S. K. Alla,
R. K. Mandal and
N. K. Prasad*
Department of Metallurgical Engineering, IIT (BHU), Varanasi – 221005, India. E-mail: nandkp.met@iitbhu.ac.in; Fax: +91-5422369478; Tel: +91-5422369346 Tel: +91-9956629843
First published on 25th October 2016
Nanocrystalline MgxCe1−xO2 (x = 0.01, 0.03, and 0.05) particles with near uniform size were synthesized by microwave refluxing method. The effects of Mg doping on the optical and magnetic properties of CeO2 are systematically studied. Particle size of all samples were demonstrated by XRD, HRTEM and Raman spectroscopy analyses. Mg doping effect on defect formation was investigated by Raman, UV-Vis and PL spectroscopy studies. Our results show that the increase of oxygen vacancies or defects was induced by Mg doping but not by particle size variation. In addition, a decrease of saturation magnetization was observed with increased Mg content. This can be due to the formation of Mg dopant complexes around oxygen vacancies which leads to formation of F2+ or F0 centers and thereby destroying the long range ferromagnetic ordering.
The possibility of mixed oxidation states of Ce and wide bandgap of CeO2 have shown strong absorbance in the UV region. Such an absorption was found to be a function of particles size or type of dopant.13,14 Even the optical bandgap of CeO2 becomes wider if the particle or crystallite size was reduced to nanometer range (1–3 nm). This has been attributed to the quantum size effect. For the particles with size between 8 to 50 nm, band gap increment was believed to be due to decrease of defect states, which are related to Ce3+ content. This causes elimination of some of the localized states within the bandgap.15–17 On the other hand, selective doping also gives rise to red shift of UV absorption indicating decrease in the bandgap. This was understood in terms of formation of defect states between the valence and conduction bands.18
Doping is one of the ways for tailoring the magnetic properties such as saturation magnetization (MS) and Curie temperature (TC) of CeO2. For example, doping of Fe-, Co-, Cu-, Ca-, Cr-, or Ni-ions significantly improves its saturation magnetization.19–24 Thin films of Co-doped CeO2 display comparatively large magnetic moment and exhibit TC around 875 K.20 The significant increase in the MS value was also observed with substitution of Ca- and Cr-ions.22,23 In these reports, authors have proposed that increased magnetization with doping of these elements was due to the formation of oxygen vacancy by the dopant that facilitates ferromagnetic interaction between these ions. In contrast, MS value is reported decreasing with increased doping concentration of Fe- or Pr- or Y-doped CeO2.25–27 The reason for decrement in MS value for these materials has been attributed to the formation of F2+ and F0 centers associated with increased oxygen vacancies. Further, more than 4 at% Ni doped CeO2 displayed a decrease in the MS value which suggests that enhancement of magnetization also depends on the level of concentration of dopant.24
Rationalizing the enhancement of magnetic properties by selective doping is less intuitive because the particle size as well as surface configurations also influence the magnetic properties of CeO2. In addition, optical properties of CeO2 were also found to be dependent on the size and nature of doping. Thus, to study the effect of doping on the optical and magnetic properties, where dopant facilitates change in size of the particles, becomes a challenging task. Moreover, some dopants causes the formation of metallic clusters or secondary phases which can also influence the RTFM.28 Keeping this in view, we have selected magnesium as dopant for cerium oxide and synthesized magnesium doped CeO2 to attain uniform particle size distribution. The effect of doping on the optical and magnetic properties of CeO2 were investigated.
Many synthesis protocols have been adopted in literature for making of undoped and doped CeO2. They refer to co-precipitation,29 sol–gel,30 hydrolysis,31 pulsed laser deposition,32 hydrothermal,14 and sonochemical process.33 These techniques are successfully utilized for doping of various elements into CeO2 and for producing desired shape and/or size of doped CeO2 nanostructures. However, most of these approaches are expensive, complex, time and energy consuming. Thus, relatively simple, low cost and single step microwave refluxing methods was utilized for the synthesis of Mg doped CeO2 nanostructures.
X-ray diffraction patterns of all samples were recorded using a Phillips X'pert Pro Advanced powder X-ray diffractometer with Cu Kα radiation (λ = 0.15431 nm). Crystallite size was estimated using Scherrer's equation after instrumental broadening corrections depending on the profile. The transmission electron microscopic (TEM) observations were carried out using TEM (FEI TECHNAI G2) with an accelerating voltage 200 kV. Reinshaw micro-Raman spectroscope with a 514.5 nm Ar+ laser source utilized for recording Raman spectra. Cary 60 UV-Vis spectrometer (Agilent Technologies) was used to collect UV-Vis absorption spectra. Photoluminescence observations were carried out on LS-45 Fluorescence spectrometer (Perkin Elmer) under an excitation wavelength of 330 nm. Magnetic measurements were taken on SQUID at 1 tesla (MPMS-XL, Quantum Design). X-ray photoelectron spectra (XPS) were recorded using PHI 5000 Versaprobe II photoelectron spectrometer (ULVAC-PHI) using Al Kα X-ray beam.
![]() | ||
Fig. 1 (a) XRD pattern of S1, S2 and S3 samples (b) variation of lattice parameter with Mg concentration. |
TEM bright field image of undoped CeO2 sample is shown in Fig. 2a and HRTEM image with d-spacing of ∼0.31 nm which corresponds to its (111) plane. The FFT pattern of a selected encircled region of Fig. 2b displays two fold symmetry (inset in Fig. 2b). Fig. 3a displays TEM bright filed image of sample S1. The inset polycrystalline diffraction rings could be indexed as conforming to reflections of FCC structure. Fig. 3b presents the high resolution TEM image with d-spacing of ∼0.31 nm. This d-spacing corresponds to interplanar spacing of (111) plane of CeO2. TEM bright field image of S3 is shown in Fig. 4a. The inset in Fig. 4a displays corresponding polycrystalline ring pattern. HRTEM image with d-spacing of ∼0.31 nm is also given in Fig. 4b akin to Fig. 3b. The inset of Fig. 4b gives FFT of encircled region. Structure thus remains invariant with respect to the level of doping.34
![]() | ||
Fig. 2 (a) TEM-BF image and corresponding SAED pattern (inset picture) (b) HRTEM image of undoped CeO2 sample. |
![]() | ||
Fig. 3 (a) TEM-BF image and corresponding SAED pattern (inset picture) (b) HRTEM image of S1 sample. |
![]() | ||
Fig. 4 (a) TEM-BF image and corresponding SAED pattern (inset picture) (b) HRTEM image of S3 sample. |
SEM micrographs and corresponding elemental mapping of S1, S2 and S3 samples are presented in Fig. 5. EDX analysis confirms the presence of Mg, Ce and O elements in all the samples. The atomic percentage of Mg in S1, S2 and S3 samples were 0.73, 0.93 and 1.91% respectively.
In order to identify the surface species and their ionic states, X-ray photoelectron spectra were recorded for sample S2. Ce 3d, Mg 1s and O 1s regions of spectra of this sample are given in Fig. 6. The deconvolution of Ce 3d spectrum (Fig. 5a) consists of mainly four major peaks. The doublet peaks located at binding energies of 882.3 and 900.9 eV correspond to Ce3+ 3d3/2 and 3d5/2 respectively. The other doublet peaks were observed at 898.2 and 916.6 eV. These are mainly attributed to Ce4+ 3d5/2 and 3d3/2 respectively. The shake-up peaks at around 907.8 eV on Ce3+ 3d3/2 and 889.5 and 885.8 eV on the Ce3+ 3d5/2 were also observed.35 These observations provide evidence of Ce in Ce3+ and Ce4+ states in sample S2. Mg 1s core level spectrum (Fig. 6b) shows a single peak at 1303.9 eV corresponds to Mg2+ species.36 Fig. 6c depicts the deconvolution of O 1s core level spectrum having mainly two peaks at 528.7 and 531 eV. The peak at 528.7 eV corresponds to lattice oxygen and other peak at 531 eV originates either from adsorbed oxygen on the surface or from oxygen vacancies.26,37,38
![]() | ||
Fig. 6 XPS spectra of (a) Ce 3d core level spectra (b) Mg 1s core level spectra (c) O 1s core level spectra of S2 sample. |
Direct evidence for formation of defect states and variations, is provided by Raman spectroscopy for S1, S2 and S3 samples (Fig. 7). The Raman active peak at 460 cm−1 is the characteristic peak for fluorite structure oxides. It has triply degenerate Raman active modes (F2g) and could be due to a symmetrical breathing mode of the oxygen atoms around cerium ions. The shift F2g mode toward lower energy and a mild variation in broadening was noticed in S2 and S3 samples. This can be due to presence of phonon relaxation and lattice strain for submicroscopic sized particles.39 On the other hand, broadening may be due to size and thus is an indicative of variation of particle size with doping.40 But, researchers argue that the broadening of Raman active bands is not only due to particle size effect but also concentration of defects in the particles because this mode is very sensitive to disorder in oxygen sub-lattice.41,42 Thus, the small variation of broadening observed in the present case might be induced due to increase of defect density in the samples.
In addition to F2g mode, a weak Raman band appeared at 570 cm−1 in all samples and their Lorentzian peak profile fitting are shown in inset Fig. 7. The peak at 570 cm−1 is due to the formation of defects or oxygen vacancy by doping.43 The band becomes prominent with increased dopant concentration which suggests that additional oxygen vacancies are created with doping of Mg2+ ions. In case of undoped cerium oxide, oxygen vacancies are created by the reduction of Ce4+ to Ce3+ when size is reduced to nanometer scale. As the size of the particles for all samples was similar so the incorporation of Mg2+ ion in CeO2 lattice seems to be responsible for creation of oxygen vacancies.
Fig. 8 presents the UV-absorption spectra of S1, S2 and S3 samples. The absorption wavelength for S1, S2 and S3 are at 297, 298 and 300 nm respectively. The red shift of absorption peaks shown in Fig. 8 indicates a decrease in the bandgap with increased Mg2+ doping. Since UV-absorption of cerium oxide resulted from charge transfer between O2− (2p) to Ce4+ (4f) bands, this decrease in bandgap could be due to formation of defect state between valence and conduction band. The bandgap was estimated by following equation18
Eg (eV) = 1240/λabs. edge |
The λabs. edge values of S1, S2 and S3 samples were 340, 343 and 347 nm respectively and corresponding bandgap values were 3.65, 3.62 and 3.57 eV respectively.
As mentioned above, the defect states are formed between valence and conduction band leading to continuous decrease of bandgap with increased Mg concentrations. Oxygen vacancies with doping may also be playing a role in this decrease. Since defects or oxygen vacancies were induced by formation of Ce3+ states, it is clear that incorporation of Mg2+ in to CeO2 lattice enables formation of Ce3+ states.13
Further, to understand the nature of defects and their corresponding transitions, photoluminescence spectroscopy analysis was carried out. Fig. 9 displays room temperature PL spectra of S1, S2 and S3 samples under an excitation wavelength of 330 nm. All spectra mainly consist of emission bands at 396, 420, 480 and 530 nm. The bands with peaks in the range of 350 to 500 nm were attributed to the presence of defect states including oxygen vacancies between Ce4+ conduction band and O2− valence band.44 The intensity of these emission bands were decreasing with increased Mg2+ concentration. This consistent decrease in the intensity with Mg2+ doping depicted the formation of defect states between valence and conduction band. These states dissipate the energy of radiation such that the efficiency of overall transition gets reduced.45 For S3 sample, the emission peaks was very weak which indicated that the concentration of defects or oxygen vacancies were much higher.46 Based on the results of Raman, UV and PL spectroscopy studies, it is well understood that incorporation Mg2+ in to CeO2 lattice can solely introduce defects or oxygen vacancies and these oxygen vacancies increase with dopant concentration.
At room temperature and an applied field up to ±1 T, magnetization (M) vs. applied field (H) curves of S1, S2 and S3 are presented in Fig. 10a. The paramagnetic component was present in all the samples and this component became prominent with increased dopant concentration (Fig. 10a). After subtraction of linear background for all the samples, a clear hysteresis loop was observed (Fig. 10b). The saturation magnetization (MS) values were taken from Fig. 10b and they are listed in Table 1. The MS values are comparable with undoped CeO2, ZnO and Al2O3 systems.5,6,47–49 Such a low values of MS indicate the weak ferromagnetic behavior for S1, S2 and S3 samples. The MS values for, 11% Y-, 15% Pr-, Fe and Eu co-doped CeO2 samples were 36 × 10−4, 1.8 × 10−4 and 1.0 × 10−4 emu g−1.26,27,39 Further, the MS values decreased with increased Mg concentration.
![]() | ||
Fig. 10 Room temperature M vs. H curves for S1, S2 and S3 samples (a) before and (b) after subtraction of linear background. |
Sample | Lattice parameter (Å) | Crystallite size (nm) | Lattice strain | Saturation magnetization (emu g−1) |
---|---|---|---|---|
S1 | 5.429 | ∼8 | ∼0.01 | 0.0011 |
S2 | 5.428 | ∼8 | ∼0.01 | 0.0007 |
S3 | 5.427 | ∼8 | ∼0.01 | 0.0003 |
It is well known that stoichiometric cerium oxide exhibits diamagnetic nature because of its Ce4+ ion having completely filled shell. But when size of the cerium oxide particles dropdown to nanometer ranges it shows RTFM nature. This is due to reduction Ce4+ ion into Ce3+ which leads to non-stoichiometry such that defects are introduced in to the cerium oxide lattice in order to compensate the charges. But, in the case of doping, magnetic properties were related to the formation of metallic clusters, secondary phases or development of defects by doping. In our case, prevalence of metallic clusters and secondary phases are ruled out by XRD, HRTEM and Raman spectroscopy analyses. Thus, magnetic properties of CeO2 in our case arise owing to doping effect only. Essentially, the RTFM nature observed for all samples could be due to their nanometric size. In addition, variation of particle size with doping is known to affect the magnetic properties of CeO2. But, XRD and TEM analyses confirm the invariance of particle size. Furthermore, increase of defects or oxygen vacancies can also enhance the magnetic properties. Although, Raman and PL spectroscopy analyses gave evidence of enhanced oxygen vacancies with increased doping. However, the MS value found to be decreasing continuously with increased dopant concentration in the present study.
Similar behavior was also observed in Y3+ doped CeO2 nanoparticles.27 The MS values were of the order of 10−4 emu g−1 for 3 and 11% of Y3+ doped samples which decreased with increased dopant concentration. The particle size for both the samples were nearly same and oxygen vacancies increased at higher doping. The authors have reported that at lower concentration, Y3+ ions uniformly distribute throughout the particles whereas at higher concentration Y3+ ions aggregated on the surface of the particles and decreased the magnetization value. Thus, it can be presumed that in case of Mg doped CeO2, Mg2+ ions might have segregated on the surface of the particles and diminished the magnetization value. In addition, the Mg substitutional defects cause the formation of shallow acceptor levels (hole states) near top of O2− valence band. In order to compensate the defects, an electron from the neighboring oxygen vacancy may transfer to these holes. As a consequence, it may lead to formation of oxygen vacancy centered defect complexes. It was assumed that possible defect complexes can be Mg2+–Vo–Ce3+ and Mg2+–Vo–Mg2+ type in the host lattice at even small amount of Mg doping. Similar type of behavior was also reported in Pr doped CeO2 nanoparticles.26 For Mg2+–Vo–Ce3+ complex, localized electron on the vacancy leads to creation of F+ center and establish ferromagnetic interaction in the samples. This might have caused weak ferromagnetic behavior in Mg doped CeO2 samples. The decrease in magnetization may be due to the formation of Mg2+–Vo–Mg2+ complex. This complex may favor formation of F2+ or F0 centers that are detrimental to long range ferromagnetic ordering. The formation of F0 center by trapping of two electrons around a vacancy whereas absence of electrons around vacancy creates F2+ center which promotes antiferromagnetic exchange interaction. Therefore, the decrement of MS values with increased Mg doping can be due to the formation of varieties of complexes. Since the magnetization values of Mg doped CeO2 are much smaller and decreased with increasing Mg content. Thus, these samples may be predominantly paramagnetic rather than having RTFM nature. Recently, Coey et al. proposed that this saturation magnetization is related to giant orbital paramagnetism which arises due to collective response of coherent domains in applied magnetic field.11 However, to understand the exact reason for the decrease of magnetization of CeO2 with increased Mg content and presence of dominant paramagnetic component need further investigations.
This journal is © The Royal Society of Chemistry 2016 |