Full Paper
similar in trend, albeit much more significant, compared to the
cases of PEMs. Interestingly, the aggregation properties of
PVMs are strongly influenced by the polarity of substituents
and whether the functional groups can donate or withdraw
electron density from the macrocyclic core appears less impor-
tant. Pronounced concentration-dependent 1H NMR chemical
shift changes were observed for PVMs 19–23 with either elec-
tron-donating or -withdrawing substituents (Figure 2), but not
cies. We assume that PVM 21 and PVM 23 are present as mon-
omeric species at the lowest studied concentration (0.13 mm,
and 0.66 mm, respectively). Our study (for details, see the Sup-
porting Information) shows that the number of aggregates of
both PVM 21 and PVM 23 are around two at the highest stud-
ied concentration (4.08 mm and 10.5 mm, respectively). There-
fore, it is mainly monomer–dimer aggregation that occurs and
formation of higher order aggregates beyond dimers is insig-
nificant in the concentration range we studied for PVM 21 and
23. We assume that the aggregations of other PVMs (19, 20,
22, 24), which show weaker self-association in CDCl3 than 21
and 23, are mainly limited to dimerization.
Since DOSY experiments support monomer–dimer equilibri-
um as the major aggregation process, the concentration-de-
pendent chemical shift data of endo- and exo-annular protons
of PVMs were analyzed by using the monomer–dimer model
[Eq. (1)][41]
!
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ð1 À 8Kassocct þ 1Þ
dobs ¼ dmonomer À D 1 þ
ð1Þ
4Kassocct
where dobs is the observed chemical shift, dmonomer is the
chemical shift of the monomer, Kassoc is the association con-
stant, ct is the molar concentration of the PVM, and D is the
chemical shift difference between monomer and dimer. Self-as-
sociation constants of PVMs were then extracted by using non-
linear least-squares regression method from the concentration-
dependent chemical shifts of exo-annular (Ha) and the endo-an-
nular (Hb) protons of PVMs at 296 K. We were able to identify
the best values of association constants, which give the small-
est standard deviations in the curve fitting, for macrocycles
showing significant aggregation (see the Supporting Informa-
tion, Figures S15–S18). The self-association constants calculated
from endo- and exo-annular protons of the PVMs are in good
agreement, within experimental error. PVM 21, functionalized
with “reverse” ester (OCOnC9H19), gave the strongest self-asso-
ciation constants (Table 1). We also observed strong aggrega-
tion of PVMs 20 (R=OnC4H9) and 23 (R=COOnC4H9). Surpris-
ingly, PVM 24, substituted with tert-butyl ester, also underwent
decent aggregation, although Kassoc was decreased four-fold
compared to that of PVM 23, substituted with n-butyl esters.
The length of alkyl chains attached to the ether or ester
groups negatively influenced the aggregation of PVMs, likely
due to the larger entropic loss upon aggregation. We observed
much stronger aggregation of PVMs 20 (R=OnC4H9) and 23
(R=COOnC4H9) compared to PVMs 19 (R=OnC10H21) and 22
(R=COOnC10H21), respectively.
Figure 2. The concentration-dependent chemical shifts of exo-annular pro-
1
tons (Ha) of PVMs 17–24. H NMR spectra were recorded in CDCl3 at 296 K.
for PVMs 17 and 18, containing nonpolar alkyl substituents. At
ambient temperature, the chemical shift of two anisochronous
aromatic protons of PVM 21 (R=OCOnC9H19) changed from
7.49 ppm to 7.31 ppm and from 7.03 ppm to 6.89 ppm respec-
tively, as the concentration was increased from 0.39 to 6.2 mm
(Figure 1). The changes in the chemical shift of exo-annular (Ha)
and endo-annular (Hb) protons were similar in both trend and
magnitude for the same PVM. The chemical shifts of aliphatic
protons remain unchanged over the same concentration
range. The aggregation behaviors of PVMs are also tempera-
ture dependent. When the temperature was raised, the reso-
nance peaks of exo-annular and endo-annular protons of PVM
21 were shifted to lower field, which indicates that dissociation
occurs at elevated temperature (see the Supporting Informa-
tion, Table S7 and Figure S12).
The aggregation of PVMs in solution was also studied by
fluorescence spectroscopy. We observed significant fluores-
cence quenching in the solution of PVM 23 when the solvent
was changed from chloroform to acetonitrile. However, such
fluorescence quenching behavior was absent in the emission
spectra of non-aggregating PVM 18 under the same experi-
mental conditions (see the Supporting Information, Figure S11).
These results further suggest the fluorescence quenching ob-
served for PVM 23 is attributed to its aggregation.
Substituent effects
To further quantify the self-association constants, we first es-
timated the size of macrocycle aggregates from the diffusion
coefficient of the stacked species in solution. The diffusion co-
efficient was obtained through diffusion-ordered spectroscopy
(DOSY) experiments. We used PVM 21 and PVM 23, which
show the strongest aggregation, as the representative exam-
ples to estimate the number of macrocycles per stacked spe-
Previously, Moore and co-workers reported that PEMs undergo
considerable aggregation only when functionalized with elec-
tron-withdrawing groups (e.g., COOnC4H9), and no obvious ag-
gregation when functionalized with electron-donating groups
(e.g., OnC4H9; Table 1). This is in good agreement with the
polar/p model,[26] which predicts that the introduction of elec-
tron-withdrawing substituents reduces the electron density of
Chem. Eur. J. 2015, 21, 16935 – 16940
16937
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim