Preferential crystallographic alignment in polycrystalline MnP

Article abstract of DOI:10.1016/j.jmmm.2010.03.022

The existence of preferential crystallographic alignment in hot pressed and die upset manganese phosphide (MnP) was investigated using magnetic measurements and electron backscatter diffraction (EBSD). Pole figures calculated from the EBSD data show that die upsetting causes the 〈1 1 0〉 directions to align preferentially along the die upset (DU) direction with the 〈0 0 1〉 direction preferentially perpendicular to the DU direction. Magnetic measurements show that the die upsetting can reduce the saturation field relative to that of a similar sample with randomly oriented grains. Since the low-field magnetocaloric effect in single crystals of MnP has been shown to be greatest along the 〈0 1 0〉 direction and smallest along the 〈1 0 0〉 direction, this technique offers a means to achieve the advantages of single crystal alignment with the economy of using bulk processing techniques on polycrystalline material. The peak magnetic entropy change measured with the field applied along the DU direction in the DU sample was 3%, 5%, and 8% greater than the peak entropy change of a randomly oriented powder at fields of 2.0, 1.0, and 0.5 T, respectively.

Full text of DOI:10.1016/j.jmmm.2010.03.022

ARTICLE IN PRESS
Journal of Magnetism and Magnetic Materials 322 (2010) 2571–2574
Contents lists available at ScienceDirect
Journal of Magnetism and Magnetic Materials
journal homepage: www.elsevier.com/locate/jmmm
Preferential crystallographic alignment in polycrystalline MnP
Ryan A. Booth a, Melania Marinescu b, JinFang Liu b, Sara A. Majetich a,n
a Department of Physics, Carnegie Mellon University, Pittsburgh, 5000 Forbes Avenue, PA 15213, USA
b Electron Energy Corporation, Landisville, PA 17538, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The existence of preferential crystallographic alignment in hot pressed and die upset manganese
phosphide (MnP) was investigated using magnetic measurements and electron backscatter diffraction
(EBSD). Pole figures calculated from the EBSD data show that die upsetting causes the /1 1 0S
directions to align preferentially along the die upset (DU) direction with the /0 0 1S direction
preferentially perpendicular to the DU direction. Magnetic measurements show that the die upsetting
can reduce the saturation field relative to that of a similar sample with randomly oriented grains. Since
the low-field magnetocaloric effect in single crystals of MnP has been shown to be greatest along
the /0 1 0S direction and smallest along the /1 0 0S direction, this technique offers a means to achieve
the advantages of single crystal alignment with the economy of using bulk processing techniques on
polycrystalline material. The peak magnetic entropy change measured with the field applied along the
DU direction in the DU sample was 3%, 5%, and 8% greater than the peak entropy change of a randomly
oriented powder at fields of 2.0, 1.0, and 0.5 T, respectively.
Received 23 November 2009
Received in revised form
16 February 2010
Available online 7 April 2010
Keywords:
Magnetocaloric effect
Manganese phosphide
Magnetocrystalline anisotropy
Texture
& 2010 Elsevier B.V. All rights reserved.
1. Introduction
The most common figure of merit used to compare MCE
materials is the magnetic entropy change per unit mass for a
given change in magnetic field. For magnetic refrigeration to be a
viable and energy efficient technology for small scale applications
such as household air-conditioning and refrigeration, this mag-
netic field needs to be provided by permanent magnets, which are
limited to a maximum strength of about 2 T. Because these
materials are more easily saturated along certain crystallographic
orientations, the MCE for low, fixed field changes is greatest along
the easy-axis. For instance, the prototype room-temperature MCE
material, gadolinium, saturates at ꢀ0.8 T along the [0 0 0 1]
direction and ꢀ1.2 T along [10–10] [6]. Single crystal studies of
Gd5(Si,Ge)4 have also shown that the magnetocrystalline aniso-
tropy has a strong effect on the MCE [7]. Since very few MCE
materials are cubic, crystallographic alignment techniques could
be advantageous in many systems. In the past use of bulk
processing techniques to achieve preferential crystallographic
alignment combined with close attention to microstructure led to
the successful development of high energy product permanent
magnet materials like Ni–Fe–B and Sm–Co [8–10].
Single crystal studies of MnP have shown that strong
magnetocrystalline anisotropy leads to a greater MCE along
the easy axis [11,12]. MnP is an attractive system for this study
of magnetic anisotropy because its rich magnetic phase diagram
has inspired a significant body of single crystal work. The
crystal structure of MnP is orthorhombic of space group Pnma
with a¼0.5260 nm, b¼0.3174 nm, and c¼ 0.5919 nm in the
paramagnetic phase at 293 K, and a¼0.5241 nm, b¼0.3181 nm,
Since the discovery of the giant magnetocaloric effect in
Gd5Si2Ge2 [1], interest in magnetocaloric effect (MCE) materials
has grown at an extraordinary pace due to the potential energy
savings of MCE refrigeration devices. To date, most of this work
has focused on the development of new materials systems, tuning
the magnetic and structural transition temperatures, and redu-
cing hysteretic losses, but in the last few years a number of
significant studies have investigated the effect of microstructure
on the magnetic and structural phase transitions. For instance, it
was shown that the magnetocaloric effect increased by up to 80%
in heat treated Gd5Si2Ge2 relative to that of arc-melted Gd5Ge2Si2
due to the redistribution of the Si and Ge atoms [2]. Lyubina et al.
[3] showed that rapid solidification of La(Fe,Si)13 alloys by melt
spinning significantly reduced the annealing time necessary to
remove a-Fe impurities while strongly reducing hysteresis. Trung
et al. [4] showed that similar techniques could reduce hysteresis
in the MnFe(P,Ge) materials. The use of micro analysis techniques
such as transmission electron microscopy, scanning electron
microscopy, and scanning Hall probe microscopy are also helping
to elucidate the effect of microstructure on the characteristics and
dynamics of phase transitions in MCE materials [5].
n Corresponding author. Tel.: +1 412 268 3105; fax: +1 412 681 0648.
E-mail address: sara@cmu.edu (S.A. Majetich).
0304-8853/$ - see front matter & 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmmm.2010.03.022

ARTICLE IN PRESS
2572
R.A. Booth et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 2571–2574
and c¼0.5903 nm in the ferromagnetic phase at 60 K [13]. MnP
undergoes a ferromagnetic to paramagnetic phase transition at a
Curie temperature of 290 K [11]. In this paper we follow the
c4a4b convention for the lattice parameters. We examined
the effect of hot pressing and die upsetting MnP. The texture of
the processed samples was analyzed using electron backscatter
diffraction (EBSD) to measure pole figures. The magnetocrystal-
line anisotropy was determined from magnetization curves, and
the magnetic entropy change was calculated from isothermal
magnetization curves using the Maxwell relation.
2. Experimental
MnP was prepared in a solid–vapor reaction by mixing Mn
(99.999%, Aldrich) and P (99%, Aldrich) in a 3:1 Mn:P molar ratio,
purging several times with high-purity argon, sealing the powder
under vacuum of 10ꢁ2 torr in a borosilicate glass ampoule, and
heating at 450 1C for 8 h. Phosphorus was added and this process
was repeated with a 3:2 molar ratio, and finally a 3:3.15 molar
ratio then heated for 3 days. The excess phosphorus was used to
assure the correct final stoichiometry. This procedure circumvents
the need for high temperatures and minimizes the amount of
initial phosphorus present at any given time to decrease the fire
and explosion hazards. The phase purity and lack of preferential
crystallographic orientation was confirmed by X-ray diffraction.
The preparation of the MnP single crystals used in this study was
described previously [11].
The powder was further processed by hot pressing 3 g of
powder under a vacuum of 10ꢁ6 torr. The pressing force was
continuously increased to 2 tons and the temperature increased
from 690 1C at the start of pressing to 740 1C at the end of
pressing. A 0.64 g piece of the hot pressed sample was die upset
with the pressure being continuously increased to 1 ton and the
temperature being increased from 705 1C at the start of pressing
to 785 1C at the end of pressing. In the die upset process a sample
is deformed under pressure in an oversized die so that there is
room for it to expand outward. This is different from hot pressing
in which a powder is compressed to fill a die completely. The die
upset deformation was calculated as the ratio of the ending height
to the starting height and was found to be 70%. We define the die
upset (DU) and hot pressed (HP) directions as the direction
pressure applied during die upsetting or hot pressing processes.
A diamond saw was used to cut a 1 ꢂ 1 ꢂ 1 mm3 piece of both
the die upset and hot pressed samples. Hysteresis loops of this
sample were measured at 250 K with the magnetic field applied
parallel and perpendicular to the DU and HP directions. The
isothermal magnetization curves used to calculate the magnetic
entropy change were measured with the field applied parallel to
the DU direction. All magnetic measurements were performed
using a Quantum Design SQUID magnetometer.
Fig. 1. A Kikuchi pattern from the polished die upset MnP sample with the Miller
indices shown.
study mathematical modeling was employed to deconvolute the
overlapping peaks. EBSD is a much simpler alternative.
To create an EBSD pattern, the SEM electrons enter the sample
and undergo multiple scattering processes as they typically do
during the formation of a normal SEM image. As they exit,
electrons diffract and form a Kikuchi pattern, which is recorded by
the EBSD camera [16]. In typical X-ray and electron diffraction
experiments with a single crystal or grain, the initial angle of the
radiation interacting with the sample is defined and as a result
diffraction spots are produced. In EBSD, the initial angle of a
scattering process varies widely due to the multiple scattering
events occurring within the sample and a Kikuchi pattern is
formed instead of diffraction spots. The Kikuchi pattern contains
information about the local crystallographic orientation, the local
chemical phase, and local strain in a region of interest. Tango
software (HLK) was used to index and analyze these patterns, an
example of which is shown in Fig. 1.
Over a 240 ꢂ 22
m
m2 area 7277 sites of 0.5 ꢂ 0.5 m2 in size
m
were sampled and indexed to obtain the local crystallographic
orientation. The sampled area on the polished face of the die upset
MnP is very close to the center of the original die upset piece. Pole
figures were created from this information to analyze the texture
of the sample.
3. Results and discussion
The percent difference in the magnetization when the
magnetic field was applied parallel and perpendicular to the
single crystal b-axis, DU, and HP directions is plotted in Fig. 2.
From single crystal measurements, the b-axis is known to be the
easy axis, the a-axis intermediate, and the c-axis the hard axis in
MnP [17]. Therefore, the magnetic measurements do indicate that
b-axis is preferentially oriented along the DU direction of the die
upset sample, and to a much lesser degree the HP direction of the
hot pressed sample.
A second 2 ꢂ 2 ꢂ 10 mm3 sample was cut from the DU sample
and the face perpendicular to the DU direction was mechanically
polished. EBSD patterns were obtained from the polished face
using an Oxford Instruments EBSD camera installed on a Phillips
XL-30 scanning electron microscope (SEM). EBSD is a micro-
analysis technique used to create texture or phase maps of
material surfaces. Alternatively, texture information could be
obtained by analysis of peak height ratios in an X-ray diffraction
2
y
scan [14] or from rocking curves or pole figures. Due to the
Since single crystal measurements shown in Fig. 2 were made
by first saturating the sample, then decreasing the field, the
textured samples were measured in the same way. Also, the HP
and DU samples have some degree of strain introduced by the hot
pressing and die upsetting processes as indicated by an increase
in the peak widths in the X-ray diffraction patterns compared to
the peak widths of the powder pattern. Analysis of the peak
existence of many overlapping peaks and lack of isolated a, b, or
c-axis diffracting planes in the X-ray diffraction pattern, these
more standard X-ray techniques could not be used on our sample.
Orthorhombic materials are not typically studied for their texture,
and previous neutron diffraction studies of orthorhombic
uranium encountered similar difficulties [15]. In the uranium

ARTICLE IN PRESS
R.A. Booth et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 2571–2574
2573
Fig. 2. Percent difference in magnetization between measurements with the field
applied parallel and perpendicular to the b-axis, DU, and HP directions for the
single crystal, die upset sample, and hot pressed samples, respectively.
Fig. 3. Magnetic entropy change of the DU sample with the field applied parallel
to the DU axis compared to the magnetic entropy change of a randomly oriented
powder.
widths using the Scherrer relation gives average grain sizes of
55 nm in MnP powder and 20 nm in both the HP and DU samples.
This suggests that strain is the source of hysteresis in the
processed samples since it is not likely that the grains shrink
during the deformation process. The measured hysteresis at 250 K
was 440 Oe in the hot pressed sample and 300 Oe in the die upset
sample. No measureable hysteresis was observed in the single
crystals; thus the anomalously high difference in magnetization in
the hot pressed sample and the negative difference in the die
upset sample at zero field can be attributed to the history of the
sample and a resulting non-zero remnant magnetization.
The difference in the magnetization between a field applied
parallel to and perpendicular to the DU directions is 7.5% and 2.5%
at 1.0 and 2.0 T, respectively. The difference in magnetization
between
a field applied parallel to the /0 1 0S direction
(or parallel to the b-axis) and perpendicular to a combination of
the /1 0 0S and /0 0 1S directions (or perpendicular to the
c- and a-axes, respectively) was 24.9% and 10.2% at 1.0 and 2.0 T,
respectively. The peak magnetic entropy change of the DU sample
with the field applied parallel to the DU axis is about 3%, 5%, and
8% greater than the peak entropy change of a randomly oriented
powder sample at fields of 2.0, 1.0, and 0.5 T, respectively as
shown in Fig. 3.
The pole figures shown in Fig. 4 represent the distribution of
the crystallographic directions /1 0 0S, /0 1 0S, /0 0 1S, and
/1 1 0S throughout the die upset sample. A pole figure is a
stereographic projection of the distribution of
a certain
Fig. 4. Pole figures of the /1 0 0S or the c-axis, /0 1 0S or the b-axis, /1 0 0S or
the a-axis, and /1 1 0S directions in MnP calculated from an EBSD map. The three
axes: vertical, horizontal, and out of the plane are the same for each pole figure
and refer to the sample coordinates. The vertical axis, DU, corresponds to the die
upset direction. The axis pointing out of the page corresponds to the axis normal to
the polished face of the sample. The lowest intensity is blue and the highest
intensity is red. (For interpretation of the references to color in this figure the
reader is referred to the web version of this article.)
crystallographic direction in the real space of the sample [18].
For instance, a pole figure for the /1 0 0S family of directions in a
cubic single crystal would have high intensity at the center, top,
bottom, left, and right of the pole figure indicating the four-fold
symmetry. In the die upset MnP sample, the /1 1 0S pole figure
indicates that the /1 1 0S direction is strongly oriented along the
die upset direction. While it is not immediately obvious why the
/1 1 0S direction is the preferred packing direction for the die
upset process, the texture is consistent with a deformation that
only breaks symmetry in one direction. Texture that only breaks
symmetry in one direction is referred to as fiber texture. The
preference for the /0 0 1S direction to come out of the plane of
the sample as indicated by the strong intensity in the center and
horizontal band of the /0 0 1S pole figure is consistent with this
fiber texture as the /0 0 1S direction is perpendicular to the
/1 1 0S direction. The preference for the /1 0 0S and /0 1 0S
directions to lie in the plane of the sample, as indicated by the
concentration of intensity around the periphery of their
respective pole figures, is also consistent with this fiber texture
since /1 1 0S, /1 0 0S, and /0 1 0S are mutually perpendicular
to the /0 0 1S direction, which is preferentially out of the plane.
The angular distribution of intensity along the DU direction in the
/1 0 0S and /0 1 0S pole figures is also consistent with the fiber
texture. Since the angle between the /1 1 0S and /1 0 0S
directions is 331 and the angle between the /1 1 0S and
/0 1 0S directions is 571, it is expected that the /0 1 0S
direction is more broadly distributed about the DU direction
than the /1 0 0S.
The preference for die upsetting to select the /1 1 0S direction
in MnP is due to subtle physics of the structure and bonding

ARTICLE IN PRESS
2574
R.A. Booth et al. / Journal of Magnetism and Magnetic Materials 322 (2010) 2571–2574
within the MnP crystal. These results show that it is not possible
to give any stronger preference to the magnetically easy b-axis
using the die upset technique under the temperature and pressure
conditions described here. However, it may be possible to use a
processing technique that breaks symmetry in two directions to
strengthen the b-axis texture. One possible technique would be to
die upset the sample in one direction, rotate 901 and die upset
again. While cold rolling was shown not to be effective on MnP
due to its brittleness, hot rolling could be. Because rolling selects
two directions in a sample, the rolling direction and transverse
direction, it is possible that rolling could also lead to stronger
b-axis texture.
physics of the magnetocaloric effect should depend more strongly
on the texture of the sample.
Acknowledgements
This work was supported by Department of Energy under
Grant no. DE-FG02-08ER46481.
References
[1] V.K. Pecharsky, K.A. Gschneidner, Phys. Rev. Lett. 78 (1997) 4494.
[2] A.O. Pecharsky, K.A. Gschneidner Jr., V.K. Pecharsky, J. Appl. Phys. 93 (2003)
4722.
[3] J. Lyubina, O. Gutfleisch, M.D. Kuz’min, M. Richter, J. Magn. Magn. Mater. 320
(2008) 2252.
[4] N.T. Trung, Z.Q. Ou, T.J. Gortenmulder, O. Tegus, K.H.J. Buschow, E. Bru¨ck,
Appl. Phys. Lett. 94 (2009) 102513.
[5] J.D. Moore, K. Morrison, G.K. Perkins, D.L. Schlagel, T.A. Lograsso, K.A.
Gschneidner Jr., V.K. Pecharsky, L.F. Cohen, Adv. Mater. 21 (2009) 3780.
[6] S. Yu., A.M. Dan’kov, V.K. Tishin, Pecharsky, K.A. Gschneidner Jr., Phys. Rev. B
57 (1998) 3478.
Although the EBSD and magnetic measurements qualitatively
agree, it should be noted that the area sampled was in the middle
of die upset sample and may not be representative of the texture
throughout the sample as has been suggested in scanning X-ray
diffraction experiments with Nd–Fe–B [10]. While EBSD only
probes the surface of a material, probing regions closer to
the edges could provide more information about distribution of
the texture.
[7] M. Nazih, A. de Visser, L. Zhang, O. Tegus, E. Bru¨ck, Solid State Commun. 126
(2003) 255.
[8] A.M. Gabay, M. Marinescu, J.F. Liu, G.C. Hadjipanayis, J. Magn. Magn. Mater.
321 (2009).
4. Conclusions
[9] W. Gru¨nberger, D. Hinz, A. Kirchner, K.H. Mu¨ller, L. Schultz, J. Alloys Compd.
257 (1997) 3318.
[10] L.H. Lewis, T.R. Thurston, V. Panchanathan, U. Wildgruber, D.O. Welch, J. Appl.
Phys. 82 (1997) 3430.
[11] R.A. Booth, S.A. Majetich, J. Appl. Phys. 105 (2009) 07A926.
[12] M.S. Reis, R.M. Rubinger, N.A. Sobolev, M.A. Valente, K. Yamada, K. Sato, Y.
Todate, A. Bouravleuv, P.J. Von Ranke, S. Gama, Phys. Rev. B 77 (2008)
104439.
[13] H. Fjellvag, A. Kjekshus, Acta Chem. Scand. A38 (1984) 563.
[14] D. Givord, A. Lie´nard, R. Perrier de la Baˆthie, P. Tenaud et, T. Viadieu, J. Phys.
Colloq. 46 (1985) C6-313.
In conclusion, we have demonstrated that preferential align-
ment in polycrystalline samples of MnP can be achieved through
die upset deformation and that the MCE of the textured sample is
slightly greater than a randomly oriented sample. As most of the
proposed MCE materials have some crystallographic and mag-
netic anisotropies, bulk processing techniques such as die
upsetting should be investigated further as a means of achieving
the advantage of single crystal properties with the economy of
polycrystalline processing techniques. These effects could be
particularly interesting in first-order transition materials as the
coupling between the structural and magnetic transitions is much
stronger than in second-order materials like MnP, and thus the
[15] Y.C. Kim, H.J. Kim, C.H. Lee, B.S. Seong, S. Kim, Met. Mater. 4 (1998) 1101.
[16] F.J. Humphreys, J. Mater. Sci. 36 (2001) 3833.
[17] Y. Shapira, N.F. Oliveira Jr., C.C. Becerra, S. Foner, Phys. Rev. B 29 (1984) 361.
[18] J.S. Kallend, U.F. Kocks, A.D. Rollett, H.-R. Wenk, Mater. Sci. Eng.
132 (1991) 1.
A