Signatures of quantum transport in self-assembled epitaxial nickel suicide nanowires

Article abstract of DOI:10.1063/1.1769583

The electrical properties of self-assembled epitaxial NiSi₂ nanowires (NW) formed on Si substrates were measured. The quantum corrections due to weak antilocalization and electron-electron interactions were also found. The analysis of magnetore

Full text of DOI:10.1063/1.1769583

Signatures of quantum transport in self-assembled epitaxialnickel silicide nanowires
J.-F. Lin, J. P. Bird, Z. He, P. A. Bennett, and D. J. Smith
Citation: Applied Physics Letters 85, 281 (2004); doi: 10.1063/1.1769583
View online: http://dx.doi.org/10.1063/1.1769583
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/85/2?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Spin relaxation in a germanium nanowire
J. Appl. Phys. 100, 044303 (2006); 10.1063/1.2230012
Nitride-mediated epitaxy of self-assembled Ni Si 2 nanowires on (001)Si
Appl. Phys. Lett. 87, 253111 (2005); 10.1063/1.2149970
Quantum interference effect in single Pt ( Ga ) ∕ C nanowire
Appl. Phys. Lett. 87, 182112 (2005); 10.1063/1.2125108
Self-assembled nanobridge formation and spontaneous growth of metal-induced nanowires
Appl. Phys. Lett. 86, 253101 (2005); 10.1063/1.1949291
Self-assembly of spatially separated silicon structures by Si heteroepitaxy on Ni disilicide
J. Appl. Phys. 91, 6077 (2002); 10.1063/1.1469205
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
153.106.95.14 On: Tue, 09 Dec 2014 19:39:51

APPLIED PHYSICS LETTERS
VOLUME 85, NUMBER 2
12 JULY 2004
Signatures of quantum transport in self-assembled epitaxial
nickel silicide nanowires
a)
J.-F. Lin and J. P. Bird
Nanostructures Research Group, Department of Electrical Engineering, Arizona State University,
Tempe, Arizona 85287-5706
Z. He
Science and Engineering of Materials Program, Arizona State University,
Tempe, Arizona 85287-1504
P. A. Bennett
Department of Physics and Astronomy, Arizona State University, Tempe, Arizona 85287-1504
D. J. Smith
Center for Solid State Science and Department of Physics and Astronomy, Arizona State University, Tempe,
Arizona 85287-1704
(
Received 13 February 2004; accepted 18 May 2004)
We have measured the electrical properties of self-assembled epitaxial NiSi nanowires (NWs)
2
formed on Si substrates. We find quantum corrections due to weak antilocalization and electron–
electron interactions. Analysis of the magnetoresistance indicates that electron phase coherence in
the NWs is limited by Nyquist dephasing below 10 K, and by electron–phonon scattering at higher
temperatures. The phase-breaking and spin–orbit scattering lengths are found to be ϳ45 nm and
3
–7 nm, at 4.2 K, respectively, similar to reports for thin NiSi films. © 2004 American Institute of
2
Physics. [DOI: 10.1063/1.1769583]
There are many barriers to the continued downscaling of
microelectronic devices. A bottom-up approach to fabrica-
tion, based on self-assembling nanostructures, may offer a
solution to these problems. In this regard, recent work on
self-assembled epitaxial silicide nanowires (NWs) on silicon
deposition ͑Ni/Au:50/200 Å͒, native oxide at the NW-
contact areas was removed by an HF-acid dip. We focus on
studies of two NWs (labeled NW1 and NW2), whose dimen-
sions (Table I) were determined by atomic force microscopy
(AFM) and TEM. We have also made four-terminal contact
pads, but these were bridged by several parallel NWs. These
measurements nonetheless allowed us to infer a contact re-
sistance of 1.5 k⍀ to each NW, much smaller than the total
resistance (Table I). The resistance between contacts un-
bridged by NWs exceeded 10 M⍀ at 4.2 K. Constant cur-
1
–10
has attracted much interest.
NW formation was discov-
1
ered for DySi2, and later found for several rare-earth and
4
–10
transition metals.
These structures may ultimately have
practical uses as nanoscale interconnects, sensor elements, or
11–13
as active devices, in analogy with carbon nanotubes.
Sil-
icide NWs offer unique advantages over other NW systems;
since they are perfect single crystals and are highly compat-
ible with silicon processing. In this letter, we present trans-
port measurements of such epitaxial NWs. Their electrical
properties are reminiscent of “dirty metals” and show pro-
nounced quantum corrections due to weak-antilocalization
(WAL) and electron–electron interactions (EEI). An analysis
of these features allows us to determine the critical length
scales for electron transport in the NWs.
NiSi NWs were formed on n-type Si͑111͒ substrates
10 ⍀ cm͒ miscut 8° in the [112] direction. Growth was per-
2
͑
formed in an ultrahigh vacuum chamber, after flashing the
substrate for 30 s at 1250 °C. Ni was deposited at a rate of 1
monolayer every 2 min, and Joule heating was used to main-
tain the substrate at 500 °C during this process. Growth re-
sults in NWs with typical dimensions of 5 nm thick, 15 nm
wide, and 1.5 ␮m long. A transmission electron microscope
(TEM) cross-sectional micrograph of one NW is shown in
the upper part of Fig. 1 and reveals a nearly perfect crystal-
line structure.
Two-terminal contacts to the NWs were formed by
electron-beam lithography and liftoff (Fig. 1). Prior to metal
FIG. 1. Upper panel: TEM cross-sectional micrograph of a NiSi2 NW.
Lower panel: AFM image of NW1 and its two electrodes.
a)
Electronic mail: bird@asu.edu
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
003-6951/2004/85(2)/281/3/$22.00 281 © 2004 American Institute of Physics
53.106.95.14 On: Tue, 09 Dec 2014 19:39:51
0
1

282
Appl. Phys. Lett., Vol. 85, No. 2, 12 July 2004
Lin et al.
TABLE I. Parameters of the two NWs studied here.
L
W
t
lso
D
␮@30 K
2
R@4 K R@30 K
Wire ͑nm͒ ͑nm͒ ͑nm͒ ͑nm͒ ͑cm2 s−1
͒
͑cm /Vs͒ ͑k⍀͒ ͑k⍀͒
NW1 790
NW2 740
22
14
6
6
7.2
3.5
0.7
0.8
54
53
51
71
49
66
rents ͑ϳ50 nA͒ and lock-in detection ͑11 Hz͒ were used to
measure the NW magnetoresistance (MR) in a variable-
temperature insert.
Figure 2 shows the MR of NW1 at several temperatures
[each curve has been normalized to its zero-field resistance,
R͑0͒]. A positive MR is seen in all traces, and its magnitude
increases with decreasing temperature. This behavior is remi-
niscent of WAL, and very similar MR curves were obtained
in studies of NW2. In one-dimension ͑1D͒, WAL gives rise
FIG. 3. Temperature dependence of l . Filled circles: NW1. Open circles:
␸
1
4,15
to a MR:
NW2. Filled squares: Results (from Ref. 16) Inset: MR of wire NW1 at
2
3
0 K. Solid line is a fit to the form R͑B͒ϰB .
−
1/2
−1/2
⌬
R
R
3 1 4 1
1
1 1
1
=
ͩ
+
+
ͪ
−
ͩ
+
ͪ
.
ͫ
ͬ
2
2
2 l2
2
R
LRT 2 l2 3 l
l
l
␸
so
h
␸
h
We have also attempted to fit the MR using the form for
1
4
͑
1͒
WAL in two-dimensions ͑2D͒. For NW1, the 2D form fits
the experiment well at 15 and 20 K, using the same values of
W and lso listed in Table I (see, for example, the data for
2
2
0.5
ͱ
Here, L is the wire length, R = 8␲ប /e , l =͑D␶ ͒ is the
phase-breaking length, ␶ is the phase-breaking time, and D
T
␸
␸
␸
2
0 K in the right-hand side inset to Fig. 4). At lower tem-
0
.5
is the diffusion constant. l =͑D␶ ͒ is the spin–orbit scat-
so
so
peratures, however, reasonable 2D fits can only be obtained
by allowing W to vary significantly from its physical value.
The suggestion, therefore, is that this NW exhibits a cross-
over from 1D to 2D WAL, which is consistent with the as-
sociated variation of l␸ with temperature. Figure 3 shows
tering length, and ␶ is the spin–orbit scattering time. lh
so
0
.5
=
=
͑D␶ ͒
is related to the magnetic field by D␶h
3ប /e AB , where A is the cross-sectional area of the wire
h
2
2
2
and B is the magnetic field applied normal to its plane. A
Wϫt, where W is the width of the wire and t is its thick-
=
that, in NW1, l becomes comparable to W when the tem-
␸
ness, which we take here to be 6 nm from TEM studies. The
solid lines through the lower three data sets of Fig. 2 are
single-parameter fits to the form of Eq. (1), which were ob-
tained in the following way. First, we performed three-
parameter ͑W,l ,l ͒ fits of the MR to Eq. (1). As expected,
perature is increased to ϳ20 K. In NW2, the MR is well
described by Eq. (1) over the entire range of temperature,
and we attribute the absence of a 1D–2D crossover in this
NW to its smaller width ͑ϳ14 nm͒.
so
␸
The data of Fig. 2 show a clear suppression of quantum
coherence with increasing temperature. At 30 K, the MR
the values of W and lso obtained in this way were insensitive
to temperature, and our estimate for W was in good agree-
ment with that obtained from AFM studies. The spin–orbit
scattering lengths inferred for the NWs are listed in Table I,
and are consistent with studies of epitaxial NiSi films ͑l
2
shows the classical form ⌬R͑B͒/R͑0͒=͑␮B͒ , where ␮ is the
mobility (see Fig. 3, inset). By fitting the MR to this form,
we infer the wire mobility, and in Table I we list the values
of ␮ for the two wires. These values are much smaller than
those quoted for crystalline metals at low temperatures, and
2
so
1
6
ϳ19 nm͒. With W and lso established in this way, we then
performed the “quasi-one-parameter” fits shown in Fig. 2,
1
7
are more reminiscent instead of metallic alloys.
using l as the fit parameter.
␸
In Fig. 3, we plot the temperature dependence of l for
␸
NW1 and NW2 and see that the two sets of data fall almost
on top of each other. We also plot the variation of this pa-
1
6
rameter, reported previously for epitaxial NiSi2 films.
These data follow the T1 dependence predicted for
electron–phonon scattering, and the variation of l␸ in the
NWs appears consistent with this effect above 10 K. At
18
1
/3
lower temperatures, however, l varies as T , as expected
␸
1
9
for Nyquist dephasing in 1D:
2
4
␯
͑E ͒␴A ប
F
T−2/3
␶N
=
.
͑2͒
ͫ
ͬ
2
2
B
2
e k
Here, ␯͑E ͒ is the density of states at the Fermi level and ␴
is the conductivity. From the T variation in Fig. 3, and
with the value of D listed for NW1 in Table I, we use Eq. (2)
F
1
/3
4
7
−1
−3
to infer ␯͑E ͒=3.5ϫ10
J
m . This is consistent with the
F
4
6
−1
−3
estimate ␯͑E ͒=7.0ϫ10
J
m , obtained by direct appli-
FIG. 2. Temperature-dependent MR of NW1. Successive curves are shifted
F
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
2
up from each other by 0.03. Solid lines are one-parameter ͑l ͒ fits to Eq. (1).
cation of the relation ␴=␯͑E ͒e D. It also agrees well with
␸
F
1
53.106.95.14 On: Tue, 09 Dec 2014 19:39:51

Appl. Phys. Lett., Vol. 85, No. 2, 12 July 2004
Lin et al.
283
2
2
of CoSi2, while ϳ60 ␮⍀ cm has been reported for 3 nm
1
6
thick NiSi films. von Kanel measured the systematic in-
2
fluence of film thickness on surface scattering and found a
surface resistivity of 100 ␮⍀ cm in 1 nm thick CoSi films.
2
Our NWs have resistivities of ϳ800 ␮⍀ cm and we believe
that this is due to interface scattering due to native oxide at
the free NW surface, which may be reduced by passivation.
2
3
Indeed, Lee et al. have studied free-standing NiSi /C com-
2
posite NWs and obtained 350 ␮⍀ cm at 4.2 K.
To conclude, we have presented transport measurements
of epitaxial silicide NWs. Electron coherence is limited by
Nyquist dephasing below 10 K, and by electron–phonon
scattering at higher temperatures. The phase-breaking and
spin–orbit scattering lengths are found to be comparable to
those in thin NiSi films, although the nanoscale dimensions
2
of the NWs allow the observation of quantum transport at
high temperatures (at least 30 K).
This work was supported by the Department of Energy
Grant No. DE-FG03-01ER45920), the National Science
FIG. 4. Main panel: R͑T͒ for NW1 at 0 and 8 T. Solid line through the 8 T
data is a one-parameter fit ͑␭͒ to Eq. (3). Left inset: 8 T data in the range
from 4–7 K, as a function of T . Solid line is a one-parameter ͑D͒ fit to
Eq. (4). Right inset: MR of NW1 at 20 K. Solid (dotted) line is a fit to the
two-dimensional (one-dimensional) WAL MR. The (two-dimensional) fit
gives better agreement.
(
1
/2
Foundation (Grant No. ECS 0304682), and the Office of Na-
val Research (Grant No. N00014-98-0594). One of the au-
thors (J.F.L.) is grateful to Professor J. J. Lin of National
Chiao Tung University, Taiwan, for helpful discussions and
also acknowledges the kind hospitalilty of Professor Y.
Ochiai of Chiba University, Japan.
4
7
−1 −3
␯
͑E ͒= ϳ10 J m , calculated for a free-electron system
F
with a Fermi energy of 5 eV.
1
Figure 4 shows the temperature dependence of the resis-
tance of NW1 at B=0 and 8 T. At 8 T, the resistance closely
follows a logarithmic temperature dependence. Since this
magnetic field should be sufficient to quench WAL, the loga-
rithmic variation is suggestive of 2D EEI. The solid line
through the 8 T data is a one-parameter ͑␭͒ fit to the theory
C. Preinesberger, V. S. R. Kalka, and M. Dahne-Prietsch, J. Phys. D 31,
L43 (1998).
C. Preinesberger, S. K. Becker, S. Vandre, T. Kalka, and M. Dahne, J.
Appl. Phys. 91, 1695 (2002).
B. Z. Liu and J. Nogami, J. Appl. Phys. 93, 593 (2003).
J. Nogami, B. Z. Liu, M. V. Katkov, C. Ohbichi, and N. O. Birge, Phys.
Rev. B 63, 233305 (2001).
Y. Chen, D. A. A. Ohlberg, and R. S. Williams, J. Appl. Phys. 91, 3213
2
3
4
2
0,21
5
of EEI in 2D:
6(
2002).
R͑T͒ − R͑To͒
R͑To͒
e2
2␲2ប
T
Z. He, M. Stevens, D. J. Smith, and P. A. Bennett, Appl. Phys. Lett. 83,
5292 (2003).
7
=
−
␭R͑T͒lnͩ ͪ
.
͑3͒
T
o
Z. He, M. Stevens, D. J. Smith, and P. A. Bennett, Surf. Sci. 524, 148
8(
2003).
Here, ␭ is a constant ͑0ഛ␭ഛ1͒ describing the effects of
Coulomb screening and T is an arbitrary reference tempera-
Z. He, D. J. Smith, and P. A. Bennett (unpublished).
M. Stevens, Z. He, D. J. Smith, and P. A. Bennett, J. Appl. Phys. 93, 5670
o
9
ture. From the 8 T data in Fig. 4, we infer ␭=0.067, within
the range expected from theory. At 0 T, the resistance varia-
tion deviates from a ln T dependence below 20 K, as the
WAL comes into play as we have discussed already.
0(
2003).
1
J. L. McChesney, A. Kirakosian, R. Bennewitz, J. N. Crain, J.-L. Lin, and
F. J. Himpsel, Nanotechnology 13, 545 (2002).
1
1
Y. Cui, Q. Wei, H. Park, and C. M. Lieber, Science 293, 1289 (2001).
Y. Huang, X. Duan, Q. Wei, and C. M. Lieber, Science 291, 630 (2001).
J. Appenzeller, R. Martel, V. Derycke, M. Radosavljevic, S. Wind, D.
Neumayer, and P. Avouris, Microelectron. Eng. 64, 391 (2002).
B. L. Altshuler and A. G. Aronov, Pis’ma Zh. Eksp. Teor. Fiz. 33, 515
12
Below ϳ7 K, the resistance at 8 T in both NWs appears
13
1
/2
to follow the T dependence predicted for EE1 in 1D (Fig.
2
0,21
14
4, left-hand side inset):
15(
1981); [JETP Lett. 33, 499 (1981)].
J. J. Lin and N. Giordano, Phys. Rev. B 35, 545 (1987).
E. Jentzsch, R. Schad, S. Heun, and M. Huzler, Phys. Rev. B 44, 8984
e2 R͑T͒LT, LT =
បD
R͑T͒ − R͑To͒
=
ͱ
.
͑4͒
16
R͑To͒
2ប L
k T
B
7(
1991).
1
Equation (4) valid at temperatures for which W is less than
G. K. White, Experimental Techniques in Low-Temperature Physics
18(
Oxford University Press, Oxford, UK, 1968).
G. Bergmann, Phys. Rep. 107, 1 (1984).
B. L. Altschuler, A. G. Aronov, and D. E. Khmel’nitskii, J. Phys. C 15,
the thermal length L . By fitting the low-temperature resis-
T
tance variation to Eq. (4), we can estimate the diffusion con-
stant in the two NWs. Our results are listed in Table I and
imply a thermal diffusion length of ϳ12 nm at 4.2 K, com-
parable to the width of the NWs. It therefore seems plausible
that, at the lowest temperatures, the signature of 1D EE1
should be observed in the resistance.
It is interesting to compare the properties of these NWs
to those of bulk and thin-film silicides. Resistivities of
Ͻ1 ␮⍀ cm have been reported for high-purity bulk crystals
1
2
2
9
0
1
7367 (1982).
B. L. Altshuler, A. G. Aronov, and P. A. Lee, Phys. Rev. Lett. 44, 1288
(1980).
B. L. Altshuler and A. G. Aronov, in Electron–Electron Interactions in
Disordered Systems, edited by M. Pollack and A. L. Efros (North–
Holland, Amsterdam, 1985), p. 1.
H. von Kanel, Mater. Sci. Rep. 8, 193 (1992).
K. S. Lee, Y. H. Mo, K. S. Nahm, H. W. Shim, E. K. Suh, J. R. Kim, and
J. J. Kim, Chem. Phys. Lett. 384, 215 (2004).
22
23
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
53.106.95.14 On: Tue, 09 Dec 2014 19:39:51
1