Structural and magnetic properties of ZnO–CoFe2O4 nanocomposites

Journal of Magnetism and Magnetic Materials 389 (2015) 27–33

Contents lists available at ScienceDirect

Journal of Magnetism and Magnetic Materials journal homepage: www.elsevier.com/locate/jmmm

Structural and magnetic properties of ZnO–CoFe2O4 nanocomposites T.J. Castro a,n, S.W. da Silva a, F. Nakagomi a, N.S. Moura b, A. Franco Jrb, P.C. Morais a,c a b c

Instituto de Física, Universidade de Brasília, C.P. 04455, Brasília, DF 70910-900, Brazil Instituto de Física, Universidade Federal de Goiás, Goiânia, GO 74001-970, Brazil Huazhong University of Science and Technology, School of Automation, Wuhan 430074, China

art ic l e i nf o

a b s t r a c t

Article history: Received 20 March 2014 Received in revised form 5 April 2015 Accepted 9 April 2015 Available online 11 April 2015

Nanocomposites with chemical composition (CoFe2O4)x þ (ZnO)1–x were prepared by the combustion reaction method. X-ray diffraction data reveal the presence of only two crystalline nanosized phases with average diameters of  55 nm (ZnO) and  60 nm (CoFe2O4). The Rietveld method revealed that the molar ratio between the two phases shifted their nominal values from x ¼0.005, 0.050, 0.100, 0.200 and 0.400 up to x ¼ 0.030, 0.090, 0.170, 0.270 and 0.570, respectively. It was also found that the change of the lattice parameters of the wurtzite and spinel phases is related with the exchange of metallic ions between the two crystal structures. Our findings show that the ion exchange has direct influence on the vibrational and magnetic properties of the as-produced nanocomposites. Evidences supporting our findings are the appearance of vibrational modes induced by disorder and the decrease of the saturation magnetization. & 2015 Elsevier B.V. All rights reserved.

Keywords: Nanocomposite Raman scattering Strain Cobalt ferrite Zinc oxide

1. Introduction In the last few years an increasing attention has been devoted to a new class of materials called nanocomposites, comprising multicomponent hybrid nanostructures containing two or more nanosized components assembled in a controlled way [1,2]. Due to the synergistic properties usually induced by the intimate contact and interaction between different components, nanocomposite materials can achieve enhanced properties and provide novel functionalities not available in the single-phased nanostructures. These advantages make hybrid nanostructures one of the most promising candidates for the exploration of new applications. Hybrid semiconducting and magnetic nanostructures are currently under intense investigation owing to their potential application in optoelectronic, spintronic, nanoelectronic, and biomedicine [3,4]. Zinc oxide (ZnO) is a typical metal-oxide semiconductor with a wide energy band gap (Eg ¼ 3.37 eV), already integrated in many optical, electronic, and acoustic devices [5,6]. On the other hand, spinel ferrites have remarkable properties, such as high electrical resistivity, enhanced mechanical hardness, and superior chemical stability, among others [7]. Therefore, combinations of CoFe2O4 and ZnO nanophases to form nanocomposites are expected to provide new materials with multiple properties. Studies involving the combination of such materials may already been found in the n

Corresponding author. E-mail address: [email protected] (T.J. Castro).

http://dx.doi.org/10.1016/j.jmmm.2015.04.036 0304-8853/& 2015 Elsevier B.V. All rights reserved.

literature [8,9]. In this study we report on the synthesis of (CoFe2O4)x þ(ZnO)1–x nanocomposites revealing both interesting optical and magnetic properties. The nanocomposites were investigated by means of scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDS), X-ray diffraction (XRD), Raman spectroscopy, Mössbauer spectroscopy, and vibration sample magnetometry (VSM).

2. Experimental procedure Initially, nanoparticulated powders of both zinc oxide (ZnO) and cobalt ferrite (CoFe2O4) were prepared by the combustion reaction method. All reagents used for the samples’ preparation, namely zinc nitrate Zn(NO3)2  6H2O, iron nitrate Fe(NO3)3  9H2O, cobalt nitrate Co(NO3)2  6H2O and urea CO(NH2)2 as fuel, are of analytical grade and were manipulated in air without Nitrogen or inert gas protection. Then, for each selected relative content, zinc oxide and cobalt ferrite powders were dispersed in 100 mL of 2-propanol. The mixture was poured into a three quarter-liter polyethylene bottle and ball milled (with yittria stabilized zirconia media) for 5 h. Next, the resulting mixture was dried in an oven at 70 °C for 24 h. Finally, (CoFe2O4)x þ(ZnO)1–x nanocomposites samples with nominal molar ratio varying from 0.5 to 40% were prepared by heating all molar ratio combinations in a resistive furnace under air atmosphere at a heating rate of 3 °C/min up to 1100 °C, soaked for 2 h, then slowly cooled down to room temperature.

28

T.J. Castro et al. / Journal of Magnetism and Magnetic Materials 389 (2015) 27–33

The as-produced powder nanocomposites were characterized by XRD using a Rigaku diffractometer (model Ultima IV) equipped with CuKα radiation (λ ¼1.5418 Å), in a wide range of Bragg angles (20° o2θ o120°) with a scanning rate of 2°/min, at room temperature. The Raman system used to record the spectra of the

nanocomposite samples was a commercial triple spectrometer (Jobin Yvon Model T64000) equipped with a charge-coupled device (CCD) detector. The 514 nm line from an Argon-ion laser was used to illuminate the samples at an optical power around 0.2 mW. All Raman measurements were performed at room

Fig. 1. (a) 0.005, (b) 0.050, (c) 0.100, (d) 0.200 and (e) 0.400 SEM micrographs. The EDS spectrum of x ¼0.400 is shown in the bottom of the figure.

T.J. Castro et al. / Journal of Magnetism and Magnetic Materials 389 (2015) 27–33

29

temperature. The nanocomposites’ morphology and chemical composition were investigated using a SEM microscope (SEM) (JEOL JEM 2100) equipped with an EDS accessory. For these measurements, each sample was dispersed in 3 mL of n-propanol and sonicated for 5 min. Droplets of this dispersion were placed over a Copper grid coated with parlodion and Gold films and dried under air. Room-temperature Mössbauer spectra were recorded using a 57 Co source in a Rh matrix whereas a thin Fe-sheet was used for calibration. Room-temperature magnetic properties (hysteresis cycles) were assessed using a VSM system (ADE Magnetics, model EV9, Westwood MA, USA) with applied magnetic fields up to 72.0 kOe.

3. Results and discussion Typical SEM images of the nanocomposites (CoFe2O4)x þ (ZnO)1–x are shown in Fig. 1. The obtained SEM micrographs revealed the presence of large aggregates, probably due TO the annealing process. In addition, it was verified that the size of the aggregates increases while increasing the CoFe2O4 content. Compositional analyses obtained by EDS as well as the nominal compositional values are shown in Table 1. The difference observed between the nominal and the obtained composition can be explained taking into account the overlap of the Kβ line from Iron (  7057 eV) with the Kα line from Cobalt (  6915 eV). As the energy difference between these two emission lines is smaller than the resolution of the equipment there is a possibility of errors while identifying and determining the concentration of these two elements. However, the EDS spectra reveal the presence of Co, Fe, Zn, O, and Au elements only, thus suggesting no sample contamination. Except for the peak related to Gold, due to sample preparation for SEM, there are no additional elements observed. The XRD patterns of pure ZnO, (CoFe2O4)x þ(ZnO)1–x nanocomposite (x¼0.200) and pure CoFe2O4 are presented in Fig. 2(a). Fig. 2(b) shows the expanded region around the (311) peak of the CoFe2O4 phase and around the (002) and (101) peaks of the ZnO phase. All observed XRD peaks are well indexed to the ZnO wurtzite structure (JCPDS Card no. 79-2205) and the CoFe2O4 spinel structure (JCPDS card No. 22-1086). The absence of extra XRD reflections in the diffraction patterns confirms the high purity of all phases. Moreover, it can be clearly seen from Fig. 2(b) that the relative intensities as well as the angular positions of the XRD peaks associated to both single-phased materials are systematically changing as the ferrite’s content increases. In order to deeper the understanding of the XRD data all spectra were analyzed using the Rietveld method via GSAS-EXPGUI programs [10– 12]. The Rietveld analyses revealed that the mean particle size is nearly the same for all samples and for the two single-phased materials; being  55 nm and  60 nm for ZnO and CoFe2O4, respectively. Analysis of the XRD patterns using the Rietveld method showed that the actual molar rate of the cubic spinel phase is higher than the nominal content. It was found that the actual Table 1 Nominal and measured (EDS) chemical compositions of the (CoFe2O4)x þ(ZnO)1–x samples. Sample

x ¼0.005 x ¼0.050 x ¼0.100 x ¼0.200 x ¼0.400

Nominal (Atoms %)

EDS (Atoms %)

Fe

Co

Zn

Fe

Co

Zn

1.0 9.0 16.6 28.6 44.4

0.5 4.5 8.3 14.3 22.2

98.5 86.5 75.0 57.1 33.3

1.8 7 0.2 8.8 7 0.4 11.2 7 0.2 20.7 7 0.5 40.9 7 0.6

1.4 70.2 4.6 70.2 7.6 70.4 13.0 70.5 22.3 70.7

977 2 877 1 817 1 66 7 1 377 1

Fig. 2. (a) XRD patterns from pure ZnO, (CoFe2O4)0.20 þ(ZnO)0.80 and pure CoFe2O4 samples. (b) XRD patterns of all samples showing the region around 35°.

molar ratio between the two phases are x ¼0.030, 0.090, 0.170, 0.270 and 0.570 instead of their respective nominal values of x¼ 0.005, 0.050, 0.100, 0.200, and 0.400. In addition, it was found that the lattice parameter of the cubic spinel phase decreases linearly as its content (x) increases up to x¼ 0.400 (see inset in Fig. 3(a)). We found that at x¼ 0.400 the lattice parameter (a ¼8.392 Å) of the cubic ferrite phase is very close to the value reported for pure CoFe2O4 (JCPDS Card no. 22-1086, a ¼8.391 Å). We claim that variations of the lattice parameter can be related to the exchange of metallic ions among spinel and wurtzite phases. While annealing at 1100 °C metallic cations may have enough energy to diffuse and react with Oxygen ions of the neighboring nanoparticles. For small x values (low cobalt ferrite content) a large amount of Zinc ions from the ZnO phase is available for exchange and, as a result, a given amount of Iron and Cobalt ions from the cubic ferrite phase are likely replaced by Zn ions. In this process the cubic ferrite phase becomes non-stoichiometric, namely Zny(CoFe2)1–yO4. Thus, the interfacial substitution of Fe3 þ (0.63 Å: A site and 0.78 Å: B site) and Co2 þ ions (0.72 Å: A site and 0.79 Å: B site) of the cubic spinel phase by Zn2 þ ions (0.74 Å: A site and 0.88 Å: B site) explains the observed increase of the lattice parameter as the x-content decreases [13]. Variation in the ZnO-phase’s lattice parameter was also observed. In this case, we found the lattice parameters (a and c) of the wurtzite phase showing the opposite behavior, i.e. a increases while c decreases as the ferrite content (x) increases (see Fig. 3 (b) and (c)) in the nanocomposite samples. Similarly to the observed behavior in the spinel phase the replacement of Zn ions by Fe and Co ions in the wurtzite (ZnO) crystalline structure is responsible for the lattice parameters (a and c) variation. Furthermore, for x ¼0.400, it was found that the values of the lattice parameters a and c diverge from the smooth trend observed at smaller x values. Though the explanation of this abrupt change in the lattice parameters is not yet fully available we claim that it can be associated with the relaxation of the crystalline lattice of the ZnO wurtzite phase at x¼ 0.400. The Fig. 3(d) shows the u parameter decreasing linearly as the ferrite content increases, up to x¼ 0.200. In the wurtzite structure the u parameter describes, along the c axis, the relative position of the anion sublattice with respect to the cation sublattice. For all compounds within the wurtzite structure the u parameter is related to the c/a ratio via uc /a = 3/8 [14,15]. Indeed, we found that this relationship was fulfilled for x¼ 0 only. However, the Rietveld data showed that the above-mentioned ideal correlation between u and c/a in the wurtzite structure is affected by the exchange of Zn ions by Fe and Co ions. Probably, the reduction of u is related to the ZnO-phase

30

T.J. Castro et al. / Journal of Magnetism and Magnetic Materials 389 (2015) 27–33

Fig. 3. Lattice parameters of cubic ferrite (a) and wurtzite phases ((b) and (c)) as a function of the nominal ferrite content (x). (d) u parameter as a function of the nominal ferrite content (x). The dashed lines are only guide to the eyes.

Table 2 Strain in the c-direction, E2high Raman shift calculated (ω(E2high)calc) and obtained experimentally (ω(E2high)exp).

εcc (10  4) ω (E2high)calc. ω (E2high)exp.

Fig. 4. (i) Raman spectra of the pure ZnO phase, (ii)–(vi) (CoFe2O4)x þ(ZnO)1–x nanocomposites with different ferrite content (x), and (vii) pure CoFe2O4 phase.

density decrease, as shown by the Rietveld data analysis. Evidence of ion exchange between the two crystalline phases may be also observed from the Raman spectra of pure ZnO (i), [(CoFe2O4)x þ(ZnO)1–x] nanocomposites with different ferrite content (x) (ii)–(vi) and pure CoFe2O4 (vii) shown in Fig. 4. Typical

x ¼0.000

x¼ 0.005

x¼ 0.050

x ¼0.100

x¼ 0.200

0.00 439.0 439.8

 0.461 438.9 438.4

 2.65 439.0 438.2

 0.768 438.9 436.7

 2.34 438.8 437.2

Raman active modes for the ZnO can be seen in all recorded spectra in the lower ferrite content range ( x r0.200). The most prominent modes are: E2high–E2low at 333 cm  1, A1(TO) at 382 cm  1, and E2high at 438 cm  1 [16]. The broad band appearing in the 500-600 cm  1 spectral region, showing a strong Raman enhancement with increasing x (very weak in pure stoichiometric ZnO) have been identified to be the silent 2B1 modes (545 cm  1) and E1(LO) phonons (585 cm  1) of wurtzite ZnO [17,18]. Similar behavior was reported for Co- and Mn-doped ZnO [19]. The increasing intensity of the vibrational band in the 500–600 cm  1 spectral region is more likely associated with the breakdown of crystal’s translational symmetry due to the exchange of Zn ions by Fe and Co ions in the wurtzite structure. For x¼ 0.400, the recorded Raman spectrum presents very different characteristics from those observed at smaller x values. In this case (x¼ 0.400) the Raman spectrum shows features very similar to cobalt ferrite. Nevertheless, the Raman modes were found shifted with respect to the pure cobalt ferrite phase (see Fig. 3 (vii)). Note that cobalt ferrite has cubic spinel structure, belonging to the space group Oh7 (Fd3m), in which five active Raman modes are expected to be observed (A1g þEg þ3T2g) around 220, 310, 470, 630, and 690 cm  1. In the case of inverse spinel ferrite (e.g. CoFe2O4, MgFe2O4, NiFe2O4, and Fe3O4) the literature reports that the Raman modes in the 650–710 cm  1 region have A1g symmetry and are related to the tetrahedral sub-lattice whereas the Raman modes peaking below of 650 cm  1 are related to the octahedral sub-lattice [20]. On the other hand, in the case of normal spinel ferrite (e.g. ZnFe2O4 and MnFe2O4), the Raman mode with A1g symmetry has been observed in the 600–650 cm  1 region [21]. Therefore, the observed enhanced intensity of the Raman peak at 650 cm  1 indicates that the Zn ions from the ZnO structure have entered substitutionally into the CoFe2O4 structure. It has been repeatedly reported that strain affects electronic states and lattice vibrations in a crystal [22]. Therefore, it is

T.J. Castro et al. / Journal of Magnetism and Magnetic Materials 389 (2015) 27–33

31

Fig. 5. (a) E2high mode fitting using the PCM approach (shaded area). The dashed lines are Lorentzians. (b) Correlation length (L) and (c) defects density as a function of cobalt ferrite content.

possible to measure strain via both electronic transition and vibrational states in a crystalline structure. In this regard Raman spectroscopy has been extensively used as it can measure the variation of vibrational frequencies of a crystalline lattice. According to the literature [23] the E2high mode of the wurtzite hexagonal structure is highly sensitive to lattice strain [16]. Thus, the measurement of the Raman shift variations of the E2high mode, in deformed and non-deformed hexagonal wurtzite-related structures, can provide an estimate of the strain variation due to the substitution of Zn ions by Co or Fe ions. However, it is not trivial to explain the variation of the E2high mode, as it can be correlated to different origins. The main sources of variation are: alloy effects (variations in mass and strength of the bonds between atoms), lattice parameters modifications due to the introduction of strange atoms and effects of the breakdown of selection rules due to the reduction of long-range order. It is possible to estimate the Raman shift (Δω) of the E2high mode due to the strain effect in the c direction (εcc) using the following relation [23]:

Considering ω0 ¼439 cm  1 the values obtained for the E2high mode (for bulk ZnO [23]) are higher than those calculated. Therefore, in our case we anticipate other factors should be taken into account. The Raman modes in nanocrystalline oxides are very sensitive to disorder in the Oxygen sub-lattice. Using the Phonon Confinement Model (PCM) it is possible to obtain important information about crystalline disorder. In ideal crystals, due the momentum conservation, only phonons on the center of the Brillouin zone (q ¼0) can be observed by Raman scattering. In a solid, phonons can be spatially confined due to potential fluctuations caused by crystalline disorder, which generates a relaxation of the selection rule q ¼0 for Raman scattering. This spatial restriction of the phonon correlation length, which defines the average size of homogeneous region within the material, will cause line shape asymmetry in the Raman mode [25]. The phonon correlation length (L) can be estimated using the model proposed by Richter et al. [26] and Campbell and Fauchet [27], from which the Raman intensity is given by the Lorentzian-like function

⎡ ⎛ C33 ⎞ ⎤ Δω = ⎢β − α ⎜ ⎟ ⎥ εcc = mεcc , ⎝ C13 ⎠ ⎦ ⎣

I (ω) = I0

(1)

where α and β are deformation potentials and C13 and C33 are elastic constants. According to Gruber et al. [23] the experimental m value in Eq. (1) for zinc oxide is given by 527 728 cm  1. Considering only the effects of strain and using the values of the lattice parameters obtained by XRD and the following expression [24]:

εcc =

c − c0 Δc = c0 c0

it is possible to estimate the values of Δω and, consequently, the ω(E2high)calc values. These values are listed in Table 2 (second row). As can be observed by comparing the ω(E2high)calc values with those obtained experimentally (ω(E2high)exp) the strain effect alone does not explain the redshift observed for this vibrational mode.

C (q)

2

(ω − ωi (q))2 +

Γ0 2 2

( )

d 3q, (3)

where Io is the intensity pre-factor, ω(q) is the dispersion phonon curve, Γ0 is the natural full width at half maximum (FWHM) and C (q) is the Fourier component of the phonon confinement function, which is assumed to be Gaussian-shaped

C (q) (2)

∫

2

⎛ q2L2 ⎞ ⎟, = exp ⎜ − ⎝ 16π 2 ⎠

(4)

The phonon optical dispersion curve can be represented by

ωi (q) = A + B cos(πq),

(5) 1

1

where A¼424.5 cm and B ¼12.5 cm are assigned to the E2high mode [28]. The variable q is given in 2π/a units whereas Γ0 is

32

T.J. Castro et al. / Journal of Magnetism and Magnetic Materials 389 (2015) 27–33

Fig. 6. Mössbauer spectra of the (CoFe2O4)x þ(ZnO)1–x with x ¼0.050, 0.100, 0.200, 0.400 and 1.000. Table 3 Mössbauer parameters: IS-isomer shift; QS-quadrupole splitting; Hf- hyperfine field. Sample

x ¼0.050 x ¼0.100 x ¼0.200 x ¼0.400 x ¼1.000

Doublet Doublet Doublet Sextet Doublet Sextet Sextet Sextet

IS (mm/s)

QS (mm/s)

Hf (kOe)

Area (%)

 0.338  0.307  0.300  0.155  0.337  0.296  298  0.387

 0.430 0.451 0.483 0.273 0.580  0.007  0.178  352

   216.3  337.9 489.2 485.8

100 100 30.6 69.4 14.5 85.5 79.1 20.9

equal to 5.05 cm  1. Fig. 5(a) shows the fittings of the E2high mode for the wurtzite phase. The shaded area represents the PCM approach whereas the dashed lines are Lorentzian fits. Fig. 5(b) shows the behavior of the correlation length as a function of the cobalt ferrite content in the nanocomposite samples. As can be seen from Fig. 5(b) the L values decreases with the increasing of the CoFe2O4 content. This indicates that the phonon confinement region decreases, which is associated with the increasing of the compositional disorder [25] due to the migration of Cobalt and Iron ions from the cubic spinel to the wurtzite phase. As a consequence, the defects density, estimated by N = 3/4πL3, also increases as shown in Fig. 5(c). Fig. 6 shows the room-temperature Mössbauer spectra

Fig. 7. (a) Typical hysteresis loop of the pure CoFe2O4 obtained at room temperature. The inset in the lower right hand-side shows the diamagnetic behavior of the pure ZnO. (b) Magnetization versus applied magnetic field of the (CoFe2O4)x þ(ZnO)1–x nanocomposites for nominal x¼ 0.005, 0.050, 0.100, 0.200 and 0.400 at room temperature. The inset in the lower right hand-side shows the details of magnetization curves.

recorded from the (CoFe2O4)x þ(ZnO)1–x nanocomposities (x ¼0.050, 0.100, 0.200, 0.400, and CoFe2O4). As can be seen in Fig. 6 there is a systematic variation in the spectra profile as the cobalt ferrite content increases. For samples with x ¼0.05 and 0.100 doublets were observed whereas for samples with x ¼0.200 and x¼0.400 the spectra show a superposition of a doublet and a sextet, the latter presenting a broad line width. For the pure CoFe2O4 sample it is observed two sextets, which is an evidence for the absence of superparamagnetic relaxation at room temperature. Yet, the observed doublets can be associated with the possibility of occurrence of small particles presenting superparamagnetic behavior. Taking this possibility into account the contribution represented by the doublet is associated with magnetic collapse, meaning the transition from the sextet to the doublet profile. However, the possibility of magnetic ions exchange by Zn2 þ ions cannot be ignored. In this case, as it can be observed in the data listed in Table 3, the hyperfine field (Hf) increases with the cobalt ferrite content. This behavior can be explained by the model proposed by Sawatzky et al. [29], which considers the different ionic distributions in the first neighborhood of the Fe3 þ /Co2 þ ions and the molecular field due to the first-neighbor interactions. These interactions can be described by exchange integrals Ji. Thus, for the B-site the hyperfine field can be described by:

T.J. Castro et al. / Journal of Magnetism and Magnetic Materials 389 (2015) 27–33

Hf (B) = 2J1 nSz (A1) + 2J2 (6 − n) Sz (A2 ),

(6)

where the B–B interaction was neglected and the presence of only two types of ions in the A-sites were admitted instead. In Eq. (6) the exchange integrals (J1 and J2) correspond to the two types of ions in the A-sites whereas n is the number of ions of type 1 and (6-n) is the number of ions of type 2 (in A-sites), both as first neighbors (in the A-sites). Sz (A) is the average value of the spin zcomponent for the two types of ions. In the present case 1 and 2 correspond to Zn and Co/Fe ions, respectively. Once the Zinc ions have null magnetic moment the superexchange interaction represented by Fe3 + (B)−O−Zn2 + (A) will be zero, dropping off the first term in Eq. (6). Thus, the hyperfine field values in the B-site will depend on the content of the Zinc ion in the A-site only. Moreover, for small values of x there is a large number of non-magnetic ions (Zn2 þ ) from the ZnO phase available to substitute magnetic ions (Co2 þ and Fe3 þ ) in the cubic spinel phase. Therefore, one can expect the hyperfine field decreasing with the decrease of the cobalt ferrite content, as can be observed in the data shown in Table 3. Fig. 7 shows the room-temperature magnetization curves of the pure CoFe2O4 and ZnO phases (a) and (CoFe2O4)x þ(ZnO)1–x composites for nominal x¼ 0.005, 0.050, 0.100, 0.200, and 0.400 (b). It is clear that the magnetic properties of both ZnO (diamagnetic) and CoFe2O4 (ferrimagnetic) compounds changed drastically while forming the nanocomposite materials, as shown in Fig. 7(b). For instance, even for a small amount of ferrite content (x ¼0.005) the obtained nanocomposite is paramagnetic, becoming ferrimagnetic (soft type) as the ferrite content (x) increases, as shown in the inset at the lower right hand-side of Fig. 7(b). For example, for the x ¼0.200 composite sample, the recorded saturation magnetization (MS), coercivity (HC) and remanence (Mr) were 19 emu/g, 34 Oe and 2 emu/g, respectively. The origin of the ferrimagnetism in the (CoFe2O4)x þ(ZnO)1–x nanocomposite is obvious and due to the presence of the CoFe2O4. However, the observed MS value for the nominal x ¼0.200 nanocomposite is 30% lower than expected, despite the larger incorporation of CoFe2O4. Even significant differences were observed for smaller x values. As a first approximation this finding can be explained as long as the magnetization is taken as the difference between magnetic ions in A- and B-sites. Since the magnetic moment of Zn2 þ is zero and given that in cobalt ferrite Co2 þ ions preferably occupy the B-site the exchange of Co2 þ and Fe3 þ by Zn2 þ ions should decrease the saturation magnetization of the as-produced nanocomposite. In addition, as the non-magnetic Zn2 þ ions have no effect on the magnetic exchange interaction the exchange integral between ions in the B-sites becomes weaker at higher Zinc content and so the nanocomposite material tends to be paramagnetic at room temperature [30]. Furthermore, the diamagnetic phase due to the ZnO phase may deteriorate the magnetic properties (Ms, Hc, and Mr) of the nanocomposite.

4. Conclusions The structural and magnetic properties of the (CoFe2O4)x þ (ZnO)1–x nanocomposities synthetized by the combustion reaction method were investigated by EDS, SEM, XRD, VSM, Raman, and Mössbauer spectroscopies. The SEM micrographs showed the presence of aggregates, probably due to the annealing process, which increases with the CoFe2O4 content. The EDS data suggest that there is no sample contamination. Also, XRD data and Raman spectroscopy evidenced the presence of two phases only: cubic spinel and hexagonal wurtzite.

33

Changes of the lattice parameters (for both phases) were associated with the exchange of ions between cobalt ferrite and zinc oxide. Also, the reduction of the hyperfine field, observed from Mössbauer spectroscopy, was an evidence of Zn substitutional in the cubic spinel phase, which indeed was observed to be dependent upon the CoFe2O4 content. The strain effect alone was not enough to explain the variation of the E2high Raman mode observed in the wurtzite phase. Therefore, to account for the E2high mode variation, the Phonon Confinement Model was used, which showed a decreasing of the phonon correlation length and an increasing of defects density with the increasing of the cobalt ferrite content. The hysteresis cycles showed drastic modifications of the magnetic properties in both crystalline phases. It was observed that the nanocomposite changes from paramagnetic to ferrimagnetic as the cobalt ferrite content increases. The differences in the saturation magnetization while changing from the (CoFe2O4)x þ(ZnO)1–x nanocomposities to the pure CoFe2O4 was explained considering the exchange of Co2 þ and Fe3 þ by Zn2 þ ions.

Acknowledgments The authors acknowledge the financial support from the Brazilian agencies CNPq, CAPES and FINEP.

References [1] R.C. Pedroza, S.W. da Silva, M.A.G. Soler, P.P.C. Sartoratto, D.R. Rezende, P.C. Morais, J. Magn. Magn. Mater. 289 (2005) 139–141. [2] H. Singh, K.L. Yadav, J. Alloys Comp. 585 (2014) 805–810. [3] Q. Zheng, F. Huang, K. Ding, J. Huang, D. Chen, Z. Zhan, Z. Lin, Appl. Phys. Lett. 98 (2011) 221112. [4] A.V. Kachynski, A.N. Kuzmin, M. Nyk, I. Roy, P.N. Prasad, J. Phys. Chem. C 112 (2008) 10721–10724. [5] W.C. Shih, T.L. Wang, Y.K. Pen, Appl. Surf. Sci. 258 (2012) 5424–5428. [6] S. Mansouri, R. Bourguiga, F. Yakuphanoglu, Curr. Appl. Phys. 12 (2012) 1619–1623. [7] S.W. da Silva, T.F.O. Melo, M.A.G. Soler, E.C.D. Lima, A.F. da Silva, P.C. Morais, IEEE Trans. Magn. 39 (2003) 2645–2647. [8] C. Borgohaina, K.K. Senapati, K.C. Sarma, P. Phukan, J. Mol. Catal. A. Chem. 363–364 (2012) 495–500. [9] I. Balti, L.S. Smiri, P. Rabu, E. Gautron, B. Viana, N. Jouini, J. Alloys Compd. 586 (2014) S476–S482. [10] H.M. Rietveld, J. Appl. Crystallogr 2 (1969) 65–71. [11] A. C. Larson, R. B. von Dreele, General Structure Analysis System (GSAS), Los Alamos National Laboratory Report LAUR 86-8748, 1994. [12] B.H. Toby, J. Appl. Cryst. 34 (2001) 210–213. [13] L. Kumar, P. Kumar, M. Kar, J. Alloys Compd. 551 (2013) 72–81. [14] E. Kisi, M.M. Elcombe, Acta Crystallogr. Sect. C Cryst. Struct. Commun. 45 (1989) 1867–1870. [15] S. Desgreniers, Phys. Rev. B 58 (1998) 14102. [16] M. Šćepanović, M. Grujić-Brojćin, K. Vojisavljević, S. Bernik, T. Srećković, J. Raman Spectrosc. 41 (2010) 914–921. [17] F.J. Manjon, B. Mari, J. Serrano, A.H. Romero, J. Appl. Phys. 97 (2005) 053516. [18] C.L. Du, Z.B. Gu, M.H. Lu, J. Wang, S.T. Zhang, J. Zhao, G.X. Cheng, H. Heng, Y.F. Chen, J. Appl. Phys. 99 (2006) 123515. [19] M. Jouanne, J.F. Morhange, W. Szuszkiewicz, Z. Golacki, A. Mycielski, Phys. Status Solidi C 3 (2006) 1205–1208. [20] F. Nakagomi, S.W. da Silva, V.K. Garg, A.C. Oliveira, P.C. Morais, A. Franco Jr., J. Solid State Chem. 182 (2009) 2423–2429. [21] Z.W. Wang, D. Schiferl, Y.S. Zhao, H.S.C. O’Neill, J. Phys. Chem. Solids 64 (2003) 2517–2523. [22] F.H. Pollak, T.P. Pearsall, Semiconduct. Semimet. 32 (1990). [23] T. Gruber, G.M. Prinz, C. Kirchner, R. Kling, F. Reuss, W. Limmer, A. Waag, J. Appl. Phys. 96 (2004) 288–293. [24] R.K. Sendi, S. Mahmud, Appl. Surf. Sci. 258 (2012) 8026–8031. [25] F.J. Wang, Z.H. Xiong, D. Wu, J. Shi, Z.V. Vardeny, Synth. Met. 155 (2005) 172–175. [26] H. Richter, Z.P. Wang, L. Ley, Solid State Commun. 39 (1981) 625–629. [27] I.H. Campbell, P.M. Fauchet, Solid State Commun. 58 (1986) 739–741. [28] Y.C. Lee, S.Y. Hu, W. Water, Y.S. Huang, M.D. Yang, J.L. Shen, K.K. Tiong, C.C. Huang, Solid State Commun. 143 (2007) 250–254. [29] G.A. Sawatzky, F.V. Der Woude, A.H. Morrish, Phys. Rev. 187 (1969) 747–757. [30] P.I. Slick, Ferrites for non-microwave application, in: E.P. Wohlfarth (Ed.), Ferromagnetic Materials. A Handbook on the Properties of Magnetically Ordered Substances, vol. 2, North Publishing Company, Amsterdam, 1980, p. 202, chapter 3.