Study on the synthesis of ε-Fe3N-based magnetic fluid

Article abstract of DOI:10.1016/j.jmmm.2006.03.067

In this work, stable high-saturation magnetization ε-Fe₃N magnetic fluid was synthesized successfully by the chemical reaction of iron carbonyl (Fe(CO)₅) and ammonia gas (NH₃). The experiment results have shown that the re

Full text of DOI:10.1016/j.jmmm.2006.03.067

ARTICLE IN PRESS
Journal of Magnetism and Magnetic Materials 307 (2006) 198–204
www.elsevier.com/locate/jmmm
Study on the synthesis of e-Fe3N-based magnetic fluid
Ã
Wei Huang, Jianmin Wu , Wei Guo, Rong Li, Liya Cui
Department of Functional Material Research, Central Iron & Steel Research Institute, Beijing 100081, PR China
Received 24 May 2005; received in revised form 23 November 2005
Available online 2 May 2006
Abstract
In this work, stable high-saturation magnetization e-Fe3N magnetic fluid was synthesized successfully by the chemical reaction of iron
carbonyl (Fe(CO)5) and ammonia gas (NH3). The experiment results have shown that the reactive conditions, such as the nitriding
temperature, the gas flux ratio of Ar1:Ar2:NH3, the reactive time, the content of surfactant and the hole size of the porous plate used,
have important effects on the phase composition, the size of magnetic particles, the magnetic properties and the stability of e-Fe3N
magnetic fluid. Also it was found that the synthetic time of stable high saturation magnetization e-Fe3N magnetic fluid could be
shortened by adding n-heptane into the carrier, and the size of e-Fe3N magnetic particles could be decreased by decreasing the pore size
of the porous plate used in our experiment. Finally, stable e-Fe3N magnetic fluid with the saturation magnetization 1663 Gs and the
mean particle size 12 nm was synthesized successfully.
r 2006 Elsevier B.V. All rights reserved.
Keywords: Magnetic fluids; Iron nitride; Saturation magnetization; Nano-material; Functional material
1. Introduction
nitriding temperature, the gas flux ratio of Ar1:Ar2:NH3,
the reactive time, the content of surfactant and the hole size
of the porous plate used on the phase composition and the
size of magnetic particles, the magnetic properties and the
stability of magnetic fluid were studied.
Magnetic fluids are stable colloidal suspension composed
of single-domain magnetic nanoparticles dispersed in
appropriate solvents [1] and this liquid material has an
interesting superparamagnetic behaviour. The research on
magnetic fluids began from 1950s. Several years later, in
1965, the magnetic fluids were used successfully for seal by
the NASA, USA. After that, they became new important
functional materials and were widely used for many areas,
such as mechanism, electron, chemist, medical treatment,
environment protection and so on.
Early in 1990s, an iron-nitride magnetic fluid was
successfully synthesized by Nakatani et al. [2] using
vapour–liquid chemical reaction of iron carbonyl and
ammonia gas. The prominent advantage of iron-nitride
magnetic fluids is its higher saturation magnetization than
that of the conventional Fe3O4 magnetic fluids, so it was
studied extensively [3–5]. In this work, the effects of the
2. Experimental procedures
e-Fe3N based magnetic fluid was synthesized by a home-
made apparatus (see Fig. 1). In the experiment, at first the
liquid iron carbonyl (Fe(CO)5) was heated to 37 1C via a
water bath, the liquid iron carbonyl evaporated into the
iron carbonyl vapour at this temperature. Then the iron
carbonyl vapour was loaded into the bottom of the reactor
(7) by argon gas (come from pipe Ar1 in Fig. 1). In the
bottom of the reactor (7) the iron carbonyl vapour mixed
with ammonia gas (come from pipe NH3) and argon gas
(come from pipe Ar2). The argon gas Ar2 was used to
control the concentration of iron carbonyl vapour and
ammonia gas in the mixed gas. And then the mixed gas
went through the porous plate (5) in the middle of the
reactor (7) into the carrier liquid (6) composed of a-olefinic
hydrocarbon synthetic oil (PAO oil with low volatility and
Ã
Corresponding author. Tel.:/fax: +86 10 62183115.
E-mail address: jm1962@vip.sina.com (J. Wu).
0304-8853/$ - see front matter r 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmmm.2006.03.067

ARTICLE IN PRESS
W. Huang et al. / Journal of Magnetism and Magnetic Materials 307 (2006) 198–204
199
7
6
5
4
2
3
1
8
1
1
Ar2
Ar1
NH3
1. Flowmeter, 2. Water Bath, 3. Fe(CO)5, 4. Furnace, 5. Porous Plate, 6. Carrier
Liquid, 7. Reactor, 8. Exhaust Gas Bottle
Fig. 1. Preparation apparatus of the iron nitride magnetic fluid.
low viscosity) and succinicimide (surfactant). The porous
plate (5) is a kind of 3 mm thickness stainless steel plate
prepared by model pressing and sintering the stainless steel
powders. The gas molecule could pass through this
stainless steel plate from one side to the other side easily
due to the existence of many fine pores inside of it. But the
carrier liquid above the porous plate found it difficult to
pass through it into the bottom of the reactor (7) due to the
existence of the higher pressure of the mixed gas in this
area. The pore size could be controlled by the prepare
process. The temperature of the carrier liquid was
controlled strictly at 182 1C. At this temperature, the
Fe(CO)5 would decompose into active ultra-fine Fe
crystals. These ultra-fine Fe particles could decrease the
decomposing activation energy of NH3 from 377 to 167 kJ/
mol [6] and promoted the NH3 decomposing into N and H.
The active N atoms decomposed from NH3 diffused into
Fe particles and formed iron-nitride particles. Then these
iron-nitride particles were coated automatically by the
surfactant and dispersed homogeneously in the carrier
liquid by the Brownian motion. The iron-nitride-based
magnetic fluid was obtained during this process. The
volatilization velocity of the mixed oil would increase if
adding some low-boiling point n-heptane into the carrier
liquid in the experiment. More carrier liquid would
volatilize during the same reactive time. Thus the
concentration of magnetic particles increased and the
magnetic fluid with higher saturation magnetization could
be obtained in the same reactive time.
properties of magnetic fluids at room temperature. And the
density was measured by the picnometer.
3. Results and discussion
3.1. Thermal decomposition temperature
The e-Fe3N magnetic fluids is synthesized by the reaction
of Fe(CO)5 and NH3 as mentioned above. Fe(CO)5 is a
kind of light yellow liquid. Its boiling point is 103 1C. It
decomposes into Fe and CO when heated to more than
60 1C under environmental pressure, and the decomposing
speed increases with increasing temperature. In our
experiment it was found that the synthetic temperature
has much more effect on obtaining e-Fe3N magnetic fluid.
The best synthetic temperature of e-Fe3N magnetic fluid is
182 1C. When the temperature is higher than this tempera-
ture, the size of Fe particles will grow quickly and the
nano-size iron-nitride particles cannot be obtained. In
addition, the stability of the mixed oil decreases. When the
temperature is lower than this temperature, it is difficult to
obtain the single-phase e-Fe3N magnetic fluid. It can be
noticed that this synthetic temperature is much lower than
the nitriding temperature of Fe (500–570 1C). It is because
the decomposing activation energy of NH3 is decreased by
the active nascent Fe particles as mentioned above.
Furthermore, the crystalline field and the binding energy
of the outer layer of nano-particles are different from that
of the inner layer [7]. The existence of a lot of dangling
bonds around the outer layer of nano-particles makes
themselves easy to combine with other atoms and form
stable compounds at lower temperature.
After synthesised, X’Pert PRO (Panalytical) X-ray
diffractometer was used to analyse the phase composition
of magnetic particles. The diffraction was performed with
˚
Coka (l ¼ 1:7889 A) and the ray was filtered by the
1
graphite. The experiment parameters used were: 40 mA,
35 kV, continuous scan, scan speed 21/min. Transmission
electron micrographs were obtained using a 2000fx
transmission electron microscope (TEM) operated at
160 keV. Samples were prepared by air-drying drops of
diluted solutions of the preparations on carbon films
supported by copper grids. LDJ9600 vibrating sample
magnetometer was carried out to measure the magnetic
Beside the synthetic temperature, the flux ratio of
Ar1:Ar2:NH3 (which decides the contents of NH3 and
Fe(CO)5 in the mix gas) is also very important to the
formation of single phase e-Fe3N magnetic fluid. It will be
discussed in the next section. Fig. 2c gives the X-ray
diffraction pattern of the magnetic particles synthesized at
182 1C under a proper flux ratio of Ar1:Ar2:NH3 ¼ 7:2:2.
It shows that single phase e-Fe3N magnetic particles can be

ARTICLE IN PRESS
W. Huang et al. / Journal of Magnetism and Magnetic Materials 307 (2006) 198–204
200
where KP is the equilibrium constant of reaction (2).
PH and PNH are the partial pressures of H2 and NH3 in
carrier, respectively. aN is the degree of activity of atom N
in Fe particle.
Fe
2400
1200
Fe N
4
2
3
Fe N
3
(a)
0
3600
2400
1200
0
3=2
Since KP is a function of temperature, r ¼ PNH =½PH
ꢀ
3
2
can be regarded as the ability of nitriding at different
temperatures. The experimental results show that the N
content of magnetic particles increase with increasing NH3
flux.
(b)
(c)
9000
The non-magnetic Fe2N phase was found when the ratio
of Ar2:NH3 decreased to less than 0.5. The existence of
non-magnetic Fe2N phase decreased the saturation mag-
netization of magnetic fluid. Thus, the proper flux ratio of
Ar2:NH3 is very important.
4500
0
20
40
60
80
100
120
2 θ (degree)
Fig. 2. X-ray diffraction patterns of the magnetic particles prepared under
different Ar2: NH3 flux rates ((a) Ar2:NH3 ¼ 1:0.25; (b) Ar2:NH3 ¼ 1:0.6;
(c) Ar2:NH3 ¼ 1:1;).
3.3. The effect of surfactant content on the stability of
magnetic fluids
The magnetic particles are easy to agglomerate and settle
due to their magnetic dipole interaction. When the
surfactant with long chained molecules is added into the
carrier liquid to coat the magnetic particles, the agglom-
eration and settling of magnetic particles can be avoided
due to the repulsion between the magnetic particles coated
by the surfactant. Of course, the surfactant must be
compatible with the carrier liquid and can coat on the
surface of magnetic particles. Thus, the choice of proper
surfactant was very important. A kind of non-ionic
surfactant (succinicimide), which is compatible very well
with the carrier liquid (a-olefinic hydrocarbon synthetic
oil), was found after many experiments.
The surfactant content is also very important to the
stability of magnetic fluids. In order to determine the best
surfactant content, different amounts of surfactant (succi-
nicimide) were added into the carrier liquid, respectively, as
shown in Table 1. And then these magnetic fluids were
synthesized under the same experimental condition. After
synthesized, each of the five magnetic fluids (sample
FN001–005) was divided into two groups. One group was
laid in the natural state and the other group was laid on the
magnet.
It can be noticed from Table 1 that the settling was
found for the samples FN001–003. The agglomeration and
settling of particles came from the coat incompletely of
surfactant due to the deficiency of surfactant content. For
the samples FN004 and FN005, in which the volume ratios
of surfactant were 20 and 23.8%, respectively, no
agglomeration and settling were found and the stable
magnetic fluid was obtained even laid on the magnet. It
shows that the magnetic particles were coated completely
by the surfactant and dispersed in the carrier liquid
homogeneously (Fig. 3) when the volume ratio of
surfactant was in the range of 20–23.8%. However, the
viscosity of magnetic fluid increased with increasing the
content of non-ionic surfactant (succinicimide). Thus, it
was thought that the optimal content of non-ionic
surfactant (succinicimide) is 20% (volume ratio).
obtained successfully at 182 1C by controlling the gas flux
ratio of Ar1:Ar 2:NH3 suitably.
3.2. The influence of gas-flux ratio
In this experiment, the relationships between the gas flux
ratios of Ar1:Ar2:NH3 and the phase compositions of the
magnetic particles are studied. In order to analyse easily,
the gas fluxes of Ar1 and Ar2 were fixed at 280 and 80 ml/
min, respectively, and the gas flux of NH3 was the only
variable during the experiment. Thus the ratio of
Ar1:Ar2:NH3 could be expressed simply by the ratio of
Ar2:NH3.
Fig. 2 gives the relationships between the flux ratios of
Ar2:NH3 and the reactive products. The results show that
when Ar2:NH3 ¼ 1:0.25, the reactive product was com-
posed of Fe and Fe4N phase (Fig. 2a). It meant that the
nitriding was incomplete. The reactive product was
composed of Fe4N and Fe3N (Fig. 2b) when the flux ratio
of Ar2:NH3 ¼ 1:0.6. The single-phase e-Fe3N was obtained
when the flux ratio of Ar2: NH3 ¼ 1: 1 (Fig. 2c). The
reactive process can be expressed by the Eqs. (1) and (2):
FeðCOÞ5 ! Fe þ 5CO ";
Fe þ NH3 ! FehNH3i ! ½Nꢀdiffuse into Fe þ 2 H2 " :
(1)
(2)
3
From Eq. (2) it can be noticed that the equilibrium
constant KP is controlled by the pressures of the NH3 and
H2. The KP under the different fluxes of NH3 can be
expressed by the equation
3=2
½PH
ꢀ
NH3
2
KP ¼
aN.
(3)
P
Thus
aN ¼ KP
P
NH3
,
(4)
3=2
½PH
ꢀ
2

ARTICLE IN PRESS
W. Huang et al. / Journal of Magnetism and Magnetic Materials 307 (2006) 198–204
201
Table 1
Effect of the contents of surfactant on the stability of the magnetic fluid
No.
Carrier (polyalphaolefin) (ml)
Surfactant (succinamide) (ml)
Laid in natural state
Laid on the magnet
FN001
FN002
FN003
FN004
FN005
80
80
80
80
80
5
10
15
20
25
Settling after 1 day
Settling after 3 day
Settling immediately
Settling gradually
Settling slowly
Settling after 9 day
No setting after 180 days
No setting after 180 days
No setting after 180 days
No setting after 180 days
side due to the existence of many fine pores. The pore size
can be changed by changing the density of the stainless
steel plate.
Fig. 3 is the TEM images of magnetic particles
synthesized using two different porous plates. It can be
noticed that the size of magnetic particles shown in Fig.
3(b) is smaller than that of Fig. 3(a). It is because that the
pore size of porous plates used in Fig. 3(b) is smaller than
that of used in Fig. 3 (a). Also it can be found from Fig. 3
that the e-Fe3N particles synthesized in our experiment are
sphericity and have very narrow size distributions.
According to Ref. [8], the particle size distributions can
be obtained by measuring the largest dimension of each of
the particle images within a defined region of a micrograph
[8]. By this method, the size distributions of the magnetic
particles synthesized using different porous plates were
measured and are given in Fig. 4. The results show that all
the fluids examined had distributions that appear approxi-
mately to be Gaussian. The mean particle sizes and the
size distributive ranges of magnetic particles are given in
Table 2.
Fig. 3. TEM images of magnetic particles prepared using different porous
plates ((a) the pore size range of the porous plate is 20–30 mm and the
volume ratio of surfactant is 20%; (b) the pore size range of the porous
plate is 5–10 mm and the volume ratio of surfactant is 20%). Both of them
show that the magnetic particles were coated completely by the surfactant
and dispersed in the carrier liquid homogeneously.
3.4. The effect of the porous plate on the size of magnetic
particles
It can be noticed from Fig. 4 that the size distribution of
the magnetic particles increases with increasing the pore
size of porous plate. It is because that the amount of vapor
Fe(CO)5, which passes through the porous plate in unit
time, increases with increasing the pore size, namely, the
content of Fe crystals in the carrier liquid increases. Thus
the collision probability of Fe crystals increases and larger
magnetic particles form. Therefore reducing the pore size
of porous plate is an effective way to obtain the e-Fe3N
magnetic fluid with small size magnetic particles.
It has known that the size of magnetic particles has much
more effects on the magnetic properties and stability of the
magnetic fluids. The size of a ferromagnetic particle should
be about 10ꢁ8 m. There are two reasons why the particles
should be small. Firstly, this size provides a mono-domain
structure of particles. Secondly, such particles can be
suspended in the fluid by Brownian motion. Thus, it is very
important to control the size of magnetic particles. The size
of magnetic particles can be effected by many different
factors during the synthesis of magnetic fluid, such as the
reactive temperature, the surfactant content, the evapora-
tion temperature of Fe(CO)5, the gas current ratio and
so on.
When all the experimental conditions mentioned above
were fixed, it was found that the pore size of the porous
plate used in our experiment has much more effect on the
size of the magnetic particles. It has been mentioned above
that the porous plate is a kind of 3 mm thickness stainless
steel plate prepared by model-press and sintered the
stainless steel powder. The gas molecule can pass through
this stainless steel plate easily from one side to the other
3.5. Effect of the reactive time on the magnetic property
Besides the magnetic properties of magnetic particles, the
content of magnetic particles (the volume fraction of
magnetic particles in the magnetic fluid) has an important
effect on the saturation magnetization (Ms) of magnetic
fluid [8]. In order to obtain high saturation magnetization
e-Fe3N magnetic fluid, two kinds of the methods were used
to increase the volume fraction of magnetic particles. One
was prolonging the reactive time; the other was adding
some n-heptane into the carrier liquid before reaction.
Fig. 5 shows the magnetization curves of six samples
(A–F) with different reactive times and the carrier liquids

ARTICLE IN PRESS
W. Huang et al. / Journal of Magnetism and Magnetic Materials 307 (2006) 198–204
202
Table 2
Mean particle sizes and the size distributive ranges of the magnetic
particles synthesized using the porous plates with different pore sizes
70
60
50
40
30
20
10
0
Pore size of the porous plate (mm)
Mean particle size (nm)
Size distributive range (nm)
5–10
12
10–20
16
12–19
20–30
20
17–24
9–15
The mean particle sizes were calculated by measured the size distributions
of the magnetic particles within a defined region of TEM image.
1800
1600
5
5
5
10
10
10
15
20
20
20
25
25
25
A
(a)
(b)
(c)
(nm)
1400
1200
1000
800
600
400
200
0
B
C
D
60
50
40
30
20
10
0
E
F
0
2500
5000
7500
10000 12500 15000
Applied magnetic field (Oe)
Fig. 5. Magnetization curves of the magnetic fluids prepared under
different reactive times (The carrier liquid of the samples A, E and F is
composed of 80 ml a-olefinic hydrocarbon synthetic oil and 20 ml
succinicimide and the carrier liquid of the samples B, C and D is
composed of 80 ml a-olefinic hydrocarbon synthetic oil, 20 ml succinici-
mide and 40 ml n-heptane. The synthetic temperature is 182 1C. The gas
fluxes of Ar1 (Argon gas), Ar2 (Argon gas) and NH3 were controlled at
280, 80 and 80 ml/min, respectively. The reactive times are B:36, C:34,
D:31, A:60, E:34, F:24 h).
15
(nm)
60
50
40
30
20
10
0
2000
A
1600
B
1200
800
400
0
C
15
D
(nm)
Fig. 4. Particle size distributions measured from TEM images. The pore
size ranges of the porous plates used are: (a) 5–10; (b) 10–20; (c) 20–30 mm.
E
Adding n-heptane
No n-heptane
F
of samples A, E and F are composed of 80 ml a-olefinic
hydrocarbon synthetic oil and 20 ml succinicimide (non-
ionic surfactant). And the carrier liquids of samples B, C
and D are composed of 80 ml a-olefinic hydrocarbon
synthetic oil, 20 ml succinicimide (non-ionic surfactant)
and 40 ml n-heptane. Fig. 6 gives the relationships between
the saturation magnetizations of iron nitride magnetic
fluids and the reactive times. The volume fractions of
magnetic particles and the saturation magnetizations of
samples A–F are given in Table 3. It can be found that the
volume fraction of magnetic particles and the saturation
20
30
40
50
60
70
Reaction time (h)
Fig. 6. Relationships between the saturation magnetizations of iron
nitride magnetic fluids and the reactive times (The carrier liquid of the
samples A, E and F is composed of 80 ml a-olefinic hydrocarbon synthetic
oil and 20 ml succinicimide and the carrier liquid of the samples B, C and
D is composed of 80 ml a-olefinic hydrocarbon synthetic oil, 20 ml
succinicimide and 40 ml n-heptane. The synthetic temperature is 182 1C.
The gas fluxes of Ar1 (Argon gas), Ar2 (Argon gas) and NH3 were
controlled at 280, 80 and 80 ml/min, respectively. The reactive times are
B:36h, C:34h, D:31h, A:60h, E:34h, F:24h).

ARTICLE IN PRESS
W. Huang et al. / Journal of Magnetism and Magnetic Materials 307 (2006) 198–204
203
Table 3
The volume fractions of the magnetic particles and the saturation magnetizations of samples A–F (The carrier liquid of the samples A, E and F is
composed of 80 ml a-olefinic hydrocarbon synthetic oil and 20 ml succinicimide and the carrier liquid of the samples B, C and D is composed of 80 ml a-
olefinic hydrocarbon synthetic oil, 20 ml succinicimide and 40 ml n-heptane
Sample
F
E
D
C
B
A
Vol. % of the magnetic particle
Ms (Gs)
1.67
285
3.43
544
6.81
963
8.03
1214
10.49
1465
10.92
1633
The synthetic temperature is 182 1C. The gas fluxes of Ar1 (Argon gas), Ar2 (Argon gas) and NH3 were controlled at 280, 80 and 80 ml/min, respectively.
The reactive times are B: 36, C: 34, D: 31, A: 60, E: 34, F: 24 h.
2000
1600
1200
800
400
0
magnetization of magnetic liquid increase with increasing
reactive time. It also can be noticed from Figs. 5 and 6 that
the magnetic fluid with higher saturation magnetization
could be obtained in shorter reactive time if some n-
heptane is added into the carrier liquid before reaction. The
reason is that the n-heptane is a kind of low boil oil (boiling
point 97 1C). The high synthetic temperature (182 1C) of e-
Fe3N magnetic fluids will result in the evaporation of n-
heptane. Since three organic solutions (a-olefinic hydro-
carbon synthetic oil, succinicimide and n-heptane) have
mixed together to form the homogeneous liquid, it is
inevitable that the evaporation of n-heptane results in some
loss of carrier and surfactant. Thus the volume fraction of
magnetic particles increases during the same reactive time,
and the saturation magnetization of magnetic liquid
increase as shown in Table 3.
A
Experimental results
Calculated result
B
C
D
E
F
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
Density (g/cm3)
Fig. 7. Relationships between the saturation magnetizations and the
densities of the magnetic liquids (’ experimental results;
result according to the equation Ms(r) ¼ ꢁ0.2025+0.2146r).
calculated
3.6. Relationship between the saturation magnetization and
the density
Because the density of iron-nitride magnetic liquid is
directly proportional to the volume fraction of magnetic
particles, the relationships between the saturation magne-
tizations and the volume fractions of magnetic particles can
also be expressed by the relationships between the
saturation magnetizations and the densities of iron-nitride
magnetic fluids as shown in Fig 7. It can be noticed that the
saturation magnetization increases linearly with increase in
the density of iron-nitride magnetic liquid. The relation-
ships between the saturation magnetizations and the
densities of iron-nitride magnetic liquids can be expressed
by the equation
the size of magnetic particles, the magnetic properties and
the stability of e-Fe3N magnetic fluid were studied.
It was found that when the heat decomposition
temperature of Fe(CO)5 and the synthetic temperature of
iron nitride magnetic fluid were controlled at 182 1C, the
gas fluxes of Argon gas (Ar1 and Ar2) were controlled at
280 and 80 ml/min, respectively, the gas flux ratio of Argon
gas (Ar2) and NH3 gas was controlled at Ar2:NH3 ¼ 1 and
the carrier liquid was composed of 80% a-olefinic
hydrocarbon synthetic oil and 20% non-ionic surfactant
(succinicimide) and stable single phase e-Fe3N magnetic
fluid was synthesized successfully. The saturation magne-
tization (Ms) of e-Fe3N magnetic fluid increased with
increasing reactive time. The size of e-Fe3N magnetic
particles decreased with decreasing the pore size of porous
plate used in experiment. The synthetic time of high
saturation magnetization (Ms) e-Fe3N magnetic fluid could
be shortened by adding some n-heptane into the carrier
liquid before reaction. The relationships between the
saturation magnetizations and the densities of e-Fe3N
magnetic fluids could be expressed by a linear equation.
Finally, stable e-Fe3N magnetic fluid with the saturation
magnetization 1663 Gs and the mean particle size 12 nm
was synthesized successfully.
MsðrÞ ¼ ꢁ0:2025 þ 0:2146r,
(5)
where Ms (T) is the saturation magnetization and r (g/cm3)
is the density of magnetic fluid. According to the Eq. (5),
when Ms ¼ 0, r ¼ 0:943 g=cm3, which equals approxi-
mately the density of carrier liquid (0.95 g/cm3).
4. Conclusion
e-Fe3N based magnetic fluid was synthesized by a home-
made apparatus. The effects of the carrier compositions
and the experimental conditions on the phase composition,

ARTICLE IN PRESS
W. Huang et al. / Journal of Magnetism and Magnetic Materials 307 (2006) 198–204
204
Acknowledgement
[2] I. Nakatani, M. Hijikata, K. Ozawa, J. Magn. Magn. Mater. 122
(1993) 10.
[3] H. Mamiya, I. Nakatani, J. Magn. Magn. Mater. 177–181 (1998) 966.
[4] H.S. Lee, I. Nakatani, J. Magn. Magn. Mater. 201 (1999) 23.
[5] L. Xuehui, Z. Jibin, A. Hong, C. Erbin, L. Jun, J. Magn. Magn.
Mater. 252 (2002) 330.
This work was supported by the national 863 Project
(no: 2002AA302608), from the Ministry of Science and
Technology, China.
[6] L. Xia, Harbin Institute of Technology Book Cencern, 1996, p. 124
(in chinese).
Reference
[7] Z. Zhikun, C. Zoulin, National Defence Industry Press, 2000, p. 34
(in chinese).
[8] V. Socoliuc, D. Bica, L. Vekas, J. Colloid Interface Sci. 264 (2003) 141.
[1] L. Shen, P.E. Laibinis, T. Alan Hatton, J. Magn. Magn. Mater. 194
(1999) 37.