*2.2.1. Synthesis*

can be multiplied by corresponding force constants even if it is not simply to derive a reliable

For this reason, an important issue in ZFO NP studies is to reveal in synthesized samples the presence of iron oxides, like hematite and maghemite [5]. These oxides can infer an effective but extrinsic magnetic behavior with problems in terms of stability and aggregation if used in biomedical applications. To characterize the spin ordering and understand the nature of magnetic behavior, Mossbauer and neutron diffraction studies have been helpfully used [55]. Nevertheless, in this frame also the Raman spectroscopy can play an important role.

cubic inverted spinel structure, thus from the point of view of Raman activity equivalent to ZFO. These differences reflect on the Raman spectra. In **Figure 3**, we report the simulated

The Raman spectra of ZFO and maghemite are obviously quite similar with a total Raman yield

tions are dominant and the peak is around 650 cm−1, while in maghemite only the iron is present and the energy peak is usually observed around or just above 700 cm−1. On the contrary, the Raman spectrum of hematite is markedly different having the main peaks in the lower energy region. It is important to note that the Raman yield of hematite is approximately 10–20 times higher than that of spinel-type ferrites, thus a small amount of hematite is enough to give clear signatures in Raman spectrum. It is difficult to quantify this small amount because it is often reported that for ZFO nanometric grains, surface shell of hematite or maghemite can be formed [37, 56]. In this case, the surface sensitivity of the Raman technique can hinder a proper estimation. A more difficult goal is to reveal the presence of few amounts of maghemite in ZFO by Raman spectroscopy. Detailed best-fitting analyses are usually performed on the A1g data, and it is possible to infer about the presence of maghemite [38]; but the variability of peak energies found both for pure and doped ZFO and the weak weight of maghemite contribution in Raman signal hinder a reliable attribution simply based on Raman data. But, RS allows to monitor the stability of maghemite to laser irradiation. Indeed, it is well-known that in the oxidation route of iron ions maghemite is an intermediate metastable phase while hematite is the terminal and stable one [57]. Using this approach, the presence of maghemite in ZFO can be indirectly evalu-

greater for the former. The main difference is the peak position of the A1g. In ZFO, the ZnO<sup>4</sup>

ated by the insurgence of hematite Raman signals under laser irradiation [58].

**2.2. Experiments from pure and doped zinc ferrites obtained by microwave-assisted** 

O4

We focused on the microwave combustion synthesis method of Ca (on Zn site) and Gd (on Fe

and used for comparison. All the samples have been investigated by X-ray powder diffraction and SEM-EDX analyses as a complement of Raman results. The magnetic behavior has been measured by SQUID magnetometry. All the experimental details are reported elsewhere [59].

We found, for all the samples, a superparamagnetic behavior with saturation magnetization between 6 and 10 emu/g at the maximum applied magnetic field of 3 T, with a more effective

and Sr, and Al doped samples were also synthesized

) belongs to the rhombohedral system while maghemite (γ-Fe<sup>2</sup>

O3

) has a

vibra-

value [39, 47].

234 Raman Spectroscopy

Hematite (*α*-Fe2

**combustion method**

site) substituted ferrites. Undoped ZnFe2

role played by Ca ions with respect to Gd ions substitution.

O3

spectra for ZFO, hematite, and maghemite.

The samples were synthesized by microwave-assisted combustion methodology (MW), a rapid and green method as previously described, and by a conventional high energy ball milling.

For MW, the starting reagents were Zn(NO<sup>3</sup> ) 2 6H<sup>2</sup> O and Fe(NO<sup>3</sup> )3 9H<sup>2</sup> O taken in stoichiometric ratio to obtain ZnFe2 O4 and mixed in an agate mortar with a proper amount of citric acid as fuel, as calculated from the propellant chemistry theory [60]. In brief, by taking into account the reducing and oxidizing valences of the involved elements (Zn = +2, N = 0, O = −2, Fe = +3, C = 4, H = 1), we can calculate a global valence value of −15, −10, and +18 for iron nitrate, zinc nitrate, and citric acid, respectively. We can write the balanced chemical reaction for the ferrite formation as 2·(−15) + 1·(−10) + x·(+18) = 0 and the mol amount of citric acid necessary to reduce the nitrates oxidant amount is calculated as 2.22.

The mixture was placed in a ceramic crucible in a microwave oven for 30 min at 800 W: this power ensures a temperature inside the oven between 450 and 500°C. This undoped sample will be named ZnFe. To obtain the doped samples, proper amount of Ca(NO<sup>3</sup> ) 2 4H<sup>2</sup> O, Sr.(NO<sup>3</sup> )2 , Gd(NO<sup>3</sup> )3 6H<sup>2</sup> O, and Al(NO<sup>3</sup> ) 3 9H<sup>2</sup> O were added to the previous reagents to obtain the desired stoichiometry Zn0.95Ca0.05Fe2 O4 , Zn0.75Ca0.25Fe2 O4 , Zn0.95Sr0.05Fe2 O4 , ZnFe1.9Gd0.1O4 , and ZnFe1.9Al0.1O4 . The citric acid amount to add to the various mixtures was calculated as explained. In the following, these samples will be named Ca005, Ca025, Sr005, Gd01, and Al01, respectively.

For comparison, an undoped ZnFe<sup>2</sup> O4 sample was also synthesized by means of a classical solid-state synthesis, starting from a stoichiometric mixture of ZnO and Fe<sup>3</sup> O4 oxides ground by ball milling in tungsten jars for 6 h at 500 rpm with intermediate periods of last. The mixture was then treated in oven in air at 650°C for 12 h (heating rate 5°C/min, spontaneous cooling to 25°C) and ground after cooling. This sample will be named ZnFe-SS.

#### *2.2.2. XRD and SEM analyses*

X-ray diffraction analysis was used to determine the sample purity, in particular to control the effective cation substitution, the crystallite sizes and the eventual inversion degree of the spinel by means of the structural and profile refinement based on the Rietveld method. In **Figure 4**, the XRD patterns of the ZnFe-SS and ZnFe undoped ZnFe<sup>2</sup> O4 samples are shown together with Ca005 one. In all the cases, a pure ferrite sample is formed, whose peaks well agree with the ZnFe2 O4 cubic spinel structure (PDF card N. 89-7412). It is well evident a different peaks broadening and crystallinity between MW and SS samples suggesting markedly different particle sizes, as can be expected due to the different experimental synthesis methodologies. For the doped samples, XRD patterns demonstrate that the doping successfully occurred, because neither traces of unreacted reagents nor phase impurities are present: only for Sr005, some low peaks reveal the presence of small traces of strontium nitrate, as a residual of the reagent. The structural refinement on the basis of the Rietveld method was performed on all the patterns by using the known cubic spinel model: the main refined structural parameters are reported in **Table 3**. It can be seen that the crystallite sizes are in the nanometer range: the values for the microwave synthesis are all lower than 11 nm, while for ZnFe-SS a value of 22 nm is found.

refinement. In all the cases, the inversion takes place, although not so markedly: only for Ca025 a value of 0.54 is reached. The easily induced inversion can be a consequence of the quickness of the MW synthesis that in only about 30 min can produce a good level of crystallinity but with atomic disorder. We also verified that Ca and Sr. ions seem to prefer the A site, while Al and Gd ions the B one. This is certainly true when the substitution is about 5 atom%, while for Ca025 a different model could be hypothesized. In fact, due to the preference of calcium for octahedral coordination it is possible that these ions could be also located on Fe sites, so inducing a higher

Raman Spectroscopy in Zinc Ferrites Nanoparticles http://dx.doi.org/10.5772/intechopen.72864 237

inversion degree with respect to Ca005 and a small contraction of the lattice parameter.

In **Figure 5a,b**, SEM images of Gd01 and ZnFe-SS clear up the morphological differences between the samples from MW synthesis and solid state, giving evidence of a higher surface/

**Figure 5.** SEM images of Gd01 (a) and ZnFe-SS (b); data (c) and color map (d) from EDS elemental analysis for Gd01

sample.

**Figure 4.** XRD patterns of undoped ZnFe<sup>2</sup> O4 obtained from microwave combustion and solid state methods together with the Ca005 pattern. The bars of the expected angular positions of the spinel phase (JCPDS card 89–7412) are also reported, together with the miller indices of the main peaks.


**Table 3.** Cation distribution, impurity phases amount, lattice parameter, and crystallite size as derived from XRD analyses.

The lattice parameters although may vary due to the different ionic radii of the dopants with respect to Zn and Fe ions, are quite similar. On the other hand, a similar behavior for the lattice parameter was found for the Sr substituted samples [10]. Only for Al doping the cubic parameter decreases with respect to the undoped sample: in fact, a value of 0.53 Å is reported for Al3+ ions radius with respect to 0.65 Å of Fe3+ in high spin configuration [61]. The inversion degree, i.e., the amount of Fe ions on Zn crystallographic sites, was also determined from the structural refinement. In all the cases, the inversion takes place, although not so markedly: only for Ca025 a value of 0.54 is reached. The easily induced inversion can be a consequence of the quickness of the MW synthesis that in only about 30 min can produce a good level of crystallinity but with atomic disorder. We also verified that Ca and Sr. ions seem to prefer the A site, while Al and Gd ions the B one. This is certainly true when the substitution is about 5 atom%, while for Ca025 a different model could be hypothesized. In fact, due to the preference of calcium for octahedral coordination it is possible that these ions could be also located on Fe sites, so inducing a higher inversion degree with respect to Ca005 and a small contraction of the lattice parameter.

In **Figure 5a,b**, SEM images of Gd01 and ZnFe-SS clear up the morphological differences between the samples from MW synthesis and solid state, giving evidence of a higher surface/

The lattice parameters although may vary due to the different ionic radii of the dopants with respect to Zn and Fe ions, are quite similar. On the other hand, a similar behavior for the lattice parameter was found for the Sr substituted samples [10]. Only for Al doping the cubic parameter decreases with respect to the undoped sample: in fact, a value of 0.53 Å is reported for Al3+ ions radius with respect to 0.65 Å of Fe3+ in high spin configuration [61]. The inversion degree, i.e., the amount of Fe ions on Zn crystallographic sites, was also determined from the structural

**Table 3.** Cation distribution, impurity phases amount, lattice parameter, and crystallite size as derived from XRD

with the Ca005 pattern. The bars of the expected angular positions of the spinel phase (JCPDS card 89–7412) are also

ZnFe-SS [Zn0.93Fe0.07]T[Fe1.93Zn0.07]O — 8.4386(6) 22.9 ZnFe [Zn0.82Fe0.18]T[Fe1.82Zn0.18]O — 8.4371(40) 11.1 Al01 [Zn0.68Fe0.32]T[Fe1.58Al0.1Zn0.32]O — 8.4221(25) 8.4 Ca005 [Zn0.90Ca0.05Fe0.05]T[Fe1.95Zn0.05]O — 8.4343(17) 8.7 Ca025 [Zn0.21Ca0.25Fe0.54]T[Fe1.46Zn0.54]O — 8.4331(57) 5.5

Gd01 [Zn0.98Fe0.02]T[Fe1.88Gd0.1Zn0.02]O — 8.4366(29) 8.4

**Cations distribution Impurities a/Å Size/nm**

)2

obtained from microwave combustion and solid state methods together

3.35 8.4380(13) 7.2

**Figure 4.** XRD patterns of undoped ZnFe<sup>2</sup>

236 Raman Spectroscopy

analyses.

reported, together with the miller indices of the main peaks.

Sr005 [Zn0.88Sr0.05Fe0.07]T[Fe1.93Zn0.07]O Sr(NO<sup>3</sup>

O4


**Figure 5.** SEM images of Gd01 (a) and ZnFe-SS (b); data (c) and color map (d) from EDS elemental analysis for Gd01 sample.

volume ratio for MW samples with respect to solid-state one. From MW, rounded particles (lower than 100 nm) aggregates with large and open pores, regardless the doping ion, can be seen. This aspect can be due to the evolution of gases from nitrates and citric acid (such as NO<sup>2</sup> and CO2 ) during the heating process in the microwave oven. The solid-state synthesis also leads to aggregates, but with larger rounded particles. The electronic microanalysis allowed us to verify that the stoichiometric ratio between the different elements was maintained in the final products. In **Figure 5c**, as an example, the EDS analysis of Gd01 sample is reported. The atomic percentages of all the elements are in excellent agreement with the stoichiometric values, within the EDS detection limit, suggesting that the ferrites possess the expected composition and no ions loss occurred. The maps of the different elements show good homogeneity, suggesting that Gd ions are well distributed in the sample. A similar behavior was observed for all the undoped and the other doped samples.

#### *2.2.3. Magnetic results*

Hysteresis cycles at RT and at 10 K have been measured as well as the zero-field and field cooled magnetization curves in temperature. The magnetic data have been carefully analyzed [59], and they provide a basic proof of the RT superparamagnetic behavior, essential requirement in magnetic hyperthermia. Examples for the hysteresis curves at RT and 10 K are reported in **Figure 6a** for the Ca005 sample, as well the relative ZFC and FC magnetization curves (**Figure 6b**).

responsible of higher MS

temperature); (iii) coercive field H<sup>c</sup>

**TB (K)** **TIrr (K)** **HC (300 K) (Oe)**

ZnFe-SS 24 24 0 0 Linear behavior

**MRem (300 K) (emu/g)**

**MSat**

(5.3)

ZnFe 42 70 Negl Negl 8.7 (S shape) 115 3.3 38.2 Al01 50 100 Negl Negl 6.7 (S shape) 350 3.4 26.7 Ca005 45 100 Negl Negl 10.0 (S shape) 250 5.2 45.0 Ca025 52 70 Negl Negl 5.8 (S shape) 100 2.3 21.5 Sr005 53 85 Negl Negl 8.1 (S shape) 90 1.6 18.2 Gd01 34 60 Negl Negl 6.7 (S shape) 115 2.8 43.7

**(300 K–30 kOe) (emu/g)**

*2.2.4. Raman results*

manner with the spinel inversion.

negligible (Negl) means lower than 5 Oe for H<sup>c</sup>

evidence on the effective role of inversion degree, even if MS

with pure ZFO sample obtained by solid state synthesis.

**Figure 7.** Raman spectra at RT in the region from 100 to 800 cm−1 for all the samples investigated.

values with respect to the regular spinel. Magnetization data give

, remnant and saturation magnetization (MRem and MSat) both at RT and 10 K. The term

**HC (10 K) (Oe)**

**MRem (10 K) (emu/g)**

Raman Spectroscopy in Zinc Ferrites Nanoparticles http://dx.doi.org/10.5772/intechopen.72864

80 0.6 30.4

**MSat(10 K–30 kOe)** 

239

**(emu/g)**

In **Figure 7**, the Raman spectra of pure and doped MW ZFO samples are reported together

**Table 4.** Different functional parameters derived from magnetization experiments: (i) the blocking temperature, TB, as defined in the introduction; (ii) the irreversibility temperature, TIrr, corresponding to the merging of ZFC and FC curves and its distance from TB (TIrr gives suggestions about spin and structural disorder of nanoparticles also come from the

, and lower than 10−2 for Mrem.

values correlate in a complex

The magnetic behavior for all the samples is summarized in **Table 4**, where all the functional parameters are reported.

We found, for all the MW samples, a superparamagnetic behavior with saturation magnetization at RT between 6 and 10 emu/g at the maximum applied magnetic field of 3 T, with a more effective role played by Ca ions with respect to Gd ions substitution. A fair agreement is found with the MS values reported in the literature for pure and doped zinc ferrite spinels having comparable grain size and prepared by means of combustion synthesis or other methods [10, 60, 62–64]. In ZnFe<sup>2</sup> O4 , the inversion promotes super exchange interactions, in turn

**Figure 6.** (a) Hysteresis loops at 300 K and at 10 K (in the inset) for Ca005; (b) ZFC and FC magnetization data for the same sample.


**Table 4.** Different functional parameters derived from magnetization experiments: (i) the blocking temperature, TB, as defined in the introduction; (ii) the irreversibility temperature, TIrr, corresponding to the merging of ZFC and FC curves and its distance from TB (TIrr gives suggestions about spin and structural disorder of nanoparticles also come from the temperature); (iii) coercive field H<sup>c</sup> , remnant and saturation magnetization (MRem and MSat) both at RT and 10 K. The term negligible (Negl) means lower than 5 Oe for H<sup>c</sup> , and lower than 10−2 for Mrem.

responsible of higher MS values with respect to the regular spinel. Magnetization data give evidence on the effective role of inversion degree, even if MS values correlate in a complex manner with the spinel inversion.

#### *2.2.4. Raman results*

**Figure 6.** (a) Hysteresis loops at 300 K and at 10 K (in the inset) for Ca005; (b) ZFC and FC magnetization data for the

volume ratio for MW samples with respect to solid-state one. From MW, rounded particles (lower than 100 nm) aggregates with large and open pores, regardless the doping ion, can be seen. This aspect can be due to the evolution of gases from nitrates and citric acid (such as NO<sup>2</sup>

leads to aggregates, but with larger rounded particles. The electronic microanalysis allowed us to verify that the stoichiometric ratio between the different elements was maintained in the final products. In **Figure 5c**, as an example, the EDS analysis of Gd01 sample is reported. The atomic percentages of all the elements are in excellent agreement with the stoichiometric values, within the EDS detection limit, suggesting that the ferrites possess the expected composition and no ions loss occurred. The maps of the different elements show good homogeneity, suggesting that Gd ions are well distributed in the sample. A similar behavior was observed

Hysteresis cycles at RT and at 10 K have been measured as well as the zero-field and field cooled magnetization curves in temperature. The magnetic data have been carefully analyzed [59], and they provide a basic proof of the RT superparamagnetic behavior, essential requirement in magnetic hyperthermia. Examples for the hysteresis curves at RT and 10 K are reported in **Figure 6a** for the Ca005 sample, as well the relative ZFC and FC magnetization curves (**Figure 6b**).

The magnetic behavior for all the samples is summarized in **Table 4**, where all the functional

We found, for all the MW samples, a superparamagnetic behavior with saturation magnetization at RT between 6 and 10 emu/g at the maximum applied magnetic field of 3 T, with a more effective role played by Ca ions with respect to Gd ions substitution. A fair agreement

having comparable grain size and prepared by means of combustion synthesis or other meth-

values reported in the literature for pure and doped zinc ferrite spinels

, the inversion promotes super exchange interactions, in turn

) during the heating process in the microwave oven. The solid-state synthesis also

same sample.

and CO2

238 Raman Spectroscopy

*2.2.3. Magnetic results*

parameters are reported.

is found with the MS

ods [10, 60, 62–64]. In ZnFe<sup>2</sup>

for all the undoped and the other doped samples.

O4

In **Figure 7**, the Raman spectra of pure and doped MW ZFO samples are reported together with pure ZFO sample obtained by solid state synthesis.

**Figure 7.** Raman spectra at RT in the region from 100 to 800 cm−1 for all the samples investigated.

For all the samples, a well-defined first order Raman scattering pertinent to the ferrite phase has been detected with the most prominent signals around 650 cm−1 (A1g mode) accompanied by the other modes at lower energy. The substantial invariance of the Raman features for all the samples indicates a good stability of the spinel structure. The lower total Raman yield and the broadening of the signals of MW samples are consistent with the lower density and smaller crystallite size of the MW powders as derived by XRD analyses. Doping seems to play a negligible role in the peak positions, with the exception of Ca025 sample. For this heavily doped sample, the higher bands move at higher energies, while the F2g(2) signal further decreases in energy. For Ca and Gd samples, a weak feature at around 220 cm−1, probably due to the F2g(1) mode, is present. By comparing the pure samples, we can notice a weak, but observable, red-shift for all the Raman features of the MW sample. This fact is compatible with the reduction of crystallite size (11 nm for ZnFe and 22.9 nm for ZnFe-SS). The peak position of A1g moves from 642.5 to 638.5 cm−1, while the F2g(2) peak is peaked at 345.0 cm−1 for ZnFe-SS sample and at 334.0 cm−1 for ZnFe one.

In particular, for the whole set of samples, the A1g band is well fitted by the overlapping of two signals: the first, centered at around 641 cm−1, gives the main contribution and the second, centered at around 685 cm−1, results in a shoulder at higher energies. The energies are in good

of these bands, by applying the simple formula I685/(I685 + I641), we could estimate the inversion degree obtaining, for ZnFe-SS sample, i.e., our standard reference, a value equal to 0.095,

The results for inversion degree are reported in **Figure 9**, where the same parameter from

Even if absolute values from the two methods do not coincide, a common trend is evidenced and the highest inversion is found for Ca025 from both the techniques. Anyway, the discrepancies can be due to the different penetration depths of the two probes leading to the detection of different structural features, especially when clustered nanometric particles are involved. XRD sampling involves the bulk of the sample while Raman measurements probe mainly the surfaces, i.e., the most defective and disordered sample zones and thus it is not surprising that for undoped and low-level doped samples higher inversion degree are derived from Raman. On the contrary, for Al01 and Ca025 samples, a slightly higher inversion degree is derived from XRD analysis. In these cases, we have the higher doping level; in addition the ionic radii of Ca and Al ions are markedly higher and lower, respectively, than that of Zn and Fe ones with respect to other samples, and the differences in ionic radii with respect to the substituted ions. In fact, Ca ions are markedly greater, while Al markedly lower than Zn and Fe ones [61]. In addition, Ca025 powders exhibit the lower crystallite size. All these elements can lead to an increase of inversion degree even in the core of nanoparticles. Finally, we remark that at higher doping level, the contribution of substituents in Raman modes can give rise to a proper vibration at a specific energy. The two Lorenztian model used to interpolate A1g modes remain effective even

with minor physical meaning and the error in inversion degree estimation can grow.

**Figure 9.** Inversion degree parameters obtained from Raman and XRD data as explained in the text, for all the synthesized ferrite samples. For each sample several Raman runs have been performed in different sample zones to obtain an average

and FeO4

units. From the intensities

241

Raman Spectroscopy in Zinc Ferrites Nanoparticles http://dx.doi.org/10.5772/intechopen.72864

agreement with those expected for modes inside ZnO<sup>4</sup>

slightly higher than 0.07, derived from XRD analysis.

XRD analyses is already shown for comparison.

inversion degree with a range of variation of about 10%.

The lattice parameter value could play a role in this red shift. According to [44], the changes in Raman line position are related to the changes in the lattice parameter by the relationship Δω*<sup>i</sup>* = −3 γ*<sup>i</sup>* (*q*)ω*<sup>i</sup>* (*q*)(Δ*a*/*a*<sup>0</sup> ), where γ is the Gruneisen mode parameter and Δω*<sup>i</sup>* is the shift of Raman line. The values of γ taken by [27] are 0.72 for A1g mode and 1.88 for the F2g lower energy mode. From the lattice parameters reported in **Table 3**, we derive an expected Raman shift of about 2.0 and 3.5 cm−1 for A1g and F2g modes, respectively. The measured shifts are greater and thus the lattice parameter by itself cannot account for the observed Raman softening, for which the role of inversion in the spinel structure must be considered.

As discussed in the previous section, all the Raman data in the region 200–800 cm−1 have been treated by best-fitting procedures. Due to the different involved masses, we used couples of oscillators for the three main Raman features with an additional curve for the low energy weaker signals. Even for higher doping levels (Al01, Gd01 and Ca025), we avoid to introduce a third oscillator. A very satisfactory agreement with experimental data has been obtained. As an example, we report the result for sample ZnFe-SS in **Figure 8**.

**Figure 8.** A representative example of best-fitting of the Raman spectrum in the range of 100–800 cm−1 for sample ZnFe-SS.

In particular, for the whole set of samples, the A1g band is well fitted by the overlapping of two signals: the first, centered at around 641 cm−1, gives the main contribution and the second, centered at around 685 cm−1, results in a shoulder at higher energies. The energies are in good agreement with those expected for modes inside ZnO<sup>4</sup> and FeO4 units. From the intensities of these bands, by applying the simple formula I685/(I685 + I641), we could estimate the inversion degree obtaining, for ZnFe-SS sample, i.e., our standard reference, a value equal to 0.095, slightly higher than 0.07, derived from XRD analysis.

For all the samples, a well-defined first order Raman scattering pertinent to the ferrite phase has been detected with the most prominent signals around 650 cm−1 (A1g mode) accompanied by the other modes at lower energy. The substantial invariance of the Raman features for all the samples indicates a good stability of the spinel structure. The lower total Raman yield and the broadening of the signals of MW samples are consistent with the lower density and smaller crystallite size of the MW powders as derived by XRD analyses. Doping seems to play a negligible role in the peak positions, with the exception of Ca025 sample. For this heavily doped sample, the higher bands move at higher energies, while the F2g(2) signal further decreases in energy. For Ca and Gd samples, a weak feature at around 220 cm−1, probably due to the F2g(1) mode, is present. By comparing the pure samples, we can notice a weak, but observable, red-shift for all the Raman features of the MW sample. This fact is compatible with the reduction of crystallite size (11 nm for ZnFe and 22.9 nm for ZnFe-SS). The peak position of A1g moves from 642.5 to 638.5 cm−1, while the F2g(2) peak is peaked at 345.0 cm−1 for

The lattice parameter value could play a role in this red shift. According to [44], the changes in Raman line position are related to the changes in the lattice parameter by the

is the shift of Raman line. The values of γ taken by [27] are 0.72 for A1g mode and 1.88 for the F2g lower energy mode. From the lattice parameters reported in **Table 3**, we derive an expected Raman shift of about 2.0 and 3.5 cm−1 for A1g and F2g modes, respectively. The measured shifts are greater and thus the lattice parameter by itself cannot account for the observed Raman softening, for which the role of inversion in the spinel structure must be

As discussed in the previous section, all the Raman data in the region 200–800 cm−1 have been treated by best-fitting procedures. Due to the different involved masses, we used couples of oscillators for the three main Raman features with an additional curve for the low energy weaker signals. Even for higher doping levels (Al01, Gd01 and Ca025), we avoid to introduce a third oscillator. A very satisfactory agreement with experimental data has been obtained. As

**Figure 8.** A representative example of best-fitting of the Raman spectrum in the range of 100–800 cm−1 for sample ZnFe-SS.

), where γ is the Gruneisen mode parameter and Δω*<sup>i</sup>*

ZnFe-SS sample and at 334.0 cm−1 for ZnFe one.

(*q*)ω*<sup>i</sup>*

(*q*)(Δ*a*/*a*<sup>0</sup>

an example, we report the result for sample ZnFe-SS in **Figure 8**.

relationship Δω*<sup>i</sup>* = −3 γ*<sup>i</sup>*

considered.

240 Raman Spectroscopy

The results for inversion degree are reported in **Figure 9**, where the same parameter from XRD analyses is already shown for comparison.

Even if absolute values from the two methods do not coincide, a common trend is evidenced and the highest inversion is found for Ca025 from both the techniques. Anyway, the discrepancies can be due to the different penetration depths of the two probes leading to the detection of different structural features, especially when clustered nanometric particles are involved. XRD sampling involves the bulk of the sample while Raman measurements probe mainly the surfaces, i.e., the most defective and disordered sample zones and thus it is not surprising that for undoped and low-level doped samples higher inversion degree are derived from Raman. On the contrary, for Al01 and Ca025 samples, a slightly higher inversion degree is derived from XRD analysis. In these cases, we have the higher doping level; in addition the ionic radii of Ca and Al ions are markedly higher and lower, respectively, than that of Zn and Fe ones with respect to other samples, and the differences in ionic radii with respect to the substituted ions. In fact, Ca ions are markedly greater, while Al markedly lower than Zn and Fe ones [61]. In addition, Ca025 powders exhibit the lower crystallite size. All these elements can lead to an increase of inversion degree even in the core of nanoparticles. Finally, we remark that at higher doping level, the contribution of substituents in Raman modes can give rise to a proper vibration at a specific energy. The two Lorenztian model used to interpolate A1g modes remain effective even with minor physical meaning and the error in inversion degree estimation can grow.

**Figure 9.** Inversion degree parameters obtained from Raman and XRD data as explained in the text, for all the synthesized ferrite samples. For each sample several Raman runs have been performed in different sample zones to obtain an average inversion degree with a range of variation of about 10%.

Nevertheless, it is important to underline, as evidenced in **Figure 10**, that the best-fitting procedure with two lorentzian curves allows a very good interpolation of the experimental data.

Apart from the estimation of inversion degree, from the fitting procedures the Raman band parameters have been derived for all the samples. **Figure 11** shows the integrated intensities and linewidths and allow to graphically appreciate what already observed above. A marked decrease in total Raman yield for Ca025, especially for the component peaked at 648 cm−1. This component is the most affected by the substitution on A site, thus confirming the attribution for this mode.

As a final remark, one can infer that all the Raman data from MW samples indicate a good crystalline quality, absence of impurities or at least below the threshold of sensitivity, but in any case in agreement with results from XRD and morphological and chemical analyses. Thus, the magnetic functional parameters should be considered as intrinsic of the ZFO matrix, eventually doped.

**Figure 11.** (a) Integrated intensities and (b) linewidths for the two lorenztian curves derived from the best fitting of A1g

Raman Spectroscopy in Zinc Ferrites Nanoparticles http://dx.doi.org/10.5772/intechopen.72864 243

**Figure 12.** Raman spectra of Al01 (a) and Sr005 (b) samples collected during a thermal cycle. The samples were irradiated

for 5 min and then measured at the selected power density.

modes for all the investigated samples.

Anyway, from what discussed in Section 2.1, the Raman experimental set-up allows us to evaluate the stability of the pure and doped MW ZFO powders under laser irradiation and thus to infer about the presence of cluster of maghemite. It is difficult to appreciate the presence of maghemite directly from the Raman spectra, even after detailed spectral analyses. Hence, all the samples were evaluated by means of thermal cycles. Neutral filters with optical density equal to 2, 1, 0.6, 0.3, and 0 were used leading to power density values from 5 × 10<sup>3</sup> to 5 × 10<sup>5</sup> W/cm<sup>2</sup> . The samples were tested in different regions.

In **Figure 12a** and **b**, we report the results from the sample Al01 and the sample Sr005, respectively. These figures show two different behaviors in term of phase stability.

**Figure 10.** Raman data and curves from best-fitting runs for (a) ZnFe and (b) Sr005.

Nevertheless, it is important to underline, as evidenced in **Figure 10**, that the best-fitting procedure with two lorentzian curves allows a very good interpolation of the experimental data. Apart from the estimation of inversion degree, from the fitting procedures the Raman band parameters have been derived for all the samples. **Figure 11** shows the integrated intensities and linewidths and allow to graphically appreciate what already observed above. A marked decrease in total Raman yield for Ca025, especially for the component peaked at 648 cm−1. This component is the most affected by the substitution on A site, thus confirming the attribution

As a final remark, one can infer that all the Raman data from MW samples indicate a good crystalline quality, absence of impurities or at least below the threshold of sensitivity, but in any case in agreement with results from XRD and morphological and chemical analyses. Thus, the magnetic functional parameters should be considered as intrinsic of the ZFO matrix,

Anyway, from what discussed in Section 2.1, the Raman experimental set-up allows us to evaluate the stability of the pure and doped MW ZFO powders under laser irradiation and thus to infer about the presence of cluster of maghemite. It is difficult to appreciate the presence of maghemite directly from the Raman spectra, even after detailed spectral analyses. Hence, all the samples were evaluated by means of thermal cycles. Neutral filters with optical density

In **Figure 12a** and **b**, we report the results from the sample Al01 and the sample Sr005, respec-

to 5 × 10<sup>5</sup>

equal to 2, 1, 0.6, 0.3, and 0 were used leading to power density values from 5 × 10<sup>3</sup>

tively. These figures show two different behaviors in term of phase stability.

. The samples were tested in different regions.

**Figure 10.** Raman data and curves from best-fitting runs for (a) ZnFe and (b) Sr005.

for this mode.

242 Raman Spectroscopy

eventually doped.

W/cm<sup>2</sup>

**Figure 11.** (a) Integrated intensities and (b) linewidths for the two lorenztian curves derived from the best fitting of A1g modes for all the investigated samples.

**Figure 12.** Raman spectra of Al01 (a) and Sr005 (b) samples collected during a thermal cycle. The samples were irradiated for 5 min and then measured at the selected power density.

The sample Sr005, even if thermally treated under laser light, remains stable showing the same Raman spectrum during the whole laser-induced thermal cycle. Small peak shifts are due to temperature effects on lattice parameters. On the contrary for Al01 sample, it is evident that upon irradiation, the Raman spectrum of the Al01 change markedly and the typical Raman features of hematite appear when the sample is maintained under laser light having a power density equal to 2 × 10<sup>5</sup> W/cm<sup>2</sup> . When the laser light is attenuated at the end of the thermal cycle, the signal of hematite is still observable indicating a stable conversion. This is an indirect proof that at the beginning of the thermal cycle nanoregions in the irradiated volume had a structure different from the proper ZFO structure. It is likely that nanoregions of maghemite or highly defective Zn deficient zinc ferrite layers were present. We underline that the grain sizes for Al01 and Sr005 are similar.

In the second part of the chapter, we presented original results from room temperature

pared by using the microwave-assisted combustion method, a rapid, green, and simple synthesis route, able to ensure good physical and chemical properties. We underline the relevance of Ca substitution due to the low toxicity of the substituent; on the other side, Gd doping is rarely reported in zinc ferrites. Raman experiments have been accompanied by detailed analyses on morphological, compositional, structural, and magnetic characterization. Thus, results from XRD, SEM-EDS are briefly presented as well as the magnetic behavior. A detailed analy-

derived from best-fitting procedures and carefully compared to literature data. The results have been discussed in relation with grain size, inversion degree and doping of zinc ferrite spinel. An optimum purity level and homogeneity and crystallite sizes lower than 11 nm were determined for the doped samples. The inversion degree of the different samples was derived by both Raman and XRD data and a well agreement has been observed. The superparamagnetism seems to be favored in Ca doped samples, while in Gd doped one it is almost negligible. This is clearly associated to the inversion induced by the doping in the tetrahedral site.

In addition by Raman studies we evaluated the stability of pure and doped ZnFe<sup>2</sup>

, Marcella Bini<sup>2</sup>

[1] Kmita A, Pribulova A, Holtzer M, Futas P, Roczniak A. Use of specific properties of zinc ferrite in innovative technology. Archives of Metallurgy and Materials. 2016;**61**(4):2141-

[2] Suchomski C, Breitung B, Witte R, Knapp M, Bauer S, Baumbach T, Reitz C, Brezesinski T.

cation in energy storage and nanomagnetics. Beilstein Journal of Nanotechnology. 2016;

laser heating just to reveal the presence of maghemite in the samples. This set of data gives evidence to the ability of Raman spectroscopy in a typical problem of solid-state material science. Finally, the present results will be the bases of further works aimed to exploit SERS effect using non-metallic nanostructured zinc ferrites eventually functionalized for specific

, 3F2g Raman modes has been performed and Raman band parameters have been

, and ZnFe1.9Al0.1O4

O4, Zn1−xCax

and Maria Cristina Mozzati<sup>1</sup>

O4

nanocrystals for appli-

Fe2 O4

doped ferrites. The samples have been pre-

Raman Spectroscopy in Zinc Ferrites Nanoparticles http://dx.doi.org/10.5772/intechopen.72864

(x = 0.05 and 0.25),

245

O4 with

micro-Raman studies on nanostructured undoped ZnFe<sup>2</sup>

O4

, Zn0.95Sr0.05Fe2

ZnFe1.9Gd0.1O4

sis of A1g, E<sup>g</sup>

biomedical applications.

\*, Benedetta Albini<sup>1</sup>

2146. DOI: 10.1515/amm-2016-0289

**7**:1350-1360. DOI: 10.3762/bjnano.7.126

\*Address all correspondence to: pietro.galinetto@unipv.it

2 Department of Chemistry, University of Pavia, Pavia, Italy

1 Department of Physics and CNISM, University of Pavia, Pavia, Italy

Microwave synthesis of high-quality and uniform 4 nm ZnFe<sup>2</sup>

**Author details**

Pietro Galinetto<sup>1</sup>

**References**

Only three samples exhibit a stable pure zinc ferrite phase with no hematite detected at the end of the thermal cycle: ZnFe-SS, Ca025, and Sr005. The first exhibited a paramagnetic behavior, while the latter were found to be superparamagnetic. The phase stability proved by monitoring the Raman spectrum under laser heating is fundamental to validate the quality of the powders with good magnetic functional parameters, in particular in view of applications in biomedical field.
