Ferromagnetism above room temperature in bulk sintered gallium phosphide doped with manganese

Article abstract of DOI:10.1016/j.jpcs.2004.10.005

Evidence for ferromagnetism in bulk sintered gallium phosphide (GaP) doped with 3% manganese, having a Curie temperature of 600 K considerably higher than previous observations, is obtained using ferromagnetic resonance (FMR) and AC magnetization measurements. The field position and line width of the resonance showed a strong temperature dependence characteristic of FMR spectra. A non-resonant derivative signal centered at zero field was also observed starting at 600 K further confirming high temperature ferromagnetism. AC magnetization measurements also show the existence of ferromagnetism at high temperature. Published by Elseiver Ltd.

Full text of DOI:10.1016/j.jpcs.2004.10.005

Journal of Physics and Chemistry of Solids 66 (2005) 793–796
www.elsevier.com/locate/jpcs
Ferromagnetism above room temperature in bulk sintered gallium
phosphide doped with manganese
Frank J. Owensa,b,
*
aArmament Research Development and Engineering Ctr., Picatinny, NJ, USA
bDepartment Physics, Hunter College, CUNY, 956 Park Ave., New York, NY 10021, USA
Received 11 June 2004; revised 2 September 2004; accepted 4 October 2004
Abstract
Evidence for ferromagnetism in bulk sintered gallium phosphide (GaP) doped with 3% manganese, having a Curie temperature of 600 K
considerably higher than previous observations, is obtained using ferromagnetic resonance (FMR) and AC magnetization measurements. The
field position and line width of the resonance showed a strong temperature dependence characteristic of FMR spectra. A non-resonant
derivative signal centered at zero field was also observed starting at 600 K further confirming high temperature ferromagnetism. AC
magnetization measurements also show the existence of ferromagnetism at high temperature.
Published by Elsevier Ltd.
Keywords: A. Semiconductors
1. Introduction
ZnO [5]. The origin offerromagnetism in these compounds is
not totally understood. It has been proposed that the exchange
interaction between the dopant spins is mediated by the holes
or electrons [6]. In the ferromagnetic state there is a splitting
of the valence and conduction band depending on the spin
orientation of the charge carriers. The model predicts that
hole doped semiconductors will have higher Curie tempera-
tures than electron doped materials.
Semiconducting switching elements in logic circuits
involve transport of holes or electrons induced at specific
voltages. On the other hand storage of information employs a
separate magnetic material in which magnetic field align-
ment of micrometer long and nanometer wide magnetic
particles is the basis of the storage mechanism. In recent
years there has been considerable interest in combining both
the storage and transport into one material by developing
magnetic semiconductors consisting of alloys of a magnetic
ion such as Mn2C and a semiconducting material. In such
materials the magnetic field induced orientation of the spin of
the charge carriers would be the basis of information storage.
Ferromagnetism at 110 K has been observed in GaAs doped
with manganese [1]. More recently, there have been reports
of ferromagnetism above room temperature in GaMnN,
GaMnP made by molecular beam epitaxy which will make
large scale production expensive [2–4]. The highest Curie
temperature reported to date is above 400 K in Mn doped
Here we report ferromagnetic resonance (FMR) and AC
magnetization evidence for ferromagnetism in GaP hole
doped with 3% Mn and made by a simple inexpensive
sintering process.The Curie temperature of the material is
600 K much higher than any previous observation. FMR is a
highly sensitive and well established method of studying and
verifying the existence of ferromagnetism in a material [7].
2. Experimental
The ferromagnetic resonance measurements were made
using a Varian E-9 spectrometer operating at 9.2 GHz with
100 KHz modulation. The temperature of the sample was
controlled by flowing heated or cold nitrogen gas through a
doubled-walled quartz tube, which is part of an ADP Heli-
Tran system, and which is inserted through the center of
* Department of Physics, Hunter College, CUNY, 956 Park Ave., New
York, NY 10021, USA.
E-mail address: fowens@pica.army.mil.
0022-3697/$ - see front matter Published by Elsevier Ltd.
doi:10.1016/j.jpcs.2004.10.005

794
F.J. Owens / Journal of Physics and Chemistry of Solids 66 (2005) 793–796
the microwave cavity. The sample was contained in the
quartz tube and located at the center of the cavity. The
magnetization was obtained by measuring the magnetic
field dependence of the AC susceptibility at 350 KHz using
a method similar to that described by Clover and Wolf [8].
The system consists of an HP 204C LC oscillator modified
to have an external coil. The sample is contained in the coil,
which is in a cryogenic dewar between the poles of a
magnetic. The change in the frequency of the oscillator,
which is proportional to the change in susceptibility, is
measured using a HP 5314 frequency counter. The
magnitude of the saturation magnetization is obtained by
comparing the measurements with frequency shifts of a
sample of known magnetization such as iron.
While Mn is not ferromagnetic in the bulk there is some
evidence that Mn nanoclusters can be ferromagnetic
although there is no evidence they can be ferromagnetic
above room temperature [9]. The ferromagnetism discussed
below is likely not from ferromagnetism in the clusters but
rather from a dilute distribution of Mn in the sample.
3. Results and analysis
Fig. 2 shows the Raman spectra of the transverse optical
(TO) mode and longitudinal optical (LO) mode in doped and
undoped GaP recorded using a JY Horiba confocal Raman
spectrometer. No extraneous lines indicative of the presence
of impurities were observed. The higher frequency LO mode
is down shifted by 4 cmK1 in the manganese doped sample. It
has been shown in other semiconductors such as GaN that the
LO mode is coupled to the plasma mode whose frequency is
proportional to the electron carrier concentration [10]. The
LO mode has been shown to shift to higher frequencies with
increasing electron carrier concentration. The observed
decrease in the frequency of the LO mode in the Mn doped
GaP indicates a decrease in the electron carrier concentration
consistent with hole doping. To further verify this interpret-
ation gallium phosphide was doped with V5C by a similar
sintering process described above using V2O5. In this
material the LO mode increased by 3 cmK1 consistent with
electron doping of the sample verifying that the direction of
the shift of the LO mode is a viable way to determine whether
GaP is electron or hole doped.
The GaP used was examined by electron paramagnetic
resonance (EPR) prior to processing to insure no magnetic
impurities were present in the material. No evidence for any
magnetic impurities was found. EPR is sensitive to
magnetic species to one part per 10 billion. The samples
were synthesized by thoroughly mixing in the ratio 0.03
molecular weight MnO2 to one molecular weight gallium
phosphide having 99.999 purity and then grinding the
mixture using a mortar and pestle. The samples in the form
of pressed pellets contained in an alumina boat were
sintered at 500 8C in an oven for 4 h in air followed by rapid
quenching to room temperature. The sintered samples were
examined by X-ray diffraction employing a Scintag X-ray
instrument using the Cu Ka line. Fig. 1 shows the powder
X-ray diffraction spectra. The lines at the top of the figure
are those expected for pure gallium phosphide. The peaks in
the doped sample occur at the same scattering angles as pure
GaP and no impurity lines are evident in the data. The
absence of impurity lines is particularly important indicat-
ing that the GaMn alloy, which is ferromagnetic at room
temperature, is not present or is MnP, which has a Curie
temperature of 291 K. Because the XRD lines occur at the
same values in the doped and undoped GaP it is suggested
that the manganese is not forming clusters in the lattice.
Clusters in the lattice would likely result in an expansion of
the lattice parameters and a change in the scattering angles
in the XRD data. This conclusion is further supported by
scanning electron microscope measurements of the samples,
which show no evidence for Mn clusters in the sample.
Fig. 3(a) and (b) show the ferromagnetic resonance
spectra at 300 and 104 K, respectively, showing a marked
broadening and shift of the of the magnetic field position of
the line to lower field as the temperature is lowered
characteristic of ferromagnetic resonance spectra. The
smaller high field resonance peak at approximately
3100 G, which is temperature independent, is probably
due to some unreacted MnO2 or Mn2C in non-ferromagnetic
regions of the sample. Fig. 4(a) and (b) are plots of the
temperature dependence of the magnetic field position and
line width of the FMR signal above room temperature.
Fig. 1. Powder X-ray diffraction of GaP:Mn. The lines at top show X-ray
diffraction pattern for pure GaP.
Fig. 2. Raman spectra of the TO and LO mode of doped (a) and undoped (b)
GaP showing a downward shift of the LO mode in the doped material.

F.J. Owens / Journal of Physics and Chemistry of Solids 66 (2005) 793–796
795
Fig. 3. Ferromagnetic resonance spectra in GaP:Mn at (a) 300 K and
(b) 104 K.
These data show clearly that ferromagnetism persists well
above room temperature. Such temperature dependent
changes in the line width and field position of the resonances
would not occur in the paramagnetic state. The strong
temperature dependent effects occur in the ferromagnetic
state because the field position and line width depend on
the magnetization and the anisotropy constant, which are
temperature dependent in a ferromagnet. The specific
functional form depends on the shape of the sample
Fig. 5. (a) Low field non-resonant derivative signal recorded at 300 K and
(b) temperature dependence of the intensity of the signal showing its
emergence at 600 K which identifies the Curie temperature of the sample.
and the orientation of the dc magnetic field. The functional
form for a powder, which has a distribution of grain shapes,
sizes and orientations, has not been determined. At the Curie
temperature the FMR line becomes an EPR line of Mn2C
having a g value near the free electron value of 2.0023
corresponding to a field of 3100 G in this experiment.
Extrapolating the temperature dependence of the field
position in Fig. 4(a) to 3100 G yields a Tc greater than 600 K
Another strong evidence for the existence of ferromag-
netism above room temperature is the presence of a non-
resonant derivative signal centered at zero field which is
shown in Fig. 5(a) at 300 K. This signal is a well established
Fig. 4. (a) Temperature dependence of the field position of the resonance
above room temperature and (b) the temperature dependence of the line
width.
Fig. 6. Magnetic field dependence of the magnetization measured at 300 K.
The insert shows a measurement of the temperature dependence of the
magnetization at 3.5 kG above room temperature.

796
F.J. Owens / Journal of Physics and Chemistry of Solids 66 (2005) 793–796
indication of the presence of ferromagnetism [11,12]. The
signal occurs because the permeability in the ferromagnetic
state depends on the applied magnetic field increasing at low
fields to a maximum and then decreasing. Since the surface
resistance depends on the square root of the permeability,
the microwave absorption depends non-linearly on the
strength of the dc magnetic field resulting in a non-resonant
derivative signal centered at zero fields. This signal is not
present in the paramagnetic state and emerges as the
temperature is lowered to below Tc. Fig. 5(b) is a plot of the
temperature dependence of the signal intensity showing an
onset of growth at 600 K which corresponds to Curie
temperature of the sample.
The onset of the low field non-resonant derivative signal
indicates a Curie temperature of 600 K considerably higher
than previous observations in materials made by molecular
beam epitaxy. The ferromagnetism is believed to arise from
a dilute distribution of Mn2C in the sample and not from
clusters of Mn2C or other impurities.
Acknowledgements
The author is grateful to Mr J Oroz and Ms S. Kerwien of
this laboratory for XRD and SEM measurements,
respectively.
Fig. 6 shows a plot of the magnetization of the sample
versus applied magnetic field at room temperature providing
further support for ferromagnetism in the sample. There is a
small hysteresis not shown in the figure indicating the
material is a relatively soft ferromagnetic at room
temperature. The insert in the figure is a measurement of
the temperature dependence of the magnetization at 3.5 kG
above room temperature. The independence of the magne-
tization on temperature up to 360 K is consistent with the
high Curie temperature. The small hysteresis rules out
the possibility that the ferromagnetism arises from the
MnxGa1Kx alloy which has a large hysteresis at room
temperature having a coercive field of 3.15 kOe [13].
Also an MnP impurity cannot account for the results
because it has a Tc of 291 K well below the Tc of 600 K
observed here [14].
References
[1] H. Ohno, Science 281 (1998) 951.
[2] M.L. Reed, et al., Appl. Phys. Lett. 79 (2001) 3473.
[3] G.T. Thaler, et al., Appl. Phys. Lett. 80 (2002) 3964.
[4] N. Theodoropoulou, et al., Phys. Rev Lett. 89 (2002) 107203.
[5] P. Sharma, et al., Nature Mater. 2 (2003) 673.
[6] T. Dietl, H. Ohno, F. Matsukura, J. Cubert, D. Ferrand, Science 287
(2000) 1019.
[7] S.V. Vonsovkii, in: S.V. Vonsovkii (Ed.), Ferromagnetic Resonance,
Pergamon Press, New York, 1966, pp. 188–208.
[8] R.P. Clover, W.P. Wolf, Rev. Sci. Instr. 41 (1970) 617.
[9] M. Knickelbein, Phys. Rev. Lett. 86 (2001) 5255.
[10] T. Kozawa, et al., Appl. Phys. Lett 75 (1994) 1098.
[11] M.D. Sastry, et al., Physica C170 (1990) 41.
[12] F.J. Owens, Physica C353 (2001) 265.
In summary both ferromagnetic resonance and AC
magnetization measurements show the existence of ferro-
magnetism in gallium phosphide hole doped with
manganese and synthesized by a bulk sintering process.
[13] M. Tanka, et al., Appl. Phys. Lett. 62 (1993) 1565.
[14] Y. Shapira, N.F. Oliveria, C.C. Becerra, S. Foner, Phys. Rev. B29
(1984) 361.