitrovic´, N. Paunovic´, B. Matovic´, P. Osiceanu, ´ and M. Radovic´, Ceram. Int., 2015, 41, R. Scurtu, S. Asˇkrabic 6970–6977. 30 M. K. Sharma, D. K. Mishra, S. Ghosh, D. Kanjilal, P. Srivastava and R. Chatterjee, J. Appl. Phys., 2011, 110, 063902. ń, G. R. Gattorno, J. Guzma ń-Mendoza, M. Garcıá31 E. N. Cero ´lito and C. Falcony, Open J. Synth. Theory Appl., 2013, 2, Hipo 73–77. 32 K. Fiaczyk, A. J. Wojtowicz and E. Zych, J. Phys. Chem. C, 2015, 119, 5026–5032. 33 S. Lange, V. Kiisk, V. Reedo, M. Kirm, J. Aarik and I. Sildos, Opt. Mater., 2006, 28, 1238–1242. ˜ez, A. Pillonnet and 34 C. LeLuyer, M. Villanueva-Iban C. Dujardin, J. Phys. Chem. A, 2008, 112, 10152–10155. 35 A. Lauria, I. Villa, M. Fasoli, M. Niederberger and A. Vedda, ACS Nano, 2013, 7, 7041–7052. 36 A. D. Furasova, A. F. Fakhardo, V. A. Milichko, E. Tervoort, M. Niederberger and V. V. Vinogradov, Colloids Surf., B, 2017, 154, 21–26. 37 A. Wiatrowska and E. Zych, J. Phys. Chem. C, 2012, 116, 6409–6419. 38 C. Jayachandraiah, K. S. Kumar, G. Krishnaiah and N. M. Rao, J. Alloys Compd., 2015, 623, 248–254. 39 G. S. Wu, Y. L. Zhuang, Z. Q. Lin, X. Y. Yuan, T. Xie and L. D. Zhang, Phys. E, 2006, 31, 5–8. 40 P. B. Devaraja, D. N. Avadhani, H. Nagabhushana, S. C. Prashantha, S. C. Sharma, B. M. Nagabhushana, H. P. Nagaswarupa and B. D. Prasad, Mater. Charact., 2014, 97, 27–36. 41 S. Das, C. Y. Yang and C. H. Lu, J. Am. Ceram. Soc., 2013, 96, 1602–1609. 42 J. Laigo, F. Christien, R. Le Gall, F. Tancret and J. Furtado, Mater. Charact., 2008, 59, 1580–1586. 43 G. E. Lloyd, Mineral. Mag., 1987, 51, 3–19. 44 G. K. Williamson and W. H. Hall, Acta Metall., 1953, 1, 22–31. ´lvez-Barboza, L. A. Gonza ´lez, B. A. Puente-Urbina, 45 S. Ga E. M. Saucedo-Salazar and L. A. Garcıá-Cerda, J. Alloys Compd., 2015, 643, S62–S66.


View Article Online


PCCP

46 P. Li, I. W. Chen and J. E. Penner-Hahn, J. Am. Ceram. Soc., 1994, 77, 118–128. ń, S. Lany, S. V. Barabash and A. Zunger, 47 J. Osorio-Guille Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 75, 184421. 48 H. S. Baik, M. Kim, G. S. Park, S. A. Song, M. Varela, A. Franceschetti, S. T. Pantelides and S. J. Pennycook, Appl. Phys. Lett., 2004, 85, 672–674. 49 H. Zhang, B. Gao, S. Yu, L. Lai, L. Zeng, B. Sun, L. Liu, X. Liu, J. Lu, R. Han and J. Kang, Simul. Semicond. Proc. Dev., Int. Conf. IEEE, 2009, pp. 1–4. 50 V. P. Singh, S. B. Rai, H. Mishra and C. Rath, Dalton Trans., 2014, 43, 5309–5316. 51 E. Tirosh and G. Markovich, Adv. Mater., 2007, 19, 2608–2612. 52 Y. Liu, Z. Lockman, A. Aziz and J. MacManus-Driscoll, J. Phys.: Condens. Matter, 2008, 20, 165201. 53 D. Gao, J. Li, Z. Li, Z. Zhang, J. Zhang, H. Shi and D. Xue, J. Phys. Chem. C, 2010, 114, 11703–11707. 54 A. Sundaresan, R. Bhargavi, N. Rangarajan, U. Siddesh and C. N. R. Rao, Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 74, 161306. 55 A. Sundaresan and C. N. R. Rao, Nano Today, 2009, 4, 96–106. 56 D. Mishra, B. P. Mandal, R. Mukherjee, R. Naik, G. Lawes and B. Nadgorny, Appl. Phys. Lett., 2013, 102, 182404. 57 P. Sharma, A. Gupta, K. V. Rao, F. J. Owens, R. Sharma, R. Ahuja, J. O. Guillen, B. Johansson and G. A. Gehring, Nat. Mater., 2003, 2, 673–677. 58 X. L. Wang, Z. X. Dai and Z. Zeng, J. Phys.: Condens. Matter, 2008, 20, 045214. 59 G. Yang, D. Gao, J. Zhang, J. Zhang, Z. Shi and D. Xue, J. Phys. Chem. C, 2011, 115, 16814–16818. 60 K. Q. Lin, L. S. Wang, Z. W. Wang, R. T. Wen, Y. Chen and D. L. Peng, Eur. Phys. J. D, 2013, 2, 1–5.


Paper

ˇevic´-Mitrovic´, R. Scurtu, S. Asˇkrabic ´, 61 N. Paunovic´, Z. Dohc M. Prekajski, B. Matovic´ and Z. V. Popovic´, Nanoscale, 2012, 4, 5469–5476. 62 J. Conesa, Surf. Sci., 1995, 339, 337–352. 63 M. Y. Ge, H. Wang, E. Z. Liu, J. F. Liu, J. Z. Jiang, Y. K. Li, Z. A. Xu and H. Y. Li, Appl. Phys. Lett., 2008, 93, 062505. 64 D. M. Ramo, J. L. Gavartin, A. L. Shluger and G. Bersuker, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 75, 205336. 65 J. M. Coey, M. Venkatesan and C. B. Fitzgerald, Nat. Mater., 2005, 4, 173–179. 66 L. R. Shah, B. Ali, H. Zhu, W. G. Wang, Y. Q. Song, H. W. Zhang, S. I. Shah and J. Q. Xiao, J. Phys.: Condens. Matter, 2009, 21, 486004. 67 G. H. Chen, Z. F. Hou, X. G. Gong and Q. Li, J. Appl. Phys., 2008, 10, 074101. 68 G. Dominiak-Dzik, W. Ryba-Romanowski, M. N. Palatnikov, N. V. Sidorov and V. T. Kalinnikov, J. Mol. Struct., 2004, 704, 139–144. 69 X. Liu, W. Xiang, F. Chen, Z. Hu and W. Zhang, Mater. Res. Bull., 2013, 48, 281–285. 70 L. A. Diaz-Torres, E. De la Rosa, P. Salas, V. H. Romero and C. Angeles-Chavez, J. Solid State Chem., 2008, 181, 75–80. ¨, G. J. Zhou, S. W. Liu, D. Xu and 71 F. Gu, S. F. Wang, M. K. Lu D. R. Yuan, Chem. Phys. Lett., 2003, 380, 185–189. 72 M. Shang, D. Geng, X. Kang, D. Yang, Y. Zhang and J. Lin, Inorg. Chem., 2012, 51, 11106–11116. 73 E. De la Rosa-Cruz, L. A. Diaz-Torres, P. Salas, V. M. Castano and J. M. Hernandez, J. Phys. D: Appl. Phys., 2001, 34, L83. 74 E. De la Rosa-Cruz, L. A. Diaz-Torres, P. Salas, R. A. Rodrıguez, G. A. Kumar, M. A. Meneses, J. F. Mosino, ńdez and O. Barbosa-Garcıa, J. Appl. Phys., J. M. Herna 2003, 94, 3509–3515.
