Journal of Magnetism and Magnetic Materials 372 (2014) 233–235

Contents lists available at ScienceDirect

Journal of Magnetism and Magnetic Materials

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

Interfaces anisotropy in single crystal V/Fe/V trilayer

D. Louis a, Ia. Lytvynenko b, T. Hauet a,n, D. Lacour a, M. Hehn a, S. Andrieu a, F. Montaigne a
a
Institut Jean Lamour, UMR CNRS 7198, Université de Lorraine, 54506 Vandoeuvre-lès-Nancy, France
b
Sumy State University, 2, Rymskogo-Korsakova Street, 40007 Sumy, Ukraine

art ic l e i nf o a b s t r a c t

Article history: The value and sign of V/Fe interface anisotropy are investigated. Epitaxial V/Fe/V/Au layers with different
Received 20 May 2014 iron thicknesses were grown on single-crystalline (001) MgO substrate by ultra-high vacuum molecular
Received in revised form beam epitaxy. Magnetometry was used to measure magnetization and out-of-plane anisotropy field.
27 June 2014
From these values, we quantify the number of dead layers due to V/Fe or Fe/V interfaces, and compare it
Available online 2 August 2014
with the literature. We deduce that dead layers occur mostly at the bottom V/Fe interface. An average
Keywords: value for V/Fe and Fe/V interface anisotropy around 0 7 0.1 erg/cm2 (mJ/m2) was thus deduced.
Nanomagnetism & 2014 Elsevier B.V. All rights reserved.
Interface anisotropy
Epitaxial films
Single crystal Iron
Magnetization

1. Introduction temperature are presented both for in-plane and out-of-plane field
orientation relative to the V/Fe/V interfaces. The magnetization at
Research on spintronic devices such as magnetic random access saturation and anisotropy field are plotted as a function of the Fe
memories (MRAM) and magnetic sensors have generated a per- thickness t from which we extract an average value for Fe/V and
petual need in original magnetic materials. As the thickness of the V/Fe interface anisotropy constant.
magnetic thin films (e.g. electrodes for giant or tunnel magne-
toresistance multilayer) have shrunk, the influence of interfaces
has become crucial. The control of the magnetic configuration 2. Experiments
using interfacial effect is still heavily studied, e.g. exchange bias
with ferromagnetic/anti-ferromagnetic interfaces [1], coercivity The samples were grown on single-crystalline MgO (100)
control with hard/soft bilayer in recording media [2], multiferro- substrate using MBE with a base-pressure lower than 10
10 Torr.
ism with ferromagnetic/ferroelectric interface [3], magnetization The V (20 nm) buffer layer was deposited at room temperature
induction at oxide/oxide interface [4], Dzyaloshinskii–Moriya and annealed at 600 1C. Fe layers of thickness t ranging from
interaction [5], etc. One of the most studied features is the 0.7 nm (5 atomic layers) to 5 nm (35 atomic layers) were grown at
interface-induced anisotropy that originates from the hybridiza- room temperature, annealed at 350 1C and capped with V (5 nm)/
tion between two layers composed of different chemical elements Au (5 nm) with no further annealing.
like Co/Pt, Co/Ni, etc. (see Ref. [6] for a review). Recently, we Fluxes were calibrated by quartz or reflection high energy
investigated V/Fe/MgO as a model system to characterize electrical electron diffraction (RHEED) oscillations in-situ during growth
control of the magnetic anisotropy by a bias voltage [7,8]. and with ex-situ X-ray small angle reflectivity. The epitaxial
Although the interface magnetization and in-plane bulk aniso- relationship, growth mode, number of deposited MLs, and surface
tropy in [Fe/V] multilayers have been widely studied in the past flatness were controlled in situ using RHEED. Fig. 1 shows RHEED
[9–12], we could not find in the literature a clear thickness pattern for annealed V and Fe layers along the (001) direction and
dependence allowing to precisely determine the interface aniso- confirms good crystalline quality of the films. Later, magnetization
tropy for V/Fe interface. curves were measured using commercial rotating sample vibrating
In the present report, we show magnetometry data obtained on sample magnetometer (VSM) and commercial SQUID-VSM.
epitaxial V/Fe (t)/V trilayer grown by ultra-high vaccum molecular
beam epitaxy (MBE). Magnetization versus field loops at room
3. Results and discussions

n
Corresponding author. Fig. 2(a) shows normalized magnetization versus field loops
E-mail address: thomas.hauet@univ-lorraine.fr (T. Hauet). measured both for in-plane magnetic field and out-of-plane

http://dx.doi.org/10.1016/j.jmmm.2014.07.018
0304-8853/& 2014 Elsevier B.V. All rights reserved.
234 D. Louis et al. / Journal of Magnetism and Magnetic Materials 372 (2014) 233–235

Fig. 1. RHEED patterns along the (001) BCC axis for (a) annealed V (20 nm) buffer layers deposited on MgO and (b) MgO/V/Fe (1 nm) along (001) direction.

observed and in-between the lobes remanent magnetization is

75% of the saturation magnetization.
From the magnetization versus field curves, one can extract
magnetization at saturation. Here this quantity has been obtained
by dividing the measured moment value measured by magneto-
metry by the nominal Fe volume. In Fig. 3, we plot saturation
magnetization times thickness (t), i.e. areal magnetic moment, as
a function of the deposited Fe thickness (t). The areal magnetic
moment versus t should be a straight line passing through 0, with
a slope equal to the bulk Fe magnetization (about 1720 emu/cm3).
Although the slope is as expected, the line crosses zero for 0.3 nm.
This result is coherent with a reduction of Fe magnetization at the
interface with V which is reported in the literature. This has been
explained by roughness, charge transfer, intermixing and anti-
parallel polarization of the V [9,11,13]. An oxygen contamination of
the starting V (001) surface should also contribute to these
Fig. 2. (a) Normalized magnetization versus field loop for in-plane along (100) Fe
magnetic dead layers in Fe grown at room temperature [10].
direction (black solid squares) and out-of-plane (open red circles) field for a V/Fe
(2 nm)/V stack, (b) zoom of the main figure around zero field show square Interestingly, the same deadlayer thickness (tdl) is found in Ref.
hysteresis cycle, and (c) normalized remanent magnetization as a function of field [7] for a MgO substrate/V/Fe/MgO stack where Fe is grown on V in
angle when applied in film plane demonstrates bulk cubic ansiotropy. (For the very same conditions (same setup). As no reduction of
interpretation of the references to color in this figure legend, the reader is referred magnetization is observed at Fe/MgO interface when MgO is
to the web version of this article.)
grown on Fe [14], this suggests that the main loss of magnetization
occurs at the bottom of the V/Fe interface. Although vanadium
anti-parallel polarization may happen at both interfaces, the main
reduction must originate from the oxygen contamination of the V
buffer layer (O segregation after the buffer layer annealing [10]),
that does not occur for the top (not annealed) V layer.
In Fig. 2, a hard axis loop is obtained when the field is applied
perpendicularly to the V/Fe interface. It means that the magneti-
zation lies preferentially in the film plane. The same behavior has
been observed for all tested samples. Same conclusion is reported
in the literature for Fe thickness as thin as three atomic layers [12].
To further investigate the role of Fe/V and Fe/V interface in
stabilizing or destabilizing in-plane magnetization, we calculate
the effective anisotropy constant from the anisotropy field Hkeff
extracted from the out-of-plane field hard axis loops as
K keff ¼ 12 M S H keff . Various origins of effective anisotropy are usually
distinguished as in Eq. (1), here in CGS and including the dead-
Fig. 3. Areal magnetization versus Fe thickness. The linear fit has positive slope
layers [6]:
value 1720 emu/cm3. It crosses zero areal moment for a non-zero-thickness
K eff ðt Fe –t dl Þ ¼ ðK V –2π M 2S Þðt Fe
t dl Þ þ K i ð1Þ
meaning that Fe magnetization is reduced in average over about 2 atomic layers.
KV is the magnetic volume anisotropy and Ki is the interfaces
magnetic field for V/Fe (2 nm)/V stack. Out-of-plane direction anisotropy acting in the Fe layer. The 2πM2S term comes from the
corresponds to a magnetic hard axis direction for the Fe layer. shape anisotropy for a thin film. The negative sign shows that this
In-plane field measurement with field applied along Fe (100) anisotropy term tends to align the magnetization in the film plane.
direction shows a square loop with full magnetization at rema- The thickness t is the nominal thickness of the film and tdl is the
nence (see zoom view in Fig. 2(b)). In Fig. 2(c) is plotted the deadlayer thickness equal to 0.3 nm.
remanent magnetization extracted from hysteresis loops obtained To quantify the different terms, we plot in Fig. 4 Keff(tFe
tdl) as
for various in-plane applied field direction. As expected for cubic a function of (tFe
tdl). All values Keff(tFe
tdl) are negative which
Fe bulk magnetic anisotropy, four lobes with full remanence are confirm that the effective anisotropy favors in-plane orientation of
 D. Louis et al. / Journal of Magnetism and Magnetic Materials 372 (2014) 233–235 235

magnetic properties of single-crystal V/Fe(t)/V trilayer grown on

bcc MgO substrate. Magnetometry allows measuring magnetiza-
tion at saturation and out-of-plane anisotropy field. A total of
0.3 nm deadlayers is found and mainly attributed to the lower V/Fe
interface. In plotting the anisotropy constant deduced from the
anisotropy field, as a function of thickness, we conclude that the
volume anisotropy is dominated by in-plane demagnetization field
whereas the interface anisotropy related to Fe/V interface is
around zero with accuracy equal to 7 0.1 erg/cm2.

Acknowledgment

Authors thank T. Ferte and L. Pasquier for helping with growth

Fig. 4. Effective anisotropy constant Keff times the corrected thickness tFe
tdl as a experiments. This work was supported by the Region Lorraine and
function of tFe
tdl at 300 K for both V/Fe/V. The line is a fit using Eq. (1) with by the French Agence Nationale de la Recherche, ANR-2010-
negative slope corresponding to a magnetization 1720 emu/cm3 and a value at zero BLANC-1006 (Elecmade) and ANR-2012-BS04-0009 (Frustrated).
thickness corresponding to a value of Ki ¼07 0.1 erg/cm2.

magnetization for all Fe thicknesses. A linear evolution is obtained

whose slope is also negative. It means that KV
2πM2S term here References
provides an in-plane anisotropy. If we consider only
2πM2S term,
[1] F. Radu, H. Zabel, Magnetic Heterostructures, Springer Tracts in Modern
one finds MS ¼1720 emu/cm3, in good agreement with value
Physics, Springer-Verlag, Berlin Heidelberg, 2008.
measured extracted from Fig. 3 and with Fe bulk magnetization. [2] T. Hauet, E. Dobisz, S. Florez, J. Park, B. Lengsfield, B.D. Terris, O. Hellwig, Appl.
It confirms that KV is small as compared with shape anisotropy Phys. Lett. 95 (2009) 262504.
[15,16]. Note that the role of strain on KV term in V/Fe multilayers [3] S. Valencia, A. Crassous, L. Bocher, V. Garcia, X. Moya, R.O. Cherifi, C. Deranlot,
K. Bouzehouane, S. Fusil, A. Zobelli, A. Gloter, N.D. Mathur, A. Gaupp,
has been studied in detail in Ref. [17]. However, we have shown in
R. Abrudan, F. Radu, A. Barthélémy, M. Bibes, Nat. Mater. 10 (2011) 753.
a previous study [18] that the critical thickness for plastic relaxa- [4] L. Li, C. Richter, J. Mannhart, R.C. Ashoori, Nat. Phys. 7 (2011) 762.
tion during Fe growth on a V (001) surface containing oxygen at [5] S. Emori., U. Bauer, S.-M. Ahn, E. Martinez, G.S.D. Beach, Nat. Mater. 12 (2013)
room temperature is lower than 1 monolayer (i.e. lower than 611.
[6] M.T. Johnson, P.J.H. Bloemen, F.J.A. den Broeder, J.J. de Vries, Rep. Prog. Phys. 59
0.14 nm). This means that Fe layer relaxes to its stable bcc (1996) 1409–1458.
structure, leading to small magnetoelastic anisotropy. [7] C.-H. Lambert, A. Rajanikanth, T. Hauet, S. Mangin, E.E. Fullerton,and,
The value of anisotropy Ki extracted from the curve Fig. 4 at S. Andrieu, Appl. Phys. Lett. 102 (2013) 122410.
[8] T. Maruyama, Y. Shiota, T. Nozaki, K. Ohta, N. Toda, M. Mizuguchi,
zero thickness is close to zero. Actually, regarding the accuracy of
A.A. Tulapurka, T. Shinjo, M. Shiraishi, S. Mizukami, Y. Ando, Y. Suzuki, Nat.
our measurements and fit, we have to consider an error bar of at Nanotechnol. 4 (2009) 158.
most 70.1 erg/cm2. This result is in agreement with Ref. [11] [9] J. Izquierdo, R. Robles, A. Vega, M. Talanana, C. Demangeat, Phys. Rev. B 64
where the anisotropy of the orbital and spin moments is extracted (2001) 060404(R).
[10] F. Dulot, P. Turban, B. Kierren, J. Euge`ne, M. Alnot, S. Andrieu, Surf. Sci. 473
from ferromagnetic resonance and compared with first-principle (2001) 172.
calculations. The value of Ki originates from the bottom V/Fe [11] A.N. Anisimov, M. Farle, P. Poulopoulos, W. Platow, K. Baberschke, P. Isberg,
interface and from the top Fe/V. Since we know that atomic R. Wäppling, A.M.N. Niklasson, O. Eriksson, Phys. Rev. Lett. 82 (1999) 2390.
[12] P. Poulopoulos, P. Isberg, W. Platow, W. Wisny, M. Farle, B. Hjörvarsson,
arrangement at V/Fe and Fe/V are different, we cannot assure that
K. Baberschke, J. Magn. Magn. Mater. 170 (1997) 57.
both interfaces lead to zero but it is most probable. [13] M. Sicot, S. Andrieu, P. Turban, Y. Fagot-Revurat, H. Cercellier, A. Tagliaferri,
C. De Nadai, N.B. Brookes, F. Bertran, F. Fortuna, Phys. Rev. B 68 (2003) 184406.
[14] M. Sicot, S. Andrieu, F. Bertran, F. Fortuna, Phys. Rev. B 72 (2005) 144414.
4. Summary and conclusions [15] C.D. Graham, Phys. Rev. 112 (1958) 1117.
[16] D. Sander, J. Phys. Condens. Matter 16 (2004) 603.
[17] A. Broddefalka, P. Nordblada, P. Blomqvist, P. Isberg, R.W. Appling, O. Le Bacq,
To fill the lack of information about Fe/V interface anisotropy in O. Eriksson, J. Magn. Magn. Mater. 241 (2002) 260.
the literature, we studied the influence of the Fe thickness on the [18] P. Turban, L. Hennet, S. Andrieu, Surf. Sci. 446 (2000) 241.