Imitating micelles as a way to control coordination self-assembly: Cage-polymer switching directed rationally by solvent polarity or pH

Article abstract of DOI:10.1039/c0cc04196c

Disguising a metal complex as a micelle by using amphiphilic phosphine ligands enables it to switch between a coordination polymer and a discrete cage in response to solvent polarity or pH; this medium-dependent behaviour of the complex is rational becaus

Full text of DOI:10.1039/c0cc04196c

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Imitating micelles as a way to control coordination self-assembly:
cage-polymer switching directed rationally by solvent polarity or pHw
Nicola Giri and Stuart L. James
Received 2nd October 2010, Accepted 17th November 2010
DOI: 10.1039/c0cc04196c
Disguising a metal complex as a micelle by using amphiphilic
phosphine ligands enables it to switch between a coordination
polymer and a discrete cage in response to solvent polarity or pH;
this medium-dependent behaviour of the complex is rational
because it parallels that of true micelles.
solvent determines, indirectly, the adoption of a particular
coordination structure in a way which mimics the response of a
micelle.
In this communication, we explore whether such micellar
imitation can be extended beyond simple mono- and di-nuclear
species to larger coordination-based self-assembly more
generally. Specifically we find that it can indeed also provide
rational solvent control over the equilibrium between a large
coordination polymer and a discrete coordination cage
(Fig. 2). The micellar imitation also results in an unexpected
intermediate during conversion of the polymer to the cage.
An amphiphilic tripodal ligand, L, was used in this study
which is based on the classical tripodal ligand ‘triphos’, but
with hydrophilic piperazine groups added to each hydrophobic
phenyl group (Fig. 2a).5 As with the bidentate ligand used in
our previous work,3 in non-acidified water free L acts as a
typical non-ionic surfactant, forming micelles as indicated by
(i) concentration dependent line-broadening in NMR spectra,
(ii) dynamic light scattering (DLS) showing monodisperse
particles of hydrodynamic diameter DH 7.0 nm (20 mM, pH
9.0), (iii) a critical micelle concentration of 1.2 mM determined
by dye inclusion,4 (iv) a cloud point (29 1C), and (v)
insensitivity of micellisation to added NaCl (up to 1.0 M)
(Fig. S1, ESIw).2 By contrast, on addition of acid (which
charges the amine head groups and destabilizes the micelle
through head-group–head-group repulsions)2 or in organic
solvents such as chloroform or methanol L was found to
exist as monomers, as shown by sharp, concentration-
independent NMR spectra as well as DLS (Fig. S1, ESIw).
By analogy with the free L micelle, a 2 : 3 mixture of L and
AgOTf (2 : 3 molar ratio) in pure water gave a large
coordination polymer. DLS revealed particles of
hydrodynamic radius 140 (Æ25) nm which corresponds to ca.
2 Â 103 repeat units.6 Its 31P NMR spectrum was extremely
broad (dP = À6.0 ppm, o1/2 = 900 Hz, see Fig. 3a, pH 9.0)
and the large coordination chemical shift confirms that the P
centres coordinate to Ag in this species. By contrast there was
little change in chemical shift of the piperazinyl protons
showing that these groups remain uncoordinated. In passing,
we note that the polymer became a gel on addition of a further
half-equivalent of AgOTf, possibly due to further cross-linking
via the piperazine groups. The polymer exhibited no changes
in its NMR spectra or DLS measurements over several days
(or on cooling to 2 1C). Overall, the observations there-
fore point to a stable macromolecular species based on
Ag–P coordination, potentially stabilised by hydrophobic
interactions.
Gaining greater control over self-assembly is a fundamental
challenge facing chemistry and related fields for the foreseeable
future. One approach is to hybridise different types of
assembling driving forces in the same system. In this way,
methods for controlling one type of assembly may be conferred
on another. The present work deals with hybridisation of metal
coordination and micellisation. In metal-directed assembly the
solvent can dictate the product obtained, but cannot normally
be used rationally and predictively as a directing influence.1 In
micellisation, by contrast, the solvent provides the critical
driving force for assembly, i.e. solvophobic interactions, in
relatively well-understood ways, and as such the solvent is a
central factor which is exploited rationally.2 We have recently
shown that metal complexes can be designed to imitate
micelles. Specifically, amphiphilic ligands can be used to
surround the metal with a hydrophobic core, which is in turn
surrounded by a shell of hydrophilic head groups (Fig. 1a).3 In
water solvent, such complexes are then stabilised by both metal
coordination and hydrophobic interactions. Following
established principles of micellisation, on addition of other
species to the water (methanol or acid) such a complex can
reduce its size (change from dinuclear to mononuclear) in a
way which parallels the shrinking or disruption of a true
micelle (Fig. 1b). This occurs, however, by a non-micelle like
mechanism based on changes in the arrangement of metal
ligand bonds (e.g. bridging to chelating). Overall, therefore, the
Fig. 1 (a) Generic structure of a pseudo-micellar metal complex, (b)
interconversion of mono- and di-silver pseudomicellar complexes by
changing solvent polarity or pH.
Centre for the Theory and Application of Catalysis, School of
Chemistry and Chemical Engineering, Queen’s University Belfast,
David Keir Building, Stranmillis Road, Belfast BT9 5AG, UK.
E-mail: s.james@qub.ac.uk
w Electronic supplementary information (ESI) available: Synthesis of
L, analytical details, details of modelling. See DOI: 10.1039/
c0cc04196c
Again, analogous to free L, in acidic water (pH 1.5) or in
organic solvents (CDCl3–CH3CN, 4 : 1) the same mixture of L
and AgOTf arranged into a much smaller species, in particular
c
1458 Chem. Commun., 2011, 47, 1458–1460
This journal is The Royal Society of Chemistry 2011

View Article Online
Fig. 3 (a) 31P NMR spectra showing interconversion between the
polymer and adamantoid cage with addition of HOTf, (b) expansion of
the spectrum at pH 3.5 showing the adamantoid cage (*) and the
postulated macrocyclic intermediate (PA and PB), (c) structural unit
consistent with the intermediate signals, (d) DLS results.
Fig. 2 (a) The amphiphilic triphosphine, L, (b) overall schematic
representation of the structures adopted by free L and the L–Ag
complex in water and acidified water.
the discrete adamantoid coordination cage [Ag6L4(OTf)4]2+
Diagnostic spectroscopic data for this cage5 include the sharp
.
diagnostic of the adamantoid cage complex, showing both the
coordinated and non-coordinated O3SC6H4-4-CH3 anions, as
well as two chemically distinct aromatic P-substituents
(consistent with related X-ray crystal structures).5 DLS in
acidified water (pH 1.5) also showed a single species with
DH = 4.0 Æ 0.5 nm, which is a reasonable value considering
the molecular model of the cage (described below).
31P NMR spectrum (D2O/H+, d = À5.7 ppm, 1J109AgÀ31P
=
=
593 Hz; CDCl3–CH3CN, 4 : 1, d = À8.0 ppm, 1J109AgÀ31P
572 Hz with well-defined coupling to 107Ag and 109Ag nuclei
whose magnitude indicates AgP2 centres (cf. d = À5.95 ppm,
1J109AgÀ31P = 588 Hz) in [Ag6{CH3C(CH2PPh2)3}4(OTf)4]2+).5a,7
Similar 31P NMR spectra were obtained on reaction of L with
other silver salts such as AgO3SC6H4-4-CH3 (d = À7.95 ppm,
1J109AgÀ31P = 570 Hz). 1H NMR spectroscopy was also
A
model of the adamantoid cage (Fig. 4) was
constructed based on the single crystal structure of
[Ag6{CH3C(CH2PPh2)3}4(OTf)4]2+ 5a
with added piperazine
,
c
This journal is The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 1458–1460 1459

View Article Online
exocyclic P–Ag–P coordination by the third triphosphine
arm. This unit is based on the ‘silver-diphos’ dinuclear ring
which occurs commonly with diphosphines.5d The DLS results
from pH 6.0–3.5 indicate that the intermediate is relatively
small (DH = 1.5 nm Æ 0.3 nm) (Fig. 3d). The smallest possible
cyclic oligomer based on this repeat unit would be for n = 2,
i.e. an Ag6L4hexasilver ring (Fig. 2b). Interestingly, this ring
would be a cyclic isomer of the adamantoid cage. It suggests
that the polymer itself may also be based on this basic repeat
unit, and in fact, this type of topology has been previously seen
with a related triphosphine in the crystalline state.6 On further
decreasing the pH (2 and 1.5), the intermediate disappeared
and the adamantoid cage became the sole species present.
A
coherent overall scheme for the {(AgOTf)3L2}
system which accounts for the NMR and DLS data is shown
in Fig. 2b. The discrete cage forms when hydrophobic
interactions are either absent (organic solvents) or are
overcome by head-group protonation (low pH) whereas a
large globular coordination polymer based on a different
connectivity forms when hydrophobic interactions are
present (in pure water). This responsiveness clearly parallels
that of true micelles as illustrated by the behaviour of free L.
Interestingly, this study shows that micellar imitation can also
give rise to unexpected types of assembly such as the
intermediate seen here.
Fig. 4 Molecular model of the pseudo-micellar cage complex
[Ag6L4(OTf)4]2+ (white spheres = Ag, yellow = P and S, green = F,
phenylene groups = cyan, piperazine groups = blue).
In summary, we demonstrate the extension of micellar
imitation by simple metal complexes to much larger
structures. As such it is shown to be a novel approach to
controlling coordination-based self-assembly.
groups (see ESIw for details). It exhibits a pseudo-micellar
structure, i.e. the six central metal ions are surrounded by a
hydrophobic core region, which is in turn surrounded by a
hydrophilic shell of 24 diamine head groups. Importantly, the
model reveals that the hydrophobic core would still remain
partially exposed to the solvent. Consistent with our previous
study,3 this helps to explain why the cage rearranges into an
alternative, polymeric structure in non-acidified water, i.e. to
provide better shielding of the hydrophobic parts of the ligands
from the solvent.
We thank EPSRC (GR/S73860/01) for partial support of
this work.
Notes and references
1 (a) M. Fujita, Chem. Soc. Rev., 1998, 27, 417; (b) R. W. Saalfrank
and B. Demleitner, Transition Metals in Supramolecular Chemistry,
in Perspectives in Supramolecular Chemistry, ed. J. P. Sauvage,
Wiley-VCH, Weinheim, 1999, vol. 5, pp. 1–51; (c) D. L. Caulder
and K. N. Raymond, Acc. Chem. Res., 1999, 32, 975;
(d) S. Leininger, B. Olenyuk and P. J. Stang, Chem. Rev., 2000,
100, 853; (e) G. F. Swiegers and T. J. Malefetse, Coord. Chem. Rev.,
2002, 225, 91; (f) S. L. James, Chem. Soc. Rev., 2009, 38, 1744.
2 (a) R. J. Stokes and D. F. Evans, Fundamentals of Interfacial
Engineering, Wiley-VCH, Inc., 1996, ch. 5; (b) J. H. Fendler,
Membrane Mimetic Chemistry, Wiley-Interscience, 1982.
A further and unexpected point of interest was that an
intermediate was observed by both 31P NMR and DLS
during the break up of the polymer when titrated with acid.
The 31P NMR spectrum at pH 3.5 (which corresponds to
addition of one H+ per piperazine group) shows a main pair
of doublets due to the adamantoid cage (marked *, Fig. 3b).
3 N. Giri and S. L. James, Chem. Commun., 2011, DOI: 10.1039/
c0cc00648c.
There are two further pairs of doublets in an approximate 2 : 1
1
intensity ratio labelled PA (d = À5.2 ppm, J(109AgÀ31P)
=
4 K. G. Furton and A. Norelus, J. Chem. Educ., 1993, 70, 254.
5 (a) S. L. James, D. M. P. Mingos, A. J. P. White and D. J. Williams,
Chem. Commun., 1998, 2323; (b) P. W. Miller, M. Nieuwenhuyzen,
X. Xu and S. L. James, Chem. Commun., 2002, 2008; (c) X. Xu,
E. J. MacLean, S. J. Teat, M. Nieuwenhuyzen, M. Chambers and
S. L. James, Chem. Commun., 2002, 78; (d) P. W. Miller,
M. Nieuwenhuyzen, J. P. H. Charmant and S. L. James, Inorg.
Chem., 2008, 47, 8367; (e) S. L. James, E. Lozano and
M. Nieuwenhuyzen, Chem. Commun., 2000, 617.
6 This is an indicative value estimated from the unit cell volume in the
coordination polymer [Ag3{1,3,5-(PPh2)3C6H3}(OTf)3] where Z = 1
(Cambridge Crystallographic Database code MIJXUZ), see
X. L. Xu, M. Nieuwenhuyzen, J. Y. Zhang and S. L. James,
J. Inorg. Organomet. Polym. Mater., 2005, 15, 431.
592 Hz) and PB (partially coincident with the signals for the
1
cage), with J(109AgÀ31P) estimated as 590 Hz. The signals of
this intermediate cannot be unambiguously assigned.
However, PA and PB appear to constitute a single spin
system since there is a smaller coupling of ca. 6 Hz, close to
the limit of resolution but which is just resolved for PA, which
can be ascribed to a four-bond PA–PB through-backbone
coupling. The approximate 2 : 1 intensity ratio between PA
and PB is also consistent with an AB2 spin system (with regard
to only the 31P nuclei). A structural unit consistent with this
spin system, the magnitudes of the P–Ag couplings (indicating
AgP2 centres)7 and the 3 : 2 Ag : L stoichiometry is shown in
Fig. 3c. It has 10-membered {Ag}2 rings connected by
7 J. G. Verkade and L. D. Quin, Methods in Stereochemical Analysis,
Phosphorus-31 NMR spectroscopy in Stereochemical Analysis,
VCH Publishers Inc., Deerfield Beach, FL, 1987, vol. 8.
c
1460 Chem. Commun., 2011, 47, 1458–1460
This journal is The Royal Society of Chemistry 2011