You are on page 1of 8

See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/231126195

Magnetic nanoparticles with enhanced Fe2O3 to -Fe2O3 phase transition temperature


Article in Nanotechnology November 2006
Impact Factor: 3.82 DOI: 10.1088/0957-4484/17/23/023

CITATIONS

READS

55

151

5 authors, including:
Ayyappan Sathya

John Philip

Istituto Italiano di Tecnologia

Indira Gandhi Centre for Atomic Research

16 PUBLICATIONS 378 CITATIONS

182 PUBLICATIONS 3,414 CITATIONS

SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate,


letting you access and read them immediately.

SEE PROFILE

Available from: John Philip


Retrieved on: 12 May 2016

INSTITUTE OF PHYSICS PUBLISHING

NANOTECHNOLOGY

Nanotechnology 17 (2006) 58515857

doi:10.1088/0957-4484/17/23/023

Magnetic nanoparticles with enhanced


-Fe2O3 to -Fe2O3 phase transition
temperature
G Gnanaprakash, S Ayyappan, T Jayakumar, John Philip1
and Baldev Raj
Metallurgy and Materials Group, Indira Gandhi Centre for Atomic Research, Kalpakkam 603
102, Tamilnadu, India
E-mail: philip@igcar.gov.in

Received 3 September 2006, in final form 19 October 2006


Published 16 November 2006
Online at stacks.iop.org/Nano/17/5851
Abstract
We report a simple method for producing magnetic nanoparticles with
enhanced maghemite ( -Fe2 O3 ) to haematite ( -Fe2 O3 ) phase transition
temperature. By controlling the properties of the solvent media, we have
been able to tune the particle size from 2.3 to 6.5 nm. Nanoparticles with
higher transition temperatures have been achieved by using NaOH as an
alkali during the co-precipitation. The maghemite to haematite phase
transition was complete at a temperature below 600 C for nanoparticles
prepared using ammonia, whereas the phase transition was not complete until
750 C for the samples prepared with NaOH. The increase in average particle
size after heat treatment at 600 C is attributed to coalescence of particles by
solid state diffusion, where the system reduces its free energy by reducing the
surface area. The final particle diameters of the haematite after the heat
treatment were 35.4, 31.38 and 26.85 nm respectively for nanoparticles of
initial diameters 5, 7 and 9.8 nm. Our studies on the effect of the initial
particle size on the transition temperature show that the transition
temperature decreases with decreasing particle size due to the reduced
activation energy of the system.

1. Introduction
Tailoring of magnetic nanocrystals of well-defined size, shape
and self-assembly has been a topic of intensive research [15]
because of the enormous scope for applications of these
materials in storage media [6, 7], biomedical devices [8, 9],
optical devices, sensors [10] and magnetic fluids [1113]. In
addition, magnetic colloids have been found to be a model
system for fundamental studies of phenomena such as the
interaction between particles and molecules [14]. Inspired
by the potential applications of magnetic nanoparticles, in
recent years, several studies were focused on enlarging the
coercivity [15, 16] and transition temperatures [1720]. The
maghemite ( -Fe2 O3 ) to haematite ( -Fe2 O3 ) phase transition
studies in the nanosize regime attracted much attention due
1 Author to whom any correspondence should be addressed.

0957-4484/06/235851+07$30.00

to the technological applications of maghemite [1721]. The


maghemite to haematite phase transition has also been studied
using differential thermal analysis [21]. Factors influencing
phase transition in maghemite are particle size, supporting
media, heating rate, preparation method, metal ion doping
etc. The reported temperature of the transition of magnetically
active cubic structure (maghemite) to magnetically inactive
and most stable rhombohedral (haematite) ranges from 200
to 500 C for nanoparticles, whereas it lies between 500
and 600 C for bulk samples [22]. The lower transition
temperatures for the nanorange are attributed to the nonequilibrium state of nanocrystal which contains a large amount
of excess free energy. Recent studies revealed that doping
of maghemite with Mn(III) enhances the phase transition
temperature to 600 C due to increase in the activation
energy, to a more compact alpha phase that requires lattice
contraction [17]. However, in inert atmosphere the magnetic

2006 IOP Publishing Ltd Printed in the UK

5851

G Gnanaprakash et al

structure of magnetite is retained even at 650 C [23]. The


highest transition temperature so far reported for maghemite
nanoparticles supported on silica is 700 C because of its
strong oxideoxide interactions [24, 25]. The pressure induced
structural phase transition studies of -Fe2 O3 show a transition
pressure increasing with decreasing nanocrystal size [20],
similar to the earlier findings for CdSe nanoparticles [26, 27].
These studies indicate that the transition path decides the
final state of the system for nanocrystals, where the surface
energy contribution to the total free energy is very large. To
meet technological challenges for memory devices, further
improvement in transition temperature with minimal effort is
desirable to keep materials magnetically active at even higher
temperatures.
Among various reproducible synthesis routes for the
preparation of nanoparticles, the precipitation technique is one
of the most convenient and versatile methods [28, 29]. In
precipitation, the nucleated particle size mostly depends on
parameters related to solubility and interfacial energy [30, 31].
In the nanoregime, physical and chemical properties are size
dependent. In order to tune these properties, various groups
have been working on optimizing particle size by varying
parameters such as temperature [32], pH and ionic strength
imposed by non-complexing salts [29]. Recent studies on
CeO2 nanoparticles reveal that the solvent polarity also affects
the crystal growth [33]. Here, we report a simple method for
producing magnetic nanoparticles with enhanced maghemite to
haematite phase transition temperature and a method for tuning
the nanoparticles diameter from 2.3 to 6.5 nm by controlling
the properties of the solvent media. We also investigate
the size dependent transition temperatures and coalescence
phenomena.

2. Experimental section

coating process. Later, particles were acidified with dilute


HCl to separate the particles from the dispersion. The top
water layer with salts was discarded. The above obtained
surfactant coated particles were washed with triply distilled
water at 60 C, until the water pH reached a value of 7, to
remove ionic salt impurities trapped or adhered to the particle
coagulum. Later, the water-washed particles were dispersed
in hexane. The hexane dispersion was treated with acetone
to induce particle aggregation. The aggregated particles were
separated by centrifugation at 2500 rpm for 30 min. Addition
of acetone to hexane increases the polarity, which leads to
aggregation and sedimentation of particles dispersed in hexane.
However, excess acetone and repeated acetone washing also
cause desorption of oleic acid from magnetite interfaces due
to the competition between the affinity of the carboxylate
group of the oleic acid to the surface of the magnetite particle
and the organic polar solvent (acetone/hexane mixture). The
washing procedure was repeated twice to eliminate excess free
surfactant molecules. The surfactant coated magnetite particles
were dried at 35 C for 48 h in inert atmosphere.
2.2.1. Preparation of particles (process 1). Particles of
different sizes were obtained by controlling the rate of
ammonium hydroxide addition to the iron salt solution during
precipitation. The mixture of 15 ml of ferric and 15 ml of
ferrous salt solutions was diluted with 170 ml of water and
25% ammonia solution was added at different speeds of 80, 20
and 1 ml s1 for different batches. After addition of alkali, the
pH reached a value of 10. At this stage, the solution turned
black, indicating the formation of magnetite particles. The
average sizes of the particles prepared by this process are 5,
7 and 9.8 nm (these samples are hereafter referred to as A1, A2
and A3, respectively).

2.1. Chemicals
For the synthesis of magnetite nanoparticles, FeSO4 7H2 O,
FeCl3 6H2 O, 25% aqueous NH3 , NaOH, ethanol, oleic acid,
35% HCl, H2 SO4 , hexane and acetone procured from E-Merck
were used. All the chemicals were GR grade (except the
hexane) and used without any further purification. The water
used in all the experiments was triply distilled and filtered with
a 0.22 m size filter. The conductivity of the water used is
below 0.2 S cm1 .
2.2. Synthesis
Magnetite nanoparticles were prepared by precipitating iron
salts with ammonium hydroxide (process 1) and sodium
hydroxide (process 2). The iron salts used were freshly
prepared with 0.8 M FeCl3 6H2 O and 0.4 M FeSO4 7H2 O in an
acidic medium of HCl and H2 SO4 , respectively. The iron salt
solutions were mixed at 1:1 ratio with constant stirring. On
vigorous stirring, the solution pH was increased rapidly, by the
addition of one of the alkalis, which leads to a black coloured
solution. After 30 min of digestion time, 5 ml oleic acid was
added to the above solution to coat the magnetite nanoparticles.
At this stage, the pH of the solution was adjusted to 9 at a
constant temperature of 70 C. The same pH, temperature
and stirring speed were maintained for 30 min to finish the
5852

2.2.2. Preparation of particles (process 2). Here, particles


of different sizes were obtained by controlling the polarity of
the solvent used in the precipitation. The mixture of 15 ml
of ferric and 15 ml of ferrous salt solutions was diluted with
different compositions of mixed solvent of ethanol and water.
To this 30 ml salt solution, 50 ml of triply distilled water and
120 ml of ethanol were added to get a total solution volume
of 200 ml, i.e. 60:40 ethanol:water ratio. 30 ml of 6 M NaOH
was added at an addition rate of 1 ml s1 . On the completion
of the NaOH addition, the pH of the solution reached a value
of 12.5 and it turned blackish. Particles with average sizes of
2.3, 3.3, 4.8 and 6.5 nm were obtained by varying the solvent
composition (these samples are hereafter referred to as S1, S2,
S3 and S4, respectively).
2.3. Characterization
2.3.1. X-ray powder diffraction (XRD). A MAC Science
MXP18 X-ray diffractometer was used for crystal structure and
average particle size analysis. 2 values were taken from 20
The XRD
to 110 using Cr K radiation ( value of 2.2897 A).
patterns of nanoparticles were verified by comparing with the
JCPDS data.

Magnetic nanoparticles with enhanced -Fe2 O3 to -Fe2 O3 phase transition temperature

2.3.3. Vibrating sample magnetometer (VSM). Magnetic


data for the nanoparticle samples were obtained with a
vibrating sample magnetometer, EG&G Princeton Model
4500.
Hysteresis measurements were taken at room
temperature under an applied field up to 7 kOe.

Ethanol/water ratio
0 : 100

Intensity (a.u)

2.3.2. Transmission electron microscopy (TEM). The TEM


instrument used was a JEOL 2011 with an acceleration voltage
of 200 kV. One drop of magnetite dispersion was placed on
a carbon coated copper grid (0.3 cm diameter, mesh size of
200 holes cm1 ) and left to dry overnight at room temperature
(25 C).

20 : 80

40: 60

60 : 40

40

50

60

70

80

90

100

110

2 (degree)

2.3.4. Atomic absorption spectrometer (AAS). The amount of


Na ions present in the samples was measured using a doublebeam atomic absorption spectrometer (AAS), GBC Model
Avanta.
2.3.5. Thermal gravimetric analysis. A Mettler Toledo
TG/SDTA-851e instrument was used for TGA and DTA
measurements. Weight loss measurements were taken from
50800 C in an inert atmosphere of argon. The constant
temperature increment of 5 C min1 was maintained.

3. Results and discussion


3.1. Crystal structure
From the XRD pattern, the average particle size is obtained
from the most intense peak of (311) in magnetite, by using
the DebyeScherrer formula. The size of the magnetite
particle measured from the peak broadening was in good
agreement with the values obtained by TEM and magnetization
data. Comparison of the XRD patterns with the JCPDS
data confirms the samples are magnetite with inverse spinel
structure. The inverse spinel structure consists of oxide ions in
the cubic close-packed arrangement in which 1/3 of tetrahedral
interstices and 2/3 of octahedral interstices coordinate with
oxygen. All Fe(II) ions occupy the octahedral interstices, and
half of the Fe(III) occupy the tetrahedral interstices with the
remaining half of the Fe(III) in octahedral interstices. Electron
spins of Fe(III) ions in octahedral interstices are aligned antiparallel to those in tetrahedral interstices. The Fe(II) ions align
their spins parallel to Fe(III) ions in adjacent octahedral sites
leading to a net magnetization. The measured values of the
lattice constants for the samples A1 and S3 are about 0.838 nm,
indicating that the samples are magnetite. This has also been
confirmed with JCPDS data. The reported values of the lattice
constants of magnetite and maghemite are 0.839 and 0.835 nm
respectively [34]. The lattice parameter contrast observed
in magnetite nanoparticles is attributed to very high surface
energy of the particles and the existence of the maghemite
impurity, especially at the surface of the particles [23]. The
surface ions Fe(II) of magnetite oxidize to more stable Fe(III)
ions and it converts to maghemite.
3.2. Control of the average particle size
The nanoparticles with different sizes are prepared by controlling the solubility and reaction rates. The equilibrium critical

Figure 1. The XRD patterns of samples obtained with different


ethanolwater ratios.

radius (r ) of particles during nucleation in supersaturated solution is defined as [30, 31]

r =

2V
kB T ln(S)

(1)

where V is the molecular volume of the precipitated species,


is the surface free energy per unit surface area, kB is the
Boltzmann constant, T is the absolute temperature and S is
the saturation ratio, which is defined as the ratio between the
solute concentrations at saturation ( X ) and equilibrium (X S ),
i.e. S = X/ X S . For a given value of S , all the particles whose
radius is below r will dissolve in the solution whereas others
will grow. The concentration X S of ions forming a saturated
solution in equilibrium with the solid is given by [35]




i
z 1 z 2 e2
X S = exp
(2)
exp
kB T
40 kB T (a+ + a )
where i is the free energy change, a+ and a are radii
of ions of charge z 1 and z 2 , respectively, e is the elementary
charge, is the solvent dielectric constant and 0 is the
permittivity in vacuum. Therefore, the saturation ratio is
inversely proportional to the solvent dielectric constant and the
equilibrium critical radius is directly proportional to the solvent
dielectric constant. By controlling the dielectric constant
of solvent with a suitable ratio of ethanol and water, it is
possible to vary the critical nuclei size and hence the particle
size. Table 1 shows the compositions of the solvent, the
corresponding particle size and the saturation magnetization.
By controlling the ratio of ethanol and water, we have been able
to vary the particle size from 2.3 to 6.5 nm, where the alkali
used for precipitation was NaOH. For those samples prepared
with ammonium hydroxide (A1, A2, and A3), we varied the
particle size by changing the rate of addition of the alkali.
However, we have also observed a similar trend in particle size
with ethanol and water ratio in those samples prepared with
ammonium hydroxide. The XRD patterns of samples obtained
with different ethanolwater ratios are shown in figure 1. The
broadening of the peaks with increasing ethanol concentration
clearly indicates a decrease in crystalline size. The TEM image
of sample S4 and the size distribution are shown in figure 2.
At the rapid addition rate, the number of nuclei formed
at the nucleation stage is large, which would result in a larger
5853

G Gnanaprakash et al

50

Distribution (%)

40

30

20

20 nm
10

0
2

10

12

14

Particle diameter (nm)

Figure 2. The TEM image of sample S4 (left) and the size distribution (right).

Table 1. The compositions of the solvent, the precipitated particle


size and saturation magnetization.
Ethanol:water
ratio

S1
S2
S3
S4

60:40
40:60
20:80
0:100

2.3
3.3
4.8
6.5

Saturation
magnetization
(emu g1 )
6.23
14.57
18.36
26.27

Weight Percentage (%)

Sample

Average
particle size
(nm)

100

2.3 nm
80

60

6.5 nm

40
0

number of nanoparticles with smaller size [36]. By varying the


ammonium hydroxide addition rate to the iron salt solution, we
have obtained particles with different sizes. For addition rates
of 80, 20 and 1 ml s1 , the average particle sizes were 5, 7 and
9.8 nm, respectively.
Figure 3 shows the thermogravimetric curves of magnetite
samples S1 (2.3 nm) and S4 (6.5 nm). The oleic acid coated
magnetite nanoparticles show a two-step weight loss, where
the first weight loss belongs to the weakly bound secondary
layer as well as some more weakly bound molecules within the
primary surfactant layer and the second weight loss belongs
to the strongly bound primary surfactant layer [37]. The
experimental values of the weight losses in sample S1 (2.3 nm)
and sample S4 (6.5 nm) were 27.6 and 13.6% respectively.
The first loss is observed at 250 C with a small drop of
weight and a large drop is observed at 375 C. Assuming
that the surfactant forms a close-packed monolayer on the
nanoparticles, the total weight loss due to the loss of surfactant
is calculated theoretically by using the equation [38]
 d 2  
p
M
100
a
N0
 d2  

1
dp3 + a p NM0
6
where dp is the diameter of the particle, the density of
the magnetite particle, a the head area per molecule of the
surfactant, M the molecular weight of the surfactant and N0
the Avogadro number. The calculated values of the weight
loss are found to be 30 and 13.17% respectively for 2.3 and
6.5 nm sized particles. The experimentally observed amount
of surfactant in sample S1 was about 2.03 times more than that
5854

100

200

300

400

500

600

700

800

Temperature ( 0 C)

Figure 3. The thermogravimetric curves of magnetite samples S1


(2.3 nm) and S4 (6.5 nm).

in S4, which was in reasonable agreement with the theoretical


calculations (i.e. 2.27). The difference between theoretical and
experimental values should be attributed to polydispersity of
the particles, the hindrance effect of surfactant molecules at the
interface of small size particles, the shape of the particles (in
the calculations particles are assumed to be spherical though
they are not perfect spheres) and the excess free surfactant
molecules available in S4.
Figure 4 shows the magnetization curves of samples S3
(4.8 nm) and A1 (5 nm) at room temperature. The saturation
magnetization of S3 was slightly lower than that of A1,
though both the samples show superparamagnetic behaviour
with zero remanence and coercivity. The observed lower
saturation magnetization value for sample S3 is attributed to
the existence of a small amount of non-magnetic compounds in
the samples prepared with NaOH alkali. Similar observations
were reported in the literature [39, 40].
3.3. Study of the maghemite to haematite transition
The XRD patterns of samples A1 and S3 obtained after
heat treatment at different temperatures of 250 C (4 h),
350 C (4 h), 450 C (1 h) and 600 C (1 h) are shown in
figures 5(a) and (b). The XRD pattern of the untreated samples
(25 C) confirms the cubic inverse spinel structure from the

Magnetic nanoparticles with enhanced -Fe2 O3 to -Fe2 O3 phase transition temperature


32

25

25 0C

A1

25 C

4.8 nm
15

16
S3

Magnetization (emu/g)

Magnetization (emu/g)

24

8
0
-8
-16
-24

600 0C
5

700 0C

-5

0.7

-0.3

-15

-32
-8

-6

-4

-2

-0.8

Magnetic Field (kOe)

-8

-25
-8

Figure 4. The magnetization curves of samples S3 and A1 at room


temperature.
(a)

700 0C

0.2

-6

-4

-2

-6

-4

-2

Magnetic field (kOe)

Figure 6. Magnetization curves obtained for the sample S3 before


heat treatment and after heat treatment at 600 and 700 C. The inset
shows a magnified view of the magnetization curve of sample S3
heat treated at 700 C.

Intensity (a.u)

600 C

450 C

350 C
0

250 C

25 C
20

30

40

50

60

70

80

90

100

110

300

214

116

024

113

110

750 C

Intensity (a.u)

012

104

(b)

018

2 (deg)

700 C

250 C

440

511

422

400

311

220

211

111

210

600 C

25 C

20

30

40

50

60

70

80

90

100

110

2 (deg)

Figure 5. (a) The XRD patterns of sample A1 obtained after heat


treatment of 25 C, 250 C (4 h), 350 C (4 h), 450 C (1 h) and
600 C (1 h). (b) XRD pattern of sample S3 heat treated at different
temperatures of 25, 250, 600, 700 and 750 C.

characteristic peaks, i.e. (220), (311), (400), (422), (511)


and (440). Since Fe2+ ions are more sensitive to oxidation
(Fe2+ + O2 Fe3+ ), even at room temperature surface ions
readily oxidize as Fe3+ ions in air which will lead to the
formation of maghemite at the surface [23]. Like in earlier
reports [41], both the samples S3 and A1 heated at 250 C for
4 h in air oxidize to maghemite structure, which is evidenced
from the emerging additional very low intensity diffraction
peaks (210) and (211) and the lattice constant decrease to
0.834 nm from 0.838 nm. The other peaks of magnetite such
as (220), (311), (400), (422), (511) and (440) are also slightly
shifted towards the higher angle. In maghemite, unequal
distribution of Fe(III) ions in tetrahedral and octahedral sites
leads to a net magnetization [42].
For the sample A1, the maghemite to haematite conversion
started at 350 C (for 4 h) and the transition was almost

complete at 450 C (for 1 h). At 600 C, the conversion was


complete. The rhombohedral crystal structure of haematite
is evident from the major characteristic peaks of (104) and
(110) with an intensity ratio of 100:70 including the other
characteristic peaks of (012), (113), (024), (116), (018), (214)
and (300) with respective intensity ratios.
Figure 5(b) shows the XRD pattern of sample S3 treated
at different temperatures of 25, 250, 600, 700 and 750 C. In
contrast to the case for sample A1, maghemite to haematite
conversion is not complete even at 700 C in sample S3,
though the magnetite to maghemite oxidation occurred at
250 C. Surprisingly, the completion of the phase transition
in sample S3 occurs only at 750 C. Atomic absorption
spectroscopic studies carried out on sample S3 show that the
amount of sodium is about 0.9 wt%, whereas no trace of
Na was observed in sample A1. The sodium ions seem to
have been trapped inside the crystal while crystallizing the
magnetite nanoparticles with rapid oxidation because atoms
or ions need lower activation energy for diffusion in the
nanocrystalline state [43]. We ruled out the possibilities of
having excess sodium ions outside the crystals because we
see the same amount of Na in the samples, even after ten
water washings. Therefore, diffused sodium ions in sample
S3 seem to have effectively suppressed the transition from
metastable maghemite to the thermodynamically stable phase,
haematite. It appears that the presence of diffused sodium
ions stabilizes the metastable phase by increasing the activation
energy to overcome the free energy barrier between gamma to
alpha phases. However, in sample A1, thermally unstable ions
of ammonium decompose and escape from the system very
easily. The magnetization studies of sample S3 show magnetic
behaviour up to 700 C. Figure 6 confirms the existence of the
magnetic phase of maghemite at 600 and 700 C, indicating the
slow conversion to the magnetically inactive phase, haematite.
A recent study on maghemite nanoparticles doped with 8.5%
Mn(III) showed that the gamma to alpha phase transition is
suppressed until 600 C [17]. The increasing coercivity value
(30300 Oe) at higher temperatures is attributed as due to an
increase in particle size due to coalescence. The narrowing
of the characteristic peaks of the XRD patterns at higher
temperatures also confirms the above fact.
5855

G Gnanaprakash et al

Intensity (a.u)

6.5 nm

4.8 nm

2.3 nm
20

30

40

50

60

70

80

90

100

110

2 (deg)

Figure 7. XRD patterns of S1 (2.3 nm), S3 (4.8 nm) and S4 (6.5 nm)
samples treated at 600 C.

8
6.5 nm
0

Magnetization (emu/g)

600 C
4.8 nm

2.3 nm

-4

-8
-8000

-6000

-4000

-2000

2000

4000

6000

8000

Field (Oe)

Figure 8. Magnetization curves of 2.3, 4.8 and 6.5nm size magnetite


nanoparticles treated at 600 C.

In order to better understand the enhancement of the phase


transition temperature and the universal nature in different
particle size ranges, the experiments were performed using
magnetite nanoparticles of average particle sizes 2.3, 4.8 and
6.5 nm, prepared with sodium hydroxide. The AAS result
shows that the amount of Na ions in these samples varies from
0.305 to 1.22 wt%.
Figure 7 shows the XRD patterns of S1, S3 and S4
after heat treatment at 600 C. The XRD pattern of the
specimen with 2.3 nm size particles (S1) after heat treatment
at 600 C for 1 h showed the characteristic peaks of haematite.
The most intense peaks of (104) and (110) are in the
intensity ratio of 100:84 instead of 100:70 indicating that
the conversion is almost complete. In the case of 4.8 nm
particles, the intensity ratio of haematite characteristic peaks
(104) and (110) is 47:100, indicating the presence of a
considerable amount of unconverted maghemite. For 6.5 nm
size particles, the intensity ratio of the characteristic peaks
of haematite (104) and (110) is 14:100, indicating very little
conversion of maghemite. Figure 8 shows that the saturation
magnetization values for the samples of average size 2.3,
4.8 and 6.5 nm treated at 600 C for 1 h are 0.45, 5.1 and
5.8 emu g1 , respectively. These results suggest that the
phase transition temperature decreases with decreasing particle
size. In other words, increase in surface to volume ratio at
reduced dimensions lowers the activation energy of the system,
facilitating phase transition at lower temperatures.
The size dependence of the phase transition temperature
was also investigated with magnetite nanoparticles of
5856

diameters 5, 7 and 9.8 nm, synthesized by using ammonia.


For 5 nm size particles (A1) the transition started at 350 C
(figure 5(a)) and was almost complete at 450 C. The most
intense peaks of (104) and (110) are in the intensity ratio of
100:89. At 600 C, complete conversion occurred. In the case
of 7 nm size particles, though the transition started at 350 C,
the conversion ratio was less than for 5 nm size particles
at 450 C and the conversion was complete at 600 C. For
9.8 nm size particles, the phase transition started at 450 C
and ended only at 600 C. Throughout the size range studied,
the phase transition was complete at 600 C. This undoubtedly
confirms that the enhancement in phase transition temperature
for samples S1, S3 and S4 is solely due to trapped Na ions.
After heat treatment at 600 C, a drastic increase in the
average particle size is also observed. The final particle
diameters of haematite particles were 35.4, 31.38 and 26.85 nm
respectively for magnetite nanoparticles of initial diameters 5,
7 and 9.8 nm. The increases in diameter in these samples
were threefold to sevenfold illustrating the significance of
the surface energy factor for coalescence phenomena. In
the case of nanoparticles prepared with NaOH, the exact
average particle size estimation is difficult due to the phase
transition being incomplete at 600 C. However, the increased
coercivity values observed in the magnetization curves after
the heat treatment at 600 C undoubtedly confirm an increase
in the average particle size. The minimum particle volume
(Vp ) required for exhibiting superparamagnetic behaviour
is 25kB T /K , where kB is the Boltzmann constant, K is
the anisotropic constant and T is temperature [42]. For
maghemite, the anisotropy constants for bulk and for 4 nm size
particles are 4.64 104 erg cm3 and 1.3 106 erg cm3
respectively [44]. The two orders of magnitude increase
of the K value for 9 nm size particles is attributed to the
surface anisotropic effect [44]. The minimum size required
for superparamagnetic nature for maghemite (for a K value
of 1.3 106 erg cm3 ) particles is 12 nm. The deviation
from superparamagnetic behaviour observed for the samples
S1, S3 and S4 after 600 C heat treatment could be due to
an increase in the particle size beyond the superparamagnetic
limit (12 nm) due to coalescence of particles. In an earlier
study, a threefold increase in particle size at various annealing
temperatures (well below its melting temperature) is observed
for haematite particles of 60 nm size [45]. The mechanism
responsible for the increase in diameter is coalescence of
particles by solid state diffusion, where the system reduces its
free energy by reducing the surface area in nanoparticles. The
coalescence time (the typical coalescence time for a 3 nm sized
cluster is about 40 ns) and the melting temperature [46, 47]
reduce drastically with decreasing particle size, which is in
agreement with the experimental observations.

4. Conclusions
In summary, we have been able to tailor magnetic nanoparticles
in the size range of 27 nm by changing the dielectric constant
of the solvent and the alkali addition rates. An enhancement
in the temperature of the phase transition from cubic -Fe2 O3
to rhombohedral -Fe2 O3 is observed for magnetite samples
prepared with NaOH. The enhancement in the phase transition
temperature by about 150 C for 5 nm size particles seems to be

Magnetic nanoparticles with enhanced -Fe2 O3 to -Fe2 O3 phase transition temperature

due to the trapped Na ions and an increase in activation energy.


It has been found that the transition temperature decreases
with decreasing particle size due to the reduced activation
energy of the system. After heat treatment, the average particle
size increases due to coalescence of particles by solid state
diffusion. As the percentage of substitution of sodium is
very low, to retain the ferrimagnetic nature of the material
up to 700 C, these findings may have significant practical
implications.

Acknowledgments
We thank Dr K S Viswanathan for the atomic absorption
spectrometry, Mr S Mahadevan for the XRD investigations.
We also thank Dr Vasudeva Rao and Mr P Kalyanasundaram
for support and encouragement.

References
[1] Roca A G, Morales M P, OGrady K and Serna C J 2006
Nanotechnology 17 2783
[2] Teng X and Yang H 2005 Nanotechnology 16 S554
[3] Wang J, Chen Q, Zeng C and Hou B 2004 Adv. Mater. 16 137
[4] Puntes V F, Krishnan K M and Alivisatos A P 2001 Science
291 2115
[5] Petit C, Taleb A and Pileni M P 1999 J. Phys. Chem. B
103 1805
[6] Todorovic M, Schultz S, Wong J and Scherer A 1999 Appl.
Phys. Lett. 74 2516
[7] Speliotis D E 1999 J. Magn. Magn. Mater. 193 29
[8] Perez J M, Simeone F J, Saeki Y, Josephson L and
Weissleder R 2003 J. Am. Chem. Soc. 125 10192
[9] Kalambur V S, Han B, Hammer B E, Shield T W and
Bischof J C 2005 Nanotechnology 16 1221
[10] Shemer G and Markovich G 2002 J. Phys. Chem. B 106 9195
[11] Raj K and Moskowitz R 1990 J. Magn. Magn. Mater. 85 233
[12] Charles S W and Popplewell J 1980 Ferromagnetic Materials
vol 2 (Amsterdam: North-Holland)
[13] Rosensweig R E 1985 Ferrohydrodynamics (Cambridge, MA:
MIT Press)
[14] Philip J, Prakash G G, Jayakumar T, Sundaram P K and
Raj B 2002 Langmuir 18 4625
Philip J, Prakash G G, Jayakumar T, Sundaram P K and
Raj B 2003 Macromolecules 36 9230
Philip J, Prakash G G, Jayakumar T, Sundaram P K and
Raj B 2002 Phys. Rev. Lett. 89 268301
[15] Jin J, Ohkoshi S-I and Hashimoto K 2004 Adv. Mater. 16 48
[16] Yan Q, Kim T, Purkayastha A, Ganesan P G, Shima M and
Ramanath G 2005 Adv. Mater. 17 2233
[17] Lai J, Shafi K V P M, Loos K, Ulman A, Lee Y, Vogt T and
Estournes C 2003 J. Am. Chem. Soc. 125 11470
[18] Sanchez R M T 1996 J. Mater. Sci. Lett. 15 461

[19] Ayyub P, Multani M, Barma M, Palkar V R and


Vijayaraghavan R 1988 J. Phys. C: Solid State Phys.
21 2229
[20] Clark S M, Prilliman S G, Erdonmez C K and
Alivisatos A P 2005 Nanotechnology 16 2813
[21] Ye X, Lin D, Jiao Z and Zhang L 1998 J. Phys. D: Appl. Phys.
31 2739
[22] Ennas G, Marongiu G, Musinu A, Falqui A, Ballirano P and
Caminiti R 1999 J. Mater. Res. 14 1570
[23] Sun S, Zeng H, Robinson D B, Raoux S, Rice P M,
Wang S X and Li G 2004 J. Am. Chem. Soc. 126 273
[24] Ennas G, Musinu A, Piccaluga G, Zedda D, Gatteschi D,
Sangregorio C, Stanger J L, Concas G and Spano G 1998
Chem. Mater. 10 495
[25] Lund C R F and Dumesic J A 1981 J. Phys. Chem. 85 3175
[26] Tolbert S H and Alivisatos A P 1994 Science 265 373
[27] Jacobs K, Zaziski D, Scher E C, Herhold A B and
Alivisatos A P 2001 Science 293 1803
[28] Khalafalla S E and Reimers G W 1980 IEEE Trans. Magn.
16 178
[29] Vayssieres L, Chaneac C, Tronc E and Jolivet J P 1998
J. Colloid Interface Sci. 205 205
[30] Mullin J W 1972 Crystallisation (London: Butterworth)
[31] Zhou X-D, Huebner W and Anderson H U 2002 Appl. Phys.
Lett. 80 3814
[32] Jorgensen J M, Erlacher K, Pedersen J S and Gothelf K V 2005
Langmuir 21 10320
[33] Chen H-I and Chang H-Y 2004 Colloids Surf. A 242 61
[34] He Y P, Miao Y M, Li C R, Wang S Q, Cao L, Xie S S,
Yang G Z, Zou B S and Burda C 2005 Phys. Rev. B
71 125411
[35] Israelachvili J N 1991 Intermolecular and Surface Forces 2nd
edn (London: Academic)
[36] Cao G 2004 Nanostructures and NanomaterialsSynthesis,
Properties and Applications (London: Imperial College
Press)
[37] Sahoo Y, Pizem H, Fried T, Golodnitsky D, Burstein L,
Sukenik C N and Markovich G 2001 Langmuir 17 7907
[38] Goodarzi A, Sahoo Y, Swihart M T and Prasad P N 2004
Mater. Res. Soc. Symp. Proc. 789 N.6.6.1
[39] Blums E, Cebers A and Maiorov M M 1996 Magnetic Fluids
(Berlin: Walter de Gruyter)
[40] Gribanov N M, Bibik E E, Buzunov O V and
Naumov V N 1990 J. Magn. Magn. Mater. 85 7
[41] Tang J, Myers M, Bosnick K A and Brus L E 2003 J. Phys.
Chem. B 107 7501
[42] Cullity B D 1972 Introduction to Magnetic Materials (Reading,
MA: Addison-Wesley)
[43] Schumacher S, Birringer R, Strau R and Gleiter H 1989 Acta
Metall. 37 2485
[44] Tartaj P, Gonzalez-Carreno T and Serna C J 2003 J. Phys.
Chem. B 107 20
[45] Suber L, Fiorani D, Imperatori P, Foglia S, Montone A and
Zysler R 1999 Nanostruct. Mater. 11 797
[46] Goldstein A N, Echer C M and Alivisatos A P 1992 Science
256 1425
[47] Lewis L J, Jensen P and Barrat J-L 1997 Phys. Rev. B 56 2248

5857

You might also like