Synergistic Anion-(π)n-π Catalysis on π-Stacked Foldamers

Article abstract of DOI:10.1021/jacs.8b00809

In this report, we demonstrate that synergistic effects between π-π stacking and anion-π interactions in π-stacked foldamers provide access to unprecedented catalytic activity. To elaborate on anion-(π)_n-π catalysis, we have designed, synthesized and evaluated a series of novel covalent oligomers with up to four face-to-face stacked naphthalenediimides (NDIs). NMR analysis including DOSY confirms folding into π stacks, cyclic voltammetry, steady-state and transient absorption spectroscopy the electronic communication within the π stacks. Catalytic activity, assessed by chemoselective catalysis of the intrinsically disfavored but biologically relevant addition reaction of malonate half thioesters to enolate acceptors, increases linearly with the length of the stacks to reach values that are otherwise beyond reach. This linear increase violates the sublinear power laws of oligomer chemistry. The comparison of catalytic activity with ratiometric changes in absorption and decreasing energy of the LUMO thus results in superlinearity, that is synergistic amplification of anion-π catalysis by remote control over the entire stack. In computational models, increasing length of the π-stacked foldamers correlates sublinearly with changes in surface potentials, chloride binding energies, and the distances between chloride and π surface and within the π stack. Computational evidence is presented that the selective acceleration of disfavored but relevant enolate chemistry by anion-π catalysis indeed originates from the discrimination of planar and bent tautomers with delocalized and localized charges, respectively, on π-acidic surfaces. Computed binding energies of keto and enol intermediates of the addition reaction as well as their difference increase with increasing length of the π stack and thus reflect experimental trends correctly. These results demonstrate that anion-(π)_n-π interactions exist and matter, ready for use as a unique new tool in catalysis and beyond.

Full text of DOI:10.1021/jacs.8b00809

Article
pubs.acs.org/JACS
Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Synergistic Anion−(π)n−π Catalysis on π‑Stacked Foldamers
Anna-Bea Bornhof,† Antonio Bauza,‡ Alexander Aster,§ Marion Pupier,† Antonio Frontera,‡
́
Eric Vauthey,§ Naomi Sakai,† and Stefan Matile*,†
§
†Department of Organic Chemistry, Department of Physical Chemistry, University of Geneva, CH-1211 Geneva, Switzerland
‡Department de Química, Universitat de les Illes Balears, Carretera de Valldemossa km 7.5, 07122 Palma de Mallorca, Baleares, Spain
S
* Supporting Information
ABSTRACT: In this report, we demonstrate that synergistic
effects between π−π stacking and anion−π interactions in π-
stacked foldamers provide access to unprecedented catalytic
activity. To elaborate on anion−(π)n−π catalysis, we have
designed, synthesized and evaluated a series of novel covalent
oligomers with up to four face-to-face stacked naphthalenedi-
imides (NDIs). NMR analysis including DOSY confirms folding
into π stacks, cyclic voltammetry, steady-state and transient
absorption spectroscopy the electronic communication within the
π stacks. Catalytic activity, assessed by chemoselective catalysis of
the intrinsically disfavored but biologically relevant addition
reaction of malonate half thioesters to enolate acceptors, increases
linearly with the length of the stacks to reach values that are
otherwise beyond reach. This linear increase violates the sublinear power laws of oligomer chemistry. The comparison of catalytic
activity with ratiometric changes in absorption and decreasing energy of the LUMO thus results in superlinearity, that is
synergistic amplification of anion−π catalysis by remote control over the entire stack. In computational models, increasing length
of the π-stacked foldamers correlates sublinearly with changes in surface potentials, chloride binding energies, and the distances
between chloride and π surface and within the π stack. Computational evidence is presented that the selective acceleration of
disfavored but relevant enolate chemistry by anion−π catalysis indeed originates from the discrimination of planar and bent
tautomers with delocalized and localized charges, respectively, on π-acidic surfaces. Computed binding energies of keto and enol
intermediates of the addition reaction as well as their difference increase with increasing length of the π stack and thus reflect
experimental trends correctly. These results demonstrate that anion−(π)n−π interactions exist and matter, ready for use as a
unique new tool in catalysis and beyond.
that synergistic anion−(π)n−π interactions would provide
access to exceptionally strong anion−π interactions already
on simple, uniform or mixed π dimers of benzene and
hexafluorobenzene, i.e., minimalist aromatics that essentially fail
to interact with ions as monomers.6
INTRODUCTION
■
Common aromatic compounds have electron clouds above and
beneath their surfaces. The result are negative quadrupole
moments perpendicular to the π plane (QZZ < 0) that naturally
attract cations rather than anions.1 If these compounds are,
however, substituted with strong electron withdrawing groups,
the quadrupole moment can be inverted, thus making them π
acidic. Anion binding to such π acids has been confirmed to
occur in solid and solution state with functional relevance in
self-assembly and transport.2 Evolving from the ground state to
anion stabilization in the transition state, anion−π catalysis has
been introduced recently.3 Using electric fields, we have
demonstrated the importance of polarizability in anion−π
catalysis.4 In a complementary approach, large and highly
polarizable fullerene surfaces supported by induced aromatic
dipoles rather than inverted quadrupole moments have been
found to provide access to anion−π interactions of highest
functional relevance.5 Toward polarizing the π-acidic surfaces
by charge delocalization over systems as large as possible, we
here introduce synergistic anion−(π)n−π interactions in π-
stacked foldamers. Theoretical studies have predicted early on
Naphthalenediimides (NDIs)7 have emerged as particularly
attractive framework to elaborate on anion−π interactions
because their intrinsic quadrupole moment is highly positive
and can be easily varied with core substituents.8 These
properties have been essential in constructing and proving
functional anion−π transporters and catalysts.3,4,8 A variety of
materials based on NDI stacks has been developed over the
past decades7,9−23 for organoelectronics,7,9 synthetic ion
channels,8,10 sensors and catalysts.11 Most of these materials
are based on noncovalent assemblies of NDIs that form
structural motifs like helices,12 nanotubes,13,14 sheets15 or
spheres.16 For a better understanding of these large assemblies,
efforts have been made to synthesize covalent, well-defined
Received: January 22, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.8b00809
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 1. Structure of NDI foldamers 1−4 made to elaborate on synergistic anion−(π)n−π catalysis.
a
structures, and electronic communication as well as spectro-
scopic behavior have been investigated.17,18 However, contrary
to extensive studies on donor−acceptor foldamers with NDIs
and π-basic partners,19 most of these covalent NDI oligomers
are limited to two units. Larger discrete stacks were synthesized
by self-assembly on DNA20 and rigid-rod templates,10
dendrimers,21 or polyrotaxane chemistry.22
Scheme 1. Foldamer Synthesis
Herein we report the design, synthesis and evaluation of
discrete, monodisperse NDI stacks containing up to four
covalently connected, face-to-face stacked NDIs. By exper-
imental and computational studies on their folding behavior,
electronic communication and catalytic activities, we aim to
demonstrate the existence and relevance of synergistic anion−
(π)n−π catalysis.
RESULTS AND DISCUSSION
■
The NDI monomer 1, dimer 2, trimer 3 and tetramer 4 were
synthesized in a linear manner (Figure 1). Starting from
dianhydride 5, NDI formation with amines 6 and 7 afforded
monomer 1 in a single step (Scheme 1). The side products with
two identical imide substituents were removed by column
chromatography. Oligomer synthesis was initiated from the
same dianhydride 5. The Boc-protected benzyl turn 8 and the
leucyl-hexyl (LH) tail 9 were introduced in a basic
condensation reaction yielding the monomeric building block
10. The byproduct 11 with two LH tails was isolated for use as
negative control in anion−(π)n−π catalysis. Deprotection of
NDI 10, reaction with another NDA 5 and condensation with
the dimethylamine headgroup 7 gave the desired NDI dimer 2.
The dimeric building block 12 was prepared similarly from the
same monomeric building block 10 using turn 8 instead of
tertiary amine 7 in the condensation reaction. This dimeric
building block 12 could be further converted to either trimer 3
or trimeric building block 13 using the same protocols as for
dimer 2 or dimeric building block 12, respectively. Finally,
tetramer 4 was prepared from trimeric building block 13 by
deprotection of the of the amine and successive condensation
with one equivalent of NDA 5 and the catalytic headgroup 7.
a
(a) 1. 6, AcOH, 120 °C, 1.5 h, 2. 7, AcOH, 120 °C, 5 h, 20%; (b) 1.
8, TEA, toluene/DMF 2:1, 120 °C, Ar, 30 min, 2. 9, AcOH, toluene/
DMF 2:1, 120 °C, Ar, 1 h, 30% (10), 5% (11); (c) 1. TFA/CH2Cl2
1:1, rt, 1 h, 2. 5, TEA, toluene/DMF 2:1, 120 °C, Ar, 1.5 h, 3. 7, TEA,
toluene/DMF 2:1, 120 °C, Ar, 30 min, 44%; (d) 1. TFA/CH2Cl2 1:1,
rt, 1 h, 2. 5, TEA, toluene/DMF 2:1, 120 °C, Ar, 1.5 h, 3. 8, AcOH,
toluene/DMF 2:1, 120 °C, Ar, 4 h, 55%; (e) 1. TFA/CH2Cl2 1:1, rt, 1
h, 2. 5, TEA, toluene/DMF 2:1, 120 °C, Ar, 2.5 h, 3. 7, toluene/DMF
2:1, 120 °C, Ar, 1 h, 57%; (f) 1. TFA/CH2Cl2 1:1, rt, 1 h, 2. 5, TEA,
toluene/DMF 2:1, 120 °C, Ar, 2.5 h, 3. 8, AcOH, toluene/DMF 2:1,
120 °C, Ar, 3 h, 41%; (g) 1. TFA/CH2Cl2 1:1, rt, 1 h, 2. 5, TEA,
toluene/DMF 2:1, 120 °C, Ar, 1 h, 3. 7, toluene/DMF 2:1, 120 °C, Ar,
30 min, 30%.
Foldamer24 Characterization. In the H NMR spectrum
1
of monomer 1, the NDI protons appeared as a single peak at
8.8 ppm (Figure 2A). This deshielded resonance nicely
illustrated the electron-deficient nature of the aromatic system.
In dimer 2, the NDI signals split into six resolved doublets, all
with nearly identical coupling constants and spread out toward
8.0 ppm. This variable upfield shift was in agreement with the
folding of dimer 2 into a face-to-face stack, with NDI protons
exposed to the ring current from the proximal second NDI.
Different shifts for six out of eight possible doublets suggested
that the environment of all but two NDI protons differs
(Figures 2A, S33−S39). This nearly maximal peak separation
confirmed expectations from computational simulations that in
dimer 2, the long axes of two cofacial NDI planes are twisted
out of colinearity (29.8°, Figure S14, see below).
With trimer 3 and tetramer 4, more doublets appeared
upfield, consistent with the new NDIs being sandwiched
between two neighbors in the π-stacked foldamers. This trend
B
DOI: 10.1021/jacs.8b00809
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
presence of only one set of signals in all spectra, in CDCl3 and
THF-d8/CDCl3 3:1 (see catalysis), demonstrated that,
independent of stack length, conformers other than the robust
π-stacked foldamers are at most a mere minority (Figure 2A).
Very strong preference for a π-stacked conformation was
further demonstrated by the independence on solvent and
temperature up to 100 °C in competing toluene in UV−vis
absorption studies (vide infra).24,26
In the DOSY 2D NMR spectra, a diffusion coefficient of D =
4.22 × 10−10 m2 s−1 was measured for trimer 3 (Figure S28).
According to Stokes−Einstein equation assuming a spherical
molecule, this corresponded to a radius of 9.65 Å.27 The
geometry-optimized structure of trimer 3 showed a spherical
radius of 9.9 Å in good agreement with the experimental value
(Figure S27, see below). The same experiment was performed
for tetramer 4. In CDCl3 at 298 K, a diffusion coefficient of 3.98
× 10−10 m2 s−1 was extracted from the DOSY spectra (Figure
S28). The slower diffusion of the larger tetramer 4 compared to
trimer 3 was as expected. The resulting spherical radius of 10.23
Å was again close to the theoretical value of 10.8 Å of the
geometry-optimized structure (Figure S27).
Cyclic voltammetry (CV) and differential pulse voltammetry
measurements (DPV) were performed in MeCN with a 0.1 M
solution of TBAPF6 (Figures 2B, S6). Monomer 1 exhibited a
one-electron reduction wave at V = −1.02 V corresponding to
the formation of the radical anion NDI•− (Figure 2B, bottom).
With dimer 2 and trimer 3, this first reduction wave split into
the separated respective reduction waves for each NDI. The
maximum of the first wave shifted to more positive potentials
with increasing foldamer length, from V = −1.02 V for
monomer 1 to V = −0.95 V for dimer 2 and V = −0.90 V for
trimer 3 (Figure 2B). These shifts of up to 120 mV calculated
to a decrease of the LUMO energy from −4.08 eV to −4.20 eV
against −5.10 eV for Fc+/Fc (Table 1). As for the similar results
obtained with mechanically bonded discrete NDI stacks,22 we
interpreted this lowering of the LUMO level to be caused by
favorable electronic communication between the NDIs in the
stacks.
1
Figure 2. (A) Diagnostic region of H NMR spectra of monomer 1,
dimer 2, trimer 3 and tetramer 4 (bottom to top) in THF-d8/CDCl3
3:1, with assignments to NDIs a (amine terminus), b, c and d (LH
terminus) for 4 (*, signals from benzyl turns). (B) Cyclic
voltammograms (CVs) of NDIs 1, 2 and 3 (bottom to top) in
MeCN with a glassy carbon working electrode, a calomel reference
electrode, TBAPF6 (0.1 M) as supporting electrolyte and a scan rate of
50 mV s−1. (C) Normalized UV−vis absorption spectra of 1, 2, 3 and
4 in THF (light to dark blue). (D) Dependence of absorbance ratio on
oligomer length, i.e., monomer per oligomer, with power fit; from
panel C. (E) Dependence of LUMO energies on oligomer length, with
power fit; from panel B.
was as expected for folding into ordered, face-to-face π-stacked
oligomers. For all foldamers, a tentative assignment of up to
four separated doublets to each individual NDI in the stacks
was possible by COSY, HMBC and HSQC 2D NMR
spectroscopy (Figures 2A, S33−S58). High-resolution 2D
spectra from computer-assisted aliasing techniques, developed
in-house,25 were needed to resolve the fine-structured signal
clusters (rather than broad peaks) that are characteristic for
oligomers in ordered (rather than disordered) conformation.
Additional signals for NDI protons were not observed. This
Identical trends were observed in THF/CH2Cl2 3:1 (almost
the solvent mixture used for catalysis, see below), but the
individual maxima were not resolved (Figure S7). The poor
resolution with trimer 3 also in MeCN (Figure 2B) and
a
Table 1. Characteristics of Anion−(π)n−π Catalysts
b
c
d
e
f
g
h
i
j
k
l
m
n
o
oligo
I0‑0/0‑1
ELUMO (eV) dπ−π
MEP (kJ mol−1
)
Eint (Cl) dCl‑π
Eint (Cl)
de‑π
Eint (enol)
dk‑π
Eint (keto) ΔEint
A/D
1
1.14
1.00
0.92
0.87
−
−4.08
−4.15
−4.20
nd
−
100.0
123.5
125.1
127.6
129.0
−0.1
−3.3
−5.8
−8.3
−12.1
3.10
−133.7
−137.5
−140.0
−141.7
−142.7
3.39
3.37
3.36
3.35
−
−183.9
−186.0
−187.3
−188.5
−
2.91
2.88
2.70
2.69
−
−176.2
−178.1
−178.3
−179.0
−
7.7
8.7
9.0
9.5
−
1.8
2.8
7.2
2
3.53
3.48
3.46
3.44
3.09
3.08
3.07
3.07
3
4
10.4
hyp
−
−
a
b
Distances d in Å, energies E in kJ mol−1. Oligomers, see Figure 1. hyp: Hypothetical pentamer. 1′−4′: Hypothetical oligomers with one methyl
c
terminus and one correct amine terminus 1″−4″: Hypothetical oligomers with two methyl termini, see Figure 5. Absorbance of the 0-0 divided by
d
the absorbance of the 0-1 vibronic transition, Figure 2C. Energy of the LUMO in eV against vacuum, assuming −5.10 eV for Fc+/Fc.
e
f
e−nComputational data. Distance between the first two π planes, Figure S13. Maximal molecular electrostatic potential on the external π surface,
g
h
Figure S15. Interaction energy between the chloride anions and the external π surface of 1″−4″ (without covalent counterion), Figure 5. Distance
i
j
between Cl− and 1′−4′ (with covalent counterion), Figure 5. Interaction energy between Cl− and 1′−4′, Figure 5. Distance between the enol
k
tautomer of conjugate base of 14′ (Me instead of PMP thioester) and 1′−4′, Figure 6. Interaction energy between enol tautomer of 14′ and 1′−4′,
l
m
Figure 6. Distance between the keto tautomer of 14′ and 1′−4′, Figure 6. Interaction energy between the keto tautomer of 14′ and 1′−4′, Figure
n
o
6. Difference in interaction energy of enol and keto tautomers with 1′−4′. Yield of addition (16)/yield of decarboxylation (17). Reactions were
1
conducted with 200 mM 14, 20 mol % catalysts 1−4 and 2 M acceptor 15 at 20 °C in THF-d8/CDCl3 3:1, and monitored by H NMR
spectroscopy. Total conversion (A + D) was always almost quantitative (>90%), reproducibility within 3.6% for nine independent experiments
(Figure 7).
C
DOI: 10.1021/jacs.8b00809
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
limitations with regard to solubility suggested that measure-
ments with tetramer 4 are meaningless.
The UV−vis absorption spectra of foldamers 1−4 in THF
exhibited the two main maxima below 400 nm that can be
assigned to the 0-0 and 0-1 vibronic transitions (Figure 2C).28
For monomer 1, the absorbance of the first 0-0 band at 381 nm
was stronger, the absorbance ratio of the two maxima was
of the instrument response function. In trimer 13, the
interaction of the stacked NDI units led to a blue shift of the
excited state absorption to a broad maximum around 560 nm,
indicating a delocalization of the excitation to an excimer-like
state (Figure 3B). This excimer state then decayed to the
NDI•−. Compared to monomer 10, this radical anion band was
spectrally broadened. As with the broadening of the CV
reduction waves in longer foldamers (Figure 2B), this finding
supported a delocalization of the anion over the stacked NDIs
in trimer 13. A detailed description of the ultrafast photo-
chemistry of π-stacked NDI foldamers will follow in an
upcoming publication.
Without substituents in their core, NDIs are essentially not
fluorescent because intersystem crossing from the first singlet
excited state to the triplet manifold is very fast (<200 fs).32
With monomer 10, electron transfer from the benzyl turn to
the NDI (Figure 3A) further weakened emission intensity
(Figure S12). With dimer 12 and trimer 13, excimer formation
(Figure 3B) resulted in excimer emission, which was also very
weak but clearly detectable in steady-state experiments (Figure
S12).
I0‑0/0‑1 = 1.14 (Table 1). With increasing length of the foldamer,
this ratio inverted to I0‑0/0‑1 = 0.87 for tetramer 4 (Figure 2C,
Table 1). The dependence of I0‑0/0‑1 on the number of
monomers per foldamer was sublinear and in good agreement
with a power fit (Figure 2D). The sublinear power dependence,
characteristic for diminishing returns with increasing oligomer
length, applied also to the decrease in LUMO energies,
although the availability of only three data points naturally
limited significance (Figure 2E).
The ratiometric change of the 0-0 and 0-1 vibronic bands has
been shown to be a good measure for the electronic coupling
within π stacks.29 An inversion of absorbance ratio has been
assigned regularly to π−π stacking in noncovalent self-
assemblies of NDIs.13,14,18 Moreover, the same phenomenon
has been observed in covalent perylenediimide stacks, thus
supporting the presence of extensive electron sharing between
the NDI monomers in the here introduced series of
foldamers.29,30
Inversion of ratiometric absorption with sublinear depend-
ence on stack length was observed in all solvents tested,
including CHCl3, THF, MeCN and toluene (Figure S8). The
ratiometric absorption of trimer 3 in toluene did not change
with increasing temperatures up to 100 °C (Figure S10). This
insensitivity toward different solvents and the resistance toward
thermal denaturation in a most competitive solvent suggested
that the π-stacked NDI foldamers are very stable.
Computational Studies. Theoretical calculations on the
BP86-D3/def2-TZVP level with dispersion correction in THF
continuum5 were initiated with simplified model oligomers 2″−
4″ with two methyl termini. Face-to-face π-stacked foldamers
were identified as most stable conformers independent of
oligomer length (Figures 4A, S25, S26). Completely unfolded
Transient absorption spectra upon excitation at 385 nm were
recorded for the monomeric building block 10 and trimeric
building block 13 (Figures 3, S11). After excitation of
Figure 4. Conformational analysis of trimer 3″ (Me termini) covering
(A) face-to-face π-stacked foldamer (0 kJ mol−1), (B) helical foldamer
with lone-pair−π interactions (+0.9 kJ mol−1) and (C) unfolded
oligomer (+95 kJ mol−1).
oligomers without any contacts between the NDI were much
less stable, e.g., +95 kJ mol−1 for trimer 3″ (Figure 4C). Helical
conformers stabilized by lone-pair−π interactions from one
imide oxygen to the aromatic surface occurred at only slightly
higher energy, e.g., +0.9 kJ mol−1 for trimer 3″, with O−π
distances down to 2.97 Å (Figure 4B). However, results from
NMR spectra, DOSY, CV and ratiometric absorption
demonstrated that only one conformer is strongly preferred
and that the NDIs of this conformer share electrons, which
rules out all but face-to-face π-stacked foldamers (Figures 1, 2, 3
and 4A). The underestimation of their stability compared to
stack-free, partially unfolded conformers (Figure 4B) in
computational models is likely to originate from entropic
contributions because largest contact areas in π-stacked
foldamers result in highest desolvation. These entropic
contributions have been identified as dominant driving force
in foldamer chemistry and self-assembly but passed unnoticed
Figure 3. Transient absorption spectra of monomeric building block
10 (A) and trimeric building block 13 (B) in acetonitrile upon
excitation at 385 nm.
monomer 10, a fast decrease of the excited state absorption
band at 597 nm and the stimulated emission at 407 nm with a
parallel rise of the NDI radical anion band at 475 nm was
consistent with photoinduced electron transfer from the benzyl
turn to the NDI (Figure 3A).31,32 In agreement with charge
recombination, the NDI•− band subsequently decayed
simultaneously with the ground state bleach. Note that the
rise of the excited state absorption from −0.1 to +0.1 ps
originated from the finite time resolution of the setup, which is
about 200 fs. The occurrence of the NDI•− band at this early
time steps showed that the electron transfer rate is on the order
D
DOI: 10.1021/jacs.8b00809
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
in calculations that represent solvent as a continuous
medium.24,26
chloride to the π surface decreased similarly from 3.10 Å in
monomer 1′ to 3.07 Å in tetramer 4′, whereas the hydrogen
bond to the counterion increased correspondingly from 1.68
to1.71 Å (Figures 5B, S19).
In the preferred face-to-face π-stacked foldamers, the
distances between the first two NDI planes decreased from
dimer 2″ with 3.53 Å to 3.46 Å for tetramer 4″ and 3.44 Å for a
hypothetical pentamer (Figure S13, Table 1). Consistent with
experimental results on electron sharing from absorbance ratio
and decreasing LUMO energy, this increasing π stacking with
increasing foldamer length was sublinear and showed good
agreement with a power fit. Axial views of the foldamers
revealed that the twist angle between adjacent NDIs gradually
decreased with increasing stack length (Figure S14). The
decrease for the peripheral twist from 29.8° in dimer 2″ to
minimal 26.9° in hypothetical pentamer 5″ was relatively
minor. The smallest twist in the center of the stacks, however,
decreased from 29.8° dimer 2″ to 21.2° in tetramer 4″ and
15.0° in hypothetical pentamer 5″. This decreasing twist with
increasing stack length correlated well with increasing electron
sharing in optoelectronic studies.
The molecular electrostatic potential (MEP) surface of
monomer 1″ showed maximal +100.0 kJ mol−1 on the aromatic
π surface (Figure S15). This maximal potential increased to
+123.5 kJ mol−1 in dimer 2″. From there, a less pronounced
increase around ∼ +1.5 kJ mol−1 per monomer was observed.
Overall, the dependence of the maximal potential to oligomer
length was again sublinear.
In the presence of covalent counterions, the contribution per
monomer to the chloride binding energy decreased with
increasing foldamer length. The resulting overall increase was
again sublinear (Figure 5B). This power dependence of
chloride binding by foldamers 1′−4′ reflected experimental
trends on electron sharing from absorption spectra and CV. In
clear contrast, the linear increase of chloride binding with
increasing length of counterion-free foldamers 1″−4″ exceeded
expectations from experiments. Overruling the intrinsic sub-
linear dependence on stack length, these results supported the
occurrence of strong synergistic effects between anion−π
interactions and π−π interactions within the π stack. The
increase of the synergism of chloride−π interactions in the
absence of covalent counterions coincided with a shortening of
the distance from anion to π surface from 3.10 Å in monomer
1′ to 2.73 Å in 1″ and from 3.07 Å in tetramer 4′ to 2.71 Å in
4″. This emergence of synergistic effects expressed in the linear
dependence of chloride binding to oligomers 1″−4″ will be of
interest with regard to anion−(π)n−π catalysis described in the
following.
Anion−(π)n−π Catalysis. The addition of malonic acid half
thioester (MAHT) 14 to enolate acceptors such as nitroolefin
15 has emerged as the reaction of choice to probe for anion−π
catalysis (Figure 6).3−5 This reaction is at the beginning of all
biosynthesis but often disfavored without enzymes.3−5 Instead
of the significant addition product 16, the irrelevant
decarboxylation product 17 is obtained. We have argued early
on3 that this chemoselectivity can be controlled on the level of
Chloride binding was negligible to NDI monomer 1″ that
does not offer support from additional interactions (Eint = −0.1
kJ mol−1 Figure 5A, Table 1). With increasing length of the π-
Figure 5. Dependence of interaction energies on foldamer length for
chloride binding to 1″−4″ and a hypothetical pentamer with terminal
methyl groups (A) and to 1′−4′ and a hypothetical pentamer with
covalent ammonium counterions at one and a Me at the other
terminus (B), together with examples for energy-minimized structures
of chloride complexes of monomers 1″ and 1′ and tetramer 4′.
stacked foldamers, chloride binding to the π surface increased
linearly to maximize for tetramer 4″ at Eint = −8.3 kJ mol−1 and
a hypothetical pentamer at Eint = −12.1 kJ mol−1 (Figures 5A,
S16, S17, Table 1). This increase in binding was reflected in a
slight shortening of the distance of the chloride to the π surface,
from 2.73 Å in monomer 1″ to 2.71 Å in tetramer 4″ (Figures
5, S16).
Chloride binding to the (almost) real oligomers 2′−4′ with a
covalently attached ammonium cation next to the π surface was
necessarily dominated by ion pairing. Chloride binding
increased with increasing foldamer length by ΔEint = −8.0 kJ
mol−1 from monomer 1′ to tetramer 4′ (Table 1). This
increase was almost identical to the increase computed in the
absence of ammonium counterions. The distance of the
Figure 6. (A) Selective catalysis of the addition of MAHT 14 to
acceptor 15 to afford addition product 16 rather than decarboxylation
product 17 is envisioned by discrimination between enol tautomers as
in reactive intermediate RIA-1 and transition state TSA-1 and keto
tautomers as in TSD-1. (B) Dependence of interaction energies on
foldamer length for binding of enol ( ) and keto ( ) tautomers (as in
TSA-1 and TSD-1) to 1′−4′, with examples for energy-minimized
structures of enol (C) and keto (D) tautomers bound to monomer 1′
(Me instead of PMP thioester, PMP = p-methoxyphenyl, models with
PMP gave the same trends, Figure S24).
●
○
E
DOI: 10.1021/jacs.8b00809
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 7. Dependence of the A/D chemoselectivity of the reaction of MAHT 14 with enolate acceptor 15, i.e., the yield of addition product 16
divided by the yield of decarboxylation product 17 (Figure 6), on (A) the length of foldamer catalysts 1−4 (with linear fit; measured in THF-d8/
●
CDCl3 3:1 ( ), compared to THF-d8/CDCl3 1:1 (+) and analogs with hexyl instead of LH tails in CDCl3 (X), reproducibility was within 3.6% for
nine independent experiments), (B) the absorbance ratio, and (C) the computed distance between first and second π plane in the π stack, with
power fits and indication of remote control in unfolded and folded reactive intermediates RIA-2 and RIA-3 of tetramer 4.
tautomers of the anionic malonate half thioester (MHT)
intermediates obtained by deprotonation of substrate 14.
Namely, “enol” tautomers as in reactive intermediate RIA-1
and transition state TSA-1 have to undergo the addition
reaction before decarboxylation, whereas “keto” tautomers as in
reactive intermediate TSD-1 can decarboxylate before addition.
With an sp2 α carbon, additive tautomers as in TSA-1 are planar,
and the negative charge is fully delocalized. Decarboxylative
tautomers as in TSD-1 have an sp3 α carbon, are thus bent and
have their negative charge localized on the carboxylate. From
the beginning,3 we have thought that π-acidic surfaces would be
perfect to feel the subtle difference between these tautomers,
that is planar against bent intermediates with delocalized
against localized negative charge. Experimental trends from
various systems were consistent with these meaningful
speculations and used to support, inter alia, existence and
significance of anion−π catalysis as such3,33 as well as the
successful creation of anion−π enzymes34 and electric-field-
mediated anion−π catalysis.4
Theoretical studies conducted here for the first time proved
that these intuitive arguments are correct. In energy-minimized
models, the recognition of the planar enol tautomer on
monomer 1′ exceeded that of the bent keto tautomer by ΔEint
= −7.7 kJ mol−1 (Figure 6C,D, Table 1). The obtained
structures feature the planar enol tautomer bound parallel to
the π surface at a distance d = 3.39 Å, comparable to that of
chloride complexes (Figure 5). In clear contrast, the bent
tautomer interacted with the π surface with an edge-to-face
carboxylate that carries the full negative charge. For practical
reasons, most calculations were done with model intermediates
with methyl thioesters (Figure 6B−D, Table 1). Controls with
p-methoxyphenyl (PMP) thioesters as in substrate 14 on
monomers 1′ gave nearly identical structures with even slightly
higher tautomer recognition (ΔEint = −10.0 kJ mol−1, d = 3.37
Å, Figure S24).
contributions per monomer with increasing stack length
applied for both tautomers (Figures 6B, S20−S23). However,
the exponent differed significantly. As a result, the selective
recognition of the planar enol tautomer over the bent keto
tautomer increased with increasing length of the foldamer from
E
int = −7.7 kJ mol−1 on monomer 1′ to Eint = −9.5 kJ mol−1 on
tetramer 4′. The resulting ΔEint = −1.8 kJ mol−1 promised that
contributions from synergistic anion−(π)n−π interactions
would exist and matter.
The catalytic activity of foldamers 1−4 was determined in a
mixture of THF-d8/CDCl3 3:1 containing 200 mM of MAHT
substrate 14 and 2.0 M of enolate acceptor 15 at 20 °C. Under
these conditions, control reactions with 20 mol % TEA in place
of anion−(π)n−π catalysts were very slow and did not reach
completion within 2 weeks. With 20 mol % foldamers 1−4, the
total conversion was always almost quantitative (>90%).33 As a
measure of chemoselectivity, the yield of the intrinsically
disfavored but significant addition (A) product 16 was divided
by the yield of the decarboxylation (D) product. A/D values <1
thus stand for dominant decarboxylation, A/D > 1 for
dominant addition.
The A/D = 1.8 obtained with monomer 1 was consistent
with similar values obtained previously with similar anion−π
catalysts (Table 1).33 With dimer 2, selectivity increased to
A/D = 2.8, then further to very significant A/D = 7.2 and A/D
= 10.4 with trimer 3 and anion−(π)3−π catalyst 4, respectively.
The overall most impressive increase from A/D = 1.8 to A/D =
10.4 from monomer to tetrameric π stacks provided strong
experimental support for existence and relevance of synergistic
anion−(π)n−π catalysis. “Double-digit” selectivity as for
tetramer 4 has so far been achieved only on the large,
polarizable π surface of fullerenes.5
Moving from THF-d8/CDCl3 3:1 to THF-d8/CDCl3 1:1, all
selectivities increased about the same extent to maximize at
A/D = 13.3 for tetramer 4 (Figure 7A+). This increase should
not be overinterpreted because catalysis with the TEA control
also becomes faster and with A/D = 2.1 more selective under
these conditions. Even higher selectivities were accessible in
CDCl3 with more hydrophobic foldamers having simple hexyl
tails in place of the LH tails in the series 1−4 (Figure 7AX).
As for chloride binding, the binding of both enol and keto
tautomers was characterized by a sublinear increase with NDI
foldamers length to reach Eint = −188.5 kJ mol−1 and Eint
=
−179.0 kJ mol−1, respectively, on tetramer 4′. The empirical
apparent power dependence observed previously for the
F
DOI: 10.1021/jacs.8b00809
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Although promising A/D = 6.0 and 12.5 were found for
monomers and dimers, respectively, trimer catalysts gave an
unchanged A/D = 12.7 probably due to their low solubility thus
the lower effective concentration in solution. Because solubility
problems increase with increasing foldamer length, the hexyl
series was discontinued (not shown). Kinetics were consistent
with previously reported trends for other systems (Figures S1,
S2).3 Reproducibility of A/D values was within 3.6% for nine
independent experiments made in parallel with identical
starting materials (catalysts, substrates, solvents, etc.). Syste-
matic errors in the comparison between different batches are
more difficult to quantify (e.g., possibly different amounts of
absorbed moisture or residual solvents that could affect
balancing, water content in solvents), but they were confirmed
to be small enough not to influence the overall trends,
particularly because the observed differences in activity were so
large.
The catalytic activity of monomer 1 (20 mol%) was also
measured in the presence of 0, 20, 40, and 60 mol% of dialkyl
NDI 11 to mimic the total NDI concentration with dimer 2,
trimer 3 and tetramer 4, respectively. However, measured in a
mixture of THF-d8/CDCl3 1:1 under standard conditions, the
selectivity remained constant at A/D ∼ 4.0 despite the
increasing NDI concentration. This result excluded the
possibility that the increase of A/D selectivity found with
increasing foldamer length would be just a consequence of
higher absolute concentration of NDI surfaces. Insensitivity of
the absorbance ratio I0‑0/0‑1 to the presence of substrate 14
supported that the foldamer structure is intact under conditions
used for catalysis (Figure S9). Molecular models agreed that
there is not enough space also for temporary malonate
intercalation into the NDI stacks, and that eventual temporary
displacement of NDIs in the stacks by malonate intermediates
is unlikely (see above).
The dependence of A/D selectivity on foldamer length was
almost linear in both THF-d8/CDCl3 3:1 (Figure 7A●) and
THF-d8/CDCl3 1:1 (Figure 7A+). Largest deviations from
linearity were found with 2, probably because of the poorer π
stacking in dimers compared to higher oligomers, as expressed
in a longer plane-to-plane distance of d = 3.53 Å (Table 1). The
apparent overperformance of dimer 2 in the more hydrophobic
environment in THF-d8/CDCl3 1:1 could thus be explained
with stronger π stacking induced by stronger anion-π
interactions with the anionic intermediates and transition states
(Figure 7A+). The complementary underperformance of dimer
2 in the less hydrophobic environment in THF-d8/CDCl3 3:1
would then originate from less support from anion-π
interactions to strengthen the fragile π stacking (Figure 7A●).
The linear dependence of anion−(π)n−π catalysis on
foldamer length (Figure 7A) was in contrast to the sublinear
dependence of absorbance ratio and LUMO energy levels
(Figure 2D,E). This suggested that diminishing return on
electron sharing with increasing stack length are compensated
or even slightly overcompensated by synergistic contributions
from anion−(π)n−π interactions with anionic intermediates
and transition states of the catalyzed reaction. This synergism
was best expressed in the dependence of the A/D selectivity on
the absorbance ratio of anion−(π)n−π catalysts 1−4 (Figure
7B). The result of this comparison was pronounced super-
linearity.
CONCLUSIONS
■
The objective of this study was to explore the possible existence
and significance of anion−(π)n−π catalysis. The idea was that
the stabilization of anionic intermediates and transition states
by anion−π interactions on π-acidic surfaces could be amplified
by electron sharing within π stacks. To elaborate on these
expectations, a concise series of π-stacked NDI foldamers was
designed, synthesized and evaluated. Their optoelectronic
properties, that is inversion of absorbance ratio and decreasing
LUMO energies, were consistent with increasing electron
sharing with increasing stack length. With diminishing returns
per monomer with increasing length, they obeyed the sublinear
power laws of oligomer chemistry. In computational models,
chloride−π interactions, enolate−π interactions, tautomer
discrimination and anion-stack and intrastack distances all
responded to increasing stack length, following the same
sublinear power dependence.
Experimental results on anion−(π)n−π catalysis violated the
sublinear power law of oligomer chemistry. The selectivity
increased with increasing stack length in a roughly linear
manner (Figure 7A). Comparison against the underlying trends
from optoelectronic properties and theoretical predictions
revealed quite remarkable superlinearity (Figure 7B,C). This
superlinearity demonstrated that the identified power of
anion−(π)n−π catalysis originates from operational synergism
between anion−π and π−π interactions, i.e., electron sharing
over a very long distance. Moving from trimers to tetramers,
the longest oligomers covered in this study, the linear increase
in activity at the active site at one terminus caused by the
addition of a monomer at the other terminus occurs over a
distance of ∼30 Å in unfolded oligomers and still over a
distance of ∼15 Å in the preferred foldamers (RIA-2, RIA-3,
Figure 7). The synergistic effect can be best imagined with a
molecular accordion. Anion−π interactions with the reactive
intermediates and transition states at one terminus and π−π
interactions with the newly added π surface at the other
terminus press the entire stack together to synergistically
amplify electron sharing and thus catalysis (Figure 7C).
The music made by these molecular accordions could not
sound better: Remote control of synergistic anion−(π)n−π
catalysis provides access to outstanding activities. This new
concept is also most attractive because it identifies unique
advantages of anion-π catalysis. Significant cation−(π)n−π
catalysis is inconceivable because the repulsive π-basic
aromatics will not fold into covalent face-to-face π stacks.
With regard to perspectives, one big question stands out:
Where does it stop?35 Violating the sublinear power laws of
diminishing returns with increasing length, linear increase of
activity with length promises access to more and more powerful
anion−(π)n−π catalysis on longer and longer π stacks. This
question is tantalizing because stack elongation beyond the
reported tetramers calls for significant synthetic efforts,
including strategies to avoid increasing solubility problems.
Studies in this direction are ongoing despite these significant
challenges, results will be reported in due course.
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.8b00809.
Detailed experimental procedures (PDF)
G
DOI: 10.1021/jacs.8b00809
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(10) Talukdar, P.; Bollot, G.; Mareda, J.; Sakai, N.; Matile, S. Chem. -
Eur. J. 2005, 11, 6525−6532.
(11) Lee, K. S.; Parquette, J. R. Chem. Commun. 2015, 51, 15653−
15656.
AUTHOR INFORMATION
■
Corresponding Author
*stefan.matile@unige.ch
(12) Schneebeli, S. T.; Frasconi, M.; Liu, Z.; Wu, Y.; Gardner, D. M.;
Strutt, N. L.; Cheng, C.; Carmieli, R.; Wasielewski, M. R.; Stoddart, J.
F. Angew. Chem., Int. Ed. 2013, 52, 13100−13104.
(13) Shao, H.; Gao, M.; Kim, S. H.; Jaroniec, C. P.; Parquette, J. R.
ORCID
Antonio Bauza: 0000-0002-5793-781X
́
Antonio Frontera: 0000-0001-7840-2139
Eric Vauthey: 0000-0002-9580-9683
Stefan Matile: 0000-0002-8537-8349
Chem. - Eur. J. 2011, 17, 12882−12885. (b) Pantos,̧ G. D.; Wietor, J.-
L.; Sanders, J. K. M. Angew. Chem., Int. Ed. 2007, 46, 2238−2240.
(14) (a) Rajdev, P.; Molla, M. R.; Ghosh, S. Langmuir 2014, 30,
1969−1976. (b) Mondal, T.; Sakurai, T.; Yoneda, S.; Seki, S.; Ghosh,
S. Macromolecules 2015, 48, 879−888.
Notes
The authors declare no competing financial interest.
(15) Avinash, M. B.; Govindaraju, T. Adv. Funct. Mater. 2011, 21,
3875−3882.
ACKNOWLEDGMENTS
■
(16) Goskulwad, S. P.; La, D. D.; Bhosale, R. S.; Al Kobaisi, M.;
Bhosale, S. V.; Bhosale, S. V. RSC Adv. 2016, 6, 39392−39395.
(17) (a) Wu, Y.; Frasconi, M.; Gardner, D. M.; McGonigal, P. R.;
Schneebeli, S. T.; Wasielewski, M. R.; Stoddart, J. F. Angew. Chem., Int.
Ed. 2014, 53, 9476−9481. (b) Gabutti, S.; Schaffner, S.; Neuburger,
We thank the NMR and the MS platforms for services, and the
University of Geneva, the Swiss National Centre of
Competence in Research (NCCR) Molecular Systems
Engineering, the NCCR Chemical Biology and the Swiss
NSF for financial support. A.B. and A.F. thank MINECO of
SPAIN for financial support (projects CTQ2014-57393-C2-1-P
and CTQ2017-85821-R, FEDER funds).
M.; Fischer, M.; Schafer, G.; Mayor, M. Org. Biomol. Chem. 2009, 7,
̈
3222−3229. (c) Sao, S.; Samanta, B. R.; Chaudhuri, D. RSC Adv.
2016, 6, 34350−34353.
(18) Takai, A.; Yasuda, T.; Ishizuka, T.; Kojima, T.; Takeuchi, M.
Angew. Chem., Int. Ed. 2013, 52, 9167−9171.
REFERENCES
■
(19) Gabriel, G. J.; Iverson, B. L. J. Am. Chem. Soc. 2002, 124,
15174−15175.
(1) (a) Dougherty, D. A. Acc. Chem. Res. 2013, 46, 885−893.
(b) Kennedy, C. R.; Lin, S.; Jacobsen, E. N. Angew. Chem., Int. Ed.
(20) Ikkanda, B. A.; Samuel, S. A.; Iverson, B. L. J. Org. Chem. 2014,
79, 2029−2037.
2016, 55, 12596−12624. (c) Brauer, T. M.; Zhang, Q.; Tiefenbacher,
̈
K. Angew. Chem., Int. Ed. 2016, 55, 7698−7701. (d) Hart-Cooper, W.
M.; Clary, K. N.; Toste, F. D.; Bergman, R. G.; Raymond, K. N. J. Am.
Chem. Soc. 2012, 134, 17873−17876.
(21) Miller, L. L.; Duan, R. G.; Tully, D. C.; Tomalia, D. A. J. Am.
Chem. Soc. 1997, 119, 1005−1010.
(22) Avestro, A.-J.; Gardner, D. M.; Vermeulen, N. A.; Wilson, E. A.;
Schneebeli, S. T.; Whalley, A. C.; Belowich, M. E.; Carmieli, R.;
Wasielewski, M. R.; Stoddart, J. F. Angew. Chem., Int. Ed. 2014, 53,
4442−4449.
(2) (a) Giese, M.; Albrecht, M.; Rissanen, K. Chem. Commun. 2016,
52, 1778−1795. (b) Chifotides, H. T.; Dunbar, K. R. Acc. Chem. Res.
2013, 46, 894−906. (c) Wang, D.-X.; Wang, M.-X. Chimia 2011, 65,
939−943. (d) Ballester, P. Acc. Chem. Res. 2013, 46, 874−884. (e) He,
Q.; Ao, Y.-F.; Huang, Z.-T.; Wang, D.-X. Angew. Chem., Int. Ed. 2015,
54, 11785−11790. (f) Liao, J.-Z.; Wu, C.; Wu, X.-Y.; Deng, S.-Q.; Lu,
C.-Z. Chem. Commun. 2016, 52, 7394−7397. (g) Liu, J.-J.; Guan, Y.-F.;
Jiao, C.; Lin, M.-J.; Huang, C.-C.; Dai, W.-X. Dalton Trans. 2015, 44,
5957−5960. (h) Kumar, S.; Ajayakumar, M. R.; Hundal, G.;
Mukhopadhyay, P. J. Am. Chem. Soc. 2014, 136, 12004−12010.
(23) Ashkenasy, N.; Horne, W. S.; Ghadiri, M. R. Small 2006, 2, 99−
102.
(24) (a) Gellman, S. H. Acc. Chem. Res. 1998, 31, 173−180. (b) Hill,
D. J.; Mio, M. J.; Prince, R. B.; Hughes, T. S.; Moore, J. S. Chem. Rev.
2001, 101, 3893−4012. (c) Schmitt, J.-L.; Stadler, A.-M.; Kyritsakas,
N.; Lehn, J.-M. Helv. Chim. Acta 2003, 86, 1598−1624. (d) Seebach,
D.; Beck, A. K.; Bierbaum, D. J. Chem. Biodiversity 2004, 1, 1111−
1239. (e) Guichard, G.; Huc, I. Chem. Commun. 2011, 47, 5933−5941.
(f) Zhang, D. W.; Zhao, X.; Hou, J. L.; Li, Z. T. Chem. Rev. 2012, 112,
5271−5316. (g) Ando, S.; Ohta, E.; Kosaka, A.; Hashizume, D.;
Koshino, H.; Fukushima, T.; Aida, T. J. Am. Chem. Soc. 2012, 134,
11084−11087. (h) Yashima, E.; Ousaka, N.; Taura, D.; Shimomura,
K.; Ikai, T.; Maeda, K. Chem. Rev. 2016, 116, 13752−13990. (i) Huo,
Y.; Zeng, H. Acc. Chem. Res. 2016, 49, 922−930. (j) Meudtner, R. M.;
Hecht, S. Angew. Chem., Int. Ed. 2008, 47, 4926−4930. (k) Saraogi, I.;
Hamilton, A. D. Chem. Soc. Rev. 2009, 38, 1726−1743. (l) Mendez-
Ardoy, A.; Markandeya, N.; Li, X.; Tsai, Y.-T.; Pecastaings, G.;
Buffeteau, T.; Maurizot, V.; Muccioli, L.; Castet, F.; Huc, I.; Bassani,
D. M. Chem. Sci. 2017, 8, 7251−7257. (m) Lamouroux, A.; Sebaoun,
L.; Wicher, B.; Kauffmann, B.; Ferrand, Y.; Maurizot, V.; Huc, I. J. Am.
Chem. Soc. 2017, 139, 14668−14675.
̀
(i) Bauza, A.; Mooibroek, T. J.; Frontera, A. ChemPhysChem 2015, 16,
2496−2517. (j) Yu, Z.; Erbas, A.; Tantakitti, F.; Palmer, L. C.;
Jackman, J. A.; Olvera de la Cruz, M.; Cho, M.-J.; Stupp, S. I. J. Am.
́
Chem. Soc. 2017, 139, 7823−7830. (k) Belanger-Chabot, G.; Ali, A.;
Gabbaï, F. P. Angew. Chem., Int. Ed. 2017, 56, 9958−9961.
(3) Zhao, Y.; Benz, S.; Sakai, N.; Matile, S. Chem. Sci. 2015, 6, 6219−
6223.
(4) Akamatsu, M.; Sakai, N.; Matile, S. J. Am. Chem. Soc. 2017, 139,
6558−6561.
́
(5) (a) Lopez-Andarias, J.; Frontera, A.; Matile, S. J. Am. Chem. Soc.
́
2017, 139, 13296−13299. (b) Zhang, X.; Liu, L.; Lopez-Andarias, J.;
Wang, C.; Sakai, N.; Matile, S. Helv. Chim. Acta 2018, 101, e1700288.
(6) (a) Frontera, A.; Quinonero, D.; Garau, C.; Costa, A.; Ballester,
P.; Deya, P. M. J. Phys. Chem. A 2006, 110, 9307−9309. (b) Frontera,
A.; Quinonero, D.; Costa, A.; Ballester, P.; Deya, P. M. New J. Chem.
2007, 31, 556−560.
(25) Jeannerat, D.; Ronan, D.; Baudry, Y.; Pinto, A.; Saulnier, J.-P.;
Matile, S. Helv. Chim. Acta 2004, 87, 2190−2207.
(7) (a) Al Kobaisi, M.; Bhosale, Si. V.; Latham, K.; Raynor, A. M.;
Bhosale, Sh. V. Chem. Rev. 2016, 116, 11685−11796. (b) Suraru, S.-L.;
(26) Ikemoto, K.; Kobayashi, R.; Sato, S.; Isobe, H. Angew. Chem., Int.
Ed. 2017, 56, 6511−6514.
Wurthner, F. Angew. Chem., Int. Ed. 2014, 53, 7428−7448.
̈
(27) Cohen, Y.; Avram, L.; Frish, L. Angew. Chem., Int. Ed. 2005, 44,
520−554.
(8) (a) Gorteau, V.; Bollot, G.; Mareda, J.; Perez-Velasco, A.; Matile,
S. J. Am. Chem. Soc. 2006, 128, 14788−14789. (b) Misek, J.; Vargas
Jentzsch, A.; Sakurai, S.; Emery, D.; Mareda, J.; Matile, S. Angew.
Chem., Int. Ed. 2010, 49, 7680−7683.
(28) Barros, T. C.; Brochsztain, S.; Toscano, V. G.; Filho, P. B.;
Politi, M. J. J. Photochem. Photobiol., A 1997, 111, 97−104.
(29) (a) Spano, F. C. Acc. Chem. Res. 2010, 43, 429−439.
(b) Hestand, N. J.; Spano, F. C. Acc. Chem. Res. 2017, 50, 341−350.
(30) (a) Wang, J.; Kulago, A.; Browne, W. R.; Feringa, B. L. J. Am.
Chem. Soc. 2010, 132, 4191−4196. (b) Fimmel, B.; Son, M.; Sung, Y.
(9) (a) Zhan, X.; Facchetti, A.; Barlow, S.; Marks, T. J.; Ratner, M. A.;
Wasielewski, M. R.; Marder, S. R. Adv. Mater. 2011, 23, 268−284.
(b) Sakai, N.; Sisson, A. L.; Burgi, T.; Matile, S. J. Am. Chem. Soc. 2007,
129, 15758−15759. (c) Sakai, N.; Matile, S. J. Am. Chem. Soc. 2011,
133, 18542−18545.
̈
M.; Grune, M.; Engels, B.; Kim, D.; Wurthner, F. Chem. - Eur. J. 2015,
̈
̈
H
DOI: 10.1021/jacs.8b00809
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
21, 615−630. (c) Giaimo, J. M.; Lockard, J. V.; Sinks, L. E.; Scott, A.
M.; Wilson, T. M.; Wasielewski, M. R. J. Phys. Chem. A 2008, 112,
2322−2330.
(31) (a) Greenfield, S. R.; Svec, W. A.; Gosztola, D.; Wasielewski, M.
R. J. Am. Chem. Soc. 1996, 118, 6767−6777.
(32) Yushchenko, O.; Licari, G.; Mosquera-Vazquez, S.; Sakai, N.;
Matile, S.; Vauthey, E. J. Phys. Chem. Lett. 2015, 6, 2096−2100.
(33) Cotelle, Y.; Benz, S.; Avestro, A.-J.; Ward, T. R.; Sakai, N.;
Matile, S. Angew. Chem., Int. Ed. 2016, 55, 4275−4279.
(34) Cotelle, Y.; Lebrun, V.; Sakai, N.; Ward, T. R.; Matile, S. ACS
Cent. Sci. 2016, 2, 388−393.
(35) (a) Le Bailly, B. A. F.; Byrne, L.; Clayden, J. Angew. Chem., Int.
̀
Ed. 2016, 55, 2132−2136. (b) Byrne, L.; Sola, J.; Boddaert, T.;
Marcelli, T.; Adams, R. W.; Morris, G. A.; Clayden, J. Angew. Chem.,
Int. Ed. 2014, 53, 151−155. (c) Diemer, V.; Fischer, L.; Kauffmann, B.;
Guichard, G. Chem. - Eur. J. 2016, 22, 15684−15692.
I
DOI: 10.1021/jacs.8b00809
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX