RESEARCH PAPER

Raman study of cations’ distribution in ZnxMg12xFe2O4

nanoparticles

S. W. da Silva • F. Nakagomi • M. S. Silva •

A. Franco Jr • V. K. Garg • A. C. Oliveira •

P. C. Morais

Received: 19 September 2011 / Accepted: 25 February 2012

� Springer Science+Business Media B.V. 2012

Abstract In a complementary way, Raman and

Mossbauer spectroscopy were successfully employed

to assess the cations’ distribution among the tetrahe-

dral (A-site) and octahedral (B-site) sites of nonosized

ZnxMg1-xFe2O4 (0 B x B 1) cubic ferrite structure,

synthesized by combustion reaction method. Nano-

particles with little change in size distributions, in the

40 nm (x = 0.0) up to 42 nm (x = 1.0) were

obtained. Mossbauer data indicated that as the Zn-

content (x) increases in the range 0 B x B 1, the Fe3?

ion monotonically increases (decreases) the A-site

(B-site) occupancy up to nearly equal values at the

highest end x value. Analysis of the Raman data,

however, confirms that the three highest energy modes

around 650, 668 and 710 cm-1 are assigned to Zn–O

(B-site), Fe–O (A-site) and Mg–O (A-site) vibrations,

respectively. Additionally, in agreement with the

Mossbauer data, the Raman data show that as

the Zn-content (x) increases in the range 0 B x B 1,

the occupancy of A-sites by Mg2? ions monotonically

reduces with concomitant increase of A- and B-sites

occupancy by Fe3? and Zn2? ions, respectively.

Indeed, combination of the two sets of spectroscopic

data (Raman and Mossbauer) provides an effective

protocol for assessing the cations’ distribution within

the crystal structure of nanosized quaternary cubic

ferrite samples running for instance from Fe3þ0:42

�

Mg2þ0:58�

AZn2þ

0:20Mg2þ0:22Fe3þ

1:58

� �BO2�

4 at x = 0.2 up to

Fe3þ1:0

� �AZn2þ

0:60Mg2þ0:40Fe3þ

1:0

� �BO2�

4 at x = 0.6.

Keywords Raman spectroscopy � Cubic ferrite �Magnetic nanoparticles � Cation’ distribution �Mossbauer spectroscopy

Introduction

The crystal structure of cubic ferrites (MFe2O4;

M = Fe, Co, Ni, Mn, Mg, Zn,���) consist of a face

centered cubic (fcc) lattice of oxygen anions within

which cations (Fe3? and M2?) occupy tetrahedral

(A-sites) and octahedral (B-sites) interstitial sites

arranged in two possible extreme patterns: normal

and inverse. In the normal cubic ferrite structure, the

A-sites are occupied exclusively by M2? cations while

the B-sites are occupied exclusively by Fe3? cations. In

bulk form ZnFe2O4 is a typical normal-like cubic ferrite

with all Fe3? ions on B-sites and all Zn2? ions on

A-sites (O’Neill 1992). In the inverse cubic ferrite

Nanoparticles with little change in size distributions, in the 40

(x = 0.0) up to 42.

S. W. da Silva (&) � F. Nakagomi �V. K. Garg � A. C. Oliveira � P. C. Morais

Instituto de Fısica, Universidade de Brasılia, C.P. 04455,

Brasılia, DF 70919-970, Brazil

e-mail: swsilva@unb.br

M. S. Silva � A. Franco Jr

Instituto de Fısica, Universidade Federal de Goias,

Goiania, GO 74001-970, Brazil

123

J Nanopart Res (2012) 14:798

DOI 10.1007/s11051-012-0798-4

structure the A-sites are occupied exclusively by Fe3?

cations while occupation of the B-sites is shared by both

M2? and Fe3? cations, possibly in a random fashion. It

has been shown that in MgFe2O4 the site preference of

the divalent ions leads to a predominantly inverse

structure, with Mg2? ions mainly on B-sites while Fe3?

ions are distributed almost equally among A and B-sites

(McCurrie (1994). In between these two ends there are

cubic ferrites in which M2? ions can occupy both sites

and, therefore, are classified according to their degree

of inversion d (relative occupancy of M2? ions on

B-sites); positioning between normal (d = 0) and

inverse (d = 1). Such changes in cationic order can

be described through the formula [M1-dFed]A[MdFe2-

d]BO4, indicating that some M2? ions in the structure

have switched with some Fe3? ions from the B-sites to

the A-sites at a d degree. Such intermediated structures

are known as mixed ferrites. It is indeed reported that

the vast majority of cubic ferrites at the nanosized scale

are mixed (Soler et al. 2007; Sepelak et al. 2006;

Makovec et al. 2011). Cubic ferrites may also contain a

mixture of divalent metal ions like Zn–Mg ferrites (da

Silva et al. 2010), Co–Zn ferrites (Gul et al. 2007), Mn–

Zn ferrites (Rath et al. 1999). These quaternary

compounds have been prepared by simple substitution

of metal cations using chemical methods and are

extremely rich in terms of magnetic and magneto-

optical properties, which depend upon their degree of

disorder (Antic et al. 2010).

There are many experimental methods that can be

used to assess the degree of disorder in quaternary

spinel systems and most of them are addressed to the

determination of the d parameter. X-ray absorption

(O’Neill 1992; Oliver et al. 2000), neutron and X-ray

diffraction have been reported among others (Braestrup

et al. 2008). In particular, Mossbauer spectroscopy has

been widely used to determine the cations’ distribution

in cubic ferrites (Sepelak et al. 2006). Nevertheless,

Mossbauer spectroscopy alone cannot be used to

determine the cations’ distribution in quaternary spi-

nels, such as in ZnxMg1-xFe2O4. This is because of the

limitation of the Mossbauer spectroscopy in probing

iron ions only. In this regard it would be very much

interesting to explore the potentialities of employing

Raman spectroscopy as a complementary tool to the

Mossbauer spectroscopy. More specifically, it would be

worth to investigate the outcomes of the analysis while

combining the two set of recorded data, namely data

from Raman and Mossbauer spectroscopy.

Vibrational spectroscopy, both infrared and Raman,

are powerful tools for direct probing of lattice dynamics

in many different compounds. They are both widely

used as qualitative and quantitative analytical tech-

niques and to assess information regarding the chemical

structure and three-dimensional supramolecular struc-

ture. Actually, Raman spectroscopy has been recently

used to determine the cations’ distribution in magne-

sium ferrite (MgFe2O4) (Nakagomi et al. 2009). In this

study Raman and Mossbauer spectroscopy are success-

fully used, in a complementary way, to determine the

cations’ distribution in a series of quaternary cubic

ferrite samples (ZnxMg1-xFe2O4, with 0.0 B x B 1.0).

Experimental

Magnesium-zinc ferrite nanoparticulate samples

(ZnxMg1-xFe2O4, with 0.0 B x B 1.0) were synthesized

by the combustion reaction method without subsequent

calcination steps (Franco et al. 2007). Analytical grade

iron nitrate Fe(NO3)3�9H2O, zinc nitrate Zn(NO3)2�6H2O, magnesium nitrate Mg(NO3)2�6H2O and urea

CO(NH2)2 were used as fuels. Materials manipulation

and chemical reactions were carried out in atmospheric

air. The stoichiometric composition of each mixture was

calculated based on the total oxidizing and reducing

valences of both oxidizer and fuel. The chemical

composition (total zinc and magnesium contents) of all

ferrite samples was assessed by atomic absorption

spectrophotometry using the commercial Perkin-Elmer

5000 system. The nominal compositions were in good

agreement with the core chemical compositions obtained

by atomic spectroscopy (see Table 1).

Information regarding the samples’ crystal struc-

ture, lattice constant and average nanoparticle sizes

were obtained by room temperature X-ray diffraction

(XRD) using a Shimadzu diffractometer (model 6000)

with Cu-Ka (k = 1.54 A) radiation while scanning the

spectra in a wide range of Bragg angles (15� \ 2h\ 80�). Low temperature (77 K) Mossbauer spectra of

all samples were recorded in the transmission geom-

etry using a 57Co source in Rh matrix. The system

velocity was calibrated with a thin natural iron foil

whereas the spectra were least-square fitted to a

combination of Lorentzian-like lines. The Raman

spectra were recorded using a commercial Jobin–

Yvon triple spectrometer (T64000) equipped with a

CCD detector. The 514 nm line of a CW Argon ion

Page 2 of 10 J Nanopart Res (2012) 14:798

123

laser was used to excite the samples whereas the

optical excitation intensity was kept around 0.2 mW.

All Raman measurements were performed at room

temperature.

Results and discussions

The XRD patterns of the as-synthesized samples

(ZnxMg1-xFe2O4, with 0.0 B x B 1.0) are collected

in Fig. 1. Analysis of all XRD patterns revealed the

cubic ferrite structure ðFd�3mÞ; in agreement with the

used synthesis route. The absence of extra reflections

in the diffraction patterns ensures phase purity. The

mean particle sizes (D) were calculated from the X-ray

line broadening of the (311) diffraction peak using the

Scherrer’s equation: D = 0.9 k/b (cosh). From the

analysis of the XRD spectra we found that the mean

particle diameter (D) is nearly the same for all

synthesized specimens, ranging from ca. 40 nm (x =

0.0) up to 42 nm (x = 1.0). This finding means that the

increasing of Zn-content in the range of 0.0 B x B 1.0

does not affect much the average grain size of the as-

synthesized nanoparticulate ferrite samples (ZnxMg1-

xFe2O4). Nevertheless, the XRD data show that the

lattice parameter (a) changes linearly with the Zn-

content (x), increasing from a = 8.365 A to a =

8.430 A with the Zn-content increasing from x = 0.0

to x = 1.0, thus following roughly the Vegard’s law

(Denton and Ashcroft 1991). This behavior can be

attributed to the larger ionic radius of Zn2? (0.74 A—

A-sites and 0.88 A—B-sites) as compared to the ionic

radius of the Mg2? (0.71 A—A-sites and 0.80 A—B-

sites). The incorporation of larger ions into the lattice

of the nanoparticles would expand the lattice and

increase the observed lattice parameter. The values

found for the lattice parameter of the as-synthesized

nanoparticulate samples are close to the values

corresponding to the bulk ferrite (MgFe2O4–JCPDS

card #73-2211, a = 8.366 A and ZnFe2O4–JCPDS

card #22-1012 a = 8.441 A).

The low temperature (77 K) Mossbauer spectra of

all synthesized samples (ZnxMg1-xFe2O4, with 0.0 B

x B 1.0) show a systematic variation as the Zn-content

(x) increases (see Fig. 2). In order to investigate the

effect of the increasing x-content all Mossbauer

spectra were curve-fitted with Lorentzian-like com-

ponents, taking into account the two sublattices

(A- and B-sites). For smaller Zn-content (0.0 B x B

0.4) the Mossbauer spectra were accomplished using

two sextets, corresponding to 57Fe ions located at the

A- and B-sites. At intermediate x values (0.5 B x B

0.7) we observed broad Mossbauer spectra (two

sextets), which can be explained considering the

breakdown of the local magnetic order due to the

increase in Zn-content. At higher Zn-content, say

Table 1 Nominal stoichiometry based on the relative amounts

of the starting materials and chemical composition obtained by

atomic spectroscopy

Nominal (x) Experimental

ZnxMg1-xFe2O4 x (Zn) (1 - x)(Mg)

0.0 0.000 0.999

0.2 0.192 0.808

0.4 0.390 0.610

0.5 0.487 0.513

0.6 0.617 0.383

0.7 0.699 0.300

0.8 0.812 0.188

1.0 0.998 0.000

20 30 40 50 60 70 80

0.5

0.4

0.2

2θ (degree)

0.6

0.7

0.0di

ffrac

tion

Inte

nsity

x = 1.0

0.8

(400

)

(422

) (511

)

(440

)

(533

)

(220

)

(311

) ZnxMg

1-xFe

2O

4

Fig. 1 X-ray diffraction patterns of the as-synthesized

ZnxMg1-xFe2O4 samples, with 0 � x � 1

J Nanopart Res (2012) 14:798 Page 3 of 10

123

x = 0.8 and 1.0, the Mossbauer spectra exhibit a

strong doublet superimposed to a weak sextet, as

expected from a magnetically disordered phase. The

Mossbauer parameters resulting from the least-squares

fitting of the recorded spectra, taking into account the

two sublattices of cubic ferrites, were plotted as

function of the Zn-content (see Fig. 3). The values of

the hyperfine field (HF), isomer shift (IS), quadruple

splitting (QS) and relative Fe-content for 0.0 B

x B 0.8 are shown in Fig. 3a–d, respectively.

As can be seen from the data plotted in Fig. 3a the HF

associated to both crystallographic sites (A- and B-sites)

systematically decrease as the Zn-content (ZnxMg1-

xFe2O4, with 0.0 B x B 1.0) increases. This finding is

partially explained by the reduction of the first nearest

neighbor (A–B) superexchange interaction strength due

to the redistribution of the magnetic Fe3? ions among

A- and B-sites. Additionally, the systematic decrease of

the HF is due to the weakening of the A–B superex-

change interaction owing to the increase in distance

between A- and B-sites as a result of the increase in the

lattice parameter with increasing Zn-content (increasing

x value) (Srivastava et al. 1976; Hamedoun et al. 2010).

Furthermore, Fig. 3b shows the isomer shift behavior

(A- and B-sites) with increasing Zn-content; monoton-

ically increasing (decreasing) for the B-site (A-site) up to

about x = 0.5. For x C 0.5 the IS values reverse their

trends; monotonically decreasing (increasing) for the

B-site (A-site) in the range of 0.5 B x B 0.8, though

keeping the IS values in the vicinity of 0.43 mm/s. The

IS behavior can be explained through the bonding nature

of the Fe3? ion while occupying both sites. With

increasing Zn-content up to about x = 0.6 the Fe3?

population in the A-site increases while decreasing in the

B-site, as shown in Fig. 3d. As the ionic radius of Fe3?

(0.63 A—A-site and 0.78 A—B-site) is smaller than the

ionic radii of Zn2? and Mg2? (see ‘‘Experimental’’

section) there is an expected shrinking of the tetrahedral

coordination (A-site) with consequent increase in the

iron–oxygen orbitals’ overlapping, thus resulting in the

decrease of the IS values associated to the A-site.

Conversely, up to about x = 0.6 the B-site experiences a

systematic reduction of the Fe3? population while

increasing the Zn2? population, thus owing to an expect

expansion of the octahedral coordination with a conse-

quent decrease in the iron–oxygen orbitals’ overlapping.

This trend results in the increase of IS values associated

to the B-site. Notice, however, that above about x = 0.6

Fig. 2 Low temperature (77 K) Mossbauer spectra of the

ZnxMg1-xFe2O4 samples, with 0 � x � 1. Open circles are the

experimental data while the solid lines represent the best fit

using one doublet or two sextet which represent A- and B-sites

(dashed lines)

20

40

60

80

0.36

0.40

0.44

0.48

0.52

350

400

450

500

550Zn

xMg

1-xFe

2O

4

A siteB site

(b)

Z n - content (x)

(c)

(a)

QS

(m

m/s

)

IS (

mm

/s)

HF

(ke

O)

Fe

popu

latio

n (%

)

0.0 0.2 0.4 0.6 0.8 1.00.0 0.2 0.4 0.6 0.8 1.0-0.3

-0.2

-0.1

0.0

0.1 (d)

Fig. 3 Zinc-content dependence of a hyperfine field (HF)

b isomer shift (IS) c quadrupolar splitting and d Fe population

(%) obtained by Mossbauer spectra recorded from the

ZnxMg1-xFe2O4 samples, with 0 � x � 1. The solid and opensymbols represent A- and B-site, respectively. The dashed linesare only guide to the eyes

Page 4 of 10 J Nanopart Res (2012) 14:798

123

little changes were observed in both IS values and

relative Fe3? ion site occupancy. Figure 3c shows that

the cubic symmetry of the oxygen coordination in both

sites (A- and B-sites) is monotonically affected as one

replaces Mg2? by Zn2? ions. Further, deviation from

zero QS values in both sites indicates that the oxygen

cubic symmetry around the Fe3? ions is indeed broken.

Nevertheless, introduction of Zn2? ions affects differ-

ently the cubic symmetry of A- and B-sites. In the whole

range of Zn-content (x) investigated small QS values

were found in the A-site, indicating that replacement of

Mg2? by Zn2? ions did not change appreciably the site

symmetry. However, we found a different behavior

associated to the B-site, since the QS value is close to

zero for x = 0 but increases (negative values) system-

atically with increasing Zn-content. This behavior

shows that the B-site is more strongly affected than the

A-site upon replacement of Mg2? by Zn2? ions. Finally,

Fig. 3d shows the evolution of the area under the

Mossbauer subspectra as the Zn-content increases in the

range of 0.0 B x B 0.7. Therefore, as shown in Fig. 3d,

increasing the Zn-content the Fe3? occupancy system-

atically increases at the A-site while decreasing at the

B-site. This finding indicates that lower x values favor

the inverse-like structure with most of the Fe3? ions

occupying B-sites. However, increasing the Zn-content

above about x = 0.6 the Fe3? ions tend to be equally

distributed in both sites. This finding indicates that for

large x values (x C 0.6) the A-site is mostly occupied by

Fe3? ions whereas the B-site is occupied by Fe3?, Zn2?,

and Mg2? ions. Unfortunately, in quaternary cubic

ferrites Mossbauer data alone do not provide enough

information for complete site occupancy determination,

as two non-Mossbauer isotopes are included in the

crystal structure, as for instance Zn2? and Mg2?.

Additionally, at higher Zn-content (x C 0.8), the pres-

ence of the Mossbauer paramagnetic doublet makes it

impossible to assign the cations’ distribution to each

available site. Nevertheless, considering the Mossbauer

data presented in Fig. 3d and taking into account charge

balance requirements it was possible to propose a partial

cations’ distribution for the as-synthesized samples.

Based on the Mossbauer data we included in Table 2 the

partial cations’ distribution for each synthesized

ZnxMg1-xFe2O4 sample (upper row). Notice, however,

that in Table 2, M means the divalent cation, with no

precise assignment to Zn2? or Mg2?.

As mentioned previously, Raman spectroscopy

has shown to be a powerful tool for direct probing

of lattice dynamics of many different compounds.

Table 2 Cations’ distribution

as obtained by Mossbauer

(upper row) and Raman

spectroscopy (lower row).

M means the divalent cation,

with no precise assignment to

Zn2? or Mg2?

(x) Zn Technique Cations’ distribution

0 Mossbauer Mg2þ0:47Fe3þ

0:53

� �AMg2þ

0:53Fe3þ1:47

� �BO2�

4

Raman Mg2þ0:48Fe3þ

0:52

� �AMg2þ

0:52Fe3þ1:48

� �BO2�

4

0.2 Mossbauer M2þ0:60Fe3þ

0:40

� �AM2þ

0:40Fe3þ1:60

� �BO2�

4

Raman Mg2þ0:58Fe3þ

0:42

� �AZn2þ

0:20Mg2þ0:22Fe3þ

1:58

� �BO2�

4

0.4 Mossbauer M2þ0:33Fe3þ

0:67

� �AM2þ

0:67Fe3þ1:33

� �BO2�

4

Raman Mg2þ0:52Fe3þ

0:48

� �AZn2þ

0:40Mg2þ0:08Fe3þ

1:52

� �BO2�

4

0.5 Mossbauer M2þ0:30Fe3þ

0:70

� �AM2þ

0:70Fe3þ1:30

� �BO2�

4

Raman Mg2þ0:28Fe3þ

0:72

� �AZn2þ

0:50Mg2þ0:22Fe3þ

1:28

� �BO2�

4

0.6 Mossbauer M2þ0:06Fe3þ

0:94

� �AM2þ

0:94Fe3þ1:06

� �BO2�

4

Raman Fe3þ1:00

� �AZn2þ

0:60Mg2þ0:40Fe3þ

1:00

� �BO2�

4

0.7 Mossbauer M2þ0:04Fe3þ

0:96

� �AM2þ

0:96Fe3þ1:04

� �BO2�

4

Raman Fe3þ1:00

� �AZn2þ

0:70Mg2þ0:30Fe3þ

1:00

� �BO2�

4

0.8 Mossbauer –

Raman Fe3þ1:00

� �AZn2þ

0:80Mg2þ0:20Fe3þ

1:00

� �BO2�

4

1.0 Mossbauer –

Raman Fe3þ1:00

� �AZn2þ

1:00Fe3þ1:00

� �BO2�

4

J Nanopart Res (2012) 14:798 Page 5 of 10

123

According to the literature (Nakagomi et al. 2009;

Seong et al. 2001), as the Raman integrated intensity is

proportional to the number of the correspondent

oscillators the technique can be successfully used to

determine quantitatively or at least semi-quantita-

tively the content of a particular element in a given

sample. Therefore, in the present study, we selected

Raman spectroscopy as a complementary tool to the

Mossbauer spectroscopy in the task to assess the

cations’ distribution in the as-synthesized ZnxMg1-

xFe2O4 quaternary samples. Actually, Raman spec-

troscopy has been used in the study of cubic ferrites for

more than three decades (Verble 1974). A symmetry

analysis and the assignment of phonon modes to

magnetite, based on the spinel structure, were carried

out by Verble (1974); Shebanova and Lazor (2003);

Degiorgi et al. (1987) and Graves et al. (1988). Results

from these studies vary significantly either in the

number of expected Raman modes or with respect to

their positions and assignments. A major difficulty in

studying this system is related to the order–disorder

phenomena, generally observed in the spinel structure.

The cations’ exchange among the available A- and

B-sites gives rise to different structural configurations.

As a result, the Raman spectra recorded from different

samples, though presenting the same stoichiometry are

qualitatively different. Therefore, a very important

step while assessing information on cation ion site-

occupancy is to make an appropriate assignment of the

vibrational modes, though this task is an old problem

while related to cubic ferrites.

The cubic ferrite’s family, herein represented by

MFe2O4, has a cubic structure (fcc) belonging to the

space group Fd�3m. The theoretical analysis based on

the factor-group approach predicts five Raman-active

bands, namely A1g;Eg; and three T2g (White and

DeAngeli 1967). Graves et al. (1988) suggested a new

irreducible representation based on a symmetry

brought about lattice defects which leads to a splitting

of one T2g mode into A1g þ Eg representation, thus

justifying the presence of an additional vibrational

mode in the Raman spectra. On the other hand,

considering the treatment of the vibrational modes of

ferrites in terms of a molecular model, as proposed by

Waldron (1955), only sites of symmetry Td and C2v;,

respectively, occupied by Fe3? (A-site) and O2- ions

effectively contribute to the Raman spectrum. There-

fore, as confirmed in the case of inverse spinel ferrite,

such as MgFe2O4, NiFe2O4 and Fe3O4, only the A-site

contribute to the Raman spectrum (Verble 1974;

Shebanova and Lazor 2003; White and DeAngeli

1967). Although the Raman spectra of the inverse

spinels are different, they have a common feature; a

strong A1g band in the 670–710 cm-1 region. This

band is observed regardless the particular bivalent

cation in the sample’s stoichiometry and has been

assigned to the stretching vibrational of the tetrahedral

FeO4 (Kreisel et al. 1998). In the case of normal spinel

ferrites, MFe2O4 (for instance M = Zn or Mn), with

M2? ions occupying A-sites while Fe3? ions fill in

B-sites, it has been observed that the A1g band arises in

the 600–620 cm-1 region (Graves et al. 1988; Kreisel

et al. 1998). No other band has been observed at higher

frequency. Furthermore, this band arises in the

600–620 cm-1 region regardless the particular biva-

lent cation. In analogy with the Raman spectrum of

hematite (highest band at about 600 cm-1), in which

only FeO6 B-sites are found in the crystalline structure,

Kreisel et al. (1998) proposed that this band is

essentially related to the Fe–O vibrational of the

FeO6 octahedron. However, Wang et al. (2003) while

reporting on zinc ferrite considered that the band

observed at 647 cm-1 is due to the oxygen motion in

the A-site. Similar statement was made by Maletin

et al. (2007) while studying ZnxIn1-xFe2O4.

Figure 4 presents the room temperature Raman

spectra of the as-synthesized ZnxMg1-xFe2O4 samples

(0.0 B x B 1.0), recorded in the range of 200–850

cm-1. The Raman modes A1g þ Eg þ 3Tg

� �; charac-

teristic of the cubic spinel Fd�3mð Þ space group, have

been assigned to the features appearing on the

200–850 cm-1 region. The Raman features indicated

in Fig. 4 (dotted lines) are the vibrational modes

typical of the nanoparticle’s crystalline structure.

Raman features other than those related to the spinel

phase (MFe2O4) were not observed, as for instance

from a-Fe2O4, FeO, MgO, or ZnO. Although the

factor-group analysis predicts the presence of only five

Raman modes our fitting procedure, using Lorentzian-

like lines, reveals seven features instead. For the

MgFe2O4 sample (see dotted lines in Fig. 4, for

x = 0), the observed vibrational modes are around

218, 335, 388, 487, 550, 668, and 710 cm-1. The

presence of seven vibrational modes, however, are

consistent with recent data reporting on magnesium

ferrite (Nakagomi et al. 2009; Wang et al. 2002).

Page 6 of 10 J Nanopart Res (2012) 14:798

123

Indeed, for cubic ferrites there is an acceptable

consistency in regard to the strongest Raman modes

peaking around 710, 487 and 335 cm-1. However, the

agreement for the remaining Raman modes is less

satisfactory (Shebanova and Lazor 2003; Gupta et al.

2002). It was shown in the work of Nakagomi et al.

(2009) that the presence of two Raman modes above

670 cm-1 are mainly associated to the motion of

oxygen atoms in the tetrahedral complex (MO4),

where some Fe3? ions are replaced by Mg2? ions.

The Raman mode featuring at 218 cm-1 is no

longer observed in the as-synthesized samples at

x C 0.2. However, for x C 0.2, a new Raman mode

appears around 650 cm-1 (gray) as can be observed in

the spectra included in Fig. 4. Note that as the

Zn-content increases in the range of 0.2 B x B 1.0

the three highest wavelength Raman modes (see Fig. 4

for light gray, gray and black emphasized Raman

lines) change in relative intensity. Additionally, a

systematic shift of the vibrational energy of the highest

wavelength Raman modes is also observed. Changes

on relative intensity and peak position can be

explained considering alloy effects, resulting from

the introduction of a new ion at increasing content

(Pusep et al. 1995). Further, the Raman frequency

depends on the Fe(Mg)–O bond length, which changes

with both the variation of the lattice parameter and

ionic radii of divalent ions. Systematic variation of the

relative integrated intensity and peak position of the

three emphasized Raman modes (see Fig. 4) are

collected in Fig. 5. Figure 5a, b clearly show that

while the intensity of the highest wavelength Raman

mode (initially around 710 cm-1) decreases with

increasing Zn-content the intensity of the Raman

mode peaking around 670 cm-1 increases proportion-

ally. Simultaneously, it was observed that the intensity

of the Raman mode peaking around 650 cm-1

increases steeply as the Zn-content (x) increases. At

this point we claim that due the large mass difference

between the two ions (Fe3? and Mg2?) leads to the

splitting of the A1g mode into two branches. Therefore,

200 300 400 500 600 700 800

0.5

0.4

0.2

Wavenumber (cm-1)

0.6

0.0

Ram

an In

tens

ity

x = 1.0

0.8

ZnxMg

1-xFe

2O

4

Fig. 4 Raman spectra of the ZnxMg1-xFe2O4 samples, with

0 � x � 1. The dashed lines show the Lorentzian fit of the data.

The structure in black (light gray) and gray represent the bonds

at the Mg(Fe)O4 at A-site and ZnO6 at B-site, respectively

(a) (d)

(b) (e)

(c) (f)

Fig. 5 The left panel shows the integrated intensity (%) of

Raman modes associated to the bonds a Mg�O4�Mð1;2Þ12 and

b Fe�O4�Mð1;2Þ12 at A-sites and c Zn�O6�M

ð1;2Þ6 at B-sites. The

right panel shows the Raman shifts associated to the bonds

d Mg�O4�Mð1;2Þ12 (5 cm-1) e Fe�O4�M

ð1;2Þ12 (23 cm-1), and

f Zn�O6�Mð1;2Þ6 (11 cm-1), where Mð1Þ= Mg and Mð2Þ = Fe.

The dashed lines are only guide to the eyes

J Nanopart Res (2012) 14:798 Page 7 of 10

123

the modes around 710 (black) and 668 cm-1 (light

gray) can be attributed to the Mg–O and Fe–O

vibrations. Furthermore, the onset of the Raman

feature around 650 cm-1 at x C 0.2, with intensity

increasing as the Zn-content (x) increases, is a strong

evidence that this mode is related to Zn–O bond

vibration. However, the task of assigning the sublat-

tice (A- or B-site) to this particular Raman mode

(650 cm-1) is not straightforward. The first hypothesis

would be related to the well-know strong A-site

preference of Zn2? ions in bulk ferrites. In this case it

would be reasonable to expect that with increasing Zn-

content the Fe3? ion tends to migrate from A- to B-site.

However, the Mossbauer data show that the Fe3? ion

population increases in the A-site with increasing x, as

represented by the solid circles in Fig. 3d. Addition-

ally, the Mossbauer data suggest that at x C 0.6 the A-

site is mainly occupied by Fe3? ions. Therefore, the

hypothesis of Zn2? ions entering the A-site of the as-

synthesized samples has no experimental support.

Alternatively, a second hypothesis for the onset of the

Raman mode around 650 cm-1 would be based on the

proposal put forward by Kreisel et al. (1998), in which

this Raman band is associated to the Fe–O vibration in

B-sites (FeO6 complex). Again, as can be observed in

Fig. 3d (see open circles), the Mossbauer data show

that the Fe3? population decreases at B-sites as the Zn-

content increases. Additionally, the Raman data show

that the Raman peak around 650 cm-1 increases with

the increasing of the Zn-content. Therefore, like the

first hypothesis, the second one has no experimental

support. Finally, we claim that the Raman band

appearing around 650 cm-1 is associated with the

presence of Zn2? ions at B-sites. This hypothesis,

generally unusual for bulk ferrites, has been recently

reported for nanosized particles (Upadhyay et al.

2004; Choi et al. 2006).

According to the above discussion, we came to the

conclusion that the Raman mode around 650 cm-1 is due

to the Zn–O vibration with Zn2? ions populating B-sites

whereas the two modes at higher energy (*710 and

*668 cm-1) present A1g symmetry and can be

assigned to stretching modes of the A-site. In short,

the Raman mode around 710 cm-1, A1g Mgð Þ; is

associated with Mg2? ions located on A-sites

which are surrounded by twelve nearest metal ions

located on B-sites (Mg�O4�Mð1;2Þ12 ; with M(1) = Mg

and M(2) = Fe). Likewise, the A1g Feð Þ Raman mode

(*668 cm-1) is claimed to be associated with Fe3?

ions located on A-site (Fe�O4�Mð1;2Þ12 ). Once the

integrated intensity of the A1g Mgð Þ and A1g Feð ÞRaman modes are, respectively, proportional to the

number of Mg�O4�Mð1;2Þ12 and Fe�O4�M

ð1;2Þ12 bonds

we use the corresponding integrated intensities (IMg)

and (IFe) to assess the Mg- and Fe-content in the A-site.

Therefore, we can obtain both the Fe-content

xARaman Feð Þ

� �and the Mg-content xA

Raman Mgð Þ ¼�

1 � xARaman Feð ÞÞxA

Raman in the A-site from the Raman

data as follows:

xARaman Feð Þ ¼ aIFe

aIFe þ bIMg

; ð1Þ

where a and b represent the oscillator strength of the

Fe�O4�Mð1;2Þ12 and Mg�O4�M

ð1;2Þ12 bonds, respec-

tively. Nakagomi et al. (2009), while using Raman

spectroscopy to study the cations’ distribution within

the system MgxFe3-xO4, showed that the oscillator

strength involving Fe�O4�Mð1;2Þ12 bonds is twice the

oscillator strength of Mg�O4�Mð1;2Þ12 bonds. There-

fore, the relative oscillator strength is given by R =

b/a = 0.5. Finally, using Eq. 1 with R = 0.5 and

taking into account charge balance requirements it was

possible to assess the total cations’ distribution of all

synthesized samples, as shown in Table 2 (lower

rows). Note form the data displayed in Table 2 that the

Fe3? ion distribution obtained by both Mossbauer and

Raman spectroscopy agree quite well. The calculated

values of xARaman Feð Þ versus the obtained values using

Mossbauer spectroscopy xAMossbauer Feð Þ

� �are plotted in

Fig. 6. Note the excellent correlation factor (1.1)

between the two sets of data presented in Fig. 6,

indicating the robustness of the approach of using the

combination of Raman and Mossbauer data to assess

the cation site distribution in quaternary cubic ferrites.

Conclusion

Nanosized ZnxMg1-xFe2O4 (0 B x B 1) quaternary

cubic ferrite particles were synthesized by the com-

bustion reaction method whereas the cations’ distri-

bution was successfully assessed by combining

Raman and Mossbauer spectroscopy while taking

into account charge balance requirements. Due to the

Page 8 of 10 J Nanopart Res (2012) 14:798

123

large mass difference between Fe3?, Zn2? and Mg2?

ions the A1g Raman mode splits into three branches,

allowing us to quantify the cations’ distribution for the

as-synthesized ZnxMg1-xFe2O4-based samples using

the Raman data. We found the cations’ distribution

obtained by Raman spectroscopy in very good agree-

ment with the partial cations’ distribution obtained by

Mossbauer spectroscopy. Analyses of the Mossbauer

data show an increase in the orbital overlapping of the

Fe3? ions in A-sites, while decreasing in B-sites,

resulting in a systematic change of the isomer shift.

Also, it was verified that the hyperfine field decreases

monotonically with increasing Zn-content (x), which

is claimed to be related to the progressive redistribu-

tion of the Fe3? ions among A- and B-sites as well as

due to the increase of the lattice parameter. The Raman

data revealed the presence of Fe3? and Mg2? ions in

the A- and B-sites, whereas Zn2? ions were found only

in the B-site. We finally came to the conclusion that the

combination of the two sets of data (Raman and

Mossbauer) provides a robust protocol for assessing

the cations’ distribution within the crystal structure of

nanosized quaternary cubic ferrite samples.

Acknowledgments The authors acknowledge the financial

support from the Brazilian agencies CNPq, CAPES, and FINEP.

References

Antic B, Jovic N, Pavlovic MB, Kremenovic A, Manojlovic D,

Vucinic-Vasic M, Nikolic AS (2010) Magnetization

enhancement in nanostructured random type MgFe(2)O(4)

spinel prepared by soft mechanochemical route. J Appl

Phys 107:043525. doi:10.1063/1.3319563

Braestrup F, Hauback BC, Hansen KK (2008) Temperature

dependence of the cation distribution in ZnFe2O4 measured

with high temperature neutron diffraction. J Solid State

Chem 181:2364–2369. doi:10.1016/j.jssc.2008.05.028

Choi EJ, Ahn Y, Song KC (2006) Mossbauer study in zinc

ferrite nanoparticles. J Magn Magn Mater 301:171–174.

doi:10.1016/j.jmmm.2005.06.016

da Silva SW, Nakagomi F, Silva MS, Franco A, Garg VK,

Oliveira AC, Morais PC, Lima ECD (2010) Effect of the

Zn content in the structural and magnetic properties of

ZnxMg1 - xFe2O4 mixed ferrites monitored by Raman and

Mossbauer spectroscopies. J Appl Phys 107:09B503. doi:

10.1063/1.3350903

Degiorgi l, Blattermorke I, Wachter P (1987) Magnetite–pho-

non modes and the Verwey transition. Phys Rev B 35:

5421–5424. doi:10.1103/PhysRevB.35.5421

Denton AR, Ashcroft NW (1991) Vegard law. Phys Rev A

43:3161–3164. doi:10.1103/PhysRevA.43.3161

Franco A, Lima ECD, Novak MA, Wells PR (2007) Synthesis of

nanoparticles of CoxFe(3-x)O4 by combustion reaction

method. J Magn Magn Mater 308:198–202. doi:10.1016/

j.jmmm.2006.05.022

Graves PR, Johnston C, Campaniello JJ (1988) Raman-scat-

tering in spinel structure ferrites. Mater Res Bull

23:1651–1660. doi:10.1016/0025-5408(88)90255-3

Gul IH, Abbasi AZ, Amin F, Anis-ur-Rehman M, Maqsood A

(2007) Structural, magnetic and electrical properties of

Co1-xZnxFe2O4 synthesized by co-precipitation method.

J Magn Magn Mater 311:494–499. doi:10.1016/j.jmmm.

2006.08.005

Gupta R, Sood AK, Metcalf P, Honig JM (2002) Raman study of

stoichiometric and Zn-doped Fe3O4. Phys Rev B 65:104

430. doi:10.1103/PhysRevB.65.104430

Hamedoun M, Benyoussef A, Bousmina M (2010) Magnetic

properties and phase diagram of Zn(x)Ni(1-x)Fe(2)O(4):

high-temperature series expansions. J Magn Magn Mater

322:3227–3235. doi:10.1016/j.jmmm.2010.05.030

Kreisel J, Lucazeau G, Vincent H (1998) Raman spectra and

vibrational analysis of BaFe12O19 hexagonal ferrite. J Solid

State Chem 137:127–137. doi:10.1006/jssc.1997.7737

Makovec D, Kodre A, Arcon I, Drofenik M (2011) The structure

of compositionally, constrained zinc-ferrite. J Nanopart

Res 13:1781–1790. doi:10.1007/s11051-010-9929-y

Maletin M, Moshopoulou EG, Kontos AG, Devlin E, Delimitis

A, Zaspalis V, Nalbandian L, Srdic VV (2007) Synthesis

and structural characterization of In-doped ZnFe2O4

nanoparticles. J Eur Ceram Soc 27:4391–4394. doi:

10.1016/j.jeurceramsoc.2007.02.165

McCurrie RA (1994) Ferromagnetic materials–structure and

properties. Academic Press Inc, New York

Nakagomi F, da Silva SW, Garg VK, Oliveira AC, Morais PC,

Franco A (2009) Influence of the Mg-content on the cation

distribution in cubic MgxFe(3-x)O(4) nanoparticles. J Solid

State Chem 182:2423–2429. doi:10.1016/j.jssc.2009.

06.036

O0Neill HS (1992) Temperature dependence of the cation dis-

tribution in zinc ferrite (ZnFe2O4) from powder XRD

structural refinements. Eur J Minerol 4:571–580

0.2

0.4

0.6

0.8

1.0

1.2

0.2 0.4 0.6 0.8 1.0 1.2

xA

Raman = 1.1 xA

Mossbauer

xA

Raman (Fe)= aI

Fe/(aI

Fe+bI

Mg)

R = b/a = 0.5

xA

Mössbauer(Fe)

xA R

aman

(Fe)

Fig. 6 Fe-content xARaman Feð Þ

� �calculated using Eq. 1 versus

Fe-content obtained from the Mossbauer xAMossbauer Feð Þ

� �data.

The solid line represents the linear fit xARaman ¼ 1:1xA

Mossbauer

J Nanopart Res (2012) 14:798 Page 9 of 10

123

Oliver SA, Harris VG, Hamdeh HH, Ho JC (2000) Large zinc

cation occupancy of octahedral sites in mechanically

activated zinc ferrite powders. Appl Phys Lett 76:2761–

2763. doi:10.1063/1.126467

Pusep YA, da Silva SW, Galzerani JC, Milekhin AG, Preo-

brazhenskii VV, Semyagin BR, Marahovka II (1995)

Spectroscopy of the optical vibrational-modes in GaAs/

AlxGa1-xAs heterostructures with monolayer-wide Alx-

Ga1-xAs barriers. Phys Rev B 52:610–2618. doi:10.1103/

PhysRevB.52.261

Rath C, Sahu KK, Anand S, Date SK, Mishra NC, Das RP

(1999) Preparation and characterization of nanosize Mn–

Zn ferrite. J Magn Magn Mater 202:77–84. doi:10.1016/

S0304-8853(99)00217-6

Seong MJ, Hanna MC, Mascarenhas A (2001) Composition

dependence of Raman intensity of the nitrogen localized

vibrational mode in GaAs1 - xNx. Appl Phys Lett

79:3974–3976. doi:10.1063/1.1424469

Sepelak V, Feldhoff A, Heitjans P, Krumeich F, Menzel D,

Litterst FJ, Bergmann I, Becker KD (2006) Nonequilibri-

um cation distribution, canted spin arrangement, and

enhanced magnetization in nanosized MgFe2O4 prepared

by a one-step mechanochemical route. Chem Mater

18:3057–3067. doi:10.1021/cm0514894

Shebanova ON, Lazor P (2003) Raman spectroscopic study of

magnetite (FeFe2O4): a new assignment for the vibrational

spectrum. J Solid State Chem 174:424–430. doi:10.1016/

S0022-4596(03)00294-9

Soler MAG, Lima ECD, da Silva SW, Melo TFO, Pimenta

ACM, Sinnecker JP, Azevedo RB, Garg VK, Oliveira AC,

Novak MA, Morais PC (2007) Aging investigation of

cobalt ferrite nanoparticles in low pH magnetic fluid.

Langmuir 23:9611–9617. doi:10.1021/la701358g

Srivastava CM, Shringi SN, Srivastava RG (1976) Mossbauer

study of relaxation phenomena in zinc–ferrous ferrites.

Phys Rev B 14:2041–2050. doi:10.1103/PhysRevB.14.

2041

Upadhyay C, Verma HC, Anand S (2004) Cation distribution in

nanosized Ni–Zn ferrites. J Appl Phys 95:5746–5751. doi:

10.1063/1.1699501

Verble JL (1974) Temperature-dependent light-scattering

studies of Verwey transition and electronic disorder in

magnetite. Phys Rev B 9:5236–5248. doi:10.1103/Phys

RevB.9.5236

Waldron RD (1955) Infrared spectra of ferrites. Phys Rev

99:1727–1735. doi:10.1103/PhysRev.99.1727

Wang ZW, Lazor P, Saxena SK, O’Neill HS (2002) High

pressure Raman spectroscopy of ferrite MgFe2O4. Mater

Res Bull 37:1589–1602

Wang ZW, Schiferl D, Zhao YS, O’Neill HSC (2003) High

pressure Raman spectroscopy of spinel-type ferrite

ZnFe2O4. J Phys Chem Solids 64:2517–2523. doi:10.1016/

j.jpcs.2003.08.005

White WB, DeAngeli BA (1967) Interpretation of vibrational

spectra of spinels. Spectrochim Acta 23A:985–995. doi:

10.1016/0584-8539(67)80023-0

Page 10 of 10 J Nanopart Res (2012) 14:798

123