Copper nanowires within the central channel of tobacco mosaic virus particles

Article abstract of DOI:10.1016/j.electacta.2006.04.007

Simple electrochemical deposition techniques can produce highly defined metal nanostructures in templates. Electroless deposition (ELD) can be effectively used for depositing metals on insulators such as biological or plastic surfaces. With biomolecular templates, metallization methods are often restricted to mild reductions, and the deposition of copper at pH values above 12 is usually not applicable. We produced copper nanowires of 3 nm in diameter and up to 150 nm in length by electroless deposition within the 4 nm wide channel of tobacco mosaic virus (TMV) particles. We employed a low pH (～7.5) copper electroless deposition solution that is compatible with biomolecules. The fabrication process of the nanowires is based on sensitization of tobacco mosaic virus with Pd(II) prior to the electroless deposition. We analyzed the chemical composition of the nanowires by energy-filtering transmission electron microscopy, and used the method also for nickel and cobalt nanowires deposited within the viral channel.

Full text of DOI:10.1016/j.electacta.2006.04.007

Electrochimica Acta 51 (2006) 6251–6257
Copper nanowires within the central channel of
tobacco mosaic virus particles
S. Balcia, A.M. Bittnera,∗, K. Hahnb, C. Scheub,1, M. Kneza,2
,
A. Kadric, C. Wegec, H. Jeskec, K. Kerna
a Max-Planck-Institut fu¨r Festko¨rperforschung, Heisenbergstr. 1, D-70569 Stuttgart, Germany
b Max-Planck-Institut fu¨r Metallforschung, Heisenbergstr. 3, D-70569 Stuttgart, Germany
c Biologisches Institut, Universita¨t Stuttgart, Pfaffenwaldring 57, D-70550 Stuttgart, Germany
Received 8 March 2006; accepted 6 April 2006
Available online 19 May 2006
Abstract
Simple electrochemical deposition techniques can produce highly defined metal nanostructures in templates. Electroless deposition (ELD) can
be effectively used for depositing metals on insulators such as biological or plastic surfaces. With biomolecular templates, metallization methods
are often restricted to mild reductions, and the deposition of copper at pH values above 12 is usually not applicable. We produced copper nanowires
of 3 nm in diameter and up to 150 nm in length by electroless deposition within the 4 nm wide channel of tobacco mosaic virus (TMV) particles. We
employed a low pH (∼7.5) copper electroless deposition solution that is compatible with biomolecules. The fabrication process of the nanowires
is based on sensitization of tobacco mosaic virus with Pd(II) prior to the electroless deposition. We analyzed the chemical composition of the
nanowires by energy-filtering transmission electron microscopy, and used the method also for nickel and cobalt nanowires deposited within the
viral channel.
© 2006 Elsevier Ltd. All rights reserved.
Keywords: Tobacco mosaic virus; Electroless deposition; Template; Cu nanowires; Energy-filtering TEM
1. Introduction
important complement to top-down techniques. One potential
bottom-up method involves using biological molecules as tem-
plates for arranging and producing circuit components. Rod-
or tube-like biological molecules, such as DNA [2], tobacco
mosaic virus (TMV) [3,4], bacteriophages [5], protein tubes [6]
and protein fibers [7,7a] are well defined in their chemical com-
position and structure. Such molecules are used as templates for
the nucleation of nanoscale inorganic structures [2–22].
Here we present the synthesis of copper nanowires of 3 nm in
diameter with electroless deposition within the 4 nm wide chan-
nel of TMV particles. Galvanic copper deposition has become
very popular after the invention of “damascene” copper elec-
trodeposition (embedded copper technology) for on-chip metal-
lization. Advantages – compared to the formerly used aluminum
– are decreased resistance, improved reliability and reduced
process complexity [23]. Biotemplates normally cannot be con-
tacted electrically; therefore electroless deposition (ELD) is
favored over electrodeposition. In addition, ELD is advanta-
geous for high-aspect ratio structures since they can be filled
or decorated without clogging up or closing an orifice [24]. Our
The size of features on integrated circuits has rapidly
decreased, due to the improvements in the lithographic materials
and techniques used to fabricate microstructures. However, fab-
rication of nanostructures (below 100 nm) with optical pattern-
ing techniques remains problematic mainly due to the ambient
absorption of hard ultraviolet radiation and expensive instru-
mentation required. An alternative method is a bottom-up, or
self-assembly, strategy, which uses chemical or physical forces
operating at the nanoscale to assemble basic units into larger
structures [1]. As component size decreases in nanofabrication
(below 100 nm), bottom-up approaches provide an increasingly
∗
Corresponding author. Tel.: +49 711 689 1433; fax: +49 711 689 1079.
E-mail address: a.bittner@fkf.mpg.de (A.M. Bittner).
Present Address: Department Physical Metallurgy and Materials Testing,
1
University of Leoben, Franz-Josef-Str. 18, A - 8700 Leoben, Austria.
2
Present Address: Max-Planck-Institut fu¨r Mikrostrukturphysik, Weinberg
2, D-06120 Halle, Germany.
0013-4686/$ – see front matter © 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2006.04.007

6252
S. Balci et al. / Electrochimica Acta 51 (2006) 6251–6257
template, TMV, is rod-shaped, 300 nm long and 18 nm in diam-
eter, with a central channel diameter of 4 nm [22]. The exact
amino acid sequence of the TMV coat protein is known, and
thus one can induce regioselective covalent linking [19,20] or
localized chemical reactions with metal ions [3,4,9,13].
nanometer range, using the information provided by inelasti-
cally scattered electrons in a spatially resolved manner [30–32].
The method uses the characteristic inner-shell energy-loss edges
of the element under study for mapping the elemental distribu-
tion over small areas with a resolution in the nanometer range
[30,31]. By elemental mapping, EFTEM was used for chemical
analysis of several different nanomaterials: Grobert et al. [33]
have reported a nickel/iron alloy (invar) inside carbon nanotubes
and Reis et al. [16] have reported FePt nanoparticles on the M13
bacteriophage.
“Sensitization” (“activation”) [3,8,9,24–28] creates noble
metal clusters on nonconductive surfaces. The details of this pre-
treatmentarenotyetknowneventhoughsometheoreticalstudies
exist [10,29]. Followed by ELD, the two-step process is com-
parable to the “enhancement” of latent photographic patterns.
We sensitized a TMV suspension with noble metal complexes
such as Pt(II) or Pd(II); only very small clusters (<1 nm diame-
ter) are formed during this procedure [9]. The TMV suspension
was then washed and mixed with the ELD solution containing
metal ions, complexants and a reductant (e.g. Cu(II), EDTA and
a borane-containing reductant). As usual in ELD, the deposi-
tion starts, at first catalyzed by the noble metal nanoparticles.
Later, the deposited metal catalyzes its own reductive deposi-
tion; therefore the metal particles can continue to grow. In this
way, we had already synthesized nickel and cobalt nanowires in
the inner channel of TMV particles with ELD [9]. An extension
to copper is not straightforward, especially for biotemplates. For
example, Demir et al. [19] have reported the photo-reduction
of copper at the outer surface (coat) of TMV particles at pH
6.5, and found a relatively low number of polydisperse cop-
per particles. Monson and coworkers have developed a method
to deposit copper metal onto surface-attached DNA, forming
nanowire-like structures that are ∼3 nm in diameter; however,
some copper complexes cleave DNA molecules and produce
short DNA molecules, which results in short copper nanowires
(10 nm in length) [21]. Therefore, DNA is not an optimal tem-
plate for fabricating long copper nanowires with a very small
diameter.
Copper ELD usually entails very high pH values (up to 13)
[24,25], which are incompatible with most biomolecules; there-
fore we used a low-pH (∼7.5) copper ELD solution that contains
– as nearly all ELD solutions – a reductant, a metal ion source,
a complexing agent and a buffering agent. The buffering agent
controls the pH during ELD, which is very important for the
deposition of metals on virions since already moderately high
pH values (>9) can denature the viral coat proteins [22]. The
complexing agent, preventing random deposition of metal ions
in the solution, changes the electrochemical reduction potential
of the corresponding metal ions and determines the stability of
the ELD solution [24,25]. The reductant reduces the metal ions
in the ELD solution to the metallic state. In order to develop
a new ELD solution for the deposition of metals on biomolec-
ular surfaces, all the parameters, such as concentration of the
components in the deposition solution, pH value and tempera-
ture of the deposition solution, have to be compatible with the
corresponding surfaces.
2. Experimental
2.1. TMV isolation
A Plasmid-DNA was employed to mechanically infect Nico-
tiana tabacum cv. Samsun nn plants (see also ref. [3] for details
of the virus preparation). TMV particles were isolated from
systemicallyinfectedleaves4weekspost inoculation [3]. Differ-
ential centrifugation including density gradient was employed
for virus isolation, yielding 6 mg TMV/g plant material. The
virus suspension was stored in phosphate buffer (pH 7.4) at 4 ◦C.
It was dialyzed before metallization using Slide-A-Lyzer 10000
MWCO units (KMF, Lohmar).
2.2. Deposition of metals
Before metallization, the TMV suspension was first sensi-
tized with aqueous Pd(II). 500 ␮l of a 0.2 mg/ml TMV suspen-
sion in water was incubated with the same amount of freshly
prepared aqueous 1.36 mM Na2PdCl4 (Aldrich, 99.998%) with
excess (1M) NaCl (Merck, ≥99.5%) at pH 5.0. After 10 min, the
suspension was centrifuged for 10 min at 14000 rpm, producing
a pale brown pellet. The supernatant was removed and the pellet
washed twice with water by resuspension and repeated centrifu-
gation. Pd(II)-incubated TMV suspension was metallized with
copper by treating the virus suspension with the correspond-
ing copper electroless deposition (ELD) solution for 20–30 min
(second step). In analogy to other electroless deposition reac-
tions based on DMAB (dimethylamine borane), gas bubbles
formed during the reaction.
The ELD copper solution described by Jagannathan et
al. was used in order to metallize TMV with copper at
lower pH level (pH ∼7.5) [28]. Pd (II)-incubated TMV sus-
pension was mixed with an equal volume of a solution
containing 0.032 M CuSO4·5H2O (Sigma, 99.999%), 0.04 M
Na2H2EDTA (ethylenediaminetetraacetic acid disodium salt,
Aldrich), 0.05 MTEA(triethanolamine, JanssenChimica, 97%),
and 0.067 M DMAB (Aldrich, 97%) in water. All metalliza-
tion steps were carried out at room temperature. Water was
purified with a Millipore Milli-Q or a Barnstead Nanopure
apparatus to 18.2 Mꢀcm. For nickel wire production, Pd(II)-
incubated TMV suspension was metallized by mixing the
suspension with an equal volume of a solution containing
0.18 M Ni(CH3COO)2·4H2O (Aldrich, 98%), 0.23 M lactic acid
(Aldrich, 95%) and 0.067 M DMAB (Aldrich, 97%). For cobalt
metallization (see Fig. 2c for the nanowire), the Pd(II)-incubated
In the present work we also apply energy-filtering transmis-
sion electron microscopy (EFTEM) to investigate the copper,
nickel, and cobalt nanowires in the inner channel of TMV.
Energy-filtering transmission electron microscopy is a powerful
tool for nanoanalytical studies. It allows the analysis of small
specimen volumes with high sensitivity and resolution in the

S. Balci et al. / Electrochimica Acta 51 (2006) 6251–6257
6253
TMV suspension was mixed with an equal volume of a solution
3. Results and discussion
containing 0.14 M CoSO4·H2O (Aldrich), 0.15 M sodium suc-
cinate (Aldrich, 99%) and 0.067 M DMAB. Metallization times
were as for copper.
3.1. Deposition of copper
In order to deposit copper, first we tested standard
formaldehyde-based high pH level ELD solutions [23–25], but
foundthatthevirionsweredestroyedsinceTMVdenaturesatpH
>9 [22]. In order to deposit metal on biological surfaces, all reac-
tions and components in the deposition solution have to be com-
patible with the template (pH, deposition rate, temperature, . . .).
For instance, the pH of the deposition solution often has to be
maintained around ∼7 to avoid denaturation of proteins—in our
case to avoid disassembly of the protein subunits during depo-
sition. In addition, we also require a highly selective deposition
of metals on biological surfaces without deposition in the bulk
solution or the enclosing reaction vessel. In order to fulfill all
the requirements for copper deposition on the nanoscale within
the TMV central cavity, we used the copper deposition solution
described by Jagannathan and Krishnan [28]. After their initial
studies with EDTA-based systems, the authors focused on mul-
tidentate nitrogen donors that form very strong complexes with
copper. In this way one can increase the stability of the ELD
solution at pH ∼9. However, we switched back to EDTA as the
complexing agent in our deposition solution (pH ∼7.5), since
we found that the solution can be stored for long times with-
out special protection. The other ingredients in the process were
copper sulfate as the metal ion source, triethanolamine as the
buffering agent, and DMAB as the reducing agent. DMAB is a
strong reducing agent, which operates especially well at basic
pH values [28,34]. A buffer, here TEA, is essential since the
electroless metal deposition as well as side reactions of DMAB
(see below) consume OH−, which would decrease the pH of the
solution and thus decrease the deposition rate.
2.3. Sample preparation
In order to analyze the chemical composition of nanowires
by EFTEM, 10 ␮l aliquots of the TMV suspension were taken
directly from the respective metallization bath, deposited on
formvar/carbon-coated copper grids (300 or 400 mesh, SPI-
Supplies), and incubated for 2–3 min, after which the suspension
was wicked off with a filter paper. Prior to the deposition of
the metallized virions, the grids were dipped into ethanol (Roth
Rotipuran, ≥99.8%) and air-dried.
2.4. Energy-filtering transmission electron microscopy
measurements
The specimens were investigated by EFTEM in order to
study the nanometer chemical composition of nanowires. All
experiments were performed on a Zeiss 912 Omega energy-
filtering microscope operated at 120 kV, except for Fig. 1 where
a Philips CM 200 at 200 kV (photographic detection) was used.
The Zeiss TEM data were acquired using a Gatan slow scan
CCD camera with 1024 × 1024 pixels and 25 ␮m pixel size. For
the experiments a window size of 50 eV, binning 2x, an objective
aperture of 60 ␮m and an exposure time of 60 s for each image
were used. The copper map was calculated using two pre-edge
images and one post-edge image at 841, 901 and 961 eV,
respectively, by using the three-window technique. In the same
way, a nickel map was recorded with two pre-edge images and
a post-edge image at 775, 820 and 885 eV. For cobalt, 699,
749 and 804 eV were used. An integration width of 20 pixels,
corresponding to 19.5 nm box length (parallel to the nanowire),
was used for all line profiles. We tried to operate the Zeiss 912
microscope at 60 kV instead of the usual 120 kV to obtain a
better contrast for the virions, but the data did not improve.
After activation of the virus particles with a catalyst (Pd(II)),
the metallization bath resulted in deposition of copper within
the 4 nm-wide internal channel of TMV particles without denat-
uration of the virus particles. Fig. 1a and b show bright-field
transmission electron microscopy (TEM) images of copper
Fig. 1. TEM micrograph of copper nanowires within the central channel of TMV particles. (a) The dark line corresponds to a copper nanowire, 3 nm in diameter,
and 115 nm in length. The wire is surrounded by dark-grey region, the viral coat proteins of two end-to-end aggregated virions, of which only one is metallized. (b)
A 85 nm copper nanowire (dark line 3 nm in diameter) within the central channel of a virion (grey region around the nanowire).

6254
S. Balci et al. / Electrochimica Acta 51 (2006) 6251–6257
Fig. 2. Elemental mapping of copper, nickel and cobalt nanowires by EFTEM. A bright field TEM image of (a) copper, (b) nickel, and (c) cobalt nanowires within
the central channel of TMV particles and the corresponding elemental distribution images ((d) copper, (e) nickel, and (f) cobalt) recorded with the three-window
technique (Cu L2,3 edge at 931 eV, Ni L2,3 edge at 854 eV, and Co L2,3 edge at 779 eV). Line profiles integrated over 20 pixels (19.5 nm) show the compositional
line traces of the elements inside the marked areas ((g) copper, (h) nickel, and (i) cobalt).
nanowires in virions, ∼3 nm in diameter. Therefore, electroless
deposition can also be used on the nanoscale (below 5 nm) for
the generation of materials. The exact diameter of the nanowire
is difficult to measure since the borders of the wire are not always
clearly visible. In four separate experiments, we observed that
between 40 and 50% of the TMV particles were metallized with
copper, and that the length of the wires varied from 5 to 150 nm.
We never found a 300 nm long copper wire within the virus
central channel.
The equilibrium is on the right hand side, hence ELD of cop-
per is thermodynamically allowed (spontaneous reaction); the
kinetic barrier is so high that spontaneous deposition does not
occurinthedepositionbath[25], andcatalystssuchasnoblemet-
als(orcopper)arerequired. ThePourbaixdiagramforthesystem
boron-water shows that HB4O7− is the dominant species at pH
values from seven to nine at nearly all relevant redox potentials
(from −1.0 to 0.8 V) [35]. Note that the (formal) oxidation state
of the boron does not change: the hydrogen atoms are oxidized.
The reaction can be separated into two-half reactions, cathodic
reduction of copper ions and anodic oxidation of DMAB on a
catalytic surface such as palladium for the first copper atoms to
be deposited, or copper during ELD [25,36-38]. The reduction
of the copper-EDTA complex
The ELD of copper can be formulated as follows, assum-
ing complete oxidation of the borane [35,38], and taking into
account the various (de)protonation equilibria:
4(CH3)2NHBH3 + 12CuEDTA2− + 9OH−
2e− + CuEDTA2− + H2O → Cu0 + OH− + HEDTA3− (2)
+
→ HB4O7− + 12Cu + 4(CH3)2NH2
requires potentials below −0.22 VSHE [25]. Note that the forma-
+ 12HEDTA3− + 2H2O
(1)
tion constant of the complex is very high (log Kf = 18.8), thus

S. Balci et al. / Electrochimica Acta 51 (2006) 6251–6257
6255
free Cu2+ is practically absent [25]. Weassume thatthe oxidation
of DMAB is described by
metals within the hollow channel of TMV particles. Excess NaCl
is very important in order to avoid formation of palladium oxo
and hydroxo complexes [26], which is favored when the pH
value is above two and when the chloride concentration is very
low. In order to initiate deposition of metals within the central
channel of TMV, the palladium complex has to reach the cen-
tral channel of the virion. It cannot bind electrostatically to the
negatively charged channel [13], however, it might form com-
plexes with the functional groups of amino acids (for example
lysine, arginine, asparagine) lining the interior channel, even
when the residues contain positively charged protonated amine
groups [26]. Prior to the ELD of copper, the Pd(II) species have
to be reduced by the reductant (DMAB) since only Pd0 can act
as a catalyst for ELD, but not Pd(II). Since small palladium (or
platinum) clusters down to a few atoms in size are catalytically
active [10,26], and the inside surface of TMV has to be coated
uniformly with the palladium clusters rather than totally filled
with them, the copper deposition can not only be confined to the
viral channel, but also be so effective that the channel is filled
with a wire instead of separate clusters.
4(CH3)2NHBH3 + 21OH−
→ HB4O7− + 4(CH3)2NH2+ + 14H2O + 24e−
(3)
which is a two-electron process (per H atom of the borane)
[38]. Note that this equation does not suggest any mechanism.
When the oxidation is combined with nickel or cobalt ion reduc-
tion, the mixed potential is around −700 to −600 mVSHE [37],
well compatible with the onset of DMAB oxidation on gold at
roughly −700 mVSHE for our pH value, 7.5 [38]. Obviously,
the reduction of the more noble copper compounds is allowed.
However, the mechanism of the anodic oxidation of DMAB has
not been completely clarified and depends strongly on the pH
value [25,36–38]. There appears to be general agreement that
BH3(OH)− is the active reducing species—clearly its forma-
tion from DMAB depends on pH. Multistep electron transfers
that each involves the production of hydrogen (one-electron oxi-
dation of the borane hydrogen atoms) have been proposed for
alkaline media [25,34]:
3.2. Deposition of nickel and cobalt
2BH4−n(OH)n− + 2OH− → 2BH3−n(OH)1+n− + H2 + 2e−
Nickel and cobalt nanowires were produced [9], again within
the central channel of TMV particles, after ELD of Pd(II)-
sensitized TMV suspension with special ELD formulations (see
also Section 2 for details). ELD mechanisms and redox poten-
tials have been discussed above; here we mention that nickel and
cobalt require more negative mixed potentials that are still above
the onset of the two-electron oxidation of DMAB. Fig. 2b and
c show bright-field TEM images of nickel and cobalt nanowires
(darklines), respectively. Thediameterofthenickelnanowiresis
∼3 nm, and their length ∼65 nm, as determined from the respec-
tive images. The grey-shaded region around the nickel wire is
interpreted to represent a destroyed TMV particle, distributed
on the carbon-coated TEM grid. The dark cobalt nanowire
within the single TMV particle (grey-shaded region around the
nanowire) shown in Fig. 2c exhibits similar dimensions as the
nickel wire in Fig. 2b. Cobalt wires of similar diameter were
recently reported by Kazakova et al. [40].
(4)
(with n = 1,2,3 in the course of the oxidation). This mechanism
predicts a strong correlation with high pH values; hence it cannot
be expected to be valid at pH values around 7. The relevant redox
potentials have been determined at several very basic pH values
[34]. An extrapolation of the pH dependence to our pH 7.5 gives
values above 0VSHE, which would not allow copper, let alone
cobalt or nickel reduction, another indication that the multistep
mechanism (4) can be responsible only for side reactions, espe-
cially hydrogen evolution. This would explain the observed gas
bubbles. In passing we note that all mechanisms can involve the
codeposition of boron in the metal structure during yet another
side reaction [36], with important consequences for mechanical
and magnetic properties of the deposit.
While all these considerations are valid for the macroscopic
scale, nanoscale ELD can change a number of parameters. For
example, the dissolution potential of nanoscale metals can shift
some mV from the bulk value [39]. The concentrations in the
vicinity of a growing copper surface deviate substantially from
bulk values. In the extreme case, one may have to be cautious
to define concentrations: a 4 nm long part of the 4 nm wide
channel of TMV contains only one copper ion and two DMAB
molecules, and the electrochemical double layer on the face of
a growing copper wire refers to an area of only π(2 nm)2 with
220 copper atoms.
3.3. Chemical analysis of nanowires
The electron spectroscopic image revealing the copper ele-
mental distribution is displayed in Fig. 2d, which was obtained
using the element-specific signal at the Cu L2,3 edge at 931 eV
and two pre-edge images for background subtraction. Fig. 2g
shows the corresponding integrated line profile from the box-
shaped area indicated in Fig. 2d. Fig. 2b, e and h are the corre-
sponding images and elemental maps for nickel (Ni L2,3 edge
at 854 eV), while Fig. 2c, f and i refer to cobalt (Co L2,3 edge at
779 eV). In addition to the measurements from the bright-field
images, the diameters of the copper, nickel, and cobalt nanowires
were also determined from the corresponding integrated line
profiles (Fig. 2g–i). The full width at half maximum (FWHM)
values in the integrated line profiles for the copper, nickel, and
cobalt nanowires are 13.5, 7.8, and 7.1 nm, respectively. These
The ELD of copper within the central channel of virions con-
sists of several steps, each of which should be carefully analyzed.
Crucially important is the surface chemistry of TMV: in the case
of virions stored in phosphate buffer, the TMV suspension is dia-
lyzed against water prior to usage, because metallization occurs
on the exterior surface of the virions due to adsorbed phos-
phate [3]. Our sensitization solution contains only Na2PdCl4
(1.36 mM) and excess NaCl (1 M) at pH 5, aiming to deposit

6256
S. Balci et al. / Electrochimica Acta 51 (2006) 6251–6257
FWHM values are higher than the ∼3 nm determined from
bright-field TEM images. Most likely this difference is due to
instrumental drift during the EFTEM measurements, and also
due to noise problems especially for the copper nanowires, as
will be discussed further below. In principle, the intensity of ele-
mental maps can be increased by using longer acquisition times
and/or a higher beam current. However, longer acquisition times
resulted in an even more pronounced drift of the samples during
the acquisition of the pre-edge and post-edge images, leading to
artefacts in the calculated elemental map. Measurements with
higher beam currents have the disadvantage that the nanowires
are damaged, again causing artefacts.
As mentioned above, the line profile intensity of the copper
nanowires shows lower counts than the nickel and cobalt line
profile intensity. The width of the profile (FWHM) is not only
larger than in the bright-field image, it is also larger than for
cobalt and nickel. For nickel and cobalt, sharp white lines occur
at the edge onset due to the excitation of 2p electrons into unoc-
cupied 3d states [31,32] (see Fig. 3). In contrast, no pronounced
features occur at the Cu L2,3 edge since the 3d shell is fully
occupied [32] (Fig. 3). In addition, the Cu L2,3 edges lies at
the highest energy-loss (931 eV) in the investigated materials,
and thus the intensity of this edge is lowest [31,32] (see Fig. 3).
Due to these reasons, the intensity in the elemental maps for
copper is lower than for nickel and cobalt, if all other param-
eters (acquisition time and beam current) are the same. The
diminished signal/noise ratio increases the FWHM, thus making
EFTEM analysis of copper nanostructures comparatively diffi-
cult. Accordingly, the FWHM value determines the resolution
of the chemical information, while the structural resolution from
the TEM image is considerably better (FWHM ∼4 nm). We also
performed measurements using a scanning transmission elec-
tron microscope (VG HB 501 UX operated at 100 kV), which is
capable to form an electron probe with a size of less than 1 nm. It
has a higher spatial resolution for analytical measurements com-
pared to the TEM used for the EFTEM investigations, however,
due to the high electron beam current density radiation damage
occurred at the nanowires. The results are thus omitted here.
4. Conclusion
In this work it is shown that copper electroless deposition
can be used to synthesize nanosized copper wires specifically
within the central channel of TMV particles, 3 nm in diame-
ter and up to 150 nm in length. Therefore, electroless deposition
methods can also be useful for nanotechnology applications. We
employed electroless copper deposition solution at physiologi-
cal conditions (pH ∼7.5) that is compatible with biomolecules.
The fabrication process of the nanowires is based on treatment
of TMV with Pd(II), followed by electroless deposition of cop-
per. Energy-filtering transmission electron microscopy proves
that the copper nanowires synthesized with the method of ELD
within the 4 nm central channel of TMV particles are indeed
composed of copper. In addition, a similar proof was obtained
also for nickel and cobalt nanowires. Length and diameter of the
elemental maps of all three metal wires reproduce the respec-
tive bright field TEM images. A closer analysis of the widths of
the EFTEM features shows a dependence on the metal: copper,
nickel, and cobalt have FWHM values of 13.5, 7.8, and 7.1 nm,
respectively. The sensitivities for cobalt and nickel are higher
than for copper in elemental maps since copper spectra do not
possess sharp “white lines” in the energy-loss spectra, and the Cu
L
2,3 edge lies at the highest energy of the materials investigated
as well.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at doi:10.1016/j.electacta.2006.04.007.
References
[1] G.M. Whitesides, J.P. Mathias, C.T. Seto, Science 254 (1991) 1312.
[2] E. Braun, Y. Eichen, U. Sivan, G. Ben-Yoseph, Nature 391 (1998) 775.
[3] M. Knez, M. Sumser, A.M. Bittner, C. Wege, H. Jeske, T.P. Martin, K.
Kern, Adv. Funct. Mater. 14 (2004) 116.
[4] W. Shenton, T. Douglas, M. Young, G. Stubbs, S. Mann, Adv. Mater. 11
(1999) 253.
[5] C. Mao, D.J. Solis, B.D. Reiss, S.T. Kottmann, R.Y. Sweeney, A. Hayhurst,
G. Georgiou, B. Iverson, A.M. Belcher, Science 303 (2004) 213.
[6] M. Reches, E. Gazit, Science 300 (2003) 625.
[7] M. Mertig, R. Kirsch, W. Pompe, Appl. Phys. A 66 (1998) S723.
[7a] T. Scheibel, R. Parthasarathy, G. Sawicki, X.M. Lin, H. Jaeger, S.L.
Lindquist, PNAS 100 (2003) 4527.
[8] A.M. Bittner, Naturwissenschaften 92 (2005) 51.
[9] M. Knez, A.M. Bittner, F. Boes, C. Wege, H. Jeske, E. Maib, K. Kern, Nano
Lett. 3 (2003) 1079.
[10] M. Mertig, L.C. Ciacchi, R. Seidel, W. Pompe, Nano Lett. 2 (2002) 841.
[11] M. Mertig, R. Wahl, M. Lehmann, P. Simon, W. Pompe, Eur. Phys. J. D.
16 (2001) 317.
[12] J. Richter, M. Mertig, W. Pompe, I. Mo¨nch, H.K. Schackert, Appl. Phys.
Lett. 78 (2001) 536.
[13] E. Dujardin, C. Peet, G. Stubbs, J.N. Culver, S. Mann, Nano Lett. 3 (2003)
413.
[14] Q. Wang, T. Lin, L. Tang, J.E. Johnson, M.G. Finn, Angew. Chem. Int. Ed.
41 (2002) 459.
[15] Z. Deng, C. Mao, Nano Lett. 3 (2003) 1545.
[16] B.D. Reiss, C. Mao, D.J. Solis, K.S. Ryan, T. Thomson, A.M. Belcher,
Nano Lett. 4 (2004) 1127.
Fig. 3. Comparison of the L2,3 white lines of cobalt with nickel and copper
electron energy-loss spectra. The nickel and copper signals are enlarged by the
indicated factors.

S. Balci et al. / Electrochimica Acta 51 (2006) 6251–6257
6257
[17] J.C. Falkner, M.E. Turner, J.K. Bosworth, T.J. Trentler, J.E. Johnson, T.
Lin, V.L. Colvin, J. Am. Chem. Soc. 127 (2005) 5274.
[18] K. Keren, R.S. Berman, E. Buchstab, U. Sivan, E. Braun, Science 302
(2003) 1380.
[19] M. Demir, M.H.B. Stowell, Nanotechnology 13 (2003) 541.
[20] T.L. Schlick, Z. Ding, E.W. Kovacs, M.B. Francis, J. Am. Chem. Soc. 127
(2005) 3718.
[21] C.F. Monson, A.T. Woolley, Nano Lett. 3 (2003) 359.
[22] M. Zaitlin, AAB Descriptions of Plant Viruses 8 (2000) 370.
[23] P.C. Andricacos, C. Uzoh, J.O. Dukovic, J. Horkans, H. Deligianni, IBM.
J. Res. Dev. 42 (1998) 567.
[24] Y. Shacham-Diamand, V. Dubin, M. Angyal, Thin Solid Films 262 (1995)
93.
[25] M. Schlesinger, M. Paunovic, Modern electroplating, John Wiley & Sons,
Inc., New York, 2000.
[26] H. Kind, A.M. Bittner, O. Cavalleri, K. Kern, J. Phys. Chem. B 102 (1998)
7582.
[27] E.A. Dobisz, R. Bass, S.L. Brandow, M.S. Chen, W. Dressick, J. Appl.
Phys. Lett. 82 (2003) 478.
[28] R. Jagannathan, M. Krishnan, IBM. J. Res. Dev. 37 (1993) 117.
[29] H. Nakai, T. Homma, I. Komatsu, T. Osaka, J. Phys. Chem. B 105 (2001)
1701.
[30] A. Berger, H. Kohl, Microsc. Microanal. Microstruct. 3 (1992) 159.
[31] D.B. Williams, C.B. Carter, Transmission electron microscopy, Plenum
Press, New York and London, 1996.
[32] R.D. Leapman, L.A. Grunes, P.L. Fejes, Phys. Rev. B 26 (1982) 614.
[33] N. Grobert, M. Mayne, M. Terrones, J. Sloan, R.E. Dunin-Borkowski, R.
Kamalakaran, T. Seeger, H. Terrones, M. Ru¨hle, D.R.M. Walton, H.W.
Kroto, J.L. Hutchison, Chem. Commun. (2001) 471.
[34] O.A. Sadik, H. Xu, A. Sargent, J. Electroanal. Chem. 583 (2005) 167.
[35] M. Pourbaix, Atlas of electrochemical equilibria in aqueous solutions,
National association of corrosion engineers, Texas, 1966.
[36] Y. Sverdlov, V. Bogush, H. Einati, Y. Shacham-Diamand, J. Electrochem.
Soc. 152 (2005) C631.
[37] T. Saito, E. Sato, M. Matsuoka, C. Iwakura, J. Appl. Electrochem. 28 (1998)
559.
[38] L.D. Burke, B.H. Lee, J. Appl. Electrochem. 22 (1992) 48.
[39] D.M. Kolb, R. Ullmann, T. Will, Science 275 (1997) 1097.
[40] O. Kazakova, D. Erts, T.A. Crowley, J.S. Kulkarni, J.D. Holmes, J. Magn.
Magn. Mater. 286 (2005) 171.