Effect of stereogenic centers on the self-sorting, depolymerization, and atropisomerization kinetics of porphyrin-based aggregates

Article abstract of DOI:10.1021/ja204543f

We present our results on the mixing of different porphyrin molecules in supramolecular assemblies. Herein, chiral amplification experiments reveal the subtle role of the structural (mis)match between these monomers. We show that according to the "sergean

Full text of DOI:10.1021/ja204543f

ARTICLE
pubs.acs.org/JACS
Effect of Stereogenic Centers on the Self-Sorting, Depolymerization,
and Atropisomerization Kinetics of Porphyrin-Based Aggregates
Floris Helmich, Maarten M. J. Smulders, Cameron C. Lee, Albertus P. H. J. Schenning,* and E. W. Meijer*
Institute for Complex Molecular Systems, Laboratory of Macromolecular and Organic Chemistry, Eindhoven University of Technology,
P.O. Box 513, 5600 MB Eindhoven, The Netherlands
S Supporting Information
b
ABSTRACT: We present our results on the mixing of different
porphyrin molecules in supramolecular assemblies. Herein,
chiral amplification experiments reveal the subtle role of the
structural (mis)match between these monomers. We show that
according to the “sergeant-and-soldiers” principle, a chiral
porphyrin “sergeant” efficiently mixes with achiral “soldiers”
in the same helical aggregate and strongly biases its handedness.
However, when we mix two porphyrin enantiomers in a
majority-rules experiment, no chiral amplification is observed
at all, which is due to their narcissistic self-sorting into conglomerate-like aggregates. The mixing between two enantiomers in the
same stack only occurs in a diluted-majority-rules experiment, in which enantiomeric mixtures of sergeants are diluted with achiral
soldiers. The different outcomes of these chiral amplification phenomena are verified by modeling studies that reveal high mismatch
penalties, which are ascribed to the high stereocenter loading of 12 methyl groups onto the monomers. Mixed-metal chiral
amplification experiments between copperÀ and zincÀporphyrins show the same distinction in their mixing behavior, which is
further supported by fluorescence measurements. The selective removal of chiral ZnÀporphyrins from these mixed-metal systems is
performed with the Lewis base quinuclidine that depolymerizes the ZnÀporphyrins upon axial ligation. This extraction process
proceeds at different time scales, depending on the mixed state: slow extraction kinetics for the mixed sergeant-and-soldiers and
diluted-majority-rules systems and an instant extraction for the phase-separated majority-rules system. By simultaneously
monitoring the supramolecular chirality during extraction, a chiral memory effect is observed for both mixed systems that show
slow extraction kinetics. For the sergeant-and-soldiers system, the remaining supramolecular backbone contains achiral monomers
only, which give rise to a long lasting chiral memory with slow, entropy-driven atropisomerization. Yet in case of the diluted-
majority-rules system, the remaining backbone contains a mixture of achiral and chiral monomers in its unpreferred helicity; giving
rise to a short chiral memory, in which the fast atropisomerization is enthalpy-driven due to the high mismatch penalty.
’
INTRODUCTION
polymerization of the appropriate ditopic monomers. Recent
studies reveal many similarities between macromolecules and
supramolecular polymers, in which the latter class is more
The self-assembly of organic molecules is an interesting1
bottom-up approach to synthesize nanometer-sized devices.
Their functionality is afforded by multiple components that
cooperate in these complex architectures, in which performance
is highly dependent on the organization of these interacting
moieties. The self-sorting of molecules provides valuable insights
into this positional control. In synthetic self-assembled systems,
self-sorting behavior is often the result of the directionality of
noncovalent interactions combined with steric constraints, as shown
for rotaxane, triangular, and dendritic structures. High-fidelity
behavior could also be achieved by enantiomeric self-recognition, in
which the expression of chiral information leads to narcissistic self-
sorting. Interestingly, the helicity of supramolecular polymers can
9
dynamic in nature. Here, we utilize the dynamic properties of
porphyrin-based aggregates to investigate if the self-sorting
behavior can be controlled by the supramolecular helicity.
Different approaches have been exploited in order to control
the helicity of racemic superstructures, for instance, by the employ-
2
1
0
11
ment of chiral auxiliaries, enantioselective physical stimuli, and
12
chiral amplification. In the latter approach, “sergeant-and-sol-
diers” experiments show that by mixing in a chiral “sergeant” with
achiral “soldier” monomers, the handedness of helical aggregates
containing predominantly achiral soldiers is biased by the chiral
sergeant. This results in a strong nonlinear chiroptical response,
which is also observed in many majority-rules and diluted-majority-
rules experiments. In these experiments, the helicity of an aggre-
gate comprising two opposite enantiomers is dominated by the
2
3
4
5
13
6
also be exploited to control the organization of molecules.
Helical supramolecular motifs are of great interest for the
understanding of the origins of homochirality of biopolymers and the
7
development of systems for chirotechnological applications. Im-
portant aspects herein are embodied by self-assembled, helical, one-
dimensional dye polymers that are the result of the supramolecular
Received: May 18, 2011
Published: June 29, 2011
r 2011 American Chemical Society
12238
dx.doi.org/10.1021/ja204543f
|J. Am. Chem. Soc. 2011, 133, 12238–12246

Journal of the American Chemical Society
ARTICLE
Scheme 1. Chiral/Achiral Amide-Functionalized TetraphenylÀZinc/CopperÀPorphyrins Used in This Study
enantiomer in excess, and in case of the diluted-majority-rules,
’ DESIGN OF THE SYSTEM
14
achiral soldiers are added to the system as well. The nonlinear
response of the optical activity in these chiral amplification
experiments is caused by the subtle interplay of secondary inter-
actions inside the supramolecular polymer. An enantiomeric im-
balance introduced by one of the monomers is sufficient to transfer
the chiral information to the supramolecular level, provided that
both monomers are assembled in the same supramolecular poly-
mer. For self-assembly of discotic molecules, the strength of these
The tetraphenyl porphyrins employed in this study have a
symmetrical functionalization pattern with amide linkages to
gallic moieties that contain achiral or branched chiral side chains
in the periphery that bias the preferred helicity of their
21
aggregates. The porphyrins are provided with either copper
or zinc ions in the center that allow fluorescence quenching
19
studies and selective depolymerizations. In methycyclohexane
(
MCH), all porphyrins show a hydrogen-bond-assisted and
15
16
secondary interactions depends on the location, number, and
highly cooperative self-assembly process into one-dimensional
H-type aggregates with a λmax ∼ 390 nm in the UVÀvis spectrum
that reveals the presence of extended cofacial aggregates, forming
supramolecular polymers with a low free monomer concen-
1
7
type of stereogenic centers in the chiral monomer. Over the years,
our group has studied these effects in detail, for instance, by using
different alkyl substituted, C -symmetrical benzene-1,3,5-tricarbox-
3
amides (BTAs). In chiral amplification studies, predictive models
were developed and employed to deduce the interplay of secondary
interactions in terms of two different energy penalties paid when
22
tration. The chiral porphyrins (R-Zn, S-Zn, and S-Cu) show
a strong CD effect in their Soret absorption, whereas achiral
porphyrins (A-Cu and A-Zn) are CD-silent. In this study, all
samples are prepared via chloroform injection of mixed mono-
mer solutions in MCH. The low amount of chloroform (0.5 vol
18
different monomers are mixed in a stack. Herein, the helix
reversal penalty (HRP) describes the energy penalty of a helix
reversal in the aggregate, whereas the mismatch penalty (MMP) is
related to the incorporation of a chiral monomer in a helical
aggregate of its unpreferred helicity.
%
) does not affect porphyrin aggregation or axial ligation of
ZnÀporphyrins with QND; hence, the injection technique
19
affords thermodynamically stable aggregates in an efficient way.
Previously, we described a novel approach to obtain a chiral
memory effect at the supramolecular level by a sergeant-and-
soldiers experiment between S-Zn and A-Cu (Scheme 1) and the
selective removal of S-Zn by the subsequent treatment with the
For clarity, we refer to the term “coaggregation” when different
porphyrin (co)monomers are mixed and share the same aggregate.
Previously, we reported on the sergeant-and-soldiers experi-
ments of mixed-metal porphyrins in dilute MCH solutions,
which revealed that chiral sergeant comonomers efficiently
transfer their own chirality to achiral comonomers, yielding a
19
Lewis base quinuclidine (QND). In this contribution, we study
chiral amplification with an expanded porphyrin library contain-
ing porphyrins with different stereochemistries in the periphery
and different metal centers. We show that within this library
highly distinctive mixed states can be achieved, which are
rationalized by modeling studies, a fully mixed sergeant-and-
soldiers system versus a narcissistically self-sorted majority-rules
system. The selective removal of ZnÀporphyrins by axial ligation
with QND reveals a remarkable difference in their kinetics, which
provides novel mechanistic insights into the mixing of porphyrins
inside supramolecular polymers. For the first time, we use the
diluted-majority-rules principle as a tool to mix opposite por-
phyrin enantiomers in a single stack by “dilution” of the system
with achiral monomers. Thereby, we tune the removal rate of the
chiral ZnÀporphyrin from the system. For the purpose of
constructing functional, multicomponent nanosized architec-
tures, we envision that it is highly important to explore these
chiral amplification phenomena since they could provide addi-
tional tools to place components at specified locations in the
assembly. Additionally, these studies could also provide novel
insights into the long lasting challenge of understanding the
19
helical aggregate with a single handedness. Modeling of these
À1
data revealed high HRPs on the order of ∼15 kJ mol , which
was attributed to the 4-fold hydrogen-bonding motif that causes a
high energy barrier to overcome a helix reversal in the aggregate.
In order to investigate the amplification of chirality in more
detail, we perform majority-rules experiments as well, in which
we can accurately determine MMPs. We start our studies with
ZnÀporphyrins where we describe the sergeant-and-soldiers
experiments between R/S-Zn and A-Zn, after which we perform
majority-rules experiments with both Zn-porphyrin enantio-
mers. In this paper, a special focus is maintained on the kinetic
stability of the created metastable assemblies.
’ CHIRAL AMPLIFICATION OF ZNÀPORPHYRINS AT
ROOM TEMPERATURE
Figure 1A shows the CD spectra of R-Zn:A-Zn coaggregates
À6
at 1.0 Â 10 M upon increasing the fraction of R-Zn. The shape
of the CD spectrum remains unchanged upon enrichment of
20
formation of racemates versus conglomerates in chiral drugs.
R-Zn, yet the CD intensity increases nonlinearly with the
1
2239
dx.doi.org/10.1021/ja204543f |J. Am. Chem. Soc. 2011, 133, 12238–12246

Journal of the American Chemical Society
ARTICLE
À6
Figure 1. Sergeant-and-soldiers experiments of R-Zn/S-Zn and A-Zn at 1.0 Â 10 M in MCH. (A) Full CD spectra of the R-Zn:A-Zn system
recorded at 20 °C. (B) The net helicity as a function of the fraction of sergeants R-Zn and S-Zn. Majority-rules experiments of R-Zn and S-Zn at 1.0 Â
À6
10
M in MCH. (C) Full CD spectra recorded at 20 °C of mixtures of different ee as indicated by the arrow from pure R-Zn through ee = 0% (1:1
À4
mixture) to pure S-Zn. (D) The net helicity as function of the ee. The blue lines in parts C and D show the overlay on both data sets at σ = 1.41 Â 10
À1
and ω = 0.135 (10.8 and 4.9 kJ mol ) for the HRP and MMP, respectively.
2
3
composition, going from the CD-silent state of pure A-Zn to a
saturated, CD-active state at ∼10% R-Zn. After conversion of the
Table 1. Overview of the Helix Reversal Penalty (HRP) and
Mismatch Penalty (MMP) after Fitting the Different Chiral
Amplification Experimental Data
CD spectra to g-values at λ = 391 nm, we construct the net
max
24
helicity versus the fraction sergeant (Figure 1B), which shows
mirror-image behavior of R-Zn with A-Zn and S-Zn with A-Zn.
Both sergeant-and-soldiers experiments were simultaneously
porphyrin
mixture
HRP
MMP
a
À1 b
À1 b
experiment (figure)
(kJ mol
)
(kJ mol )
18a
fitted with a modified model developed by van der Schoot.
S&S (1B)
MR (1D)
R/S-Zn:A-Zn
R-Zn:S-Zn
16.6
5.6
0.2
3.4
The least-squares fit on the experimental data in Figure 1B yields
two dimensionless energy penalties, σ and ω, that represent the
HRP and MMP, respectively, via σ = exp[À2HRP/RT] and ω =
exp[ÀMMP/RT]. From the contour plot of the sum of squared
average (1B/D)
S&S (3A)
R/A/S-Zn
10.8
14.3
15.4
9.5
4.9
S-Cu:A-Zn
0.7
MR (3B)
R-Zn:S-Cu
11.2
2.6
À1
residuals we can determine a HRP of ∼14 kJ mol quite
DMR (4B)
R-Zn:S-Cu:A-Cu
accurately; however, an accurate determination of the MMP is
a
S&S, sergeant-and-soldiers experiment; MR, majority-rules experi-
difficult because of the wide spread in the plot for this penalty
Supporting Information Figure S3A). The minimum of the
contour plot of the sum of squared residuals is found at 16.6 and
.2 kJ mol for the HRP and MMP, respectively (Table 1). The
high value for the HRP corresponds to the values we obtained in
the sergeant-and-soldiers experiments for the mixed metal
porphyrins described earlier, which indicates that the nature
of the center metal ion does not affect chiral amplification.
In order to obtain a more accurate determination of the MMP,
majority-rules experiments are performed between R-Zn and S-
Zn. Figure 1C shows the CD spectra of different enantiomeric
b
ment; DMR, diluted-majority-rules experiment. The values for the HRP
and MMP are obtained from the minimum of the contour plot of squared
residuals. All contour plots are shown in the Supporting Information.
21
(
À1
0
1
8a
which is in contrast to what has been observed for BTAs and
14a,25
other supramolecular polymers.
Here, saturated CD effects
19
were observed at ee values far below 100%, indicative of a
homochiral supramolecular system containing enantiomeric
1
8b
mixtures. The data are fitted with the majority-rules model,
and from a similar contour plot a wide spread in the HRP is
found, which makes an accurate estimation of this penalty
À6
mixtures at 1.0 Â 10 M, going from pure R-Zn (ee = 100%)
difficult (Supporting Information Figure S5A). The MMP on
À1
through R-Zn:S-Zn (1:1) (ee = 0%) to pure S-Zn (ee = À100%).
the other hand could be accurately determined at ∼4 kJ mol
,
The g-values at λ
= 392 nm are used to construct the net
while the overall minimum of the contour plot is found at 5.6 and
max
À1
21
helicity versus the ee (Figure 1D) and these data points show
almost a linear trend. Apart from the homochiral samples, the
linear relationship indicates that chiral amplification is absent,
3.4 kJ mol for the HRP and MMP, respectively (Table 1).
18a
Similarly as shown for BTA’s, we can obtain a single value
17b
for both energy penalties by considering the contour plots of
1
2240
dx.doi.org/10.1021/ja204543f |J. Am. Chem. Soc. 2011, 133, 12238–12246

Journal of the American Chemical Society
ARTICLE
À6
Figure 2. Cooling curves performed from the molecularly dissolved state at 65À20 °C at 60 °C/h in MCH at 1.0 Â 10 M. (A) T
e
analysis of the pure
R-Zn and S-Zn enantiomers probed by CD at 392 nm (black, left axis) and the enantiomeric (1:1) mixture at ee = 0% probed by UVÀvis at 390 nm (red,
right axis). (B) Same experiment for the R-Zn and S-Cu system, all probed by CD at 392 nm.
both sergeant-and-soldiers and majority-rules experiments. This
the high structural mismatch. With this finding, we exceed an
upper boundary for the mismatch penalty in chiral amplification,
which is in contrast with the lower boundary we observed earlier
for chiral deuterium-substituted BTAs that do not show a
majority-rules effect due to the lack of helical bias induced by
À1
overall minimum is estimated at 10.8 and 4.9 kJ mol for the
18
HRP and MMP, respectively. The simulation corresponding to
these energy penalties overlays both chiral amplification data sets
very well (Figure 1B/D, Table 1). These values are quite close to
each other and reveal considerable energy penalties for both the
HRP and MMP, indicating that it is highly unfavorable to both
reverse a stack from one handedness to the other and to
incorporate a chiral monomer in a stack of its opposite chirality.
In these analyses, both energy penalties seem to be related to
structural aspects of the monomer, with strong hydrogen bond-
17a
the isotope.
’ CHIRAL AMPLIFICATION OF MIXED-METAL ZN/
CUÀPORPHYRINS AT ROOM TEMPERATURE
In the present library, CuÀporphyrins are nonemissive,
1
8a,19
ing in the former
and a high structural mismatch originating
whereas ZnÀporphyrins are highly emissive, albeit their fluores-
19
from 12 methyl groups in the latter. As a consequence, the
likelihood that opposite enantiomers coaggregate would be
significantly reduced when both energy penalties are high in this
system, ultimately leading to narcissistic self-sorting into con-
glomerates comprising stacks with R-Zn or S-Zn only.
cence is partially quenched upon aggregation. Still, the remain-
ing aggregate fluorescence is more than sufficient for quenching
studies with CuÀporphyrins that act as an energy trap when the
two metal porphyrins are coaggregated, thereby monitoring the
mixing of porphyrins with fluorescence. Quenching of the
aggregate fluorescence was clearly observed in the previously
described sergeant-and-soldiers between S-Cu and A-Zn, which
showed that this process follows the amplification of chirality
(Figure 3A). It can be observed that a fully homochiral system is
obtained at 10% S-Cu and that ∼80% of the aggregate fluores-
The extent of self-sorting herein is tested by the investigation
of cooling curves performed on solutions of R-Zn and S-Zn at
À6
ee = 100%, 0%, and À100% at 1.0 Â 10 M. Figure 2A shows
the CD intensity at 392 nm upon cooling R-Zn (ee = 100%) and
S-Zn (ee = À100%) from the molecularly dissolved state at
28
6
5À20 °C at 60 °C/h, which reveals nonsigmoidal cooling curves,
cence is quenched. Since no metal dependency was found for
the chiral amplification behavior (vide supra, Table 1) and aggre-
gate fluorescence can be used as a probe to investigate porphyrin
coaggregation, we investigated the mixed-metal majority-rules
experiment between R-Zn and S-Cu.
22
indicating a highly cooperative self-assembly process. The
cooling curves for pure R-Zn and S-Zn are exact mirror images;
both have a sharp transition at T = 54.2 °C and saturate at 79
e
and À78 mdeg, respectively. After converting the CD data to the
dimensionless net helicity, both data sets are simultaneously
fitted with a temperature-dependent nucleationÀelongation
Analogously to the majority-rules experiments between R-Zn
and S-Zn, we prepared samples in MCH by injection of mono-
mer mixtures containing R-Zn and S-Cu at different ee in
2
6
model from which we estimate an enthalpy release of 130 kJ
À1 27
À6
mol
.
In the hypothesis of the formation of supramolecular
chloroform at 1.0 Â 10 M. Since we prepared enantiomeric
conglomerates of R-Zn and S-Zn, no coaggregation of opposite
enantiomers takes place, which implies an orthogonal self-
assembly process upon cooling the mixture at ee = 0%. When
mixtures with the copper analogue of the S-Zn enantiomer, we
observed slight differences in the UVÀvis and CD spectra
between S-Cu and S-Zn. As a result, R-Zn and S-Cu are not
exactly mirror images and the ee = 0% sample does not reveal a
À6
cooling this 1:1 mixture of R-Zn and S-Zn at 1.0 Â 10 M, no
21
CD activity is observed, yet when probing the self-assembly
completely CD-silent spectrum. Still, when the CD intensity at
392 nm is converted to the g-value, the obtained net helicity
versus ee shows a straight line (Figure 3B).
process by UVÀvis at λ = 390 nm, a similar, nonsigmoidal
max
trend is observed with T = 49.8 °C (Figure 2A). By using the
e
Van’t Hoff relation, in which we correlate the total concentration
The linear relationship between the net helicity and ee is even
more pronounced than for the homometal analogue of the
majority-rules experiments, which is also expressed by the energy
À1
to T at a fixed enthalpy value of 130 kJ mol , we obtain a
e
À7
porphyrin concentration of 5 Â 10 M that matches the T of
e
À1
the 1:1 mixture, exactly half the concentration of the enantio-
penalties obtained after fitting the data, 15.4 and 11.2 kJ mol
21
21
merically pure systems. This temperature analysis clearly
indicates that both porphyrin enantiomers prefer the formation
of their own stack with their own handedness, which is caused by
for the HRP and MMP, respectively (Table 1). These chiral
amplification data, in particular the high value for the MMP,
clearly indicates that coaggregation between enantiomers is highly
1
2241
dx.doi.org/10.1021/ja204543f |J. Am. Chem. Soc. 2011, 133, 12238–12246

Journal of the American Chemical Society
ARTICLE
À6
Figure 3. (A) Sergeant-and-soldiers experiments of S-Cu:A-Zn at 1.0 Â 10 M in MCH at 20 °C. (B) Majority-rules experiments of S-Cu:R-Zn at
À6
1
.0 Â 10 M in MCH at 20 °C. The net helicity (black, left axis) and fluorescence quenching (red, right axis) are plotted versus the fraction S-Cu (in A)
and ee (in B), and the blue lines represent the model fits.
À6
unfavorable, which can also be observed in the fluorescence
response. Figure 3B shows only a weak deviation from linearity in
the aggregate fluorescence, which indicates that quenching
occurs only to a small extent. This is probably caused by
concentration of 1.0 Â 10 M. Figure 4A shows the sergeant-
and-soldiers representation of this experiment with the g-value at
max = 390 nm versus the fraction sergeant, which represents
enantiomeric mixtures of R-Zn and S-Cu at ee = 25, 50, and
75%. The ee = 100% series herein represents the sergeant-and-
λ
21
(
monomeric) CuÀporphyrin energy acceptors that approach
the ZnÀporphyrin stacks within the F €o ster distance. Quenching
due to face-to-face stacking at the ZnÀCuÀporphyrin hetero-
junction would significantly enhance fluorescence quenching, as
we observe for the mixed-metal sergeant-and-soldiers experi-
ments (Figure 3A). In this majority-rules experiment, only ∼20%
quenching is observed at 10% S-Cu (ee = 80), which is
significantly less than the ∼80% observed in the former experi-
ment. This excludes a significant contribution to the quenching
soldiers of R-Zn and A-Cu, which is considered as the mirror
30
image of S-Zn and A-Cu (ee = À100%). Similar to the
sergeant-and-soldiers principle, the diluted-majority-rules experi-
ments show nonlinear amplification curves toward positive
g-values originating from the majority of R-Zn, in which higher ee’s
result in higher g-values. The data at 100% sergeant represents
the majority-rules experiment between R-Zn and S-Cu, which
3
1
shows considerably lower g-values at this wavelength. Related
to the sergeant-and-soldiers experiment at ee = 100%, it can also
be observed that a higher fraction of sergeant is necessary in order
to obtain a fully homochiral system when the ee is lowered. From
the amplification curves in Figure 4A it remains difficult to
estimate whether coaggregation between both enantiomers takes
place; however, when these diluted-majority-rules data are
represented in a majority-rules plot, an obvious deviation from
linearity is observed. This representation in Figure 4B shows the
linear response between the net helicity and ee at 0% A-Cu,
corresponding to the (“nondiluted”) majority-rules experiment
between R-Zn and S-Cu (Figure 3B). The net helicity for the
highly diluted samples, e.g., at 80% A-Cu, is obtained from the g-
values at 20% sergeant in Figure 4A. When these values at 25, 50,
and 75% ee are normalized to the sergeant-and-soldiers at 100%
ee, the obtained net helicity shows a positive deviation from
linearity. This qualitative piece of evidence for chiral amplifica-
tion is identical to the trend observed for most majority-rules
systems, indicating that both enantiomers coaggregate in one
We estimate from the fit on this deviation a HRP of
.5 kJ mol and a MMP of 2.6 kJ mol (Table 1). As
expected, the HRP remains high since it refers to a similar H-type
aggregate with strong hydrogen bonds that penalize a helix
reversal. However, the MMP has considerably been reduced
upon “dilution” with soldiers, which suggests that the structural
mismatch between the opposite enantiomers is suppressed by
the structural intrusion of achiral comonomers, hence resulting
in a coaggregate containing A-Cu, R-Zn, and S-Cu.
due to cofacial stacking. Additionally, the T analysis performed
e
on this system indicates the formation of supramolecular con-
glomerates of R-Zn and S-Cu as well. The cooling curves show
that R-Zn is more stable than S-Cu, as evidenced by their
2
9
difference in T at 53.7 and 51.6 °C, respectively (Figure 2B).
e
Analogously to the T analysis for R/S-Zn, the enthalpy release
e
for the pure enantiomers is obtained after fitting both cooling
curves, and we probe the cooling of the R-Zn:S-Cu (1:1) mixture
21
with CD at 392 nm. When cooling this mixture at ee = 0% at
À6
1
.0 Â 10 M, a positive CD response is observed at 49.6 °C;
corresponding to the aggregation of R-Zn. Upon further cooling,
the growth in optical activity ceases due to the aggregation of S-
Cu, which starts at 47.0 °C. In their Van’t Hoff analyses, both
transition temperatures correspond to the predicted T ’s at 5 Â
e
À7
1
0
M for R-Zn and S-Cu, respectively. At lower temperatures,
a nearly CD-silent state is obtained again due to the equal
abundance of both aggregates with opposite helicity.
The chiral amplification experiments between mixed-metal
porphyrins reveal the same features of porphyrin coaggregation
as observed for the ZnÀporphyrins; under the applied condi-
tions, coaggregation of achiral and chiral porphyrins is feasible,
unlike the coaggregation of opposite enantiomers. Considering
the ease of helical induction of sergeant comonomers to the
achiral backbone and the high energy barrier caused by the
structural mismatch between the opposite enantiomers, it is
highly interesting to investigate the possibility if the opposite
enantiomers coaggregate in a stack that contains achiral soldiers
as well. In order to investigate this possibility, we use the diluted-
14a,18a,25
stack.
9
À1
À1
21
18a,14c,14d
’ SELECTIVE REMOVAL OF ZNÀPORPHYRINS FROM
majority-rules principle,
in which enantiomeric mix-
CU/ZN-COAGGREGATES WITH QND
tures of R-Zn and S-Cu sergeants are “diluted” with A-Cu
soldiers. The samples are prepared by chloroform injections of
R-Zn/S-Cu/A-Cu monomer mixtures in MCH to a porphyrin
We reported on the selective depolymerization of ZnÀpor-
19
phyrins from coaggregated Cu/Zn systems, which is performed
1
2242
dx.doi.org/10.1021/ja204543f |J. Am. Chem. Soc. 2011, 133, 12238–12246

Journal of the American Chemical Society
ARTICLE
À6
Figure 4. Diluted-majority-rules of R-Zn + S-Cu and A-Cu at 1.0 Â 10 M in MCH at 20 °C. (A) g-Value at 390 nm versus the fraction [R-Zn + S-Cu]
sergeant for ee values of 25, 50, and 75%. The 100% ee data set equals the sergeant-and-soldiers experiment of R-Zn and A-Cu, which is considered as the
mirror image of S-Zn and A-Cu, and the boxed data points at 100% sergeant equal the majority-rules experiment of R-Zn and S-Cu at 390 nm
(Supporting Information Figure S8D). (B) The same experiment represented as the net helicity versus ee with the corresponding fits on the nondiluted
and 80% diluted state.
À6
Figure 5. Kinetic profile of the extraction of S-Zn of S-Zn:A-Cu (1:9) coaggregates at a porphyrin concentration of 1.0 Â 10 M and a QND
À2
concentration of 1.0 Â 10 M in MCH at 20 °C; the conversion of the S-Zn:QND as probed by UVÀvis at λmax ∼ 431 nm (black axis, left) and CD
(
red axis, right) versus time (A). Similar profile for the extraction of R-Zn from the R-Zn:S-Cu (1:1) system under identical conditions (B).
with quinuclidine (QND) that caps free ZnÀmonomers that
absorb at λmax = 431 nm; hence, by time-dependent monitoring
of this wavelength we can probe this removal process.
22
are involved in the aggregateÀmonomer equilibrium. Upon
ZnÀQND complexation, one porphyrin plane gets sterically
blocked, which inhibits the formation of hydrogen bonds be-
tween adjacent porphyrins. By this scavenging process, free
ZnÀmonomers are selectively extracted and removed from the
self-assembling system. The mixed-metal sergeant-and-soldiers
experiments between chiral and achiral comonomers reveal
coaggregation with amplification of chirality, whereas no ampli-
fication of chirality is observed in the mixed-metal majority-rules
reaction as a result of narcissistic self-sorting. By the combination
of both chiral amplification phenomena, coaggregation of en-
antiomers becomes feasible when the stacks are “diluted” with
achiral comonomers. By proper choice of center metal and side
chain identity, we can selectively extract ZnÀporphyrins from
Figure 5A shows the kinetic plot of the removal of S-Zn from
S-Zn:A-Cu (1:9) coaggregates, in which the receptive sergeant is
À7
present at a concentration of 1.0 Â 10 M. This time-dependent
UVÀvis measurement at 431 nm reveals extremely slow extrac-
tion kinetics; only after ∼15 h all ZnÀporphyrins are converted
to S-Zn:QND complexes, as evidenced by the UVÀvis absorp-
tion. Earlier, we attributed this slow kinetics to the low concen-
tration of receptive S-Zn comonomers and the slow exchange
19
between monomers and aggregates. Upon the extraction of
sergeant, the preferred helicity of the remaining A-Cu backbone
remains present and the full optical activity after ∼15 h indicates
a chiral memory effect.
In the narcissistically self-sorted majority-rules system with R-
Zn and S-Cu, we selected the sample at ee = 0%, which is CD-
19
these mixed-metal sergeant-and-soldiers and (diluted-)major-
ity-rules systems. By simultaneously monitoring the supramole-
cular chirality and extraction kinetics, we can study the observed
differences in these porphyrin coaggregates in more detail. In our
kinetic investigations on mixed-metal sergeant-and-soldiers and
silent due to equal abundance of R-Zn stacks and S-Cu stacks in
À7
solution. When we add QND to this system with 5.0 Â 10
M
receptive R-Zn, we observe instant depolymerization; ∼1 min
À2
after the addition of 1.0 Â 10 M QND the absorption band of
(
diluted-)majority-rules systems, we perform the ZnÀporphyrin
the R-Zn:QND band at 431 nm has fully developed (Figure 5B).
Similar kinetic behavior has been observed with pyridine titration
to S-Zn homoaggregates, indicating rapid depolymerization of
À6
extractions on a total porphyrin concentration of 1.0 Â 10
M
and we use a 10 000 times excess of QND. Under these condi-
tions, we fully remove all ZnÀporphyrins from the coaggregates
and ensure the exclusive formation of monomeric complexes that
22
R-Zn stacks. Upon the removal of R-Zn, it can be observed
that the CD response follows UV; going from a CD-silent state at
1
2243
dx.doi.org/10.1021/ja204543f |J. Am. Chem. Soc. 2011, 133, 12238–12246

Journal of the American Chemical Society
ARTICLE
À6
Figure 6. Kinetic profile of the extraction of R-Zn from the R-Zn:S-Cu:A-Cu (12.5:7.5:80) system at a porphyrin concentration of 1.0 Â 10 M and a
À2
QND concentration of 1.0 Â 10 M in MCH at 20 °C; the conversion of the R-Zn:QND as probed by UVÀvis at λmax ∼ 431 nm (black axis, left) and
CD (red axis, right) versus time (A). Schematic depiction of the selective removal of R-Zn from the diluted-majority-rules system with a reversal of the
supramolecular chirality (B).
ee = 0% to a CD active state originating from ee = À100% from
A-Cu (12.5:7.5:80, ee = 25% at 20% sergeant) coaggregate. At a
concentration of 1.25 Â 10 M, the R-Zn sergeant dominates
À7
À7
S-Cu stacks at 5.0 Â 10 M. The simultaneous response of CD
and UV reveals the absence of a chiral memory effect, and it
supports the self-sorted character of the R-Zn:S-Cu system.
Under the assumption that the dynamics for monomerÀ
aggregate exchange does not strongly depend on the metal/side
chain of the porphyrin, the remarkable difference in extraction
kinetics between the coaggregated S-Zn:A-Cu (1:9) and the
narcissistically self-sorted R-Zn:S-Cu (1:1) system can be ratio-
nalized by a shielding effect that occurs when ZnÀporphyrins are
coaggregated with CuÀporphyrins. This hypothesis can be
explained by the exposure of porphyrin monomers when they
detach from the end from the stack in order to preserve the
dynamic equilibrium between monomers and aggregates. In the
case of efficient coaggregation, the ZnÀporphyrins are homo-
geneously distributed over the mixed stack, and the monomers
that keep the whole coaggregate in the equilibrated state
comprise ZnÀ as well as CuÀporphyrins. As a result, ZnÀmo-
nomer exposure occurs to a considerably lower extent than when
the ZnÀporphyrins form their own homoaggregate, in which
only ZnÀmonomers are involved in the equilibrium. In the latter,
each monomer exposure event leads to monomer scavenging,
which causes a shift in the equilibrium, hence leading to depolym-
erization. In this way, CuÀporphyrins “protect” the ZnÀporphyr-
ins by consistent participation in the dynamic equilibrium. We
therefore postulate that ZnÀmonomers are removed more slowly
when they are coaggregated with CuÀmonomers.
the helical sense of this coaggregate, which shows a positive CD
signal at 392 nm (Figure 4A). Upon the removal of R-Zn from
this diluted-majority-rules system we observe slow extraction
kinetics; after ∼200 min all ZnÀporphyrins are converted to R-
Zn:QND complexes (Figure 6A). The extraction kinetics is in
line with the sergeant-and-soldiers experiment, yet in this coaggre-
gate, S-Cu and A-Cu comonomers are predominantly involved
in the dynamic equilibrium, thereby protecting R-Zn from axial
ligation (Figure 6B). Remarkably, when we consider the chirality
of the supramolecular backbone during extraction, we observe an
optical response between sergeant-and-soldiers and majority-
rules systems. Upon R-Zn removal, a direct response in CD is
observed; however, the CD response does not follow the
absorbance of R-Zn:QND. When all R-Zn comonomers are
removed at ∼200 min, the chirality is halfway the transition from
a positive to a negative helical sense, which originates from the
remaining S-Cu:A-Cu (7.5:80) coaggregate. The observed chiral
memory effect fully disappears after 1 day, which is in contrast
with chiral memory effect of the remaining A-Cu backbone in the
1
9
S-Zn:A-Cu (1:9) coaggregate that lasts over months.
The difference between these memory effects is obviously
related to frustration caused by S-Cu, which is initially incor-
porated in a stack of the wrong handedness. Therefore, the
initial R-Zn:S-Cu:A-Cu (12.5:7.5:80) coaggregates are higher
in energy than the final state, where R-Zn has been removed,
which causes a rapid, enthalpy-driven atropisomerization.
Since A-Cu lacks the preference for any helical direction, this
energy difference caused by initial structural frustration is
absent; as a result, the atropisomerization is entropy driven
and slow due to the conformational inertness of the A-Cu
backbone.
The chiral amplification experiments on the diluted-majority-
rules system indicate that coaggregation of enantiomers is enab-
led upon “dilution” with soldier comonomers that encounter a
lower structural mismatch (Figure 4B). This behavior between
sergeant-and-soldiers and majority-rules is also observed in its
extraction process with QND, as we show for the R-Zn:S-Cu:
1
2244
dx.doi.org/10.1021/ja204543f |J. Am. Chem. Soc. 2011, 133, 12238–12246

Journal of the American Chemical Society
ARTICLE
’
CONCLUSION
of stereochemical control may also provide mechanistic insights
into dynamic mixing.
The coaggregation behavior of porphyrins with different
metals and side chains was investigated by chiral amplification.
Efficient coaggregation was observed between achiral and chiral
porphyrins, as evidenced by a strong sergeant-and-soldiers effect.
On the other hand, no chiral amplification was observed in the
majority-rules experiment. The distinctive behavior in chiral
amplification was quantified by modeling studies that revealed
a combination of high helix reversal and high mismatch penalties,
’
ASSOCIATED CONTENT
S
Supporting Information. The synthesis of the R-Zn and
b
UVÀvis/CD/PL spectra of all chiral amplification experiments
with corresponding contour plots and calculations on cooling
curves. This material is available free of charge via the Internet at
http://pubs.acs.org.
À1
on the order of 10 and 5 kJ mol , respectively. The high
mismatch penalty explains the absence of a majority-rules effect;
the conformational mismatch between the opposite enantiomers
does not allow coaggregation. Rather, narcissistic self-sorting of
the enantiomers is observed, as shown by cooling studies.
The chiral amplification behavior does not depend on the
metal center inside the porphyrin. Mixed-metal sergeants-and-
soldiers and majority-rules studies also showed highly distinctive
mixing, which was also supported with the fluorescence quench-
ing between ZnÀ and CuÀporphyrins. The limits of coaggrega-
tion were further explored with diluted-majority-rules measure-
ments, which showed a chiral amplification effect, indicating that
both enantiomers coaggregate when the stacks contain achiral
comonomers as well. Compared to the mixed-metal majority-
rules experiment, lower mismatch penalties were estimated after
fitting the diluted-majority-rules experiment.
’ AUTHOR INFORMATION
Corresponding Author
A.P.H.J.Schenning@tue.nl; E.W.Meijer@tue.nl
’
ACKNOWLEDGMENT
This work was supported by the Council of Chemical Sciences
of The Netherlands Organization for Scientific Research
(CW-NWO). We thank Marko Nieuwenhuizen and Peter Korevaar
for stimulating discussions, and Dr. Martin Wolffs and Juli €e n van
Velthoven for early contributions to the project.
’
REFERENCES
Besides chirality and fluorescence, we performed kinetic
experiments in order to investigate the coaggregation behavior
in more detail. Studies on the selective extraction of ZnÀpor-
phyrins from mixed-metal sergeant-and-soldiers and (diluted-)
majority-rules systems showed remarkable differences in their
rate of chiral ZnÀporphyrin removal. For the coaggregated
systems, slow extraction kinetics was found, whereas an instant
ZnÀporphyrin removal was found for the narcissistically self-
sorted system. A mechanistic explanation for this rate difference
relies on the susceptibility of monomers that are in equilibrium
with the (co-)aggregates; CuÀporphyrins “protect” ZnÀpor-
phyrins by their monomer exposure in order to retain the
equilibrium of the coaggregate.
The chiral response upon the removal of chiral ZnÀporphyr-
ins reveals a chiral memory effect for the coaggregated systems
due to the kinetic inertness of the achiral backbone. For the
sergeant-and-soldiers system, a long lasting memory is found,
whereas a short memory was found for the diluted-majority-
rules system. The memory difference is caused by the opposite
CuÀenantiomer that remains coaggregated with the achiral
backbone in the latter, which frustrates the helical imprint and
causes an energy difference; hence, a fast, enthalpy-driven
atropisomerization is observed. Yet for the sergeant-and-sol-
diers system, this energy difference before and after selective
extraction is not present, which results in a slow, entropy-driven
atropisomerization.
(
1) (a) Kay, E. R.; Leigh, D. A.; Zerbetto, F. Angew. Chem., Int. Ed.
2
2
007, 46, 72–191. (b) Schenning, A.; Meijer, E. W. Chem. Commun.
005, 3245–3258.
(2) (a) Jiang, W.; Schalley, C. A. Proc. Natl. Acad. Sci. U. S. A. 2009,
1
06, 10425–10429. (b) Jiang, W.; Winkler, H. D. F.; Schalley, C. A.
J. Am. Chem. Soc. 2008, 130, 13852–13853.
3) (a) Schmittel, M.; Mahata, K. Chem. Commun. 2010,
(
46, 4163–4165. (b) Mahata, K.; Saha, M. L.; Schmittel, M. J. Am. Chem.
Soc. 2010, 132, 15933–15935.
(
4) Corbin, P. S.; Lawless, L. J.; Li, Z. T.; Ma, Y. G.; Witmer, M. J.;
Zimmerman, S. C. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 5099–5104.
5) (a) Hutin, M.; Cramer, C. J.; Gagliardi, L.; Shahi, A. R. M.;
Bernardinelli, G.; Cerny, R.; Nitschke, J. R. J. Am. Chem. Soc. 2007,
29, 8774–8780. (b) Shi, X. D.; Fettinger, J. C.; Davis, J. T. J. Am. Chem.
Soc. 2001, 123, 6738–6739.
6) (a) Taylor, P. N.; Anderson, H. L. J. Am. Chem. Soc. 1999,
21, 11538–11545. (b) Wu, A. X.; Isaacs, L. J. Am. Chem. Soc. 2003,
(
1
(
1
125, 4831–4835.
(7) (a) Mateos-Timoneda, M. A.; Crego-Calama, M.; Reinhoudt,
D. N. Chem. Soc. Rev. 2004, 33, 363–372. (b) Okamoto, Y.; Nakano, T.
Chem. Rev. 1994, 94, 349–372. (c) Verbiest, T.; Van Elshocht, S.;
Kauranen, M.; Hellemans, L.; Snauwaert, J.; Nuckolls, C.; Katz, T. J.;
Persoons, A. Science 1998, 282, 913–915. (d) Maeda, K.; Yashima, E.
Top. Curr. Chem. 2006, 265, 47–88.
(
8) Lee, C. C.; Grenier, C.; Meijer, E. W.; Schenning, A. Chem. Soc.
Rev. 2009, 38, 671–683.
(9) Yashima, E.; Maeda, K.; Nishimura, T. Chem.—Eur. J. 2004,
10, 42–51.
(10) (a) George, S. J.; Tomovic, Z.; Smulders, M. M. J.; de Greef,
T. F. A.; Leclere, P.; Meijer, E. W.; Schenning, A. Angew. Chem., Int. Ed.
007, 46, 8206–8211. (b) Fenniri, H.; Deng, B. L.; Ribbe, A. E. J. Am.
Chem. Soc. 2002, 124, 11064–11072. (c) Ishi-i, T.; Crego-Calama, M.;
Timmerman, P.; Reinhoudt, D. N.; Shinkai, S. Angew. Chem., Int. Ed.
002, 41, 1924–1929. (d) Oda, R.; Huc, I.; Schmutz, M.; Candau, S. J.;
MacKintosh, F. C. Nature 1999, 399, 566–569. (e) El-Hachemi, Z.;
Mancini, G.; Ribo, J. M.; Sorrenti, A. J. Am. Chem. Soc. 2008,
As evidenced by the chiral amplification phenomena, we
achieved highly distinctive mixed states within the porphyrin
assemblies by only subtle peripherical changes in the monomer.
Within the same library, we can use the diluted-majority-rules
principle as a tool to control the extent of mixing. For the purpose
of constructing functional, multicomponent nanoarchitectures,
we envision that it is highly interesting to explore these chiral
amplification phenomena, since they could provide additional
tools to control stoichiometry and place the components at
specified locations in the assembly. On the other hand, this type
2
2
1
30, 15176–15184. (f) Jiang, S. G.; Zhang, L.; Liu, M. H. Chem.
Commun. 2009, 6252–6254. (g) Rosaria, L.; D’Urso, A.; Mammana,
A.; Purrello, R. Chirality 2008, 20, 411–419. (h) Toyofuku, K.; Alam,
1
2245
dx.doi.org/10.1021/ja204543f |J. Am. Chem. Soc. 2011, 133, 12238–12246

Journal of the American Chemical Society
ARTICLE
M. A.; Tsuda, A.; Fujita, N.; Sakamoto, S.; Yamaguchi, K.; Aida, T.
Angew. Chem., Int. Ed. 2007, 46, 6476–6480.
(
11) (a) Janssen, P. G. A.; Ruiz-Carretero, A.; Gonzalez-Rodriguez,
D.; Meijer, E. W.; Schenning, A. Angew. Chem., Int. Ed. 2009,
8, 8103–8106. (b) El-Hachemi, Z.; Arteaga, O.; Canillas, A.; Crusats,
4
J.; Escudero, C.; Kuroda, R.; Harada, T.; Rosa, M.; Ribo, J. M. Chem.—
Eur. J. 2008, 14, 6438–6443. (c) Tejedor, R. M.; Oriol, L.; Serrano, J. L.;
Urena, F. P.; Gonzalez, J. J. L. Adv. Funct. Mater. 2007, 17, 3486–3492.
(
12) Palmans, A. R. A.; Meijer, E. W. Angew. Chem., Int. Ed. 2007,
6, 8948–8968.
13) (a) Palmans, A. R. A.; Vekemans, J.; Havinga, E. E.; Meijer,
4
(
E. W. Angew. Chem., Int. Ed. 1997, 36, 2648–2651. (b) Green, M. M.;
Reidy, M. P.; Johnson, R. J.; Darling, G.; Oleary, D. J.; Willson, G. J. Am.
Chem. Soc. 1989, 111, 6452–6454.
(14) (a) van Gestel, J.; Palmans, A. R. A.; Titulaer, B.; Vekemans, J.;
Meijer, E. W. J. Am. Chem. Soc. 2005, 127, 5490–5494. (b) Green, M. M.;
Garetz, B. A.; Munoz, B.; Chang, H. P.; Hoke, S.; Cooks, R. G. J. Am.
Chem. Soc. 1995, 117, 4181–4182. (c) Wilson, A. J.; van Gestel, J.;
Sijbesma, R. P.; Meijer, E. W. Chem. Commun. 2006, 4404–4406.
(
d) Selinger, J. V.; Selinger, R. L. B. Macromolecules 1998, 31, 2488–2492.
15) Stals, P. J. M.; Smulders, M. M. J.; Martin-Rapun, R.; Palmans,
A. R. A.; Meijer, E. W. Chem.—Eur. J. 2009, 15, 2071–2080.
16) (a) Iavicoli, P.; Xu, H.; Feldborg, L. N.; Linares, M.; Paradinas,
(
(
M.; Stafstrom, S.; Ocal, C.; Nieto-Ortega, B. L.; Casado, J.; Navarrete,
J. T. L.; Lazzaroni, R.; De Feyter, S.; Amabilino, D. B. J. Am. Chem. Soc.
2
010, 132, 9350–9362. (b) Smulders, M. M. J.; Stals, P. J. M.; Mes, T.;
Paffen, T. F. E.; Schenning, A. P. H. J.; Palmans, A. R. A.; Meijer, E. W.
J. Am. Chem. Soc. 2010, 132, 620–6.
(17) (a) Cantekin, S.; Balkenende, D. W. R.; Smulders, M. M. J.;
Palmans, A. R. A.; Meijer, E. W. Nature Chem. 2011, 3, 42–46. (b) Veld,
M. A. J.; Haveman, D.; Palmans, A. R. A.; Meijer, E. W. Soft Matter 2011,
7, 524–531.
(
18) (a) Smulders, M. M. J.; Filot, I. A. W.; Leenders, J. M. A.; van
der Schoot, P.; Palmans, A. R. A.; Schenning, A. P. H. J.; Meijer, E. W.
J. Am. Chem. Soc. 2010, 132, 611–619. (b) van Gestel, J. Macromolecules
2
004, 37, 3894–3898.
19) Helmich, F.; Lee, C. C.; Schenning, A. P. H. J.; Meijer, E. W.
J. Am. Chem. Soc. 2010, 132, 16753–16755.
20) Wang, Y. L.; Chen, A. M. Org. Process Res. Dev. 2008, 12,
82–290.
(
(
2
(21) See Supporting Information
(22) Helmich, F.; Lee, C. C.; Nieuwenhuizen, M. M. L.; Gielen, J. C.;
Christianen, P. C. M.; Larsen, A.; Fytas, G.; Leclere, P.; Schenning, A.;
Meijer, E. W. Angew. Chem., Int. Ed. 2010, 49, 3939–3942.
(
23) The CD spectrum of A-Zn shows a small CD effect, which is
most likely due to enantiomeric impurities.
24) The net helicity is obtained from dividing the g-value at a certain
(
composition by the maximum g-value at λmax. The g-value is the anisotropy
factor, which is calculated via g=Δε/ε=CD-effect [mdeg]/(32 980 Â OD).
(25) (a) Jin, W.; Fukushima, T.; Niki, M.; Kosaka, A.; Ishii, N.; Aida,
T. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 10801–10806. (b) Lohr, A.;
Wurthner, F. Angew. Chem., Int. Ed. 2008, 47, 1232–1236.
(
26) Smulders, M. M. J.; Schenning, A. P. H. J.; Meijer, E. W. J. Am.
Chem. Soc. 2008, 130, 606–610.
27) The enthalpy value differs slightly from the one we published
(
22
earlier, which is most likely related to monomer purity and the
preparation method with chloroform.
(
28) The percentage of quenching is determined by the difference in
luminescence between pure A-Zn (100%) and pure S-Cu (0%) at the
21
emission maximum at λmax = 604 nm.
29) The difference in T for R-Zn in Figure 2A,B is caused by a
concentration effect, as evidenced by a lower absorption for the latter at λmax
30) Considering the mirror-image behavior in chiral amplification
(
e
.
(
of R/S-Zn with A-Zn (Figure 1B), we assume the same behavior for the
amplification of A-Cu.
(31) The CD spectra of amplified, achiral porphyrins observed in
sergeant-and-soldiers experiments have a different shape than the CD
21
spectra of intrinsically chiral aggregates.
1
2246
dx.doi.org/10.1021/ja204543f |J. Am. Chem. Soc. 2011, 133, 12238–12246