Fabrication of asymmetric molecular junctions by the oriented assembly of dithiocarbamate rectifiers

Article abstract of DOI:10.1021/ja110244j

The oriented assembly of molecules on metals is a requirement for rectification in planar metal-molecule-metal junctions. Here, we demonstrate how the difference in adsorption kinetics between dithiocarbamate and thioacetate anchor groups can be utilized

Full text of DOI:10.1021/ja110244j

ARTICLE
pubs.acs.org/JACS
Fabrication of Asymmetric Molecular Junctions by the Oriented Assembly
of Dithiocarbamate Rectifiers
Deqing Gao,† Frank Scholz,† Heinz-Georg Nothofer,† William E. Ford,† Ullrich Scherf,§ Jurina M. Wessels,†
Akio Yasuda,† and Florian von Wrochem*,†
†Sony Deutschland GmbH, Materials Science Laboratory, Hedelfinger Strasse 61, 70327 Stuttgart, Germany
§Bergische Universit€at Wuppertal, Makromolekulare Chemie und Institut f€ur Polymertechnologie, Gauss-Strasse 20, 42097 Wuppertal,
Germany
S Supporting Information
b
ABSTRACT: The oriented assembly of molecules on metals is a
requirement for rectification in planar metalꢀmoleculeꢀmetal
junctions. Here, we demonstrate how the difference in adsorption
kinetics between dithiocarbamate and thioacetate anchor groups
can be utilized to form oriented assemblies of asymmetric mol-
ecules that are bound to Au through the dithiocarbamate moiety.
The free thioactate group is then used as a ligand to bind Au
nanoparticles and to form the desired metalꢀmoleculeꢀmetal
junction. Besides allowing an asymmetric coupling to the electrodes, the molecules exhibit an asymmetric molecular backbone where the
length of the alkyl chains separating the electrodes from a central, para-substituted phenyl ring differs by two methylene units. Throughout
the junction fabrication, the layers were characterized by photoelectron spectroscopy, infrared spectroscopy, and scanning tunneling
microscopy. Large area junctions using a conducting polymer interlayer between a mercury-drop electrode and the self-assembled
monolayer prove the relationship between electrical data and molecular structure.
’ INTRODUCTION
For the fabrication of lithographically defined, micrometer- to
nanometer-scale (μm- to nm-scale) molecular junctions, the for-
mation of ultrathin self-assembled monolayers in a solution
deposition process is one of the most promising assembly
strategies.7,12,13 However, the realization of diodic junctions
featuring a well-defined electrical response requires an oriented
growth of asymmetric molecules at metal surfaces.14 Moreover, a
well-defined packing density and tilt angle of the molecular
backbone within the monolayer are essential to obtain reprodu-
cible currentꢀvoltage characteristics. Recently, the oriented
growth of asymmetric molecules on noble metals has been
achieved with dithiols having different, orthogonal protective
groups, such as trimethylsilylethyl,15 acetyl,16 and cyanoethyl.14
These protective groups can be removed sequentially, allowing
selective binding to the substrate via the desired thiolate end
group. Nevertheless, the synthesis of such asymmetrically pro-
tected dithiol derivatives is demanding, and processing condi-
tions for the sequential removal of orthogonal protective groups
have to be accurately controlled. Furthermore, no asymmetry in
the electrical coupling to the electrodes is possible if thiolates are
used on both sides of the junction.
The physical limits encountered in the miniaturization of
CMOS devices and the exponential increase in the production
cost of future CMOS-based electronics1 have triggered a growing
research effort in the area of molecular electronics.2ꢀ4 Conse-
quently, the development of concepts and experimental test-beds
for the implementation of organic molecules in electronic devices
has attracted significant interest.5 For the realization of organic-
based memories6,7 and field-programmable gate arrays,8 the
development of molecular rectifiers is an issue of fundamental
importance. By definition, a rectifier (diode) is a two-terminal device
with a unipolar currentꢀvoltage characteristic. The first concept of
a unipolar molecular diode, based on a donor-insulator-acceptor
(D-σ-A) model, was proposed by Aviram and Ratner.9 According to
this model, a recombination of opposite charges previously transferred
from the electrodes into the highest occupied molecular orbital
(HOMO) and the lowest unoccupied molecular orbital
(LUMO) of two decoupled units within the rectifier (Dþ-σ-
Aꢀ zwitterion state) takes place.10 Williams et al. described a
different mechanism for molecular rectification where a single
electroactive unit (e.g., a phenyl ring) is asymmetrically posi-
tioned between two electrodes.11 In this case, the alignment of
the HOMO and LUMO of the phenyl ring with the Fermi levels
of the contacts determines the asymmetry of the currentꢀvol-
tage characteristics, and rectification ratios of up to 500 were
predicted by appropriately tuning the tunneling barrier lengths,
which are given by two insulating alkyl chains.11
An alternative pathway for the directional assembly at sur-
faces is the usage of two different anchor groups. In the
past, dithiocarbamate anchor groups have been investigated in
Received: November 15, 2010
Published: March 28, 2011
r
2011 American Chemical Society
5921
dx.doi.org/10.1021/ja110244j J. Am. Chem. Soc. 2011, 133, 5921–5930
|

Journal of the American Chemical Society
ARTICLE
was reacted with acetic anhydride to give amide 4 which was then
used as starting material for the synthesis of 1 and 2. Compounds
5 and 6 were prepared by FriedelꢀCrafts acylation of 4. Borane is
known as a good reducing agent for amides without reacting with
halides.28 First, borane-tetrahydrofuran (BH3 THF) was chosen
3
to carry out the reduction of the two ketone groups of 5 or 6.
However, characterization of the resulting products by 1H NMR
analysis showed that, in case of both compounds, the amide
group was reduced to amine, whereas the carbonyl group next to
the aryl was only reduced to alcohol. Thus, a two-step reduction
was carried out. In the first step, triethylsilane (Et3SiH) in
trifluoroacetic acid (TFA)29 was used to reduce the aryl ketone
to methylene in high yield. In a second step, the amide was re-
Figure 1. Molecular structures of dithiocarbamate/thiol compounds 1
and 2 and of compound 3, which was used as a reference in this study.
duced to amine by using borane-dimethyl sulfide (BMS, BH3
3
S(CH3)2), which is easier to handle than BH3 THF.30 The
3
metalꢀnanoparticle networks17 and self-assembled monolayers.18ꢀ22
It has been shown experimentally22 and theoretically22,23 that the
conjugated dithiocarbamate anchor group strongly couples to
coinage metals, causing a mixing of discrete molecular levels with
metal states, thereby leading to a significant reduction of the
metalꢀmolecule charge injection barrier. This results in a lower
contact resistance than thiols at the metalꢀmolecule interface.
Furthermore, it was shown that dithiocarbamate monolayers are
thermally more stable than comparable thiolate monolayers.22
In this paper, we target the fabrication of monolayer-based
rectifier junctions using molecules having two different anchor
groups. On one hand, this enables a directional molecular
assembly on metal surfaces. On the other hand, an asymmetric
electrical coupling to the electrodes is possible. On the basis of the
theoretical work of Williams et al.,11 we designed and synthesized
two asymmetric compounds, 1 and 2 (Figure 1), having a central
phenyl ring and two alkyl chains in the para-positions as insulating
barriers. Whereas derivative 1 has an n-propyl chain on each side,
2 contains an n-propyl chain and n-pentyl chain at the two para
positions of the phenyl ring, thus being also asymmetricin terms of
tunneling barrier length. Instead of two identical thiol-based anchor
groups,11 the present derivatives contain a dithiocarbamate (NCS2)
and an acetyl protected thiol group (SCOCH3), respectively. The
difference in length of the two tunneling barriers separating the central
phenyl ring from the electrodes and the combination of different
anchor groups connecting the molecule with the leads has shown to
dramatically increase rectification.24 The dithiocarbamate group is a
strong chelating agent,25 and reacts faster with the Au surface than the
acetyl protected thiol.26,27 As a result, an oriented monolayer was
obtained. Subsequent removal of the acetyl protective groups allowed
the reaction of Au nanoparticles from solution with the thiol surface
for top electrode formation in a mild deposition step. With this
procedure, a simple way of realizing an asymmetric metalꢀ
moleculeꢀmetal junction without using different thiol protective
groups is disclosed. The whole process, that is, monolayer formation,
deprotection, and nanoparticle growth, was monitored by X-ray
photoelectron spectroscopy (XPS), polarization modulation infrared
reflectionꢀabsorption spectroscopy (PM-IRRAS), scanning tunnel-
ing microscopy (STM), and contact angle goniometry. The electrical
properties of asymmetric molecular junctions were determined using
a polymeric interlayer between a liquid top electrode and the oriented
self-assembled monolayer.
products 9 and 10 were then subjected to nucleophilic substitu-
tion by reacting with potassium thioacetate31 in the presence of
sodium iodide as the halide-exchange agent. A slow addition of
potassium thioacetate was necessary to prevent β-elimination of HCl
as a competing reaction. The acetyl-protected thiol compounds 11
and 12 were treated with carbon disulfide and sodium methoxide
resulting in the corresponding dithiocarbamate salts 1 and 2.
The synthesis of compound 3 is depicted in Scheme 2. Tri-
fluoroacetyl was selected to protect the amine group of N-methyl
benzylamine because it can be removed more easily than the
acetyl group. Thus, N-methyl benzylamine was first acetylated
with trifluoroacetic anhydride.32 Through a stepwise Friedelꢀ
Crafts acylation, reduction of the aryl ketone, and deprotection of
the trifluoroacetyl group, the amine 16 was obtained. In a sub-
sequent reaction, compound 16 was treated with carbon disulfide
and sodium methoxide, yielding the dithiocarbamate salt 3. It is
worth noting that the reduction of compound 14 followed by the
deprotection33 of 15 was successful, whereas the reverse proce-
dure was not. This might be due to deactivation of the reducing
agent by the amine. Further details concerning the syntheses and
characterization of 1ꢀ3 can be found in the Supporting
Information.
Junction Fabrication. A sequence of steps is required to
realize asymmetric molecular junctions within a “bottom-up”
assembly strategy, and in each of these steps, different analysis
techniques provide the necessary information about surface
chemistry and layer structure. Photoelectron spectroscopy, infra-
red spectroscopy, scanning probe microscopy, and contact angle
goniometry were employed, yielding information on monolayer
and electrode structure, binding chemistry at the interface, and
molecular orientation.
The junction fabrication can be divided into three main pro-
cessing steps: (a) oriented monolayer formation in a self-assembly
process, (b) deprotection of the exposed thiol end groups, and (c)
gold nanoparticle (NP) deposition for electrode formation. The
scheme in Figure 2 shows these steps with the asymmetric rectifier 2
as an example.
Monolayer Formation. The adsorption of dithiocarbamates
to Au was first studied using a simple compound (3, see Figure 1)
whose structure is closely related to 1 and 2, but lacking the
terminal thiol group. It further has a methyl substituent rather
than an ethyl substituent on the dithiocarbamate anchor group.
Compound 3 was selected as a reference because the interaction
of the molecule with the substrate is reduced to the dithio-
carbamate anchor group only, excluding any additional influ-
ence/interaction arising from the thioacetate moiety during self-
assembly.
’ RESULTS AND DISCUSSION
Synthesis and Characterization. The syntheses of com-
pounds 1 and 2 are depicted in Scheme 1. 3-Phenyl-1-propylamine
5922
dx.doi.org/10.1021/ja110244j |J. Am. Chem. Soc. 2011, 133, 5921–5930

Journal of the American Chemical Society
ARTICLE
Scheme 1. Synthesis of Compounds 1 and 2a
a Conditions: (a) AcOH, heating; (b) AlCl3, CS2, r.t.; (c) Et3SiH, TFA, r.t.; (d) BH3 S(CH3)2, THF, reflux; (e) KSCOCH3, NaI, DMF, r.t.; (f) NaOMe,
3
CS2, CHCl3, r.t.
Scheme 2. Synthesis of Compound 3a
a Conditions: (a) diethyl ether, r.t.; (b) AlCl3, CS2, reflux; (c) Et3SiH, TFA, r.t.; (d) K2CO3, CH3OH/H2O, r.t.; (e) NaOMe, CS2, CHCl3, r.t.
Figure 2. Schematic illustration of the fabrication steps for the formation of asymmetric metalꢀmonolayerꢀmetal junctions using compound 2:
(a) self-assembly on Au; (b) deprotection; (c) deposition of Au nanoparticles by incubation.
The adsorption characteristics of 3 to the Au substrate are
accessible through X-ray photoelectron spectroscopy in the S
2p core level region (Figure 3), which shows a unique signal
(S 2p3/2/S 2p1/2 peak doublet) attributed to chemisorbed
sulfur (161.94 eV)34 at binding energies consistent with repor-
ted values for dialkyl dithiocarbamates on Au18 and slightly lower
5923
dx.doi.org/10.1021/ja110244j |J. Am. Chem. Soc. 2011, 133, 5921–5930

Journal of the American Chemical Society
ARTICLE
Figure 3. XP spectrum of monolayer 3 on Au(111), showing a S 2p3/2 core-level peak at 161.94 eV. A single component (S 2p3/2/S 2p1/2 doublet)
appeared, representative of the bidentate bond of the dithiocarbamate moiety to Au. Single fit curves to the data are drawn in red; the sum of the fit curves
and the linear background are drawn in black. The scheme on the left shows the molecular arrangement of 3 on Au.
Table 1. Structural Properties of Monolayers 1ꢀ3a
elemental ratio
AFb
O/N
C/N
S(bound S)/S(unbound S)
S/Aud
MA (Å2)e
c
monolayer composition
dS(DTC)-S(thiol) (Å)
1
2
3
protected
C17H24NOS3
C15H22NS3
C19H28NOS3
C17H26NS3
C15H22NS2
13.44
13.36
15.79
15.96
-
0.73
0.73
0.69
0.68
-
1.94
0.59
1.64
0.74
0.68
19.4
15.7
19.1
18.4
17.8
1.74 (2.38)
1.61 (2.21)
1.24 (1.80)
1.57 (2.31)
-
0.032
0.033
0.031
0.033
0.031
30.4
29.5
31.4
29.5
31.4
deprotected
protected
deprotected
a The molecular length dS(DTC)-S(thiol) was determined from DFT gas-phase calculations. The attenuation factor AF was derived from the molecular
length and the tilt angle of the molecules relative to the surface (28.6°), and was used for calculation of the corrected S(bound S)/S(unbound S) ratios
(corrected values in parentheses). Elemental ratios from XPS peak areas yielded the chemical composition for each self-assembled monolayer. The
molecular area (MA) was obtained from the S/Au ratios by referencing to hexagonally close packed dodecanethiol monolayers. XP sensitivity factors
were gained from the Kratos database and calibrated using known reference samples (SFS 2p = 0.668, SFN 1s = 0.477, SFO 1s = 0.780, SFC 1s = 0.330,
SFAu 4f = 6.250). b Attenuation factor AF = exp(ꢀx/λ), using an electron attenuation length (EAL) of λ = 3.7 nm, as computed with the program “NIST
standard reference database 82”. The monolayer thickness (x) was obtained from x = cos RdS(DTC)-S(thiol), where dS(DTC)-S(thiol) is the molecular length
and R is the molecular tilt angle from the surface normal (R = 28.6°, see DFT calculations in the Supporting Information). c S(bound S)/S(unbound S) is the
ratio of the two S 2p peak areas with the S2p3/2 components at ∼162 eV and ∼164 eV, respectively. Values in the parentheses were corrected by the
attenuation factor (AF). d Ratio of the peak areas of bound S 2p (S) and of Au 4f (Au) divided by 2 to account for the presence of the two sulfur atoms in
the dithiocarbamate group. e MA obtained from S/Au peak area ratios by reference to densely packed alkanethiol monolayers (S/Au ∼ 0.045, MA =
21.6 Å2/molecule39).
(by ∼0.2 eV) than those found for thiolates on Au (162.1 eV for
octanethiol monolayers). Dithiocarbamate monolayers of this
kind have been studied previously,18,19 and it is established that
they bind to Au with both sulfur atoms in a bidentate configura-
tion, providing outstanding contacts, both electrically and me-
chanically.22 The molecular surface density of monolayer 3
(estimated via S 2p/Au 4f peak area ratios) is ∼0.7 times the
density of a close packed dodecanethiol monolayer (Table 1),
which can be rationalized in terms of a higher bulkiness of the
dithiocarbamate anchor group. The chemical shift and molecular
density were consistent with previous results obtained from
[1,10;40,100]terphenyl-400-yl-methane-dithiocarbamate monolayers
on Au.22 Notably, the purity of the adsorbed monolayers is
evident from C 1s/N 1s peak area ratios, which are very close to
stoichiometric values for 3 as well as for 1 and 2 (see Table 1).
In case of rectifiers 1 and 2, the situation is more complex, as
there is a competition in the adsorption of the thiolate and di-
thiocarbamate moiety to the metal surface. To obtain a direc-
tional assembly of rectifiers 1 and 2 on Au, we took advantage of
the fact that the rate constant for the molecular adsorption of
acetyl-protected thiols on Au is 2 orders of magnitude lower than
that for thiols on Au,27,35 whereas the adsorption rate constant
for dithiocarbamates is quite close to that of thiols (∼4 times
lower compared to thiols).26 These rate constants determine the
orientational order of the resulting monolayers, as they make
sure that rectifiers 1 and 2 attach to Au with the dithiocarbamate
anchor group first. A stability test of the thioacetate group in basic
dithiocarbamate solutions showed that the thioacetate moiety
was stable during the assembly time; however, this stability
decreases with the water content in solution (Supporting In-
formation).
Upon self-assembly of monolayers of 1 and 2, XPS allowed
to distinguish between different sulfur species contributing to
the S 2p spectrum. Furthermore, the oxygen O 1s signal
indicated the presence of an acetyl protective group on the
thiol moiety. In Figure 4, both the S 2p and the O 1s signals are
shown for monolayers of 2 before and after deprotection with
ammonium hydroxide, respectively. The monolayers feature
two emission peaks at a BE of ∼162 eV and ∼164 eV
(Figure 4a and Table 1 in Supporting Information). The peak
at ∼162 eV is well-known to arise from thiolate34 or
dithiocarbamate18 chemically bound to Au, whereas the peak
at ∼164 eV is attributed either to unbound thiols34,36 or to
acetyl protected thiols.37
5924
dx.doi.org/10.1021/ja110244j |J. Am. Chem. Soc. 2011, 133, 5921–5930

Journal of the American Chemical Society
ARTICLE
Figure 4. XP spectra of monolayer 2 on Au(111) acquired before deprotection (top) and after deprotection (bottom) in the S 2p (a) and the O 1s (b)
spectral region. As shown in (a), the S 2p3/2 components at 161.95 and 164.04 eV were not affected by the treatment, whereas in (b), a clear reduction in
the O 1s peak intensity at 532.6 eV was observed. Fit curves to the data are drawn in red and blue, corresponding to the two sulfur species shown in the
left scheme. The sum of the fit curves and the linear background are drawn in black. The scheme on the left side shows the monolayer structure before
and after deprotection.
To clarify the structure of monolayers 1 and 2, a quantitative
evaluation of XP signals was done, resulting in the figures listed in
Table 1. From the S 2p/Au 4f peak area ratios, where only the
bound sulfur component was taken into account, an average
molecular density of ∼0.73 times that of a dense dodecanethiol
monolayer was obtained, which is very close to the density found
for monolayer 3. From DFT calculations utilizing experimental
surface densities, an upright orientation of molecule 2 with an
average tilt angle of ∼28.6° toward the surface normal was
derived (see Supporting Information for computational details).
On the basis of the two XP peaks S 2p(162 eV) and S 2p(164
eV), we obtained a ratio of their areas of A(162 eV)/A(164 eV) =
1.8ꢀ2.3. Here a correction factor that accounts for the photo-
electron attenuation was applied (see Table 1 and Supporting
Information). The result is in reasonable agreement with the ex-
pected relative amount of bound and unbound sulfur (A(bound)/
A(unbound) = 2), when assuming an adsorption model as
presented in Figure 2. According to this model, both sulfur
atoms of the dithiocarbamate moiety attach to the substrate,
whereas the protected thiol functionality is exposed to the
surface. Unfortunately, when determining the ratio of the two
areas, the instrumental error due to limited XPS resolution and
partial overlap of the two components is ∼10ꢀ15%; thus, we
cannot exclude that a certain amount of rectifiers might attach to
the surface through the thiolate or even bind in a “bridge”
configuration. In any case, by exploiting the different adsorption
rates of thioacetates and dithiocarbamates, a shorter assembly
time (30ꢀ60 min) might help to reduce thiolate adsorption and
thus provide a maximum yield of dithiocarbamate-linked
molecules on Au.
In contact angle measurements with deionized water, mono-
layer 3 showed values (∼90°) that were fairly close to those from
alkanethiols and dialkyldithiocarbamates (103ꢀ114°),20 whereas
monolayers 1 and 2 (∼68° and ∼74°, respectively) exhibited a
higher polarity, characteristic of thiol-terminated surfaces (Sup-
porting Information, Table 1).
From the above data, we conclude that the oriented assembly
of target compounds 1 and 2, as expected based on the slow rate
of adsorption of thioacetates to Au, is the main feature of the
studied monolayers. The orientational order is a prerequisite for
the formation of rectifier-type junctions, in analogy to previous
studies where the oriented growth could be achieved using
monoacetylated dithiols.38
Deacetylation of the Thioacetate. The acetyl protection
group was removed from the monolayer surface by treating the
monolayer with ammonium hydroxide. Thiol and thioacetate
exhibit the same chemical shift for the S 2p signal; thus, the dea-
cetylation can only be monitored via changes in the O 1s signal.
XP data showed that most of the oxygen was removed upon the
treatment (Figure 4b), that is, its amount was decreased by about
1 oxygen atom/molecule (Table 1). Figure 4 shows that the
treatment did not affect the monolayer structure as far as the
bonding environment of the sulfur is concerned. Specifically, the
BE of bound and unbound sulfur species, their relative peak areas
and the molecular surface density (29ꢀ31 Å2/molecule, see
Table 1) were unchanged.
Further evidence for the deacetylation of the terminal thioe-
ster comes from PM-IRRAS, a technique which is highly specific
to the detection of chemical bonds, as it relies on the analysis
of vibrational modes that are excited by infrared radiation.
The changes in the intensities of the asymmetric and symmetric
in-plane CꢀH stretching modes of the methyl group, located at
2965 cmꢀ1 (υa(CH3)) and 2879 cmꢀ1 (υs(CH3)),40,41 were
used to monitor the presence of the methyl group in the acetyl
moiety (Figure 5). Compounds 1 and 2 include two methyl
groups, one of which is coordinated to the amine and the other
one is part of the thioester functionality. In addition, there are
Further supporting this scenario, the two protected mono-
layers show a consistent oxygen peak (O 1s at ∼532.6 eV),
indicative of the acetyl protective groups. The signal is strongly
reduced upon deprotection. Note that a slight oxygen back-
ground (expected stoichiometry for 1 and 2: O/N = 1), observed
for all samples (including monolayer 3), presumably indicates
the presence of water adsorbed to the surface.
5925
dx.doi.org/10.1021/ja110244j |J. Am. Chem. Soc. 2011, 133, 5921–5930

Journal of the American Chemical Society
ARTICLE
7 and 9 methylene groups for derivatives 1 and 2, respectively,
which appear in the asymmetric and symmetric CꢀH stretching
modes at 2932 cmꢀ1 (υa(CH2)) and 2854 cmꢀ1 (υs(CH2)).40
The position of the methylene peaks is consistent with mono-
layers existing in the liquid phase rather than in the crystalline
phase.41 This result was expected considering the short alkyl
chains in molecules 1 and 2. As shown in Figure 5, a strong
decrease in the methyl υa(CH3) signal intensity was observed for
monolayer 2 upon treatment with ammonium hydroxide, thus,
demonstrating the removal of the acetyl group from the thiol
surface. The low intensities of the υs(CH3) mode and of the
thioacetyl CdO stretching vibration at 1680 cmꢀ1 (not shown)
can be explained by a parallel orientation of the acetyl SꢀCꢀC
plane relative to the surface plane, because in this case both the
υs(CdO) and υs(CH3) transition dipoles are perpendicular to the
surface normal (and thus not excited by the electrical field vector).42
Au Nanoparticle Deposition. The metallization of the thiol-
terminated surface was accomplished by incubation of the
monolayer in a nanoparticle solution. This process was se-
lected in view of the high sensitivity of thin self-assembled
monolayers, which exclude any conventional thermal deposi-
tion of metals on them. During immersion of the monolayers
in the NP solution, an exchange of the dodecylamine ligands
linked to the NPs with the thiolate end groups at the mono-
layer surface occurred. Thus, a layer of NPs, chemically linked
to the underlying monolayer, was formed. As dithiocarbamates
1 and 2 are completely insoluble in toluene, which is used as a
solvent during nanoparticle incubation, a ligand exchange, that
is, the desorption of dithiocarbamates from the surface with
subsequent re-adsorption on the Au NP, can be excluded.
Figure 6a shows an STM image of the template-stripped gold
(TSG) surface modified with monolayer 1, and Figures 6b and
7 show the same surface after deprotection and incubation
with the Au NPs. Whereas in Figure 6a the characteristic
features of a TSG surface are seen (despite the low roughness
of Δhrms ∼0.3 nm, the grains of the polycrystalline Au surface
appeared), in Figures 6b and 7 a close packed layer of Au
clusters and NPs covers the surface. Here, we found peak-to-
valley STM height differences of 6ꢀ20 nm, corresponding to
the diameters of the Au clusters or of their agglomerates. The
Au NPs were firmly attached to the surface because exten-
sive rinsing of the sample in different solvents did not change
the NP coverage. Control samples with monolayer 3, lacking
the thiol end group, were subjected to the same treatment, but
did not show any difference in surface topography upon NP
incubation (Figure 6c,d).
Figure 5. PM-IRRAS spectra of representative monolayer 2 in the
2700ꢀ3100 cmꢀ1 region before and after removal of the acetyl protective
group. The arrows indicate the position of the asymmetric and symmetric
stretching modes of the methyl group at 2965 and 2879 cmꢀ1, respectively.
Figure 6. STM images of monolayer 1 on TSG before (a) and after (b) deprotection and NP incubation. (a) The rms roughness of the surface was
∼0.3 nm, but it showed grains (100ꢀ200 nm in size) from the polycrystalline TSG surface. (b) After deprotection and NP incubation, the surface was
densely covered with large gold clusters and NPs. Monolayer 3 on TSG before (c) and after (d) ammonium hydroxide treatment and NP incubation,
demonstrating that the NPs did not bind to the surface. All images were acquired at a bias voltage of 400 mV and a tunneling current of 2 pA. The height
scale was 3 nm for images a, c, and d and 20 nm for image b.
5926
dx.doi.org/10.1021/ja110244j |J. Am. Chem. Soc. 2011, 133, 5921–5930

Journal of the American Chemical Society
ARTICLE
The efficiency of the metallization step, that is, the full che-
misorption of the Au NPs to the thiol-terminated surface, was
verified by XPS and showed the disappearance of one of the two
thiol components in the S 2p spectral region (Figure 8).
Although the total S 2p signal intensity was strongly attenuated
by the presence of the Au NPs on top of the monolayer, a
prevalent suppression of the S 2p component related to unbound
sulfur at ∼164 eV could be observed (see Figure 4 for comparison).
Thus, most of the thiols at the surface are converted to thiolates
upon reaction with the Au NPs in solution, and a thin metal layer is
formed at the monolayer surface. The whole process delivers a
metalꢀmoleculeꢀmetal junction that is intrinsically asymmetric
due to the choice of the two different anchor groups employed for
molecular self-assembly.
Electrical Characterization. An alternative approach for con-
tacting monolayers is the deposition of a highly conducting poly-
meric interlayer between the monolayer of interest and, for
example, a thermally evaporated top electrode. This technique
enabled the measurement of large area molecular junctions with
diameters up to 100 μm.13 We used monolayers as described in
the section “Deacetylation of the Thioacetate” and deposited a
poly(3,4-ethylene-dioxythiophene):poly(4-styrenesulphonic-acid)
(PEDOT:PSS) film on these surfaces for further electrical
characterization with a hanging Hg-drop apparatus (see Experi-
mental Section). Because of the polar character of the water-
soluble PEDOT:PSS suspension, these films are highly adhesive
to the thiol end groups of the monolayers. PEDOT:PSS was spin
coated onto monolayers of 1 and 2, forming polymeric films with
a homogeneous thickness of about 100 nm. The currentꢀvoltage
characteristics were acquired in the bias range from ꢀ1 to 1 V,
and measurements comparing curves from 1 and 2 with data
from alkanedithiol monolayers having different chain length
(1,8-octyldithiol (ODT) and 1,12-dodecanedithiol (DDT))
allowed us to determine the conductance per unit area of mono-
layers of 1 and 2. The curves were very reproducible on different
samples and positions. The average current densities in Figure 9
Figure 7. STM zoom image of monolayer 1 on TSG after deprotec-
tion and NP incubation. The image was acquired at a bias voltage
of 400 mV and a tunneling current of 2 pA. The height scale
was 15 nm.
Figure 8. XP S 2p spectrum of monolayer 1 after deprotection and NP
incubation. The unbound thiol component at ∼164 eV almost
vanished, showing that most of the terminal thiols in the monolayer
reacted with the Au NPs (∼162 eV). Fit curves to the data are drawn in
red and blue, while the sum of the fit curves and the linear background
are in black.
Figure 9. Average current density versus bias voltage for monolayer junctions of 1 and 2 having the structure Au//Monolayer/PEDOT:PSS/Hg. The
curves are compared with data from ODT and DDT in analog devices. At a bias of 1 V, the current densities for DDT:2:1:ODT follow the ratio 1:3:7:13.
The error bars indicate the standard deviation for the measured curves. The scans are acquired from 0 to 1 V, down to ꢀ1 V and back to 0 V. Both scan
directions are shown in the graph. The individual scans in a linear representation are shown in the Supporting Information (Figures S4 and S5).
5927
dx.doi.org/10.1021/ja110244j |J. Am. Chem. Soc. 2011, 133, 5921–5930

Journal of the American Chemical Society
ARTICLE
show that at a bias of 1 V the conductance of a monolayer of 2 is
3.1 times lower than that of a monolayer of 1 and 2.2 times higher
than that of a DDT monolayer. The ratio in the conductance of
ODT and DDT monolayers was 13. These proportions in the
conductance of the layers are in good agreement with values
obtained if a simple tunnelling barrier model5,43 with an attenua-
tion factor of β ∼ 0.5/Å is applied. Referred to the molecular
structure (see Figure 1), this means that the current density
scales exponentially with the length of the barrier x,
asymmetric electrodeꢀmonolayerꢀelectrode structure, an issue
of high relevance to obtain diodes in molecular electronics.
’ EXPERIMENTAL SECTION
Monolayer Preparation. For XPS, FT-IR, STM, and contact
angle characterization, template-stripped gold (TSG) surfaces were
prepared using known procedures.44 In short, Au was deposited by
thermal evaporation (100 nm) onto a polished silicon wafer (Si-Mat
GmbH). A glass slide covered with epoxy glue (EPO-TEC 353, Polytec)
was then deposited on the Au surface. Subsequently, the glue was cross-
linked at 150 °C (for 1 h). The TSG surface was formed by stripping the
Au on the glass from the silicon wafer with the aid of a razor, yielding
substrates with a rms roughness of 0.3 nm, as observed by atomic force
microscopy. The self-assembled monolayers were prepared in an argon
environment (glovebox) by immersion of the freshly cleaved TSG
samples into 0.2 mM solutions of dithiocarbamate salts 1, 2, or 3 in
absolute ethanol for 18ꢀ24 h. Such a long assembly time was chosen
because a long exposure of Au surfaces to thiols or dithiocarbamates is
known to provide denser, higher quality monolayers. However, we also
tested the assembly of dithiocarbamates for times as short as 1 h, without
significant loss in monolayer density as far as XPS spectra are concerned.
A shorter soaking time probably slightly reduces the fraction of
molecules linked to Au with the thiol moiety. After self-assembly, all
samples were thoroughly rinsed with absolute ethanol, blown dry with
argon, and immediately used for characterization. The samples were
then transferred for XPS analysis in argon filled containers. To avoid
oxidation, all solvents used for monolayer preparation and processing
(ethanol and methanol) were saturated with argon before use.
J µ expðꢀβxÞ
that is, with the number of aliphatic bonds within the molecular
junction. The current densities for ODT and DDT junctions are
in excellent agreement with values reported by de Boer et al.,13
even though in our case a Hg drop electrode rather than a
thermally evaporated Au contact was employed. Also the mea-
sured tunnelling decay constant, even though lower than re-
ported with other experimental test beds, is very close to previous
results gained with monolayer/PEDOT:PSS junctions.13
Interestingly, above a bias of ∼0.4 V, the rise in current density
for 1 and 2 is significantly more pronounced than for alkane-
dithiols. This is probably caused by the frontier orbitals of the
phenyl ring in molecular rectifiers 1 and 2, which above a certain
bias voltage are aligned with the Fermi level of one of the
electrodes and thus become accessible for resonant tunnelling.
In addition, at a bias of at 1 V, for both rectifiers a moderate
rectification (rectification ratio = 1.32) is found. This effect,
observed for the first time in monolayer/polymer type junctions,
does not show up in alkanedithiol junctions that are intrinsically
symmetric. Hence, the observed rectification can be attributed to
the asymmetry of the monolayer junctions consisting of the or-
iented self-assembled rectifiers 1 and 2.
In summary, the presented electrical data show a clear signa-
ture from the molecular layers and confirm that the current
density reflects the nature of the molecules in the respective
junction, both concerning molecular length and structure. It also
shows that, as for ODT and DDT monolayers, the packing
density and orientation of rectifiers 1 and 2 is sufficiently well-
defined to avoid electrical shorts in the molecular junctions.
However, it should be mentioned that polymeric interlayers do
not allow a systematic study of molecular rectification since the
electrical contact with the terminal thiol groups is not well
understood, as it is for thiolateꢀAu interfaces. For a quantitative
investigation, identical metal electrodes on both sides of the
junction are required, and single-molecule measurements are
currently ongoing to correlate theoretical predictions with ex-
perimental rectification ratios.
Deacetylation and Au Nanoparticle Incubation. The acetyl
protective groups were removed by exposing the monolayers of 1 and 2
to an ethanolic solution of ammonium hydroxide (0.52 M) for 12 h. The
resulting samples were thoroughly rinsed with ethanol and immersed
into a freshly prepared solution of dodecylamine-stabilized Au nano-
particles45 (average particle diameter: 4 nm) in toluene for 18 h. The
concentration of this nanoparticle solution was adjusted to an absor-
bance of 0.4 at the maximum of the plasmon absorption band (λmax
=
514 nm, 2 mm path length). After incubation, excess nanoparticles were
removed by rinsing the samples with toluene and ethanol and drying
them in a stream of N2.
UVꢀVis and Infrared Spectroscopy. UV/vis absorption spec-
tra were recorded using a Varian Cary 50 scan spectrophotometer
using a cell path length of 1 cm. Polarization modulation infrared
reflectionꢀabsorption spectroscopy (PM-IRRAS) data were acquired
with a Bruker Vertex 70 spectrometer equipped with a PMA50 module
including a nitrogen cooled mercuryꢀcadmiumꢀtelluride (MCT)
detector and a ZnSe photoelastic modulator. The IR beam was set at
80° incident to the TSG surface, and the spectra were collected over 15
min at a resolution of 4 cmꢀ1
.
Contact Angle Goniometry. To determine the surface energy of
the monolayer, contact angle measurements were done with a KSV
CAM 100 at room temperature. A drop of deionized water from a
’ CONCLUSIONS
The fabrication of metalꢀmoleculeꢀmetal junctions consist-
ing of monolayers of aligned, self-assembled rectifier candidates
has been reported in this study. The oriented growth of these
rectifiers has been achieved by making use of the different
adsorption rate constants of dithiocarbamates and acetyl pro-
tected thiols on Au surfaces. The critical top electrode deposition
was demonstrated both by nanoparticle incubation and by for-
mation of conducting polymeric films. Electrical measurements
with large area polymeric junctions reflect the conductance of the
molecular layers, demonstrating the feasibility of monolayer-
based junctions. The presented method is a simple and repro-
ducible procedure for the realization of devices with an
Millipore system (18.2 MΩ cm) was dispensed on the surface after
3
rinsing it with ethanol and drying it in nitrogen.
X-ray Photoelectron Spectroscopy (XPS). High-resolution
XPS spectra were recorded with a Kratos Axis Ultra instrument using
an Al KR (1486.6 eV) source operated at 15 kV and 180 W. With an
X-ray monochromator and a pass energy of 40 eV for the analyzer, an
instrumental energy resolution of 0.5 eV was achieved. A survey
spectrum and high resolution spectra of the S 2p, C 1s, N 1s, O 1s
and Au 4f regions were acquired. The binding energies were calibrated
against the Au 4f7/2 core-level peak at 84 eV. The spectra were fitted
using Voigt functions with a 50/50 Lorentz/Gaussian ratio and a
linear background. The line shape parameters were determined by
5928
dx.doi.org/10.1021/ja110244j |J. Am. Chem. Soc. 2011, 133, 5921–5930

Journal of the American Chemical Society
ARTICLE
least-square fitting to carbon and sulfur core-level signals of known
alkanethiol reference samples. The S 2p spectrum, consisting of the
two components S 2p3/2 and S 2p1/2 with a fixed separation of 1.18 eV,
was fitted with a relative ratio of 2/1 for the S 2p3/2/S 2p1/2 areas. The
molecular surface density was obtained from the relative S/Au signal
intensity by consideration of the bound S 2p component only
(∼162 eV), which allowed a direct comparison with the surface density
of other dithiocarbamate or thiol derivatives. The S/Au intensity ratio is
independent on the specific molecular structure of various organosulfur
compounds that are chemically linked to Au, thus a reliable indicator for
molecular densities.
Scanning Tunneling Microscopy (STM). The structural prop-
erties of the modified TSG surfaces were characterized using a
scanning tunneling microscope (Multimode Nanoscope IIIa, Digital
Instruments) equipped with a low current amplifier. Self-cut Pt/Ir (80/
20) tips were used as a probe. The STM scans were recorded
in constant-current mode under ambient conditions at room
temperature.
Electrical Measurements. For sample preparation, the self-as-
sembly of rectifier molecules and subsequent deacetylation were
carried out following the same procedure as described above. The
monolayers of 1,8-octyldithiol and 1,12-dodecanedithiol were pre-
pared by immersion of the TSG substrates into 3 mM ethanolic
solutions of the dithiols for 36 h.46 After monolayer preparation, the
substrates were rinsed intensely with ethanol. A poly(3,4-ethylene-
dioxythiophene):poly(4-styrenesulphonic-acid) (PEDOT:PSS) solu-
tion was prepared by diluting 0.25 mL of the water-based source
suspension (Clevios FE, H.C. Starck GmbH & Co.) with 0.35 mL of
methanol and by subsequent filtration through a nylon filter having a
pore size of 0.1 μm. The PEDOT:PSS solution was then spin coated
onto the TSG substrates at 1500 rpm for 50 s and dried in a vacuum
chamber, resulting in a polymer film with a thickness of ∼100 nm. For
electrical characterization, a home-built Hg-drop setup with a Hamil-
ton syringe as Hg-dispenser was employed. Technical details about the
experimental setup are reported elsewhere.22 The measurements were
carried out under ambient conditions, that is, the Hg droplet was
extruded and suspended in air above the sample. Subsequently, the
surface was approached using a piezo-table until a large area contact
(0.39 ( 0.03 μm in diameter) between sample and Hg droplet was
established. The contact of the Hg droplet with the sample was
monitored from side-view using a CCD-camera, and based on the
junction diameter, the contact area was determined. Currentꢀvoltage
curves were acquired on 2 different samples and on 6 positions in total.
For each position, a series of 5 curves was recorded with a maximum
bias value at 0.1, 0.2, 0.5, 0.6, and 1 V. The data was acquired using a
voltage ramp with a bin size of 10 mV and an elapsed time between
steps of 0.1 s. Averaged curves were created from 6 single scans on 2
different samples. The Hg electrode was on ground and the bias
applied to the bottom TSG electrode. Current densities were obtained
by normalizing the currents with the contact area of the Hg droplet on
the sample.
’ ACKNOWLEDGMENT
We thank Dr. Nadja Krasteva and Dr. Yvonne Joseph for
providing Au nanoparticles and Dr. Armin Gembus from Bruker
Optics for the acquisition of PM-IRRAS spectra.
’ REFERENCES
(1) 2009 Semiconductor Industry Association, International Tech-
nology Roadmap for Semiconductors 2009 Edition, Emerging Research
Devices; http://www.itrs.net/Links/2009ITRS/Home2009.htm. 2009.
(2) Carroll, R. L.; Gorman, C. B. Angew. Chem., Int.Ed. 2002,
41, 4379–4400.
(3) Joachim, C.; Gimzewski, J. K.; Aviram, A. Nature 2000,
408, 541–548.
(4) Tao, N. J. Nat. Nanotechnol. 2006, 1, 173–181.
(5) Akkerman, H. B.; de Boer, B. J. Phys.: Condens. Matter 2008,
20, 013001.
(6) Campbell Scott, J.; Bozano, L. D. Adv. Mater. 2007, 19, 1452–1463.
(7) Green, J. E.; Choi, J. W.; Boukai, A.; Bunimovich, Y.; John-
ston-Halperin, E.; Delonno, E.; Luo, Y.; Sheriff, B. A.; Xu, K.; Shin,
Y. S.; Tseng, H.-R.; Stoddard, F.; Heath, J. R. Nature 2007, 445, 441–444.
(8) Snider, G. S.; Williams, R. S. Nanotechnology 2007, 18.
(9) Aviram, A.; Ratner, M. A. Chem. Phys. Lett. 1974, 29, 277–283.
(10) Metzger, R. M. Chem. Rev. 2003, 103, 3803–3834.
(11) Kornilovitch, P. E.; Bratkovsky, A. M.; Williams, R. S. Phys. Rev.
B 2002, 66, 165436ꢀ1–165436ꢀ11.
(12) Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides,
G. M. Chem. Rev. 2005, 105, 1103–1169.
(13) Akkerman, H. B.; Blom, P. W. M.; de Leeuw, D. M.; de Boer, B.
Nature 2006, 441, 69–72.
(14) Jiang, P.; Morales, G. M.; You, W.; Yu, L. P. Angew.Chem., Int.
Ed. 2004, 43, 4471–4475.
(15) Yu, C. J.; Chong, Y.; Kayyem, J. F.; Gozin, M. J. Org. Chem.
1999, 64, 2070–2079.
(16) Pollack, S. K.; Naciri, J.; Mastrangelo, J.; Patterson, C. H.;
Torres, J.; Moore, M.; Shashidhar, R.; Kushmerick, J. G. Langmuir 2004,
20, 1838–1842.
(17) Wessels, J. M.; Nothofer, H. G.; Ford, W. E.; von Wrochem, F.;
Scholz, F.; Vossmeyer, T.; Schroedter, A.; Weller, H.; Yasuda, A. J. Am.
Chem. Soc. 2004, 126, 3349–3356.
(18) Morf, P.; Raimondi, F.; Nothofer, H.-G.; Schnyder, B.; Yasuda,
A.; Wessels, J. M.; Jung, T. A. Langmuir 2006, 22, 658–663.
(19) Zhao, Y.; Pꢀerez-Segarra, W.; Shi, Q.; Wei, A. J. Am. Chem. Soc.
2005, 127, 7328–7329.
(20) Weinstein, R. D.; Richards, J.; Thai, S. D.; Omiatek, D. M.;
Bessel, C. A.; Faulkner, C. J.; Othman, S.; Jennings, G. K. Langmuir
2007, 23, 2887–2891.
(21) Cao, R.; Diaz, A.; Cao, R.; Otero, A.; Cea, R.; Rodriguez-Arguelles,
M. C.; Serra, C. J. Am. Chem. Soc. 2007, 129, 6927–6930.
(22) von Wrochem, F.; Gao, D.; Scholz, F.; Nothofer, H. G.; Nelles,
G.; Wessels, J. M. Nat. Nanotechnol. 2010, 5, 618–624.
(23) Li, Z.; Kosov, D. S. J. Phys. Chem. B 2006, 110, 9893–9898.
(24) Loertscher, E.; Gao, D.; Schwarz, F.; Gotsmann, B.; Ford, W.
E.; von Wrochem, F.; Riel, H. Unpublished results 2011.
(25) McCubbin, Q. J.; Stoddart, F. J.; Welton, T.; White, A. J. P.;
Williams, D. J. Inorg. Chem. 1998, 37, 3753–3758.
’ ASSOCIATED CONTENT
Supporting Information. Synthesis procedures and 1H
(26) Zhu, H.; Coleman, D. M.; Dehen, C. J.; Geisler, I. M.;
Zemlyanov, D.; Chmielewski, J.; Simpson, G. J.; Wei, A. Langmuir
2008, 24, 8660–8666.
(27) Schwartz, D. K. Annu. Rev. Phys. Chem. 2001, 52, 107–137.
(28) Lane, C. F. Chem. Rev. 1976, 76, 773–799.
(29) Kiuchi, M.; Adachi, K.; Kohara, T.; Minoguchi, M.; Hanano, T.;
Aoki, Y.; Mishina, T.; Arita, M.; Nakao, N.; Ohtsuki, M.; Hoshino, Y.;
Teshima, K.; Chiba, K.; Sasaki, S.; Fujita, T. J. Med. Chem. 2000,
43, 2946–2961.
(30) Brown, H. C.; Choi, Y. M.; Narasimhan, S. J. Org. Chem. 1982,
47, 3153–3163.
S
b
NMR spectra of compounds 1 and 3 and the intermediates,
molecular modeling results, XPS and electrical data. This
material is available free of charge via the Internet at http://
pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
Florian.vonWrochem@eu.sony.com
5929
dx.doi.org/10.1021/ja110244j |J. Am. Chem. Soc. 2011, 133, 5921–5930

Journal of the American Chemical Society
ARTICLE
(31) Zheng, T.-C.; Burkart, M.; Richardson, D. E. Tetrahedron Lett.
1999, 40, 603–606.
(32) Plenio, H.; Hermann, J.; Diodone, R. Inorg. Chem. 1997,
36, 5722–5729.
(33) Bergeron, R. J.; McManis, J. S. J. Org. Chem. 1988,
53, 3108–3111.
(34) Laibinis, P. E.; Whitesides, G. M.; Allara, D. L.; Tao, Y.-T.;
Parikh, A. N.; Nuzzo, R. G. J. Am. Chem. Soc. 1991, 113, 7152–7167.
(35) Lau, K. H. A.; Huang, C.; Yakovlev, N.; Chen, Z. K.; O’Shea,
S. J. Langmuir 2006, 22, 2968–2971.
(36) Nuzzo, R. G.; Zegarski, R.; Dubois, L. H. J. Am. Chem. Soc. 1987,
109, 733–740.
(37) Tour, J. M.; Jones, L., II; Pearson, D. J.; Lamba, J. J. S.; Burgin,
T. P.; Whitesides, G. M.; Allara, D. L.; Parikh, A. N.; Atre, S. V. J. Am.
Chem. Soc. 1995, 117, 9529–9534.
(38) Niklewski, A.; Azzam, W.; Strunskus, T.; Fischer, R. A.; Woll, C.
Langmuir 2004, 20, 8620–8624.
(39) Schreiber, F. Prog. Surf. Sci. 2000, 65, 151–256.
(40) Allara, D. L.; Swalen, J. D. J. Phys. Chem. 1982, 86, 2700–2704.
(41) Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D.
J. Am. Chem. Soc. 1987, 109, 3559–3568.
(42) Ulman, A. An Introducton to Ultrathin Organic Films; Academic
Press: San Diego, CA, 1991.
(43) Wang, W.; Lee, T.; Reed, M. A. Phys. Rev. B 2003,
68, 035416-1–035416-7.
(44) Naumann, R.; Schiller, S. M.; Giess, F.; Grohe, B.; Hartman,
K. B.; Karcher, I.; Koper, I.; Lubben, J.; Vasilev, K.; Knoll, W. Langmuir
2003, 19, 5435–5443.
(45) Leff, D. V.; Brandt, L.; Heath, J. R. Langmuir 1996,
12, 4723–4730.
(46) Akkerman, H. B.; Kronemeijer, A. J.; van Hal, P. A.; de Leeuw,
D. M.; Blom, P. W. M.; de Boer, B. Small 2008, 4, 100–104.
5930
dx.doi.org/10.1021/ja110244j |J. Am. Chem. Soc. 2011, 133, 5921–5930