Magnetic properties of Ni/Pt multilayers
R. Krishnan, H. Lassri, Shiva Prasad, M. Porte, and M. Tessier
Citation: J. Appl. Phys. 73, 6433 (1993); doi: 10.1063/1.352623
View online: http://dx.doi.org/10.1063/1.352623
View Table of Contents: http://jap.aip.org/resource/1/JAPIAU/v73/i10
Published by the American Institute of Physics.
Related Articles
Spin transfer switching characteristics in a [Pd/Co]m/Cu/[Co/Pd]n pseudo spin-valve nanopillar with
perpendicular anisotropy
J. Appl. Phys. 111, 07C910 (2012)
Magnetization reversal mechanisms in 35-nm diameter Fe1-xGax/Cu multilayered nanowires
J. Appl. Phys. 111, 07A920 (2012)
Promotion of L10 ordering of FePd films with amorphous CoFeB thin interlayer
J. Appl. Phys. 111, 07C112 (2012)
Effect of thickness of MgO, Co-Fe-B, and Ta layers on perpendicular magnetic anisotropy of
[Ta/Co60Fe20B20/MgO]5 multilayered films
J. Appl. Phys. 111, 07C111 (2012)
Effect of the number of interfaces on the magnetic properties of [SnO2/Cu-Zn ferrite] multilayer
J. Appl. Phys. 111, 07C110 (2012)
Additional information on J. Appl. Phys.
Journal Homepage: http://jap.aip.org/
Journal Information: http://jap.aip.org/about/about_the_journal
Top downloads: http://jap.aip.org/features/most_downloaded
Information for Authors: http://jap.aip.org/authors
Downloaded 01 Mar 2012 to 14.139.97.76. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
Magnetic properties
of Ni/Pt multilayers
FL Krishnan and H. Lassri
Magt&istne
et Mat&iaux
MagnL;tiques, CNRS, 92195 Meudon, France
Shiva Prasad
Physics Department, Indian Institute of Technology, 400076 Bombay, India
M. Porte and M. iessier
Magnkisme
et iliatC;riaux Mugnetiques, CNRS, 92195 Meudon, France
We have grown Ni/Pt multilayers (ML) with and without Pt buffer layer, by evaporation under
UHV conditions on glass and Si substrates maintained at 30 and 200 “C!. The magnetization
decreases with t(Ni) and no induced moment is found on Pt from polareation effects even at 5
K. The perpendicular magnetization loops for samples with t(Ni)<l7
A are rectangular and
coercivities as high as 4.7 kOe is obtained for t(Ni) =9 A at 5 K. For the samples deposited at
20 “C, due to dispersion in the results, KS could not be calculated with accuracy though it is
positive. But for the samples deposited on a Pt buffer layer at 200 “C, KS is found to be $0.17
erg cm -’ at 5 K and the volume anisotropy is also higher with respect to those deposited at
30 32.
INTRODUCTION
Recently much attention has been paid to multilayers
based on Pt, such as Co/I?, which show perpendicular
anisotropy for Co layers thinner than about I5 A.‘-s Besides the fundamental interest, these materials are also
promising candidates for magnetooptical storage media capable of performing at shorter wavelengths. Recently perpendicular anisotropy was reported also in Fe/Pt multilayem4 However, reports on Ni-based systems are relatively
very few. We had reported that in Ni/Ag (Ref. 5) multilayers there was no contribution to the surface anisotropy,
even though a contribution to the surface magnetostriction
from the surface Ni atoms was found.6 Therefore it is interesting, from the fundamental point of view, to investigate if surffce anisotropy could be found in other Ni-based
systems. Our preliminary investigations indeed showed
that perpendicular magnetization with a remanence ratio
of one could be obtained in Ni/Pt multilayers for relatively
thin Ni layers.7 Earlier, rectangular hysteresis loops had
been observed also for Ni/Pd multilayers even though the
remanence ratio was much smaller than one.8p9In this paper we describe our study of the magnetization and the
anisotropy in Ni/Pt multilayers. We show that the samples
with Ni layers thinner than about. 15 A show distinctly
different behavior from those with thicker Ni layers. The
magneto-optical Kerr speetroseopy of these samples is reported elsewhere.‘”
EXPERIMENTAL
RESULTS AND DJSCUSSION
Low-angle x-ray diffraction of ah the samples revealed
peaks typical of the modulated structure, and the thickness
calculated from these peaks agree within 3% with that
obtained from the quartz oscillator after calibration. Figures l(a) and l(b) show the low- and high-angle x-ray
diffractions for a sample [9,20] x 32 grown at 200 “C on a
lOO-A-thick Pt layer. The (111) peak from the Pt buffer
DETAILS
NiPPt multilayers (ML) were grown by evaporation in
ultrahigh vacuum under controlled conditions, and the
pressure during the film deposition was maintained in the
range 3-5~ lOem Torr. The rate of deposition (about 0.3
A/s) and the flnal thickness were monitored by precalibrated quartz oscillators. The Ni-layer thickness t( Ni) was
varied from 6 to 40 A and that of t(Pt) was kept fixed at
20 A. The number of bilayers in the range 10-32 was adjusted to get a total t(Ni) of about 250 A. Samples were
6433
deposited on glass substrates at 30 “C. Samples with Ni
layer thinner than about 15 A were also deposited on a
platinum buffer 100 A thick at 200 “C for the reasons explained later. The top layer in all the samples was Pt 20 A
thick. The growth parameters would be designated as
[r(Ni),r(Pt)]Xn,
where it indicates the number of Ni layers.
Low-angle x-ray diffraction studies were made to
check the periodicity and the thickness of the bilayer. Magnetization (M) and the M-H loops were measured with a
vibrating sample magnetometer (VSM) and the anisotropy
with a torque magnetometer, in the temperature range
5-295 K under a maximum field of 15 kOe.
.I. Appi. Phys. 73 (IO), 15 May 1993
0
4
8
-40
50
60
‘20, IUi)
FIG. 1. Low-angle (a) and high-angle (b) x-ray diffraction patterns for
the sample [9,20] ~32 deposited on IO@k-thick Pt buffer layer and at
2CXfT.
0021-6979/93/106433-03$06.00
@ 1993 American institute of Physics
Downloaded 01 Mar 2012 to 14.139.97.76. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
6433
M
500
L
t
H
//
/
F
5 koe
FTG. 2. The t(Ni)
I
I:
t
0
10
20
,
,
30
40
tpJi (A)
, .
50
dependence of the magnetization at 295 and 5 K.
layer and the satellite peaks are clearly seen in Fig. 1 (b). It
can be concluded that Ni layers have a (111) texture.
Figure 2 shows the t(Ni) dependence of the magnetization both at 295 and 5 K. At 295 K the samples with
f(Ni) < 15 A are very weakly magnetic because the Curie
temperature for these samples is close to 300 K. The relatively lower M values and the decrease in M with decreasing r(Ni) indicate the formation of the less magnetic or
nonmagnetic Ni-Pt alloy at the interface. Assuming a dead
layer at the interface, one can express the magnetization in
the multilayers M as M=Me( 1 - 26/t), where Ma stands
for the magnetization of bulk Ni and 6 is the dead-layer
thickness at each interface. An analysis of a plot of
MXt(Ni)
as a function of t(Ni) at 5 K gives Me= 500
~20 emu/cm3 which is close to the value of 520 emu/cm3
of bulk Ni and 2S in the range S-6 A. This means that at
each interface about a monolayer of Ni is nonmagnetic. It
seems, therefore, that in Ni/Pt system, Ni atoms do not
induce any moment on Pt. It is recalled that in Co/Pt
multilayers Co has been shown to induce a moment on
neighboring Pt atoms. ‘*
For the samples with t(Ni) 220 A the in-plane (H
applied parallel to the tllm plane) M-H loops at 295 K
indicating
show a high remanence ratio (R =MR/MS),
that the magnetization lies in the film plane. The in-plane
coercivity increases at low temperatures. For example, the
coercivity of the sample [28,20] X 11 deposited at 300 K on
glass increases from 17 to 180 Oe as the temperature is
decreased from 300 to 5 K. On the contrary, for the samples with f( Ni) d 17 A, the in-plane loops are typical of the
hard axis one, with practically no remanence at all and one
observes rectangular M-H loops along the film normal indicating the presence of a strong uniaxial anisotropy. Figure 3 shows the perpendicular M-H loop at 5 K for the
sample [9,201X 32, deposited on lOO-A-thick Pt buffer
layer and at 475 K. It is noteworthy that the loop is rectangular with R = 1 with a coercivity as high as 4.7 kOe. As
the temperature is increased, the remanence ratio and the
coercivity start decreasing.
Let us now discuss the results on anisotropy. It is well
known that the layer thickness dependence of the anisotropy in multilayers can be described, based on the phenomenological model, by the equation Kerr= K V+ 2&/r, where,
6434
J. Appl. Phys., Vol. 73, No. 10, 15 May 1993
FIG. 3. The perpendicular loop at 5 K for the sample, [9,20] X 32, deposited on E&&-thick
Pt buffer layer and at 200 ‘C.
&,, is the measured anisotropy, and KY, K, are the volume
and the surface anisotropies. The term K, consists of three
contributions: ( 1) the demagnetizing energy 2?rM”, (2)
the intrinsic crystalline anisotropy Kerr, and (3) the magnetoelastic anisotropy KhllE, which arises from the interaction of the magnetos&i&ion d with the stress in the film.
This is given by the term KME= - f aA, where (z is the
stress. In polycrystallme layers in the absence of any crystalline anisotropy, one can expect a contribution to KY
from KCYLp We had mentioned that in our samples some
( 111) texture is observed. Therefore the film normal corresponds to [l 1 I] direction, which is the easy axis for bulk
nickel. So one might argue that this could lead to the observed perpendicular anisotropy. But this is not so. For
instance, Takahashi er aL9 found in Ni/Pd multilayers that
perpendicular anisotropy was observed only for the [lOO]
orientation and not for the [Ill]. Therefore we consider
that the intrinsic anisotropy observed here arises only from
the magnetoelastic interactions.
The product of K&X t(Ni) versus t( Ni) at 5 K is
plotted for the samples with 6 < t( Ni) < 40 A, deposited at
30 “C in Fig. 4 and one finds that for t(Ni) < 10 A Keff is
positive. It is seen that while the experimental points align
well in a straight line for t(Ni)> 12 A, those for thinner
layers show some dispersion, though the deposition param-
t
0.3
1-.
\,
‘.
*\
‘\
03
0‘1
2
t
-.
l
3aJo
.
5r.,,,
..._
~~-.JbtNIIA)
&+
i ‘._-...ia<o
3 l-----
CJJ
0.1.
- 0,s
0,s
Y
0
o
.
- \-.
J----Y
FIG. 4. The t(Ni) dependence of the product K,,x f(Ni) for the samples
deposited both at 30 “C (open circles) and at 200 “C on LOO-A-thick Pt
buffer layer (closed circles).
Krishnan et al.
Downloaded 01 Mar 2012 to 14.139.97.76. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
6434
eters were carefully controlled. Nevertheless, one can remark that there are two distinct straight lines with two
different slopes, the one for t ( Ni) < 10 8, being higher. This
suggested that the change in the properties observed for
thinner layers could be either due to a change in the structure or to some coherency strains. The extrapolation of the
data for samples with t(Ni) > 10 A shows that there is a
small contribution from the surface anisotropy with KS > 0.
As it was difficult to calculate the anisotropies for the thinner Ni layers with any accuracy, we tried to deposit the
samples at a higher temperature, hoping that this could
stabilize the changes that occurred for the thinner layers.
Figure 4 also shows the results at 5 K for the samples
deposited at 200 “C on a Pt buffer layer 100 A thick. It is
seen that the experimental points are now aligned better
and the &is positive for all the samples (9 to 17 A thick).
The extrapolation of the straight line gives KS= 3-0.17
erg cm ‘. This is certainly smaller than l-O.6 erg cmd2
found for Co/Pt multilayers. Nevertheless this is the first
time such a positive surface anisotropy has been reported
for N&based multilayers.
The slopes of the straight lines in Fig. 4 confirm the
two dist.inct behaviors both for the samples deposited at 30
and 200 “C. RV values of -1.9X 10’ and -0.9X lo6
erg cm -3 are found for the thinner and thicker Ni layers,
respectively. This increase in K, for the thinner samples
[deposited both at 30 and 200 “C) with respect to the
thicker ones suggest that KME contributions are not the
same in both the cases.
In order to calculate KkjE it is necessary to know the
magnetostriction constants (A) in these materials. We
have shown that 3, is also strongly dependent on Ni layer
thickness as, for example, in Ni/Ag, where we have
shown” that the absolute value of A decreased with r(Ni)
and tended to become posit.ive due to the contribution from
surhace magnetos&i&ion. This would mean that even if the
stress (LT) in the tilms remains the same (as a first approximation) XhiE could still be different for sufficiently thinner
Ni layers. But on the one hand, the t dependence of/z in
Ni/Pt could be different from that observed in NVAg, and
on the other hand, o also could be quite different for very
thin layers. In the absence of any measurements of R in
NVPt, it would be a speculation to discuss any further.
6435
J. Appl. Phys., Vol. 73, No. 10, 15 May 1993
Nevertheless our results do show the important role played
by the magnetostriction in multilayers, which is often neglected and one assumes the bulk value of L to interpret the
anisotropy in very thin layers. Magnetos&i&ion measurements in Ni/Pt are necessary to understand our results
further and work is in progress. It is also interesting to
study the influence of the substrate temperature on the
anisotropy.
In conclusion, we have prepared Ni/Pt multilayers by
evaporation under ultrahigh vacuum conditions and characterized them. The effective anisotropy for layers thinner
than about 1.5 A deposited at 30 “C is positive but show a
dispersion. On the contrary, for the samples with Ni layers
thinner than 15 A deposited at 200 “C on Pt buffer layer it
is found clearly that the surface anisotropy at 5 K was
j-O.17 erg cm-z. The perpendicular loops for these thinner
samples show a remanence ration of 1.0 and the coercivity
for the samples with t(Ni) =9 A was as high as 4.7 kOe
at 5 K.
The partial supports of this work from Brite Euram
Contract BREU-0153 and from Indo-French Center for
the Promotion of the Advanced Research (Centre FrancoIndien pour la Promotion de la Recherche Avancee) are
gratefully acknowledged.
‘P. F. Car&, A. D. Meinhaldt, and A. Suna, Appt. Phys. Lett. 47, 178
(1985).
*W. 3. Zeper, F. J. A, Greidanus, P. F. Garcia, and P. R. Fincher, J.
Appl. Phys. 65, 4971 ( 1989).
3R. Krishnan, M. Porte, and M. Tessier, IEEE Trans. Magn. 26, 2727
(1989); J. Magn. Sot. Jpn. 15, 21 (1991 j.
“5. Iwata, S. S. P. Parkin, H. Nuri, and T. Suzuki, Mater. Res. Sot.
Symp. Proc. 232, 85 (1991).
‘R. Krishnan, M. Porte, M. Tessier, H. Szymczak, and R. Zuherek,
Proceedings of the Ffth Internationai Conference on Physics of d-iagnetic
&iureriaZs, edited by W. Gorzkowskii M. Gutowski, H. K. Lachowicz,
and H. Szymczak (World Scientific, Singapore, 1990), p. 294.
“R. Zuberek, H. Szymczak, R. Krishnan, and hl. Tess&, J. Phys.
(Paris) Colioq. Suppl.. 49, CS-1761 (1988).
‘R. Krishnan, H. Laasri, M. Porte, M. Tessier, and P. Renaudin, Appl.
Phys. L&t. 59, 3649 (1991).
*N. K. Flevaris, Appl. Phys. Lett. 5g, 2177 (1991).
“H. Takahashi, S. Fukatsu, S. Tsunashima, and S. Uchiyama, J. Magn.
Magn. Mater. 93, 469 (1991).
‘OP. Kieler, R. Krishnan, M. Nyvlt, V. Parizek, V. Presser, M. Tessier,
and S. Visnovsky (these proceedings).
“S. Riiegg, G. Schultz, P. Fischer, R. Wienke, W. B. &per. H. Ebert, J.
Appl. Phys. 69, 5655 (1990).
Krishnan et al.
Downloaded 01 Mar 2012 to 14.139.97.76. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions
6435