MOCVD growth, structure and magnetic properties of Fe films grown on GaAs (001) substrates

Article abstract of DOI:10.1016/j.ssc.2006.07.031

Thin iron films have been grown on (001) GaAs substrates by low pressure metal organic chemical vapor deposition (LP-MOCVD) at different temperatures with the pressure of 150 Torr. X-ray diffraction (XRD) analysis showed that all films have only one strong diffraction peak (110). The surface of Fe film became smooth with increasing the growth temperature. Magnetization measurements showed that the Fe films grown at different temperatures were ferromagnetic with easy axis parallel to the film surface and hard axis perpendicular to the substrates. The field dependence of magnetization along two axes showed a remarkable difference, implying that the samples have strong magnetic anisotropy. Furthermore, when the applied magnetic field is perpendicular to the Fe surface, a sharp jump in the hysteresis loop could be observed, followed by a broad shoulder, which is related to the interface effect, the existence of carbon and the formation of 180^{{ring operator}}/90^{{ring operator}} magnetic domains.

Full text of DOI:10.1016/j.ssc.2006.07.031

Solid State Communications 140 (2006) 33–37
www.elsevier.com/locate/ssc
MOCVD growth, structure and magnetic properties of Fe films grown on
GaAs (001) substrates
a,b,∗
a
a
a,b
a
a
Kewei Liu
, Dezhen Shen , Jiying Zhang , Xiaojie Wu , Binghui Li , Bingsheng Li ,
a
a
Youming Lu , Xiwu Fan
a
Key Laboratory of Excited State Processes, Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, Changchun 130033, China
b
Graduate School of the Chinese Academy of Sciences, Beijing 100049, China
Received 23 November 2005; received in revised form 22 April 2006; accepted 23 July 2006 by B.-F. Zhu
Available online 10 August 2006
Abstract
Thin iron films have been grown on (001) GaAs substrates by low pressure metal organic chemical vapor deposition (LP-MOCVD) at different
temperatures with the pressure of 150 Torr. X-ray diffraction (XRD) analysis showed that all films have only one strong diffraction peak (110). The
surface of Fe film became smooth with increasing the growth temperature. Magnetization measurements showed that the Fe films grown at different
temperatures were ferromagnetic with easy axis parallel to the film surface and hard axis perpendicular to the substrates. The field dependence
of magnetization along two axes showed a remarkable difference, implying that the samples have strong magnetic anisotropy. Furthermore, when
the applied magnetic field is perpendicular to the Fe surface, a sharp jump in the hysteresis loop could be observed, followed by a broad shoulder,
◦
◦
which is related to the interface effect, the existence of carbon and the formation of 180 /90 magnetic domains.
c 2006 Elsevier Ltd. All rights reserved.
ꢀ
PACS: 75.50.Bb; 75.70.Ak; 75.30.Gw; 68.55.-a
Keywords: A. Iron films; D. Magnetic properties; D. Interface effect; D. MOCVD
1
. Introduction
60 ML or 9 nm, exhibit a remarkable and unexpected in-
plane uniaxial magnetic anisotropy (UMA) with an [110] easy
axis [10,14,21–23]. This is dramatically different from the
cubic magnetic anisotropy of bulk bcc Fe with h100i easy
axes. The transformation from superparamagnetism (SP) to
ferromagnetism (FM) is observed for ultrathin Fe films [18].
Up to now, most of the Fe/GaAs structures were deposited by
molecular beam epitaxy (MBE) [17,20,23–26]. However, the
growth of an Fe epitaxial layer on a GaAs substrate with a high-
quality interface using MOCVD has been also reported [27].
The hysteresis loop along the [100] and [110] axes has been
shown at low temperature (6 K) for the sample thickness below
Ferromagnetic thin films on semiconductor substrates have
been of great interest in recent years because of the possible
future development of spintronics, e.g., spin injection through
the ferromagnetic thin-film semiconductor interfaces [1–5].
The Fe/GaAs system has attracted much attention in this area
of study because of its room temperature ferromagnetism and
small lattice mismatch (1.3%) [6]. Based on the small lattice
mismatch, the stable bcc phase of Fe grew epitaxially on GaAs
(
001), GaAs (110), GaAs (113)A, and GaAs (133)A substrates,
and numerous studies optimizing its growth conditions and
analyzing its interface structure have been reported [7–16].
Much attention has been concentrated on the magnetic behavior
of Fe films with the thickness below 100 nm. The thin
films of Fe/GaAs (001), with the thickness less than about
60 nm. It can be observed that the easy axis was not clear for the
thinner films (<20 nm). By comparison with that prepared by
MBE, it seemed that the different growth methods could lead to
different results in magnetic properties of the films.
In this paper, high quality Fe films have been prepared by
using LP-MOCVD at different temperatures. The surface mor-
phology of the Fe films was investigated with an atomic force
microscope (AFM). The magnetic properties were measured
∗
Corresponding author at: Key Laboratory of Excited State Processes,
Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy
of Sciences, Changchun 130033, China. Tel.: +86 431 6176322.
E-mail address: liukewei23@sohu.com (K. Liu).
0
038-1098/$ - see front matter ꢀc 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ssc.2006.07.031

3
4
K. Liu et al. / Solid State Communications 140 (2006) 33–37
Fig. 2. Depth profile of Fe films measured by Auger electron spectroscopy. The
inset is energy dispersive spectroscopy of the samples.
Fig. 1. XRD spectra of Fe films with different growth temperatures: (a) 593 K;
b) 573 K; (c) 553 K.
(
substrates with the Ts ranging from 553 to 593 K. It is well
known that the lattice constant of Fe is slightly bigger than
half of the lattice constant of GaAs and the lattice mismatch
between Fe and GaAs is 1.3%. XRD measurement with Cu Kα
radiation (λ = 0.154178 nm) was used to make θ–2θ scans
to determine (110) d-spacing lattice constant for the Fe films,
just as shown in Fig. 1. In this figure, the diffraction peak
by vibrating sample magnetometer (VSM). A clear magnetic
anisotropy of the Fe films was observed at room temperature.
2
. Experiments
Growth was carried out in a LP-MOCVD system with
a horizontal rectangular quartz reactor at low pressure.
The ironpentacarbonyl (Fe(CO)5) carried by ultrahigh purity
hydrogen was used as material source. The source was kept
at 283 K during deposition. The GaAs substrates were semi-
insulating with (001) orientation. The substrates were cleaned
in trichloroethylene, acetone and ethanol for 5 min in an
ultrasonic bath, and etched in a sulfuric peroxide solution
◦
at 2θ ≈ 44.7 can be associated with the (110) plane of
the bcc Fe phase, and no other diffraction peak was detected.
Three samples have similar results except that the peak width
and intensity are different. Furthermore, the intensity of Fe
(110) diffraction peak is increasing with the increase of the
growth temperature. The full width at half-maximum (FWHM)
(
3H2SO4 + 1H2O2 + 1H2O) for 5 min before loaded into the
◦
◦
of the (110) diffraction peaks are 0.21 for sample a, 0.3
reactor. The substrates were placed at the center of a graphite
boat, and then inserted in the reactor. The substrates were
heated to 873 K for 10 min in H2 gas flow. The deposition
was carried out at the pressure of 150 Torr with different
temperatures (553 K, 573 K and 593 K). The flow rate of
Fe(CO)5 was kept at 8 sccm/min.
◦
for sample b and 0.24 for sample c. From Fig. 1, it can be
calculated that the (110) d-spacing lattice constant is 0.20242±
0
.00004 nm, which is close to the bulk α-Fe of 0.20265 nm.
Fig. 2 shows the depth profiles of Fe film measured by Auger
electron spectroscopy (AES). The sample was prepared at 593
K with the thickness of about 550 nm. As shown in Fig. 2,
the profile pattern consists of two regions: region (a) and
region (b). They correspond to the regions from the surface
to the depth of 510 nm and from the depth of 510 nm to
the substrate, respectively. In region (a) a signal from Fe is
detected and the intensity of the signal is constant. In region
The structure and composition of the films were studied by
XRD. As for the magnetic property of the samples, the field
dependences of magnetization (M–H curves) were measured
by the VSM system at room temperature.
3
. Results and discussion
(b) the signal of Fe decreases toward the interface of the
In order to study the effect of substrate temperature (Ts)
on crystallinity, the iron films were grown on (001) GaAs
Fe/GaAs substrate. The decrease of the signal intensity of Fe
is caused by the diffusion of Ga or As atoms. The composition

K. Liu et al. / Solid State Communications 140 (2006) 33–37
35
differences in roughness of three samples. From Fig. 3, it
can be observed that the root mean square (rms) surface
roughness (σ) of samples is 5.102 nm (a: 593 K), 8.100 nm
(b: 573 K) and 11.259 nm (c: 553 K), respectively. The surface
of Fe film is smooth (rough) when the growth temperature
is high (low). The magnetic properties of Fe films grown at
different temperatures were studied with a vibrating sample
magnetometer system. The sample’s size is approximately
2
0
.3 × 0.3 cm . Fig. 4 shows the results of the magnetic
moment measurement when the applied magnetic field is
parallel to or perpendicular to the surface of Fe films at room
temperature. When the applied field is parallel to the surface,
the magnetization curves along these directions exhibit a quasi-
rectangular shape for all the samples, just as shown in Fig. 4.
However, when the applied magnetic field is perpendicular
to the Fe films, a sharp jump in the hysteresis loop could
be observed, followed by a broad shoulder. This process can
be divided into two distinct dynamic regimes, as shown in
the figure: the low dynamic regime, extending to magnetic
field <100Oe, and the high dynamic regime when magnetic
field >100Oe. According to Raquet et al., [19] magnetization
reversal in the low dynamic regime is predominantly by
domain-wall motion, and in the high dynamic regime is
predominantly by domain nucleation. A number of reasons
could be responsible for this trend: these include surface partial
oxidation because of the exposure to the air, interface effect,
surface roughness, the existence of carbon or the formation of
◦ ◦
80 /90 magnetic domains [20]. We discounted the surface
1
partial oxidation because no obvious O signal was detected by
EDS and this effect would be very weak if there is any. As
for the surface roughness, we did not find large difference in
M–H curves (applied field perpendicular to the film plane) of
the samples grown at different temperatures with root mean
square (rms) surface roughness (σ). So this reason can also
be excluded. Therefore, the resulting broad shoulder can be
attributed to the interface effect, the existence of carbon and
◦
◦
the formation of 180 /90 magnetic domains, but it is not
clear which is the dominating reason. Furthermore, from Fig. 4,
the square loop behavior in-plane and the slow approach
to saturation out-of-plane for the samples indicated that the
Fe films grown at different temperatures are ferromagnetic
with easy axis along the film surface. The magnetic moment
3
3
per unit volume (emu/cm ) of sample a is 1620 emu/cm ,
3
which is smaller than that of bulk α-Fe1707 emu/cm at
room temperature. The growth-temperature dependence of
the coercivity (Hc) and the saturation magnetization (Ms) is
shown in Fig. 5, when the applied field is parallel to the
film surface. By comparison, it can be seen that the 593 K
Fig. 3. AFM images of Fe films grown at the temperature of (a) 593 K, (b)
73 K and (c) 553 K. Image size is 0.5 × 0.5 µm with 150 nm high.
5
(573 K) sample has the lowest (highest) Hc but the highest
lowest) Ms among the three samples. This is in agreement
(
of Fe film was studied by EDS (see the inset of Fig. 2).
No obvious O signal has been detected and little of carbon
with the XRD observation that sample a (b) has the narrowest
(widest) FWHM value. According to Scherrer’s formula, the
grain size of sample a (b) will be much larger (smaller)
than the other two samples. Since the ferromagnetic coupling
is strongly dependent on the exchange interaction between
neighboring magnetic ions, decrease of the grain size (or
domain) can even change a system from the ferromagnetic
(
weight ratio ≈ 1.2%) may come from the decomposition
of Fe(CO)5. Fig. 3 shows the surface morphology of the Fe
films prepared at 553 K, 573 K and 593 K, respectively.
The thickness of the Fe films is about 550 nm. The scanned
2
area is 5 × 5 µm . The AFM pictures clearly show the

3
6
K. Liu et al. / Solid State Communications 140 (2006) 33–37
Fig. 5. The growth temperature dependence of Ms and coercivity measured at
room temperature, when the applied field was parallel to the sample surface.
The inset shows the growth-temperature dependence of Mr /Ms.
4. Conclusion
Fe films with strong preferred orientation (001) were
prepared on GaAs (001) substrates by LP-MOCVD with the
pressure of 150 Torr at different temperatures. The surface
roughness of Fe films decreases with increasing growth
temperature. Magnetization measurements showed that there
were remarkable differences between in-plane and out-of-plane
magnetic properties of Fe films. The sample prepared at 593 K
has the biggest Ms and Mr /Ms ratio but the lowest Hc due
to its high crystal quality. Furthermore, when the applied
magnetic field is perpendicular to the Fe surface, one can clearly
observe a sharp jump in the hysteresis loop, followed by a
broad shoulder. It is considered that the broad shoulder may be
caused by the interface effect, the existence of carbon and the
◦
◦
formation of 180 /90 magnetic domains. The detailed reasons
need to be further studied.
Acknowledgements
This work is supported by the National Natural Science
Foundation of China under Grant No. 50402016 and No.
Fig. 4. Magnetization vs. applied field along and perpendicular to the surface
of the films grown at different temperatures: (a) 593 K; (b) 573 K; (c) 553 K.
60501025; the Key Project of National Natural Science
Foundation of China under Grant No. 60336020, the Innovation
Project of Chinese Academy of Sciences.
state to the paramagnetic state. Thus, samples with larger
grains will present a larger saturation magnetization (Ms) at
low magnetic field. Similarly, a smaller coercivity (Hc) value
will also be expected in samples with larger grains due to
the easier inversion for larger magnetic domains. It is also
found that the ratio between the remnant magnetization and
saturation magnetization (Mr /Ms) increased from 0.76 to 0.97
with the increase of the growth temperature (see the inset
of Fig. 5), when the applied field is parallel to the film
plane.
References
[
1] S. Datta, B. Das, Appl. Phys. Lett. 56 (1990) 665.
[2] R. Sato, K. Mizushima, Appl. Phys. Lett. 79 (2001) 1157.
[3] H.J. Zhu et al., Phys. Rev. Lett. 87 (2001) 016601.
[
[
4] A.T. Hanbicki et al., Appl. Phys. Lett. 80 (2002) 1240.
5] O. Thomas, Q. Shen, P. Schieffer, N. Tournerie, B. Le’pine, Phys. Rev.
Lett. 90 (2003) 017205-1.
[
6] Y. Chye, V. Huard, M.E. White, P.M. Petroff, Appl. Phys. Lett. 80 (2002)
449.

K. Liu et al. / Solid State Communications 140 (2006) 33–37
37
[
[
[
7] J.R. Waldrop, R.W. Grant, Appl. Phys. Lett. 34 (1979) 630.
8] G.A. Prinz, J.J. Krebs, Appl. Phys. Lett. 39 (1981) 397.
9] G.A. Prinz, G.T. Rado, J.J. Krebs, J. Appl. Phys. 53 (1982) 2087.
[18] Y. Shiratsuchi, Y. Endo, M. Yamamoto, S.D. Bader, J. Appl. Phys. 97
(2005) 10J106.
[19] B. Raquet, R. Mamy, J.C. Ousset, Phys. Rev. B 54 (1996) 4128.
[20] J. Swerts, K. Temst, M.J. Van Bael, C. Van Haesendonck, Y.
Bruynseraede, Appl. Phys. Lett. 82 (2003) 1239.
[21] M. Gester, C. Daboo, R.J. Hicken, S.J. Gray, A. Ercole, J.A.C. Bland,
J. Appl. Phys. 80 (1996) 347.
[
10] E.M. Kneedler, B.T. Jonker, P.M. Thibado, R.J. Wagner, B.V. Shanabrook,
L.J. Whitman, Phys. Rev. B 56 (1997) 8163.
11] J.M. Florczak, E.D. Dahlberg, Phys. Rev. B 44 (1991) 9338.
12] C. Daboo et al., Phys. Rev. B 51 (1995) 15964.
[
[
[
[
13] A. Filipe, A. Schuhl, P. Galtier, Appl. Phys. Lett. 70 (1997) 129.
14] M. Z o¨ lfl, M. Brockmann, M. K o¨ hler, S. Kreuzer, T. Schweinb o¨ ck,
S. Miethaner, F. Bensch, G. Bayreuther, J. Magn. Magn. Mater. 175
[22] B. L e´ pine et al., J. Cryst. Growth 201–202 (1999) 702.
[23] Y. Chye, V. Huard, M.E. White, P.M. Petroff, Appl. Phys. Lett. 80 (2002)
449.
(
1997) 16.
[24] X. Jin, C. McEvoy, I.V. Shvets, J. Appl. Phys. 97 (2005) 10M103.
[25] W. Kipferl, M. Sperl, T. Hagler, R. Meier, G. Bayreuther, J. Appl. Phys.
97 (2005) 10B313.
[
[
[
15] Y.B. Xu, E.T.M. Kernohan, D.J. Freeland, A. Ercole, M. Tselepi, J.A.C.
Bland, Phys. Rev. B 58 (1998) 890.
16] H.-P. Sch o¨ nherr, R. N o¨ tzel, W. Ma, K. Ploog, J. Appl. Pyhs. 89 (2001)
[26] T.A. Moore, M.J. Walker, A.S. Middleton, J.A.C. Blanda, J. Appl. Phys.
97 (2005) 053903.
[27] H.J. Haugan, B.D. McCombe, P.G. Mattocks, J. Magn. Magn. Mater. 247
(2002) 296.
1
69.
17] P.K. Muduli, J. Herfort, H.-P. Sch o¨ nherr, K.H. Ploog, J. Appl. Phys. 97
2005) 123904.
(