Magnetostriction and magnetoresistance in nanocontacts

Article abstract of DOI:10.1016/j.jmmm.2005.08.020

Ni nanocontacts have been grown by electrodeposition using a self-terminating technique in a single electrolyte bath based on nickel sulfate, nickel chloride and boric acid. Resistance measurements performed on different samples presented two kinds of obviously different magnetoresistance effects. The analysis of the data sets showed that magnetostriction might play a key role in magnetoresistance of the electrodeposited Ni nanocontacts.

Full text of DOI:10.1016/j.jmmm.2005.08.020

ARTICLE IN PRESS
Journal of Magnetism and Magnetic Materials 302 (2006) 82–85
www.elsevier.com/locate/jmmm
Magnetostriction and magnetoresistance in nanocontacts
Ã
Y.F. Zhu , G.Q. Di
Key Laboratory of Thin Solid Film Materials of Jiangsu province, Department of Physics, Suzhou University, Suzhou 215006, China
Received 26 May 2005; received in revised form 3 July 2005
Available online 26 September 2005
Abstract
Ni nanocontacts have been grown by electrodeposition using a self-terminating technique in a single electrolyte bath based on nickel
sulfate, nickel chloride and boric acid. Resistance measurements performed on different samples presented two kinds of obviously
different magnetoresistance effects. The analysis of the data sets showed that magnetostriction might play a key role in magnetoresistance
of the electrodeposited Ni nanocontacts.
r 2005 Elsevier B.V. All rights reserved.
PACS: 75.75.+a; 73.63.-b
Keywords: Nanocontact; Magnetoresistance; Magnetostriction
1. Introduction
report our own data sets and propose a mechanism to
explain them.
The study of magnetoresistance (MR) is one of the most
exciting topics, due to its great importance for technical
applications as hard-disk read heads and magnetic storage
media like magnetoresistance random access memory.
Renewed interest in submicron scale ferromagnetic nano-
contacts has been stimulated greatly by the discovery of
ballistic magnetoresistance (BMR) in 1999 [1]. The
obtained BMR in mechanically formed nanocontacts
reaches few hundred percents [1–4], which is attributed to
the scattering of conducting electrons by a constricted
domain wall (DW) within the nanocontacts. Recently,
Chung et al. [3] have shown that the MR oscillates as a
function of the nanocontacts’ resistance. Many electro-
deposited nanocontacts were also used to investigate BMR
[5–11], some of which demonstrate even larger MR than
the mechanically formed ones [9,10]. There is no correla-
tion between MR and resistance of the electrodeposited
nanocontacts [8], and there are some major issues in the
electrodeposited nanocontacts, such as poor reproducibil-
ity and low stability [12]. So far, there is no widely accepted
explanation for these phenomena [13]. In this paper, we
2. Experiment
Samples with geometry schematically shown in Fig. 1
were used: Two 99.9% pure Ni wires having diameter
in 200 mm, which were mounted on a piece of glass
substrate with epoxy. The wire I and wire II were posited
perpendicular to each other. The distance between wire I
and the tip of wire II was about 20 mm [8,9]. The details to
make samples are given elsewhere [7–10].
The nanocontacts were grown by electrodeposition from
a single electrolyte containing Ni2+ ions. The composition
of the electrolyte is as follows: 2.5 g NiSO4 ꢀ 6H2O and 1.4 g
NiCl2 ꢀ 6H2O and 0.3 g H3BO3 added at 10 ml H2O. The pH
of this solution is about 3–4. The self-terminating electro-
chemical method of Boussaad and Tao was used in the
electrodeposition process [14]. The deposition of nanocon-
tacts was carried out at room temperature. We applied a
constant potential of 1.05 V to deposit Ni from wire I to the
top of wire II. A lower pH of the electrolyte and/or a
higher potential between the wires will cause the generation
of hydrogen, which may affect the rat of deposition and
prevent the formation of Ni tip. A MR measuring system
was also employed to investigate nanocontacts. The
Ã
Corresponding author. Tel.: +86 512 65112251;
fax: +86 512 65112597.
E-mail address: zhu3512@163.com (Y.F. Zhu).
0304-8853/$ - see front matter r 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmmm.2005.08.020

ARTICLE IN PRESS
Y.F. Zhu, G.Q. Di / Journal of Magnetism and Magnetic Materials 302 (2006) 82–85
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
83
1000=R ðOÞ (in nm) [16], we can get the diameter of this
sample is 23.9 nm. Thus the predicted MR due to DW
scattering is E4.5%, which is not in agreement with the
experimental result (MR ¼ 477%). Therefore, the DW
scattering theory can not be used to explain the electro-
deposited nanocontact’s MR. What can be the explanation
of the huge MR in these contacts?
Nickel has negative magnetostriction and the saturation
magnetostriction constant ls ¼ ꢃ34 ꢂ 10ꢃ6 [17]. As shown
in Fig. 1, the length of the electrodeposited Ni between
wires is about 20 mm [8] (even reaches few tens of microns
[9]). Therefore, a magnetic field parallel to the wire II axis
will induce magnetostriction, which causes pulling nano-
contact away from the central part of wire I. For a length
of the electrodeposited Ni of 20 mm, the contraction due to
˚
magnetostriction is lsd ¼ ꢃ6.8 A. Because the epoxy covers
Fig. 1. Schematic drawing of the experimental setup. The length of the
electrodeposited Ni: dE20 mm. The diameter of Ni wire: D ¼ 200 mm.
the wire I except for the central part [8,9], there is a small
free part in wire I (see Fig. 1). A magnetic field
perpendicular to wire I will cause the free part to shrink
by less than 6.8 nm. Combination of the shrinking of wire I
and shortening of the electrodeposited Ni will cause a
displacement of between 0.68 and 7.5 nm. This displace-
ment would have a profound effect on the structure and
diameter of the nanocontact.
resistance of nanocontact was continuously monitored in a
constant current mode with a voltage measurement
technique, while magnetic field between 75.0 kOe was
applied parallel to the wire II axis.
If the displacement is very small, it may deform the
nanocontact. While the nanocontact is slowly stretched, its
cross-section is necking down. Since the conductivity
decreases in proportion to the cross-sectional area, the
resistance of nanocontact increases with increasing the
magnetic field, as the typical MR curves shown in Fig. 2.
When decreasing the magnetic field, the resistance of
nanocontact cannot drop to the initial value; it may be due
to that the nanocontact is distorted, which causes the poor
reproducibility of the electrodeposited nanocontacts (see
Fig. 2(b)). The resistance drops to minimum while the
magnetic field is not at zero (see inset at Fig. 2(a)), which
may be caused by the hysteresis of Ni wires.
Different length changes near the nanocontact result
in different MR effects. If the length change is very
large, it may break the nanocontact. Since we used a
constant current to measure the resistance of the nano-
contact, breaking the nanocontact will cause a very high
electric field between the gap formed by displacement,
strong electric field make the atoms in the contact move
[18], and then a discrete jump is obtained, as shown in
Fig. 3. Decreasing the magnetic field leads the wire I and
wire II to an intimate situation, thus there is no discrete
jump during this period. Therefore, we conclude that the
huge MR may partly due to the magnetostriction of the
electrodeposited Ni and the bulk Ni wire I. Moreover, the
MR could range between few hundred percents [5,8,16] and
100 000% [10].
3. Results and discussion
Fig. 2 shows consecutive magnetoresistance curves in a
sample whose initial zero-field contact resistance was
1.75 O after electrodepostion. It is seen from the first loop
that, with increase in the field in the positive direction, the
resistance increases with field until it reaches a saturation
value (10.1 O) which represents ꢁ477% MR (Fig. 2(a)).
The MR curves are similar to the previous reports [7–10].
During the period of experimentation, we made hun-
dreds of samples. In addition to Fig. 2, Fig. 3 is another
typical MR loop we often obtained. When the external
magnetic field increases from 0 to 75.0 KOe, the resistance
of nanocontact increases. When the nanocontact’s resis-
tance arrives at a maximum value, it jumps to about the
initial value. However, when the magnetic field decreases,
the resistance of nanocontact decreases continuously.
There is no jump during this period.
Since the pioneering MR measurements of the mechani-
cally formed nanocontacts by Garcıa et al. [1], The MR has
´
been attributed to the scattering of spin-polarized electrons
on DW formed in the constriction. Lepadatu and Xu [15]
gave the empirical relations between the percentage
changes in resistance and the cross-sectional area at the
constriction:
Dr
r
K
Sa
¼
,
Recent work by Yang et al. [6] has demonstrated little
MR effect, in which they eliminated magnetostriction, and
they also provide an independent confirmation that
magnetostriction may play a key role in the previous
reported huge BMR.
where S is the cross-sectional area at the constriction.
For Ni, the constant K and a take the values of 1.12 ꢂ 10ꢃ6
and 0.3, respectively. The resistance of the nanocontact
after electrodeposition is 1.75 O (see Fig. 2). Using d ¼

ARTICLE IN PRESS
Y.F. Zhu, G.Q. Di / Journal of Magnetism and Magnetic Materials 302 (2006) 82–85
84
↓+5 KOe
10
10
8
↓-5 KOe
↓+5 KOe
↓-5 KOe
↓ +5 KOe
10
9
8
8
MR = 477%
7
6
6
6
5
4
4
4
3
2
-300 -200 -100
0
100 200 300
2
2
H (Oe)
-8 -7 -6 -5 -4 -3 -2 -1
H (KOe)
0
1
2
3
4
5
0
400
800
1200 1600 2000 2400
Time (s)
(a)
(b)
Fig. 2. (a) Several typical MR curves recorded from the same Ni nanocontact. R0 ¼ 1:75 O, RH ¼ 10:1 O, MR ¼ 477%. The inset shows the details of the
MR curves at field value ranging between –300 Oe and +300 Oe. It demonstrates that the resistance of nanocontact drops to minimum while the magnetic
field is not zero. (b) The same MR curves are arranged by time order to demonstrate the unproducible nature of electrodeposited Ni nanocontact clearly.
20
18
16
14
12
10
8
20
18
16
14
12
10
8
MR = 104100%
6
6
4
4
2
2
↓+5 KOe
↓-5 KOe
0
0
↑0 Oe
↑0 Oe
-2
-2
-5
-4
-3
-2
-1
0
1
2
3
4
5
0
200
400
600
1000
800
H (KOe)
(a)
(b)
Time (s)
Fig. 3. (a) Another typical MR curve recorded from a Ni nanocontact, Rmin ¼ 16:3 O, Rmax ¼ 17 000 O, MR ¼ 104 100%. Rmax412 900 O, which
indicates the contact has broken. (b) The same MR curve is arranged by time order to show clearly. The arrows in (a) and (b) indicate a significant decrease
in the resistance of nanocontact, corresponding to the movement of atoms in the nanocontact (see the text for details).
4. Conclusion
tance effects of nanocontacts after the fabrication
were studied. Contrary to previous reports, in some
samples, there are discrete jumps during the measuring
process. We attribute them to the magnetic-field-induced
To summarize, nanocontacts between two bulk Ni wires
were fabricated by electrodeposition, the magnetoresis-

ARTICLE IN PRESS
Y.F. Zhu, G.Q. Di / Journal of Magnetism and Magnetic Materials 302 (2006) 82–85
85
electrodeposited Ni shortening and bulk Ni wire shrinking.
Using the mechanism proposed in this report, we can
explain the previously reported huge BMR, poor reprodu-
cibility and low stability of the electrochemically made
nanocontacts.
[4] S.H. Chung, M. Mun
J. Appl. Phys. 93 (2003) 7939.
[5] N. Garcıa, I.G. Saveliev, Y.-W. Zhao, A. Zlatkine. J. Magn. Magn.
˜
oz, N. Garcı
´
a, W. Egelhoff, R.D. Gomez,
´
Mater. 214 (2000) 7.
[6] C.-S. Yang, et al., J. Magn. Magn. Mater. 286 (2005) 186.
[7] N. Garcı
(2000) 3053.
[8] N. Garcıa, G.G. Qiang, I.G. Saveliev, Appl. Phys. Lett. 80 (2002)
1785.
´
a, H. Rohrer, I.G. Saveliev, Y.-W. Zhao, Phys. Rev. Lett. 85
Acknowledgments
´
[9] H.D. Chopra, S.Z. Hua, Phys. Rev. B 66 (2002) 020403(R).
[10] S.Z. Hua, H.D. Chopra, Phys. Rev. B 67 (2003) 060401(R).
[11] Erik B. Svedberg, Jonathan J. Mallett, et al., Appl. Phys. Lett. 84
(2004) 236.
The authors would like to acknowledge many useful
conversations with K.H. Chu, J.Y. Gu and D.H. Fan.
Thanks are extended to D. L. Sun for his laboratory
hospitality.
[12] S. Khizroev, et al., Appl. Phys. Lett. 86 (2005) 042502 (and references
therein).
[13] Ge Yi, Phys. Rev. B 69 (2004) 132405.
References
[14] S. Boussaad, N.J. Tao, Appl. Phys. Lett. 80 (2002) 2398.
[15] S. Lepadatu, Y.B. Xu, Phys. Rev. Lett. 92 (2004) 127201.
˜
[1] N. Garcıa, M. Munoz, Y.-W. Zhao, Phys. Rev. Lett. 82 (1999) 2923.
´
˜
[16] N. Garcıa, M. Munoz, G.G. Qian, H. Rohrer, I.G. Saveliev, Y.-W.
´
Zhao, Appl. Phys. Lett. 79 (2001) 4550.
[2] J.J. Versluijs, M.A. Bari, J.M.D. Coey, Phys. Rev. Lett. 87 (2001)
026601.
˜
[3] S.H. Chung, M. Munoz, N. Garcıa, W.F. Egelhoff Jr., R.D. Gomez,
´
[17] E.W. Lee, Rep. Prog. Phys. 18 (1955) 184.
[18] M. Viret, et al., Phys. Rev. B 66 (2002) 220401(R).
Phys. Rev. Lett. 89 (2002) 287203.