Synthesis of platinum cluster chains on DNA templates: Conditions for a template-controlled cluster growth

Article abstract of DOI:10.1021/jp037800r

We investigate the conditions which should be fulfilled to grow chains of nanosized noble metal clusters on DNA templates according to a selectively heterogeneous, template-controlled mechanism. A long incubation of double-stranded DNA molecules with Pt(II) complexes is necessary to obtain a template-directed formation of thin and uniform cluster chains after chemical reduction of the DNA/salt solution. Without this activation step, DNA acts as a nonspecific capping agent for the formed clusters and does not hinder the formation of random cluster aggregates. The effect of binding Pt(II) complexes to the DNA is investigated by UV-vis spectroscopy, electrophoresis experiments, and scanning force microscopy, revealing that the base stacking along the DNA molecule is significantly distorted but the double-stranded DNA configuration is retained. Citrate ions can be used as additional stabilizers for the heterogeneously grown metal clusters, leading to significantly more regular metal cluster chains. After a systematic variation of the absolute concentration of the reactants (Pt salt, DNA, reducing agent), we can conclude that there is an optimum concentration value for the fabrication of cluster chains and that small variations around the optimum value do not have noticeable effects on the quality of the metallization products. The described metallization procedure can be resolved into a series of simple and efficient steps, which is essential for a biomimetic fabrication of nanostructures in a reproducible way.

Full text of DOI:10.1021/jp037800r

J. Phys. Chem. B 2004, 108, 10801-10811
10801
Synthesis of Platinum Cluster Chains on DNA Templates: Conditions for a
Template-Controlled Cluster Growth
Ralf Seidel,†,‡ Lucio Colombi Ciacchi,†,§ Michael Weigel,† Wolfgang Pompe,† and
Michael Mertig*,†
Max Bergmann Zentrum fu¨r Biomaterialien and Institut fu¨r Werkstoffwissenschaft, Technische UniVersita¨t
Dresden, D-01069 Dresden, Germany, Department of Nanoscience, Delft UniVersity of Technology,
2628 CJ, Netherlands, and Theory of Condensed Matter Group, CaVendish Laboratory, UniVersity of
Cambridge, CB3 0HE Cambridge, United Kingdom
ReceiVed: December 11, 2003; In Final Form: May 5, 2004
We investigate the conditions which should be fulfilled to grow chains of nanosized noble metal clusters on
DNA templates according to a selectively heterogeneous, template-controlled mechanism. A long incubation
of double-stranded DNA molecules with Pt(II) complexes is necessary to obtain a template-directed formation
of thin and uniform cluster chains after chemical reduction of the DNA/salt solution. Without this “activation”
step, DNA acts as a nonspecific capping agent for the formed clusters and does not hinder the formation of
random cluster aggregates. The effect of binding Pt(II) complexes to the DNA is investigated by UV-vis
spectroscopy, electrophoresis experiments, and scanning force microscopy, revealing that the base stacking
along the DNA molecule is significantly distorted but the double-stranded DNA configuration is retained.
Citrate ions can be used as additional stabilizers for the heterogeneously grown metal clusters, leading to
significantly more regular metal cluster chains. After a systematic variation of the absolute concentration of
the reactants (Pt salt, DNA, reducing agent), we can conclude that there is an optimum concentration value
for the fabrication of cluster chains and that small variations around the optimum value do not have noticeable
effects on the quality of the metallization products. The described metallization procedure can be resolved
into a series of simple and efficient steps, which is essential for a biomimetic fabrication of nanostructures
in a reproducible way.
I. Introduction
biomimetic metallization techniques are the difficulties associ-
ated with the reproducible assembly of individual structures into
Metallic wires fabricated via the selective growth of metal
on DNA templates are very promising candidates for use in
microelectronic devices of future generation.1-13 The aim of
this approach is to create novel devices where quantum size
effects and the large number of surface atoms influence their
chemical, electronic, magnetic, and optical behavior.14,15 Since
DNA molecules have a diameter of only two nanometers and
can be arbitrarily long, very thin (below 10 nm) and regular
wire-like metallic structures can be efficiently produced.6
Moreover, both the ends of DNA molecules and the surface of
solid substrates can be functionalized through a variety of
methods to create specific links between DNA and substrate,
allowing the integration of DNA molecules into specific,
predefined sites of microstructured electronic circuits.1,13,16,17
Furthermore, by exploiting the specific recognition of comple-
mentary nucleotide sequences, complex structures made of DNA
can be fabricated without much effort9,16,13,18 and metallized in
a later stage.9
working devices and with the control of their desired physical
properties such as, e.g., electric conductivity. Moreover, it is
not clear at the present stage whether the recently developed
techniques for the metallization of biomolecules can be ef-
ficiently employed in large scale applications. To be suitable
for industrial processing, the methods for the fabrication of wires
(and other metallic structures) should be both (i) simple and
(ii) robust. Namely, (i) the overall process should be composed
by a few, clearly defined elementary steps which can be
performed in sequence with high degree of automatizm, and
(ii) the method should be as little as possible sensitive to small
fluctuations of the set of parameters which define the optimum
conditions for the fabrication of the products. In this work, our
goal is to define the conditions which permit a selectively
heterogeneous metallization of DNA molecules to be achieved
in a way which is simple, robust, and reproducible.
The assembling of metallic nanoparticles along DNA tem-
plates can be performed in several ways.19 For instance, chains
of separated particles can be prepared by binding positively
charged colloids to the negatively charged backbone of DNA
via electrostatic interactions.11 Moreover, particles can be grown
directly on the DNA template after the previous formation of
nucleation sites along the DNA followed by chemical reduction
of dissolved metal salts. Through this technique, coarse pal-
ladium wires with a diameter of about 50 nm have been
obtained,2 which possess peculiar electric conduction properties
both at room and at low temperatures.3,8
Despite of the progress which has been made in the past years,
the current techniques of biopolymer metallization for the
production of nanoscopic conductive elements are still at an
early stage of development. The major drawbacks of the today’s
* To whom correspondence should be addressed. E-mail: mertig@
tmfs.mpgfk.tu-dresden.de.
† Technische Universita¨t Dresden.
‡ Delft University of Technology.
§ University of Cambridge.
10.1021/jp037800r CCC: $27.50 © 2004 American Chemical Society
Published on Web 06/25/2004

10802 J. Phys. Chem. B, Vol. 108, No. 30, 2004
Seidel et al.
More recently, by accurate tuning of the metallization
conditions, continuous chains of platinum clusters no thicker
than a few nanometers and extended over several micrometers
have been realized by the present authors.6 The key feature of
the employed technique is the long incubation of DNA with an
aged solution of K2PtCl4 to allow the binding of Pt(II)
complexes to the biomolecule. The bound Pt(II) complexes act
in the subsequent reduction step as preferential nucleation
centers for a selectively heterogeneous cluster growth.5,6 In fact,
through a series of first principles investigations, we demon-
strated that the presence of the nucleotide ligands both enhance
the electronic affinity of the complexes and stabilize the
formation of Pt-Pt bonds during the first reduction stages.6,20
Since cluster growth is an autocatalytic process,21 the first
formed nuclei quickly develop to larger clusters and consume
the dissolved Pt(II) complexes.6,22 In this way, the homogeneous
nucleation process is hindered. Therefore, the achievement of
a selective DNA metallization without homogeneous formation
of clusters in solution is the result of a delicate balance between
the kinetics of reduction in the homogeneous and heterogeneous
case.
into the influence of the DNA composition on the overall
reaction kinetics.16 Finally, we give a summary of problems
and open questions to stimulate future work in the field of DNA
metallization.
II. Experimental Section
A. Metallization of DNA Molecules. If not otherwise stated,
the reduction experiments were performed by aging a 10 mM
solution of K2PtCl4 or K2PdCl4 (both purchased from Fluka)
for at least 24 h, diluting the aged solution to 1 mM, and mixing
it with 5 µg/mL of λ-DNA (New England Biolabs) or DNA
from salmon testes (SIGMA). Both types of DNA exhibit
statistically random sequences and GC base pairs contents of
49.9% and 41.2%, respectively. In this way we can exclude
such effects of DNA metallization, which arise from the specific
composition of DNA sequence. The concentration of DNA was
varied between 0 and 100 µg/mL in the investigations described
in section III.B. The kinetics of DNA activation was investigated
by diluting four stock solutions of K2PtCl4 at different initial
concentrations which were previously aged for at least 24 h to
a final concentration of 1 mM and mixing them with 5 µg/mL
of λ-DNA. The reduction of the dissolved complexes to form
metal cluster was started by heating the solution to 37 °C and
adding 100 µL of a 10 mM solution of dimethylaminoborane
(DMAB, Fluka) to 1 mL of the DNA/Pt solution. In some cases,
sodium citrate was added to the DNA solution in the concentra-
tion of 0.5 mM before starting the reduction.
Here we show how this balance is shifted by changing the
absolute and relative concentrations of reactants, and the type
of metal complexes. This can have a dramatic effect on the final
products, leading in most cases to the formation of random
cluster agglomerates or to a very coarse DNA metallization.
However, we find that there is an optimum interval of
parameters where thin and uniform cluster chains form selec-
tively on DNA molecules. Moreover, small variations around
the optimum set of parameters have only little or no influence
on the quality of the metallization products. This makes our
metallization technique robust, in the sense that once an
optimum set of parameters is found for a given metal complex
uncontrolled small fluctuations of the reducing conditions are
very unlikely to lead to faulty products. We stress that, despite
the fact that the actual molecular mechanisms leading to cluster
formation are rather complicated, the overall process can be
resolved into a series of very simple steps. In particular,
uncomfortable processes such as dialysis, ion exchange, or
changing of the solvent are avoided, which would be very
expensive and inefficient at a level of massive production.
The kinetics of both the activation process and the cluster
formation process were studied using a Cary 50 UV/VIS
photospectrometer (Varian), recording the absorbance spectra
at wavelengths above 240 nm and above 550 nm, respectively,
during the reaction course.
B. Microscopic Investigations. Samples for scanning force
microscopy (SFM) and transmission electron microscopy (TEM)
were prepared 30 min after starting the reduction. SFM samples
were prepared as described by Bezanilla et al.23 by diluting the
DNA/cluster suspension with HEPES/Mg buffer (40mM HEPES,
5mMMgCl2, pH 7.6) to a final DNA concentration of 1 µg/mL
and placing a droplet onto freshly cleaved mica for 2 min.
Subsequently, the sample was rinsed with water and blown dry
with nitrogen. Imaging was performed with a Nanoscope IIIa
(Digital Instruments) in tapping mode in air. SFM samples from
activated DNA were prepared following the same protocol. TEM
samples were prepared by placing a droplet of the DNA/cluster
suspension onto a carbon-coated copper grid for 5 min.
Afterward, the sample was rinsed with water and excess water
was removed with filter paper. The TEM investigations were
carried out using a Philips CM 200 with tungsten cathode at
160 kV for low-resolution imaging, whereas a Philips CM 200
microscope with a field emission cathode was employed for
the high resolution work.
Defining the conditions which are necessary for a template-
controlled cluster growth on DNA requires the detailed knowl-
edge of all steps of the metallization process. This begins with
the dissolution of metal salt in water and with the reaction of
the dissolved metal complexes with the DNA bases and ends
with the formation of the metallized products after the addition
of a reducing agent to the solution in order to achieve the
formation of metal clusters. Although the essential steps of the
process have been presented in ref 6, a number of questions
remain to be answered in more detail. In particular, we
investigate here (i) the effect of the hydrolysis degree of
dissolved the Pt(II) complexes on the kinetics of binding to the
DNA bases during the activation step; (ii) whether the addition
of citrate ions to stabilize the formed metal clusters improve
the quality of metallization; (iii) to which extent variations of
the concentrations of reactants influence the formation of regular
and thin metal wires; (iv) the importance of the double-stranded
DNA configuration with respect to the single-stranded DNA
configuration for the formation of the reduction products; and
(v) how the balance between homogeneous and heterogeneous
nucleation can be shifted by changing the overall reduction
kinetics using Pd(II) complexes instead of Pt(II) ones. Further-
more, we briefly discuss the results of previous investigations
C. Electrophoresis Experiments. In the electrophoresis
investigations, 5 µg of native λ-DNA and of λ-DNA digested
with Hind III (New England Biolabs) were activated for 20 h
in 1 mL of a 1mM K2PtCl4 solution previously aged for 24 h.
Single-stranded DNA (ss-DNA) was obtained by thermal
denaturation of double-stranded DNA (ds-DNA) at 95 °C and
subsequent cooling on ice. After activation, 20 µL of each
sample were mixed with 1 µL of a 1 mM DMAB solution to
cover the DNA with nucleation sites for the subsequent gold
stain. Electrophoresis was performed in a 0.5% agarose gel for
1.5 h in 1× TAE-buffer at 4.2V/cm. Subsequently the gel was
covered with a gold stain solution7 until the DNA bands

Platinum Cluster Chains on DNA Templates
J. Phys. Chem. B, Vol. 108, No. 30, 2004 10803
as follows (see, e.g., ref 29)
(CH3)2HNBH3 + PtCl2(H2O)2 + 2H2O f
PtV + (CH3)2HNH+ + H3BO3 + 2HCl + OH- + 2H2 (3)
In the absence of any capping agent, the formed metal clusters
agglomerate into ∼50 nm large, polycrystalline particles (Figure
1a). When a capping agent like sodium citrate is added to the
solution prior to the reduction, small metal clusters are stabilized
by a citrate ligand shell and are visible in the TEM images
together with bigger agglomerates (Figure 1b).30 The isolated
particles are of polyhedral shape (tetrahedrons, cubes, and
cuboctahedrons can be distinguished on the sample’s surface),
which is typical for Pt clusters stabilized by citrate.31 Interest-
ingly, the morphology of the formed cluster aggregates is in
part chainlike. However, the typical length of these randomly
formed chains is much shorter than the length of any hetero-
geneously grown chain of particles on DNA (see below and
refs 5, 6, and 16).
Since the reduction is performed under air, the presence of
dissolved oxygen is expected to lead to formation of PtOx phases
at the surface of the metallic particles.26,33 Superficial platinum
oxide phases have been in fact detected by time-of-flight
secondary ion mass spectrometry (TOF-SIMS) in a previous
analysis of the platinum powders obtained following a reduction
protocol analogous to the one used in this work.34 Furthermore,
an XPS analysis of the same material confirmed the metallic
character of the obtained particles, revealing the presence of
oxygen and carbon adsorbates, together with chlorine and
potassium ions.35 Although no boron impurities could be
detected in these analyses, the boron content in particles
produced by reduction of Pt complexes with boron compounds
(such as, e.g., DMAB) has been found to be about 0.05 wt %
by other authors.29,36
Figure 1. TEM micrographs of Pt particles formed after reduction of
a 1 mM solution of PtCl4 with 1 mM DMAB. (a) In the absence of
citrate; (b) in the presence of 0.5 mM citrate. The scale bar in the inset
is 10 nm.
2-
B. Pt Cluster Synthesis in the Presence of Nonactivated
2-
DNA. We turn now to study the reduction of dissolved PtCl4
ions in the presence of native DNA. To this aim, an aged
solution of K2PtCl4 is mixed with dissolved DNA, and DMAB
is immediately added to the solution. Under these conditions,
the characteristic reduction time is shorter than the characteristic
binding time of Pt(II) complexes to the DNA bases,37 so that
only a few complexes have time to bind to the DNA before the
reduction is fully completed. We refer to these conditions as
reduction in the presence of “nonactivated” DNA.
The overall kinetics of the reduction process in the presence
of DNA in different concentrations is monitored by means of
UV-vis spectroscopy at the wavelength of 600 nm (Figure 2).
At this wavelength, there is no absorbance due to either
dissolved [PtCl4]2- complexes or DNA. Far from the plasmon
resonance frequency (which is at about 215 nm in the case of
Pt colloids26), the absorption of light depends mostly on the
concentration of the formed metal particles, and only to a minor
extent on their shape. According to ref 26, the major contribution
to the increase of absorption in this region of the spectrum is
due to dipole-dipole interactions of small particles when they
aggregate into bigger structures during the reduction reaction.
Therefore, the measured curves represent the overall kinetics
of the reduction process.
appeared. The reaction was stopped by rinsing the gel with
abundant water.
III. Results
A. Pt Cluster Synthesis in the Absence of DNA. When K2-
PtCl4 salt is dissolved in water solution, it dissociates, and after
2-
the dissociation, the PtCl4 ions undergo the following sol-
volysis reactions:24,25
PtCl42- + H2O h PtCl3(H2O)- + Cl-
(1)
(2)
PtCl3(H2O)- + H2O h PtCl2(H2O)2 + Cl-
At the equilibrium, a 1 mM solution of K2PtCl4, which is usually
considered as standard concentration to obtain suspensions of
metal nanoparticles in reduction experiments,26,27 consists of 5
PtCl42-, 53% PtCl3(H2O)-, and 42% PtCl2(H2O)2.28 At room
temperature, the aging time necessary for reaching the equilib-
rium is about 1 day. The solution is then stable for some days.
Within this time, the formation of higher-order solvolysis
products can be neglected because the corresponding formation
rates are 3 orders of magnitude lower.25 After mixing an aged
solution of K2PtCl4 with DMAB, the dissolved Pt(II) complexes
are reduced to metallic platinum particles (Figure 1). Consider-
ing, for instance, doubly hydrolyzed complexes, an overall
chemical reaction of this preparation can formally be written
In the presence of DNA, the overall reduction kinetics is
slowed (Figure 2). The maximum effect is reached at a DNA
concentration of 100 µg/mL, and no further retardation is
measured by increasing the DNA amount. The metallization
products obtained at 5 µg/mL are small crystallites with a
diameter of about 5 nm, as well as a few larger polycrystalline

10804 J. Phys. Chem. B, Vol. 108, No. 30, 2004
Seidel et al.
plexes, respectively.39 In fact, the characteristic time for the
binding of non-hydrolyzed complexes is about 2 h (which is
comparable with the characteristic time of the chlorine/water
substitution reaction), whereas 2-fold hydrolyzed complexes
react with the DNA bases within a few minutes.39 Although
monofunctional adducts are formed fast, the formation of
bifunctional adducts is much slower. In the case of cisplatin,
the latter is completed within about 10 h.39
In the case of dissolved [PtCl4]2- ions at high Pt(II)/base
ratios, it is known that three Pt(II) complexes bind on average
to each base pair.37 This reaction is completed in about 10 h.40
In analogy with the reaction mechanism proposed for hydrolyzed
cisplatin complexes, the reaction of hydrolyzed Pt(II) complexes
with DNA is expected to begin with the formation of a covalent
bond between the N7 atom of a purine DNA base (A or G) and
the Pt atom after loss of a water ligand. This reaction can be
formally written as
2-
Figure 2. Reduction kinetics of a solution of PtCl4 in the presence
of different amounts of nonactivated DNA, measured by UV-vis
spectroscopy at 600 nm. The DNA concentration is given in µg/mL.
2-
The concentrations of PtCl4 (1mM) and DMAB (1 mM) are kept
constant.
G + PtCl2(H2O)2 f (G)PtCl2(H2O) + H2O
(4)
Then, after the loss of a second water ligand, bifunctional
complexes like, e.g., (AG)PtCl2 are expected to form, where a
Pt(II) ion binds to two stacked bases. Similarly to the hydrolysis-
limited binding process of cisplatin, we expect that the hydroly-
sis degree of the [PtCl4]2- solution (see section III.A) will affect
the kinetics of binding to DNA. To gain more information about
the characteristic binding time of dissolved Pt(II) complexes
under the conditions used in our metallization experiments, we
2-
measure the binding kinetics of PtCl4 to DNA at different
hydrolysis degrees by means of UV-vis spectroscopy.
1. Kinetics of [PtCl4]2- Binding to the DNA. DNA molecules
have a maximum in the absorbance spectrum at ∼260 nm, which
is due to electronic excitations of the nitrogenous bases. The
260 nm absorbance increases passing from an ordered stack of
bases to disordered conformations because of the diminished
electronic interactions between the bases. This is evident, for
instance, during the melting of DNA, which is the transition
from the double-stranded to the single-stranded structure. Also
the binding of dissolved [PtCl4]2- ions to the DNA is known
to lead to a significant distortion of the base stacking.5,38 This
allows us to investigate the kinetics of the binding process
looking at the changes of the 260 nm absorbance peak during
the incubation of DNA with [PtCl4]2- (Figure 4). To separate
the contribution of the DNA spectrum from that of the Pt(II)
complexes, which also have a strong absorbance in the 260 nm
region, we subtract from the spectrum of the Pt(II)/DNA solution
Figure 3. TEM micrograph showing the metallization products
obtained after reduction of a 1 mM aged solution of PtCl4 with 1
mM DMAB in the presence of 5 µg/mL nonactivated DNA.
2-
particle agglomerates (Figure 3). Note that TEM analysis alone
does not permit the possibly present nonmetallized DNA
molecules to be visualized. However, previously performed SFM
studies5,6 revealed that under similar conditions the large metal
agglomerates are surrounded by free DNA molecules and that
a few small crystallites appear to be located on the DNA strands.
The addition of citrate to the Pt(II)/DNA solution prior to
reduction led to similar aggregates (not shown) and was not
effective in stabilizing long chains of particles aligned along
the DNA molecules. From these results, we conclude that, to
produce thin metallic wires, the reduction has to start necessarily
after the completion of the binding reaction of Pt(II) complexes
to the bases. The details of this process are investigated in the
next section.
C. Activation of the DNA Template. The binding process
of Pt(II) complexes to DNA is well investigated in the case of
cisplatin (cis-[Pt(NH3)2Cl2]) which is widely used as an
anticancer drug.38 When DNA is incubated with Pt(II) com-
plexes such as cisplatin, the Pt(II) atom binds to one or two
stacked DNA bases forming monfunctional and bifunctional
DNA‚Pt(II) adducts, respectively. The most favorable binding
site for cisplatin to the DNA is the N7 position of guanine,
followed by the N7 position of adenine.38 After these sites are
occupied, the binding reaction proceeds more slowly and
indiscriminately with other metal-binding sites of all bases.37 It
is known that the limiting step for the binding of [Pt(NH3)2Cl2]
complexes to guanine or adenine is the hydrolysis of the
complexes with loss of one or two chlorine ions, leading to the
formation of [Pt(NH3)2Cl(H2O)]+ and [Pt(NH3)2(H2O)2]2+ com-
2-
the spectrum of a reference solution containing only PtCl4
ions at the same concentration (Figure 4a).
As expected, the binding of the Pt(II) complexes to DNA
leads to an increase in 260 nm absorbance, i.e., to a strong
distortion of the DNA base stacking, with increasing incubation
time (Figure 4b). For each obtained DNA spectrum, we measure
both the sample and the reference solution to account for the
time-dependent changes of the [PtCl4]2- spectrum due to
ongoing hydrolysis processes. We start from a freshly prepared
1 mM solution and three aged stock solutions with different
concentrations (100, 10, and 1.5 mM), which were diluted to a
concentration of 1 mM immediately before mixing with the
DNA solution. In the four samples, the absolute concentration
of Pt(II) complexes is the same but the initial relative concentra-
tions of non-hydrolyzed, 1-fold hydrolyzed, and 2-fold hydro-
lyzed complexes are different (Table 1). The measured kinetics
curves are reported in Figure 4c. Initially, we observe a rapid
increase of the absorbance, which slows down at long incubation
times. However, the absorbance does not reach a saturation value

Platinum Cluster Chains on DNA Templates
J. Phys. Chem. B, Vol. 108, No. 30, 2004 10805
Figure 4. Distortion of the DNA base stacking due to the binding of PtCl42- complexes as a function of the incubation time. (a) Spectra of a 1 mM
solution of PtCl42-, and of a solution containing 1 mM PtCl42- and 5 µg/mL λ-DNA immediately after mixing (incubation time t ) 0). Subtraction
of the first spectrum form the second one provides a spectrum for the DNA only. The characteristic maximum at about 260 nm is visible. (b)
Spectra of DNA incubated with PtCl42- obtained at increasing incubation times. Note the increase of the absorbance in the region around 260 nm,
indicating the ongoing distortion of the DNA base stacking. (c) Distortion of the DNA base stacking recorded at 260 nm for differently hydrolyzed
solutions of PtCl42-, which were prepared by dilution of one fresh and three aged stock solutions with initial concentrations of 100, 10, and 1.5 mM
to a final concentration of 1 mM. The percentages of the hydrolyzed species are reported in Table 1.
TABLE 1: Relative Equilibrium Concentrations of the
Hydrolyzed Pt(II) Species in One Fresh and Three Stock
Solutions of K2PtCl4 Aged for 16 ha
t∆Abs/2
Pt(H2O)2Cl2 (min)
2 -
-
PtCl4
1
0.700
0.350
0.170
Pt(H2O)Cl3
2-
fresh PtCl4
aged 100mM PtCl4
0
0
440
350
190
70
2-
2-
0.292
0.585
0.562
0.008
0.065
0.268
aged 10mM PtCl4
aged 1.5mM PtCl4
2-
a The data were obtained using the equilibrium constants from ref
25. The right column in the table gives the characteristic time for the
distortion of the base stacking of DNA (see text and Figure 4) upon
incubation with the four differently hydrolyzed stock solutions after
their dilution to a final concentration of 1 mM.
within the whole duration of the experiment, indicating that the
binding of Pt(II) to the DNA still continues after more than 16
h of incubation. The time after which half of the maximum
absorbance is reached for each solution is reported in Table 1.
The more the [PtCl4]2- solution is hydrolyzed, the faster the
helix-breakdown process is, indicating a faster binding of the
complexes to the DNA. Like in the case of cis-Pt(NH3)2Cl2,
the binding process appears thus to be limited by the hydrolysis
of the complexes.
Figure 5. Electrophoresis experiments of single and double stranded
DNA after 16 h of activation with PtCl42-. Lane 1: single-stranded
λ-DNA. Lane 2: double-stranded λ-DNA. Lane 3: single-stranded
λ-DNA digested with Hind III. Lane 4: double-stranded λ-DNA
digested with Hind III.
indicates that the process of Pt(II) binding to ds-DNA does not
break the double strand configuration.41
2-
3. DNA Morphology after Pt(II) Binding. The binding of Pt(II)
complexes to ds-DNA is known to cause an overall shrinkage
of the molecule length, as it was previously noticed in SFM
experiments at low Pt/base ratios.42 Here we investigate by
means of SFM the structural changes which DNA undergoes
2. Double-Stranded DNA Structure after PtCl4 Binding.
From the kinetics measurements presented above, it is unclear
whether the DNA still retains its double-stranded structure after
the binding of Pt(II) complexes. Indeed, the binding of Pt(II)
complexes could break the hydrogen bonds between the strands
causing a transition from double-stranded to single-stranded
DNA, which could as well lead to an increase of the absorbance
peak at 260 nm. To rule out this possibility, we perform a series
of gel electrophoresis experiments (Figure 5). Solutions of
double strand λ-DNA, thermally denatured λ-DNA, λ-DNA
digested with Hind III, and thermally denatured λ-DNA digested
with Hind III are incubated with a 1 mM aged solution of
K2PtCl4. After 20 h of incubation, the four solutions are treated
with a highly diluted DMAB solution and let run in an agarose
gel. Subsequent gold staining (see Experimental Section) reveals
the appearance of clear bands in the case of both full length
and digested λ-DNA (lanes 2 and 4 of Figure 5). However, no
bands appear in the lanes of the previously thermally denaturated
ss-DNA samples (lanes 1 and 3 of Figure 5). This clearly
2-
upon PtCl4 binding at a Pt/base ratio of about 65:1 (Figure
6). Compared with native λ-DNA (Figure 6a), the incubation
with hydrolyzed PtCl42- ions reveals the formation of kinks and
coils in the DNA structure (Figure 6, parts b and c). This results
in an overall shortening of the DNA molecules from a native
length of about 16 µm to lengths in the range of 1.5-2.9 µm.
The appearance of these short segments cannot be due to rupture
of long molecules, as this would not result in the appearance
of a single, sharp λ-DNA band in the electrophoresis experi-
ments (see Figure 5, lane 2), which is indicative of a unique
molecule weight.
D. Synthesis of Pt Cluster Chains in the Presence of
Activated DNA. In section B we have shown that the reduction
of metal salt in the presence of nonactivated DNA leads to the

10806 J. Phys. Chem. B, Vol. 108, No. 30, 2004
Seidel et al.
Figure 6. AFM images of DNA after incubation with PtCl42-. (a) Native λ-DNA. (b) λ-DNA after 1 h of incubation. (c) λ-DNA after 16 h of
incubation.
formation of particle aggregates. They are in part bound to the
DNA5,6 but also form bigger agglomerates of homogeneously
nucleated particles. As demonstrated in ref 6, the formation of
long cluster chains via a clean, purely heterogeneous cluster
formation process necessarily requires the activation of the DNA
for about 20 h prior to the reduction.5,6,16 At fixed concentration
of Pt(II) complexes (1 mM), analysis of metallization products
at different concentrations of activated DNA (5, 20, and 100
µg/mL) revealed that continuous chains were obtained at a DNA
concentration of 5 µg/mL. These concentrations correspond to
a Pt(II)/base ratio of about 65:1, and will be considered as the
“standard” concentrations in our metallization experiments. After
long activation, about 1.5 complexes are expected to be bound
to each base on average,40 which means that less than 3% of
the total complexes are bound to the DNA. Despite the large
amount of free complexes in solution, no large agglomerates
of homogeneously formed particles are present after reduction
with DMAB (see Figure 7a and refs 5, 6, and 16). In other
words, the balance between the homogeneous and heterogeneous
nucleation processes is shifted from mixed homogeneous and
heterogeneous nucleation in the case of nonactivated DNA to
purely heterogeneous nucleation in the case of activated DNA.6
A numbers of parameters which control this balance are
investigated below.
The parameters which in principle can influence the metal-
lization process are the absolute and relative concentrations of
the reactants (metal complexes, DNA, reducing agent); the type
of metal salt and its hydrolysis degree; the strength of the
reducing agent (i.e., both the reduction potential between metal
complexes and reducing agent, and the kinetics of electron
transfer); the temperature and the pH of the reducing bath; the
composition of the DNA sequence.16 Finally, variations of the
activation time influence both the number of complexes which
bind to the DNA prior to reduction and, possibly, the hydrolysis
degree of the bound complexes. Here we study how changing
the absolute concentrations of reactants and the metal type
influence the metallization process. Furthermore, we analyze
the metallization products obtained with different DNA molec-
ular structures (ds-DNA, ss-DNA, and single nucleotides) and
in the presence of citrate as a capping agent for the formed
clusters. The effect of the guanine content of the DNA template
on the kinetics of metallization has been studied in ref 16 and
will be discussed later in this paper. The strength of the reducing
agent, the temperature, and the pH of the reducing bath can be
considered parameters for a “fine-tuning” of the metallization
procedure, and a systematic study of their effects on the
metallization products will require further investigations.
1. Variation of the Absolute Concentration of the Reactants.
Varying the absolute concentrations of the reactants (Pt salt,
DNA, and reducing agent) can have a dramatic effect on the
reaction products. Therefore, we perform metallization experi-
ments (activation followed by reduction) using Pt(II) concentra-
tions of 1/4, 1/2, 2, and 4 times the standard concentration, while
keeping constant the concentration ratios of the reactants. A
chain of clusters obtained under the standard conditions is shown
as a reference in Figure 7a. The inset of Figure 7b shows a
high resolution TEM image of the same preparation, which
confirms the metallic character of the obtained particles (see
the caption of Figure 7b). Remarkably, also when the absolute
concentrations of the reactants are either half or twice the
standard ones, we obtain very clean samples with purely
heterogeneously formed cluster chains (Figure 7c). Instead, in
the case of either one-fourth or four times the standard
concentrations, only random cluster agglomerates are found on
the sample’s surface (not shown).
2. Variation of the DNA Molecular Structure. To investigate
the influence of the DNA structure on the morphology of the
metallization products, we perform reduction experiments in the
presence of ss-DNA. The persistence length of ss-DNA is about
1 nm, i.e., much smaller than the persistence length of ds-DNA
which is ∼100 nm in solutions with ionic strength of about 1
mM.43 Thermally denaturated λ-DNA is activated for 16 h, and
subsequently, DMAB is added to the solution. The metallization
products are shown in Figure 7d. In contrast with the case of
ds-DNA, only very short cluster chains and random aggregates
of small clusters are obtained. The small clusters appear to be
stabilized to a certain extent by the ss-DNA, which act as a
capping agent in a similar way as nonactivated ds-DNA (cf.
Figure 3).
To study the capping action of DNA for the formed clusters
in more detail, we perform reduction experiments in the presence
of a mixture of single nucleotides, dATP, dTTP, dCTP, and
dGTP. The relative ratios of the different nucleotides are chosen
to match the base composition of λ-DNA. The nucleotide
mixture is incubated for 16 h with an aged solution of K2PtCl4
at the standard reaction concentration. Subsequently, DMAB
is added to the solution, and the metallization products are
imaged with TEM. The formed clusters are shown in Figure
7e. The surface of the TEM grid is covered by separated small
polyhedral crystallites surrounded by a capping layer, indicating
that single nucleotides efficiently stabilize the formed clusters.
A few bigger agglomerates are also present, which could
presumably be avoided by more carefully tuning the concentra-
tions of the reactants (here they are kept equal to those of the
standard DNA metallization protocol).
3. Effect of Citrate. Citrate ions are known to be good capping
agents for metal clusters31 (cf. Figure 1b). With the aim of
improving the quality of the obtained cluster chains, we add
citrate to a solution of activated DNA immediately before
starting the reduction.44 In the presence of citrate, the metal-
lization kinetics is slightly slowed, as expected in the presence
of capping agents.31 The obtained products are extremely regular

Platinum Cluster Chains on DNA Templates
J. Phys. Chem. B, Vol. 108, No. 30, 2004 10807
Figure 8. TEM micrograph of a chain of Pt clusters grown on λ-DNA
in the presence of citrate (see text).
Figure 9. TEM micrographs of chains of (a) Pt and (b) Pd clusters
grown on λ-DNA in the presence of citrate at our standard concentra-
tions (see text). Note the presence of homogeneously nucleated particles
in the case of palladium.
Figure 7. TEM micrographs of the metallization products formed after
reduction of a PtCl42- solution with DMAB at different concentrations
in the presence of ds-DNA, ss-DNA, and dissolved nucleotides. (a) 5
µg/mL λ-DNA, 1 mM PtCl42-, and 1 mM DMAB. (b) as in (a), at
higher resolution. Power spectrum analysis of multiple particles revealed
interlayer distances correspondent to the [111], [200], and [202]
crystallographic planes of bulk Pt, thus confirming the metallic character
of the obtained particles. (c) 10 µg/mL λ-DNA, 2 mM PtCl42-, and 2
mM DMAB. (d) 5 µg/mL single-stranded λ-DNA, 1 mM PtCl42-, and
1 mM DMAB. (e) 5 µg/mL of nucleotides at a composition corre-
sponding to that of λ-DNA, 1 mM PtCl42-, and 1 mM DMAB.
together and form continuous wires with a typical length of 50-
100 nm (cf. Figure 7b), in the presence of citrate, they are clearly
separated by a ligand shell (see also Figure 9a).
4. Comparison Pd Vs Pt. As demonstrated in ref 6, increasing
the activation time leads to an acceleration of the overall kinetics
of metallization. At the same time, the nucleation behavior
changes from mixed heterogeneous and homogeneous nucleation
to purely heterogeneous nucleation (see ref 6). However, since
only less than 3% of the Pt(II) complexes are bound to the DNA,
this can only happen if the heterogeneous nucleation channel
is considerably faster than the homogeneous nucleation channel.
If, on the other hand, the homogeneous process itself is too
fast, then the formation of homogeneously nucleated particles
remains highly probable. This would lead to undesired ag-
glomeration processes and formation of thicker metallic wires
and/or cluster agglomerates. We investigate this issue using K2-
PdCl4 instead of K2PtCl4 as the salt precursor for both the
activation and the subsequent metallization. Indeed, the reduc-
tion of Pd(II) complexes is known to be faster than the reduction
of the corresponding Pt(II) complexes.45
chains of clusters of approximately the same size (∼5 nm),
which uniformly cover the DNA template over its whole length
(Figure 8). Neither bigger agglomerates of homogeneously
formed clusters nor random cluster aggregates are present in
any of the investigated samples, whose backgrounds appear to
be consistently clean of any spurious product of homogeneous
metallization processes. In comparison with the chains obtained
in the absence of citrate, the chains obtained here are signifi-
cantly less curled (compare, for instance, Figure 8 with Figure
7a). Most of the crystallites are of regular shape, which is a
direct influence of the capping action of citrate. Whereas without
citrate the single clusters composing the chains tend to grow

10808 J. Phys. Chem. B, Vol. 108, No. 30, 2004
Seidel et al.
In reduction experiments in the absence of DNA, the
characteristic time for the reaction of an aged 1 mM solution
of [PdCl4]2- ions with DMAB lies below one second, and only
large cluster agglomerates could be found in TEM investigations
(not shown). The addition of citrate has a similar effect as shown
in the case of Pt(II) complexes, leading to separate, crystalline
clusters, as well as a number of bigger agglomerates (cf. Figure
1b). When Pd(II) ions are reduced in the presence of activated
DNA and citrate, the characteristic time for metal formation is
about 30 s. In comparison, the metallization starting from
PtCl42- in the presence of citrate is significantly slower, being
completed after about 8 min. TEM images of the obtained
metallization products are shown in Figure 9. In the case of Pt,
all obtained clusters appear to be grown specifically onto the
DNA template (Figure 9a) forming long and ordered chains over
the whole DNA length. Instead, in the case of Pd, a significant
number of isolated clusters are present. These are clearly the
result of homogeneous nucleation and growth processes (Figure
9b). Chains of Pd clusters along the DNA template are still
obtained, but the metallization is coarser than in the case of Pt
due to the fact that homogeneously formed clusters form
aggregates on the heterogeneously formed chains (see also ref
2).
that the binding of Pt(II) complexes to DNA bases prior to
chemical reduction of the Pt(II)/DNA solution is necessary to
promote a uniform, template-controlled growth of clusters along
the DNA templates.6 We call this step ‘activation’ of the DNA
template.
B. Activation of the DNA Template. From the spectroscopic
experiments reported in section III.C, it is clear that the binding
kinetics of dissolved [PtCl4]2- ions to λ-DNA depends on the
hydrolysis degree of the Pt(II) complexes. This is in agreement
with the data for cisplatin at low Pt/base ratios.39 At the Pt(II)
concentration used in our standard metallization protocol (10
mM stock solution diluted to 1mM prior to reduction), the
absorbance reaches half of its maximum value after a charac-
teristic time of about 3 h (Table 1), and still increases after more
than 16 h. This behavior is in general consistent with the results
of the metallization experiments reported in refs 6 and 16.
However, in those works, the reduction rate (which is mainly
related with the number of Pt(II) complexes bound to the DNA)
was found to increase even after more than 30 h of activation.
This result indicates that the DNA metallization kinetics is also
influenced by processes which take place after the number of
Pt(II) complexes bound to DNA has reached the saturation value
of 1.5 complexes per base.40 Processes characterized by
extremely slow kinetics are, for instance, the formation of
bifunctional complexes and the further hydrolysis of bound
complexes.39 Evidence that they can influence the kinetics of
DNA metallization comes from previous theoretical work, where
it was found that the hydrolysis of DNA‚Pt(II) complexes have
a great influence on the stability of the first-formed Pt nuclei.20
In the same work, bifunctional (and fully hydrolyzed) complexes
were found to have higher electron affinity with respect to
monofunctional (and less hydrolyzed) complexes.
IV. Discussion
A. Reduction in the Presence of Nonactivated DNA. In
the presence of nonactivated DNA, the overall kinetics of the
process of metal cluster formation is slowed (see Figure 2). This
effect is associated with the hindering of the aggregation of small
metal crystallites into bigger agglomerates (cf. Figure 1a with
Figure 3), strongly suggesting that nonactivated DNA acts like
a capping agent for the metal clusters. Indeed, the reduction
kinetics during the formation of metal particles is known to be
retarded with increasing concentration of capping agents (see,
e.g., ref 31). The capping action of DNA is mostly evident for
C. Formation of Cluster Chains. The results of our
metallization experiments show that the Pt(II) complexes which
are bound to the DNA act as very efficient nucleation centers
for the growth of small crystalline particles. The mechanisms
of the initial nucleation process of noble metal clusters in
reducing baths have been extensively studied by means of first
principles molecular dynamics simulation, both in the homo-
geneous case47,22 and in the heterogeneous case.6,20 In these
works, it was found that the initial formation of Pt dimers
between unreduced Pt(II) complexes and singly reduced Pt(I)
complexes is the limiting step of the whole cluster formation
process, especially at high concentration of metal salt and in
the presence of mild reducing agents.47 After the initial
nucleation steps, the nucleus (which is not necessarily com-
pletely reduced to the zerovalent state) continues to grow via
steps of addition of unreduced Pt(II) complexes22 and reduction
of the whole cluster, according to an autocatalytic mechanism.21
2-
the reduction of PtCl4 in the presence of single nucleotides.
In these experiments, the Pt/nucleotide ratio is chosen to be
about 65:1. Thus, only very few complexes are bound to
nucleotides before the chemical reduction begins. Despite this,
the obtained metal particles appear to be perfectly stabilized
and form small crystallites surrounded by a ligand shell (see
Figure 7e). This clearly proves the great affinity of nucleotides
to the metal surface, to which they are expected to bind mainly
via covalent bonds between surface metal atoms and deproto-
nated nitrogen atoms of the bases.
Although the aggregation of small clusters is hindered by
the presence of DNA, the processes of nucleation and growth
of the small crystallites are expected to be influenced only to a
minor extent. The reason for this is that the metal reduction is
relatively fast. It is completed within a few minutes (Figure 2),
whereas the characteristic time for binding Pt(II) to the DNA
bases varies from 1 to 10 h (Figure 4), depending on the
hydrolysis degree of Pt(II). Therefore, homogeneous metal
reduction is expected to take place before Pt(II) complexes will
form adducts with DNA.46 In fact, if the metal salt is reduced
immediately after addition of dissolved DNA, particles grow
according to a prevalently homogeneous nucleation channel, as
it was observed in previous SFM experiments.5,6 The chainlike
aggregate structures visible in Figure 3 are most likely due to
aggregation processes after the formation of small clusters,
similarly to the case of absence of DNA but presence of citrate
(see Figure 1b). These results show that uncontrolled adsorption
of homogeneously formed clusters to the DNA does not lead
to the formation of long, ordered cluster chains. This means
In the presence of activated DNA, the initial reduction
reaction can involve either dissolved Pt(II) complexes or DNA‚
Pt(II) complexes. On one hand, our previously performed
theoretical work6,20 suggests an enhancement of the electronic
affinity of Pt(II) complexes after binding to the DNA (see, in
particular, ref 20 for a thorough discussion about possible DNA‚
Pt(II) species involved in the reduction reaction). This is due
to the presence of the heterocyclic DNA bases as ligands for
Pt(II), which can “accommodate” the additional electrons in a
more favorable way. On the other hand, reduction in solution
cannot be excluded a priori because of the large number of
dissolved complexes. However, our first principles molecular
dynamics studies led to the conclusion that the heterogeneous
formation of a Pt(I)-Pt(II) dimer at DNA is considerably
favored over the analogous homogeneous reaction in solution

Platinum Cluster Chains on DNA Templates
J. Phys. Chem. B, Vol. 108, No. 30, 2004 10809
irrespective whether the initial reduction involve a free or a
bound Pt(II) complex.6,20 In other words, the presence of DNA
bases as ligands for Pt(II) complexes has both the effect of
enhancing the electron affinity of the metal complexes and
increasing the bond energy of the first-formed Pt-Pt bonds
during the reduction process. This means that the formation of
heterogeneous nucleation centers (the DNA‚Pt(II) adducts
formed during the activation) leads to a locally enhanced
nucleation kinetics along the DNA. Since the clusters grow in
an autocatalytic, nucleation-limited way, this leads in turn to
an oVerall enhancement of the metallization kinetics, as observed
in the experiments reported in ref 6. Therefore, enhancement
of the stability of the first-formed nuclei due to the complexation
of Pt(II) by heterocyclic systems such as the DNA bases leads
to a purely heterogeneous growth of particles on the DNA under
the right reaction conditions. In particular, the concentration of
metal complexes and the strength of the reducing agents should
be carefully tuned to ensure that the autocatalytic growth started
from the heterogeneous nuclei quickly consume the feedstock
of free metal complexes before homogeneous processes lead
to the formation of free clusters in the solution. If this condition
is fulfilled, then the metallization occurs via a selectively
heterogeneous, DNA-controlled mechanism.
the homogeneous nucleation channel, resulting in the formation
of random aggregates of homogeneously nucleated particles.
3. Metallization of ss-DNA. In section III. D.2, we have shown
that regular chains of particles cannot be grown in solution on
single stranded DNA with our method.49 Compared with double
stranded DNA, the extent of random aggregation is increased,
because the more flexible structure is not able to separate
efficiently the heterogeneously formed clusters. Indeed, this can
only be achieved if the persistence length of the molecular
template is larger than the cluster diameter. Therefore, it is
necessary to use ds-DNA as metallization template to obtain
the formation of regular wire-like metallic structures in solu-
tion.50
4. Effect of Citrate. Citrate is known to act as an efficient
capping agent for noble metal clusters (see ref 31 and Figure
1b). Therefore, we have added citrate prior to the reduction of
platinum in the presence of DNA to fully suppress the local
agglomeration of heterogeneously nucleated particles. Indeed,
the addition of citrate ions leads to less curled and, therefore,
longer cluster chains than those obtained in the absence of citrate
(see Figure 8). However, the clusters obtained in this way are
no longer in metallic contact. They are surrounded by a ligand
shell, so that the obtained cluster necklace is not expected to
be conductive. A thermal treatment after formation of the
structure may be possibly sufficient to decompose the citrate
layer restoring the metallic continuity.
1. Effect of the Kinetics of Metal Reduction. Consistently with
the discussion above, the balance between heterogeneous and
homogeneous nucleation appears to be directly influenced by
the kinetics of the reduction process. Indeed, in the metallization
experiments performed with Pd(II) complexes, which are
reduced considerably faster than Pt(II) complexes, both the
heterogeneous formation of cluster chains on the DNA and the
homogeneous formation of metal particles in solution are
simultaneously observed (see Figure 9). Especially in the
absence of additional stabilizers such as citrate, these homo-
geneously formed particles can form random aggregates. This
generally results in a much coarser and irregular metallization,
as reported in refs 2 and 3, where the obtained wires have a
diameter of about 50 nm and the formation of bigger, poly-
crystalline agglomerates of clusters is evident.48
On the other hand, when citrate is not added as a stabilizer,
the produced wires appear to be continuous over distances of
about 50 to 100 nm (see Figure 7b). However, small gaps are
still present, which are expected to interrupt the metallic
conductivity over longer distances. We expect that a further
selective reduction of metal complexes on the cluster chains
(similarly to the protocol used to produce “core-shell” nano-
particles51) will fill the gaps without increasing the thickness
of the wire of more than a few nm (thus keeping the wire
diameter below 10 nm). Investigations of this possibility, as
well as measurements of the I/V characteristics of the obtained
structures are currently under investigation in our laboratory.
2. Effect of Changing the Concentration of Reactants. In our
investigations, we have identified a standard set of concentra-
tions which results in a purely heterogeneous growth of platinum
on DNA templates. In addition, we have shown that small
deviations from the standard conditions (from 1/2 up to 2 times
the standard concentration) have no effect on the quality of the
metallization products. However, larger deviations (1/4 and 4
times the standard concentration) have a dramatic effect on the
metallization products, and in both cases random agglomerates
of particles are obtained. On one hand, when the concentration
of Pt and DNA is too high, the formation of agglomerates is
expected to happen both as a result of aggregation of the formed
metallic chains, and, possibly, of the formation of links between
different DNA strands by Pt(II) complexes during the activation
step. On the other hand, when the concentration of Pt and DNA
is too low, the formation of metal clusters is likely to occur
through the classical mechanism of full reduction of the metal
salt to the zerovalent state, followed by agglomeration of Pt(0)
complexes. Namely, as discussed in ref 47, at very low
concentration of Pt salt unreduced complexes are expected to
have little effect on the mechanism of cluster formation.
However, to observe a catalytic activity of DNA in promoting
heterogeneous metal formation, the nucleation process neces-
sarily has to involved unreduced complexes. This means that
at too low concentration of metal salt the balance between
homogeneous and heterogeneous nucleation is shifted toward
5. Purity of the Pt Nanoparticles. Pt powders produced by
reduction of Pt(II) complexes with DMAB under air predomi-
nantly reveal metallic character. PtOx phases are present but
are limited to the outermost atomic layers of the particles (see
section III.A and refs 26, 34, and 35). Boron impurities are
expected to be present in the bulk of the particles at a
concentration of about 0.05 wt %.29,36 We assume that the
presence of DNA has little or no influence on the amount of
impurities in the particle chains. In fact, HR-TEM investigations
(see Figure 7b) confirm the metallic character of the synthesized
clusters. Moreover, evidence of the predominantly metallic
character of structures obtained using a similar protocol comes
from previous work on 50 nm thick Pd wires grown on DNA
templates after the reduction of Pd complexes with DMAB.3,8
Conductivity measurements showed that these wires present
ohmic behavior, and that their specific resistivity is only 10
times larger than that of bulk palladium.3,8 Electron scattering
at the grain boundaries of the clusters was identified as the
dominant scattering mechanism.3 Whether this effect is also
relevant for the conductivity of the ultrathin cluster chains
composed by series of single clusters remains to be investigated.
Certainly, the present impurities will affect the conductivity of
the particle chains in some way. To reduce the amount of
impurities to the minimum, the cluster chains could be produced
under exclusion of oxygen, or using hydrogen instead of DMAB
as reducing agent.

10810 J. Phys. Chem. B, Vol. 108, No. 30, 2004
Seidel et al.
6. Effects of the DNA Composition. All of the investigations
presented here are carried out using DNA molecules with a
random sequence and a content of GC base-pair between 40
and 50%. Thus, the influence of sequence and composition of
the DNA template is excluded from the present study. However,
we briefly discuss in this section results of previous investiga-
tions showing a peculiar influence of the content of GC vs AT
base pairs on the kinetics of the reduction reaction in the
presence of DNA.16
The reaction conditions described in this work are tuned to
fabricate chains of metal clusters in the bulk solution. However,
application in electronic circuits or sensors will require the in-
situ metallization of molecules bound to the surface of a solid
substrate. The optimal reaction conditions for the metallization
of molecules on a solid surface should be the subject of further
investigations. Moreover, although continuous metal wires are
obtained over relatively short distances, long chains of clusters
invariably present small gaps that are expected to limit their
conductivity. Additional work, including systematic measure-
ments of the electric characteristic of the cluster chains and the
continuous metal wires, should be devoted to address the
problem of closing these gaps. If a thermal annealing of the
metallized DNA is not sufficient, then the selective deposition
of a thin metal layer on the formed clusters should lead to
continuous wires with a diameter below 10 nm. Moreover, the
influence of the impurities present in the bulk and on the surface
of the obtained metallic structures on their conductivity remains
to be investigated.
Finally, previously performed metallization experiments using
DNA with different GC contents revealed that the metallization
kinetics can be remarkably influenced by the nucleotide
composition of the DNA sequence under particular conditions.16
We stress that the metallization procedure presented in this paper
is not designed to be affected by the DNA sequence and
composition. However, we expect that the present metallization
protocol can be modified by carefully acting on the kinetics of
Pt(II) binding and reduction to develop a procedure for a
sequence-specific metallization of DNA, where only guanine-
rich sequences become fully metallized. This would enable a
direct control of the metal growth by the DNA template at an
unprecedented level.
Since for short interaction times Pt(II) complexes are known
to bind primarily to the guanine bases,37 one would expect that
the reaction kinetics becomes increasingly accelerated with
increasing content of GC base pairs in the DNA used. This issue
was investigated performing metallization experiments using
three different types of DNA, containing 27%, 42%, and 72%
of GC pairs.16 The metallization rate was found to be indepen-
dent of the DNA content when no complexes are bound to the
DNA bases (no activation), or after very long activation times,
where all metal-binding sites along the DNA are indiscriminately
occupied by Pt(II) complexes. However, for a window of
activation time ranging from 30 to 300 min, the reduction
process was found to be faster for higher G contents. Indeed,
in this time window the kinetics of Pt(II) binding to guanine is
substantially faster than the binding to other bases.37 Thus, in
this time window, the number of complexes bound to the DNA
depends directly on the GC content of the DNA. This confirms
the active role of the formed Pt(II)‚DNA adducts in promoting
an acceleration of the reduction reaction. In particular, these
experiments suggest that the DNA sequence could be designed
and the metallization performed in such a way that only G-rich
DNA sequences become fully metallized, while A-rich parts of
the template remain in their native form or are decorated by a
few, isolated clusters.
Acknowledgment. We are thankful to R. Goldberg, M.
Lehmann, H. Lichte, and R. Wahl for valuable discussions and
co-work. We gratefully acknowledge P. Rossbach and J.S.
Becker for the TOF-SIMS and XPS analysis of the Pt powders.
This work was partially supported by the Deutsche Forschungs-
gemeinschaft (Grants: FOR 335/2 and PO 392/18-1). The work
of L.C.C. is supported by the EPSRC, U.K.
V. Future Perspectives
In conclusion, we have described in detail a procedure for
the selective growth of metal on DNA templates according to
a heterogeneous nucleation and growth mechanism which is
controlled by the template itself. The growth of metal clusters
on other organic polymers, in particular on proteins, can be
achieved according to similar procedures.52-56 The described
metallization method is both simple and robust and thus in
principle suitable to be extended to larger scale processing.
Namely, although the molecular mechanisms of the processes
of heterogeneous cluster formation on DNA templates are rather
complicated, the sequence of steps of the metallization procedure
which permit the fabrication of uniform and thin cluster chains
appears to be astonishingly simple. Moreover, we stress that
small variation of the reactant concentrations are extremely
unlikely to affect the quality of the metallization products, once
the “standard” concentrations have been carefully chosen.
Therefore, we expect that the metallization procedure presented
here will be a good starting point toward the development of
large scale biopolymer metallization techniques. The overcoming
of the difficulties that will invariably be associated with the
scaling-up of the DNA metallization is a fundamental issue to
be addressed in future work.
References and Notes
(1) Braun, E.; Eichen, Y.; Sivan, U.; Ben-Yoseph, G. Nature 1998,
391, 775-778.
(2) Richter, J.; Seidel, R.; Kirsch, R.; Mertig, M.; Pompe, W.; Plaschke,
J.; Schackert, H. K. AdV. Mater. 2000, 12, 507-510.
(3) Richter, J.; Mertig, M.; Pompe, W.; Mo¨nch, I.; Schackert, H. K.
Appl. Phys. Lett. 2001, 78, 536-538.
(4) Harnack, O.; Ford, W. E.; Yasuda, A.; Wessels, J. M. Nano Lett.
2002, 2, 919-923. Ford, W. E.; Harnack, O.; Yasuda, A.; Wessels, J. M.
AdV. Mater. 2001, 13, 1793-1797.
(5) Seidel, R.; Mertig, M.; Pompe, W. Surf. Interface Anal. 2002, 33,
151-154.
(6) Mertig, M.; Colombi Ciacchi, L.; Seidel, R.; Pompe, W.; De Vita,
A. Nano Lett. 2002, 2, 841-844.
(7) Keren, K.; Krueger, M.; Gilad, R.; Ben-Yoseph, G.; Sivan, U.;
Braun, E. Science 2002, 297, 72-75.
(8) Richter, J.; Mertig, M.; Pompe, W.; Vinzelberg, H. Appl. Phys. A
2002, 74, 725-728.
(9) Yan, H.; Park, S. H.; Finkelstein, G.; Reif, J. H.; LaBean, T. H.
Science 2003, 301, 1882-1884.
(10) Deng, Z.; Mao, C. Nano Lett. 2003, 3, 1545-1548.
(11) Nakao, H.; Shiigi, H.; Yamamoto, Y.; Tokonami, S.; Nagaoka, T.;
Sugiyama, S.; Ohtani, T. Nano Lett. 2003, 3, 1391-1394.
(12) Monson, C. F.; Woolley, A. T. Nano Lett. 2003, 3, 359-363.
(13) Mertig, M.; Pompe, W. in: Nanobiotechnology - Concepts,
Applications and PerspectiVes; Niemeyer, C. M., Mirkin, C., Eds.; Wiley-
VCH: Weinheim, Germany, 2004; pp 256-277.
Various mechanistic aspects of the single steps composing
the metallization protocol has been addressed and discussed also
taking into account previously performed first principles ato-
mistic simulations. While the global metallization mechanisms
are now sufficiently clear, a number of more specific questions
still remain to be answered, which are expected to stimulate
future basic research work in the field of DNA metallization.
(14) Alivisatos, A. P.; Johnsson, K. P.; Peng, X.; Wilson, T. E.; Loweth,
C. J.; Bruchez, M. P.; Schultz, P. G. Nature 1996, 382, 609-611.

Platinum Cluster Chains on DNA Templates
J. Phys. Chem. B, Vol. 108, No. 30, 2004 10811
(15) Mirkin, C. A.; Letsinger, R. L.; Mucic, R. C.; Storhoff, J. J. Nature
1996, 382, 607-609.
with 4 bp long single stranded ends (sticky ends) having sequences defined
by the site of the enzymatic cut, which do not re-hybridize at room
temperature. Only the extremities of the molecule, that is, the 5′ end of the
23130 bp fragment and the 3′ end of the 4361 bp fragment are terminated
by 12 bp long sticky ends, which are known to hybridize at room
temperature. In fact, in the gel a band correspondent to a fragment of about
27500 bp is present, which derives from hybridization of a number of 4361
bp fragments with 23130 bp fragments. These fragments are held together
only by the hydrogen bonds between the complementary sticky ends, since
the linking of the backbones requires specific enzymatic action. The fact
that these long fragments are found after incubation with Pt(II) complexes
means that the double stranded configuration is maintained over a length
as short as 12 bp.
(16) Mertig, M.; Seidel, R.; Colombi Ciacchi, L.; Pompe, W. AIP Conf.
Proc. 2002, 633, 449-453.
(17) Zhang, Y.; Austin, R. H.; Kraeft, J.; Cox, E. C.; Ong, N. P. Phys.
ReV. Lett. 2002, 89, 198102-1-4.
(18) Seeman, N. C. Curr. Opin. Struct. Biol. 1996, 6, 519-526.
(19) Richter, J. Physica E 2003, 16, 157-173.
(20) Colombi Ciacchi, L.; Mertig, M.; Seidel, R.; Pompe, W.; De Vita,
A. Nanotechnology 2003, 14, 840-848.
(21) Watzky, M. A.; Finke, R. G. J. Am. Chem. Soc. 1997, 119, 10382-
10400.
(22) Colombi Ciacchi, L.; Pompe, W.; De Vita, A. J. Phys. Chem. B
2003, 107, 1755-1764.
(42) Onoa, G. B.; Cervantes, G.; Moreno, V.; Prieto, M. J. Nucleic Acids
Res. 1998, 26, 1473-1480.
(43) Baumann, C. G.; Smith, S. B.; Bloomfield, V. A.; Bustamante, C.
(23) Bezanilla, M.; Manne, S.; Laney, D. E.; Lyubchenko, Y. L.;
Hansma, H. G. Langmuir 1995, 11, 655-659.
(24) Sanders, C. I.; Martin, Don, S., Jr. J. Am. Chem. Soc. 1961, 83,
807-810.
Proc. Natl. Acad. Sci. 1997, 94, 6185-6190.
(44) Note that citrate is used in this work exclusively as a capping agent
and not as a reducing agent. This would require much higher reaction
temperatures and longer reaction times.
(45) Wu, M. L.; Chen, D. H.; Huang, T. C. J. Colloid Interface Sci.
2001, 243, 102-108.
(25) Elding, L. I. Acta Chem. Scand. 1970, 24, 1527-1540.
(26) Henglein, A.; Ershov, B. G.; Malow, M. J. Phys. Chem. 1995, 99,
14129-14136.
(27) Duff, D. G.; Edwards, P. P.; Johnson, B. F. G. J. Phys. Chem.
1995, 99, 15934-15944.
(28) Cotton, F. A.; Wilkinson, G.; Murillo, C. A.; Bochmann, M.
AdVanced Inorganic Chemistry, 6th ed.; John Wiley: New York, 1999.
(29) van Rheenen, P. R.; McKelvy, M. J.; Glaunsinger, W. S. J. Solid
State Chem. 1987, 67, 151-169.
(46) In fact, in an SFM analysis of metallization products in the presence
of 20 µg/mL of DNA, we found the presence of isolated small clusters
with a height of about 5 nm on the surface of the sample after 3 min of
reduction (image not shown), that is, when the measured turbidity is still
negligible.
(47) Colombi Ciacchi, L.; Pompe, W.; De Vita, A. J. Am. Chem. Soc.
2001, 123, 7371-7380.
(48) We note that the DNA metallization with Pd reported in ref 2 and
3 was performed at a much higher Pd:base ratio (about 2700:1) and with
shorter activation time (2 h).
(49) Note, however, that chains could actually be grown on single
stranded DNA when the molecules are stretched and fixed between two
metal contacts, which would avoid the formation of tight random coils.
(50) A way to keep ss-DNA linear over long distances could be the
attachment of its ends to a solid substrate prior to the activation and reduction
steps. This, however, will certainly require adjustments of the reaction
parameters and needs further investigations.
(51) Hodak, J. H.; Henglein, A.; Giersig, M.; Hartland, G. V. J. Phys.
Chem. B 2000, 104, 11708-11718.
(52) Shenton, W.; Pum, D.; Sleytr, U. B.; Mann, S. Nature 1997, 389,
585-587.
(53) Mertig, M.; Kirsch, R.; Pompe, W.; Engelhardt, H. Eur. Phys. J.
D 1999, 9, 45-48. Mertig, M.; Wahl, R.; Lehmann, M.; Simon, P.; Pompe,
W. Eur. Phys. J. D 2001, 16, 317-320.
(30) In principle, it is possible to tune the concentrations of Pt(II) and
capping agent and the strength of the reducing agent to obtain monodisperse
colloidal solution of metal clusters.29,31,32 The conditions applied here mainly
serve as a reference for the later described experiments of DNA metalliza-
2-
tion, where the same concentrations of PtCl4 and DMAB are used.
(31) Henglein, A.; Giersig, M. J. Phys. Chem. B 2000, 104, 6767-
6772.
(32) Henglein, A.; Giersig, M. J. Phys. Chem. B 1999, 103, 9533-
9539.
(33) Henglein, A.; Lindig, B.; Westerhausen, J. Radiat. Phys. Chem.
1984, 23, 199.
(34) Schindler, A.; Ko¨nig, R.; Herrmannsdo¨rfer, T.; Braun, H. F.; Eska,
G.; Gu¨nther, D.; Meissner, M.; Mertig, M.; Wahl, R.; Pompe, W. Europhys.
Lett. 2002, 58, 885-891.
(35) Rossbach, P.; Becker, J. S. private communication.
(36) Bo¨nnemann, H.; Brijoux, W.; Brinkmann, R.; Fretzen, R.; Joussen,
T.; Ko¨ppler, R.; Korall, B.; Neiteler, P.; Richter, J. J. Mol. Catal. 1994,
86, 129-177.
(37) Macquet, J. P.; Theophanides, T. Biopolymers 1975, 14, 781-799.
Macquet, J. P.; Theophanides, T. Inorg. Chim. Acta 1976, 18, 189-194.
Macquet, J. P.; Butour, J. L. Eur. J. Biochem 1978, 83, 375-387.
(38) Cisplatin: chemistry and biochemistry of a leading anticancer drug;
Lippert, B., Ed.; Wiley-VCH: Weinheim, Germany, 1999.
(39) Bancroft, D. P.; Lepre, C. A.; Lippard, S. J. J. Am. Chem. Soc.
1990, 112, 6860-6871.
(54) Kirsch, R.; Mertig, M.; Pompe, W.; Wahl, R.; Sadowski, G.;
Boehm, K. J.; Unger, E. Thin Solid Films 1997, 305, 248-253. Mertig,
M.; Kirsch, R.; Pompe, W. Appl. Phys. A 1998, 196, S723-S727.
(55) Behrens, S.; Rahn, K.; Habicht, W.; Bo¨hm, K.-J.; Ro¨sner, H.;
Dinjus, E.; Unger, E. AdV. Mater. 2002, 14, 1621-1625.
(40) Sacharenko, E. T.; Moschkowski, Yu. S. Biofizika 1972, 7, 373-
378, in Russian.
(41) The Hind III digested double stranded λ-DNA consists of fragments
composed by 23130, 2027, 2322, 9416, 564, 6682, 4361 bp, each terminated
(56) Dujardin, E.; Peet, C.; Stubbs, G.; Culver, J. N.; Mann, S. Nano
Lett. 2003, 3, 413-417.