# Effect of the concentration of precursors on the microwave absorbent properties of Zn/Fe oxide nanopowders

## Authors

- First Online:

- Received:
- Accepted:

DOI: 10.1007/s11051-010-0032-1

- Cite this article as:
- Fannin, P.C., Marin, C.N., Malaescu, I. et al. J Nanopart Res (2011) 13: 311. doi:10.1007/s11051-010-0032-1

- 2 Citations
- 138 Views

## Abstract

Zn/Fe oxide compound powders were obtained by the hydrothermal method using ferric nitrate Fe(NO_{3})_{3}·9H_{2}O and zinc nitrate Zn(NO_{3})_{2}·6H_{2}O at 200 °C and different precursor molar ratios* x* = Fe^{3+}/Zn^{2+} equal to 2.8/0.2, 2.5/0.5, 1.8/1.2 and 1.5/1.5. The samples were characterized by X-ray diffraction (XRD) and scanning electron microscopy/energy dispersive X-ray analysis (SEM–EDAX). Room temperature measurements of the frequency dependence of the complex magnetic permeability and complex dielectric permittivity, over the frequency range from 0.1 to 6 GHz, were performed. For precursor molar ratios* x* = 2.8/0.2,* x* = 1.8/1.2 and* x* = 1.5/1.5 the obtained samples showed a ferromagnetic-like resonance behaviour. This behaviour was assigned to the prevalent compounds in the obtained samples, Fe_{2}O_{3} (for* x* = 2.8/0.2) and ZnFe_{2}O_{4} (for* x* = 1.8/1.2 and* x* = 1.5/1.5). Based on the magnetic and dielectric measurements, the microwave absorbent properties of the four samples were analysed, and the sample containing mostly of ZnFe_{2}O_{4} (for* x* = 1.8/1.2) was found to be the best electromagnetic absorber in the frequency range 1.36–6 GHz.

### Keywords

Microwave absorberZn/Fe oxideHydrothermal methodComplex magnetic permeabilityComplex dielectric permittivityNanoparticles## Introduction

With the progress made in the practical applications of microwaves, the problem of electromagnetic interference has become of utmost importance, increasing the need for electromagnetic absorbent materials. An electromagnetic absorber must allow electromagnetic waves to penetrate inside it, where the electromagnetic wave must experience high electric and magnetic losses, within a layer as thin as possible.

*V*of material, which is characterized by the imaginary part of the complex permeability,

*μ*′′, the imaginary part of the complex permittivity,

*ε*′′, and the conductivity,

*σ*, is given by [1]:

*f*is the frequency of the electromagnetic wave propagating through the medium, \( \varepsilon^{\prime\prime}_{\text{eff}} = {\frac{\sigma }{{2\pi f\varepsilon_{0} }}} + \varepsilon^{\prime\prime} \) is the imaginary part of the effective complex permittivity,

*μ*

_{0}is the free space permeability,

*ε*

_{0}is the free space permittivity, \( \vec{H} \) is the magnetic field strength, \( \vec{H}^{*} \) is the complex conjugated magnetic field strength, \( \vec{E} \) is the electric field strength and \( \vec{E}^{*} \) is the complex conjugated electric field strength.

A good absorbent material must exhibit no reflection at the air–absorber interface and a high value of the time-average of the electromagnetic loss power within its volume. From Eq. 1 it can be observed that the conductors have a high loss power, *P*_{loss}, but it is known that, at the same time, they reflect the electromagnetic waves (Collin 1966). Also, *P*_{loss} of a specific material depends on its magnetic and electric parameters (*μ*′′, *ε*′′ and *σ*) and on the frequency of the electromagnetic wave, limiting its operation range. Hence, one way to design an electromagnetic absorber working in a desired frequency range is by the use of composite materials.

The first step in realizing composite absorbing systems is the synthesis of nano-microparticles, and one of the methods often used is the hydrothermal method (Lijun Zhao et al. 2008). In brief, in the hydrothermal method the precursors (i.e. the starting raw materials) are mixed in an autoclave for the precipitation process and then the autoclave is sealed and heated for a period of time (Lijun Zhao et al. 2008). As it is known, the initial molar ratios of the precursors may not be the same as the molar ratios of the ions in the resulting compounds (El-Shobaky et al. 2010; Franco et al. 2007). Depending on the concentration of the precursors, on the oxidation and reduction numbers of the precursor ions, on the pH value of the precipitation solution and on the temperature of the reaction, it is possible to get the desired nano-microparticle powder together with a mixture of other oxides.

Various ferrite nanoparticle systems were analysed from the microwave properties point of view (Sunny et al. 2010; Dong-Lin Zhao et al. 2009; Ghasemi and Morisako 2008; Lixi Wang et al. 2009; Duan Yuping et al. 2010; Jin Hai-bo et al. 2010; Guozhi Xie et al. 2010; Cheng et al. 2010). Zn/Fe oxide compound powders were analysed both for basic research in magnetism (Zhu et al. 2009) and various technological applications, including gas sensors (Chu Xiangfeng et al. 1999; Wen-Hui Zhang et al. 2010), catalysts (Valenzuela et al. 2002; Jianjun Liu et al. 1996; Toledo-Antonio et al. 2002; Jianxun Qiu et al. 2004; Yuan and Zhang 2001) and as desulfurization sorbents (Mayumi Tsukada et al. 2008).

In this article we analyse the effect of the concentration of the precursors on the microwave absorbent properties of the resulting Zn/Fe oxide compound powders, as obtained by the hydrothermal method.

## Obtaining method

Four samples were prepared, using the same precursors (i.e. ferric nitrate (Fe(NO_{3})_{3}·9H_{2}O and zinc nitrate Zn(NO_{3})_{2}·6H_{2}O) in different designated molar ratios x = Fe^{3+}/Zn^{2+} equal to 2.8/0.2, 2.5/0.5, 1.8/1.2 and 1.5/1.5. The reactants were mixed under continuous stirring. The precipitation was made in an alkaline medium, using 5 M NaOH aqueous solution, at a pH value of approximately 12. The colloidal suspension was transferred to a Teflon-lined stainless steel autoclave and sealed tightly. The autoclave was put in an air oven and maintained at 200 °C for 4 h. After decantation and filtration, the resulting precipitate was washed several times with distilled water on a filter paper and then laid out to dry in an air oven, at 105 °C. The powder samples obtained as described above were denoted by sample A, B, C and D corresponding to the molar ratios x = 2.8/0.2, 2.5/0.5, 1.8/1.2 and 1.5/1.5, respectively.

## Structural characterization of the samples

Based on the XRD pattern of sample A, one can assert that this sample consists of a mixture of Fe_{2}O_{3} and ZnFe_{2}O_{4}, with Fe_{2}O_{3} being prevalent, as all the characteristic peaks of Fe_{2}O_{3} are present and only some of the characteristic peaks of ZnFe_{2}O_{4} are slightly visible. From the EDAX quantification of the same sample (Fig. 3a) one obtains the atomic fraction Fe/Zn = 14.02, almost equal to the molar ratio of the precursors (Fe^{3+}/Zn^{2+} = 2.8/0.2). Taking into account the 5% measurement error in the EDAX quantification, as given by the apparatus manufacturer (FEI Company), one can conclude that the entire zinc and iron ions in solution were converted to oxides during the obtaining process. Bearing in mind the possible composition of sample A (i.e. ZnFe_{2}O_{4}, Fe_{2}O_{3} and ZnO), the EDAX quantification result is correlated to the result obtained from the XRD pattern, namely that sample A is a mixture of oxides with high content of Fe_{2}O_{3}.

The XRD pattern of sample B also exhibits all the characteristic peaks of Fe_{2}O_{3}, and more and stronger peaks of ZnFe_{2}O_{4} (in comparison to sample A)_{,} leading to the conclusion that sample B has a higher content of ZnFe_{2}O_{4} than sample A. The atomic fraction Fe/Zn, as determined from the EDAX quantification of sample B (Fig. 3b), is 3.29, being larger than the atomic fraction Fe/Zn of the single phase ZnFe_{2}O_{4}. Thus, XRD analysis and EDAX quantification results are in agreement, leading to the conclusion that sample B is a mixture of oxides (i.e. Fe_{2}O_{3} and ZnFe_{2}O_{4}). We can also notice that the atomic fraction Fe/Zn as determined by EDAX quantification is smaller than the molar ratio of the precursors (Fe^{3+}/Zn^{2+} = 2.5/0.5), meaning that a part of the iron ions in solution has not been converted to iron oxides; they probably remained in the form of ferric nitrate. On the other hand, bearing in mind the fact that the SEM/EDAX analysis provides rather a surface compositional analysis of the particles, another possible explanation is that the atomic fraction on the surface of the particles in the sample B may be different from the atomic fraction inside the particles.

The EDAX quantification shows that the atomic fraction Fe/Zn is 1.51 for sample C and 1.59 for sample D, showing a higher content of Zn than in the case of the single phase ZnFe_{2}O_{4.} This leads to the conclusion that in these samples there are some other Zn compounds besides ZnFe_{2}O_{4}. The XRD analysis of samples C and D presents all the characteristic peaks of the ZnFe_{2}O_{4} spinel, leading to the conclusion that this is the predominant phase, and the characteristic peaks of Fe_{2}O_{3} are no longer visible. In the case of both samples, one can notice a few small peaks characteristic to ZnO, besides the characteristic peaks of the ZnFe_{2}O_{4} spinel. Thus, the result obtained from the XRD analysis is correlated to the one obtained by EDAX quantification. Bearing in mind the 5% measurement error in the EDAX quantification, in the case of sample C one can notice that the atomic fraction Fe/Zn as determined by EDAX quantification is the same as the molar ratio of the precursors (Fe^{3+}/Zn^{2+} = 1.8/1.2), whilst in the case of sample D the atomic fraction Fe/Zn is larger than the molar ratio of the precursors (Fe^{3+}/Zn^{2+} = 1.5/1.5). Thus, a part of the zinc ions in solution has not been converted to zinc oxides in the case of sample D; they probably remained in the form of zinc nitrate. As explained in the case of sample B, another possible explanation is that the atomic fraction on the surface of the particles in the sample D may be different from the atomic fraction inside the particles.

*d*, in the analysed samples was calculated from the diffraction peaks using the Debye–Scherrer formula [4]:

*λ*is the X-ray wavelength,

*β*is the half peak reflection and

*θ*is the Bragg angle. In our case, for Cu–Kα radiation,

*λ*= 0.15418 nm. The results are given in Table 1.

Diameter of crystallites in the analysed samples

Sample A | Sample B | Sample C | Sample D | |
---|---|---|---|---|

Diameter of Fe | 26.8 | 26.7 | – | – |

Diameter of ZnFe | – | 14.8 | 11.4 | 9.8 |

## Microwave absorbent properties

*α,*and the reflection coefficient,

*R,*are usually involved in the analysis of the absorbent properties of a propagation medium (Collin 1966). As shown in (Fannin et al. 2009),

*α*and

*R*can be expressed by means of the complex magnetic permeability (\( \mu_{r} = \mu^{\prime} - j\mu^{\prime\prime} \)) and of the effective complex dielectric permittivity (\( \varepsilon_{{r,{\text{eff}}}} = \varepsilon ^{\prime} - j\left( {{\frac{\sigma }{{2\pi f\varepsilon_{0} }}} + \varepsilon ^{\prime\prime}} \right) = \varepsilon ^{\prime} - j\varepsilon ^{\prime\prime}_{\text{eff}} \)) of the analysed material. The attenuation constant is given by Eq. 3 and, in the case of normal incidence of the electromagnetic wave coming from the air and reflecting on the absorbent material surface, the reflection coefficient is given by Eq. 4 (Fannin et al. 2009).

*μ*′ is the real part of the complex magnetic permeability and

*ε*′ is the real part of the complex magnetic permittivity of the absorbent material, and the significance of the other parameters has already been mentioned.

*α*, and the reflection coefficient,

*R*, the frequency dependencies of the complex magnetic permeability and of the effective complex dielectric permittivity of the investigated samples were measured, over the frequency range 0.1–6 GHz, at room temperature. Measurements were performed by means of the coaxial line technique as detailed in Ref. (Fannin et al. 1999), and the results are given in Figs. 4 and 5.

As can be observed from Fig. 4, the frequency dependence of the complex magnetic permeability of samples A, C and D has a ferromagnetic resonance profile, distinguished by the transition in the real part, *μ*′, from values larger than unity to values smaller than unity, at the resonance frequency, *f*_{res} (Fig 4b) and by an absorption peak of the imaginary part, *μ*′′, at a frequency in the vicinity of *f*_{res} (Fig. 4a). Sample B also exhibits a maximum of the imaginary part, *μ*′′, but the real part remains larger than unity, behaviour which is characteristic of the magnetic relaxation processes.

As revealed by the XRD analysis and EDAX quantification result, Fe_{2}O_{3} is the prevalent compound in sample A. In its bulk form, it has two magnetic transition temperatures, one at 250 K and the other one at 950 K (Vonsovski 1974). Below 250 K (this temperature is referred to as the Morin transition (Morin 1950)) Fe_{2}O_{3} is in the antiferromagnetic phase and above 250 K up to the Curie temperature, of 950 K, it exhibits weak ferromagnetism (Vonsovski 1974). Even though the magnetic properties of antiferromagnetic nanoparticle systems strongly depend on the particle size (Wesselinowa 2010; Zysler et al. 2004), ferromagnetic resonance was still observed in the Fe_{2}O_{3} nanoparticles at room temperature (Owens 2009). Therefore, the ferromagnetic resonance observed in the case of sample A is due to the weak ferromagnetism of Fe_{2}O_{3}.

The XRD analysis and EDAX quantification result of sample B showed that it is a mixture of Fe_{2}O_{3} and ZnFe_{2}O_{4.} In its stoichiometric composition, zinc ferrite has a normal spinel structure with non-magnetic Zn^{2+} cations occupying tetrahedral sites and magnetic Fe^{3+} cations occupying octahedral sites and exhibits antiferromagnetic behaviour below 10 K (Katsuhisa Tanaka et al. 1998), meaning that ZnFe_{2}O_{4} is paramagnetic at room temperature. Even so, it was found that some zinc ferrite ultrafine particle systems show unusually high magnetization even at room temperature (Katsuhisa Tanaka et al. 1998; Hofmann et al. 2004). This behaviour can be explained in terms of an inversion, which may occur in the occupancy of the tetrahedral and octahedral sites by Fe^{3+} and Zn^{2+} ions, on the surface of the nanoparticles (Kamiyama et al. 1992; Jeyadevan et al. 1994). Due to the large value of the Fe/Zn ratio in the precursors used in obtaining sample B and due to the preference of the Zn ions to occupy tetrahedral sites, we may speculate that the Zn ferrite crystallite in sample B did not experience occupancy inversion, exhibiting thus no ferromagnetic-like resonance behaviour (see Fig. 4).

In sample C the predominant phase is the ZnFe_{2}O_{4} spinel, as shown by the XRD analysis and the EDAX quantification. This sample displays ferromagnetic-like resonance (see Fig. 4) which can be explained by the inversion in the spinel structure (i.e. the occupancy of the tetrahedral sites by Fe^{3+} ions accompanied with Zn^{2+} ions in the octahedral sites).

In sample D, the prevalent phase is still the ZnFe_{2}O_{4} spinel, but additional ZnO was revealed by the XRD analysis and the EDAX quantification. ZnO nanoparticle systems display ferromagnetic behaviour at room temperature (Gao Daqiang et al. 2009; Sharma Prashant et al. 2009). The ferromagnetic-like resonance behaviour of sample D is a result of the combined contributions of zinc ferrite and ZnO crystallite (see Fig. 4). The shift in the resonance frequency from 2.86 GHz in sample C to 3.55 GHz in sample D is due to the presence of ZnO in sample D and due to the different crystallite diameter of zinc ferrite (*d* = 11.4 nm in sample C and *d* = 9.8 nm in sample D).

Figure 5 presents the frequency dependence of the effective complex dielectric permittivity of the samples. All samples are mixtures of oxides; hence, the imaginary component \( \varepsilon^{\prime\prime}_{\text{eff}} \) of the complex dielectric permittivity displays more than one peak in the case of all samples, corresponding to different possible relaxation processes. The cations of Zn^{2+} and Fe^{3+} at their respective positions in the crystalline structure form electric dipoles with the surrounding O^{2−} ions, contributing to the complex dielectric permittivity within the crystalline grains. Interfacial polarization and hopping electron relaxation can also contribute to the complex dielectric permittivity in all samples, because the samples are powders.

*α*and the reflection coefficient

*R*were computed by means of Eqs. 3 and 4 using the measured frequency dependence of the complex magnetic permeability and the effective complex dielectric permittivity. The results are presented in Figs. 6 and 7.

From Fig. 6 one can observe that the attenuation constant of sample C is the largest for frequencies between 0.5 and 5 GHz, and that samples C and D present one maximum whilst samples A and B display two maxima.

*α*can be rewritten as:

From Eq. 5 a correlation between the frequency dependence of the dielectric and magnetic parameters and the frequency dependence of *α* can be established. Hence, in the case of samples A and B the term \( {\frac{{\mu^{\prime}}}{{\varepsilon^{\prime}}}}\varepsilon^{\prime \prime 2} \) gives approximately the same contribution as does the term \( {\frac{{\varepsilon^{\prime}}}{{\mu^{\prime}}}}\mu^{\prime \prime 2} \)in Eq. 5, whilst in the case of samples C and D the term \( {\frac{{\varepsilon^{\prime}}}{{\mu^{\prime}}}}\mu^{\prime \prime 2} \) prevails. Therefore, the maximum of *α* in the case of samples C and D is due to the maximum of *μ*′′, whilst in the case of samples A and B both *ε*′′ and *μ*′′ contribute to the maximum of *α*.

From Fig. 7 it can be noticed that the wave reflection at the air–sample interface is low in all four samples (less than 22% of the incident wave is reflected).

*R*and an attenuation constant

*α*) deposited on a total reflective support, the overall reflection coefficient at normal incidence is given by:

*w*is the thickness of the absorber. In Eq. 6 the first term gives the wave reflection on the incidence interface and the second one is given by the transmission through the incidence interface and the absorption of the wave whilst passing through the absorber before and after totally reflecting on the reflective support.

*R*

_{w}was computed by means of Eq. 6, for all four samples considering an absorber layer thickness

*w*= 2 mm. The results are presented in Fig. 8.

One can notice from Fig. 8 that sample A is the poorest absorber, reflecting approximately 90% of the incident wave. Within the frequency range 0.1–1.36 GHz, sample B reflects the least of the incident wave, whilst for frequencies larger than 1.36 GHz sample C is the best absorber of the four samples. The minimum of the total reflection coefficient *R*_{w} is achieved at a frequency of 3.5 GHz, where sample C reflects 78% of the incident wave. In other words, in the microwave S-band (2–4 GHz), ZnFe_{2}O_{4} is the best absorbing material from the Zn/Fe oxide powders analysed here. Also, a mixture of Fe_{2}O_{3} and ZnFe_{2}O_{4} (sample B) proves to be more absorbing than a mixture of ZnFe_{2}O_{4} and ZnO (sample D), over the same microwave S-band.

This result is of potential interest for microwave absorbers working in the microwave S-band.

## Conclusions

Four samples were prepared by the hydrothermal method, using the same precursors (i.e. ferric nitrate and zinc nitrate) in different designated molar ratios x = Fe^{3+}/Zn^{2+}. For the highest Fe^{3+}/Zn^{2+} molar ratio (x = 2.8/0.2), the prevalent compound obtained was Fe_{2}O_{3}. For smaller Fe^{3+}/Zn^{2+} molar ratios (x = 1.8/1.2 and x = 1.5/1.5), the dominant compound obtained was ZnFe_{2}O_{4}.

Frequency dependencies of the complex magnetic permeability and of the effective complex dielectric permittivity of the investigated samples were measured, over the frequency range 0.1–6 GHz, at room temperature. In samples A, C and D ferromagnetic-like resonance behaviour was observed.

Based on the dielectric and magnetic measurements, the attenuation constant *α* and the reflection coefficient *R* were computed over the investigated frequency range. The overall reflection coefficient at normal incidence on a thin coating of an absorber deposited on a total reflective support was computed. For the frequency range 0.1–1.36 GHz, sample B was found to be the best absorber, whilst for higher frequencies sample C has the best absorbent characteristics, of the four investigated samples.

## Acknowledgements

P. C. Fannin acknowledges ESA for part funding of this study, whilst the other authors gratefully acknowledge partial financial support from CNMP grant no. 1-32155/2008. Acknowledgements are also due to A. Bucur for the support in performing the XRD measurements.