Angewandte
Chemie
observed in the 50–500 nm range and are inferred to be
polymer aggregates;[3] these aggregates were also observed by
multi-angle static light scattering (SLS) (SI Figure S11).
Polymer 1 dissolved in dimethyl formamide (DMF) to
give a deep red solution. The strong absorption peaks
observed in the UV/Vis spectra of 1 up to 400 nm were
attributed to the p-p* transitions of the polymer backbone
(N^N) and POP ligands, whereas absorption beyond 400 nm
was attributed to the metal-to-ligand charge transfer
(MLCT)[16–18] (Figure 1). The polymer has a photolumines-
several months. Scanning electron microscopy images of the
gels showed the formation of fibrous aggregates ca. 10 nm in
width and hundreds of micrometers in length (SI Figure S19).
The sol–gel transition was followed in a 1% [D6]DMSO
solution by variable temperature NMR (VT-NMR) spectros-
copy in order to elucidate the molecular basis for gelation (SI
1
Section 2.4). The broad H NMR spectrum of the polymer
disappeared upon heating to 908C due to gel formation[10] and
was replaced by sharp peaks corresponding to those of
a [Cu(POP)]+ complex. Comparison with the integrated
intensities of peaks from an added adamantane reference
indicated total displacement of the [Cu(POP)]+ complex from
the polymer upon gel formation. Minimal change was
observed in the NMR spectra of the DMSO sample after
standing at 258C for three days. The transition was further
investigated by 31P NMR and mass spectrometry, which
provided data consistent with the liberation of [Cu(POP)]+
upon heating. Similar behavior was observed to occur in DMF
by NMR spectroscopy (SI Figure S15). Temperatures greater
than 1408C were required to induce gel formation, and the
samples had to be heated outside of the spectrometer and
measured rapidly for [Cu(POP)]+ formation to be observed.
After leaving the sample to cool to room temperature for
30 min the spectrum had reverted to that of the polymer
solution, and after 24 h it matched that of the sample before
heating. Repeated heating and cooling of the sample resulted
in the gradual build-up of an insoluble yellow precipitate.
The formation of polymer 1 was driven by the tendency of
copper(I) to form heteroleptic complexes of the type
CuIP2N2.[8,16] However, heating this polymer in high boiling,
coordinating solvents such as DMSO and DMF results in
Figure 1. UV/Vis absorption (black) and photoluminescence (red) of
the polymer 1 in solution and photoluminescence spectrum of the
polymer in the gel state (orange). The blue line shows the photo-
luminescence of the isolated gel matrix. Inset: Photograph showing
reversible formation of a yellow gel upon heating a red polymer
solution of 1 and the reversible gel forming upon heating.
ꢀ
breaking of the weaker Cu N bonds and release of [Cu-
(POP)]+ into the solution leaving behind the imine-containing
polymer backbone [Eq. (1)].
cence (PL) maximum at 780 nm, also attributed to the MLCT
state. Heating a solution of the polymer (1 wt.%) to temper-
atures above 1608C resulted in its transformation to an
opaque yellow state which did not flow when its container was
inverted, indicating gel formation (Figure 1-inset). The PL
spectrum of the gel showed an emission maximum at 580 nm,
significantly blue-shifted compared to the polymer solution.
The color change, shift in photoluminescence, and gel
formation reversed upon cooling the sample to room temper-
ature. This type of reversible gel formation upon temperature
rise[4,8–10] is the opposite behavior to that observed for most
gelators, which undergo a sol–gel transition upon cooling. We
had previously reported a similar effect in a series of related
polymers, however, the high temperature of gel formation and
poor solubility of the polymers inhibited further study or
device fabrication.[3c]
Heat-set gel formation was also observed to occur in
DMSO at lower temperatures than in DMF, allowing the
process to be monitored using rheometry. When the sample
was heated to 1408C, the elastic modulus (G’) became an
order of magnitude greater than the viscous modulus (G’’)
and behaved linearly over a wide range of frequencies,
indicating gel formation (SI Section S3.1). The DMSO gels
remained stable in the rheometer upon cooling to room
temperature and were observed to be stable in a vial over
½CuðN^NÞðPOPÞꢁþ Ð ½ðN^NÞꢁn þ n ½CuðPOPÞꢁþ
ð1Þ
n
Without the bulky phosphine groups and cationic charge
of the copper, the polymers are inferred to aggregate into the
fibrous structure of the gel matrix observed by scanning
electron microscopy (SEM; Figure S19). In DMF, cooling of
the solution appears to allow the [Cu(POP)]+ groups to
reattach to the vacant pyridyl imine sites, restoring the
polymers and dissolving the gel. In DMSO, we infer that the
lower solubility of the polymers results in phase separation
due to greater aggregation, rendering the process irreversible.
This mechanism is different than the one that we proposed to
explain the heat-set gelation in a related system,[3c] in that we
do not propose here that CuI remains in the polymer.
In order to characterize the PL observed at high temper-
atures, the [Cu(POP)]+ complex and metal-free polymer
backbone [(N^N)] were isolated and their photolumines-
cence spectra were measured separately. Compound 3, [Cu-
(POP)(CH3CN)2]BF4, was synthesized according to a modi-
fied literature procedure (SI Section S1.4) and was found to
show no luminescence. The demetallated polymer backbone
was isolated by forming a gel (1.5 wt.% 1 in DMSO), then
washing the gel repeatedly with DMSO until no [Cu(POP)]+
was detected in the washings by NMR. The spectrum provides
Angew. Chem. Int. Ed. 2014, 53, 8388 –8391
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8389