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

net/publication/237299063

Magnetic properties of stainless steels at room and cryogenic temperatures

Article  in  Journal of Magnetism and Magnetic Materials · July 2009

DOI: 10.1016/j.jmmm.2009.01.002

CITATIONS READS

16 6,141

3 authors, including:

Paul Oxley
College of the Holy Cross
32 PUBLICATIONS   1,131 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Paul Oxley on 22 November 2017.

The user has requested enhancement of the downloaded file.

 ARTICLE IN PRESS

Journal of Magnetism and Magnetic Materials 321 (2009) 2107–2114

Contents lists available at ScienceDirect

Journal of Magnetism and Magnetic Materials

journal homepage: www.elsevier.com/locate/jmmm

Magnetic properties of stainless steels at room and cryogenic temperatures

Paul Oxley , Jennifer Goodell, Robert Molt
Physics Department, The College of the Holy Cross, 1 College Street, Worcester, MA 01610, USA

a r t i c l e in fo abstract

Article history: The magnetic properties of ten types of ferritic and martensitic stainless steels have been measured at
Received 10 September 2008 room temperature and at 77 K. The steel samples studied were in the annealed state as received from
Received in revised form the manufacturer. Our room temperature measurements indicate significantly harder magnetic
21 November 2008
properties than those quoted in the ASM International Handbook, which studied fully annealed
Available online 21 January 2009
stainless steel samples. Despite having harder magnetic properties than fully annealed steels some of
PACS: the as-received steels still display soft magnetic properties adequate for magnetic applications. The
75.50.Bb carbon content of the steels was found to affect the permeability and coercive force, with lower-carbon
steels displaying significantly higher permeability and lower coercive force. The decrease in coercive
Keywords: force with reduced carbon content is attributed to fewer carbide inclusions which inhibit domain wall
Magnetic measurement
motion. Cooling to 77 K resulted in harder magnetic properties. Averaged over the ten steels tested the
Ferromagnetic property
maximum permeability decreased by 8%, the coercive force increased by 14%, and the residual and
Stainless steel
Martensitic saturation flux densities increased by 4% and 3%, respectively. The change in coercive force when cooled
Ferritic is comparable to the theoretical prediction for iron, based on a model of domain wall motion inhibited
by inclusions. The modest changes of the magnetic properties indicate that the stainless steels can still
be used in magnetic applications at very low temperatures.
& 2009 Elsevier B.V. All rights reserved.

1. Introduction magnetic stainless steels have previously been made, although

some measurements have been made with martensite austenite
Ferritic and martensitic steels can have good soft magnetic mixtures [8] and austenitic steels [9].
properties showing high permeability, low coercive force, low There are two main purposes of the work presented here. The
residual magnetic flux density, and high magnetic saturation flux first is to measure the magnetic properties of a range of magnetic
density. These steels also have high electrical resistivity which is stainless steels in the annealed condition as-received from the
essential to minimize eddy current losses in a.c. magnet applica- manufacturer. The magnetic properties measured are the max-
tions. The steels can be manufactured to large dimensions when imum permeability (mmax), the coercive force (Hc), the saturation
required for particular applications, they can be machined easily, magnetic flux density (Bs), and the residual magnetic flux density
are widely available, and are relatively inexpensive. The combina- (Br) for ten stainless steels in round rod form. The second purpose
tion of these properties has meant that ferritic and martensitic of this work is to determine how these magnetic properties
stainless steels have been used in a broad range of magnetic change at very low temperature (77 K (196 1C)). Magnetic
applications [1]. They can be used in fuel injectors [2] and fuel stainless steels are particularly well suited to low-temperature
pumps on automobiles, in relays, motors, and generators, and they a.c. applications since eddy current heating losses are reduced by
also have potential for use in nuclear power plants [3]. the high electrical resistivity of the steel (typically 70 mO cm at
Despite the usefulness of ferritic and martensitic stainless room temperature and decreasing by only 40% at 77 K [10]).
steels in magnetic applications there has been only limited study Minimizing heating losses is essential when working at cryogenic
of the magnetic properties of these steels. Most studies have used temperature.
steels heat treated after purchase from the manufacturer [2,4–6].
Room temperature measurements of coercive force and maximum
permeability of selected fully annealed stainless steels are also 2.. Experiments
reported in the ASM Handbook published nearly 20 years ago [7].
To our knowledge no studies of the cryogenic properties of The ten stainless steels tested in this study were purchased in
round rod form in the annealed condition from the Fry Steel
Company (Santa Fe Springs, CA). Their chemical compositions are
 Corresponding author. Tel.: +1 508 793 2473; fax: +1 508 793 3367. listed in Table 1 and their rod diameters, Brinell hardnesses, and
E-mail address: poxley@holycross.edu (P. Oxley). electrical resistivities are given in Table 2.

0304-8853/$ - see front matter & 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmmm.2009.01.002
 ARTICLE IN PRESS

2108 P. Oxley et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 2107–2114

Table 1 B, is calculated by
Chemical composition of the stainless steels used in this study as provided by the Z t
manufacturer. 1
BðtÞ ¼ B dt (1)
N B AB 0
Steel Type Chemical composition (% by weight)
Due to the small size of the e.m.f. induced in the field coil,
Cr C Mn Si S P N Ni Mo eH, this e.m.f. was amplified by a factor of 190 before being
recorded by the oscilloscope. The magnetic field strength, H, is
410 M 11.7 0.13 0.45 0.28 0.003 0.019 0.047 0.27 0.14
416 M 12.2 0.12 1.14 0.45 0.32 0.017 0.026 0.15 0.11
calculated by
420 M 12.2 0.34 0.45 0.32 0.001 0.16 0.053 0.47 0.07 Z t
1
422 M 11.7 0.21 0.07 0.40 0.001 0.016 0.029 0.83 1.05 HðtÞ ¼  dt (2)
431 M 16.1 0.16 0.64 0.39 0.02 0.024 2.18 0.05 m0 NH AH 0 H
440A M 15.1 0.55 0.78 0.40 0.005 0.019 0.038 0.38 0.04
In all experiments: NH ¼ 300 turns, AH ¼ 4  105 m2, NB ¼ 100
430 F 16.4 0.017 0.43 0.27 0.008 0.023 0.019 0.18 0.02
430F F 17.5 0.027 0.45 0.40 0.34 0.027 0.043 0.34 0.36 turns, and AB was equal to the cross-sectional area of the rod
430FR F 17.6 0.029 0.42 1.34 0.31 0.014 0.13 0.31 under test. The current in the magnetizing solenoid oscillated
446 F 25.8 0.063 0.72 0.53 0.001 0.024 0.067 at a frequency of 4 Hz and could produce magnetic fields up to
4600 A/m which were sufficient to bring the rod samples close to
Type M ¼ martensitic, F ¼ ferritic.
saturation. A frequency of 4 Hz was chosen since this was high
enough to provide sufficiently large e.m.f’s to allow reliable
Table 2 measurement of B(t) and H(t), and was low enough to accurately
Physical properties of the stainless steels used in this study hardness values are give the d.c. magnetic properties. Prior to taking measurements in
quoted by the manufacturer. this way each sample was demagnetized by slowly reducing the
amplitude of a 4 Hz applied field from 5400 A/m to zero. For
Steel Diameter (inch) Brinell hardness (HB) Electrical resistivity (mO cm) [11]
different current amplitudes and, therefore, different applied
410 0.188 180 57 magnetic fields a series of hysteresis loops could then be
416 0.188 235 57 measured. A single hysteresis loop for the 430 stainless steel is
420 0.188 279 55 shown in Fig. 2. The tips of several loops with different applied
422 0.188 215 61
field amplitudes can be joined to produce the magnetization curve
431 0.265 270 72
440A 0.188 222 60 for the sample, as seen in Fig. 3 for the 430 steel. From the
430 0.188 153 60 magnetization curve the maximum permeability, mmax, is found
430F 0.188 184 60a simply from the maximum value of the quantity B/m0H, as
430FR 0.281 162 76a 
446 0.315 217 64 B 
mmax ¼ (3)
m0 Hmax
a
Indicates resistivity value from Dietrich [7].
Values for Hc and Br were found from hysteresis loops in which
the sample reached a magnetic flux density of 1.00 T. A value for
the saturation magnetic flux density, Bs, was calculated by finding
the value of B at which the slope of the magnetization curve
equaled m0, i.e.,
dB
at B ¼ Bs : ¼1 (4)
dðm0 HÞ
Since the samples were not driven fully into saturation linear
extrapolation was used to infer the flux density at which Eq. (4)
was satisfied (Fig. 4). Measurements of the magnetic properties at
cryogenic temperatures were made in exactly the same way as at
room temperature, except that the magnetizing solenoid, steel
sample, and flux and field coils were cooled in liquid nitrogen.

Fig. 1. Experimental apparatus used to measure the magnetic properties of

stainless steel rod samples. An a.c. current in the magnetizing solenoid produces
an oscillating magnetic field at the rod sample. The magnetic field is found by
integrating the amplified e.m.f. induced in the field coil and the magnetic flux
density is found by integrating the e.m.f. induced in the flux coil. From these
measurements the magnetic properties of the rods are calculated.

The experimental arrangement used to determine the mag-

netic properties of the rod samples is illustrated in Fig. 1. A 30-cm-
long rod sample was placed inside a 52-cm-long magnetizing
solenoid. Around the rod was wound a flux coil with NB turns
and cross-sectional area AB, and adjacent to this was the field coil
with NH turns and cross sectional area AH. Both the flux and field
coils were 3 cm long. A synthesizer controlled a home-built
current amplifier to produce an oscillating current in the
magnetizing coil which generated an oscillating field, H, at the
sample location. The e.m.f. induced in the flux coil, eB, was Fig. 2. Room-temperature hysteresis loop for the 430 stainless steel sample at
recorded with a digital oscilloscope and the magnetic flux density, magnetic fields up to 4200 A/m.
 ARTICLE IN PRESS

P. Oxley et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 2107–2114 2109

Fig. 3. Room-temperature magnetization curve (solid line) for the 430 stainless
Fig. 5. 300 and 77 K values of maximum permeability for all ten steel samples
steel sample. The tips of the hysteresis loops (dashed lines) are joined to produce
tested. The error bars represent the typical variation in the measured permeability
the magnetization curve. For clarity only some of the measured hysteresis loops
for multiple measurements.
are shown.

Fig. 6. 300 and 77 K values of coercive force for all ten steel samples tested. The
error bars represent the typical variation in the measured coercive force for
Fig. 4. Variation of the quantity dB/d(m0H) with magnetic flux density at room multiple measurements.
temperature for the 430 stainless steel sample. The saturation flux density (Bs) is
defined as the flux density at which dB/d(m0H) is equal to one. The inset shows a
linear fit to the last five data points and the extrapolation of the fit to determine a
value for Bs of 1.51 T for the 430 sample.

Two additional experiments were also performed. In the first

values of mmax were measured for the 430FR sample with applied
magnetic fields oscillating at frequencies in the range 2–30 Hz,
instead of simply at 4 Hz as was used for all other data. The
purpose of this test was to probe how eddy current shielding
affects the measured magnetic permeability and confirm that our
4 Hz measurements correctly represent the d.c. magnetic perme-
abilities of the steels. The second additional experiment was to
anneal the type 430 and 430F samples to 780 1C ourselves and
compare their magnetic properties to those steels in the as-
received condition and to those listed in the ASM International
Handbook.

3.. Results and discussion

Fig. 7. 300 and 77 K values of residual flux density for all ten steel samples tested.
The error bars represent the typical variation in the measured residual flux density
3.1. Room-temperature tests
for multiple measurements.

Our measured values of mmax, Hc, Br, and Bs at room

temperature and 77 K are shown in Figs. 5–8 and we begin by magnetic material was the 430FR sample. It had a maximum
discussing the room-temperature results. Considering all four permeability of 960, a coercive force of 265 A/m, a residual
magnetic properties measured at room temperature the softest magnetic flux density of 0.49 T, and a saturation magnetic flux
 ARTICLE IN PRESS

2110 P. Oxley et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 2107–2114

density of 1.47 T. Type 430FR stainless steel is designed to have resisted by inclusions for three reasons. Firstly, the wall area and
good soft magnetic properties and an especially low residual flux therefore wall energy is reduced when it resides on an inclusion
density [12], as we observed. Type 430FR also has the highest (the ‘‘surface tension’’ effect) [13]. Secondly, the intersection of a
electrical resistivity (see Table 2) of any of the steels tested due to wall at an inclusion site allows a rearrangement of magnetic poles
its high silicon content. Type 430FR steel would therefore be which reduces the energy of the poles [17]. Thirdly, when the
preferred for a.c. applications where eddy current losses are a inclusion is several times the wall thickness (0.1 mm for iron),
particular concern. spiked closure domains form around the inclusion. These domains
Exact comparisons between the measured magnetic properties act to reduce the free pole energy at the inclusion, but they also
of the steels are difficult due to different trace element levels and inhibit wall motion [14,16–18]. In annealed stainless steels coarse
thermo-mechanical histories of our samples. However, two carbides are significantly larger than 0.1 mm [19]. For these large
general trends can be identified and understood. Firstly, we inclusions spike domains, rather than the surface tension effect, is
observed a trend for the stainless steels with higher carbon the dominant effect inhibiting wall motion and the coercive force
content to have increased coercive forces and lower maximum is proportional to the volume fraction of inclusions [14]. In Fig. 9, a
permeabilities, as shown in Fig. 9. We interpret these changes as linear fit to the variation of coercive force with carbon content
due to higher concentrations of chromium carbide inclusions shows reasonable agreement. In Section 3.2, we also interpret
present in higher carbon steels. These inclusions inhibit domain changes in coercive force at low temperature with decreased wall
wall motion leading to harder magnetic properties. The influence mobility.
of inclusions on the coercive force of ferromagnetic materials has The effect of carbon on the magnetic properties of iron-carbon
been studied by many researchers [13–16]. Domain wall motion is alloys has previously been measured [20–22]. These experiments
showed that as carbon content is increased permeability
decreases, coercive force increases, while residual flux density
remains largely unchanged. These three dependencies are also
found in our tests of stainless steels and are seen in Fig. 9.
Experimentally, we found that saturation flux density is un-
affected by carbon content. This is expected since saturation flux
density is a structure insensitive property, little affected by the
presence of trace elements, such as the low levels of carbon in our
steels.
The second general trend we have observed is a reduction of
saturation flux density with chromium content. This reduction
has previously been reported for ferritic stainless steels [1] and for
iron chromium alloys [23] and in both cases a linear decrease in
saturation flux density was observed. Our data, displaying an
approximate linear dependence, is shown in Fig. 10. To reduce the
influence of differing carbon contents on Bs two graphs are shown
in Fig. 10, one for the ferritic steels with low carbon content and
one for the martensitic, higher carbon steels. Since the chromium
provides stainless steel with its corrosion resistance, any given
Fig. 8. 300 and 77 K values of saturation flux density for all ten steel samples
magnetic application in a corrosive environment should balance
tested. The error bars represent the typical variation in the measured saturation
flux density for multiple measurements and the uncertainty in the extrapolation to increased corrosion resistance against reduced saturation flux
complete saturation. density.

Fig. 9. Magnetic softness, as determined by a high permeability mmax and a low coercive force Hc, increases with reduced carbon content. The solid line in the coercive force
graph is a linear fit to the data, as predicted if coercive force is determined by large inclusions [14]. Residual and saturation flux densities (Br and Bs) are seen to be
insensitive to carbon content up to a value of 0.55% by weight, the maximum carbon content of any of the steels tested. The ferritic steels are the four data points with the
lowest carbon contents and are indicated by the hollow points. The measurements were taken at room temperature and each data point represents one of the stainless steel
samples tested.
 ARTICLE IN PRESS

P. Oxley et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 2107–2114 2111

Values for mmax and Hc for a subset of the steels tested here are increase grain size and so reduce grain boundary resistance to
also listed in the ASM International Handbook [7] and a wall motion. Upon slow cooling a low density of coarse chromium
comparison of the two sets of measurements is made in Table 3. carbides form, which inhibit domain wall motion less than a high
As can be seen in the table our measurements display lower density of fine carbides which typically exist before annealing. A
permeability and higher coercive force than those in the ASM process anneal does not raise the steel temperature high enough
Handbook. The major difference between the two studies is the to enter the austenite phase, and hence does not reach as high a
heat treatment performed on the steels prior to testing: the ASM level of carbon solubility as does a full anneal. This leads to finer
Handbook samples were fully annealed while our samples were carbides and harder magnetic properties than fully annealed
tested in the annealed state as received from the manufacturer steels. This is observed in Table 3 where the process annealed
(process annealed). steels used in this work display harder magnetic properties than
During full annealing the steel is heated to a temperature the fully annealed steels in the ASM study.
above the critical point where an austenite phase in the iron If the softer magnetic properties of the steels listed in the ASM
occurs. This dramatically increases the solubility of carbon in the Handbook are to be attributed to the full annealing cycle
iron’s ferrite phase [24] and relieves internal strains which would undergone by the ASM steels, then we would expect those steels
otherwise impede magnetic domain wall motion. It can also to also be mechanically softer than the steels tested in our work.
This is because the more homogenous, coarser carbide micro-
structure produced by full annealing resists mechanical deforma-
tion less than the finer microstructure of process annealed steels.
Experimentally there is clear evidence that magnetic hardness of
stainless steels is correlated with mechanical hardness [3,5,25,26].
Table 3 shows that, indeed, in every case the steels reported in the
ASM Handbook have a lower Brinell hardness than the steels
tested in our work. The magnetic properties of as-received
annealed stainless steels presented in our work, therefore,
supplement and extend those of fully annealed steels listed in
the ASM Handbook. Anyone wishing to use magnetic stainless
steels without the inconvenience of performing a full anneal
should, therefore, look to our results for information on their
expected magnetic properties.
Two additional samples of our 430 and 430F steels were
annealed after purchase in order to compare to the fully annealed
ASM samples. The 430 and 430F steel samples were chosen for
annealing since they have the lowest annealing temperature of
any of the steels tested and therefore took the shortest time to
anneal. Even so, the annealing process required a full day to
complete. The steel samples were heated to 780 1C in a hydrogen
atmosphere, held at this temperature for 1 h, and then cooled to
600 1C at a rate of 36 1C per hour. They were then air cooled. Fig. 11
shows how mmax and Hc are changed by the additional annealing
heat treatment. A substantial improvement results, with approxi-
mately a factor of two increase in mmax and a factor of two
decrease in Hc for both types of steel. After this additional
annealing our 430F sample did not display as soft magnetic
properties as the 430F sample quoted in the ASM Handbook. Our
430 sample did, however, achieve similar magnetic properties as
the Handbook 430F sample (the Handbook does not provide data
for 430 steel). The main difference between these two types of
steel is that type 430F has a greater amount of sulfur, which
Fig. 10. Dependence of saturation flux density on chromium content for should little affect the magnetic properties. We conclude then that
martensitic steels and ferritic steels. Each data point represents one of the
stainless steel samples tested and the solid lines are linear fits to the data. An
additional annealing of as-received steels can provide as soft
approximate linear dependence is seen, as has been previously observed [1]. The magnetic properties as those of fully annealed steels in the
measurements were taken at room temperature. ASM Handbook. We presume that our 430F sample remained

Table 3
Comparison of the maximum permeability and coercive force for the stainless steels used in this work and in the ASM Handbook [7].

Steel mmax (this work) mmax (ASM) Hc (this work) Hc (ASM) Brinell hardness Brinell hardness
(this work) (ASM)

410 360 750 770 480 180 163

416 490 750 900 480 235 163
420 450 950 820 800 279 183
430F 530 2000 560 160 184 140
430FR 960 2600 270 130 162 153
446 450 1000 670 360 217 163

Also shown is the mechanical hardness of the steel samples used. In this work the mechanical hardness was measured and provided by the steel manufacturer. The steels
tested in this work display harder magnetic and mechanical properties than those in the ASM Handbook due to their different annealing procedures.
 ARTICLE IN PRESS

2112 P. Oxley et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 2107–2114

Fig. 12. Variation of mmax with the frequency of the applied magnetic field,
normalized to the value of mmax at 2 Hz (data points). The solid line is the
prediction of Eq. (5) and is not a fit to the data. As the field frequency is increased
eddy currents lead to a reduction in the measured permeability, in agreement with
the theoretical prediction. At a frequency of 60 Hz, Eq. (5) predicts an effective
permeability equal to 50% of the d.c. permeability.

A comparison between Eq. (5) and our measurements of the

variation of mmax with frequency for the as-received 430FR
sample is shown in Fig. 12. In our data the d.c. permeability,
m, is taken to be the permeability measured at the lowest
frequency used, i.e., 2 Hz. There is good agreement between the
Fig. 11. Effect of annealing two of the as-received samples at 780 1C and cooling theory and the experiment, both indicating a reduction in
slowly. Substantially softer magnetic properties result and the 430 sample is as permeability as the frequency of the applied field is increased.
soft as the fully annealed 430F sample listed in the ASM Handbook. The
Whether or not this reduction is acceptable will depend on the
measurements were taken at room temperature.
specific magnetic application. This data and the theory also
confirm that our measurements using a 4 Hz oscillating field
accurately represent the d.c. magnetic permeabilities of our
magnetically harder than our 430 sample due to its higher initial stainless steels since the 4 Hz data point is in good agreement
mechanical hardness and carbon content (Tables 1 and 2). with the d.c. value.
One final test was performed at room temperature to measure
mmax of the 430FR sample with applied magnetic fields oscillating 3.2. Cryogenic tests
at frequencies in the range 2–30 Hz, instead of simply at 4 Hz as
was used for all data previously mentioned. The purpose of this At 77 K the steels displayed harder magnetic properties than at
test was two-fold: firstly to probe how eddy current shielding 300 K, as shown in Figs. 5 to 8. On average over the ten steels
affects the measured magnetic properties when the steels are tested there was a decrease in mmax of 8%, an increase in Hc of 14%,
used in low-frequency a.c. applications, and secondly to confirm an increase in Br of 4%, and an increase in Bs of 3% when the steels
that 4 Hz measurements accurately represent the d.c. magnetic were cooled to 77 K. These modest changes in magnetic properties
properties. Eddy current shielding of the applied magnetic field in indicate that the stainless steels can still be used in magnetic
the interior of the steel rods leads to an effective permeability, m0 , applications at very low temperature.
which is lower than the d.c. permeability, m. The theoretical In Section 3.1, we discussed how an increase in coercive force
relationship between m and m0 is given by [20] with carbon content can be interpreted in terms of large carbide
! inclusions which impede domain wall motion. Now we seek to
m0 2 BerðfÞ  Bei0 ðfÞ  BeiðfÞ  Ber0 ðfÞ interpret how temperature affects coercive force, again based on a
¼ (5)
m f Ber 2 ðfÞ þ Bei2 ðfÞ model of domain wall motion impeded by large inclusions. The
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi coercive force due to large inclusions around which closure
where f ¼ d pm0 mf =2r, d is the rod diameter, f is the frequency domains exist is proportional to the domain wall energy, Ew, and
of the applied field, and r is the electrical resistivity of the steel. inversely proportional to the saturation magnetization, Ms [14,15]:
The Kelvin functions Ber, Bei, Ber0, and Bei0 are given by
Ew
 4n Hc / (7)
X
1
1 f Ms
n
BerðfÞ ¼ ð1Þ
n¼0 ½ð2nÞ!2 2 At a domain wall the magnetization vector must change
X
1  4nþ2 direction, and since crystal anisotropy energy depends on the
1 f
BeiðfÞ ¼ ð1Þn orientation of the magnetization vector relative to the crystal’s
½ð2n þ 1Þ!2 2
n¼0 easy axis there is an energy associated with this direction change.
X1  4n1
2n f Domain wall energy, therefore, depends on the crystal anisotropy
Ber 0 ðfÞ ¼ ð1Þn
½ð2nÞ! 2 2 through the anisotropy constant, K, which is temperature
n¼0
X1  4nþ1 dependent. The wall energy also depends on the stress in the
2n þ 1 f
Bei0 ðfÞ ¼ 2
ð1Þn (6) material and the material’s magnetostriction constant, l, but this
n¼0 ½ð2n þ 1Þ!
2
contribution is usually considered small for iron [14,15]. The
 ARTICLE IN PRESS

P. Oxley et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 2107–2114 2113

domain wall energy is, therefore, given by be inferred from measurements of the coercive force of marten-
pffiffiffiffiffiffiffiffiffiffi site/austenite two phase steels [8]. In that study, the coercive force
Ew / KðTÞ (8) of a mixture containing 36% martensite was observed to increase
The Akulov–Zener law for the variation of K(T) with tempera- by approximately 15% when the mixture was cooled from 300 to
ture is given by [27] 80 K. Our results for the change in coercive force with temperature
  are, therefore, consistent with these previous experiments. To our
Ms ðTÞ 10 knowledge our study is the first comprehensive set of experi-
KðTÞ ¼ K 0 (9)
M s0 ments investigating the low temperature magnetic properties of
for a cubic material such as iron. K0 and Ms0 are the anisotropy magnetic stainless steels.
constant and saturation magnetization at absolute zero, and Ms(T)
is the saturation magnetization at temperature T. The temperature
4.. Conclusions
dependence of Ms is given by Bloch’s T3/2 law [28]
"  3=2 #
T We have measured the magnetic properties of a range of
M s ðTÞ ¼ M s0 1  A (10) ferritic and martensitic stainless steels and studied how these
y
properties are affected by cooling to 77 K. This work provides
where y is the Curie temperature and the constant A has an information about the four most important magnetic parameters
empirical value of 0.118 for iron [20]. Strictly, Eq. (10) is only (mmax, Hc, Bs, and Br) which can be used when selecting stainless
valid at temperatures near absolute zero, however it has been steels for particular magnet applications. A typical application of
tested for iron and nickel and the T3/2 scaling was confirmed this information would be when choosing a steel to construct an
up to temperatures as high as 300 K [20]. Gathering together electromagnet in which minimizing eddy current losses is
Eqs. (7)–(10) the temperature dependence of the coercive force is important. We tested steels in the as-received annealed condition
given by and so our work would be particularly important for researchers
"  3=2 #4 who wish to use magnetic steels but who cannot, or, for
T convenience, do not wish to fully anneal the steels after purchase.
Hc ðTÞ / M 4s ðTÞ / 1  A (11)
y The modest changes in the magnetic properties we observe when
Applying Eq. (11) for iron we find the steels are cooled to 77 K indicate that they can still be used in
magnetic applications at very low temperatures, and our results
Hc ðT ¼ 77Þ provide information as to their expected low-temperature
¼ 1:07 (12)
Hc ðT ¼ 300Þ magnetic properties.
An additional temperature-dependent contribution to Hc, not
included in Eq. (11), is that from grain boundaries. This
Acknowledgements
contribution is proportional to Ms(T) and gives rise to an increase
of 2% above the value in Eq. (12) for iron [15]. Averaged over the
The authors wish to thank John Oxley for constructing the
ten steels we tested the ratio of the coercive force at 77 K to
current and field coil amplifiers, Dick Miller for machining
the coercive force at 300 K was 1.14. Given the complexity of the
expertise used in constructing the experimental set up, and Drs.
microstructure of stainless steels in comparison with iron and
Alan Fox and De-Ping Yang for useful discussions and for reading
given the inherent difficulties of interpreting macroscopic proper-
the manuscript. We thank the referee for their helpful comments
ties from microscopic effects, we consider the agreement between
on the manuscript. We also thank the Research Corporation for
the theory of Eq. (11) and experiment to be fair. Neglected in the
funding through Cottrell College Science Award CC6761.
theory of Eq. (11) is any contribution to the coercive force from
internal stresses which inhibit wall motion. These stresses are
produced by crystal dislocations and inclusions and may not be References
negligible in our steels as they are in iron. Whether or not these
[1] D.A. DeAntonio, Adv. Mater. Process. 161 (2003) 29.
will give rise to a temperature-dependent coercive force is not
[2] L. Battistini, R. Benasciutti, A. Tassi, J. Mag. Magn. Mater. 133 (1994) 603.
clear however. The magnetostriction constant l100 of iron has [3] K. Ara, N. Ebine, N. Nakajima, ASME Trans. J. Pressure Vessel Technol. 118
negligible change between room temperature and 77 K [29], and (1996) 447.
so for iron one would not expect a change in Hc with temperature [4] K. Ara, IEEE Trans. Magn. 25 (1989) 2617.
[5] H. Kwun, G.L. Burkhardt, J. Appl. Phys. 15 (1987) 1576.
due to stress. [6] S.S.M. Tavares, D. Fruchart, S. Miraglia, D. Laborie, J. Alloys Compd. 312 (2000)
Eq. (10) predicts a 2% increase in saturation magnetization for 307.
iron when cooled to 77 K. This increase is close to the average [7] Douglas W. Dietrich, Magnetically Soft Materials, ASM Handbook, vol. 2,
Properties and Selection: Nonferrous Alloys and Special-Purpose Materials,
measured increase in Bs, for the ten steels tested. The measured ASM, Materials Park, OH, 1990.
increase was 2.6 (1.3)%, where the number in parenthesis [8] J. Ding, H. Huang, P.G. McCormick, R. Street, J. Mag. Magn. Mater. 139 (1995)
represents the uncertainty in the average measurement 109.
[9] D.C. Larbalestier, H.W. King, Cryogenics 13 (1973) 160.
due to the 74% uncertainty in each of the ten individual [10] J.E. Jensen, W.A. Tuttle, R.B. Stewart, H. Brechna, A.G. Prodell (Ed.), Brookhaven
measurements of Bs. The measurements of saturation flux National Laboratory Selected Cryogenic Data Notebook, Brookhaven National
density in our steels, therefore, agree with the theoretical Laboratory Report Number, BNL-10200-R, 1980.
[11] Philip Harvey (Ed.), Engineering Properties of Steel, ASM, Materials Park, OH,
prediction for iron.
1982.
Previous measurements of the temperature dependence of the [12] Gunvant N. Maniar, Terry A. DeBold, Technical Article 00002, Carpenter
coercive force for materials more similar to stainless steel than Technology Corporation, Reading, PA, 1997.
[13] M. Kersten, Z. Physik 44 (1943) 63.
iron are not numerous. Experiments on iron–chromium alloys
[14] L.J. Dijkstra, C. Wert, Phys. Rev. 76 (1950) 979.
[23] show a change in Ms by a factor of 1.03 between room [15] John B. Goodenough, Phys. Rev. 95 (1954) 917.
temperature and 77 K for 15 at% Cr. When raised to the fourth [16] William D. Nix, Robert A. Huggins, Phys. Rev. 135 (1964) A401.
power as in Eq. (11) this indicates a 13% increase in Hc at 77 K. The [17] L. Néel, Cah. Phys. 25 (1944) 21.
[18] H.J. Williams, Phys. Rev. 71 (1947) 646.
only previous measurements of the temperature dependence of [19] J.R. Davis, Stainless Steels ASM Specialty Handbook, ASM, Materials Park, OH,
the magnetic properties of magnetic steels that we have found can 1994.
 ARTICLE IN PRESS

2114 P. Oxley et al. / Journal of Magnetism and Magnetic Materials 321 (2009) 2107–2114

[20] R.M. Bozorth, Ferromagnetism, D. Van Nostrand Company, New York, 1961. [25] Noriya Ebine, Katsuyuki Ara, IEEE Trans. Magn. 35 (1999) 3928.
[21] E. Gumlich, Wiss. Abhandl. Physik-tech. Reichsanstalt 4 (1918) 267. [26] P.D.S. Pedrosa, J.R. Teodosio, S.S.M. Tavares, J.M. Neto, M.R. da Silva, J. Alloys
[22] T.D. Yensen, Trans. Am. Inst. Elec. Eng. 43 (1924) 145. Compd. 329 (2001) L14.
[23] A.T. Aldred, Phys. Rev. B 14 (1976) 219. [27] C. Zener, Phys. Rev. 96 (1954) 1335.
[24] B.L. Bramfitt, 3rd Ed, Carbon and Alloy Steels Mechanical Engineers Hand- [28] F. Bloch, Z. Phys. 61 (1930) 206.
book, vol. 1, Wiley, Hoboken, NJ, 2006. [29] G.M. Williams, A.S. Pavlovic, J. Appl. Phys. 39 (1968) 571.

View publication stats