Effects of Alkyl Chain Length and Hydrogen Bonds on the Cooperative Self-Assembly of 2-Thienyl-Type Diarylethenes at a Liquid/Highly Oriented Pyrolytic Graphite (HOPG) Interface

Article abstract of DOI:10.1002/chem.201500707

An appropriate understanding of the process of self-assembly is of critical importance to tailor nanostructured order on 2D surfaces with functional molecules. Photochromic compounds are promising candidates for building blocks of advanced photoresponsive surfaces. To investigate the relationship between molecular structure and the mechanism of ordering formation, 2-thienyl-type diarylethenes with various lengths of alkyl side chains linked through an amide or ester group were synthesized. Their self-assemblies at a liquid/solid interface were investigated by scanning tunneling microscopy (STM). The concentration dependence of the surface coverage was analyzed by using a cooperative model for a 2D surface based on two characteristic parameters: the nucleation equilibrium constant (K_n) and the elongation equilibrium constant (K_e). The following conclusions can be drawn. 1)The concentration at which a stable 2D molecular ordering is observed by STM exponentially decreases with increasing length of the alkyl chain. 2)Compounds bearing amide groups have higher degrees of cooperativity in self-assembly on 2D surfaces (i.e., σ, which is defined as K_n/K_e) than compounds with ester groups. 3)The self-assembly process of the open-ring isomer of an ester derivative is close to isodesmic, whereas that of the closed-ring isomer is cooperative because of the difference in equilibrium constants for the nucleation step (i.e., K_n) between the two isomers. Getting along: A scanning tunneling microscopy (STM) investigation revealed a cooperative self-assembly process for photochromic 2-thienyl-type diarylethenes. The lengths of the alkyl chains, hydrogen bonds, and photoisomerization have a significant effect on the adsorption parameters at a liquid/highly oriented pyrolytic graphite (HOPG) interface (see figure).

Full text of DOI:10.1002/chem.201500707

DOI: 10.1002/chem.201500707
Full Paper
&
Surface Analysis |Hot Paper|
Effects of Alkyl Chain Length and Hydrogen Bonds on the
Cooperative Self-Assembly of 2-Thienyl-Type Diarylethenes at
a Liquid/Highly Oriented Pyrolytic Graphite (HOPG) Interface
[
a]
Soichi Yokoyama, Takashi Hirose, and Kenji Matsuda*
Abstract: An appropriate understanding of the process of
self-assembly is of critical importance to tailor nanostruc-
tured order on 2D surfaces with functional molecules. Photo-
chromic compounds are promising candidates for building
blocks of advanced photoresponsive surfaces. To investigate
the relationship between molecular structure and the mech-
anism of ordering formation, 2-thienyl-type diarylethenes
with various lengths of alkyl side chains linked through an
amide or ester group were synthesized. Their self-assemblies
at a liquid/solid interface were investigated by scanning tun-
neling microscopy (STM). The concentration dependence of
the surface coverage was analyzed by using a cooperative
model for a 2D surface based on two characteristic parame-
ters: the nucleation equilibrium constant (K ) and the elon-
n
gation equilibrium constant (K ). The following conclusions
e
can be drawn. 1) The concentration at which a stable 2D
molecular ordering is observed by STM exponentially de-
creases with increasing length of the alkyl chain. 2) Com-
pounds bearing amide groups have higher degrees of coop-
erativity in self-assembly on 2D surfaces (i.e., s, which is de-
fined as K /K ) than compounds with ester groups. 3) The
n
e
self-assembly process of the open-ring isomer of an ester
derivative is close to isodesmic, whereas that of the closed-
ring isomer is cooperative because of the difference in equi-
librium constants for the nucleation step (i.e., K ) between
n
the two isomers.
Introduction
ence in the concentration of the solution phase has attracted
considerable attention, since it provides important insights
into thermodynamic properties of surface-confined self-assem-
Scanning tunneling microscopy (STM) is a powerful tool to elu-
cidate spatial alignment, orientation, and conformation of indi-
vidual molecules on a surface because of its outstanding reso-
lution at the sub-nanometer scale, even under ambient condi-
[
41–45]
blies.
De Feyter et al. have rationally interpreted an origin
of polymorphs by considering the surface density of two differ-
[
46]
ent phases and the adsorption energies per molecule. Al-
though the concentration dependence of such phase transi-
tions between two or more different orderings has been well
documented, much less is known about the mechanism that
underlies ordering formation that occurs on an unoccupied
surface with increasing concentration of solute.
[
1–8]
tions.
Two-dimensional crystal engineering at liquid/solid in-
[
9–15]
terfaces has attracted growing attention in the last decade.
To efficiently design and predict the formation of thermody-
namically stable molecular ordering formed through self-as-
sembly processes, interplays of various interactions, such as
molecule–substrate interactions, intermolecular interac-
Stimuli responsiveness is a key factor to develop sophisticat-
[
16–20]
[21–25]
[1,26,29,47–52]
tion,
entropy terms,
tems employing these interactions, such as chiral assem-
solvent effects,
and the balance of enthalpy and
ed molecular nanodevices.
For precise control over
[
26–29]
must be taken into account. Various sys-
the dynamics of stimuli-responsive nanostructures, an insight
[
53]
into the self-assembly process is of particular significance.
[
21,30–33]
[34–37]
blies,
post-assembly manipulation,
and porous net-
have recently been intensively investigated by
many groups.
We previously reported the photoinduced formation and dis-
appearance of 2D molecular ordering composed of a photo-
[
2,17,38]
works,
[
54]
chromic 2-thienyl-type diarylethene derivative.
A dynamic
Mechanism of ordering formation has been investigated
from various aspects. One important aspect is to study kinetic
and thermodynamic factors in the formation of 2D order-
photoresponsive surface that is highly stimuli-responsive can
be realized by the precise design of a cooperative self-assem-
bly process. Herein, we have investigated the relationship be-
tween the chemical structure and adsorption parameters for
ordering formation by using STM under ambient conditions.
The observed results provide a fundamental insight into the
mechanism of the cooperative self-assembly process on a 2D
surface. We also discuss the effects of 1) the length of the alkyl
chain, 2) the hydrogen bond of the amide functional group,
and 3) photoisomerization of the core moiety on various ad-
sorption parameters.
[
39,40]
ings.
Polymorphs on 2D surfaces that arise from a differ-
[
a] S. Yokoyama, Dr. T. Hirose, Prof. K. Matsuda
Department of Synthetic Chemistry and Biological Chemistry
Graduate School of Engineering, Kyoto University
Katsura, Nishikyo-ku, Kyoto 615-8510 (Japan)
E-mail: kmatsuda@sbchem.kyoto-u.ac.jp
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201500707.
Chem. Eur. J. 2015, 21, 13569 – 13576
13569
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Results and Discussion
Kn, followed by subsequent stepwise elongation processes de-
scribed by the elongation constant Ke.
At the nucleation step in Equation (1), a molecule in a super-
Molecular design
sol
natant solution, A , adsorbs on the substrate to make a nu-
sub
To investigate the effect of alkyl chain length on the self-as-
sembly process on a highly oriented pyrolytic graphite (HOPG)
surface, a series of amide derivatives 1o-Cn (n=12, 14, 16, 18,
cleus on the substrate, A1 . The temporal differentiation of
sub
[A1 ] is given by Equation (2):
2
0, 22), with alkyl chains (ÀC H2n+1) of different lengths at
n
either end, were systematically designed and synthesized
(
Figure 1).
1,2-Bis(3-methyl-2-thienyl)hexafluorocyclopentene
was chosen as the photochromic core moiety because of its
ease of synthesis and sufficient photostability. An ethylene
spacer was introduced between the photochromic core and
the side chains to provide flexibility to the amide functional
groups, which are involved in the formation of an intermolecu-
lar hydrogen-bond network. To investigate the effect of amide
groups on the ordering process, ester derivatives 2o-Cn (n=
sub
d½A1
Š
Asub
knð1 À qÞ S ½A Š À kÀn½A1 ŠLNA
sol
sub
ð2Þ
¼
dt
in which k and k are nucleation rate constants of adsorption
n
Àn
À1 À1
and desorption, respectively, the units of which are m
s
and
À1
s , respectively; q is fractional coverage of the surface area;
Asub is the total area of the substrate; S is the unit area occu-
pied by one molecule of adsorbate A; L is the volume of the
supernatant solution; and NA is the Avogadro constant. The
first and second terms on the right-hand side of Equation (2)
represent the formation and desorption of nuclei, respectively.
1
6, 22) with hexadecyl and docosyl groups were also synthe-
sized.
sub
By assuming a steady-state approximation, [A1 ] is given by
Equation (3):
ð1 À qÞ A
sub
sub
sol
½
A1 Š ¼
K ½A Š
ð3Þ
n
LN
S
A
in which the nucleation equilibrium constant, K , is defined as
n
the ratio of k to k (k /k ).
n
Àn
n
Àn
In a similar manner to the nucleation process, the molar
concentration of N-mer domains on the substrate formed by
subsequent elongation processes is given by Equation (4):
Figure 1. Molecular structure of 2-thienyl-type diarylethenes featuring an
amide group (1o-Cn) and an ester group (2o-Cn).
ð1 À qÞ A
sub
sub
sol
N
½
AN Š ¼
sðK ½A ŠÞ
ð4Þ
e
LN
S
A
Cooperative model for self-assembly on a 2D surface
in which s is the degree of cooperativity, which is defined by
the ratio of two equilibrium constants (K /K ). Notably, Equa-
Inspired by the cooperative model for self-assembly in solu-
n
e
[
55–58]
tion,
we recently developed a cooperative equilibrium
tion (3) can be obtained from Equation (4) by substituting N
with one. The total molar concentration of component A on
the substrate is given by Equation (5):
[54]
model for 2D surfaces. In our model study, we postulated
that our system was thermodynamically controlled, that is, ad-
sorption/desorption equilibrium was achieved on the timescale
of the STM measurements, as determined from the following
experimental results: 1) molecular ordering of the synthesized
compounds (1o-Cn and 2o-Cn) can be observed shortly after
dropping the solution on the HOPG substrate (within 5 min);
sub
sub
sub
sub
sub
ct ¼½A1 Š þ 2½A2 Š þ 3½A3 Š þ ::: þ N½AN Š þ :::
1
X
ð1 À qÞ A
sub
sol
i
¼
i
sðK ½A ŠÞ
e
LN
S
ð5Þ
A
i¼1
sol
2
) reorientation and size changes of domains were observed
ð1 À qÞ A
sK ½A Š
sub
e
¼
during in situ STM measurements (Figure S2 in the Supporting
Information); and 3) the average value of surface coverage did
not change, even after standing for several hours (Table S1 in
sol
2
LN
S ð1 À K ½A ŠÞ
A
e
[59]
sol
Supporting Information).
in which K ½A Š<1 is assumed to approximate the form of in-
e
To reproduce an experimentally observed concentration de-
pendence of surface coverage, we assumed the presence of
two processes with different equilibrium constants: initial nu-
cleation on a substrate described by the nucleation constant
finite series.
Because the fractional coverage, q, is the ratio of the area
occupied by component A to the total area of substrate, q is
given by Equation (6):
Chem. Eur. J. 2015, 21, 13569 – 13576
www.chemeurj.org
13570
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
sub
ct LNAS
q ¼
ð6Þ
Asub
The total concentration of A in the system is given by Equa-
tion (7):
sol
sub
t
c ¼ ½A Š þ c
ð7Þ
t
Notably, concentration changes in the supernatant solution
caused by adsorption/desorption events are taken into ac-
count by introducing the mass balance [Eq. (7)] and volume of
[60]
the supernatant solution, L [Eq. (2)] into this model. Simulta-
neous Equations (5), (6), and (7) give Equation (8):
Figure 2. High-resolution STM images and molecular models of the ordering
of a) 1o-C14, and b) 1o-C20 at the octanoic acid/HOPG interface (c=2 mm
and 10 mm for 1o-C14 and 1o-C20, respectively; Iset =30 pA; Vbias =800 mV).
The orientations of the main symmetry axis of the underlying graphite are
indicated by white arrows. The directions of the row of bright spots and
sK ðc À aqÞ
e
t
q ¼ ð1 À qÞ
ð8Þ
2
f1 À K ðc À aqÞg
e
t
alkyl chains are almost parallel to the main symmetry axis of graphite (i.e.,
the <11 2¯ 0> direction). Enlarged molecular models for: c) 1o-C14, and
d) 1o-C20. The hydrogen-bond network of the amide groups is highlighted
in which a is given by Equation (9):
in green.
Asub
a ¼
ð9Þ
LNAS
Lattice parameters of the ordering of 1o-C14-C22 obtained
from the STM measurements are summarized in Table 1. The
length of the a axis increased by about 0.4 nm with an in-
crease in the alkyl chain length of 2 carbon atoms, whereas
the length of the b axis and the angle a were independent of
the alkyl chain length. This tendency is consistent with preser-
vation of the ordering pattern for the amide derivatives 1o-Cn.
The concentration dependence of surface coverage provides
a valuable insight into the mechanism of ordering formation
on 2D surfaces. At first, we confirmed that the evaporation of
octanoic acid on the HOPG substrate was less than 1% after
Since Asub, L, and S can be determined from experimental
conditions, a was treated as a constant in the model simula-
tion.
By using Equation (8), we could successfully reproduce the
experimentally observed concentration dependence of surface
coverage for compound 1o-C16, which has two hexadecyl
[49]
chains linked by an amide group.
This compound shows
a high degree of cooperativity (s) and a large K value in the
e
self-assembly process at the octanoic acid/HOPG interface;
however, the relationship between molecular structure and
these adsorption parameters (i.e., K and s) has not been in-
e
5
h under ambient conditions; therefore, the concentration
vestigated and remains unclear.
change by evaporation of the solvent is negligible during STM
measurements (Figure S4 in the Supporting Information). Then,
to determine the average surface coverage for the ordering of
Dependence of adsorption parameters on alkyl chain length
1
o-C14–C22 at various concentrations, we collected more
Molecular ordering of the open-ring isomer of the amide deriv-
atives 1o-Cn was observed at the octanoic acid/HOPG inter-
face. Regardless of the alkyl chain length, the ordering patterns
than 10 STM images at each concentration (Figure 3a and Fig-
ures S5–S39 in the Supporting Information). The critical con-
centration, defined as the concentration at which the adsorp-
tion saturated, increased by approximately an order of magni-
[61]
of the amide derivatives 1o-Cn were similar. High-resolution
[
62]
STM images of 1o-C14 and 1o-C20 are shown in Figure 2.
The molecular ordering pattern of 1o-C14–C22 was consistent
with that previously determined for 1o-C16 by using a molecu-
lar mechanics/molecular dynamics (MM/MD) model. The pat-
terns show that the open-ring isomer is folded into the parallel
conformer and the conformers are arranged in a head-to-head
orientation on the substrate to form a striped lamellar arrange-
ment comprising twofold networks of alternate intra- and in-
Table 1. Lattice parameters of the 2D ordering obtained from STM im-
ages.
[a]
2
[b]
a [nm]
b [nm]
a [8]
S [nm ]
1o-C14
6.0Æ0.1
6.3Æ0.2
6.8Æ0.1
7.2Æ0.1
7.6Æ0.1
0.99Æ0.04
1.04Æ0.04
0.98Æ0.02
0.98Æ0.03
1.00Æ0.01
88Æ2
3.0
3.3
3.3
3.5
3.8
1
1
1
o-C16
o-C18
o-C20
89Æ1
88Æ1
[54]
termolecular hydrogen bonding.
Because the open-ring
87Æ2
isomer adopts the photoinactive parallel conformation, no
photoisomerization is considered to proceed at a liquid/solid
1o-C22
89.4Æ0.2
[a] At the octanoic acid/HOPG interface. Determination of the lattice pa-
[
63]
interface.
A very similar striped ordering pattern was ob-
rameters for compound 1o-C12 was difficult due to the instability of or-
dering under STM measurement conditions (Figure S3 in the Supporting
Information). [b] The unit area occupied by one molecule on the sub-
strate determined from the lattice parameters.
served when phenyloctane was used as a solvent. This result is
reasonable because no solvent molecule coadsorbs on HOPG
substrate, according to the MM/MD simulation.
Chem. Eur. J. 2015, 21, 13569 – 13576
www.chemeurj.org
13571
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tude as the alkyl chain length increased by two carbon atoms.
This is indicative of a significant effect of dispersion interac-
tions between the alkyl chain and HOPG substrate on the asso-
ciation constant of the ordering formation process for com-
pounds 1o-C14–C22. This indicates that, depending on the
molecular structure, the appropriate concentration for the STM
observation of a stable self-assembled monolayer (SAM) on
HOPG substrates, can differ by several orders of magnitude,
and that the length of the alkyl chain is a crucial factor in opti-
mizing the affinity for the HOPG surface.
Table 2. Adsorption parameters of compounds 1o-C14–C22 with alkyl
groups of various chain lengths.
[a]
À1
À1
K
n
[m
]
K
e
[m
]
s
À1
À4
À2
1o-C14
80Æ120
520Æ150
(2Æ2)”10
(5Æ2)”10
(3Æ2)”10
[
b]
1
1
1
1
o-C16
o-C18
o-C20
o-C22
2.1Æ0.9
4600Æ20
3
4
(1.2Æ0.8) ”10
(3.6Æ0.2)”10
(5.7Æ1.7)”10
[c]
5[c]
À6[c]
<0.6
ꢀ10
[d]
[d]
[d]
full adsorption
full adsorption
full adsorption
[a] At the octanoic acid/HOPG interface. The margin of error denotes
a 95% confidence interval for probability distribution parameters ob-
tained by nonlinear regression analysis. [b] Adsorption parameters for
The experimentally observed concentration dependence of
surface coverage was analyzed by using the cooperative equi-
e
1o-C16 were reanalyzed in this work and K and s were corrected from
À1
À4 [64]
[54]
previously reported values (K =(4600Æ20)m , s=(2.8Æ1.1)”10 ).
e
librium model for 2D surfaces [Eq. (8)].
The experimental
À6
À6
[
c] s was fixed at 10 . When s is below 10 , the simulated curves are no
data for 1o-C14–C20 were successfully reproduced by model
simulations (Figure 3a). The resulting adsorption parameters
are summarized in Table 2. The observed surface coverage for
compound 1o-C22, with docosyl chains, was near the theoreti-
cal maximum, which means that almost all molecules available
in the supernatant solution were used in the formation of mo-
lecular ordering on the HOPG surface (Figure S39 in the Sup-
porting Information).
longer significantly dependent on s. Therefore, a s value of less than
10 cannot be determined precisely (Figure S32 in the Supporting Infor-
À6
mation). [d] The simulated curve also became less dependent on s when
6
the surface coverage was near the theoretical maximum (K
e
ꢁ10 ).
pendence on the critical concentration, 2) K and s reflect the
n
steepness of the concentration dependence of surface cover-
age with change in concentration, and most importantly 3) the
critical concentration in simulated curves is not largely affected
In terms of the precision of the calculated adsorption param-
eters shown in Table 2, K has a smaller confidence interval
e
than those of K and s. This is attributed to the following in-
by the value of K or s (Figure S32 in the Supporting Informa-
n
n
trinsic features of the model simulation: 1) K has a strong de-
tion). Just below the critical concentration (0<q<1), a large
dispersion was usually observed in the fractional coverage in
the STM measurements for different areas of the substrate. The
large dispersion inevitably gives rise to a large error in the
averaged fractional coverage in this concentration region,
e
which leads to a large confidence interval for K and s. Conse-
n
quently, K can be determined more precisely than K or s by
e
n
using this model simulation.
The natural logarithm of K was linearly dependent on the
e
alkyl chain length (Figure 3b). The relationship can be rational-
ized in terms of the Gibbs free energy of adsorption, DG , in
e
accordance with Equation (10):
DG ¼ ÀRT ln K
ð10Þ
e
e
in which R is the gas constant and T is the temperature. In our
model, Kn denotes the equilibrium between the free mono-
mers in solution and the nuclei of molecular orderings on the
surface, whereas K denotes the equilibrium between the free
e
monomers in solution and the molecule incorporated within
molecular ordering already present on the 2D surface. There-
fore, DG should reflect the energy gained by the incorpora-
e
tion of free monomers in solution within existing molecular or-
derings on a HOPG substrate. The plot of experimentally ob-
served DG values versus the number of carbon atoms in the
e
Figure 3. a) Concentration dependence of surface coverage of the amide de-
rivatives 1o-C14 (*
), 1o-C16 (
), 1o-C18 (~), 1o-C20 (!), and 1o-C22 (^)
&
alkyl chain is shown in Figure S40 in the Supporting Informa-
tion. Linear regression gives Equation (11):
at the octanoic acid/HOPG interface. Solid lines denote the curves of best fit
obtained from the cooperative equilibrium model. The dashed line denotes
the curve based on a theoretical maximum for a monolayer. The error bars
denote standard deviations obtained from a histogram consisting of more
than ten experimental STM images (see the Supporting Information). b) The
DG ¼ À2:76n þ 23:3½kJmolÀ1Š
ð11Þ
e
plot of the logarithm of K
alkyl side chain. The solid line denotes the best-fitting line obtained by
a linear regression (lnK
e
against the number of carbon atoms in each
in which n is the number of carbon atoms in the alkyl chain.
Because the diarylethene derivatives 1o-Cn have two alkyl side
chains, the total number of carbon atoms increased by two
e
=1.12nÀ9.4, in which n is the number of carbon
atoms in the alkyl group).
Chem. Eur. J. 2015, 21, 13569 – 13576
www.chemeurj.org
13572
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
when n was increased by one. Therefore, Equation (11) means
The difference in the ordering patterns may be due to the
absence of hydrogen-bond networks in the ordering of 2o-
C22. It is noted that a bright spot corresponding to the diaryl-
ethene moiety has two alkyl chains extending in the same di-
rection in Figure 4b, which suggests that the open-ring isomer
of the ester derivative 2o-C22 also adopts a parallel conforma-
tion in its molecular ordering on the HOPG surface.
that the stabilization free energy for the elongation of molecu-
À1
lar ordering increases by 2.8 kJmol as the length of the alkyl
chains increases by two carbon atoms for a molecular adsor-
bate. By determining the slope of DG , depending on chemical
e
structures for various systems, the relationship between molec-
ular structure and critical concentration at which surface cover-
age saturates can be quantitatively evaluated. A similar effect
is commonly exploited in supramolecular systems in which
alkyl chains interact with each other, such as the critical micelle
The effect of an amide group on ordering formation at
a liquid/solid interface became clear upon comparing the con-
centration dependence of surface coverage for 1o-C22 and
2o-C22 (Figure 5). The critical concentration of 1o-C22 was
about 5 mm, which was near the theoretical maximum, whereas
that of 2o-C22 was 25 times higher at 125 mm. Moreover, the
curve of the concentration dependence of 2o-C22 was less
steep, which was indicative of a nearly isodesmic process for
ordering formation. The observed adsorption parameters for
[65–67]
concentration of surfactants
and hydrophobic interaction
[
68,69]
chromatography.
Effect of amide groups on adsorption parameters
3
À1
3
À1
2
o-C22 were determined to be K =3”10 m , K =9”10 m ,
n e
As described above, increasing the alkyl chain length is a good
approach to stabilize the molecular ordering on the HOPG sur-
face. We have previously reported that the ester derivative 2o-
C16, with a hexadecyl chain, does not form any 2D ordering
and s=0.3. The isodesmic nature of 2o-C22 contrasts sharply
with the cooperative nature of 1o-C14–C22 shown in Figure 3.
[70]
pattern at the octanoic acid/HOPG interface.
a stable ordering pattern was observed for the ester derivative
o-C22, with a docosyl chain. Interestingly, although both 1o-
In contrast,
2
C22 and 2o-C22 contain alkyl chains of equal length, we ob-
served a striking difference between their ordering patterns in
the high-resolution STM images shown in Figure 4. A striped
ordering pattern was observed for the amide derivative 1o-
C22 (Figure 4a), whereas an oblique primitive lattice com-
prised of a pair of bright spots was observed for the ester de-
rivative 2o-C22 (Figure 4b).
Figure 5. Concentration dependence of surface coverage of amide derivative
1
o-C22 (^) and ester derivative 2o-C22 (*) at the octanoic acid/HOPG inter-
face. The solid line denotes a simulated curve of best fit obtained by the
3
À1
3
À1
model simulation (2o-C22, K
s=0.3Æ0.3).
n
=(3Æ3)”10 m , K
e
=(9Æ2)”10 m
,
Effect of the core structure on adsorption parameters
Photoisomerization enables us to have precise control over the
concentrations of molecular components. The concentration
dependence of the formation of molecular ordering is particu-
larly important in photoresponsive 2D self-assembly systems,
wherein the concentrations of photochromic components
change upon photoisomerization. Diarylethene derivatives 1-
Cn and 2-Cn undergo photoisomerization between the open-
and closed-ring isomers upon photoirradiation with UV and
visible light (Figure 6a). We previously reported that no order-
ing formation was observed for the closed-ring isomer 1c-C16,
which was the principle behind the phototriggered formation
and disappearance of molecular ordering upon photoisomeri-
Figure 4. High-resolution STM images and molecular models of the 2D or-
dering of: a) amide derivative 1o-C22 (a=(7.6Æ0.1) nm,
[54]
zation.
b=(1.00Æ0.01) nm, a=(89.4Æ0.2)8), and b) ester derivative 2o-C22
We found that molecular ordering was observed for the
ester derivative 2c-C22, which prompted us to investigate the
difference in the adsorption parameters between the two iso-
mers. The closed-ring isomer 2c-C22 was isolated by means of
HPLC from a UV-irradiated solution of 2o-C22, and its molecu-
lar ordering on the HOPG surface was carefully investigated.
(
a=(3.6Æ0.1) nm, b=(2.1Æ0.1) nm, a=(78Æ3)8) at the octanoic acid/
HOPG interface (c=20 and 250 mm for 1o-C22 and 2o-C22, respectively;
set =30 pA; Vbias =800 V). The orientations of the main symmetry axis of un-
I
derlying graphite are indicated by white arrows. Enlarged molecular models
of: c) 1o-C22, and d) 2o-C22. The hydrogen-bond network of the amide
groups is highlighted in green in the ordering of 1o-C22. No hydrogen-
bond network was observed in the ordering of 2o-C22.
Chem. Eur. J. 2015, 21, 13569 – 13576
www.chemeurj.org
13573
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 3. Adsorption parameters of 2-C22 at the octanoic acid/HOPG in-
terface.
À1
À1
K
n
[m
]
K
e
[m
]
s
3
3
2
2
o-C22 (open)
c-C22 (closed)
(3Æ3)”10
2Æ40
(9Æ2)”10
0.3Æ0.3
4
À3
(1.27Æ0.04)”10
(2Æ3)”10
C22 should be attributed to the difference in the energy
gained during the nucleation process; judging from the small
À1
value of K (2Æ40m ) for the closed-ring isomer, the stabiliza-
n
tion energy for the formation of nuclei during ordering forma-
tion on HOPG is significantly smaller for the closed-ring isomer
2
c-C22.
Figure 6. a) Scheme of the photochromic reaction for diarylethene 2o-C22
with an ester group. b) A high-resolution STM image of 2D ordering of
closed-ring isomer 2c-C22 at the octanoic acid/HOPG interface (c=125 mm,
Conclusion
I
t
=10 pA, Vbias =À1.0 V). The lattice parameters are a=(4.0Æ0.1),
b=(1.82Æ0.04) nm, and a=(84Æ3)8. The orientations of the main symme-
In this study, we investigated the relationship between the mo-
lecular structure of 2-thienyl-type diarylethenes and coopera-
tive adsorption parameters of these compounds. As the length
try axis of underlying graphite are indicated by white arrows. For closed-ring
isomer 2c-C22, the direction of extended alkyl chains is almost parallel to
the main symmetry axis of graphite (i.e., the <11 2¯ 0> direction), whereas
¯
0> direction.
of the alkyl chain of diarylethene increased, the K value in-
e
that of the row of bright spots is aligned along the <101
c) Surface coverage as a function of concentration for 2o-C22 (*) and 2c-
C22 ( ). The solid lines denote simulated curves of best fit obtained by the
creased exponentially, although the pattern of molecular or-
dering did not change. This behavior resulted from enthalpy
gain with increasing alkyl chain length, which was associated
with the concentration dependence of micelle formation.
When amide groups at the root of the alkyl chain were re-
placed by ester groups, the cooperative parameter, s, ap-
proached unity and the mechanism resembled an isodesmic
process. This was likely to be due to the absence of hydrogen-
bonding interactions for the ester derivative.
&
model simulation.
The closed-ring isomer of the ester derivative was highly solu-
ble in octanoic acid and suitable for STM measurements at var-
ious concentrations. A high-resolution STM image of 2c-C22 is
shown in Figure 6.
Lattice parameters of the ordering for 2c-C22 were deter-
mined to be a=(4.0Æ0.1) nm, b=(1.82Æ0.04) nm, and a=
In terms of the effect of core structure for the ester deriva-
tives of 2o-C22 and 2c-C22, the s value of the closed-ring
isomer was significantly smaller than that of the open-ring
(
84Æ3)8. Based on the length of the b axis, the center-to-
center distance between adjacent diarylethene cores was de-
termined to be 0.91 nm, which was significantly larger than
the typical p–p distance (0.34 nm) and suggested a face-on ar-
e
isomer, but the K values were similar. This suggests that Ke
values of diarylethenes 2o-C22 and 2c-C22 are regulated by
dispersion interactions between the alkyl chain and HOPG sub-
[70]
rangement of 2c-C22. The length of the a axis (4.0 nm), the
direction of which is almost parallel to that of the alkyl chain,
corresponds to the sum of a docosyl chain (ꢂ2.8 nm) and the
long axis of the diarylethene core moiety (ꢂ1.3 nm). The high-
resolution STM image of 2c-C22 was well reproduced by
a MM/MD model; the closed-ring isomer 2c-C22 extends two
alkyl chains in opposite directions and the alkyl chains of adja-
cent rows are interdigitated with each other.
strate, irrespective of the core structure; on the other hand, Kn
was significantly affected by the core structure. The energy
gain for nucleation on substrates for the closed-ring isomer
2
c-C22 was lower than that of the open-ring isomer 2o-C22,
which gave rise to a large difference in cooperative self-assem-
bly on the HOPG surface.
These findings enable the precise control of cooperativity in
self-assembly on 2D surfaces by careful design of the molecu-
lar structure. To clarify the contribution of solvent and thermo-
dynamic parameters (i.e., enthalpy or entropy) toward the pro-
cess of self-assembly at the liquid/solid interface (i.e., K , K ,
A plot of surface coverage of 2c-C22 against concentration
is shown in Figure 6c. Interestingly, the coverage of 2c-C22
shows a steeper change around 80 mm than that of 2o-C22,
which suggests that the cooperativity in ordering formation
for the closed-ring isomer 2c-C22 is higher than that of the
open-ring isomer 2o-C22. The adsorption parameters for 2c-
n
e
and s), we are currently investigating the role of the solvent
and temperature dependence of these parameters.
C22, simulated by our cooperative model, were K =1.27”
e
4
À1
À3
1
0 m and s=2”10 (Table 3). Interestingly, the K values
e
were similar for the open- and the closed-ring isomers, which
suggested that the energy gain for the elongation of molecu-
lar ordering was independent of the core structure of the two
photochemical isomers. Therefore, the difference in the coop-
erativity of 2D ordering formation between 2o-C22 and 2c-
Experimental Section
General
Detailed syntheses of the compounds reported here are given in
the Supporting Information. The final products were purified by
Chem. Eur. J. 2015, 21, 13569 – 13576
www.chemeurj.org
13574
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
preparative gel permeation chromatography (GPC) and character-
ized by NMR and IR spectroscopy and mass spectrometry.
[3] S. De Feyter, F. C. De Schryver, J. Phys. Chem. B 2005, 109, 4290–4302.
[
[
4] S. De Feyter, F. C. De Schryver, Chem. Soc. Rev. 2003, 32, 139–150.
5] P. Samorí, A. Fechtenkçtter, F. Jäckel, T. Bçhme, K. Müllen, J. P. Rabe, J.
Am. Chem. Soc. 2001, 123, 11462–11467.
STM measurements
[
[
[
[
6] L. C. Giancarlo, G. W. Flynn, Acc. Chem. Res. 2000, 33, 491–501.
7] J. K. Gimzewski, C. Joachim, Science 1999, 283, 1683–1688.
8] J. P. Rabe, S. Buchholz, Science 1991, 253, 424–427.
All STM experiments were performed at room temperature under
ambient conditions. The STM images were acquired with a PicoSPM
instrument (Molecular Imaging Co.) or an Agilent technologies
9] C.-A. Palma, M. Cecchini, P. Samorì, Chem. Soc. Rev. 2012, 41, 3713–
3730.
5
500 scanning probe microscope in the constant-current mode.
[10] Y. Xue, M. B. Zimmt, J. Am. Chem. Soc. 2012, 134, 4513–4516.
The STM tips used in this research were mechanically cut from
a Pt/Ir wire (80/20, diameter 0.25 mm). HOPG (purchased from
Bruker Co.) was used as the substrate. Homogeneous solutions of
[11] A. Ciesielski, C.-A. Palma, M. Bonini, P. Samorì, Adv. Mater. 2010, 22,
3506–3520.
[12] J. Adisoejoso, K. Tahara, S. Okuhata, S. Lei, Y. Tobe, S. De Feyter, Angew.
Chem. Int. Ed. 2009, 48, 7353–7357; Angew. Chem. 2009, 121, 7489–
1
o-Cn and 2o-Cn in octanoic acid were prepared by heating the
7493.
solutions or ultrasonication before STM measurements were re-
corded. A drop of the solution (10 mL) was deposited onto a freshly
cleaved HOPG surface, and the tip was immersed into the solution
and then the image was scanned. STM images were analyzed by
using the graphite substrate as a calibration grid. The total STM
scans were performed 10–30 times to determine an averaged sur-
face coverage at a concentration of the sample solution.
[
[
[
13] A. Llanes-Pallas, C.-A. Palma, L. Piot, A. Belbakra, A. Listorti, M. Prato, P.
Samorì, N. Armaroli, D. Bonifazi, J. Am. Chem. Soc. 2009, 131, 509–520.
14] J. A. A. W. Elemans, S. Lei, S. De Feyter, Angew. Chem. Int. Ed. 2009, 48,
7298–7332; Angew. Chem. 2009, 121, 7434–7469.
15] K. G. Nath, O. Ivasenko, J. M. MacLeod, J. A. Miwa, J. D. Wuest, A. Nanci,
D. F. Perepichka, F. Rosei, J. Phys. Chem. C 2007, 111, 16996–17007.
[16] K. Tahara, K. Katayama, M. O. Blunt, K. Iritani, S. De Feyter, Y. Tobe, ACS
Nano 2014, 8, 8683–8694.
[
[
[
17] X. Zhang, Q. Zeng, C. Wang, RSC Adv. 2013, 3, 11351–11366.
18] L. Xu, X. Miao, X. Ying, W. Deng, J. Phys. Chem. C 2012, 116, 1061–1069.
19] L. E. Heller, J. Whitleigh, D. F. Roth, E. M. Oherlein, F. R. Lucci, K. J. Kolon-
ko, K. E. Plass, Langmuir 2012, 28, 14855–14859.
Molecular modeling
The molecular ordering adsorbed on the HOPG surface was mod-
eled by a MM/MD approach by using the Materials Studio v6.1.0
program from Accelrys Software Inc. The Dreiding force field im-
plemented in the Forcite module was used for MM and MD calcu-
lations. The initial geometries were inspired from experimentally
observed high-resolution STM images for each ordering. For the
HOPG substrate, only one layer of graphene sheet (CÀC bond
length of 1.42 Š, flat geometry with hexagonal symmetry) was as-
sumed. The Cartesian position of the graphene sheet was fixed
during MM/MD calculations to suppress deformation/distortion of
the substrate.
[
[
20] D. M. Cyr, B. Venkataraman, G. W. Flynn, Chem. Mater. 1996, 8, 1600–
1615.
21] I. Destoop, E. Ghijsens, K. Katayama, K. Tahara, K. S. Mali, Y. Tobe, S. De
Feyter, J. Am. Chem. Soc. 2012, 134, 19568–19571.
[22] J. Adisoejoso, K. Tahara, S. Lei, P. Szabelski, W. R z˙ ysko, K. Inukai, M. O.
Blunt, Y. Tobe, S. De Feyter, ACS Nano 2012, 6, 897–903.
[23] L. Xu, X. Miao, B. Zha, W. Deng, J. Phys. Chem. C 2012, 116, 16014–
1
6022.
24] T. Takami, U. Mazur, K. W. Hipps, J. Phys. Chem. C 2009, 113, 17479–
7483.
[
1
[
[
25] Y. Yang, C. Wang, Curr. Opin. Colloid Interface Sci. 2009, 14, 135–147.
26] M. O. Blunt, J. Adisoejoso, K. Tahara, K. Katayama, M. Van der Auweraer,
Y. Tobe, S. De Feyter, J. Am. Chem. Soc. 2013, 135, 12068–12075.
UV/Vis spectroscopy and photochemical reactions
[
27] W. Song, N. Martsinovich, W. M. Heckl, M. Lackinger, J. Am. Chem. Soc.
Absorption spectra were measured on
a spectrophotometer
2013, 135, 14854–14862.
(
JASCO V670 spectrophotometer). The optical lengths of the
[28] Y. Miyake, T. Nagata, H. Tanaka, M. Yamazaki, M. Ohta, R. Kokawa, T.
Ogawa, ACS Nano 2012, 6, 3876–3887.
[29] R. Gutzler, T. Sirtl, J. F. Dienstmaier, K. Mahata, W. M. Heckl, M. Schmittel,
M. Lackinger, J. Am. Chem. Soc. 2010, 132, 5084–5090.
quartz cells were 10 or 5.0 mm. Conversion-ratio-determined solu-
tions of 2-C22 were prepared by using a USHIO 500 W super-high-
pressure mercury lamp with a combination of a heat absorbing
filter (HA30), sharp cut filters (U-29 and UV-31), and a band-pass
filter (UV-D35).
[30] Z. Guo, I. De Cat, B. Van Averbeke, E. Ghijsens, J. Lin, H. Xu, G. Wang,
ˇ
´
F. J. M. Hoeben, Z. Tomovic, R. Lazzaroni, D. Beljonne, E. W. Meijer,
A. P. H. J. Schenning, S. De Feyter, J. Am. Chem. Soc. 2013, 135, 9811–
9819.
[
[
[
31] I. Destoop, H. Xu, C. Oliveras-Gonzµlez, E. Ghijsens, D. B. Amabilino, S.
De Feyter, Chem. Commun. 2013, 49, 7477–7479.
32] N. Katsonis, E. Lacaze, B. L. Feringa, J. Mater. Chem. 2008, 18, 2065–
2073.
33] S. De Feyter, P. C. M. Grim, M. Rücker, P. Vanoppen, C. Meiners, M. Sief-
fert, S. Valiyaveettil, K. Müllen, F. C. De Schryver, Angew. Chem. Int. Ed.
Acknowledgements
This research was supported by the Funding Program for Next
Generation World-Leading Researchers (NEXT program, no.
GR062), a Grant-in-Aid for Scientific Research on Innovative
Areas “Photosynergetics” (no. 26107008) from the Ministry of
Education, Culture, Sports, Science and Technology (MEXT),
and a Grant-in-Aid for Young Scientists (B) (no. 25810048) from
the Japan Society for the Promotion of Science (JSPS), Japan.
1998, 37, 1223–1226; Angew. Chem. 1998, 110, 1281–1284.
[
[
[
[
34] M. Li, D. den Boer, P. Iavicoli, J. Adisoejoso, H. Uji-i, M. Van der Auweraer,
D. B. Amabilino, J. A. A. W. Elemans, S. De Feyter, J. Am. Chem. Soc.
2014, 136, 17418–17421.
35] D. den Boer, M. Li, T. Habets, P. Iavicoli, A. E. Rowan, R. J. M. Nolte, S.
Speller, D. B. Amabilino, S. De Feyter, J. A. A. W. Elemans, Nat. Chem.
2013, 5, 621–627.
36] A. Miura, S. De Feyter, M. M. S. Abdel-Mottaleb, A. Gesqui›re, P. C. M.
Grim, G. Moessner, M. Sieffert, M. Klapper, K. Müllen, F. C. De Schryver,
Langmuir 2003, 19, 6474–6482.
Keywords: cooperative effects
tunneling microscopy · self-assembly · thermodynamics
· monolayers · scanning
37] Y. Okawa, M. Aono, Nature 2001, 409, 683–684.
[
[
1] K. S. Mali, J. Adisoejoso, E. Ghijsens, I. De Cat, S. De Feyter, Acc. Chem.
Res. 2012, 45, 1309–1320.
2] K. Tahara, S. Lei, J. Adisoejoso, S. De Feyter, Y. Tobe, Chem. Commun.
[38] D. Bonifazi, S. Mohnani, A. Llanes-Pallas, Chem. Eur. J. 2009, 15, 7004–
7025.
[39] U. Mazur, K. W. Hipps, Chem. Commun. 2015, 51, 4737–4749.
[40] R. Gutzler, L. Cardenas, F. Rosei, Chem. Sci. 2011, 2, 2290–2300.
2010, 46, 8507–8525.
Chem. Eur. J. 2015, 21, 13569 – 13576
www.chemeurj.org
13575
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[
[
41] A. Ciesielski, P. J. Szabelski, W. R z˙ ysko, A. Cadeddu, T. R. Cook, P. J. Stang,
P. Samorì, J. Am. Chem. Soc. 2013, 135, 6942–6950.
42] A. Bellec, C. Arrigoni, G. Schull, L. Douillard, C. Fiorini-Debuisschert, F.
Mathevet, D. Kreher, A.-J. Attias, F. Charra, J. Chem. Phys. 2011, 134,
open-ring isomer (1o-C16) to the closed-ring isomer (1c-C16); see
ref. [54].
[60] Several thermodynamic models based on chemical potential have been
reported on the origin of polymorphs on a 2D surface; however, most
of them assume complete coverage on the substrate and a constant
concentration during order formation; see refs. [39] and [40], and refer-
ences therein.
[61] Polymorphs depending on alkyl chain length have been reported by
many groups; see, for example: a) E. Ghijsens, O. Ivasenko, K. Tahara, H.
Yamaga, S. Itano, T. Balandina, Y. Tobe, S. A. De Feyter, ACS Nano 2013,
7, 8031–8042; b) X.-Y. Wang, T.-F. Jiao, Z.-X. Zhang, T. Chen, M.-h. Liu, L.-
J. Wan, D. Wang, J. Phys. Chem. C 2013, 117, 16392–16396; c) L. Xu, X.
Miao, B. Zha, K. Miao, W. Deng, J. Phys. Chem. C 2013, 117, 12707–
1
24702.
43] N. T. N. Ha, T. G. Gopakumar, M. Hietschold, J. Phys. Chem. C 2011, 115,
1743–21749.
[
[
[
[
2
44] C.-A. Palma, J. Bjork, M. Bonini, M. S. Dyer, A. Llanes-Pallas, D. Bonifazi,
M. Persson, P. Samorì, J. Am. Chem. Soc. 2009, 131, 13062–13071.
45] L. Kampschulte, T. L. Werblowsky, R. S. K. Kishore, M. Schmittel, W. M.
Heckl, M. Lackinger, J. Am. Chem. Soc. 2008, 130, 8502–8507.
46] S. Lei, K. Tahara, F. C. De Schryver, M. Van der Auweraer, Y. Tobe, S. De
Feyter, Angew. Chem. Int. Ed. 2008, 47, 2964–2968; Angew. Chem. 2008,
12714; d) K. Tahara, S. Okuhata, J. Adisoejoso, S. Lei, T. Fujita, S. De Feyt-
1
20, 3006–3010.
er, Y. Tobe, J. Am. Chem. Soc. 2009, 131, 17583–17590.
[
47] K. Tahara, K. Inukai, J. Adisoejoso, H. Yamaga, T. Balandina, M. O. Blunt,
S. De Feyter, Y. Tobe, Angew. Chem. Int. Ed. 2013, 52, 8373–8376;
Angew. Chem. 2013, 125, 8531–8534.
[
62] Determination of the lattice parameters for compound 1o-C12 was dif-
ficult due to instability of the ordering under STM measurements; the
domain size of the ordering of 1o-C12 (observed at cꢂ50 mm) tended
to rapidly decrease during sequential scans in the same area (Figure S3
in the Supporting Information), which suggested that the dodecyl
chain was too short to impart sufficient stability of molecular ordering
of the 2-thienyl-type diarylethenes on HOPG substrate.
63] The open-ring isomer in the supernatant solution can undergo photo-
isomerization to give the closed-ring isomer. Through the adsorption/
desorption equilibrium of the molecular components, the reversible
photoinduced formation and disappearance of 2D molecular ordering
of the open-ring isomer was observed for 1o-C16; see ref. [54].
64] An incorrect value for the volume of supernatant solution (10 mL) was
used in the model simulation conducted in the previous report; see
ref. [54]. It was corrected to 15 mL in this research.
[
[
[
48] X.-M. Zhang, Q.-D. Zeng, C. Wang, Chem. Asian J. 2013, 8, 2330–2340.
49] S. Ahn, A. J. Matzger, J. Am. Chem. Soc. 2012, 134, 3208–3214.
50] M. O. Blunt, J. C. Russell, M. D. Gimenez-Lopez, N. Taleb, X. Lin, M.
Schrçder, N. R. Champness, P. H. Beton, Nat. Chem. 2011, 3, 74–78.
51] L.-P. Xu, L.-J. Wan, J. Phys. Chem. B 2006, 110, 3185–3188.
52] R. Heinz, A. Stabel, J. P. Rabe, G. Wegner, F. C. De Schryver, D. Corens, W.
Dehaen, C. Süling, Angew. Chem. Int. Ed. Engl. 1994, 33, 2080–2083;
Angew. Chem. 1994, 106, 2154–2157.
[
[
[
[
[
53] P. A. Korevaar, S. J. George, A. J. Markvoort, M. M. J. Smulders, P. A. J.
Hilbers, A. P. H. J. Schenning, T. F. A. de Greef, E. W. Meijer, Nature 2012,
4
81, 492–496.
54] S. Yokoyama, T. Hirose, K. Matsuda, Chem. Commun. 2014, 50, 5964–
966.
55] T. Hirose, F. Helmich, E. W. Meijer, Angew. Chem. Int. Ed. 2013, 52, 304–
09; Angew. Chem. 2013, 125, 322–327.
[
[
[
5
[
[
65] R. Zana, Adv. Colloid Interface Sci. 2002, 97, 205–253.
66] J. M. Del Rio, C. Pombo, G. Prieto, V. Mosquera, F. Sarmiento, J. Colloid
Interface Sci. 1995, 172, 137–141.
67] R. Zana, M. Benrraou, R. Rueff, Langmuir 1991, 7, 1072–1075.
68] F.-Y. Lin, W.-Y. Chen, R.-C. Ruaan, H.-M. Huang, J. Chromatogr. A 2000,
3
56] M. M. J. Smulders, M. M. L. Nieuwenhuizen, T. F. A. de Greef, P. van der
[
[
Schoot, A. P. H. J. Schenning, E. W. Meijer, Chem. Eur. J. 2010, 16, 362–
3
67.
8
72, 37–47.
69] D. Haidacher, A. Vailaya, C. Horvµth, Proc. Natl. Acad. Sci. USA 1996, 93,
290–2295.
70] T. Sakano, Y. Imaizumi, T. Hirose, K. Matsuda, Chem. Lett. 2013, 42,
537–1539.
[
57] T. F. A. de Greef, M. M. J. Smulders, M. Wolffs, A. P. H. J. Schenning, R. P.
Sijbesma, E. W. Meijer, Chem. Rev. 2009, 109, 5687–5754.
[
[
2
[
[
58] D. Zhao, J. S. Moore, Org. Biomol. Chem. 2003, 1, 2471–3491.
59] More experimental evidence that supports an adsorption/desorption
equilibrium at the octanoic acid/HOPG interface in our system is the
phototriggered reversible formation and disappearance of molecular
ordering of 1o-C16. In this system, desorption of ordering of the open-
ring isomer (1o-C16) can be reasonably explained by the decrease in
concentration in the solution phase upon photoisomerization from the
1
Received: February 18, 2015
Revised: June 15, 2015
Published online on July 28, 2015
Chem. Eur. J. 2015, 21, 13569 – 13576
www.chemeurj.org
13576
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim