Journal of ELECTRICAL ENGINEERING, VOL 59.

NO 7/s, 2008, 15-20

CHARACTERISATION OF HARD MAGNETIC MATERIALS

Roland Grössinger*
Different methods determining the hysteresis loop of permanent magnets are summarised. In industry the use of a
hysteresograph (based on fluxmeters) within an electromagnet as a field source is most common. This is a magnetic closed system
where no demagnetizing effects have to be considered. A Vibrating Sample Magnetometer (VSM) is a magnetically open system,
but it offers a significantly higher sensitivity. If as a field source a superconducting coil is used fields up to 10T are easy to get.
The system with the highest sensitivity, which allows the measurement of single grains or thin films, is a SQUID-magnetometer.
SQUID magnetometers are available with external fields up to 7 T, however such instruments are more common in scientific
laboratories. A Pulsed Field Magnetometer (PFM) offers sufficient high external fields (10 T and more) together with a fast
measuring system. Disadvantageous is that in such system time dependent effects such as eddy currents and magnetic viscosity
may limit the achievable accuracy. However for a quality control a PFM is a very suitable instrument.

K e y w o r d s : hard magnetic materials, magnetometers, pulsed fields

1 INTRODUCTION Table 1. Magnetic parameters of industrial permanent magnet

materials; WT means the maximum working temperature of the
material
The determination of the hysteresis loop of modern,
industrial permanent magnet materials needs sufficient Material (BH)max Coercivity WT (°C)
high magnetic fields combined with an easy to handle (MGOe) IH C (kOe)
measuring procedure. Since decades of years exists Nd-Fe-B
up to 45 up to 30 180
hysteresis measuring systems, however the development sintered
of modern electronic devices together with data Nd-Fe-B
up to 10 up to 11 150
bonded
acquisition systems allowed a dramatic improvement of
Sm-Co
the sensitivity, reproducibility and handling of magneto- up to 30 up to 50 350
sintered
meters. Additionally the use of superconducting magnets Sm-Co bonded
opened the door to sufficient high external fields. In the up to 12 up to 10 150
present paper an overview of generally used hysteresis AlNiCo up to 10 up to 2 350
measuring systems will be given. Special emphasis will be Hard Ferrite up to 4 up to 5 300
Flexible up to 2 up to 3 100
laid on industrial applications [1,2].
The hysteresis loop of hard magnetic materials can be
For defining the behaviour close to the working point
given in M(H) or B(H) (see Fig.1) – M(H) is important for
the shape of the demagnetizing curve in the second
characterising a material, B(H) is important for the
quadrant of the B(H) loop is necessary. In this case it is
optimization of a magnetic circuit.
sufficient to magnetize first the magnet (pulse magnetizer)
M and measure afterward only B(H) in the second quadrant.
B
MS
In the following the most common used magneto-
BR meters will be described and discussed with the target
HC MR
H H
B C H how suitable they are for determining the hysteresis loop
of (industrial) permanent magnets.
minor loop
virgin curve (BH)max 2 MAGNETOMETERS

Magnetometers can be subdivided by three main

Fig. 1 Left side M(H) including all important points such as
remanence and coercivity; right side B(H) of a permanent magnet;
here also the stored energy density (BH)max is shown.
The accurate determination of the loop needs a
sufficient high external magnetic field. Generally this
field should be two or three times larger than the coercive
field, however this depends also on the type of the
magnetization process (pinning or nucleation). In table 1
an overview of important properties of industrial
permanent magnets is given.
⎯⎯⎯⎯⎯
* Institute for Solid State Physics, Technical University Vienna, Wiedner Hauptstr. 8-10 Vienna, A-1040 Austria; rgroess@ifp.tuwien.ac.at
differences in technology:
(i) Magnetic open or magnetic closed system
Fig. 2 shows schematically the difference between a
magnetic closed and a magnetically open system.
Generally the internal field Hint is related to the external
field Hext according to:
H int = H ext − NM ( H )
N (0 ≤ N ≤ 1) is the demagnetizing factor which is in a
magnetic closed system zero, whereas in a magnetic open

ISSN 1335-3632 © 2008 FEI STU

16 R. Grössinger: CHARACTERISATION OF HARD MAGNETIC MATERIALS
system the sample geometry determines the value of N. In
order to get the “true” shape of the hysteresis the
measured loop M(Hext) has to be corrected to M(Hint) point
by point as shown in Fig. 3
Measurements are made in
“closed“ or “open“ magnetic
circuits
Open Circuit
Closed Circuit
Fig. 2. Left: Set-up for M(H) measurement in a magnetically
closed system; right: stray field as occurs at a M(H)
measurement in a magnetically open system.
Hard
Typical
Fig. 3. Typical M(Hext) blue curve and M(Hint) red curve of a
hard magnetic material
(ii) Measuring method for determining the magneti-
zation
Here the signal conditioning methods (dc or ac-method)
and the type of integration is important.
(iii) External field production
The field production method determines the achievable
maximum field as well as the measuring velocity.
2.1 Electromagnet
A hysterezisgraph using an electromagnet is most
common in industry. The reason is that this magnet needs
now sophisticated cooling facilities like eg liquid He. If it
is used just for a magnetometer at room temperature it is
generally a magnetic closed system as shown in Fig. 2
(left side). The advantages of the use of an electromagnet
are:
(i) Magnetically closed system – N = 0.
(ii) No sophisticated cooling facility necessary – simple
to use
The disadvantages are:
(i) Careful sample preparation – to plane surfaces.
(ii) Limited maximum field (up to 2T) – therefore pulse
magnetization for saturating the material is
necessary.
(iii) DC-integration – drift problems, limited sensitivity.
Figure 3 shows, as an example, the commercial
permegraph C300 from Magnet-Physik using an electro-
magnet with a maximum field of 1700 kA/m (2.1T).
Fig. 4. Typical permegraph C300 from Magnet-Physik [12]
2.2 Vibrating Sample Magnetometer
The Vibrating Sample Magnetometer (VSM) was
invented already more than 50 years ago and was a break
through in magnetic measuring methods [3]. A vibrating
sample magnetometer can be installed in an electromagnet
as well as in a superconducting magnet. Consequently it
can be operated over a broad temperature range. It is
always a magnetically open system, which means the
sample geometry has to be considered. The main
improvement is that here a different magnetization
measuring principle is used. The sample vibrates
periodically at the magnet gap or bore. As vibrating
device a motor, a loudspeaker or a piezoelectric
transducer can be used. Consequently a periodic induction
voltage is generated in the pick-up coil. This allows the
use of a lock-in amplifier with an enhanced signal to noise
ratio – see Fig. 5. There are nowadays easy to handle
systems available such as the PPMS from Quantum
Design (see: [13]). The advantages of such a VSM are:
(i) High sensitivity – up to 10-6 emu
(ii) Fully automated
(iii) Broad temperature range possible
(iv) Sufficient high fields possible – especially if a
superconducting magnet is used.
The disadvantages are:
(i) Magnetic open system – needs correction for
demagnetizing field
(ii) In the case of a superconducting magnet liquid He is
generally necessary – however nowadays some
companies offers already cryogen free magnets up to
16 T (see eg [14])
(iii) Generally small samples
(iv) Slow – limited dH/dt.
2.3 SQUID magnetometry
A SQUID consist of a superconducting ring which has
on one (rf-SQUID) or on two points (ac-SQUID)
Josephson contacts. Due to quantum-mechanics only
Journal of ELECTRICAL ENGINEERING, VOL 59. NO 7/s, 2008 17

multiples of Φ0 = 2.07×10-15 Vs can diffuse into the ring.

The compensating shielding current leads to voltage drop
at the SQUID. The current-voltage characteristic is
periodic, with a periodicity which corresponds to one flux
quantum. The principle is shown in Fig. 6.

Function Audio System

Computer control
Generator Amplifier
control electronics
system (individual
Vibrator
CAN
Pick-up coils Reference modules)
Signal
Hall probe Generator
Dewar
Electromagnet Sample soft iron
shield
Power Motor
Supply Controller
7 Tesla
vapor-cooled Nitrogen
superconducting jacketed
magnet dewar

Rotating Table
Stepping
Gauss Pre- Motor
meter Amplifier
Lock-In
Amplifier
Fig. 7. Cutaway view of the MPMS SQUID VSM, [16]
Personal
Computer The disadvantages are:
(i) Magnetically open system
System diagram of (ii) Complex handling
Vibrating Sample Magnetometer (iii) Small samples
(iv) Slow measuring system
Fig. 5. Principle of a vibrating sample magnetometer using an
electromagnet for generating the field and a lock in amplifier
for the signal detection. 2.4 Pulsed Field Magnetometer (PFM)
Pulsed field magnetometers allow a very fast measure-
Magnetic field ment of the hysteresis loop of hard magnetic materials in
sufficient high external fields.

Superconductor Power semiconductors

Biasing Biasing
current current Pickup coil

Field coil
Josephson
junction Instrumentation
Capacitor bank
Magnetic
Sample
0
=φ

voltage variation
for steady Fig. 8. Block diagram of a PFM.
- Δφ

increasing
magnetic flux
Fig. 6. Principle of a SQUID device, [15]
Figure 7 shows a sketch of a modern SQUID-
magnetometer from Quantum Design which allows
measurements of a very high sensitivity (10-8 emu) in
external fields up to 7 T. Due to its extremely high
sensitivity a SQUID system is generally used in scientific
laboratories but not in industry.
The advantages of a SQUID magnetometer are:
(i) Highest sensitivity (10-8 emu) – this allows the
measurement of thin films or single grains
(ii) Sufficient high field available
(iii) Temperature variable system possible
Fig.8 shows a block diagram of such a PFM. It has as
an energy source a condenser battery, a charger, a high
field magnet and a measuring device.
The condenser battery is charged up to a voltage between
1 kV and several kV (the stored energy is equal to:
CU2/2), then the condensers are discharged through the
pulse magnet. The capacitance C and the inductivity L
determines the pulse time, which lies typically between 1
– 100 ms. The current I0 exp(-tR/2L)sin(ωt) flows through
the pulse magnet and generates high pulsed field of the
shape H0 exp(-tR/2L)sin(ωt). With this technique fields up
to 20 T can be generated in magnet bores of up to
100 mm. Therefore such systems are well suited for the
characterisation as well as for the quality control of
18 R. Grössinger: CHARACTERISATION OF HARD MAGNETIC MATERIALS

modern permanent magnets [2,4]. Figure 9 shows a hysteresis plot; this is known as the f /2f method [7]. The
diagram of a commercial PFM as developed by Hirst Ind. equations used in the f /2f calculations are as follows:
J eddy = −σ dB / dt , M eddy ∝ J eddy
Controller
M magnet = μ H + M eddy
‘

Capacitor Bank As the measurement occurs at two frequencies it is possible

to deduce Meddy and to remove its component from Mmagnet.
Figure 10 shows the linear extrapolation procedure which has
Diode
Computer to be done point by point. Figure 11 shows the two M(H)
Thyristor
C++ software loops as measured with two different pulse durations as well
Pulsed magnet
as the corrected curve.
Coolant
J (T)
Pickup coil

1.28
J and H Integrators Data aquisition card
1.24

Fig. 9. Block diagram of the Hirst - PFM. 1.20

The sensitivity of a PFM is mainly determined by the 1.16
type of pick-up measuring coils which are chosen.
The advantages of a PFM for characterising permanent 1.14
magnets are:
(i) Sufficient high external fields - 10 T and more (even 10 20 30 40
at room temperature) are possible. Frequency (Hz)
(ii) The system can be operated at room temperature
Fig. 10. Extrapolation procedure for determining the “true”
(iii) Large sample volume is possible polarization value from two different PFM measurements.
(iv) Fast measuring process
The disadvantages of a PFM for characterising permanent
Magnetization (T)
magnets are: 1.6
(i) It is a magnetic open system Short pulse
Long pulse
(ii) Limited sensitivity 1.4

(iii) Due to high dH/dt two measuring errors have to be 1.2

considered:
Eddy current error and magnetic viscosity.
Because of their importance for a PFM these effect
will be discussed shortly.
Magnetization measurements in transient fields on
metallic conducting samples cause eddy currents. These
eddy currents influence the shape of the loop and the
value of the obtained magnetization. Therefore, eddy
currents and the by these eddy currents caused dynamic
magnetization was measured and analyzed theoretically
[5]. The influence of the eddy current error on the shape
of the hysteresis loop of hard magnetic materials can be
either calculated using a finite element method [6]. Or – if
this error is small – it can be corrected using the so-called
f /2f method which allows estimating the “true” hysteresis
loop [7]. The pulsed magnet is divided electrically into
two coils that could be pulsed independently, providing a
long and short duration pulse. The two pulse durations are
conventionally named f and 2f. The two frequency pulses
are used to measure each sample. Both magnets generate
the same M signal with respect to the applied field but
with the addition of eddy currents which are typically for
the different dH/dt. As the eddy currents are related to
frequency, by processing the two measurements, it is
possible to mathematically remove the eddy currents
effects, producing the direct equivalent of a static
1.0 Corrected
0.8
0.6
0.4
0.2
-1.8 -1.3 -0.8 -0.3
Applied field (T)
Fig. 11. M(H) measurement as measured with two
different pulse durations (40 and 57 ms) as well as the
corrected curve.
Generally it can be concluded that eddy current effects
depend on dH/dt, on the electrical conductivity of the
sample material and on the sample geometry – they scale
quadratic with the sample diameter [5]. This means that in
metallic samples large diameters call large problems.
The second time dependent effect is the so-called
magnetic viscosity. The effects of time dependence of the
magnetization process in magnetic materials have attracted
much attention for its importance in both theory and
application. Usually, the time dependence of the
magnetization reversal at constant applied field, the magnetic
after-effect MAE, is interpreted as originating from thermally
Journal of ELECTRICAL ENGINEERING, VOL 59. NO 7/s, 2008
activated barrier hopping processes, which in the case of a
flat energy barrier distribution can be described by:
M ( t ) = M 0 − S ln ( t / t0 )
Generally the investigation of the magnetic viscosity (S ─
viscosity coefficient) is used to get information concerning
the magnetization process [8]. In our case however the
magnetic viscosity has to be considered if the hysteresis loop
M(H) is measured in a transient field. In this case it causes
also a change of the shape of the loop and its important
parameters (such as coercivity) [9]. Figure 12 shows the
dependence of the coercivity of Sm(Co,Cu)5 on dH/dt [10].
The importance of the magnetic viscosity on the coercivity
becomes evident.
Hc (MA/m)
1.6
* Cu2.5
* Cu2.0 ° Cu2.5
1.2
* Cu3.0
0.8 * Cu1.5
° Cu2.0
SmCo5- xCux
0.4
0 5 10 15 20
dH/dt (GA/m/s)
Fig. 12. Coercive field as a function of the sweep rate
dH/dt, measured in the as cast (°) and (*) annealed
SmCo5-xCux alloys [10].
The magnetic viscosity depends mainly on the
microstructure (different phases, grain size, interfaces
etc). It is not geometry dependent which means
independent of the sample size and shape. It depends
exponentially on the field sweep rate dH/dt and it
increases with increasing coercivity. Therefore it exists
principally in all hard magnetic materials.
If one compares now the possibilities correcting for
time dependent effects the situation is as follows:
(i) The eddy current error is geometry dependent and
scales with the diameter of the sample. It can be
either corrected mathematically or by using powder
as samples.
(ii) The viscosity error is independent of geometry; it
cannot be corrected using a mathematical model. It
depends on the magnetization process and scales
with the coercivity. At room temperature the effect is
some times sufficient small that it can be neglected.
3 CALIBRATION OF MAGNETOMETERS
In a magnetometer the field and the magnetization has
to be calibrated.
FIELD CALIBRATION: in a static magnetometer –
where the field is produced either with an electromagnet
19
or with a superconducting coil - the field can be calibrated
with a hall probe and/or with an induction coil. In both
cases these field probes itself are calibrated using an
NMR-probe. In a PFM the field can also be calibrated
using the law of induction in combination with an
analytically describable H(t):
dH
U i = − NA
dt
However, also here, the main problem is the
determination of the “effective winding area” NA. If in a
PFM a condenser discharge is used for generating the
field the time dependence of the field H(t) is well known
and consequently the time dependence of the induction
voltage can be described analytically:
dB d
ui (t ) = − NA = − NAμ0 H 0 ( e − at sin ωt )
dt dt
This curve can be fitted to the measured data ui(t) and
consequently the field in a PFM can be calibrated with an
absolute error of ± 1% [11].
MAGNETIZATION CALIBRATION: in a static
magnetometer the magnetization is usually calibrated
using “well-known” materials such as pure Ni or
paramagnetic salts. In a PFM metallic samples cause an
eddy current error therefore insulating materials such as
soft magnetic ferrites are better suited. As an example in
Fig. 13 the magnetization of pure Fe3O4 calibration
sample as measured at room temperature is shown.
0.75
μ0 M (T)
0.50
Fe3O4 sphere
0.25 d = 5.5 mm
-1000 -500 500 1000
μ0 Hext (kA/m)
-0.25
-0.50
-0.75
Fig. 13. M(H) PFM measurement at room temperature on a
pure Fe3O4 spherical sample.
One problem using soft magnetic ferrites for a room
temperature calibration is, that these materials exhibit
rather large temperature dependence as shown in Fig. 14.
This means that the exact temperature has to be taken into
account. Comparisons with additional calibration experi-
ments using pure Ni and pure Fe samples show, that a
reliable calibration with an absolute error of ± 1.5 % also
for a PFM is possible [11].
In Fig. 15 a hysteresis measurement on a commercial
sintered Nd-F-B magnet is shown. There were two
measurements performed using two different pulse times.
The agreement of the data obtained for the coercivity and
20 R. Grössinger: CHARACTERISATION OF HARD MAGNETIC MATERIALS

the remanence shown in the inset demonstrate clearly the temperature). The use of a transient field can cause time
reliability of these measurements. dependent errors such as eddy currents as well as
magnetic viscosity. The eddy current error can be
corrected mathematically (f/2f method), the magnetic
Magnetization (Am2/kg)
viscosity error is at room temperature often sufficient
95 95
small. A PFM is for industrial purpose a well suited
90 Fe3O4 (table) 90 instrument for the quality control of permanent magnets.

85 Comparison between the temperature 85 REFERENCES

behavior of the magnetisation
80 of Iron-oxyde and the ferrite 3C30 80 [1] GRÖSSINGER, R. ─ SATO TURTELLI, R.: Encyclopedia of Materials:
TU Vienna Science and Technology, Elsevier, 2001, Magnetic Measurements:
75 75 Quasistatic and AC, (2001), 4832 – 4841.
[2] CORNELIUS, R.N. ─ DUDDING, J. ─ KNELL, P.A. ─ GRÖSSIN-
70 70
3C30 Philips (measured) GER, R. ─ ENZBERG-MAHLKE, B. ─ FERNENGEL, W. ─ KÜPF-
ERLING, M. ─ TARABA, M. ─ TOUSSAINT, J.C. ─ WIMMER, A. ─
65 65 EDWARDS, D.: Pulsed Field Magnetometer for Industrial Use, IEEE
10 20 30 40 50 60 70
Temperature (°C) Transactions on Magnetics, Vol. 38 (2002), No. 5, S. 2462 – 2464.
[3] FONER, S., “Versatile and sensitive vibrating-sample magneto-
Fig. 14. Temperature dependence of the saturation magneti- meter” Rev. Scient. Instr. 30 (1959) 548 – 557.
zation of commercial ferrite (3C30) in comparison with that of [4] DUDDING, J. ─ KNELL, P.A. ─ CORNELIUS, R.N. ─ ENZBERG-
Fe3O4. MAHLKE, B. ─ FERNENGEL, W. ─ GRÖSSINGER, R. ─
KÜPFERLING, M. ─ LETHUILLIER, P. ─ REYNE, G. ─ TARABA,
Kuaida N42H, Zn KW809 RA11x15x3.5 # 3 M. ─TOUSSAINT, J.C. ─ WIMMER, A. ─ EDWARDS, D.: A Pulsed
6
Long Pulse
Field Magnetometer for the quality control of permanent magnets, J. of
B (T) Magn. Magn. Mat., 242-245 (2002), 1402 - 1404.
Short Pulse
4 [5] GRÖSSINGER, R. ─ KÜPFERLING, M., KASPERKOVITZ, P. ─
WIMMER, A. ─ TARABA, M. ─ SCHOLZ, W. ─ DUDDING, J. ─
LETHUILLIER, P. ─ TOUSSAINT, J.C. ─ ENZBERG-MAHLKE, B. ─
2 FERNENGEL, W. ─ REYNE, G.: EDDY currents in pulsed field
measurements, J. of Magn. Magn. Mat., 242-245 (2002), 911 - 914.
0 [6] GOLOVANOV, C. ─ REYNE, G. ─ MEUNIER, G. ─
Short Pulse GRÖSSINGER, R. ─ DUDDING, J.: Finite Element Modelling of
Br = 1.27 T Permanent Magnets Under Pulsed Fields", IEEE Trans on Magn. Vol.
-2 Hcb= 0.885 MA/m 36 (2000) 1222-1225.
Long Pulse
Br = 1.26 T [7] JEWELL, G.W. ─ HOWE, D. ─ SCHOTZKO, C. ─ GRÖSSINGER,
-4 Hcb= 0.888 MA/m R.: A method for assessing eddy current effects in pulsed
magnetometry, IEEE Trans on Magn. Vol. 28 (1992) 3114-3116.
-3 -2 -1 0 1 2 3 4 [8] GIVORD, D. ─ TENAUD, P. ─ VADIEU, T.: Coercivity
Hint (MA/m) mechanism in Ferrite and Rare Earth transition metal sintered magnets
Fig. 15. Room temperature hysteresis measurements on a (SmCo5, Nd-Fe-B), IEEE Trans on Magn. MAG-24(2) (1988) 1921-
1923.
commercial Nd-Fe-B performed in a PFM using two
different pulse durations (40 and 57 ms Sin Full wave). [9] TELLEZ-BLANCO, J.C. ─ KOU, X.C. ─ GRÖSSINGER, R. ─
ESTEVEZ-RAMS, E. ─ FIDLER J. ─ MA, B.M.: Study of coercivity and
magnetic anisotropy of sintered Sm2Co17 type permanent magnets in the
4 SUMARY temperature range from 4.2K to 1000K, Proc. of 14th International
Workshop on Rare-Earth Magnets and their Applications, World Scientific
For measuring the full hysteresis loop of modern Ed. F.P. Missel, V. Villas-Boas, H.R. Rechenberg, F.J.G. Landgraf; (1996)
Sao Paulo Brasil; 707 – 716.
permanent magnets sufficient high external magnetic
[10] GRÖSSINGER, R. ─ TELLEZ-BLANCO, SATO ─ TURTELLI, R. ─
fields are necessary. For Rare Earth based magnets these HAUSER, R. ─ REITERER, K. ─ SASSIK, H. ─ CHOUTEAU, G.:
fields can be as high as 7 T. The industrially often used Determination of the magnetic viscosity in SmCo5-xCux alloys by pulsed-
electromagnet allows only the measurement of the field method; Physica B 294 –295 (2001) 194 – 198.
demagnetizing curve (up to 2 T), however after a pulse [11] GRÖSSINGER, R. ─ TARABA, M. ─ WIMMER, A. ─
magnetizing procedure. The use of a VSM in combination DUDDING, J. ─ CORNELIUS, R. ─ KNELL, P. ─ BISSEL, P. ─
ENZBERG-MAHLKE, B. ─ FERNENGEL, W. ─ TOUSSAINT, J. ─
with a superconducting magnet (which offers much higher EDWARDS, D.: Calibration of a Pulsed Field Magnetometer, IEEE
fields) is a very reliable solution for determining the full Trans on Magn. Vol. 38(5) (2002) 2982-2984.
loop. For industrial applications the generally small [12] http://www.magnet-physik.de/magnetic-measuring-technique
sample size together with the rather low measurement _permagraph-c-300.html).
speed is prohibitive. A SQUID magnetometer has the [13] http://www.qdusa.com/products/ppms.html
highest sensitivity however this is generally not necessary [14] http://www.cryogenic.co.uk/
for hard magnetic materials with a sufficient high [15] http://hyperphysics.phy-str.gsu.edu/Hbase/solids/squid.html
moment. [16] http://www.qdusa.com/products/squidvsm.html
A pulsed field magnetometer (PFM) is a fast system
which offers external fields up to 10 T (even at room
Received 28 September 2008