Constitutional dynamic systems of ionic and molecular liquids

Article abstract of DOI:10.1039/b901899a

The ionic liquids N-methylpyridinium-2-carboxaldehyde Tf₂N ^- and N-methylpyridinium-3-carboxaldehyde Tf₂N^- form readily-reversible covalent bonds with protic nucleophiles, creating all-liquid constitutionally dynamic materials. The Royal Society of Chemistry 2009.

Full text of DOI:10.1039/b901899a

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Constitutional dynamic systems of ionic and molecular liquidswz
Morgan D. Soutullo, Richard A. O’Brien, Kyle E. Gaines and James H. Davis, Jr.*
Received (in Austin, TX, USA) 30th January 2009, Accepted 3rd March 2009
First published as an Advance Article on the web 23rd March 2009
DOI: 10.1039/b901899a
The ionic liquids N-methylpyridinium-2-carboxaldehyde Tf2Nꢀ
and N-methylpyridinium-3-carboxaldehyde Tf2Nꢀ form readily-
reversible covalent bonds with protic nucleophiles, creating
all-liquid constitutionally dynamic materials.
regenerated by the formation of bis(hemialdal). In some cases
this net ‘free’ H2O can be independently quantified by direct
NMR observation.
As expected, the proportion of aldehyde transformed in
these systems increases steadily as the amount of H2O is
increased (Fig. 3). For example, at 25 1C in the system
1–20 mol% H2O, B20% of the aldehyde reacts—the
maximum possible if (within error/detection limits) all of the
H2O is sequestered. At 25 1C and added H2O equivalent to
40 and 60 mol% of 1, virtually all is likewise bound. Even at
100 mol% added H2O, the conversion of ca. 70% of the
aldehyde to hydrates means that B50% of this large H2O
bolus is covalently sequestered.
Dynamism—as embodied in the equilibrium formation and
scission of covalent bonds under mild conditions—is central to
constitutional dynamic chemistry (CDC), dynamic combina-
torial chemistry (DCC), and systems chemistry.1 From these
decidedly interconnected paradigms emerges a conceptual
model1 of stimulus-responsive soft materials, a prospective
use of ionic liquids (ILs) that we believe is ripe for exploration.2
Demonstrating the viability of the concept of IL-based
dynamic chemistry,3 we describe here the reactions of protic
nucleophiles with the new ILs N-methylpyridinium-2-
carboxaldehyde bis(trifluoromethanesulfonyl)amide, 1, and
N-methylpyridinium-3-carboxaldehyde bis(trifluoromethane-
sulfonyl)amide, 2. The cations of these materials readily and
reversibly react in a holo (self-contained)1d fashion with protic
nucleophiles via their ‘dynamism-disposed’ aldehyde groups.
In contrast to typical dynamic systems where the bulk of any
liquid present is an inert solvent for relatively low concentra-
tions of reactive components, here the liquids are the systems.
Both 1 and 2 are colorless liquids that react with H2O to
form gem diols and diastereomeric bis(hemialdal)s (Fig. 1).
These derivatives are distinguishable in the 1H-NMR by
resonances between d 6.7 and d 7.3 arising from the unique
gem diol and bis(hemialdal) C–H protons. The occurrence of
these in a range devoid of other resonances is invaluable to the
study of dynamic systems of 1 and 2 with H2O or alcohols.
To examine CD systems of 1 and H2O, freshly dried IL
samples were charged with the latter in molar increments of
20%, 40%, 60%, 80%, 100% and 120% relative to 1. Each of
these neat, homogeneous liquids was studied by 1H- and
13C-NMR at temperatures of 25, 35, 45, 55 and 85 1C (coaxial
tube, external d6-dmso lock and reference). In all cases, the
1–H2O combination quickly reaches equilibrium with the gem
diol and bis(hemialdal) covalent conjugates. The extent of
hydration is easily quantified by comparing the intensities
of the unique diol and bis(hemialdal) C–H resonances to that
of CHO in the unreacted cation (Fig. 2). In turn, the residual
‘unbound’ H2O is quantified by accounting for the amount of
the added water bolus needed to form the gem diol and that
The increasing hydrate contribution to the system constitu-
tion as a function of increasing added H2O is observed
regardless of temperature. To illustrate, at 85 1C, the conver-
sion of 1 to hydrates increases from 14% to 42% when
progressing from 20 mol% added H2O to 100 mol%.
However, in all cases the degree of hydration at equilibrium
is less for any given system as the temperature is increased. For
example, in the system 1 with 100 mol% added H2O, the
amount of aldehyde hydrated at equilibrium is B70% at
25 1C, and about 43% at 85 1C. This is wholly consistent with
the thermodynamic nature of aldehyde hydration in general.4
It should be noted as well that as system temperatures
are increased, their constitutions also change because the
proportion of bis(hemialdal)s decreases relative to gem diol.
Like the foregoing, this is reasonable since more ‘free’ H2O
becomes available at higher T, suppressing the condensation
reaction in which the bis(hemialdal)s are formed and H2O is
regenerated.
Department of Chemistry, University of South Alabama, Mobile, AL,
36688, USA. E-mail: jdavis@jaguar1.usouthal.edu;
Fax: +1 251 460 7359; Tel: +1 251 460 7427
w In memory of Professor Dmitry Rudkevich (1963–2007).
z Electronic supplementary information (ESI) available: Synthetic
procedures; spectra comprising a representative NMR study. See
DOI: 10.1039/b901899a
Fig.
1 Ionic components of the ionic liquid–molecular liquid
dynamic systems.
ꢁc
This journal is The Royal Society of Chemistry 2009
Chem. Commun., 2009, 2529–2531 | 2529

View Article Online
Fig. 2 Left: NMR excerpt showing the unique aldehyde, gem diol and bis(hemialdal) C–H resonances in the dynamic system 1–40% H2O; right:
NMR excerpt showing typical changes in the resonance of ‘free’ water in the system 2–20% H2O. It is independently verifiable that this is not
adventitious water in the d6-dmso lock/reference.
Fig. 3 Conversion of 1 to H2O and t-BuOH addition products as a function of both temperature and quantity of protic nucleophile. The system
1–t-BuOH was not studied at 85 1C because of the bp of the alcohol (83 1C).12
Building on these results, we initiated a parallel study using
2,2-dimethylpropan-1-ol (tert-butanol, t-BuOH). While H2O
is the least sterically demanding protic nucleophile reversibly
reactive with 1 or 2, this alcohol provides an opportunity to
probe the dynamic nature of systems with a participant at the
opposite steric extreme. Despite its bulky nature, alcoholysis
was observed at all proportions of t-BuOH and 1. At 25 1C
and 20 mol% of added alcohol, the degree of addition was the
same (within error) as that in the system 1–20 mol%
added H2O. However, all of the systems with larger alcohol
As a case in point, at 25 1C the system 1–80 mol% H2O sees
approximately 64% of 1 being hydrated. Under identical
conditions in the system containing 2, only 20% is converted
to hydration products. The superficially minor difference in
cation structure in 2 versus 1 clearly leads to substantial
differences in the constitutions of the dynamic systems they
form, emphasizing the ease with which tuning can be achieved.
But, this is not the only way in which the systems of 1 and 2
differ.
In many of the 1/2–H2O systems a portion of H2O is plainly
not sequestered at equilibrium. As mentioned, this ‘free’ H2O
can sometimes be observed in the 1H-NMR (Fig. 2), its
identity as such supported by the disappearance of its reso-
nance on D2O doping. In a representative system where free
water is observed, that of 2–20 mol% added H2O, the
unsequestered H2O appears as a broad peak at ca. d 4.2
(25 1C), which simultaneously shifts and sharpens as T is
increased to 55 1C (d 4.0), then 85 1C (d 3.6). In a 1-based
system where the resonance of ‘free’ H2O is cleanly observed,
1–80 mol% H2O, the same progression is seen though the
chemical shift values are quite different: d 5.6 (25 1C); d 5.2
(55 1C); and d 4.6 (85 1C). Significantly, when H2O is observed
in any of the 1-based systems it is always in this general
chemical shift range, as is likewise true for any system of 2
and the range specified for the 2–20 mol% added H2O
example. This non-correspondence may indicate that the
residual H2O is in appreciably different environments in
systems derived from the two different cations.7 Indeed, since
1 and 2 have the same anion, and since gem diols are known to
proportions relative to
1 evidenced measurably lower
magnitudes of addition than the analogous aqueous systems.
Since aldehyde alcoholysis is preferred over hydration,5 we
conclude that sterics play a meaningful role in modulating
these dynamic systems. Accordingly, we see hinted at in this
result the possibility of designing dynamism-disposed ILs
useful in molecular recognition and sorting.
Having established the dynamic parameters of systems of 1
and H2O/alcohols, we turned to the study of systems based on
2, the cation of which features the aldehyde group at the
pyridinium ring 3-position. Since iodide salts of the latter
cation do not, as solids, tend to form hydrates6 (unlike their
2-analogs), we speculated that the corresponding ILs would
form systems with H2O with different constitutions (proportion
of aldehyde plus H2O vs. hydrates) under identical conditions.
This proved to be the case. On a same-temperature basis, at all
molar proportions of 2–H2O, we observed the formation of
CD systems which invariably manifested lower hydrate
proportions than those in the same-condition systems of 1.
ꢁc
This journal is The Royal Society of Chemistry 2009
2530 | Chem. Commun., 2009, 2529–2531

View Article Online
have ‘‘a higher requirement for solvation by water than does
water itself,’’8 our data might suggest that the cations of 1 and
2 exert a significant effect on the structure of the ostensibly free
H2O, albeit to observably different degrees. It is important to
note, however, that in the absence of additional data the basis
of this behavior does remain an open issue; for example, we
cannot presently exclude the possibility that the observed
water signal in these spectra is not the product of signal
averaging with that of (the unobserved) gem diol OH protons.
Having established the feasibility of designing dynamic IL
systems, we set out to determine if we could exploit one in a
model application, the slow release of an odorant. As a proof-
of-concept, a system was prepared from 1 and 20 mol% added
menthol,9 a volatile alcohol used as a palliative in conjunction
with nasal and chest congestion. A control of menthol in the
non-functionalized IL [BMIM]Tf2N was likewise prepared.
Samples of each system were placed in a vacuum desiccator
and their masses checked periodically over four days, the
materials being maintained under active vacuum in the
interim. Over this period the control mass diminished essen-
tially to that of the base IL, the menthol loss being verified by
1H-NMR. In contrast, B40% of the initial menthol charge
remained in the system based on 1. This is consistent with
dynamic covalent bonding between the non-volatile IL and the
highly volatile menthol resulting in the dampening of the
evaporation of the latter. These results further cement findings
by Lehn et al. concerning the utility of dynamic systems for
slow-release applications.10 However, unlike the systems in the
foregoing work, the IL-based system contains only one
volatile component, the fragrance compound.
components are ionic, systems based on dynamic reactions
beyond those of aldehydes with protic nucleophiles, and
applications of dynamic IL systems in separations and in
absorptive cooling.11
JHD thanks Chevron for support of this research.
Notes and references
1 Select references: (a) R. F. Ludlow and S. Otto, Chem. Soc. Rev.,
2008, 37, 101; (b) J.-M. Lehn, Chem. Soc. Rev., 2007, 36, 151;
(c) P. T. Corbett, J. Leclaire, L. Vial, K. R. West, J.-L. Wietor,
J. K. M. Sanders and S. Otto, Chem. Rev., 2006, 106, 3652;
(d) J.-M. Lehn, Prog. Polym. Sci., 2005, 30, 814; (e) S. Otto,
J. Mater. Chem., 2005, 15, 3357.
2 M. D. Soutullo, C. I. Odom, B. F. Wicker, C. N. Henderson,
A. C. Stenson and J. H. Davis, Jr., Chem. Mater., 2007, 19, 3581.
3 A few previously described ILs engage in chemistry that is clearly
dynamic, though it has not been discussed in these terms, nor do
broad potential implications for this behavior appear to have been
heretofore proposed. For example, see ref. 2 and: (a) E. D. Bates,
R. D. Mayton, I. Ntai and J. H. Davis, Jr., J. Am. Chem. Soc.,
2002, 124, 4194; (b) D. J. Tempel, P. B. Henderson,
J. R. Brzozowski, R. M. Pearlstein and H. Cheng, J. Am. Chem.
Soc., 2008, 130, 400.
4 E. G. Sander and W. P. Jencks, J. Am. Chem. Soc., 1968, 90, 6154.
5 P. Gianni and E. Matteoli, Gazz. Chim. Ital., 1975, 105, 125; and
references therein.
6 Highly variable physical properties on the part of N-methyl-
pyridinium-2-carboxaldehyde iodide were earlier linked to its
exposure to H2O or alcohols and ascribed to the formation of
hydrates and hemialdals. See: G. M. Steinberg, E. J. Poziomek and
B. E. Hackley, Jr., J. Org. Chem., 1961, 26, 368, and references
therein. For a later treatment of the reactivity of this salt with
protic substances, see: J. Cabal, F. Hampl, F. Liska, J. Patocka,
F. Riedl and K. Sevcikova, Collect. Czech. Chem. Commun., 1998,
63, 1021.
7 (a) R. J. Hooley, P. Restorp, T. Iwasawa and J. Rebek, J. Am.
Chem. Soc., 2007, 129, 15639; (b) E. Dutkiewicz and
A. Jakubowska, J. Phys. Chem. B, 1999, 103, 9898.
8 R. Stewart and J. D. Van Dyke, Can. J. Chem., 1972, 50, 1992.
9 R. Eccles, Appetite, 2000, 34, 29.
10 B. Levrand, Y. Ruff, J.-M. Lehn and A. Herrmann, Chem.
Commun., 2006, 2965.
11 M. B. Shiflett and A. Yokozeki, 22nd International Congress of
Refrigeration: Refrigeration Creates the Future, Beijing, China,
Aug 21–26, 2007, b1119/1–b1119/7, International Institute of
Refrigeration, Paris, 2007.
In summary, we have shown that constitutionally dynamic
systems can be easily created using simply functionalized ionic
liquids as reactive components. The constitutions of these
systems are tunable and thermally responsive. They require
no solvent per se but instead comprise stimulus-responsive
liquid materials. In the case of the IL–H2O systems, the IL
component can be regarded as ‘self-drying’ (and as such merits
investigation as an additive to other ILs expressly for the
purpose of water sequestration). In conjunction with volatile
alcohols of interest, effective materials for the slow release of
the latter can be prepared. Among obvious extensions of
this work are the creation of systems in which all dynamic
12 Despite minimizing handling in air, the possible hydration of dried
1 by adventitious moisture during NMR sample preparation
prompted us to estimate an error of 3%, reflected here in the
assigned error bars.
ꢁc
This journal is The Royal Society of Chemistry 2009
Chem. Commun., 2009, 2529–2531 | 2531