Disorder-order phase change of ω-(N-pyrrolyl)alkanethiol self-assembled monolayers on gold induced by STM scans and thermal activation

Article abstract of DOI:10.1039/b800239h

Molecular ordering of pyrrolyl-terminated alkanethiol self-assembled monolayers (PyC_nSH SAMs) on Au(111) substrates (n = 11 or 12) was investigated by scanning tunneling microscopy (STM) and various spectroscopic methods. The SAMs, which were in a disordered state when formed at room temperature, could be ordered either globally by thermal annealing at 70°C, or locally via stimulation with repetitive STM scans. The ordered phase was characterized by small domains of molecular rows formed along 〈112〉 directional set with an inter-row corrugation period close to 1.44 nm, in which defects were abundant. Based on the experimental results, the molecular arrangement in the ordered PyC_nSH SAM was proposed to be a (5×√3)rect structure with a molecular deficiency ≥10%. While mechanical interactions between molecules and scanning probe tips had been pointed out as the major cause of scan-induced phase transformations in other SAM systems, electronic or electrostatic factors were thought to affect considerably the scan-induced ordering process in this SAM system. From comparison of surface molecular coverage between disordered and thermally ordered SAMs of PyC₁₂SH, it was inferred that the disorder could be ascribed to both kinetic and thermodynamic factors. The kinetic barrier to the ordered phase was supposed to result from strong dipole-dipole interactions among the pyrrolyl endgroups. the Owner Societies.

Full text of DOI:10.1039/b800239h

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/pccp | Physical Chemistry Chemical Physics
Disorder–order phase change of x-(N-pyrrolyl)alkanethiol self-assembled
monolayers on gold induced by STM scans and thermal activation
Joon Sung Lee,ab Young Shik Chi,a Jinhee Kim,b Wan Soo Yunb and Insung S. Choi*a
Received 7th January 2008, Accepted 13th March 2008
First published as an Advance Article on the web 14th April 2008
DOI: 10.1039/b800239h
Molecular ordering of pyrrolyl-terminated alkanethiol self-assembled monolayers (PyCnSH
SAMs) on Au(111) substrates (n = 11 or 12) was investigated by scanning tunneling microscopy
(STM) and various spectroscopic methods. The SAMs, which were in a disordered state when
formed at room temperature, could be ordered either globally by thermal annealing at 70 1C, or
locally via stimulation with repetitive STM scans. The ordered phase was characterized by small
ꢀ
domains of molecular rows formed along h112i directional set with an inter-row corrugation
period close to 1.44 nm, in which defects were abundant. Based on the experimental results, the
molecular arrangement in the ordered PyCnSH SAM was proposed to be a (5ꢀO3)rect structure
with a molecular deficiency Z 10%. While mechanical interactions between molecules and
scanning probe tips had been pointed out as the major cause of scan-induced phase
transformations in other SAM systems, electronic or electrostatic factors were thought to affect
considerably the scan-induced ordering process in this SAM system. From comparison of surface
molecular coverage between disordered and thermally ordered SAMs of PyC12SH, it was inferred
that the disorder could be ascribed to both kinetic and thermodynamic factors. The kinetic
barrier to the ordered phase was supposed to result from strong dipole–dipole interactions among
the pyrrolyl endgroups.
stituted systems by aromatic moieties have been most widely
1. Introduction
studied, because these systems usually give a variety of well-
Self-assembled monolayers (SAMs) of organic thiols on metal
defined structural phases that sensitively depend on packing
surfaces have been a subject of wide and intensive studies, for
configurations of the aromatic moieties. For example, struc-
their potential applications in functional modifications of
tures of phenyl- or methylbiphenyl-terminated short alkane-
metal surfaces as well as in molecular electronics.1 Much work
thiol SAMs have been known to be dependent on the alkyl
linker length by STM studies.5–7 It has also become well
on various aspects of organic thiol SAMs has been performed
not only for practical applications, but also for tackling
known that various structural phases of SAMs may exist for
fundamental issues such as correlations between structures
a single molecular species, depending on the formation con-
ditions and post-treatments.4,8–10 These studies have provided
of SAMs and the corresponding molecular species. SAMs of
normal alkanethiols on Au(111) have been most intensely
a fundamental insight on the correlation between SAM struc-
studied as a model system of organic thiol SAMs, for their
tures and various factors that govern the intermolecular and
simplicity and ease of sample preparation. However, even for
molecule–substrate interactions.
this simple model system, there are several unresolved issues
McCarley and Willicut have made a major contribution to
studies on pyrrolyl-terminated alkanethiol (PyCnSH) SAMs
on Au(111) by reporting an electrochemical polymerization of
the pyrrolyl endgroups,11–13 which implied possible applica-
tion of this system to molecular electronics and chemical
surface engineering.14 Although they proposed formation of
polymerized patterns on those SAMs by scanning probe-based
methods to be viable, there has not yet been a report of
success. SAMs of PyCnSH with n r 10 have been shown by
the authors to have an almost fully disordered nature, with
vibration mode frequencies and intensity ratios from infrared
spectra of the methylene linkers in the SAMs similar to those
of liquid molecules, and with a much smaller surface density of
molecules (B50%) than that of close-packed alkanethiol
SAMs. However, there have been disparities in characteristics
of the SAM system such as surface molecular coverage and
such as identification of Au–S binding sites and possible Au
surface relaxation in many structural phases identified so far.2
Substitution of the methyl terminal group in alkanethiols by
other functional groups can alter the properties of their SAMs
dramatically. The substitution opens up many practical pos-
sibilities such as applications in molecular electronics and soft
lithography, as well as enhancement of tribological properties
and stability.3,4 Variation of structural phases dependent on
terminal groups and alkyl linker length in endgroup-substi-
tuted alkanethiol SAM systems has also been a subject of
many in-depth studies. In this research field, endgroup-sub-
a Department of Chemistry and School of Molecular Science (BK21),
KAIST, Daejeon 305-701, Korea. E-mail: ischoi@kaist.ac.kr; Fax:
+82 (0)42 869 2810
b Korea Research Institute of Standards and Science (KRISS),
Daejeon 305-600, Korea
ꢁc
This journal is the Owner Societies 2008
3138 | Phys. Chem. Chem. Phys., 2008, 10, 3138–3149

View Article Online
pyrrole ring orientation in later reports.15–17 Moreover, there
has not been an SPM-based study on the system with mole-
cular or lattice resolution.11
In this article, we present the first scanning probe micro-
scopy (SPM) study on the structures of PyCnSH SAMs. It is
shown that the SAMs (n = 11 or 12) formed at room
temperature have a partially disordered structure, while they
can transform into an ordered phase by STM scans or thermal
annealing. Structural phase transformations by SPM scans or
thermal activation have been reported in various molecular
monolayer systems.4,18–21 For example, there has been a report
that a metastable (2ꢀO3)rect phase of hexadecanethiol SAM
was turned into a more stable (O3ꢀO3)R301 phase by me-
chanical perturbation with an AFM tip.20 There also has been
a report of phase conversion at a low temperature of 35 K
induced by STM scans between low-coverage phases of phy-
sisorbed azobenzene molecular layer on Au(111).22 To our
knowledge, however, there has been no report of phase
transformation from disordered to ordered phases in any
molecular SAM system.
Fig. 1 An STM topograph of pristine PyC12SH SAM assembled at
room temperature. Image was taken with It = 30 pA, Vtip = 0.5 V.
The scanned area is 30 ꢀ 30 nm2.
In the scan-induced ordering of the SAMs, no odd–even
effect was observed and the ordered phases of the PyCnSH
SAMs with n = 11 or 12 appeared to be alike. The ordered
phase was composed of small domains of molecular rows
2.2 Scan-induced ordering of PyC12SH SAM
ꢀ
The initially disordered RT-SAM of PyC12SH tended to be
ordered with repetitive STM scans. This scan-induced order-
ing was more facile with more invasive scan conditions, such
as higher tunneling current setpoints or smaller tip biases.
Fig. 2 shows the ordering of PyC12SH SAM on a Au/mica
substrate by repetitive STM scans under invasive scan condi-
tions. As is shown in Fig. 2a, the SAM was initially in a
disordered phase characterized by smaller Au vacancies and
fuzzy STM topography. Faint signs of partial molecular
ordering could be observed already at this initial scan. After
40 min of consecutive scans, the scan-induced molecular
ordering became evident (Fig. 2b). The ordering was usually
accompanied by Au vacancy coarsening, or even by Au
sublayer erosion.23 The ordering efficiency—the time required
for ordering of a unit area—was largely erratic, but dependent
on scan conditions and tip properties to some extent.
formed along the h112i directional set with an inter-row
corrugation period of 1.47 ꢂ 0.07 nm. Since a molecularly
resolved STM image of the ordered phase could not be
obtained from the SAMs, we used other experimental methods
such as cyclic voltammetry (CV), X-ray photoelectron spectro-
scopy (XPS), ellipsometry, polarized infrared external reflec-
tance spectroscopy (PIERS), and contact angle goniometry to
investigate structural characteristics of the disordered and
ordered phases of the SAMs. Based on the experimental
results, the molecular arrangement in the SAM could be
deduced to be a (5ꢀO3)rect structure.
2. Results and discussion
2.1 The pristine SAM of PyC12SH
The ordered monolayer was composed of small domains of
molecular rows oriented by B1201 with respect to each other.
These molecular row directions were related to three equiva-
lent surface directions that came from the three-fold symmetry
of the unreconstructed Au(111) surface. Traces of roughly
triangular facets can be seen in Fig. 2a. It is well known that
The PyC12SH SAM formed at room temperature (RT) was in
a disordered phase, characterized by 0.24-nm-deep etch pits
typical of thiol SAMs on Au(111).5,23 While molecular resolu-
tion in the STM images was mostly not achievable in this
phase, some distinct features suggestive of partially ordered
molecules could be observed at very scarce instances. Six
protrusions in a row could be seen in the encircled area of
Fig. 1. The average distance between the protrusions was
measured to be B0.48 nm, which was quite close to the
next-nearest-neighbor (NNN) distance of Au atoms on un-
reconstructed (111) surface (0.4995 nm). This intermolecular
distance of B0.5 nm had already been shown to manifest itself
in the (O3ꢀO3)R301-based close-packed structures of alka-
nethiol SAMs,2,18 as well as in the ordered (O3ꢀ2O3)R301
structures of phenyl-terminated alkanethiol SAMs.5 However,
it is not clear whether this partial ordering would reveal a
characteristic of the ordered phase, which is discussed below.
We further note that the disordered RT-SAM did not show
any sign of ordering after weeks of storage at RT.
ꢀ
these facets run preferentially along h110i equivalent direc-
tions on thermally annealed Au/mica substrates.24 Therefore,
ꢀ
the molecular row directions corresponded to an h112i equiva-
lent set, which comprised the NNN directions on the Au(111)
lattice. The domain sizes mostly did not exceed B10 nm, and
small fragments of the rows were often observed between the
domains as well as in the middle of the domains. These small
segments did not seem to obey the 1201 rule strictly. In
addition to these irregularities, other defects such as disloca-
tions hindered an exact determination of the corrugation
period of the row structure, which was observed to be close
to 1.5 nm. Despite our efforts, molecular resolution of the
ordered phase could not be acquired, leaving the lateral
ꢁc
This journal is the Owner Societies 2008
Phys. Chem. Chem. Phys., 2008, 10, 3138–3149 | 3139

View Article Online
Fig. 3 STM topographs of the PyC12SH SAM showing dynamic
pattern evolutions: (a) obtained after 20 min of scans with It = 50 pA
and Vtip = ꢃ0.8 V, (b) with It = 500 pA and Vtip = 1.0 V after (a), (c)
after 10 min of scans with It = 50 pA and Vtip = ꢃ0.8 V, (d) after 12 h
of (c) with It = 50 pA and Vtip = ꢃ0.8 V, (e) with It = 500 pA and
Vtip = ꢃ1.0 V after a scan with It = 500 pA and Vtip = 1.0 V. The
selected area is 50 ꢀ 50 nm2.
substrate. After an additional 10 min of scans, the molecules
became ordered better, perhaps commensurable with the sub-
strate. The row structures in the encircled area of Fig. 3c were
evolved from the less-compliant rows of Fig. 3b to comply
with the 1201 rule imposed by the symmetry of the Au(111)
substrate. Through these processes, Au vacancy coarsening
occurred to a lesser extent than that in Fig. 2b, presumably
due to less invasive scan conditions.
Fig. 2 STM topographs of PyC12SH SAM: (a) pristine RT-SAM
imaged with It = 30 pA and Vtip = 0.3 V, (b) the same area imaged
with It = 30 pA and Vtip = ꢃ1.0 V after 40 min of consecutive scans
with the previous scan conditions. Corrugation height of the molecular
rows in (b) is B0.07 nm. The scanned area is 100 ꢀ 100 nm2. Contrast
scales are the same in both images.
arrangement of the molecules in the row structure undeter-
mined.
The same area was scanned after being left for 12 h to check
the stability of the scan-ordered phase (Fig. 3d). Au vacancy
coarsening occurred to a large extent in this area. A similar
phenomenon for dodecanethiol SAMs has been previously
reported and attributed to a relaxation of residual stress
induced in the system upon scanning.23 The row structures
seen in Fig. 3c were also disrupted by this relaxation process,
leading to random cluster-like structures, such as those seen in
Fig. 3a. However, this observation did not suffice to conclude
that the ordered phase was metastable, because this disruption
of row structures was accompanied or driven by the Au
sublayer relaxation. After these processes, the area was
scanned with a positive tip bias, and then imaged with a
negative tip bias. This scanning with a positive tip bias caused
the row structures to be re-formed (Fig. 3e). It is evident that
the formation of row structures did not prevent the SAM from
being relaxed and reorganized. Therefore, any interpretation
2.3 Gradual ordering and stability of the scan-ordered phase
The scan-induced ordering of PyC12SH SAM was investigated
step by step for its mechanism and stability. Fig. 3a was
obtained after 20 min of scans with a relatively less invasive
scan condition (It = 50 pA, Vtip = ꢃ0.8 V). The STM
topography in this phase was characterized by randomly
distributed small corrugations with lateral sizes smaller than
a few nm. After this area was scanned with It = 500 pA and
Vtip = 1.0 V, molecular row patterns suddenly appeared.
However, the characteristic three-fold rotational symmetry
seen in Fig. 2b was not evident at this stage (the encircled
area of Fig. 3b). We assume that at this stage, the scanned area
was in its intermediately ordered phase, where the S–Au head
parts of the PyC12S molecules were not yet periodically ar-
ranged and ordered in registry with the underlying Au(111)
ꢁc
This journal is the Owner Societies 2008
3140 | Phys. Chem. Chem. Phys., 2008, 10, 3138–3149

View Article Online
of the row structure formation with possible polymerization of
pyrrolyl endgroups could be definitely excluded.
We found that negative tip biases gave crisper and more
stable images, while positive tip biases gave quicker ordering.
This phenomenon may indicate that electronic or electrostatic
contribution was considerable in the scan-induced ordering
process. The influence of tip bias polarity on the stability of
alkanethiol SAMs has been discussed by Yang and Liu.5 In the
report, desorption of the molecules from the SAMs was
observed at large negative tip biases, while no desorption
was observed at positive tip biases. It has been argued that
negative tip biases tend to weaken the Au–S bonds via field-
induced charge redistribution. This property has been inver-
sely utilized by many experimenters: positively pulsed tip
biases have been used for in situ tip cleaning on alkanethiol
SAMs. However, PyC12SH SAMs showed a different prop-
erty: negative tip bias pulses gave better results in tip cleaning,
while scanning with positive tip biases tended to contaminate
the tip.
Fig. 4 An STM topograph of PyC11SH SAM ordered by scanning
for 10 min with It = 30 pA and Vtip = 0.4 V. (Imaging conditions:
It = 30 pA and Vtip = 0.874 V).
It is possible that this polarity-reversed response was caused
by a difference in molecular dipole moments.25 If this was the
case, weakening of Au–S bonds might not be the reason for
the facile rearrangements in the PyC12SH SAM by positive tip
biases. Additionally, it is also possible that packing of pyrrolyl
endgroups was promoted by positive tip biases. With a
positively biased tip engaged, pyrrolyl moieties near the tip
would become more negatively charged. Delocalization of this
excess negative charge may be facilitated by close-packing of
the pyrrolyl moieties, thus lowering the overall system energy.
The appearance of molecular row structures, which did not
follow the symmetry of Au(111) substrate (in Fig. 3b), may
imply the importance of pyrrolyl endgroup packing as a
driving force of the ordering. Finally, it is hardly probable
that the difference in the ordering efficiency caused by tip bias
polarity change resulted from a difference in relative tip
heights. Although typically measured I–Vtip curves on
PyC12SH SAMs were asymmetric, which indicated that the
magnitude of tunneling current was changed with tip bias
polarity, the difference was not substantial to make a signifi-
cant change in tip height.
headgroups (the S–C bonds being sp3-hybridized). Therefore,
the SAMs are formed in various molecular arrangement
structures with larger unit cells. While a striking variety of
SAM structures has been obtained in those systems with
shorter alkyl linker lengths, such a variation in SAM struc-
tures was not observed in atomic force microscopy (AFM)
measurements of PhCnSH SAMs with long alkyl linkers (n Z
12).3 This is thought to be caused by an enhanced stress
relaxation via long alkyl linkers, which may enable a herring-
bone packing among the phenyl endgroups irrespective of the
number of methylene linkers being odd or even. The same
explanation may be applied to this PyCnSH SAM system. The
absence of odd–even effect in scan-induced ordering of PyC-
nSH SAMs might imply that (1) the row structure in the
ordered phase was largely determined by an ordering between
the pyrrolyl endgroups as the hexagonal structure in PhCnSH
SAMs was determined by the herringbone ordering between
the phenyl groups,27 and (2) the PyCnS molecules (with either
odd or even n) adapted themselves for the packing of the
pyrrolyl endgroups into the row structure, possibly via some
deformation within the alkyl chain linkers.
2.4 Odd–even effect in scan-induced ordering
The scan-induced ordering effect was also observed in
PyC11SH SAM, which was also in a disordered phase when
assembled at RT. After being stimulated by repetitive STM
scans, the SAM became ordered as shown in Fig. 4. All
characteristics of the ordered PyC11SH SAM, such as the
inter-row corrugation period, were similar to those of the
scan-ordered PyC12SH SAM.
2.5 Ordering by thermal annealing at 70 1C
After being annealed at a higher temperature (B70 1C), the
PyC12SH SAM turned into a quasi-ordered phase (Fig. 5a).
The corrugation period of the row structure was measured to
be 1.47 ꢂ 0.07 nm. Symmetry of the domain orientations was
also the same as that of the scan-ordered phase. Therefore, we
conclude that this ordered phase in thermally annealed SAM
(HT-SAM) had the same molecular arrangement as the scan-
ordered phase.
Distinguished odd–even effects have been observed in many
alkanethiol SAM systems on Au(111) with phenyl-based end-
groups.3,5–8,26 It is well known that phenyl- or methylbiphenyl-
terminated SAMs with odd numbers of methylene linkers
(PhCnSH or BPn SAMs with odd n) form (O3ꢀ2O3)R301-
based structures with the aromatic endgroups packed in a
herringbone fashion. For even n, this endgroup packing is
hindered as the tilt angle of the aromatic rings from surface
normal increases due to a fixed bonding geometry of the thiol
Although average domain size in the HT-SAM was larger
than that of the scan-ordered phase shown in Fig. 2b, there
still were many defects in each domain. These defects included
discontinuities in row structures and dislocations in some of
the parallel rows, implying that the commensurability between
the ordered rows of the pyrrolyl endgroups and the Au–S
ꢁc
This journal is the Owner Societies 2008
Phys. Chem. Chem. Phys., 2008, 10, 3138–3149 | 3141

View Article Online
tors with small indices on Au(111). Some of the surface vectors
were indicated in Fig. 5b. It was often observed that etch pits
in thiol SAMs on Au(111) with (O3ꢀO3)R301-based struc-
tures had hexagonal (or triangular) shapes with edges parallel
8,9
to the h112i or h110i29 directional sets which were shown by
the black and the white triangle in Fig. 5a, respectively. Since
the two directional sets correspond to NN directions of the
molecules and of the top Au atoms, such directional formation
of step boundaries should contribute to minimize the system
energy, depending on the relative strength of intermolecular
interactions to that of interatomic interactions in the substrate.
However, coalesced etch pits in the HT-SAM of PyC12SH did
not show such a preference in edge directions. This observa-
ꢀ
ꢀ
tion indicates that the ordered structure was not
a
(O3ꢀO3)R301-based one, and/or the ordered phase had such
a high defect density that faceting in some preferential direc-
tions would not give even a marginal reduction in the system
energy.
The corrugation height of molecular rows in this ordered
phase was measured to be about 0.07 nm under the scan
conditions. This corrugation height is typical for ordered thiol
SAMs terminated by aromatic moieties.6,30 There have been
reports of corrugations of similar magnitudes (0.05–0.1 nm) in
superstructures of ordered phases of H3C–C6H4–C6H4–(CH2)n–
SH (BPn) SAMs on Au(111).7,8 In these systems, each of the
ordered phases had a superstructure that was superimposed
over a simpler base structure, and the molecular packings
inside lower features turned out to be the same as those in
higher features of the superstructures. The periodic topo-
graphic features in STM images were explained to be caused
by a small mismatch between ideal lattice dimensions for the
BPn monolayers and the underlying Au substrate. In other
words, the superstructures were attributed to S atoms being
adsorbed at different sites on Au substrates. The authors also
acknowledged possible reconstructions in the Au top layer,
which may contribute to lower the overall system energy. The
same explanations might be applicable to the ordered phase of
PyC12SH SAM.
Fig. 5 STM topographs of PyC12SH SAM annealed at 70 1C. (a) A
close-up scan image from the boxed area of (b). Triangles with edges
parallel to NNN (black) or NN (white) directions of the Au(111)
substrate were drawn as guides. The scanned area is 100 ꢀ 100 nm2.
(b) A large scale image from an area of 400 ꢀ 400 nm2. Most of the
coalesced etch pits had rounded-off polygonal shapes with edges
parallel to surface vectors with small indices on Au(111). Some of
the edge directions were indicated by indexed arrows. Images were
taken with It = 20 pA, Vtip = ꢃ1.0 V.
2.6 Molecular coverage of PyC12SH SAM
The molecular coverages of PyC12SH RT- and HT-SAMs
were measured by cyclic voltammetry (CV). A series of
successive voltammetric scans for the HT-SAM is given in
Fig. 6. An irreversible anodic peak was observed near the
electrode potential of 0.8 V, which was attributed to oxidation
of pyrrolyl units in the SAM.11–14 No polymeric wave could be
seen in voltammetry cycles subsequent to the first cycle,
probably because no special efforts were carried out to remove
residual water from the electrolyte solution.11 Because neither
polymeric wave nor pyrrolyl monomer oxidation31 prior to the
CV scans was detected, the number of molecules in the CV-
active area could be counted by a ratio of 2eꢃ per mole-
cule,12,13 using the shaded area in Fig. 6 as the total charge
transferred from the pyrrolyl units. The average surface
molecular densities of RT- and HT-SAMs were calculated to
headgroup parts was not perfect, and thus a stress, which was
high enough to limit the molecular row length to r10 nm, was
developed along the row direction. There were also small areas
of indefinite shapes that appeared to be higher than the
ordered area. These areas were thought to be covered by
molecules that had smaller tilt angles and larger lateral
molecular densities than those in the ordered area. The degree
of long range order or the integrity of molecular arrangement
in the ordered phase did not appear to affect the free energy of
the system significantly.
As a result of Ostwald ripening process during the thermal
annealing,7,28 the Au vacancies coalesced into larger ones with
a lower number density compared to that of the RT-SAM
(Fig. 5b). The coalesced etch pits mostly had rounded-off
polygonal shapes, with edges parallel to various surface vec-
be (6.0 ꢂ 0.2)ꢀ10ꢃ10 and (5.9 ꢂ 0.2)ꢀ10ꢃ10 mol cmꢃ2
,
respectively. These values corresponded to B77% of the
experimental alkanethiol SAM coverage (7.8ꢀ10ꢃ10 mol
ꢁc
This journal is the Owner Societies 2008
3142 | Phys. Chem. Chem. Phys., 2008, 10, 3138–3149

View Article Online
SAMs were calculated to be 79.7 and 78.9% of the alkanethiol
reference system, respectively.34 This molecular coverage of
B79% agreed well with the value of B77% obtained from the
CV measurements within experimental error.
2.7 Thickness of PyC12SH SAM
The thickness of the PyC12SH SAM was measured by ellipso-
metry. Ellipsometric measurements gave thicknesses t = 1.5 ꢂ
0.03 nm for the RT-SAM, and t = 1.44 ꢂ 0.04 nm for the HT-
SAM. Compared to t = 1.78 nm of phenyl-terminated
dodecanethiol (PhC12SH) SAM, where the molecules are
close-packed,3 the thickness of the PyC12SH SAM corre-
sponded to B84% of that of PhC12SH SAM. Since there
should be no substantial size difference between phenyl and
pyrrolyl moieties, this difference in SAM thickness must be
caused by a difference in lateral molecular densities, which
again agrees with the previous results from CV and XPS.
We have also tried to measure the thickness of the PyC12SH
RT-SAM by scanning tunneling spectroscopy (STS). During
STM scans with a relatively high impedance condition (It = 30
pA, Vtip = ꢃ1.0 V), tunneling currents with respect to tip
height were obtained at arbitrary points on the sample. A
typical tunneling current versus tip height (I–z) relation is
shown in Fig. 8.35 There was no kink in the linear fit of the
data, indicating that the tip was in contact with the SAM
within the range of ꢃ0.5 r z r 0.2 nm,5 and the tip was
partially penetrating or at least touching the SAM throughout
the STM measurements in this work. At z Z 0.2 nm, noisier
datapoints could be seen which were caused by instrumental
limitation in these measurements. Assuming tunneling resis-
tance R E 10 kO for a mechanical contact between the tip and
the gold substrate,36 the SAM thickness was estimated to be
Z 1.5 nm. However, this lower limit of the SAM thickness was
probably overestimated, because SAM thicknesses measured
by I–z STS tend to be larger than the true SAM thicknesses
due to strong interactions between the biased tip and the
molecules.9
Fig. 6 Cyclic voltammograms obtained with a scan rate of 0.1 V sꢃ1
from Au electrode surfaces covered by PyC12SH SAM annealed at
70 1C. The shaded area corresponds to the oxidation peak area
of pyrrolyl units in the SAM.
cmꢃ2, which is compatible with the (O3ꢀO3)R301-based
close-packing),32 and there was no appreciable difference in
molecular coverage between RT- and HT-SAM of PyC12SH.
We crosschecked the molecular coverage of the PyC12SH
SAMs by using XPS. It has been well established that photo-
electron intensity from Au substrates covered with alkanethiol
SAMs decreases logarithmically with increasing the alkyl
chain length.33 Effective thicknesses of the SAMs were esti-
mated by comparing the attenuation of Au 4f XPS signals with
those from a reference system of simple alkanethiol SAMs
(CnSH SAM, with n = 10, 12, 15, 16, 18). The least-squares fit
from the reference system was used to determine the effective
thicknesses and lateral packing densities of the PyC12SH
SAMs (Fig. 7). The effective thicknesses of the RT- and HT-
SAMs of PyC12SH were determined to be equivalent to those
of C12.5SH and C12.4SH, respectively. From these values, the
lateral packing densities of PyC12S molecules in RT- and HT-
Fig. 7 Attenuation of the substrate photoelectron intensity (Au 4f) of
the PyC12SH SAMs. Values from simple alkanethiol SAMs (CnSH
SAMs, n = 10, 12, 15, 16, 18) were used to obtain effective molecular
lengths of the PyC12SH SAMs. The effective molecular lengths were
used to estimate surface molecular coverages of the SAMs.
Fig. 8 Tunneling current versus tip height (I–z) of the PyC12SH RT-
SAM. By a linear fit of log(I) versus z, the SAM thickness was
estimated to be Z 1.5 nm. The feedback setpoint (It = 30 pA, Vtip
= ꢃ1.0 V) was used for the reference tip height (z = 0).
ꢁc
This journal is the Owner Societies 2008
Phys. Chem. Chem. Phys., 2008, 10, 3138–3149 | 3143

View Article Online
The effective tunneling barrier height in I–z spectroscopy is
given as F = ꢃ0.952[Dln(I)/Dz]2 (in eV and A).5 From the
slopes of linear fits, a barrier height F = 1.31 ꢂ 0.07 eV was
obtained for the PyC12SH SAM. Compared with F = 1.39 ꢂ
0.01 eV for long alkanethiol SAMs,37 this value seemed
reasonable, considering that the pyrrolyl endgroups may con-
tribute to decrease the measured tunneling barrier height with
a lower energy gap or by enhancing lateral charge hopping.
normal to the chain axis. Although both of the mode inten-
sities depend on the chain tilt angle as well as on the twist
angle, we can safely assume that there is no substantial
difference in the chain tilt angle between the RT- and HT-
SAMs, based on the results that the molecular coverages and
thicknesses showed no substantial difference between the RT-
and HT-SAMs. The ratios of peak areas ns(CH2)/na(CH2)
were 0.51 and 0.28 for the RT- and HT-SAMs, respectively.
These ratios indicate that the average twist angle of the HT-
SAM was larger (closer to ꢂ901) than that of the RT-SAM.
This result was analogous to the results of Lee et al., in
which larger twist angles were obtained for PhCnSH SAMs
with even n, and smaller twist angles for odd n.3 The authors
attributed this phenomenon to different degrees of interactions
of terminal phenyl group in odd- versus even-numbered films,
which resulted from different tilt angles of the terminal phenyl
groups (larger tilt angles for n = even). In other words, twist
angles of the alkane chains in PhCnSH SAMs with even n
should be larger compared to the SAMs with odd n, to
accomodate terminal phenyl groups into an ordered state by
reducing their vertically projected size. The major difference
between our results and those from PhCnSH SAMs is that in
our results, the twist angle change came only from thermal
annealing, which was not accompanied by an appreciable
change in the surface molecular coverage.
2.8 Crystallinity of PyC12SH SAM
We qualitatively analyzed the degree of packing and orienta-
tion of PyC12S molecules on the RT- and HT-SAMs through
PIERS measurements (Fig. 9). It is well known that informa-
tion on alkyl chain conformation and packing in alkanethiol
SAMs can be deduced from the peak positions of the CH2
stretching vibrations.38 In the IR spectra, the peaks from the
symmetric methylene stretching mode ns(CH2) appeared at
2848 cmꢃ1, and those from the antisymmetric stretching mode
na(CH2) appeared at 2918–2919 cmꢃ1. These values match
with those from well-ordered and close-packed long alka-
nethiol SAMs.38 In addition, it has been suggested that a
frequency increase in these vibration modes may happen with
reduced chain–chain interactions, even when the alkyl chains
are in an all-trans conformation.38 Therefore, it can be con-
cluded that the alkyl chain parts in both the RT- and HT-
SAMs of PyC12SH were not only in almost all-trans config-
urations, but also compactly packed with each other.
Spectral intensities of vibrational modes in the pyrrolyl
endgroups were also compared between RT- and HT-SAMs
for orientational information in a similar way. In the range of
900–1600 cmꢃ1, three absorption peaks were well discrimi-
nated from the background: two of them were from in-plane
modes n(CQC) and d(C–H), and one from n(C–N),13 which
was assumed to be also in the ring plane by sp2-hybridization
of the N atom. All three peaks from the HT-SAM were smaller
than those from the RT-SAM, indicating that the pyrrole rings
in the HT-SAM were oriented more parallel to the substrate
on average. This result may seem contradictory to the result
that the average twist angle of the alkyl backbones in the
HT-SAM was larger than that of the RT-SAM. This contra-
diction could be resolved if it is assumed that there are some
Since only the components of the vibrational transition
dipole moments perpendicular to the surface plane contribute
to the absorption spectra, the relative intensities of these
methylene vibration peaks reveal additional information on
twist angle of the alkyl chains.3 The twist angle is defined as
the rotation angle of the plane containing the –C–C–C– alkyl
backbone with respect to the plane normal to the surface and
containing the chain axis, on the assumption of an all-trans
configuration. The transition dipole direction of the antisym-
metric mode is orthogonal to the backbone plane, and that of
the symmetric mode is parallel to the backbone plane and
Fig. 9 PIERS spectra of the RT- and HT-SAMs of PyC12SH.
ꢁc
This journal is the Owner Societies 2008
3144 | Phys. Chem. Chem. Phys., 2008, 10, 3138–3149

View Article Online
gauche defects near the pyrrolyl endgroups in the RT-SAM or
some of the thiols in the RT-SAM are bonded to the Au
substrate with a twist angle near 1801 (via sp-hybridization or
different adsorption sites),39,40 allowing the pyrrole rings in the
RT-SAM to be more upright than those in the HT-SAM
without affecting the infrared spectral properties of the alkyl
chains (as shown in Fig. 10a). This model also explains the
disordered nature of RT-SAM of PyC12SH observed by STM
(Table 1).
Table 1 Static contact angles (ys) of dimethyl formamide (DMF),
methylene iodide (MI), and nitrobenzene (NB) on the PyC12SH SAMs
ys (DMF)
ys (MI)
ys (NB)
PyC12SH-RT
PyC12SH-HT
39.3 ꢂ 1.41
38.7 ꢂ 1.51
33.7 ꢂ 1.11
30.7 ꢂ 1.21
21.1 ꢂ 0.91
16.2 ꢂ 0.91
p(3ꢀ1) superlattice of (O3ꢀO3)R301 structure), let alone the
difference in surface molecular coverage.
From the fact that the PyC12SH HT-SAM had a molecular
coverage of Z 77% of the close-packed alkanethiol SAMs
while having a compactly-packed alkyl linker layer, it can be
deduced that the tilt angle of the PyC12S was larger than that
of the close-packed alkanethiol SAMs. Barrena et al. have
reported various below-saturation phases of alkanethiol
SAMs that have larger molecular tilt angles than B351 of
the (O3ꢀO3)R301-based structures.19 Among them,
(2ꢀO3)rect and (4ꢀO3)rect phases with a tilt angle of B501
had a surface molecular coverage of 75% with respect to that
of saturated (O3ꢀO3)R301-based structures. A (2ꢀ2)-based
structure with the same 75% coverage also has been predicted
for alkanethiol SAM systems by a chain-interlocking model,21
which was not observed experimentally.20 These structures or
their variants could be possible candidates for the Au–S
registry of the ordered PyC12SH SAM. However, although
the molecular coverage of these models is close to the experi-
mental values for the ordered PyC12SH SAM, these structural
models are not quite compatible with the row structure along
We verified the orientational properties of the pyrrolyl
endgroups in the PyC12SH SAMs via contact angle measure-
ments. For all of the three probe liquids used (dimethyl
formamide, methylene iodide, and nitrobenzene), contact
angles were lower on the HT-SAM than on the RT-SAM of
PyC12SH. This trend was also consistent with the results of
Lee et al.,3 in which advancing contact angles of the probe
liquids on PhCnSH SAMs were shown to be lower for even n
than for odd n, probably due to a larger tilt angle of the
terminal phenyl groups in the SAMs with even n. Therefore,
this result supports the observation that the pyrrole rings in
the HT-SAM were oriented more parallel to the substrate with
larger tilt angles than those in the RT-SAM.
2.9 Ordered phase of PyC12SH SAM
Although molecularly resolved STM images could not be
obtained for the ordered PyC12SH SAM, a feasible molecular
structure could be proposed based on the coverage, thickness,
and orientational properties obtained for the SAM. The row
structure in the ordered PyC12SH SAM bore a superficial
resemblance with the AFM scan-induced p(3ꢀ1) superstruc-
ture observed in decanethiol SAM.18 However, there was a
difference in the corrugation period between the p(3ꢀ1) phase
and the ordered phase of the PyC12SH SAM (1.33 nm for the
ꢀ
h112i directions with a corrugation period of B1.47 nm.
Fig. 10c shows a (5ꢀO3)rect structure on Au(111) that has a
row corrugation period of 1.44 nm and a surface molecular
coverage of 90% of the close-packed alkanethiol SAMs. This
ꢀ
structure has row directions parallel to the h112i directional set
and its row corrugation period is similar to the experimental
value, 1.47 ꢂ 0.07 nm. We propose that the ordered PyC12SH
SAM has a molecular arrangement based on this structure.
The disparity in surface molecular coverages obtained from
the experiments (B77–79%) and of the model (90%) can be
attributed to molecular defects which are thought to be
manifested in the ordered PyC12SH SAM shown in Fig. 5a.
A similar molecular row pattern has been reported to occur in
microcontact-printed dodecanethiol SAMs on Au(111) by
Larsen et al.10 The authors assumed this phase to have a
(5ꢀO3)rect structure with 2 molecules per unit cell. However,
this assignment of the surface molecular density (60%) had
not been firmly based on observations.
In a recent molecular dynamics simulation, the liquid state
of pyrrole has been described as a complex aggregate of
pyrrole dimers oscillating between hydrogen-bonded T-shapes
and p-bonded stacked forms.41 This report may imply that a
near parallel-displaced arrangement between pyrrolyl groups
can be one of possible structural features in the ordered phase
of the PyC12SH SAM, in addition to the usual herringbone-
like arrangement between small aromatic moieties such as
phenyl or biphenyl groups. A hypothetical arrangement be-
tween the pyrrolyl endgroups is drawn in Fig. 10c, which
pertain to those structural elements aforementioned: herring-
bone-like and parallel-displaced arrangements between the
pyrrolyl endgroups that are heavily tilted from the surface
Fig. 10 Schematic drawings of the PyC12SH SAMs: (a) disordered
RT-SAM, and (b) ordered SAM. The alkyl chains in both SAMs are
thought to be closely packed, although the substrate–molecule registry
might be different. (c) The proposed structure of the ordered PyCnSH
SAM ((5ꢀO3)rect phase). Small circles represent top layer Au atoms.
A hypothetical arrangement between the pyrrolyl endgroups is illu-
strated on the right side of the drawing. Choice of S–Au adsorption
sites is arbitrary, and possible substrate reconstructions are not
considered in the drawing.
ꢁc
This journal is the Owner Societies 2008
Phys. Chem. Chem. Phys., 2008, 10, 3138–3149 | 3145

View Article Online
normal. In a recent study of 9-mercaptoanthracene SAM on
Au(111), a distance of B0.56 nm between p–p near-parallel-
displaced anthryl groups has been reported.42 In the proposed
(5ꢀO3)rect structure, intermolecular distance along the row
direction should be the NNN distance of Au(111), which is
0.4995 nm. This mismatch in the intermolecular distance may
be the source of stress, which is thought to have caused the
molecular defects in the ordered phase of the PyCnSH SAM.
strong dipole of the pyrrolyl group still disturbs the usual
herringbone packing between aromatic moieties,46,47 yielding
disordered SAMs for these molecules. This dipole-induced
disorder seems to persist even in the PyC12SH RT-SAM, which
was shown to have a compactly packed alkyl linker layer by the
PIERS analysis and a higher surface molecular coverage by the
CV and XPS analyses, in contrast to PyCnSH SAMs with
smaller n. As was suggested in section 2.8, this disorder is
probably enabled by gauche defects near the pyrrolyl endgroups
as depicted in Fig. 10a. A parallel-shifted packing between
pyrrolyl groups such as shown in Fig. 10b may help to relieve
the destabilizing dipole–dipole interaction, which is probably
the main reason for the disorder in RT-SAMs of PyCnSH. Such
endgroup packing also conforms to the molecular orientational
properties deduced from the PIERS data.
2.10 Disorder and order in PyCnSH SAMs
The ordering in PyCnSH SAMs had similar aspects to the
(O3ꢀO3)R301 - c(4ꢀ2) superlattice transformation ob-
served in decanethiol SAM on Au(111).18 It can be induced
by both STM scan and thermal annealing, and causes a change
in the average twist angle of the alkane chains, while not
causing a practical change in lateral molecular coverage. There
also have been other reports of phase transformations in thiol
SAM systems on Au(111) which are induced by thermal
annealing or scanning probes (with STM or AFM).8,19–21
For example, Cyganik and Buck have reported a thermally
induced irreversible phase transformation in BP4 SAM on
Au(111).4 In this system, the SAM, initially in a (5O3ꢀ3)rect
phase with an area per molecule of 27 A2 (a-phase) at room
temperature, was transformed into a (6O3ꢀ2O3)oblique
phase with an area per molecule of 32.4 A2 (b-phase), which
was less dense and more stable, after annealing at 373 K. They
attributed this phenomenon to an energy hyperspace which
had a number of pronounced local minima caused by a
competition between different factors contributing to the total
energy of the system. Factors such as Au–S–C bonding geo-
metry, molecular adsorption sites, interaction between alkyl
chains, and interactions between aromatic endgroups (such as
p-stacking, steric repulsion, and dipole interaction) were
thought to determine the energy hyperspace. The major differ-
ence between the ordering in PyCnSH SAMs and the phase
transformations in decanethiol and BP4 SAMs was that the
PyCnSH SAMs were formed in a disordered phase at RT,
whereas the other SAMs were already in ordered phases at RT.
To date, there has not been any report of disorder–order
phase transformation in molecular SAM systems. We suspect
that this special characteristic of PyCnSH SAMs is caused by a
much stronger and complex interaction between the pyrrolyl
endgroups that have a heterogeneity within the aromatic ring:
the N atom. The disordered nature of the SAMs at RT may be
caused by a dipole–dipole interaction between the pyrrolyl
groups that disturbs an orderly packing of the endgroups and
imposes a high energy barrier to the ordered phase. As was
mentioned previously, PyCnSH molecules with n r 10 also
form disordered SAMs. On the contrary, it is known that
PhCnSH SAMs with odd n are ordered via a herringbone
arrangement of phenyl endgroups in the (O3ꢀ2O3)R301
structure.3,5 It is quite noteworthy that a small difference in
the aromatic endgroups (phenyl versus pyrrolyl) leads to such
a change in molecular arrangement.43
It seems highly probable that kinetic factors rather than
thermodynamic ones determine the phase of the RT-SAMs.
However, although only an insignificant difference in surface
molecular coverage was found between the RT- and HT-
SAM, this did not necessarily mean that surface molecular
densities were identical between the ordered and the disor-
dered phases. Strictly speaking, the HT-SAM consisted of two
different phases: the ordered phase, and a disordered phase
which had a higher surface molecular density than that of the
ordered phase, as judged by higher STM heights of small
disordered areas in Fig. 5a. This means that surface molecular
density of the disordered phase may be slightly higher than
that of the ordered phase. Therefore, it is possible that the
disordered state was not just a kinetically limited state, but a
thermodynamically favored one over the ordered state at RT
in terms of total enthalpy of the self-assembly system.
As has been discussed by Touzov and Gorman,18 mechan-
ical interactions between scanning probe tips and samples are
thought to be mostly responsible for scan-induced phase
changes observed in many thiol SAM systems. The same
mechanism must be also responsible for the scan-induced
ordering in this work, because the STM tips partially penetrate
into the SAM under the scan conditions used here. The
mechanical agitation exercised by the STM tips must have
helped the disordered PyCnS molecules to overcome the
energy barrier to the ordered phase. However, as was dis-
cussed in the previous section, contributions from electrostatic
and/or electronic interactions between the tip and the PyCnS
molecules may also be considerable. In addition, there might
be another factor which was related to the bias polarity-
dependent ordering efficiency: transfer of PyCnS molecules
from the SAMs to the tip. As was stated earlier, the disordered
RT-SAM is thought to have a slightly higher surface mole-
cular density than the ordered phase. It is possible that
molecular transfer from the SAM to the tip, which was
enhanced at positive tip biases, had accelerated the ordering
process.
3. Experimental
Pyrrole molecules are known to form N–Hꢄ ꢄ ꢄp hydrogen
bonds readily with each other.44,45 Although such an inter-
molecular bonding is decapacitated in the alkanethiol-deriva-
tized PyCnSH molecules, it is highly probable that an inherent
Materials
Pyrrole (98%, Aldrich), 1,11-dibromoundecane (98+%, Al-
drich), 1,12-dibromododecane (98%, Aldrich), thiourea
ꢁc
This journal is the Owner Societies 2008
3146 | Phys. Chem. Chem. Phys., 2008, 10, 3138–3149

View Article Online
(99%, Sigma–Aldrich), potassium hydroxide (KOH, 85%,
Junsei), sodium hydroxide (97%, Sigma–Aldrich), hydrochlo-
ric acid (HCl, 35%, Junsei), 1-decanethiol (96%, Aldrich), 1-
dodecanethiol (98+%, Aldrich), 1-pentadecanethiol (98%,
Aldrich), 1-hexadecanethiol (95%, Fluka), 1-octadecanethiol
(98%, Aldrich), tetrabutylammonium perchlorate (99%, Flu-
ka), absolute ethanol (EtOH, 99.8+%, Merck), dichloro-
methane (CH2Cl2, 100.0%, J. T. Baker, HPLC grade),
anhydrous N,N-dimethylformamide (DMF, 99.8%, Sig-
ma–Aldrich), methylene iodide (99%, Sigma–Aldrich), nitro-
benzene (99+%, Sigma–Aldrich), anhydrous acetonitrile
(99.8%, Sigma–Aldrich) were used as received. Ultrapure
water (18.3 MO cm) from Human Ultra Pure System (Human
Corp., Korea) was used. Au/mica substrates were purchased
from Agilent Technologies AFM (Arizona, USA), and Au/Cr/
glass substrates were purchased from Arrandee (Germany).
formed samples in the same solution at 70 1C for an hour.49
Afterwards, the solution temperature was gradually lowered
to room temperature before the sample was taken out from the
solution. After the incubation, the substrates were rinsed with
pure ethanol, and dried with an Ar stream. All experiments
were conducted with freshly prepared samples.
Scanning tunneling microscopy (STM)
All STM measurements were carried out under an ultra-high
vacuum (UHV) using a UHV-VT-SPM from Omicron Nano-
technology GmbH (Germany). All STM images were obtained
in constant current mode, and tip biases with respect to
grounded samples were presented throughout. Electrochemi-
cally etched W tips were used in the experiment. To compen-
sate for errors in determining lateral feature sizes in the STM
topographies caused by non-proportionality in piezo motion
of the scanner and by differences in tip length, a highly
oriented pyrolytic graphite sample was measured using the
same scan size and speed of the SAM sample measurements
and analyzed as a calibration reference.
Synthesis of molecules for the SAMs
12-(1H-Pyrrol-1-yl)dodecane-1-thiol (PyC12SH) and 11-(1H-
pyrrol-1-yl)undecane-1-thiol (PyC11SH) were synthesized fol-
lowing the reported procedure.14,15 In brief, pyrrole (5.7
mmol) and KOH (6.5 mmol) were added to anhydrous
DMF solution (40 mL) containing 1,12-dibromododecane or
1,11-dibromoundecane (16 mmol), and the mixture was stirred
to react at room temperature overnight. The reaction product
was extracted with diethyl ether after addition of water (40
mL). After drying with MgSO4, concentration of the product
in vacuo followed by column chromatography provided 1-(12-
bromododecyl)-1H-pyrrole (47% yield) or 1-(11-bromounde-
cyl)-1H-pyrrole (44% yield) as a pale yellow liquid. The
synthesized compound (2.0 mmol) was mixed with thiourea
(6 mmol) in ethanol (50 mL) and the solution was heated to
reflux for 8 h. The reaction product was allowed to cool down
to room temperature under an Ar purge, and then NaOH
solution (4 mL, 10% w/w) was added to the product. The
solution was heated to reflux for 4 h. Upon cooling to room
temperature, the hydrolyzed solution was titrated with dilute
HCl to pH 7. After extraction with diethyl ether, the organic
phase was rinsed with water, dried with MgSO4, and then
purified by column chromatography to give PyC11SH (80%
Cyclic voltammetry (CV)
Cyclic voltammograms were acquired using a BAS 100B
(Bioanalytical Systems, Inc.). The three-electrode electroche-
mical cell consisted of a modified Au electrode, a Pt wire
counter electrode, and an Ag/Ag+ reference electrode. Experi-
ments were carried out in degassed and anhydrous acetonitrile
solution containing tetrabutylammonium perchlorate (0.1 M)
as a carrier electrolyte. The active area of the gold electrode
was 0.283 cm2, and the surface coverage values were corrected
for surface roughness, assuming a roughness factor of 1.2.32,50
X-Ray photoelectron spectroscopy (XPS)
The XPS spectra were obtained by using a VG-Scientific
ESCALAB 250 spectrometer (UK) with a monochromatized
Al Ka X-ray source (1486.6 eV) at an incidence angle of 581.
Emitted photoelectrons were detected by a multichannel de-
tector at a takeoff angle of 901 relative to the surface. During
the measurements, the base pressure was 10ꢃ9–10ꢃ10 Torr. The
lateral packing densities of the SAMs were calculated by
investigating the attenuation of the Au 4f XPS signal. As a
1
yield) and PyC12SH (76% yield) as a pale yellow liquid. H
NMR (300 MHz, CDCl3): PyC11SH, d 1.27 (m, 15H), 1.66 (m,
2H), 1.76 (m, 2H), 2.53 (q, 2H), 3.86 (t, 2H), 6.14 (t, 2H), 6.65
(t, 2H); PyC11SH, d 1.27 (m, 15H), 1.64 (m, 2H), 1.76 (m, 2H),
2.53 (q, 2H), 3.87 (t, 2H), 6.14 (t, 2H), 6.65 (t, 2H). MS(EI) for
C15H27NS (PyC11SH): calcd 253.1864, found 253.1512.
MS(EI) for C16H29NS (PyC12SH): calcd 267.2021, found
267.1428.
reference system,
a series of SAMs of n-alkanethiols
(CH3(CH2)nꢃ1SH, n = 10, 12, 15, 16, 18) on gold were used
to determine the attenuation of the Au 4f signals as a function
of effective molecular length.
Ellipsometry
The thicknesses of the monolayer films were measured with a
Gaertner L116s ellipsometer (Gaertner Scientific Corp., IL,
USA) equipped with a HeNe laser (632.8 nm) at a 701 angle of
incidence. A refractive index of 1.46 was used for the films.
Preparation of the SAMs
The SAMs of PyCnSH (n = 11, 12) were formed on Au/mica,
Au/Cr/glass (for STM measurements), or Au/Cr/Si substrates
(for other measurements). The Au/Cr/glass substrates were
hydrogen flame-annealed prior to the SAM formation. The
SAMs were formed by immersing the gold substrates into
ethanolic solutions (0.1 mM) of the thiol compounds under Ar
atmosphere overnight at room temperature (RT).48 Thermal
annealing of PyC12SH SAM was done by immersing the RT-
Polarized infrared external reflectance spectroscopy (PIERS)
The PIERS spectra were recorded with a nitrogen-purged
Thermo Nicolet Fourier transform infrared spectrometer
(model: NEXUS). The instrument was equipped with a liquid
nitrogen-cooled mercury–cadmium–telluride (MCT) detector
and a smart apertured grazing angle (Smart SAGA) apparatus
ꢁc
This journal is the Owner Societies 2008
Phys. Chem. Chem. Phys., 2008, 10, 3138–3149 | 3147

View Article Online
for grazing-angle reflectance IR spectroscopy. The p-polarized
light was incident at 801 relative to the surface normal of the
substrate. The spectra were taken by averaging approximately
2000 scans for background and 1000 scans for the samples
(resolution of 2 cmꢃ1), and the final spectra were obtained
with minimal baseline correction.
control structural properties of SAMs of various endgroups
with diverse functionalities.
References
1 J. C. Love, L. A. Estroff, J. K. Kriebel, R. G. Nuzzo and G. M.
Whitesides, Chem. Rev., 2005, 105, 1103.
2 C. Vericat, M. E. Vela and R. C. Salvarezza, Phys. Chem. Chem.
Phys., 2005, 7, 3258.
Contact angle measurement
3 S. Lee, A. Puck, M. Graupe, R. Colorado, Jr, Y.-S. Shon, T. R.
Lee and S. S. Perry, Langmuir, 2001, 17, 7364.
4 P. Cyganik and M. Buck, J. Am. Chem. Soc., 2004, 126, 5960.
5 G. Yang and G. Liu, J. Phys. Chem. B, 2003, 107, 8746.
Static contact angles were measured by using a Phoenix 300
goniometer (Surface Electro Optics Co., Ltd., Korea) at room
temperature under ambient conditions. Dimethyl formamide,
methylene iodide and nitrobenzene were used as probe liquids.
6 W. Azzam, P. Cyganik, G. Witte, M. Buck and C. Woll, Langmuir,
¨
2003, 19, 8262.
7 P. Cyganic, M. Buck, W. Azzam and C. Woll, J. Phys. Chem. B,
¨
2004, 108, 4989.
8 P. Cyganik, M. Buck, J. D. E. T. Wilton-Ely and C. Woll, J. Phys.
Chem. B, 2005, 109, 10902.
9 G. J. Su, R. Aguilar-Sanchez, Z. Li, I. Pobelov, M. Homberger, U.
Simon and T. Wandlowski, ChemPhysChem, 2007, 8, 1037.
10 N. B. Larsen, H. Biebuyck, E. Delamarche and B. Michel, J. Am.
Chem. Soc., 1997, 119, 3017.
11 R. J. Willicut and R. L. McCarley, Anal. Chim. Acta, 1995, 307,
269.
12 R. J. Willicut and R. L. McCarley, J. Am. Chem. Soc., 1994, 116,
10823.
13 R. L. McCarley and R. J. Willicut, J. Am. Chem. Soc., 1998, 120,
9296.
14 R. J. Willicut and R. L. McCarley, Langmuir, 1995, 11, 296.
15 C.-G. Wu, S.-C. Chiang and C.-H. Wu, Langmuir, 2002, 18, 7473.
16 Y.-H. Kim, D. B. Wurm, M. W. Kim and Y.-T. Kim, Thin Solid
Films, 1999, 352, 138.
4. Conclusions
¨
We studied structural phases of PyC11SH and PyC12SH SAMs
on Au(111). By STM topographic measurements, the SAMs
were found to be in a disordered phase when formed at room
temperature. However, the PIERS results on the PyC12SH
RT-SAM showed that the alkyl linker parts were in nearly
all-trans configuration and compactly packed with each other.
The PIERS data implied that the disorder was mostly confined
to near-terminal parts of the molecules or to S–Au bonds,
because of strong interchain interactions between the long
alkyl linkers. The disorder was supposed to be caused by
strong dipole–dipole interactions between the pyrrolyl
endgroups.
17 C. N. Sayre and D. M. Collard, Langmuir, 1995, 11, 302.
18 I. Touzov and C. B. Gorman, J. Phys. Chem. B, 1997, 101, 5263.
19 E. Barrena, E. Palacios-Lidon, C. Munuera, X. Torrelles, S.
´
Ferrer, U. Jonas, M. Salmeron and C. Ocal, J. Am. Chem. Soc.,
2004, 126, 385.
20 E. Barrena, C. Ocal and M. Salmeron, J. Chem. Phys., 2001, 114,
4210.
21 E. Barrena, C. Ocal and M. Salmeron, J. Chem. Phys., 2000, 113,
2413.
22 A. Kirakosian, M. J. Comstock, J. Cho and M. F. Crommie, Phys.
Rev. B: Condens. Matter Mater. Phys., 2005, 71, 113409.
23 J. A. M. Sondag-Huethorst, C. Schonenberger and L. G. J.
¨
The PyCnSH SAMs underwent a disorder–order phase
change by repetitive STM scans with invasive scan conditions.
This scan-induced ordering was more facile with positive tip
biases than with negative tip biases. We believe that electro-
static or electronic contributions to the ordering process were
considerable, in addition to those from mechanical perturba-
tion by the tip. It was found that the molecular ordering also
could be accomplished by thermal annealing at 70 1C. The
thermally ordered phase of the PyC12SH SAM had the same
characteristics as the scan-ordered phase, such as a row
corrugation period near 1.5 nm and abundance of defects.
On the basis of the experimental results, the molecular
arrangement in the ordered PyCnSH SAM was proposed to
be based on a (5ꢀO3)rect structure.
Fokkink, J. Phys. Chem., 1994, 98, 6826.
24 B. Lussem, S. Karthauser, H. Haselier and R. Waser, Appl. Surf.
¨
Sci., 2005, 249, 197.
¨
25 The dipole moments of C12S and PyC12S radicals have been
calculated to be 2.05 and ꢃ0.36 D from S heads to terminal
groups, respectively. These values were obtained by density-func-
tional calculations using B3PW91 with the 6-31G(d,p) basis set,
after geometry optimizations of the respective free thiol molecules.
We note that other methods such as ab initio calculations yielded
similar values. Although these values do not truly represent partial
charge distribution in the SAMs, they make a good basis for a
qualitative comparison.For studies of dipole moments in thiol
SAMs, see the following: I. H. Campbell, J. D. Kress, R. L.
Martin, D. L. Smith, N. N. Barashkov and J. P. Ferraris, Appl.
Phys. Lett., 1997, 71, 3528; R. W. Zehner, B. F. Parsons, R. P.
Hsung and L. R. Sita, Langmuir, 1999, 15, 1121. For the above
calculations, the following software was used: M. J. Frisch, G. W.
Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheese-
man, J. A. Montgomery, Jr., T. Vreven, K. N. Kudin, J. C. Burant,
J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci,
M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji,
M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M.
Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X.
Li, J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C.
Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. Ochterski, P. Y. Ayala, K.
Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G.
Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas,
D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman,
There was practically no difference in surface molecular
coverage between RT-formed and thermally ordered SAMs of
PyC12SH. From this result, it could be inferred that the
disorder in the SAMs formed at RT should be ascribed to
kinetic factors rather than thermodynamic ones. However,
since it is possible that the disordered phase had a slightly
higher surface molecular density than that of the ordered
phase, the disordered state may be not only a kinetically
limited state, but also a thermodynamically favored one over
the ordered state at RT, in terms of total enthalpy of the self-
assembly system.
The main focus of this work is on the effect of the dipolar
pyrrolyl endgroups to the structural phases of the correspond-
ing thiol SAM on gold. Our study not only filled in the current
library of correlations between molecular species and their
SAM structures, but also presented a novel instance of dis-
order–order phase transformation in molecular SAMs. We
believe that this work can be beneficial for future studies to
ꢁc
This journal is the Owner Societies 2008
3148 | Phys. Chem. Chem. Phys., 2008, 10, 3138–3149

View Article Online
J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B.
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L.
Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A.
Nanayakkara, M. Challacombe, P. M. W. Gill, B. G. Johnson, W.
Chen, M. W. Wong, C. Gonzalez and J. A. Pople, GAUSSIAN 03
(Revision B.03), Gaussian, Inc., Wallingford, CT, 2004.
26 T. Ishida, W. Mizutani, N. Choi, U. Akiba, M. Fujihira and H.
Tokumoto, J. Phys. Chem. B, 2000, 104, 11680.
37 W. Wang, T. Lee and M. A. Reed, Phys. Rev. B: Condens. Matter
Mater. Phys., 2003, 68, 035416.
38 R. G. Nuzzo, F. A. Fusco and D. L. Allara, J. Am. Chem. Soc.,
1987, 109, 2358.
39 C. Zeng, B. Li, B. Wang, H. Wang, K. Wang, J. Yang, J. G. Hou
and Q. Zhu, J. Chem. Phys., 2002, 117, 851.
40 J. Nara, S. Higai, Y. Morikawa and T. Ohno, J. Chem. Phys.,
2004, 120, 6705.
27 H. H. Jung, Y. D. Won, S. Shin and K. Kim, Langmuir, 1999, 15,
1147.
41 L. Gontrani, F. Ramondo and R. Caminiti, Chem. Phys. Lett.,
2006, 417, 200.
28 O. Cavalleri, A. Hirstein and K. Kern, Surf. Sci., 1995, 340
L960.
29 J.-P. Bucher, L. Santesson and K. Kern, Langmuir, 1994, 10, 979;
J. Zhang, Q. Chi and J. Ulstrup, Langmuir, 2006, 22, 6203.
30 G. Yang, Y. Qian, C. Engtrakul, L. R. Sita and G. Liu, J. Phys.
Chem. B, 2000, 104, 9059.
31 It is known that autoxidation of N-alkylpyrrole produces carbo-
nyl-containing moieties such as succinimide. Since no carbonyl
peak was observed near 1710 cmꢃ1 in PIERS measurements of the
PyC12SH SAMs (Fig. 9), the possibility of pyrrolyl autoxidation
prior to the CV measurements could be excluded.
42 R.-F. Dou, X.-C. Ma, L. Xi, H. L. Yip, K. Y. Wong, W. M. Lau,
J.-F. Jia, Q.-K. Xue, W.-S. Yang, H. Ma and A. K.-Y. Jen,
Langmuir, 2006, 22, 3049.
43 Since the N atom in the pyrrolyl group is sp2-hybridized as is C1
atom in the phenyl group, the molecular conformation of an
isolated pyrrolyl-terminated alkanethiol is expected to be similar
to that of a phenyl-terminated alkanethiol molecule.
´
44 A. Gomez-Zavaglia and R. Fausto, J. Phys. Chem. A, 2004, 108,
6953.
45 R. Goddard, O. Heinemann and C. Kruger, Acta Crystallogr.,
¨
Sect. C: Cryst. Struct. Commun., 1997, 53, 1846.
32 C. A. Widrig, C. Chung and M. D. Porter, J. Electroanal. Chem.,
1991, 310, 335.
33 P. E. Laibinis, C. D. Bain and G. M. Whitesides, J. Phys. Chem.,
1991, 95, 7017.
46 The influence of dipole–dipole or quadrupolar interactions be-
tween terminal groups on molecular ordering has been well
demonstrated using a phenoxy-alkanethiol-based SAM system in
the following work: Y.-Y. Luk, N. L. Abbott, J. N. Crain and F. J.
Himpsel, J. Chem. Phys., 2004, 120, 10792.
47 The dipole moment of 1-methylpyrrole is B2.13 D from aromatic
ring to methyl group, while that of toluene is B0.36 D. These
values, obtained by density-functional calculations using B3PW91
with the 6-31G(d,p) basis set, were excerpted from the following:
NIST Computational Chemistry Comparison and Benchmark
Database, NIST Standard Reference Database Number 101
Release 12, Aug 2005, Editor: Russell D. Johnson III, http://
srdata.nist.gov/cccbdb.
48 Usage of other solvents, such as toluene and DMF, resulted in the
formation of small aggregates of the thiols physisorbed on the
formed SAMs, which caused difficulties in STM measurements.
49 Increasing the annealing time also tended to intensify the contam-
ination of the sample surface by the molecular aggregates, which
was found to emerge even after a short annealing time of 1 h as
shown in Fig. 5.
50 D. D. Popenoe, R. S. Deinhammer and M. D. Porter, Langmuir,
1992, 8, 2521; S.-K. Oh, L. A. Baker and R. M. Crooks, Langmuir,
2002, 18, 6981.
34 It has been reported that inelastic photoelectron scattering in
surface layers of organic compounds can be predicted from the
numbers of valence electrons, molecular densities, and bandgap
energies of the compounds (S. Tanuma, C. J. Powell and D. R.
Penn, Surf. Interface Anal., 1993, 21, 165). In this calculation, we
assumed that the difference in the photoelectron attenuation
between PyC12S and CnS is solely governed by the number of
valence electrons, nve (nve = 103 for PyC12S, and nve = (13 + 6n)
for CnS). For validity of this assumption, see the following: P.
Harder, M. Grunze, R. Dahint, G. M. Whitesides and P. E.
Laibinis, J. Phys. Chem. B, 1998, 102, 426.
35 Since the samples frequently contaminated the tip while being
scanned, the quality of STM image was constantly checked to
exclude erroneous results due to tip contamination. Because con-
taminated tips or local contaminations on the surface would give
noisier linear fits with less steep slopes or convexly curved fits, it
could be assumed that these data were obtained with a clean tip.
36 J. K. Gimzewski and R. Moller, Phys. Rev. B: Condens. Matter
¨
Mater. Phys., 1987, 36, 1284.
ꢁc
This journal is the Owner Societies 2008
Phys. Chem. Chem. Phys., 2008, 10, 3138–3149 | 3149