In the past decade, several strategies
have been proposed to direct the assembly
of nanocomponents from the bottom
up to fabricate two- and three-dimensional
integrated structures. In these
strategies, either the surface or the bulk
of a nanocomponent is functionalized to
facilitate positive and negative interactions
based on molecular, electrostatic,
magnetic, or capillary forces, to enable
the components to interact with one
another in a fluidic medium and form
thermodynamically stable structures.
However, in many cases, the assembled
structures are not well bonded. This
paper summarizes results on directing
the assembly of metallic rod-shaped
(30200 nm diameter) components
with one another and with substrates to
form structures that can be bonded by
adhesive or solder. The methodology can
be adapted with other self-assembling
strategies to form mechanically stable,
and in certain instances electrically
conductive, assemblies composed of
nanoscale componenents.
INTRODUCTION
In recent years, the extreme miniaturization
of microelectronic, photonic, and
micro-electromechanical systems has
pushed the capabilities of conventional
top-down microfabrication to its limits in
terms of cost-effective mass production
of devices and integrated systems. New
nanofabrication techniques are needed
to address the challenges of decreasing
dimensions in order to enable the era of
nanotechnology.
One promising strategy
is to direct the assembly of nanoscale
components to form two-dimensional
(2-D) or three-dimensional (3-D) structures
from the bottom up.110 The directed
assembly of engineered structures is
inspired by biological self-assembly;
nature can mass produce complex 2-D
and 3-D structures with sizes ranging
from the sub-nanometer to the millimeter
and beyond. The nanocomponents used
in directed assembly are fabricated using
either conventional nanolithography11,12 or non-lithographic methods including
chemical vapor deposition,13,14 electrodeposition
in templates,1517 and molecule-templated
growth.18,19
The nanocomponents are engineered
so they are capable of interacting with
one another, usually in a liquid medium,
to form organized integrated structures.
The complexity of the assembled structure
can be increased by utilizing complex
components whose surfaces have
been engineered with different recognition
sites that facilitate a variety of
simultaneous orthogonal interactions of
different magnitudes. Several interaction
forces have been used to direct the
assembly of nanocomponents including
molecular recognition,2023 electrostatic
forces,24,25 dielectrophoresis,26,27 magnetic
forces,2830 or surface-tension-based
forces.31,32
Although the aforementioned strategies
have succeeded in directing the
assembly of nanocomponents into 2-D
and 3-D integrated structures, in many cases, the structures formed are not well
bonded to one another (i.e., the assemblies,
although held together in the fluidic
medium in which they are assembled, fall
apart when removed from the medium
or during mild sonication).33,34 Also,
in directed assembly between rigid
nanocomponents, the strength and the
extent of binding is proportional to the
overlap area at the binding site between
components.
|
"The directed assembly of engineered structures is inspired by biological self-assembly; nature can mass produce complex 2-D and 3-D structures with sizes ranging from the sub-nanometer to the millimeter and beyond." |
|
Any local roughness of the
components (especially when the size
approaches 100 nm) reduces the effective
binding contact area due to asperities,
and consequently decreases the strength
and extent of binding. Hence, assemblies
often consist of only a few bonded nanocomponents,
and large-scale integration
is not possible. It should be noted that
in biological self-assembly, most of the
components utilized in the assemblies
are soft and deformable. This allows
the mating surfaces to conform to one
another, resulting in large contact areas
for optimum binding.
This study demonstrated the fabrication
of 2-D and 3-D structures composed
of rod-shaped metallic nanocomponents
(referred to as nanowires) that were
bonded to each other and to substrates
using a curable adhesive or solder.
Liquid layers of an organic adhesive or
molten solder on specific regions of the
nanowire facilitated binding between
components. The adhesive and solder
were subsequently hardened by curing
(polymerization) and cooling, respectively.
In the case of the adhesive, the
2-D and 3-D structures formed survived
mild sonication and could be taken out of
the fluidic medium without disruption. In the case of solder, it was possible to
form ohmic low-resistance contacts that
point to the feasibility of using these
joints as nanoscale electrical contacts.
FABRICATION AND
SURFACE ENGINEERING
OF NANOWIRES
The use of nanowires as components
for directed assembly was motivated by
the fact that, in a relatively straight forward manner, large numbers of patterned
nanowires could be fabricated ranging
in diameter from 30200 nm and with
lengths up to 20 ΅m.1517 Electrodeposition
in nanoporous alumina or polycarbonate
membranes (Figure 1a) was
used to fabricate approximately 108109
nanowires (per membrane, Figure 1b).
First, a thin seed layer of metal was
evaporated on one side of the membrane
to seal the pores. Electrodeposition of
the wire was carried out sequentially
using different electrolytes to pattern
the wire with different metals along its
length (Figure 1c). Since the nanowire
contained different segments, it was
possible to selectively functionalize parts
of the wire using organic molecular self assembly
(Figure 1d).35,36 Although
the results in this article describe the use
of metallic nanowires, the same strategy
can be used to fabricate nanowires containing
semiconducting and insulating
segments15,16,37 to build electronically
functional nanocomponents such as
diodes or transistors.37,38
|
"Although the aforementioned strategies have succeeded in directing the assembly of nanocomponents into 2-D and 3-D integrated structures, in many cases, the structures formed are not well bonded to one another." |
|
Since the surface-area-to-volume ratio
of nanowires is high, there was a natural
tendency of the wires to adhere to one another due to van der Waals forces.
During dissolution of the nanoporous
membrane containing gold wires, large
bundle-like structures were observed
(Figure 2a and 2b). Although these
bundles are well ordered, they break apart
readily on sonication and hence are of
limited use in real-world applications.
Directed Assembly of Nanowires
Two forces were used to direct assembly
of the nanowires: surface-tension based
forces and magnetic forces.
Surface-tension-based assembly31,32
involved the modification of the surface
energy of specific segments of the
nanowires using hydrophobic organic
molecules that attached preferentially
to specific segments; precipitation of a
hydrophobic liquid layer on the modified
segments; and agitation of the nanowires
in a hydrophilic medium to facilitate favorable interactions between wires
and direct assembly.
When nanowires
patterned with hydrophobic liquid layers
collided with one another in a hydrophilic
liquid, there was a tendency of the liquid
layers on different nanowires to fuse
with one another on contact in order to
minimize their surface free energy. This
surface tension force between liquid
layers on colliding nanowires was large
enough to hold the wires together in the
liquid. Since this assembly involved
binding between liquid layers (that were
soft and deformable) patterned on the
nanowires, the roughness of the wires did
not hamper binding and it was possible
to get large-scale integration.
The magnetic-force-based assembly
of nanowires was relatively straightforward.2830 When nanowires containing
nickel segments (nickel segment length
>> wire diameter) were tumbled in a
magnetic field they bound to one another
and aligned in the direction of the magnetic
field (Figure 3a). The assembly of
segmented nickel nanowires was also
directed on patterned nickel substrates
in magnetic fields. The wires aligned
between nickel pads and the number of
wires between adjacent pads could be
controlled to some extent by controlling
the nickel pattern on the substrate, the
nanowire concentration, and the strength
of the magnetic field (Figure 3b and 3c).
Although the nanowires could be positioned
on the patterned contact pads using
magnetic fields, the nanowires were not
strongly bound and were disrupted by
mild sonication on removal from the
magnetic field.
Forming Bonded Nanowire
Structures
Two- and three-dimensional assemblies
of nanowires that were held together
by adhesive or solder were fabricated.
The nanowires utilized were composed
of either all gold or Au-Ni-Au segments.
When the wires were immersed
in a hydrophobic thiol (hexadecane
thiol [HDT]) solution (in ethanol), the
HDT bound preferentially to the gold
segments on the nanowires, but did
not adhere well to the nickel segments,
probably due to the formation of an
oxide layer on nickel during processing.
This allowed segments to be selectively
patterned on the same nanowire to be
either hydrophobic (gold segment) or hydrophilic (nickel segment). To
facilitate self-assembly, a small drop of a
dilute solution of a hydrophobic polymerizable
adhesive that was composed
of a monomer (lauryl methacrylate), a
crosslinker (1,6-hexanediol diacrylate),
and a polymerization initiator (either a
thermal initiator benzoyl peroxide or a
photo-initiator, benzoin isobutyl ether)
was introduced into a vial containing
nanowires in ethanol. When water was
added to the nanowire-adhesive-ethanol
suspension, the hydrophobic adhesive
preferentially precipitated on the hydrophobic
gold segments.
|
"Liquid layers of an organic adhesive or molten solder on specific regions of the nanowire facilitated binding between components." |
|
On agitation, the
nanowires (with gold segments coated
with adhesive) aggregated. Depending
on how the assembly was carried out,
either 2-D assemblies were obtained at
the water-air interface or 3-D assemblies
were obtained in the bulk. After the
assembly, the monomeric adhesive was
cured (polymerized) using ultraviolet
light or heat to harden the adhesive and
bond the assemblies. Structures with different
topologies were fabricated using
this methodology: 2-D rafts and 3-D
bundles with pure gold wires (Figure 4a,
b, and c) and 2-D Au-Ni-Au networks
with predominantly end-to-end connectivity
between gold segments (Figure
4d and e).39 On polymerization of the
adhesive, the self-assembled aggregates
formed were mechanically bonded and
survived mild sonication.
Nanowires were also soldered to
one another and to substrates by either
incorporating a solder segment within
the nanowire itself or by directing the
assembly of nanowires on substrates that
were coated with a thin film (100 nm)
of solder. Tin/lead (Sn/Pb, most widely
used) solder or tin (a model lead-free
solder) solder were deposited electrolytically.
Unlike macroscopic soldering,
since the solder volume present on the
nanowires or substrate was very small
it was crucial to protect the nanosolder
surface from corrosion, intermetallic
diffusion, and oxidation. During dissolution
of the membrane in NaOH, the tin
solder segments were readily corroded
(Figure 5a).
To minimize corrosion, a
small quantity of the organic molecule
benzatriazole (BTA) was added to the
solution during membrane dissolution.
The adsorption of BTA molecules on
the surface of the nanowire protected
the nanowire during dissolution and significantly improved corrosion resistance
(Figure 5a). The impact of intermetallic
diffusion was dramatic in nanowires
that contained gold segments directly in
contact with tin solder segments (Figure
5b). In this case, on heating above the
melting point of the solder (reflow), the
intermetallic diffusion between gold and
tin was very dramatic as gold diffuses
rapidly into tin. This intermetallic diffusion
was minimized for optimum solder
bonding by including a nickel spacer
segment between the gold contacts and
the tin solder.40
Any oxidation of the solder during
reflow greatly reduces electrical conductivity
of the joint formed. In these
experiments, special care was taken to
reduce oxidation during reflow using
a tube furnace that was purged with
nitrogen before, during, and after reflow.
In the presence of air, the solder joint
was extensively oxidized and it was
difficult to form ohmic low-resistance
electrical contacts. Figure 5c shows
scanning-electron microscopy (SEM)
energy-dispersive spectroscopy (EDS)
maps of one such oxidized joint at room
temperature after solder reflow, showing
a large presence of oxygen in the joint
region.
Au-Ni-Au-Sn nanowires containing
tin solder segments built into the wire
were soldered (Figure 6a). A drop of a
suspension of wires in ethanol was placed
on a substrate and allowed to evaporate.
Upon evaporation, the wires tended to
aggregate at the edges of the drop and
several wires lay near one another. When
a silicon chip containing a high density
of wires was heated in the tube furnace,
the solder from adjacent wires lying next
to one another fused into one another to
form solder joints. On cooling, some of
the wires were soldered to each other.
To electrically characterize the solder
joints, contact pads were lithographically
aligned at the ends of the solder-bonded
wires (Figure 6b).40 Current-voltage
curves of the solder joints were measured,
and it was observed that it was possible to
form low-resistance ohmic solder joints.
The measured resistances of soldered
nanowires were in the range of 30200
Ω for 200 nm diameter wires.40 Au-Ni-Au nanowires were soldered
onto nickel contact pads (on an oxidecoated
silicon chip) that were covered
with a thin layer of copper and Sn/Pb solder (solder ~100 nm thick). Nanowires
used were composed of ferromagnetic
nickel central segments terminated with
gold contacts.41 The wires were first
positioned on patterned nickel contact
pads using magnetic fields as described
earlier. After assembly, the wires were
heated in a tube furnace in the presence
of a nitrogen gas purge. The solder
reflowed and bonded the nanowires to the
contact pads (Figure 6c). The resistance
between nanowire bridged contact pads
dropped from 300106 Ω before reflow
to about 13 Ω after reflow (for 200 nm
diameter wires), indicating the formation
of solder contacts with good electrical
conductivity.41
CONCLUSIONS
This study demonstrated that it is possible
to bond nanowires to one another
and to substrates using adhesives or
solder. The strategies outlined can be
extended to other nanocomponents to
form mechanically stable assemblies
fabricated from the bottom up. It is also
possible to use adhesive or solder bonding
as a final step in the directed assembly
of nanocomponents assembled by other
forces such as molecular recognition or
electrostatic forces. The authors believe
that these strategies of forming mechanical
and in some cases electrically conductive
bonds between nanocomponents
will find utility in nanomanufacturing
and in developing structures engineered
from the bottom up for use in real-world
applications.
ACKNOWLEDGEMENTS
We acknowledge financial support
from the National Science Foundation
(CAREER/NSF-DMII) and the American
Chemical Society Petroleum Research
Fund. We thank Dr. Kenneth Livi for
assistance with SEM EDS element
analysis.
REFERENCES
1. S.I. Stupp et al., Science, 276 (1997), pp. 384389.
2. D.H. Gracias et al., Science, 289 (2000), pp. 11701172.
3. Y. Huang et al., Science, 291 (2001), pp. 630633.
4. D.H. Gracias et al., Adv. Mater., 14 (2002), pp.
235238.
5. H.O. Jacobs et al., Science, 296 (2002), pp. 323
325.
6. N.I. Kovtyukhova and T.E. Mallouk, Chem.A
European J., 8 (2002), pp. 43544363.
7. J.M. Lehn, Science, 295 (2002), pp. 24002403.
8. G.M. Whitesides and B. Grzybowski, Science, 295
(2002), pp. 24182421.
9. H. Fan et al., Science, 304 (2004), 567571.
10. Y. Lin et al., Nature, 434 (2005), pp. 5559.
11. J.M. Gibson and J. Murray, Phys. Today, 50 (1997),
pp. 5661.
12. J.A. Stroscio and D.M. Eigler, Science, 254 (1991),
pp. 13191326.
13. Z.F. Ren et al., Science, 282 (1998), pp. 11051107.
14. Y. Wu et al., Nature, 430 (2004), pp. 6165.
15. C.R. Martin, Science, 266 (1994), pp. 19611966.
16. C.R. Martin, Chem. Mater., 8 (1996), pp. 17391746.
17. B.R. Martin et al., Adv. Mater., 11 (1999), pp.
10211025.
18. C.J. Murphy, Science, 298 (2002), pp. 21392141.
19. Y. Sun and Y. Xia, Science, 298 (2002), pp. 2176
2179.
20. C.A. Mirkin, Inorg. Chem., 39 (2000), pp. 2258
2272.
21. H. Mattoussi et al., J. Am. Chem. Soc., 122 (2000),
pp. 1214212150.
22. J.K.N. Mbindyo et al., Adv. Mater., 13 (2001), pp.
249251.
23. A.K. Salem et al., Adv. Mater., 16 (2004), pp.
268271.
24. H.O. Jacobs, S.A. Campbell, and M.G. Steward,
Adv. Mater., 14 (2002), pp. 15531557.
25. B.A. Grzybowski et al., Nature Mater., 2 (2003), pp.
241245.
26. K.D. Hermanson et al., Science, 294 (2001), pp.
10821086.
27. S. Evoy et al., Microelectronic Eng., 75 (2004), pp.
3142.
28. L. Bauer et al., Nano Lett., 1 (2001), pp. 155158.
29. C.J. Love et al., J. Am. Chem. Soc., 125 (2003), pp.
1269612697.
30. C.M. Hangarter and N.V. Myung, Chem. Mater., 17
(2005), pp. 13203124.
31. T.D. Clark et al., J. Am. Chem. Soc., 123 (2001),
pp. 76777682.
32. R.R.A. Syms et al., J. Microelectromechan. Sys.,
12 (2003), pp. 387417.
33. Y. Lin et al., Science, 299 (2003), pp. 226229.
34. S. Park et al., Science, 303 (2004), pp. 348351.
35. A. Ulman, Chem. Rev., 96 (1996), pp. 15331554.
36. P.F. Nealey et al., Mol. Electron. (1997), pp. 343
367.
37. S. Park, S.W. Chung, and C.A. Mirkin, J. Am. Chem.
Soc., 126 (2004), pp. 1177211773.
38. N.I. Kovtyukhova, B.K. Kelley, and T.E. Mallouk, J.
Am. Chem. Soc., 126 (2004), pp. 1273812739.
39. Z. Gu, Y. Chen, and D.H. Gracias, Langmuir, 20
(2004), pp. 1130811311.
40. Z. Gu et al., Reflow and Electrical Characteristics
of Nanoscale Solder, Small (2005), In press.
41. H. Ye et al., Integrating Nanowires with Substrates
Using Directed Assembly and Nanoscale Soldering, IEEE Trans. Nanotech. (2005), In press.
Zhiyong Gu (postdoctoral fellow), Hongke Ye
(postdoctoral fellow), and David H. Gracias
(assistant professor) are with the Department of
Chemical and Biomolecular Engineering at Johns
Hopkins University in Baltimore, Maryland.
David Gracias is also with the Department of
Chemistry at Johns Hopkins University.
For more information, contact David H. Gracias,
Department of Chemical and Biomolecular
Engineering and Department of Chemistry,
Johns Hopkins University, 3400 N. Charles Street,
Baltimore, MD 21218; (410) 516-5284, fax: (410)
516-5510, e-mail: dgracias@jhu.edu.
|