PIERS Proceedings, Beijing, China, March 2327, 2009

1348

On the Use Complex Susceptibility Measurements in Investigating

the Field Dependence of Resonance and After-effect Function of
Nano-particle Colloids
P. C. Fannin
Department of Electronic and Electrical Engineering, Trinity College, Dublin 2, Ireland

Abstract Measurement of the field and frequency dependent magnetic complex susceptibility,
(, H) = 0 (, H) i00 (, H), of nano-particle colloids, such as magnetic fluids, has proven to
be a reliable method for investigating a number of important properties of such fluids including
ferromagnetic resonance, relaxation mechanisms and other important parameters. Also, because
of the inverse Fourier transform relationship which exists between the after-effect function, b(t),
and 00 (), b(t) may be obtainedgnd used in determining a value of the precessional decay time,
0 .
Here, examples of results obtained from measurements obtained by means of the transmission
line technique in the MHz-GHz range, are presented.

1. INTRODUCTION

Magnetic fluids are stable colloidal systems consisting of single-domain magnetic particles covered
by a surfactant in order to prevent particle agglomeration and dispersed in a carrier liquid. The
particles have radii ranging from approximately 210 nm and when in suspension their magnetic
properties can be described by the paramagnetic theory of Langevin. The particles are considered
to be in a state of uniform magnetization with a magnetic moment m = Ms v, where Ms denotes
the saturation magnetization and v is the magnetic volume of the particle.
There are three characteristic times which govern the behaviour of a particle and two of these are
associated with the relaxation of the particles magnetic moment, namely the Brownian relaxation
time (B ) and the Neel relaxation time (N ). The third is the decay time associated with precession
(0 ) of the magnetic moment. All three times can be determined by means of complex susceptibility
measurement, () = 0 () i00 (), as can the presence of ferromagnetic resonance which occurs
at a frequency fres . However the time of particular interest here is 0 , which can be determined
by means of the unique relationship which exists between the after-effect function, b(t), and the
imaginary susceptibility component 00 ().
Here it is demonstrated how b(t), fres and 0 are influence by the application of a polarizing
field, H.
2. COMPLEX SUSCEPTIBILITY

(), of an assembly of single domain particles can also be described in terms of its parallel, || (),
and perpendicular, (), components, with [1]
() =

1
|| () + 2 () .
3

(1)
(0)

||
, where || (0) is the static
|| (), can be described by the Debye equation, with || () = 1+i
||
parallel susceptibility and || is the parallel relaxation time.
Over the frequency range considered here relaxation due to Brownian rotational diffusion of the
particles will, in general, be ignored and thus the || component, will be considered to be dominated
by the Neel relaxation mechanism with relaxation time N .
The perpendicular or transverse susceptibility, (), can have a resonant character, this effect
being indicated by a transition in the value of 0 () from a +ve to a ve quantity at an angular
frequency, res = 2fres .
Under equilibrium conditions, the magnetic moment, m, and the anisotropy field, HA , of a
particle are parallel and any deviation of the magnetic moment from the easy axis direction results
in the precession of the magnetic moment about this axis. If the polar angle is small, res , is
given by [1],
res = 2fres = gHA
(2)

Progress In Electromagnetics Research Symposium, Beijing, China, March 2327, 2009

1349

HA = 2K/Ms , where K is the anisotropy constant in J/m3 and is the gyromagnetic ratio. If
a radio frequency field is applied perpendicular to HA , the motion of the magnetic moment has
a typical resonant character which is commonly described by means of the Landau and Lifshitz
equations, namely,

1 + 2 02 + i0
()
=
(3)
(0)
(1 + 2 ) 02 2 + 2i0
where is a damping parameter and the precessional decay time, 0 = (0 )1 .
The after-effect function, b(t), represents the decay of magnetization after the sudden removal
of an external polarizing magnetic field, and () and b(t) are related by the expression [2],

00

1 ()
b(t) = 2Re F
(4)

where F 1 denotes the inverse Fourier transform.

Scaife [3] has shown that b(t) for the Landau and Lifshitz equations, has the form,

t
b(t) = b(0) exp
cos 0 t
0
If the area under b(t),

b(t)dt = B say, then it follows that, B = b(0)

resulting in a normalised value of [4],

(1/0 )
B
0
=
=
2
2
b(0)
1 + 02 02
(1/0 ) + 0

(5)

dt cos 0 t exp(t/0 ),

(6)

from whence, by knowing B/b(0) and 0 , 0 can be determined.

3. SUSCEPTIBILITY MEASUREMENT

Measurements reported here were made by means of the short-circuited, coaxial transmission line
technique [5, 6] using a Hewlett Packard (HP) 50 coaxial line incorporating a co-axial cell, in
conjunction with an HP 8753C network analyser. To obtain polarised measurements the coaxial cell
containing the ferrofluid sample, terminated in a standard HP short circuit load, was placed between
the pole faces of an electromagnet, the axis of the cell being perpendicular to the biasing field. The
biasing field, H, was altered between 0 and 104 KAm1 . Automatic swept measurements of the
input impedance of the line containing the sample were measured and from these measurements
the complex components, 0 () and 00 () were determined.
4. RESULTS

Measurements are presented for a 330 Gauss fluid consisting of Mn0.5 Fe0.5 Fe2 O4 particles suspended
in Isopar M. Measurements were performed over the range 100 MHz to 6 GHz and for 10 values of H.
The curves where then fitted [7] up to 10 GHz in order to facilitate a more accurate determination
of the inverse Fourier transform of 00 ().
Figure 1 shows a plot of the 0 () and 00 () components of the sample and from these plots it
can be seen that when H = 0, fres = 1.45 GHz, and when H = 104 KAm1 , fres rises to 5.7 GHz.
The 00 () component has a high frequency loss peak that occurs at a frequency of fmax = 0.92 GHz
at H = 0. This shifts to 5.7 GHz at H = 104 KAm1 . Thus the value of fmax approaches the value
of fres as resonance becomes the dominant process.
A plot of fres against H is shown in Fig. 2 and as res = 2fres = (H + HA ), the value of HA
is found from the intercept of Fig. 2 and determined as being equal to 37 KAm1 , corresponding
to a mean value of K, of 3.5 103 J/m3 . Form the slope of Fig. 2. is found to be 2.5 105 s1 A1 m.
Figure 3 shows the 00 () data which was transformed to obtain the b(t) profiles shown in Fig. 4
and one can observe how, over the polarizing field range, b(t) changes from an exponential type
decay to an oscillatory one. This transition arises because with increasing H, the parallel relaxation
component diminishes its contribution to the overall susceptibility and b(t) becomes similar to that
of the Landau Lifshitz form.

PIERS Proceedings, Beijing, China, March 2327, 2009

1350

Figure 1: Plot of 0 and 00 against f (Hz).

Figure 2: Plot of fres against H.

Figure 3: Plot of fit to 00 against f (Hz).

Figure 4: Plot of b(t) against t sec.

Figure 5: Plot of 1 and 2 against H.

Figure 6: Plot of fmax /fres against H.

Computing the area, B, under the b(t) curves and using the fres values, 0 as a function of H,
was determined by use of Eq. (6). The results are shown plotted in Fig. 5. As can be observed
from Fig. 5, for the analysed sample, Eq. (6) has real solutions only for polarizing fields larger
than 60 KA/m. This is due to the fact that the approximation used in computing the solutions of
Eq. (6) (i.e., 0 = 2fres ) is valid only for strong polarizing fields. One also notes that, Eq. (6) is
a quadratic equation and has two solutions, 1 and 2 The correct values of 0 can be chosen by
simply testing the values of 1 and 2 with the relation = (0 0 )1 , where cannot be larger
that one. Performing this test for all solutions of 1 and 2 , we determine that the correct solutions
of Eq. (6) are those of 1 (see Fig. 5) and at large values of H a mean value of 0 = 2.05 1010 s
is obtained.
In order to determine that the value of for our sample is indeed less than 1, we have also

Progress In Electromagnetics Research Symposium, Beijing, China, March 2327, 2009

1351

determined by an alternative method, which is described in Ref. [8]. Here it is shown that in the
region of strong polarizing fields,
s
1 (fmax /fres )2
=
(7)
1 + (fmax /fres )2
From Fig. 6, one can observe that in polarizing fields larger than 60 KA/m, the ratio fmax /fres
tends to have a constant value of 0.98 and using this value in Eq. (7) results in = 0.14. Thus
verifying the criteria that of the sample should be less than 1.
5. CONCLUSIONS

Using the frequency and polarizing field dependencies of the complex magnetic susceptibility,
(, H) = 0 (, H) i00 (, H), and the corresponding after-effect functions, the determination
and subsequent observation of the behaviour of the resonant frequency, fres , and the precessional
decay time, 0 , under the strong polarizing fields has been presented. The significance of the area,
B, under the after-effect functions has been highlighted and it has been shown that 0 can be
B
determined from the expression b(0)
= 1+20 2 ; this method being based on the analysis of the
res 0
after effect function of a magnetic fluid at resonance.
The determined value of 0 = 2.05 1010 s lies within the generally accepted range for 0 .
ACKNOWLEDGMENT

B. K. P. Scaife and C. Couper are acknowledged for useful discussions and also ESA for funding
this work.
REFERENCES

1. Raikher, Y. L and M. I. Shliomis, Sov. Phys. JETP, Vol. 40, 526, 1975.
2. Fannin, P. C., Wideband measurement and analysis techniques for the determination of the
frequency-dependent, complex susceptibility, Adv. in Chem. Phys., Vol. 104, 181, 1988.
3. Scaife, B. K. P., On the theory of the complex frequency-dependent susceptibility of magnetic
fluids, Adv. in Chem. Phys., Vol. 109, 1, 1999.
4. Fannin, P. C., C. N. Marin, and C. Couper, J. Magn. Magn. Mater, In Press.
5. Roberts, S. and A. R. von Hippel, J. App. Phys., Vol. 17, 610, 1946.
6. Fannin, P. C., T. Relihan, and S. W. Charles, J. Phys. D: Appl. Phys., Vol. 28, 10, 2003, 1995.
7. Fannin, P. C., T. Relihan, and S. W. Charles, Phys. Rev. B, Vol. 55, 14423, 1997.
8. Fannin, P. C. and C. N. Marin, J. Magn. Magn. Mater, Vol. 299, 425, 2006.