Melville Laboratory for Polymer Synthesis
Department of Chemistry
Lensfield Road, Cambridge, CB2 1EW
United Kingdom
Aqueous Self-Assembly with
Cucurbit[n ]urils:
From Solution to Emulsion
Alexander S. Groombridge
Hughes Hall
Supervisors:
Professor Oren A. Scherman
& Professor Chris Abell
This thesis is submitted for the degree of Doctor of Philosophy, September 2017

Declaration
This report is submitted in fulfilment of the requirements for the Doctorate of Philosophy
in Chemistry. Except where indicated to the contrary, either directly or by reference, the
work described in this thesis is solely the work of the author. This work by no means
surpasses the word limit of 60000, as specified by the Degree Committee for Physics and
Chemistry.
Alexander S. Groombridge
2017
3

Abstract
Making use of the non-covalent bond to make materials is of great interest in many fields
of research. This PhD thesis describes a variety of highly interdisciplinary research un-
dertaken at the interface between chemistry, materials science, physics and engineering.
Chapter 1 is an introductory chapter into the core concepts underlying this thesis.
Supramolecular chemistry as a broad research field is briefly reviewed, followed by a focus
on host-guest chemistry. The macrocyclic cucurbit[n]urils (CB[n]s) in particular are high-
lighted with a discussion on their recent applications since their discovery. Emulsions and
their controlled generation with microfluidic techniques are then reviewed, as they have
been used as templates for self-assembly processes throughout this thesis.
A study into the synthesis of extended polymer networks composed entirely from small
molecules held together by non-covalent interactions is described in Chapter 2. These
highly dynamic and responsive supramolecular polymer networks have not yet been con-
structed with CB[n] host-guest chemistry. The ability of the larger CB[8] macrocycle
to encapsulate multiple guest molecules in a stepwise fashion was taken advantage on in
designing the synthesis of branching monomers. The monomers had two (A2) or three
(B3) terminal guest moieties for CB[8], which upon combination formed branching supra-
molecular polymers that were multi-stimuli responsive. However, the polymers precipit-
ated from solution at high concentrations rather than form a cross-linked network, due to
competing intra-chain cyclisation and the limited water solubility of CB[8]. By confining
these polymers to microfluidic droplets, directed assembly to the liquid-liquid interface
could drive polymerisation to form an interfacial cross-linked gel that was both elastic and
self-healing.
Chapter 3 follows on from these results, describing attempts into constructing hy-
perbranched supramolecular polymers from an AB2 guest molecule and CB[8] that would
form globular polymers. Intramolecular complexation dominated with the guest molecules
synthesised (A and B complexing within the molecule), evidenced by a variety of charac-
terisation. Compared to previous works that relied on linear molecules to form a folded
conformation for intramolecular complexes, these molecules were pre-organised with a
unique cooperative complexation pathway. The stimuli-responsiveness of the complexes
was probed, and the formation of self-sorting mixtures was demonstrated with multiple
CB[n] and additional guest molecules.
Controlling the self-assembly of semi-conducting nanocrystals with CB[7] is detailed
in Chapter 4, a process that typically requires harsh conditions or extensive time-scales.
Semi-conducting nanocrystals could be assembled instantaneously from water into ex-
5
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
tended networks that were highly porous with excess CB[7], retaining their nanoscale
properties. Limiting quantities of CB[7] could then form nanoscale aggregates that re-
mained in solution. Confinement of these assemblies within microfluidic droplets allowed
the synthesis of dense microparticles, that retained their shape after redispersal in water.
By simply including metallic nanocrystals as a minor component, mixed aggregates could
be synthesised analogously.
Finally, Chapter 5 draws overall conclusions from the results of this thesis, looking
broadly at the potential for future prospects in these areas of research.
6
Acknowledgements
A PhD is never completed alone, and I would like to say thank you for the kind help and
support I have received from so many people in these last few years in Cambridge. Firstly
I would like to thank everyone associated with the NanoDTC programme, for funding
me during my research and supporting me throughout. I am very grateful to both Prof.
Oren Scherman and Prof. Chris Abell for their supervision and support in helping me to
direct my ideas into tangible results. The experiences that Oren has provided to me in his
laboratory and at many international meetings have been unforgettable. Chris has been
very kind in allowing me to also join his laboratory and broaden my scientific horizons.
I would also like to say a big thank you to everybody in the Melville and Microdroplets
laboratories, whom have made my PhD an incredible experience. In particular I would like
to thank Matthew Rowland for helping me a lot in the early stages of my PhD; Cindy Tan
and Nathalia Carneiro for being fantastic welcoming labmates; Steven Barrow, Aniello
Palma and David Clarke for being experienced voices of reason whom I always went to
with my ideas and problems; Richard Parker for teaching me so much about microfluidics
and for driving me to become a better scientist; Ziyi Yu for his varied help with microfluidic
techniques; Dominique Hoogland and Magdalena Olesin´ska for being amazing friends and
for countless collaborations and discussions; and Kamil Sokolowski and Guanglu Wu for
inspiring me with their love of science and their hard work. Dominique, Magdalena, Kamil,
Gaunglu, Vijay Rana, Yuchao Wu, and Marlous Kamp are further acknowledged for their
essential help in proof-reading my thesis.
My friends and family also deserve a lot of gratitude, for putting up with me and sup-
porting me throughout my PhD. I sincerely thank Latika for joining me on this adventure,
for always being there for me, and for being very strict in stopping me from working too
hard! Lastly, I thank my cat, Callie, for getting me out of bed in the morning whether I
liked it or not, and helping me to recover after the more difficult days in the lab.
7
Abbreviations
ACN: Acetonitrile
ADA.HCl: 1-adamantylamine hydrochloride
AuNC: Gold Nanocrystal
AgNC: Silver Nanocrystal
BTA: 1,3,5-benzene triamide
CAD: Computer-Aided Design
CB[n]: Cucurbit[n]uril
CD: Cyclodextrin
CDI: Carboxydiimidazole
COSY: Homonuclear Correlation Spectroscopy
DCM: Dichloromethane
DIPEA: N,N -diisopropylethylamine
DLS: Dynamic Light Scattering
DMF: N,N -dimethylformamide
DOSY: Diffusion-Ordered NMR Spectroscopy
DP: Degree of Polymerisation
FT-IR: Fourier Transform Infrared Spectroscopy
HMBC: Heteronuclear Multiple-Bond Correlation Spectroscopy
HRMS: High Resolution Mass Spectrometry
HSQC: Heteronuclear Single-Quantum Correlation Spectroscopy
IFT: Interfacial Tension
ITC: Isothermal Calorimetry
LED: Light-Emitting Diode
LSPR: Localised Surface Plasmon Resonance
MeOH: Methanol
MPA: 3-mercaptopropionic acid
NC: Nanocrystal
NIR: Near-Infrared Light
NMR: Nuclear Magnetic Resonance
PDI: Polydispersity Index
PDMS: Poly(dimethylsiloxane)
PFOH: Perfluorooctanol
PLQE: Photoluminescence Quantum Efficiency
QD: Quantum Dot
8
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
SEM: Scanning Electron Microscopy
SERS: Surface-Enhanced Raman Spectroscopy
SLS: Static Light Scattering
SPR: Surface Plasmon Resonance
STEM: Scanning TEM
TEM: Transmission Electron Microscopy
THF: Tetrahydrofuran
TLC: Thin Layer Chromatography
UPy: 2-ureido-4-pyrimidinone
XPS: X-ray Photoelectron Spectroscopy
9
Contents
1 Introduction 26
1.1 Supramolecular Chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.1.1 Host-Guest Chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . 27
1.1.2 Cucurbit[n]urils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.2 Emulsions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
1.2.1 Droplet Microfluidics . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
1.3 Aim of This Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2 Aqueous Interfacial Gels Assembled from Small Molecule Supramolecu-
lar Polymers 40
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.1.1 Branching Polymerisations . . . . . . . . . . . . . . . . . . . . . . . 40
2.1.2 Supramolecular Polymers . . . . . . . . . . . . . . . . . . . . . . . . 43
2.1.3 Host-Guest Supramolecular Polymers . . . . . . . . . . . . . . . . . 49
2.2 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.2.1 Supramolecular Polymerisation . . . . . . . . . . . . . . . . . . . . . 52
2.2.2 Stimuli-Responsiveness . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.2.3 Gelation at Liquid-Liquid Interfaces . . . . . . . . . . . . . . . . . . 60
2.3 Conclusions and Future Outlook . . . . . . . . . . . . . . . . . . . . . . . . 68
2.4 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
2.4.1 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . 69
2.5 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3 Cooperative Intramolecular Host-Guest Complexes 74
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.2 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.2.1 Synthetic Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.2.2 NMR Titrations with CB[7] and CB[8] . . . . . . . . . . . . . . . . . 81
3.2.3 DOSY Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
3.2.4 UV-vis Spectrophotometry . . . . . . . . . . . . . . . . . . . . . . . 97
3.2.5 Isothermal Titration Calorimetry . . . . . . . . . . . . . . . . . . . . 98
3.2.6 Advanced Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
3.2.7 Rigid Architectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
3.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
10
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
3.4 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3.4.1 Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3.4.2 2D NMR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
3.4.3 NpVio2 Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
3.4.4 AzoVio2 Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
3.4.5 Np2Vio Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
3.4.6 Towards Rigid Asymmetric Guests . . . . . . . . . . . . . . . . . . . 123
3.5 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
4 Multiscale Self-Assembly of Semi-Conducting and Plasmonic Nanocrys-
tals in Water 124
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
4.1.1 Nanocrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
4.1.2 Self-Assembly of Nanocrystals . . . . . . . . . . . . . . . . . . . . . . 127
4.2 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
4.2.1 Bulk Aggregation of InP/ZnS QDs and AuNCs . . . . . . . . . . . . 135
4.2.2 Supraparticles of InP/ZnS QDs and AuNCs . . . . . . . . . . . . . . 140
4.2.3 Microstructures of InP/ZnS QDs and AuNCs . . . . . . . . . . . . . 143
4.2.4 Photoluminescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
4.3 Conclusions and Future Directions . . . . . . . . . . . . . . . . . . . . . . . 161
4.4 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
4.4.1 Materials and Instrumentation . . . . . . . . . . . . . . . . . . . . . 162
4.4.2 Nanocrystal Preparation and Characterisation . . . . . . . . . . . . 162
4.4.3 Bulk Aggregates and Supraparticles . . . . . . . . . . . . . . . . . . 164
4.4.4 Microfluidic Droplet Generation and Analysis . . . . . . . . . . . . . 164
4.4.5 Photoluminescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
4.5 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
5 Conclusions 166
A Supplementary Information 182
A.1 Supplementary Data for Chapter 3 . . . . . . . . . . . . . . . . . . . . . . . 183
A.1.1 2D NMR Spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
A.1.2 ITC Titrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
A.1.3 AzoVio2 Photo-isomerisation
1H NMR Stacks . . . . . . . . . . . . . 198
A.2 Supplementary Information for Chapter 4 . . . . . . . . . . . . . . . . . . . 202
A.2.1 Ostwald Ripening Fitting . . . . . . . . . . . . . . . . . . . . . . . . 202
11
List of Figures
1.1 Chemical structures of some common synthetic macrocyclic host molecules. 28
1.2 The synthesis of a mixture of CB[n], and space-filling models demonstrat-
ing the increasing size and constant height as n increases. Adapted from
references.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.3 A scheme showing the different types of host-guest binding architectures
available with CB[n]s. Adapted from references.1 . . . . . . . . . . . . . . . 31
1.4 A scheme showing the formation of hydrogels from polysaccharide poly-
mers functionalised with phenylalanine amino acid guests via homoternary
complexation. Adapted from references.2 . . . . . . . . . . . . . . . . . . . 31
1.5 A scheme showing the formation of hybrid polymer-nanoparticle microcap-
sules from water-in-oil emulsions. Adapted from references.3 . . . . . . . . . 32
1.6 A scheme showing the thermodynamically driven destabilisation pathways
of macroemulsions. Adapted from references.4 . . . . . . . . . . . . . . . . . 33
1.7 Triple emulsions generated by coupling of multiple glass capillary micro-
fluidic devices. Adapted from references.5 . . . . . . . . . . . . . . . . . . . 34
1.8 A microfluidic chemostat used to study the growth of microbial populations.
Intricate design and incorporation of pneumatic valves allows complex func-
tion. Adapted from references.6,7 . . . . . . . . . . . . . . . . . . . . . . . . 35
1.9 A schematic of the soft lithographic replica moulding technique to construct
PDMS microfluidic devices. (a) Contact photolithography is carried out
through a printed mask onto a thin film of photoresist on a Si wafer. (b)
After developing the photoresist, the desired pattern will be present on the
Si wafer to form a ‘master’. (c) PDMS precursor is poured onto the master
and subsequently cross-linked. (d) The cross-linked PDMS can be peeled
off the master chip, and then bonded to a substrate of choice. . . . . . . . . 36
1.10 (a) T-junction, (b) planar flow-focussing and (c) capillary flow-focussing
geometries. Adapted from references.3,8 . . . . . . . . . . . . . . . . . . . . 37
1.11 Structures of developed triblock copolymers to stabilise water-in-perfluorocarbon
emulsions.9,10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.1 A series of schematic diagrams showing some of the range of dendritic and
hyperbranching architectures possible. Adapted from references.11 . . . . . 41
12
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
2.2 A graph showing the calculated branching coefficient, α, as the conversion
of A functionality progresses. The solid line shows the polymerisation of A2
with B3, where the stoichiometry of A to B is 1. The dashed line shows the
polymerisation of a single branching monomer, AB2. The red line shows
the point at which gelation will be observed, at α = 0.5. . . . . . . . . . . . 42
2.3 A schematic to show the potential growth mechanisms for supramolecu-
lar polymerisation. (a) Isodesmic polymerisation, where each addition of
monomer is independent; (b) ring-chain mediated polymerisation, where
cyclisation is favoured below a critical concentration; (c) cooperative poly-
merisation, where monomer addition is promoted after a critical nucleation
step. Adapted from references.12,13 . . . . . . . . . . . . . . . . . . . . . . . 44
2.4 The first H -bonded supramolecular polymers based upon triple, quadruple,
or sextuple interactions. Adapted from references.14 . . . . . . . . . . . . . 46
2.5 Schematic representation of a (a) 2D network or a (b) linear polymer obtain-
able in water through control over guest topology. Adapted from references.15 49
2.6 Schematic representation of the photo-controlled supramolecular polymer-
isation of branching CD-based polymers. Adapted from references.16 . . . . 50
2.7 Supramolecular polymerisation with CB[8] based upon an ABBAmonomer,
thus inhibiting cyclisation and intramolecular complexation. Adapted from
references.14,17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.8 (a) Chemical structures and schematic representations of CB[8], and the
ditopic (A2) and tritopic (B3) monomers used in this work. (b) Overview
of the proposed assembly process; A2 is first complexed with CB[8] (2 eq.),
and then mixed with B3 (0.67 eq.) to form the polymer. This polymer can
be disassembled in response to either (i) reversible photo-isomerisation of
the azobenzene with UVA light (left), or (ii) introduction of a competitive
guest 1-adamantylamine hydrochloride (ADA) (right). . . . . . . . . . . . . 52
2.9 Stacked 1H NMR spectra in D2O of A2 (a) upon addition of 1 equiv. (b) and
2 equiv. (c) of CB[8]. The complexation proceeds as previously reported.18 53
2.10 Stacked 1H NMR spectra in D2O (a) and d6-DMSO (b) of B3. The aro-
matic region in D2O is characteristic of when pi − pi stacking interactions
are occurring, this is not the case when in d6-DMSO as this solvates the
molecule well. R represents further azobenzene imidazolium groups. . . . . 54
2.11 1H NMR spectra in D2O of (a) B3, (b) A2-CB[8] (1:2), (c) A2-B3-CB[8]
(1.5:1:3). Concentrations in all cases were A2 = 0.55 mM, B3 = 0.37 mM,
CB[8] = 1.11 mM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.12 A photo to illustrate the loss in water solubility occurring upon assembly of
the supramolecular polymer at higher concentrations. Before combination
at equivalent concentrations, the separate components remain completely
dissolved. Concentrations are as follows: A2 1.10 mM, B3 0.73 mM, CB[8]
2.2 mM. Concentrations used in 1H NMR measurements were half of this,
at the limit before the solution became turbid. . . . . . . . . . . . . . . . . 56
13
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
2.13 Stacked 1H NMR spectra in d6-DMSO of B3 in response to UVA light and
heat treatment. (a) State immediately after dissolution of solid B3; (b)
after exposure to UVA light for 1 h; (c) after heating at 80 ◦C for 6 h;
(d) after leaving to reach equilibrium state in ambient conditions for 12 h.
The proportion of the E and Z isomers were calculated from comparing
integration of the aromatic peaks corresponding to each isomer, and that
of the CH2 at ca. 5.5 ppm. Blue circles represent peaks that change for the
E isomer, and red squares for the Z isomer. . . . . . . . . . . . . . . . . . . 57
2.14 UV-vis measurements taken of the B3 monomer in water (0.7 v/v% DMSO).
As is characteristic for azobenzene photoisomerisations, the n to pi∗ trans-
ition at 324 nm is altered dramatically upon UVA-triggered isomerisation,
while the pi to pi∗ transition absorbance at 426 nm increases. After heating,
the photoisomerisation can be completely reversed.19 . . . . . . . . . . . . . 57
2.15 1H NMR spectra in D2O of (a) A2-B3-CB[8] branched supramolecular poly-
mer in its equilibrium state; (b) after 1 h of UVA exposure; (c) after sub-
sequent heating at 80 ◦C for 6 h. Note, upon heating the azobenzene is >99
% E isomer, higher than that originally present in (a). . . . . . . . . . . . . 58
2.16 UV-vis measurements taken of the A2-B3-CB[8] polymer. A similar trend
in photoisomerisation is observed as for the B3 monomer alone. At room
temperature in the dark, a small amount of recovery of the equilibrium is
observed after 45 mins. Here a blue LED light exposure for 10 mins is
sufficient to drive the recovery of the E isomer due to the low concentration
used ([A2] = 55.0 µM). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.17 1H NMR spectra in D2O of A2-B3-CB[8] polymer (a) upon titration with
0.5 (b), and 1 (c) equiv. of ADA relative to CB[8]. . . . . . . . . . . . . . . 59
2.18 A schematic outline of the assembly process from dilute solution to an
interfacial gel. At dilute solution there are no precipitates, but there is
also no extended polymer network. After emulsification and subsequent
electrostatic attraction and accumulation to the liquid-liquid interface, the
concentration and density of the polymer rapidly increases over ca. 2 s,
leading to inter-chain cross-linking and gelation (Fig. 2.19b).20,21 . . . . . . 60
2.19 Transmission optical micrographs of: (a) an example flow-focussing junction
resulting in monodisperse microdroplets, typically with flow rates of 150 and
100 µL h−1 for the oil and water flows respectively; (b) top panel shows
aqueous microdroplets containing A2-B3-CB[8] and oil phase containing 2
wt.% neutral triblock and 1 wt.% RFCOOH, and bottom panel is at higher
magnification; (c) the same as in (b) but with 1 wt.% RFNH2. . . . . . . . 61
2.20 Transmission optical micrographs of emulsions on glass slides containing:
(a) aqueous droplets containing A2-B3 and no CB[8]; (b) aqueous droplets
containing A2-B3-CB[7]. The continuous phase was 2 wt.% neutral triblock
and 1 wt.% RFCOOH in FC-40. . . . . . . . . . . . . . . . . . . . . . . . . 62
14
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
2.21 Transmission optical micrographs of emulsions deposited onto a glass slide
containing: (a) aqueous droplets containing 500 kDa dextran and (b) aqueous
droplets containing A2-B3-CB[8] and 500 kDa dextran. Both continuous
phases comprised FC-40 containing 2 wt.% neutral triblock and 1 wt.%
RFCOOH. Upon addition of PFOH complete coalescence is observed over
a few seconds in case (a), and in case (b) the network stabilises the droplet
interface to coalescence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
2.22 Plots showing how γ changes with equilibration time for different experi-
ments. RFCOOH was in FC-40 at 0.001 wt.%, RFNH2 was in FC-40 at 0.01
wt.%, A2-B3-CB[8] was diluted 20x from the microfluidic stock solution to
[A2] = 0.55 µM. ADA was added in 10x stoichiometric excess to CB[8] to
facilitate rapid displacement. . . . . . . . . . . . . . . . . . . . . . . . . . . 64
2.23 Transmission optical micrographs of the pendant droplet during pendant
droplet measurements of continuous aqueous phase containing A2-B3-CB[8],
and oil droplet of 0.001 wt.% RFCOOH in FC-40. From left to right, (a) an
equilibrated 5 µL droplet pumped in and exhibiting buckling; (b) the same
droplet after pumping out to 8.5 µL to stretch the interfacial gel followed
by immediate pumping in with no equilibration time. . . . . . . . . . . . . . 65
2.24 Transmission micrographs of the pendant droplet during pendant droplet
measurements at equilibrium (top) and then after subsequent pumping in
(bottom). (a) Aqueous A2-B3 and droplet of RFCOOH in FC-40; (b)
aqueous A2-B3-CB[8] and droplet of RFNH2 in FC-40; (c) aqueous A2-
B3-CB[8] and 10x excess of ADA and droplet of RFCOOH in FC-40; (d)
aqueous A2-B3-CB[8] after 1 h UVA irradiation and droplet of RFCOOH
in FC-40. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
2.25 Transmission micrographs of the pendant droplet during pendant droplet
measurements after 45 min equilibration and subsequent pumping in at
different concentrations of aqueous A2-B3-CB[8]. Here buckling can be
observed at [A2] = 0.55 µM (a), 0.055 µM (b), and 0.0055 µM (c). It can
be clearly observed that there is a lower density of wrinkling in (b), implying
a thinner crosslinked network had been formed, and no network is formed
at (c). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
2.26 Shape fitting of pendant droplet experiments carried out with the Open-
Capsule software.22 (a) The reference droplet before droplet compression
after 45 mins of equilibration; (b) calculated shape fits based upon im-
ages of subsequent droplet compression leading to buckling; (c) repeated
compression after expanding the buckled droplet to 8.5 µL volume with no
equilibration time as in Fig. 2.23b. . . . . . . . . . . . . . . . . . . . . . . . 67
3.1 Example structures of AB2 monomers for constructing hyperbranched poly-
mers. Adapted from references.23–25 . . . . . . . . . . . . . . . . . . . . . . 74
15
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
3.2 The formation of an intramolecular heteroternary complex between naph-
thyl and viologen units. Only the 1:1 complex was observed. Adapted from
references.26 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
3.3 Intramolecular complexes being formed with CB[8] in response to a reducing
agent, based upon viologen homodimerisation. Adapted from references.27 . 75
3.4 A [10]pseudorotaxane dendrimer composed of 13 different molecules. The
structure was formed with high selectivity, employing different types of
CB[n] in water. Adapted from references.28 . . . . . . . . . . . . . . . . . . 76
3.5 A mass spectrum and molecular model of a linear guest molecule back-folded
on itself to form an intramolecular heteroternary complex with CB[8] and
a binary complex with CB[7]. Adapted from references.29 . . . . . . . . . . 77
3.6 Proposed schematic self-assembly of supramolecular hyperbranched poly-
mers from CB[8] and an AB2 molecule based on two viologens and one
naphthyl guest. Here, the ratio of AB2 to CB[8] would be 1:2. . . . . . . . . 78
3.7 The synthetic pathway to AB2 guest molecules based on two viologens and
either a naphthyl or azobenzene guest (NpVio2 or AzoVio2). (a) MsCl,
DIPEA, DCM, r.t. 4 h; (b) DIPEA, diethanolamine, THF, reflux 3 days;
(c) CBr4, PPh3, DCM, r.t. 16 h; (d) MeI, DCM, r.t. 16 h; (e) Me-Bipy,
DMF, 80 ◦ C 3 days; (f) H2O, NH4PF6; (g) acetone, Bu4NCl. . . . . . . . . 79
3.8 Stacked 1H NMR spectra of NpVio2 in D2O (bottom spectrum), followed
by addition of 0.8 (middle spectrum) and then 2.0 equiv. of CB[8] (top
spectrum). The peak assignments in red show proton environments outside
of the CB[8] cavity or forming a 1:1 binary complex with viologen, and
those in blue are inside the cavity as a heteroternary 1:1:1 complex. . . . . 81
3.9 The proposed complex of NpVio2 with 2 CB[8] molecules forming an asym-
metric complex. The positions of CB[8] were derived from the 1H NMR
experiments and optimisation was carried out with MMFF94 molecular
mechanics in Avogadro v1.2. . . . . . . . . . . . . . . . . . . . . . . . . . . 83
3.10 Stacked 1H NMR spectra of NpVio2 in D2O (bottom spectrum), followed
by addition of 1.0 (middle spectrum) and then 2.0 equiv. of CB[7] (top
spectrum). The peak assignments in red show proton environments outside
of the CB[7] cavity, and those in blue are inside the cavity. . . . . . . . . . . 84
3.11 The proposed complex of NpVio2 with 2 CB[7] molecules forming a sym-
metric complex. The positions of CB[7] are derived from the 1H NMR
experiments and optimisation was carried out with MMFF94 molecular
mechanics in Avogadro v1.2. . . . . . . . . . . . . . . . . . . . . . . . . . . 85
3.12 1H NMR spectrum of NpVio2 in D2O with 1 equiv. each of CB[8] and
of CB[7]. The initially symmetric molecule is now asymmetric, so peaks
corresponding to complexation with CB[7] or CB[8] can be assigned with
the prefix 7 or 8 respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . 85
16
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
3.13 The proposed complex of NpVio2 with CB[8] and CB[7] molecules forming
a self-sorting complex. The positions of CB[7] and CB[8] are derived from
the 1H NMR experiments and optimisation was carried out with MMFF94
molecular mechanics in Avogadro v1.2. . . . . . . . . . . . . . . . . . . . . . 86
3.14 Stacked 1H NMR spectra of AzoVio2 in D2O (bottom spectrum), followed
by addition of 1.0 (middle spectrum) and then 2.0 equiv. of CB[8] (top
spectrum). The peak assignments in red show proton environments outside
of the CB[8] cavity or forming a 1:1 binary complex with viologen, and
those in blue are inside the cavity as a heteroternary 1:1:1 complex. . . . . 87
3.15 The proposed complex of AzoVio2 with 2 CB[8] molecules forming an asym-
metric complex. The positions of CB[8] are derived from the 1H NMR
experiments and optimisation was carried out with MMFF94 molecular
mechanics in Avogadro v1.2. . . . . . . . . . . . . . . . . . . . . . . . . . . 88
3.16 Stacked 1H NMR spectra of AzoVio2 in D2O (bottom spectrum), followed
by addition of 1.0 (middle spectrum) and then 2.0 equiv. of CB[7] (top
spectrum). The peak assignments in red show proton environments outside
of the CB[7] cavity, and those in blue are inside the cavity. Two different
complexes with the Z isomer could be observed, denoted Zx and Zx’ where
applicable. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.17 The proposed complex of AzoVio2 with 2 CB[7] molecules forming a sym-
metric complex. The positions of CB[7] are derived from the 1H NMR
experiments and optimisation was carried out with MMFF94 molecular
mechanics in Avogadro v1.2. Azobenzene being twisted or planar in polar
solvents is still under debate.30 . . . . . . . . . . . . . . . . . . . . . . . . . 90
3.18 1H NMR spectrum of AzoVio2 in D2O with 1 equiv. of CB[8] and of CB[7].
The initially symmetric molecule is now asymmetric, so peaks corresponding
to complexation with CB[7] or CB[8] can be assigned with the prefix 7 or
8 respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
3.19 The proposed complex of AzoVio2 with CB[8] and CB[7] molecules forming
a self-sorting complex. The positions of CB[7] and CB[8] are derived from
the 1H NMR experiments and optimisation was carried out with MMFF94
molecular mechanics in Avogadro v1.2. . . . . . . . . . . . . . . . . . . . . . 91
3.20 1H NMR spectrum of Np2Vio in D2O with 1 equiv. of CB[8]. The initially
symmetric molecule is now asymmetric. . . . . . . . . . . . . . . . . . . . . 92
3.21 The proposed complexes of Np2Vio with 1 CB[8] molecule forming an asym-
metric complex. The position of CB[8] was derived from the 1H NMR
experiments and optimisation was carried out with MMFF94 molecular
mechanics in Avogadro v1.2. . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.22 1H NMR spectrum of Np2Vio in D2O (bottom) with 0.2 (middle) and 1.0
equiv. (top) of CB[7]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
3.23 The proposed complex of Np2Vio with 1 CB[7] molecule. Optimisation was
carried out with MMFF94 molecular mechanics in Avogadro v1.2.. . . . . . 94
17
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
3.24 An overview scheme of the proposed assembly process of the guest molecules
with CB[7] and CB[8] evidenced by NMR spectroscopy. . . . . . . . . . . . 95
3.25 UV-vis spectra of each AB2 molecule followed by titration with CB[8] or
CB[7]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
3.26 A schematic diagram showing the proposed binding events observed in the
ITC measurements shown in Fig. 3.27 for NpVio2 and AzoVio2. . . . . . . 99
3.27 ITC isotherms for cells containing CB[7] (61 µM) or CB[8] (51 µM) titrated
with each AB2 (1 mM) molecule in H2O. . . . . . . . . . . . . . . . . . . . 100
3.28 ITC isotherms for cells containing CB[8] (51 µM) and ADA.HCl (1 mM)
titrated with each AB2 (1 mM) molecule in H2O. . . . . . . . . . . . . . . . 101
3.29 Thermodynamic data for CB[8]-based host guest complexation determined
by ITC. Homoternary 2:1 complexes (blue), heteroternary 1:1:1 complexes
(red) and binary 1:1 or quaternary 2:2 complexes (green) are shown. Ad-
apted from references.31 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
3.30 ITC isotherms for cells containing AB2 (51 µM) titrated with CB[7] (0.69
mM) in H2O. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
3.31 UV-vis spectra demonstrating the photo-physical properties of AzoVio2,
AzoVio2 with 2 equiv. CB[8], and AzoVio2 with 2 equiv. CB[7]. The
initial photo-stationary state is shown, followed by UVA exposure until the
spectrum remains unchanged, then heated at 70 ◦ C until the spectrum
remains unchanged. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
3.32 A plot showing the kinetics of the photo-isomerisation of AzoVio2 present
in different complexes of CB[7] and CB[8]. Error bars have been calculated
by the standard deviation of the integration of 3 different peaks relating to
the E and Z isomers within the same spectrum. . . . . . . . . . . . . . . . 106
3.33 Stacked 1H NMR spectra of AzoVio2 following its photo-isomerisation.
From the bottom the equilibrium state is shown, then predominately the Z
isomer after 4 h UVA irradiation, then predominately the E isomer after
heating. Peaks labelled red correspond to the E isomer, and those in blue
to the Z isomer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.34 A schematic showing the potential for the NpVio2-2CB[8] complex to bind
with another guest molecule to form multiple heteroternary complexes. . . . 108
3.35 Stacked 1H NMR spectra of NpVio2-2CB[8] in D2O (bottom spectrum),
followed by addition of 1 equiv. of 2-naphthol (top spectrum). The peak
assignments in red show proton environments outside of the CB[8] cavity
or binding with 2-naphthol, and those in blue are inside the cavity as an
intramolecular heteroternary 1:1:1 complex. . . . . . . . . . . . . . . . . . . 108
3.36 The DOSY NMR spectrum for the NpVio2-2CB[8] complex with 2-naphthol
(left), and the optimised model carried out with MMFF94 molecular mech-
anics in Avogadro v1.2. (right). . . . . . . . . . . . . . . . . . . . . . . . . . 108
18
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
3.37 Stacked 1H NMR spectra of the addition of excess ADA.HCl to (a) Np2Vio-
CB[8], (b) AzoVio2-2CB[8], and (c) NpVio2-2CB[8]. ADA.HCl could only
displace the binary complexes and not the intramolecular heteroternary
complexes. Schematics of the complexes present have been included for
each step, and peaks corresponding to ADA.HCl have been highlighted
with green squares. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
3.38 Stacked 1H NMR spectra of the addition of excess ADA.HCl to (a) Np2Vio-
2CB[7], (b) AzoVio2-2CB[7], and (c) NpVio2-2CB[7]. In each case the initial
AB2-2CB[7] complex was displaced resulting in an ADA.HCl-CB[7] complex
and the unbound AB2 molecule. Schematics of the complexes present have
been included for each step, and peaks corresponding to ADA.HCl have
been highlighted with green squares. . . . . . . . . . . . . . . . . . . . . . . 111
3.39 Stacked 1H NMR spectra of the addition of Np2Vio (a), AzoVio2 (b),
NpVio2 (c) to the binary complex of ADA.HCl with CB[8] (d) in D2O.
Schematics of the complexes present have been included for each step, and
peaks corresponding to ADA.HCl have been highlighted with green squares. 112
3.40 A proposed reaction pathway to an AB2 molecule with a rigid and water-
soluble core. The first step has been carried out successfully. Hydrophobic
guests could be coupled by either amide coupling to give an amide, or by
reductive amination to give a secondary amine. . . . . . . . . . . . . . . . . 113
3.41 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the NpVio2-1CB[8] complex. For COSY, HSQC and HMBC
positive peaks are shown in blue and negative peaks are shown in red. . . . 116
3.42 Scheme of NpVio2 synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
3.43 Scheme of AzoVio2 synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . 119
3.44 Scheme of Np2Vio synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
4.1 The calculated structure of a 5 nm PbS QD stabilised by oleic acid. Adapted
from references.32 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
4.2 A schematic diagram showing the relation between bulk semiconductor
band structure to that of quantum dots. Eg is the band gap energy, and
the cluster relates to a QD. Adapted from references.33 . . . . . . . . . . . 127
4.3 A schematic diagram showing NCs stabilised by (top) steric hindrance or
(bottom) electric double layer repulsion. Adapted from references.32 . . . . 128
4.4 An overview of NC composition and their self-assembly pathways to gener-
ate superlattices (colloidal crystals). Adapted from references.34 . . . . . . 129
4.5 The evolution of new electronic states upon QD aggregation. (a) A schem-
atic of two CdSe NCs (top) and their electronic states (bottom) separated
by an energy barrier of ∆E and interparticle distance ∆x. (b) Schematic
showing the effect of ∆x on the energetic states. Adapted from references.34 130
19
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
4.6 The variety of potential non-additive effects from the direct coupling metal-
lic and semi-conducting NCs in superlattices. (a,b) Control over lumines-
cence properties, either enhancing of quenching; (c,d) emergence of new
magnetic properties of hybrid materials; (e,f) conductivity enhancements
in mixed lattices; (g,h) enhancement in catalytic activity or the ordered
lattice. Adapted from references.34 . . . . . . . . . . . . . . . . . . . . . . . 131
4.7 (a,d) SEM, (b,e) TEM, and (c,f) size distribution of CdSe supraparticle
samples obtained at different reaction times: CdSe-20 (a-c) 20mins, and
CdSe-30 (d-f) 120 mins. Adapted from references.35 . . . . . . . . . . . . . 133
4.8 TEM images of the aggregated products in the diffusion-limited colloidal
growth regime, highlighting the uniform 0.9 nm interparticle distance. Ad-
apted from references.36 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
4.9 (a) The crystal structure of CB[7]. (b) Schematic representation of the
water-soluble AuNCs and InP/ZnS QDs. (c) TEM micrographs of AuNCs
and QDs. (d) DLS measurements of AuNCs and QDs by volume percentage.
(e) Overview of the sequential assembly process of the bulk aggregated
precipitant from aqueous dispersions of AuNCs and QDs in different orders
of addition. (f) Photographs under visible and UVA light illumination of
the bulk precipitant, showing colours dependent on their assembly process.
(g) STEM micrographs showing the nanoscale structure of the aggregates
with AuNCs easily resolvable from the QDs. (h) TEM micrographs showing
the nanoscale structure of the aggregates. . . . . . . . . . . . . . . . . . . . 136
4.10 (a) The UV-vis spectra of a aqueous solutions of QDs, AuNCs and a 1:1
mixture of both. (b) The steady-state fluorescence spectrum of QDs with
λex = 360 nm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
4.11 Photos of the CB[7] induced aggregation of QDs and AuNCs in bright white
light, and under UVA illumination. (a,b) QDs & CB[7]: from left to right,
an initial suspension of QDs (1 mL) has excess CB[7] added (0.5 mL) and
the aggregation leading to precipitation was followed in both white light (a)
and UVA light (b). (c,d) AuNCs, QDs & CB[7]: an initial AuNC suspension
(0.5 mL) is mixed with a QD suspension (0.5 mL), followed by excess CB[7]
(0.5 mL). (e,f) AuNCs, CB[7], & QDs: AuNCs were premixed with CB[7],
followed by addition of QDs. . . . . . . . . . . . . . . . . . . . . . . . . . . 138
4.12 1H NMR spectra in D2O of (a) QDs and of titration of CB[7] into QDs in
nmol quantities (b). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
4.13 1H NMR spectrum of the QD-CB[7] precipitant after dissolution of the
QDs with DCl, and addition of a pyridine reference (1 µmol). This allowed
calculation of the ratio of QD:CB[7]:MPA in the precipitant to be calculated
as 1:25:65. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
20
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
4.14 DLS kinetics data showing the two stage growth mechanism for CB[7]-
based aggregation of QDs (a,b) and AuNC-QDs (c,d) plotted on a normal
and semi-log scale. Fits have been calculated for the second stage of growth
with an Ostwald ripening mechanism.37 . . . . . . . . . . . . . . . . . . . . 140
4.15 A schematic diagram of the two stage mechanism of aggregation observed
for QD-CB[7] and AuNC-QD-CB[7]. . . . . . . . . . . . . . . . . . . . . . . 141
4.16 TEM micrographs of supraparticles made from QD-CB[7] (a,b,c) and AuNC-
QD-CB[7] (d,e,f) at low concentrations of CB[7] (a,b,d,e) and higher con-
centrations (c,f). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
4.17 (a) A single inlet flow-focussing droplet generation junction and (b) the
resultant emulsion. (c) A double inlet flow-focussing droplet generation
junction and (e) the resultant emulsion. Both emulsions were generated
with overall flow rates 150/100 µL h−1 of oil and water respectively. . . . . 144
4.18 Optical micrographs of aqueous droplets containing (a) AuNCs, (b) QDs, or
(c) AuNCs and QDs dispersed in FC-40 containing 2 wt.% neutral triblock
and 1 wt.% RFCOOH. From left to right the evaporation of the emulsion
is imaged, followed by dispersal of the washed and dried structures into an
all aqueous environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
4.19 Photographs of solutions of QDs at pH7 and pH11. At pH 2 the QDs
precipitate due to their ligands becoming protonated. If base is added to
this precipitant, the QDs could be redispersed immediately. . . . . . . . . . 145
4.20 Optical micrographs of aqueous droplets containing (a) AuNCs and CB[7],
(b) QDs and CB[7], or (c) AuNCs, QDs and CB[7] dispersed in FC-40
containing 2 wt.% neutral triblock and 1 wt.% RFCOOH. From left to right
the evaporation of the emulsion is imaged, followed by dispersal of the
washed and dried structures into an all aqueous environment. . . . . . . . . 146
4.21 Optical micrographs of aqueous droplets containing (a) AuNCs, (b) QDs, or
(c) AuNCs and QDs dispersed in FC-40 containing 2 wt.% neutral triblock
surfactant and 1 wt.% RFNH2. From left to right the evaporation of the
emulsion is imaged, followed by dispersal of the washed and dried structures
into an all aqueous environment. . . . . . . . . . . . . . . . . . . . . . . . . 147
4.22 Optical micrographs of aqueous droplets containing (a) AuNCs and CB[7],
(b) QDs and CB[7], or (c) AuNCs, QDs and CB[7] dispersed in FC-40
containing 2 wt.% neutral triblock surfactant and 1 wt.% RFNH2. From
left to right the evaporation of the emulsion is imaged, followed by dispersal
of the washed and dried structures into an all aqueous environment. . . . . 148
4.23 Optical micrographs of aqueous droplets containing (a) AuNCs, (b) QDs, or
(c) AuNCs and QDs dispersed in FC-40 containing 2 wt.% neutral triblock
surfactant. From left to right the evaporation of the emulsion is imaged,
followed by dispersal of the washed and dried structures into an all aqueous
environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
21
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
4.24 Optical micrographs of aqueous droplets containing (a) AuNCs and CB[7],
(b) QDs and CB[7], or (c) AuNCs, QDs and CB[7] dispersed in FC-40
containing 2 wt.% neutral triblock surfactant. From left to right the evap-
oration of the emulsion is imaged, followed by dispersal of the washed and
dried structures into an all aqueous environment. . . . . . . . . . . . . . . . 150
4.25 Photographs of the pH indicator methyl red as aqueous solutions in neutral
and acidic water. Photographs are also shown of the emulsion generated
from microfluidics using either the neutral triblock surfactant (2 wt.%), or
the neutral triblock and RFCOOH (1 wt.%). . . . . . . . . . . . . . . . . . 151
4.26 Optical micrographs showing the repeated hydration and dehydration of
the dried microparticles. Water (100 µL) was added by micropipette and
allowed to evaporate under ambient conditions. The microparticles were
composed of either (a) QDs and CB[7], or (b) QDs, AuNCs, and CB[7].
They were generated with the FC-40 oil phase containing 2 wt.% neutral
triblock and 1 wt.% RFCOOH. . . . . . . . . . . . . . . . . . . . . . . . . . 152
4.27 SEM images of the microparticles after five hydration and dehydration
cycles. (a) microparticles of QDs and CB[7] and (b) a close-up of the
highlighted crack; (c) a microparticle of QDs, AuNCs and CB[7] with a (d)
close-up of the highlighted crack. . . . . . . . . . . . . . . . . . . . . . . . . 153
4.28 BSE images of the (a) QD-CB[7] and (b) QD-AuNC-CB[7] microparticles
with a beam power of 10 kV. . . . . . . . . . . . . . . . . . . . . . . . . . . 153
4.29 A summary table showing SEM images for various microstructures. . . . . . 154
4.30 A summary table showing close-up SEM images of the surface of various
microstructures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
4.31 Selected SEM images and their corresponding EDX elemental maps of rel-
evant elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
4.32 Extracts of EDX spectra obtained for each microstructure imaged by SEM,
with assigned peaks corresponding to each element observed. . . . . . . . . 157
4.33 (a) Optical micrographs of the QD-CB[7] microparticles generated as a
suspension in FC-40. (b) Optical micrographs and (c) fluorescence imaging
of a dried and washed microparticle before and after addition of water. . . . 158
4.34 Fluorescence micrographs of a glass slide, and of the QD suspension on the
glass slide. Blue light (λ = 460 - 490 nm band-pass) was used for excitation,
and a FITC filter (λ = 515 - 550 nm) was used in detection. . . . . . . . . . 158
4.35 Optical and fluorescence micrographs following the self-assembly process
of (a,b) QDs and (c,d) QDs and AuNCs within droplets upon drying, and
their subsequent rehydration. The FC-40 oil phase contained 2 wt.% neutral
triblock and 1 wt.% RFCOOH. . . . . . . . . . . . . . . . . . . . . . . . . . 159
4.36 Optical and fluorescence micrographs following the self-assembly process of
(a,b) QDs and CB[7], and (c,d) AuNCs, QDs and CB[7] within droplets
upon drying, and their subsequent rehydration. The FC-40 oil phase con-
tained 2 wt.% neutral triblock and 1 wt.% RFCOOH. . . . . . . . . . . . . 160
22
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
4.37 The PL spectra for the QDs before and after addition of small amounts of
CB[7], and the same for the AuNC-QD mixed dispersions (left). A zoomed
in spectrum of just the AuNC-QD aggregates (right) is also shown. . . . . . 161
A.1 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for NpVio2. For COSY, HSQC and HMBC positive peaks are
shown in blue and negative peaks are shown in red. . . . . . . . . . . . . . . 183
A.2 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the NpVio2-2CB[8] complex. For COSY, HSQC and HMBC
positive peaks are shown in blue and negative peaks are shown in red. . . . 184
A.3 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the NpVio2-1CB[7] complex. For COSY, HSQC and HMBC
positive peaks are shown in blue and negative peaks are shown in red. . . . 185
A.4 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the NpVio2-2CB[7] complex. For COSY, HSQC and HMBC
positive peaks are shown in blue and negative peaks are shown in red. . . . 186
A.5 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the NpVio2-1CB[8]-1CB[7] complex. For COSY, HSQC and
HMBC positive peaks are shown in blue and negative peaks are shown in
red. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
A.6 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the NpVio2-2CB[8] and 2-naphthol complex. For COSY,
HSQC and HMBC positive peaks are shown in blue and negative peaks are
shown in red. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
A.7 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for AzoVio2. For COSY, HSQC and HMBC positive peaks are
shown in blue and negative peaks are shown in red. . . . . . . . . . . . . . . 189
A.8 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the AzoVio2-1CB[8] complex. For COSY, HSQC and HMBC
positive peaks are shown in blue and negative peaks are shown in red. . . . 190
A.9 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the AzoVio2-2CB[8] complex. For COSY, HSQC and HMBC
positive peaks are shown in blue and negative peaks are shown in red. . . . 191
A.10 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the AzoVio2-1CB[7] complex. For COSY, HSQC and HMBC
positive peaks are shown in blue and negative peaks are shown in red. . . . 192
A.11 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the AzoVio2-2CB[7] complex. For COSY, HSQC and HMBC
positive peaks are shown in blue and negative peaks are shown in red. . . . 193
A.12 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the AzoVio2-1CB[8]-1CB[7] complex. For COSY, HSQC and
HMBC positive peaks are shown in blue and negative peaks are shown in
red. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
23
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
A.13 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for Np2Vio. For COSY, HSQC and HMBC positive peaks are
shown in blue and negative peaks are shown in red. . . . . . . . . . . . . . . 195
A.14 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the Np2Vio-CB[8] complex. For COSY, HSQC and HMBC
positive peaks are shown in blue and negative peaks are shown in red. . . . 196
A.15 ITC titrations of methyl viologen (1 mM) into CB[8] (51 µM), and a com-
petitive binding titration of ADA.HCl (1 mM) into methyl viologen-CB[8]
(1 mM and 51 µM). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
A.16 Stacked 1H NMR spectra of the AzoVio2-1CB[8] complex following its
photo-isomerisation. From the bottom the equilibrium state is shown, then
predominately the Z isomer after 4 h UVA irradiation, then predominately
the E isomer after heating. Peaks labelled blue correspond to the Z isomer. 198
A.17 Stacked 1H NMR spectra of the AzoVio2-2CB[8] complex following its
photo-isomerisation. From the bottom the equilibrium state is shown, then
predominately the Z isomer after 4 h UVA irradiation, then predominately
the E isomer after heating. Peaks labelled red correspond to the E isomer,
and those in blue to the Z isomer. . . . . . . . . . . . . . . . . . . . . . . . 199
A.18 Stacked 1H NMR spectra of the AzoVio2-1CB[7] complex following its
photo-isomerisation. From the bottom the equilibrium state is shown, then
predominately the Z isomer after 4 h UVA irradiation, then predominately
the E isomer after heating. Peaks labelled red correspond to the E isomer,
and those in blue to the Z isomer. . . . . . . . . . . . . . . . . . . . . . . . 200
A.19 Stacked 1H NMR spectra of the AzoVio2-2CB[7] complex following its
photo-isomerisation. From the bottom the equilibrium state is shown, then
predominately the Z isomer after 4 h UVA irradiation, then predominately
the E isomer after heating. Peaks labelled red correspond to the E isomer,
and those in blue to the Z isomer. . . . . . . . . . . . . . . . . . . . . . . . 201
24
List of Tables
1.1 A summary table of the different types of non-covalent interactions, with
their typical energies for a single interaction.38 . . . . . . . . . . . . . . . . 26
1.2 A table of the dimensions of different CB[n]s.1 . . . . . . . . . . . . . . . . 30
2.1 A summary table of the calculated material properties of the images shown
in Fig. 2.26 using the OpenCapsule software.22 . . . . . . . . . . . . . . . . 67
3.1 A table showing the calculated ρ, DP , and polymer molecular weight (Mn)
for the NpVio2 molecule complexed with two CB[8] molecules (3352 g
mol−1), assuming a Ka of 106.24 . . . . . . . . . . . . . . . . . . . . . . . . 80
3.2 Collated DOSY results for the supramolecular assemblies formed in units
of 10−10 m2 s−1. The diffusion coefficient (D) has been derived from the
most intense peak in the spectrum, with highest error being ±0.01. Each
measurement was undertaken at 5 mM concentration, with the exception
of *CB[8] being a filtered saturated solution c.a. 0.5 mM. . . . . . . . . . . 96
3.3 A table of hydrodynamic diameters (nm) of the complexes formed calculated
from D, derived from DOSY NMR experiments. Calculations were carried
out with the Stokes-Einstein equation, at 25 ◦ C. . . . . . . . . . . . . . . . 96
3.4 A table summarising the binding constants for CB[7] or CB[8] to each AB2
molecule studied with Ka in units of M
−1, ∆H and T∆S in units of kJ
mol−1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.1 A table showing the calculated PLQE for QD-CB[7] and AuNC-QD-CB[7]
aggregates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
A.1 A table of the fitting parameters used in an Ostwald ripening model to fit
DLS kinetics data of supraparticle growth. . . . . . . . . . . . . . . . . . . . 202
25
Chapter 1
Introduction
In this introductory Chapter, the fundamental principles behind this thesis have been
outlined. More specific introductions have been provided at the start of each results
chapter to give a comprehensive literature review of each research topic.
1.1 Supramolecular Chemistry
Supramolecular chemistry transcends that of atoms and molecules, describing their inter-
actions with each other and their environment. Derived from the latin ‘supra’ meaning
above or beyond, it is a highly complex field that is essential to almost every biological
process. Whilst the individual structure of molecules is highly important, it is the interac-
tions with their surroundings and other molecules that drive their function. For example,
protein structures are a complex combination of supramolecular interactions in water,
derived from their precise amino acid composition.38
The importance of supramolecular interactions has been highlighted many times, with
the 1987 and 2016 Nobel Prizes in chemistry being awarded for significant advances in
this field. In 1987 the award was for the “syntheses of molecules that mimic important
biological processes”. The first molecular receptors were designed for the highly specific
binding of a range of materials, entirely by non-covalent interactions.39 In 2016 the award
was for the development of “the world’s smallest machines”. Significant advances in the
understanding of molecular architectures have allowed true nanomachines to be fabricated
that could carry out real mechanical work on a molecular level.40
Supramolecular chemistry relies on the intermolecular bond, or more specifically non-
covalent interactions.38,41 Non-covalent bonds are not as directional or stable as covalent
bonds, which have bond strengths in the region of 150-450 kJ mol−1. A summary of the
different interactions available to the supramolecular chemist are shown in Table 1.1 with
their typical interaction strengths.
van der Waals pi − pi Dipole-Dipole Cation-pi H -bonding Ion-Dipole Ion-Ion
Energy/kJ mol−1 <5 0-50 5-50 5-80 4-120 50-200 200-300
Table 1.1 A summary table of the different types of non-covalent interactions, with their
typical energies for a single interaction.38
26
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
The weaker non-covalent bonds can be reversibly broken and reformed either spon-
taneously or in response to changes in conditions, such as temperature. This allows the
design of dynamic chemical systems, that can perform specific tasks on demand, among
many other unique properties.38,41 Other factors previously unconsidered, such as the
continuous phase for the supramolecular assemblies, also play an important role. For
example, hydrogen-bonds (H -bonds) between molecules in water will encounter much
competition from surrounding water molecules. In comparison, traditional covalent bonds
are relatively unaffected by their environment, often requiring harsh conditions to drive
macroscopic changes in properties. It becomes clear, however, that in order to construct
useful materials the bond directionality and stability generally needs to be improved for
non-covalent bonds as they can be stochastically orientated. In nature, and now synthet-
ically, this is achieved by co-operativity where the additive effect of multiple non-covalent
bonds gives rise to strong and directional interactions.38
1.1.1 Host-Guest Chemistry
One of the most effective methods of achieving co-operativity is through the use of ‘host-
guest’ chemistry. By considering a binding event between molecules (c.f. complexation),
the ‘host’ will bind the ‘guest’ to form a complex entirely based upon non-covalent inter-
actions.38 The co-operativity of binding is a result of multiple non-covalent interactions
driving host-guest complexation, and these can be a combination of different non-covalent
effects. Synthetic host molecules have mainly been macrocyclic in nature, with a cavity
volume of distinctly different chemical environment to the bulk media where guests can
bind. As a result, host-guest binding is an additive effect of co-operativity and the pre-
organisation effect. The pre-organisation of macrocycles results in the entropic cost to be
in the ideal conformation for host-guest binding being accounted for in the macrocycle
synthesis, thereby not being an energy barrier to host-guest binding.38
Host-guest complexation, whilst considerably stronger than individual non-covalent
interactions, remains in dynamic equilibrium. The binding strength is different for every
combination of host and guest, relying on the spatial fit of guest within the host, and a
combination of favourable non-covalent interactions. Examples include hydrophobic guests
being encapsulated in a hydrophobic cavity, or electrostatically-charged guests forming ion-
dipole pairs with the host. The equilibrium can be defined as in Eq. 1.1 for the formation of
a simple 1:1 complex, where the strength of binding is defined as the association constant,
Ka.
Hostsolvated +Guestsolvated −−⇀↽− Host ·Guest+ xSolvent
Ka =
[H ·G]
[H][G]
(1.1)
From a thermodynamic standpoint, the enthalpy of binding is typically high due to
the formation of multiple co-operative non-covalent bonds. Entropy can also be a key
factor in complexation, by releasing multiple solvent molecules associated with both the
host and guest, or triggering significant conformational changes in guest or host.
27
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Measurement of these binding constants has been carried out by varied techniques such
as NMR spectroscopy, UV-vis spectroscopy, and isothermal titration calorimetry (ITC). In
each case, a concentration range that contains significant amounts of bound and free host
and guest is required. If binding is too strong to be measurable for a particular technique,
then a poorer binding guest of known Ka can be introduced to reduce the observed Ka in
a competitive binding titration.38
NMR spectroscopy is also affected by the rate of association and disassociation, requir-
ing extra considerations. Analysis is then typically carried out with least squares fitting
with an appropriate binding model. However, it is only useful for binding constants in
the region of 10 − 104 M−1, otherwise competing guests of known binding are required.
UV-vis spectroscopy relies on a UV active species being involved in the binding of host
and guest. A plot of the absorbance intensity versus concentration of guest titrated an
then be fitted to various models, allowing determination of Ka.
38
ITC operates by careful measurement of the heat (i.e. enthalpy) that is released from
a mixture of host and titrated guest. The gradient of the ITC binding isotherm can then
be fit with various models to provide Ka, ∆G, ∆H, ∆S, and the stoichiometry of binding.
ITC is useful for binding constants within the range 102 − 107 M−1, and sees widespread
use in biochemistry.38,42
Figure 1.1 Chemical structures of some common synthetic macrocyclic host molecules.
The most common organic macrocycles are shown in Fig. 1.1. Some of the first synthes-
ised macrocycles were crown ethers43 and cryptands,44 which see industrial application
in the selective sensing of metal ions. These are used primarily in organic solvents as
their binding is interrupted by highly polar water, and bind selectively to metal ions by
a combination of size and ion-dipole interactions. Calixarenes45 and pillararenes46 have
been investigated in both organic and aqueous media through the introduction of solu-
bilising groups. These operate on a combination of van der Waals, spatial confinement,
solvophobic, and pi electron based interactions.38,45,46
Cyclodextrins (CDs)47 have seen a huge amount of interest as they are based on sugars,
are water-soluble, and are biocompatible. Villiers and Schardinger pioneered the study of
CDs in 1891 and the early 1900s, with their host-guest chemistry seeing much focus from
the 1940s.38,47 In water, they have a hydrophobic cavity, so a combination of solvophobic
and van der Waals effects drives the binding. Different sizes exist (6, 7 or 8 membered
rings) with varying binding properties as the cavity increases in size. The rich host-guest
chemistry of cucurbit[n]urils48 in water has been the focus of this thesis.
28
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
1.1.2 Cucurbit[n ]urils
The first synthesis of the host macrocycle cucurbit[6]uril (CB[6]) is dated to 1905 by
Behrend, but the first characterisation and structural identification was carried out by
Freeman et al. in 1981.49 The authors named this macrocycle, constructed from six
glycoluril units linked by twelve methylene bridges, cucurbituril for its resemblance to
a pumpkin (which belongs to the Cucurbitaceae family). Further investigations by Kim
and Day led to a family of CB[n]s being discovered: CB[5], CB[6], CB[7], CB[8] and
CB[10].1,48,50–52
The new synthesis was milder than that by Behrend, proceeding by the condensation
of glycouril and formaldehyde in concentrated HCl to give a mixture of all CB[n]s (Fig.
1.2).48 This showed that CB[6] is the thermodynamic product for the condensation, and
other homologues are kinetically trapped. The mechanism for its synthesis likely proceeds
by a templated reaction, where H3O
+ H -bonds to the carbonyl on the top and bottom of
glycouril units, and oligomerisation occurs before a ring closure step.48,51,52
Figure 1.2 The synthesis of a mixture of CB[n], and space-filling models demonstrating
the increasing size and constant height as n increases. Adapted from references.1
The CB[n] molecules are chemically inert, rigid and thermally stable. They exhibit
partial solubility in water, higher for CB[5] and CB[7] (20-30 mM) than for CB[6] and
CB[8] (< 0.5 mM), and higher when encapsulating hydrophilic guests.1,51 As n increases,
the diameter and cavity volume also increase, whereas the height remains constant at
9.1 A˚ including the van der Waals radii of the oxygen portals (see Table 1.2).1 The cavity
is hydrophobic providing a van der Waals and solvophobic driving force for binding. It is
surrounded by highly polar and electronegative carbonyl portals, to which H -bonding or
electrostatic interactions can occur.1 The cavity is also very rigid, allowing crystal packing
(ideal packing coefficient of 55 %) to be a further driving force.53 Lastly, there is the
displacement of water from within the CB[n] cavity and associated with guest molecules
29
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
as a key driving force of binding, due to the enthalpic recovery of H -bonds with water and
the entropic release of these multiple conformationally-restricted solvent molecules.54–56
CB[5] CB[6] CB[7] CB[8] CB[10]
Portal diameter (A˚) 2.4 3.9 5.4 6.9 9.5-10.6
Cavity diameter (A˚) 4.4 5.8 7.3 8,8 11.3-12.4
Cavity volume (A˚3) 82 164 279 479 870
Outer diameter (A˚) 13.1 14.4 16.0 17.5 20.0
Height (A˚) 9.1 9.1 9.1 9.1 9.1
Table 1.2 A table of the dimensions of different CB[n]s.1
The smallest CB[5] can encapsulate gases such as N2, and can bind metal or small
organic cations to its carbonyl portals.1 CB[6] and CB[7] can bind small molecules within
their cavity and portals, with CB[6] also able to encapsulate some gases.1 Freeman first
demonstrated the formation of 1:1 host-guest inclusion complexes between CB[6] and
aliphatic amines (Ka = 10
6 M−1),49 and there now exists hundreds of examples of highly
specific interactions with high Ka for both CB[6] and CB[7].
1 The highest binding con-
stants are reported for dicationic molecules that can pack well within the cavity, and
form ion-dipole interactions with the portals such as adamantane derivatives or ferro-
cenes (Ka = 10
15 M−1).57,58 This strong binding in water is on the order of the biotin-
streptavidin association often used in biochemistry, but through a single binary complex-
ation.
CB[8] has a cavity volume similar to that of γ-CD, and as a result can bind multiple
small molecules within its cavity along with having similar binding properties of single mo-
lecules to CB[6] and CB[7]. This allows the formation of host-guest complexes with either
two different guests, heteroternary, or two of the same guests, homoternary. Two compon-
ents inside the cavity can interact to further drive complexation through favourable pi−pi
stacking of aromatics inside the cavity, and in the formation of charge transfer complexes
between electron-poor and electron-rich aromatic guests.1,59 In general, the electron-poor
guest will contain a cationic quaternary nitrogen, where a symmetric dicationic guest such
as viologen will have favourable electrostatic interactions with the carbonyl portals. An
electron rich guest such as naphthalene or anthracene derivatives will then bind to form
a heteroternary charge-transfer complex. If the first guest is instead asymmetric with one
cationic species, then a homoternary complex will be formed. There exist a further range
of host-guest binding architectures as shown in Fig. 1.3.
The strong (Ka up to 10
14 M−2)1 binding of multiple guests inside CB[8] has allowed
the dynamic coupling of materials for diverse applications. If guests for CB[8] are cova-
lently attached to polymer backbones, then the addition of CB[8] as a cross-linker can
lead to dynamic supramolecular polymer-based networks.2,60–62 In water these networks
form hydrogels, gels that are over 95 % water in their composition. The resultant mater-
ial properties are a combination of the entanglement of the covalent polymer chains, but
primarily of the CB[8] cross-links. The gels have been shown to exhibit several favourable
properties of supramolecular systems, such as fast dynamics at higher temperatures that
allows for facile processing.60 They can also be shear-thinning, where high shear forces
30
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 1.3 A scheme showing the different types of host-guest binding architectures
available with CB[n]s. Adapted from references.1
will reversibly break apart the network, which is highly favourable in the field of inject-
able hydrogels for efficient drug delivery, as shown in Fig. 1.4.2 In addition, CB[8]-based
materials can be responsive to multiple different stimuli that can trigger changes in the
material properties.19
Figure 1.4 A scheme showing the formation of hydrogels from polysaccharide polymers
functionalised with phenylalanine amino acid guests via homoternary complexation. Ad-
apted from references.2
If stimuli-responsive guests are used, then the polymer networks can be broken apart
on demand. Azobenzene is a photo-responsive electron-rich guest for CB[8], where it will
isomerise from its E to Z isomer in response to UVA light (λ ∼ 350 nm) and can be
returned to its initial thermodynamic state by blue light (λ ∼ 420 nm) or by heat.63–65
The elongated E isomer fits well in the presence of an electron deficient guest, however
it becomes too bulky upon isomerisation for both guests to fit in the cavity, causing the
breakdown of the complex. Another example is the chemical response of CB[8] complexes,
where introduction of a guest with higher Ka can compete and irreversibly break the
existing complex.62
Guests have also been prepared on inorganic surfaces and nanoparticles for a variety
of applications relying on significantly altering the surface chemistry versus that of the
bulk material.1 CB[n] can complex with guests on surfaces to form pseudorotaxanes,
modulating access to the surface for electroactive species.66 They have also been used
in the construction of nanovalves, where CB[6] psuedorotaxanes prepared on mesoporous
silica can block the release of entrapped small molecules.67 After increasing the pH, the
quaternary amine guests holding CB[6] in place lose their charge, releasing the CB[6] and
31
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
the trapped molecules. Furthermore, by forming a CB[8] rotaxane on a surface, even
micelles can be immobilised by heteroternary complexation.68 CB[n] itself can associate
with metal surfaces through its carbonyl portals acting as a multidentate ligand, which
has been explored primarily with Au surfaces and in sensing applications. This has lead
to controllable surface modification with CB[n] for further construction of supramolecular
materials, reviewed in more detail in Chapter 4, section 4.1.2.
Figure 1.5 A scheme showing the formation of hybrid polymer-nanoparticle microcapsules
from water-in-oil emulsions. Adapted from references.3
Liquid-liquid interfaces have also been explored for templating CB[n] host-guest chem-
istry. The formation of polymer-based supramolecular networks at the interface of water-
in-oil emulsions has been utilised in the formation of microcapsules for encapsulation of
macromolecules.3,20,21,69–72 Utilising droplet emulsions generated with microfluidic devices,
dilute semi-viscous solutions of CB[8] and polymers/nanoparticles with pendant guests
were introduced into aqueous droplets in a continuous oil phase.3 The dilute CB[8] con-
jugates were then localised to the droplet interface, whereupon the concentration reached a
critical gelation point and a transition to an interfacial gel occurred. This is shown schem-
atically in Fig. 1.5. After subsequent drying, the microcapsules could be redispersed
into an entirely aqueous environment. The localisation of material to the interface has
been carried out by different methods, by surfactant-induced electrostatic assembly,20,21
by Pickering emulsion,3,69 or by solvophobicity.70 This method holds significant advant-
ages over commercial microencapsulation technologies, such as 99 wt.% cargo loadings,
and fabrication being achievable in one step with controllable permeability.3
1.2 Emulsions
Emulsions are dispersions of one immiscible liquid in another, being extremely common
in nature and around us in our daily lives.4 There are two distinct types of emulsion: mi-
croemulsions and macroemulsions. Microemulsions are homogeneous and thermodynamic-
ally stable mixtures of two immiscible fluids and surfactants. In contrast, macroemulsions
are thermodynamically unstable mixtures of two immiscible fluids that are only kinetically
stable. The degradation pathways of macroemulsions are shown in Fig. 1.6. Microemul-
sions are composed of very small droplets (d ∼ 10 nm), drawing an analogy to micelles
swollen with liquid, and are under fast dynamic exchange of their stabilising surfactants.
In contrast, macroemulsions are much larger (d > 1 µm), and the surfactants exchange
with solution with much slower dynamics. The amount of surfactant present is much lower
32
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
in macroemulsions.4 From here, macroemulsions will be referred to as emulsions.
Figure 1.6 A scheme showing the thermodynamically driven destabilisation pathways of
macroemulsions. Adapted from references.4
Surfactants, short for surface-active agents, are amphiphilic molecules that adsorb to
immiscible interfaces.4,38 They show these properties as different parts of the molecule have
opposite solvophobicities. For example, a surfactant to stabilise a water-in-oil emulsion
will have a long non-polar hydrocarbon tail, and a small polar hydrophilic head group.
This results in the hydrophilic portion interacting with water, and the hydrophobic portion
interacting with the oil. The driving force for surfactant assembly is a result of multiple
effects: allowing the amphiphilic molecule to be well solvated across its structure, and
acting to prevent the energetically unfavourable direct interaction of immiscible phases.
The direct interaction of immiscible phases is extremely energetically unfavourable in
emulsions due to the high interfacial surface area of droplets. The interfacial tension (γ) is
often used to describe the interfacial free energy, where it represents the amount of work
required to expand the interfacial area. It originates from the imbalance of attractive
forces between solvent molecules at the liquid-liquid interface, being a direct reflection of
the cohesive forces in each liquid. For instance, the water-octane interface has a higher
γ (and thus energetically unfavourable) of 51 mN m−1 than that of water-octanol at 8
mN m−1. This is because the hydroxyl group in octanol can face the water interface and
H -bond directly, whereas octane remains completely hydrophobic.4,73 Surfactants reduce
γ in a similar way by facilitating the interaction between immiscible phases.
Surfactants can broadly be grouped as electrostatic, steric, or particle-based.4,74 Their
assembly to the interface, and the way in which they act to prevent coalescence, differs.
Electrostatic stabilisation occurs with ionic surfactants, where a charged head group will
stabilise a polar/apolar interface. Coalescence is then prevented by the electric double
layer effect causing electrostatic repulsion between emulsion droplets (c.f. DLVO the-
ory).4 Steric stabilisation occurs with polymeric surfactants, often block copolymers, where
blocks of different solvophobicities are employed. Coalescence is prevented by decreases
in entropy when surfactant chains on neighbouring droplets begin to interact. These two
stabilisation methods are quite similar to those observed with colloidal dispersions of solid
33
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
nanoparticles.32 Particle based stabilisation (c.f. Pickering emulsions), however, rely on
solid particles much smaller than the emulsion droplets that are not well solvated by either
phase (ideal contact angle = 90◦) and so assemble at the interface.4,74 Their interaction
energies scale with their size, and are often irreversibly adsorbed. Commonly, multiple
surfactants of different types are employed to provide the best droplet stability.4
Some quantification of surfactant strength can be readily defined by the empirical
hydrophilic-lipophilic balance (HLB) as defined in Eq. 1.2, where Hw is the molecular
weight of the hydrophilic part of the molecule, and Lw is the molecular weight of the
hydrophobic part of the molecule.4,75 This results in an arbitrary value between 0 and 20,
with values below 10 allowing the formation of a water-in-oil emulsion and those above 10
allowing formation of an oil-in-water emulsion. If a value close to 0 or 20 is obtained then
the molecule is not amphiphilic and will not be a good surfactant.
HLB = 20× Hw
Hw + Lw
(1.2)
In order to form emulsions efficiently, surfactants must both be excellent at stabil-
ising the specific interface in question, but must also be able to assemble rapidly.4 The
process by which emulsions are generated significantly affects their uniformity and res-
ultant properties. The simplest way by which an emulsion can be formed is through
mechanical stirring or shaking in the presence of suitable surfactants.76 Whilst this is fast
and cheap, the emulsion droplets are highly varied in size, and droplets-within-droplets
or multiple emulsions are easily formed. A more controlled variant on this is through a
high shear homogeniser, but a distribution of droplet sizes remains. The formation of
exclusively hierarchical emulsions is also difficult to control effectively with bulk droplet
generation methods. The most controlled droplet generation achievable on the micro-
metre scale available is by microfluidic droplet generation, exemplified in Fig. 1.7 where
monodisperse hierarchical emulsions can be generated.5
Figure 1.7 Triple emulsions generated by coupling of multiple glass capillary microfluidic
devices. Adapted from references.5
1.2.1 Droplet Microfluidics
Microfluidic techniques are based on the manipulation of liquid flows on a micrometre
length scale, dealing with sub nanolitre volumes.6 Microfluidics have seen a huge surge
34
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
in research and their applications in many diverse areas due to the very small amount of
reagents required, low cost, short process times, and high resolution obtainable.77 High-
lighted applications include as highly controllable microreactors for chemical synthesis,
in the encapsulation and manipulation of cells, and in nanomaterials synthesis.6,77–79 A
particularly intricate ‘lab on a chip’ is shown in Fig. 1.8, highlighting the complexity
achievable to control cell populations down to individual cells.7
Figure 1.8 A microfluidic chemostat used to study the growth of microbial populations.
Intricate design and incorporation of pneumatic valves allows complex function. Adapted
from references.6,7
On this length scale fluid mechanics can be modelled and controlled, with ‘laminar flow’
being commonplace. At large length scales fluids will mix in a convective manner, where
inertia of liquids is the dominant factor in fluid behaviour.80 In microfluidic channels,
however, fluids do not mix, they will flow in parallel without any turbulence (laminar
flow), with the only cause of mixing being slower diffusion processes. This laminar flow
allows facile control over liquids, but if mixing is required then specific geometries to force
turbulence, such as with winding channels, must be incorporated.80 This process becomes
clear when considering the Reynolds number (Re), which relates inertia to viscosity by the
relation shown in Eq. 1.3, where ρ is the density of the fluid, U0 is the typical velocity, L0 is
the length scale in question, and η is the shear viscosity of the fluid.80 On the macroscale,
fluids have a high Re due to fast flows and large length scales and experience turbulent
flow as these inertial forces outweighs viscosity, whereas on the micrometre scale a very
low Re is observed and no turbulent flow will be seen.
Re =
ρU0L0
η
(1.3)
Microfluidic chips can be constructed from poly(dimethylsiloxane) (PDMS), glass or
other plastics.81–83 Optically transparent materials are generally required to allow monit-
oring and analysis by microscopy or spectroscopy. PDMS in particular is a soft elastomeric
material that is biocompatible, but sees most use due to its ease of synthesis for rapid
prototyping.81 Glass sees much use when harsher conditions are required such as chemical
or thermal stability, however it is much harder to rapidly prototype and to obtain smaller
features with most manufacturing achieved with micromachining.83 If a chip is to be used
35
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
in the long term, glass is also preferred as it is robust. Other plastic designs see use in
scaling up chip designs for industrial uses due to their cheap manufacture, but also lose
resolution in their features and may not be optically transparent.82
Figure 1.9 A schematic of the soft lithographic replica moulding technique to construct
PDMS microfluidic devices. (a) Contact photolithography is carried out through a printed
mask onto a thin film of photoresist on a Si wafer. (b) After developing the photoresist,
the desired pattern will be present on the Si wafer to form a ‘master’. (c) PDMS precursor
is poured onto the master and subsequently cross-linked. (d) The cross-linked PDMS can
be peeled off the master chip, and then bonded to a substrate of choice.
PDMS devices are constructed by soft lithographic replica moulding, shown schemat-
ically in Fig. 1.9.81,84,85 A design is made with computer-aided design software programs
(CAD), and either printed on transparent plastic at high resolution (5060 dpi for 20 µm
resolution) or a chrome mask is made (expensive).85 This is then used as a mask for con-
tact photolithography on a Si wafer coated with a photoresist (normally SU-8 negative
resist for microfluidic devices). After developing and removal of uncross-linked photores-
ist, the Si wafer will have a positive relief of the chip design, termed the ‘master’. PDMS
precursor is then casted on the positive relief and cured at high temperature to form a
PDMS replica. Being a soft elastomer, it is easily removed from the Si surface without
damaging the master, and sample introduction and exit points can be introduced by e.g. a
biopsy punch. Subsequently exposing the PDMS and an appropriate surface (e.g. a glass
microscope slide) to an air or O2 plasma and then pressing them together forms an irre-
versible seal, strengthened by further curing at high temperature. The surface chemistry
of the channels can be easily modified with the appropriate activated silane. Overall, this
means new ideas can be designed and tested within a few days, rather than in weeks and
at great expense if a chrome photomask is required.85
Droplet generating devices were first investigated in 2001 by Thorsen, with two immis-
cible phases intersecting at a T-junction of microchannels (see Fig. 1.10).8,76,77,86 Here, a
flow of one immiscible phase shears another to form a continuous stream of monodisperse
droplets. In order for droplet formation to take place the capillary number (Ca) must be
considered, a dimensionless quantity that relates viscous and interfacial forces.80 Contrary
to miscible phases, there now exist competing stresses at the interface between the liquids
acting to reduce the interfacial surface area, and viscous stresses acting to extend the
36
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
interface with the direction of flow. Ca is defined in Eq. 1.4, where droplet formation
occurs at a high Ca either through increased velocity (U0) or decreased interfacial tension
(γ). Each parameter can be readily controlled in a microfluidic device, allowing the gen-
eration of monodisperse droplets of controllable size, as shown in Eq. 1.5 where ddroplet
is the generated droplet diameter.80 Further control over shear forces can be gained by
altering the geometry for higher initial contact surface area between the immiscible fluids,
i.e. with flow-focussing devices.
Ca =
ηU0
γ
(1.4)
ddroplet ∝ 1
Ca
=
γ
ηU0
(1.5)
Figure 1.10 (a) T-junction, (b) planar flow-focussing and (c) capillary flow-focussing
geometries. Adapted from references.3,8
For the generation of droplets in microfluidic devices there exist two commonly used
methods: T-junctions and flow-focussing techniques.8,76,77 T-junction geometries rely on
the perpendicular injection of one immiscible phase into another, with droplet forma-
tion resulting from the induced shear forces and surfactant-based stabilisation.8,86 Flow-
focussing geometries rely on a liquid flowing in a central channel being sheared on the
outside by flows of an immiscible phase.8 The flow of droplets are then output (focussed)
into a smaller downstream channel. In this case the applied pressure and viscous stresses
result in the shear forces pulling the central fluid into a narrow strand, which breaks off
into droplets. These axisymmetric droplet generation techniques allow smaller droplets
to be generated due to higher shear forces.8 Overall this allows generation of droplets at
rates above 1 kHz, and with size distributions within 3 % of the average.8,76 The different
geometries are shown in Fig. 1.10. By having multiple droplet generating junctions in
series, hierarchical emulsions can easily be generated.
Of particular note is the abundance of perfluorocarbon oils in use for droplet micro-
fluidics.9 This is because of the near complete immiscibility of water and perfluorocarbon
37
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
oils, as they are both hydrophobic and lipophobic. They are also chemically inert and have
very low solubilities of non-fluorinated materials that may be encapsulated in the droplets,
making them an ideal carrier phase. In addition, they have a high gas permeability, which
is of significant interest for the encapsulation of cell cultures.9,10,75,87,88
Figure 1.11 Structures of developed triblock copolymers to stabilise water-in-
perfluorocarbon emulsions.9,10
With PDMS microfluidic chips, perfluorocarbon oils can easily be employed in droplet
generation by coating the internal surface of the chip with a perfluorocarbon silane. Al-
most all of the most commonly used perfluorocarbon surfactants are based on a widely
available perfluoropolyether backbone, with most effective droplet stabilisation achieved
with triblock copolymers with poly(ethylene glycol) or poly(glycerol) as shown in Fig.
1.11.9,10 The perflourinated blocks form a steric barrier to droplet coalescence, and the
hydrophilic block provides a chemically inert inner droplet surface.
1.3 Aim of This Thesis
CB[n]s have been the focus of varied research since the discovery of their different mac-
rocycle sizes in 2000 in various different fields due to the possibility of highly controlled
supramolecular chemistry in water.1,48 Initial steps have been made towards gaining con-
trol in the synthesis of functional supramolecular materials, taking advantage of the strong,
specific, yet responsive nature available from their host-guest interactions. In addition,
PDMS-based droplet microfluidics is now a widely available technique, and the study of
self-assembly processes within droplet emulsions has not yet been fully explored. The
aim of this thesis is to further push the boundaries of current knowledge surrounding
CB[n] in order to construct new materials with desirable properties, such as self-healing
supramolecular networks and aggregates of nanocrystals that retain their activity.
Only a few studies have been made in the generation of small molecule-based supra-
molecular polymers from CB[n]s, primarily focussed on linear architectures, and often
with limitations. In Chapter 2, an in-depth investigation into multicomponent branched
supramolecular polymers was carried out, in order to form truly extended polymer net-
works. The limitations of these assemblies in bulk, and the use of emulsions to drive the
supramolecular polymerisation past these limitations is described. Leading directly on
from this work, Chapter 3 details a study into the effect of molecular architecture on the
formation of intermolecular or intramolecular complexes in bulk solution. The specific ar-
chitecture employed resulted in the discovery of cooperative intramolecular complexes, and
38
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
their application in self-sorting mixtures and response to various stimuli was investigated.
Finally, Chapter 4 describes the use of CB[n]s in generating new hybrid materials from
semi-conducting nanocrystals by directed assembly. The assemblies were studied in bulk
solution, and confined within emulsions in order to generate new materials at the nano-,
micro- and macro-scale. The discovered process held many advantages over the current
state-of-the-art for controllable assemblies of nanocrystals, such as being fast (limited by
mixing time), carried out in water, and under mild conditions.
39
Chapter 2
Aqueous Interfacial Gels
Assembled from Small Molecule
Supramolecular Polymers
This work has been published in the following article:
A. S. Groombridge, A. Palma, R. M. Parker, C. Abell and O. A. Scherman, Chem.
Sci., 2017, 8, 1350–1355.
Many of the foundations for this work were established in the following article:
A. R. Salmon, R. M. Parker, A. S. Groombridge, A. Maestro, R. J. Coulston, J. Hege-
mann, J. Kierfield, P. Cicuta, O. A. Scherman, C. Abell, Langmuir, 2016, 32, 10987-10994.
2.1 Introduction
2.1.1 Branching Polymerisations
Covalent polymers are macromolecules that consist of repeating units, called monomers.
They were first recognised in 1832 by Berzelius,11 and synthetic polymers have completely
transformed the world we live in today. Their long chain structure gives rise to unique fa-
vourable physical properties, such as increased strength, elasticity and inertness.89 Nature
has exploited these macromolecules for millennia in all their essential functions, being
DNA, proteins, and in plant structures such as wood.
The physical and chemical properties of polymers can be infinitely modulated, with
ground-breaking new advances in synthetic polymers still being made today. The depar-
ture from linear architectures in particular has been of great interest for many years.11
Branching in polymers brings about a significant change in polymer properties, primarily
as they decrease the degree of crystallinity in a polymer.11,90 This is because branched
polymers can not pack as easily into ordered lattices as their linear homologues, and
thereby gives rise to other non-linear properties. Branching can be taken to two extremes,
40
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
one where all polymer chains are branched and linked together resulting in a completely
cross-linked polymer (i.e. a gel). Taken to a different extreme, where each branch res-
ults in more branches, tendency towards a dendritic architecture occurs that does not
gelate.11,89–91 This is exemplified in Fig. 2.1, where different dendritic structures are
shown.
Figure 2.1 A series of schematic diagrams showing some of the range of dendritic and
hyperbranching architectures possible. Adapted from references.11
Dendrimers display many favourable properties as a result of their highly globular and
non-entangling structure. They also posses a large proportion of terminal functional groups
exposed to the external chemical environment. The non-entangling structure gives rise to
a much reduced solution viscosity as compared to their linear analogues. This results in
much higher solubility in various solvents due to the loss of crystallinity and rigidity in the
polymer, making them much more processable for use in various industries.11 Tuning the
terminal groups also provides control over solubility. Further applications lie as additives in
the oil industry for viscosity control of formulations, and in non-linear optical materials.90
Their key limitation, however, is that they are prepared by multi-step syntheses with
tedious isolation and purification.92 On the other hand, hyperbranched polymers can be
prepared in a single step, retaining many of the favourable properties of dendrimers.11,90
For step growth (c.f. condensation growth) branching polymerisations, whereby a
terminal group A will react with another terminal group B to form a bond A−B between
monomers, the theory behind branching and cross-linked polymers has been well-described
and can be readily applied to supramolecular systems. In the 1940s, Flory developed a
statistical mechanical method to describe these types of polymerisations, with the relevant
material outlined below.11,89–91
αc =
1
(f − 1) (2.1)
The gelation point is the extent of reaction (ρ, c.f. conversion) at which the polymer
molecular weight approaches infinity, forming a cross-linked network of macroscopic size.
41
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
If two monomers are used with a single branching point, i.e. A2 and B3, the degree of
branching and the gelation point can be calculated if it is assumed that intramolecular
reactions do not occur, and A can only react with B. The branching coefficient α is
defined as the probability that a chain will end in a branching unit, and f is defined as the
number of reactive functionalities on a branching monomer. When α(f − 1) > 1, a cross-
linked network or gel is obtained. Therefore the point at which the degree of branching
observed will result in a gel can be defined as the critical branching coefficient, αc, as in
Eq. 2.1.89–91 Thus, when using monomers A2 and B3, f = 3 and αc = 0.5. α can then be
related to the conversion of each functional group (i.e. ρa and ρb), as in Eq. 2.2.
90,91
α =
ρ2a
r
= rρ2b , r =
[B]
[A]
(2.2)
If the stoichiometry of A and B functionalities is the same, then α = 0.5 at ρ = 0.71.
This means that after 71 % of A groups have reacted with B groups, a fully cross-linked
network will be formed. Importantly, if the polymerisation is stopped prior to this point,
a hyperbranching architecture will be observed. In contrast, if a similar single branching
monomer is used containing both A and B, such as AB2, then Eq. 2.3 applies.
90,91
α =
ρa
(f − 1) = ρb (2.3)
In this case, at 100 % conversion of the A functionality, f = 3, ρb = 0.5 and α = 0.5.
This means that statistically, the critical branching coefficient is never reached and gel
formation does not occur.89–91 This is shown graphically in Fig. 2.2.
Figure 2.2 A graph showing the calculated branching coefficient, α, as the conversion
of A functionality progresses. The solid line shows the polymerisation of A2 with B3,
where the stoichiometry of A to B is 1. The dashed line shows the polymerisation of a
single branching monomer, AB2. The red line shows the point at which gelation will be
observed, at α = 0.5.
Since this theory has been outlined in the 1940s, in covalent chemistry there has been
extensive work towards gaining precise control over polymer 3D architecture, with count-
less examples of cross-linked networks and hyperbranching polymers.11,89,90 The growing
field of non-covalent, supramolecular polymers has largely considered linear architectures
thus far, with more recent developments focussing on branching and cross-linked polymers.
42
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
2.1.2 Supramolecular Polymers
The field of supramolecular polymers developed at the interface of polymer science and
supramolecular chemistry from the idea of developing programmable and adaptive ma-
terials.93 ‘Supramolecular polymers’ are polymers that are composed of monomers held
together purely by directional non-covalent interactions.12–14,93–95 These supramolecular
bonds between monomers are weaker than covalent bonds, resulting in less chemically and
mechanically robust polymers. A key advantage to the supramolecular bond, however, is
the access to adaptive and responsive materials.94 As the supramolecular bonds are inher-
ently reversible, stimuli-responsive and reversible, supramolecular polymers will provide
access to the next generation of advanced materials.
Design of the monomer structure used in supramolecular polymerisation is key to con-
trolling the overall polymer structure, with new factors that should be taken into account
compared to traditional polymerisations. These include the lifetime of the supramolecu-
lar interactions (i.e. kinetics), their bond strengths (i.e. thermodynamics) and potential
conformations (i.e. entropy). The bond strength is defined by the equilibrium association
constant between the terminal groups on monomers, Ka. As a result, supramolecular
polymerisations are under thermodynamic equilibrium. In contrast to covalent polymers,
the degree of polymerisation (DP ) for supramolecular polymers heavily depends on Ka
and thus the concentration.12 For example, by simply increasing the concentration for a
supramolecular polymerisation of a certain Ka, a higher degree of polymerisation (DP )
will be expected. This is a simple example of the inherent adaptability of supramolecular
polymers.12,93,94
As Ka is highly temperature sensitive, a strong temperature dependence on mechanical
properties, and thus extent of polymerisation, will be observed. At higher temperatures,
faster dynamics of binding will lead to faster stress relaxation. This means that supra-
molecular polymers have the potential to be viscoelastic at low temperature and liquid-like
at high temperature, so they can be easily processed into materials.93 Their response to
applied stress contrasts to covalent polymers, which typically relieve stress by reptation
in their chains. However, supramolecular polymers can simply break and reform at high
stress points, resulting in viscoelastic characteristics such as shear-thinning rather than
fracturing.95 Another result of their reversible bonding is the access to self-healing mater-
ials, where simply pressing broken polymers back together and waiting for equilibration
can be enough to heal the polymer back to its original mechanical properties.13,93
The mechanism of supramolecular polymerisations shares some similarities with those
of covalent polymerisations, allowing a parallel to be drawn between step-growth poly-
merisation pathways. However, covalent polymerisations are predominantly under kinetic
control, where reactions between monomers are considered irreversible. Supramolecular
polymerisations are under thermodynamic control due to their reversibility, allowing re-
configuration into more thermodynamically favourable polymers.12 This allows facile syn-
thesis of copolymers through introduction of new monomers to a supramolecular polymer
by waiting for equilibration.96 A delicate balance of bond lifetime and Ka is also needed,
if the bonds are too strong then the polymers lose their supramolecular properties and act
43
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
more covalent, c.f. metal coordination polymers, if they lack sufficient strength then only
oligomers can be formed under reasonable concentrations.12,13,95
Figure 2.3 A schematic to show the potential growth mechanisms for supramolecular
polymerisation. (a) Isodesmic polymerisation, where each addition of monomer is inde-
pendent; (b) ring-chain mediated polymerisation, where cyclisation is favoured below a
critical concentration; (c) cooperative polymerisation, where monomer addition is pro-
moted after a critical nucleation step. Adapted from references.12,13
Supramolecular polymerisation mechanisms broadly fall into three categories: isode-
smic, ring-chain mediated, and cooperative pathways, as shown schematically in Fig. 2.3.12
Isodesmic polymerisation is the most similar to Carothers’ step growth polymerisation
(c.f. polyesters),97 characterised with high polydispersity (PDI = 1 + ρ) and a DP that
is highly dependent on the Ka between components.
12 This mechanism relies on each
bond formation to be the same and independent, i.e. there is no effect from neighbouring
groups (e.g. electronic or steric effects). DP can be defined by the following relationship:
DP ∼ (Ka[M ])λ, where [M ] is the concentration of initial monomer, and λ is a constant
of around 0.5.12 Therefore, at dilute concentrations, the only way to observe a high DP is
to have a high Ka > 10
6 M−1.12 Synthetic polymers based upon quadruple H -bonding 2-
ureido-4-pyrimidinones (UPy) have shown the ideal Carothers relationship (PDI = 1 +ρ)
is achievable, making supramolecular polymers ideal for fundamental studies.89,95,97,98
Key features resulting in an isodesmic pathway rely on monomers not needing to un-
dergo high energy conformational changes to form a polymer, as this would result in a
cooperative pathway with the limiting step being the initial formation of short oligomers.
Flexibility and length of spacers between binding motifs is another key factor that can
result in a so-called ring-chain type of polymerisation through the formation of intra-chain
cycles.12 With a heteroditopic monomer AB as an example, the concept of effective con-
centration takes precedent. If the local concentration of unreacted A units in an oligomer
is higher than those of other unreacted AB monomers, cyclisation will be preferential to
44
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
intermolecular associations.12,89,99 Therefore, a critical linking chain flexibility and length
will be present, along with a critical concentration for polymerisation.
Any suitably strong non-covalent interaction can be utilised in the formation of supra-
molecular polymers, with weaker interactions becoming viable by making them multident-
ate. Isodesmic polymerisation can occur by pi − pi stacking as shown with electronically-
coupled phenyl acetylene macrocycles, and with graphene derivatives.12,100,101 UPy and
the melamine-cyanurate pairs are a good example of possible isodesmic pathways in H -
bonded polymers.102,103 Host-guest interactions such as with di-calix[4]arenes, or various
architectures of cyclodextrins have exhibited isodesmic mechanisms, although ring-chain
is more common.12,94,104 Coulombic interactions can also be used in highly polar solvents
even though they lack directionality, such as those available on Zwitterions, through
entropy-driven polymerisations.105 In addition, metal coordination can be considered, al-
though more kinetically labile metal-ligand interactions need to be selected.12,106
With ring-chain polymerisation mechanisms, Monte Carlo simulations revealed the
rigidity and length of the spacer is the dominant factor in cyclisation, rather than the
rigidity of the reactive end groups.99 As supramolecular systems are mostly under ther-
modynamic control rather than kinetic, the degree of cyclisation can be controlled in-situ.
Well-studied examples include those with different alkyl spacer linkers for the UPy binding
motif,107 and those of crown ether host-guest complexes.108–110 In water with cyclodextrin
host-guest chemistry, it has been shown that if flexible linkers are used in a homoditopic
A2−B2 configuration this will preferentially result in cyclic oligomers as opposed to poly-
mers.111
Cooperative polymerisations have two distinct steps in their mechanism, effectively
being two isodesmic polymerisations with different Ka, and they see no equivalent in co-
valent polymerisations.12,95 Initially, polymerisation proceeds until the propagating chain
reaches a critical nucleation point at which further polymerisation proceeds with an in-
creased Ka. Further distinction can be made depending on if the initial nucleation has a
positive or negative ∆G.12,95 At high concentrations of monomers, this initial growth of
the nucleus can be switched from disfavoured to favoured. The main causes for cooperat-
ive polymerisation lie in electronic, structural and solvophobic effects.12 Electronic effects
occur where nucleation causes a change in the monomer electronic structure, for example
through inducing polarisation. Structural effects are a result of polymerisation only be-
coming possible when the monomer significantly alters its conformation; after a degree
of nucleation the energy barrier to conformational changes reduces. Solvophobic effects
occur when, after a degree of nucleation has occurred, solvophobic interactions can be
minimised through altering conformation, thus leading to cooperative polymerisation.12
Examples include H -bonding urea or amide-based assemblies where H -bonds are formed
on the top and bottom of the urea/amide bond. This is cooperative because initial dimer-
isation leads to a change in the electronic structure of the bond thus promoting further
monomer association, in addition to significant structural rearrangement for multidentate
systems as seen in helical polymerisations.12,112,113
Anti-cooperative supramolecular polymerisation can also occur, whereby nucleation of
45
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
an oligomer is highly favoured but further propagation is disfavoured. A good example of
this type of polymerisation is that of surfactant self-assembly, where micellar structures
are highly favoured but further assembly is not.12 Anti-cooperativity has also been shown
with pi − pi stacked polymers, where bulky side groups inhibit the formation of polymers
with a 10 times difference in Ka for dimerisation versus polymerisation.
114
Supramolecular polymers based upon H -bonds have seen a lot of development over
the years, with historical examples shown in Fig. 2.4. The first H -bonded main chain
supramolecular polymer was developed by Lehn et al on homoditopic monomers (A2−B2),
however the window for polymer formation lied in the liquid crystalline phase.115 X-ray
characterisation also suggested the formation of columnar superstructures such as a triple
helix, analogous to DNA. In addition, it did show properties unique to polymers, such as
the ability to draw fibres. H -bonding needs to be multidentate to be strong enough to
form anisotropic polymers as each bond is on the order of 5 - 30 kJ mol−1.12,14,102 With
the quadruple H -bonding UPy Ka = 10
7 M−1 in chloroform and a long lifetime of 0.1
- 1 seconds, though with further optimisation this can reach up to 1012 M−1.14,116 This
allowed construction of polymers up to 500 kDa in size, with easy synthesis of the relevant
monomers.95,102 In-situ copolymerisation could also be achieved with UPy as predicted,
even resulting in purely alternating copolymers, provided Ka between comonomers was
higher than self-association.96,117 Branching networks and hyperbranched polymers have
also been prepared by introducing branching points, but as the network can reassemble
compared to covalent systems, it will form denser, more thermodynamically stable net-
works given time.118
Figure 2.4 The first H -bonded supramolecular polymers based upon triple, quadruple,
or sextuple interactions. Adapted from references.14
Favourable pi−pi stacking has been investigated at great length, typically forming poly-
mers with a cooperative pathway. On its own, it is a relatively non-specific non-covalent
interaction, where control needs to be introduced by solvophobicity and steric hindrance.
46
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Monomers are selected with a rigid planar core, such as an aromatic ring, and various side
chain substitutions have been used for liquid crystalline applications.93 Linear stacking
interactions are strongest as the presence of suitable side chains results in efficient phase
separation. It is possible to use alternating donor-acceptor charge-transfer to enhance
the stacking, with the introduction of chirality achievable through side-group directed
helical twisting.12,93,119 Coupling pi − pi stacking with H -bonding further increases the
directionality, exemplified with 1,3,5-benzene triamides (BTAs) being ideal for columnar
polymerisations.120 Helical columnar stacks have been formed with BTAs, which then form
fibrous higher-ordered assemblies in organic solvents (Ka = 5×108 M−1 in hexane). Ana-
logous columnar structures have been constructed with guanine and pterine derivatives,
whereby assembly begins with the formation of barrel-like structures that subsequently
stack to form short oligomers.121 Many variations thereof have since been reported.12,93
Soluble metal coordination polymers require a kinetically labile metal-ligand interac-
tion to form truly supramolecular polymers that are completely reversible. This is nor-
mally shown by a concentration dependence on the polymerisation and other condition-
dependent properties as previously discussed.14,93 To achieve a reasonable binding con-
stant multidentate ligands must be used, analogous to multiple H -bonding arrays.14 Ex-
amples include those based upon Cu(I) or Ag(I) with multidentate pyridinium ligands.106
Cobalt-based porphyrins with two pendant pyridines can also polymerise well in organic
solvent, giving a concentration-dependent DP up to 100 via a ring-chain polymerisation
pathway.122 Supramolecular coordination polymers give access to otherwise unobtainable
electrochromic and redox responses, as has been shown by the redox dependent polymer-
isation of Cu(I/II) systems.123
Applications of supramolecular polymers to aqueous environments merits special con-
sideration, due to the potential applications in sustainable, biocompatible, and hydro-
gelating materials.94 As water is a highly polar H -bonding solvent, polymers based upon
H -bonding are much weaker due to competition, and hydrophobic effects are very strong.94
This has lead to the development of a host of different monomers that provide a hydro-
phobic pocket in which efficient H -bonding can occur, such as BTA units functionalised
with alkyl chains.94,124,125 Aromatic stacking can also be quite strong in water, with many
developments based upon large aromatic rings such as perylene diimides to drive polymer-
isations.94,126,127 Host-guest chemistry can provide an easy method for strong and highly
selective supramolecular interactions in water, and is discussed in the next section.94
Whilst supramolecular polymer materials cannot compete with covalent materials in
their mechanical properties under harsh conditions, they give access to several unique ap-
plicable properties. As they display a distinct temperature and concentration dependence
on their DP , they can be easily processed for manufacturing and have good mechanical
properties at ambient conditions allowing construction of elastomers and semi-crystalline
polymers.102,128 Stacked conjugated pi systems have resulted in several conductive ma-
terials with applications in supramolecular electronics, combining the favourable con-
ductive properties exhibited by purely molecular crystals with the processability of poly-
mers.12,95,129 In addition, optoelectronic applications have been explored due to the high
47
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
tunability of copolymerisation allowing precise design of energy levels, such as in the gen-
eration of white light LEDs with multiple UPy monomers.95,130 Furthermore, there is
increasing interest in supramolecular polymers that are biologically active, such as self-
assembling peptide amphiphiles as 3D scaffolds for cell cultures or for in vivo regenerative
medicine.95,131
Characterisation of supramolecular polymers is not an easy task. Estimation of the
extent of polymerisation can be done if the mechanism is isodesmic, to DP = (Ka[M ])
0.5,
however this does not hold for other mechanisms.12,14 Established methods for character-
isation of covalent polymers do not transfer well either, as most rely on polymers being
unaffected by changes in e.g. concentration or temperature.14
Size-exclusion chromatography is often used to determine molecular weight distribu-
tions by chromatographic separation coupled with various detectors. However, it relies
on very high degrees of dilution and strong shear forces from being run at high pressure.
With supramolecular polymers, this results in degradation.14 Viscometry cannot be used
directly as the Mark-Houwink equation requires calibrated empirical constants for each
polymer in question,89 however it has been used extensively for deducing critical polymer-
isation concentrations, for example.14,24,25 NMR end-group analysis can be used, but this
only applies for small polymers with slow kinetics and with significantly separated peaks
for the end group.
D =
kBT
3piηDh
(2.4)
Diffusion-ordered NMR spectroscopy (DOSY, also termed pulsed field gradient NMR
spectroscopy) is a technique that correlates the diffusion coefficient with each peak across
an NMR spectrum. This is achieved by applying a gradient field to spatially label mo-
lecules, and following them over time. The diffusion coefficient is inversely proportional to
molecular size via the Stokes-Einstein equation as shown in Eq. 2.4 where Dh is the hy-
drodynamic diameter, kB is the Boltzmann constant, T is temperature, η is the viscosity,
and D is the diffusion coefficient.14,132–134 However, η will change significantly at different
stages of polymerisation and must be accounted for. The polymer has also been assumed
to be completely spherical. This means DOSY can be a useful qualitative technique to
show comparative differences between monomer and polymer, but direct quantification
will be inaccurate without significant adjustments to calculations.14
Dynamic light scattering also relies on the Stokes-Einstein equation, instead calculat-
ing D from temporal changes in light scattering. It is normally quicker and easier than
DOSY measurements, but is highly prone to contamination from larger species, as larger
particles scatter light much more than smaller ones. Static light scattering can be used for
direct characterisation of molecular weight with no assumptions, but accurate results via
techniques such as a Zimm plot rely on varying concentration.14,132 Mass spectrometry
techniques are often used, however these are exclusively in the gas phase and is therefore
not an accurate representation of solution-state polymers but more indicative of potential
structures.14
48
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
2.1.3 Host-Guest Supramolecular Polymers
Host-guest interactions are highly interesting for the construction of supramolecular poly-
mers due to their typically high binding constants, and very high specificity for certain
guest molecules. They were first investigated in the late 1990s, where the association of
crown ethers and secondary ammonium ions (Ka = 2.7× 104 M−1 in CDCl3) and the as-
sociation of calix[4]arenes and solvent (Ka = 10
6−7 M−1) was carried out.12,24,94,104,109,135
Some of the macrocyclic host molecules investigated include cyclodextrins (CDs), cucur-
bit[n]urils (CB[n]s), pillararenes, calixarenes, and crown ethers.14,94 Guest molecules can
be attached covalently to the hosts to make polymers from single monomers, or multitopic
guests can be combined with a separate host molecule to form supramolecular polymers.93
Crown ethers have been used for supramolecular polymerisation with ring-chain poly-
merisation pathways. Cyclisation could be inhibited via a heteroditopic AB monomer
where the guest and host were connected with a single CH2 linker.
14,109 Homoditopic
arrangements have also been investigated, showing distinct trends in the length and flex-
ibility of the linker required to inhibit cyclisation. Interestingly, if linkers are mismatched
in length, then dimerisation can be effectively inhibited.110,135 Due to the ease of pendant
functionalisation of crown ethers, hyperbranching topologies have also been readily pre-
pared in an AB2 architecture but require high concentrations due to their low Ka.
24,136
Calixarenes have been investigated in a similar fashion based upon the complexation of
ammonium ions, utilising either heteroditopic or homoditopic monomers with control over
cyclisation.137,138 Polymerisation with sulfonate-functionalised calixarenes has also been
achieved in water with pyridinium guests, with the formation of 2D networks achieved
with a tetratopic rigid porphyrin guest as shown in Fig. 2.5.15 Pillararenes were utilised
in a similar fashion to construct polymers via heteroditopic and homoditopic architec-
tures with gel formation also possible from intertwining of linear polymers into fibrous
networks.139–141 Branching topologies have been investigated, with depolymerisation in
response to a competing guest molecule demonstrated.25
Figure 2.5 Schematic representation of a (a) 2D network or a (b) linear polymer obtain-
able in water through control over guest topology. Adapted from references.15
If a heteroditopic AB monomer is prepared from CDs, this typically leads to in-
tramolecular complexation or dimers.14,142 Harada has pioneered the design and use of
49
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
CDs for supramolecular polymer synthesis, with first efforts resulting in trimers from
cinnamoyl-functionalised CDs.142 After optimising the location of covalent attachment of
the guest on the CD, formation of intramolecular complexes could be impeded and poly-
mers in solution could be obtained.143 If homoditopic monomers are used (A2−B2), where
two CDs are covalently linked, polymers can be readily prepared with control over the
DP with the length and hydrophobicity of the linker.111,144 More recently, the reversible
cross-linking photochemistry of coumarin guest molecules has been used for the revers-
ible switching of supramolecular polymers into covalent polymers.145 Stimuli-responsive
branching polymers could be synthesised with the incorporation of the photo-responsive
azobenzene guest as shown in Fig. 2.6.16 In addition, hyperbranching topologies have been
investigated in detail based on an A2−B3 architecture, even forming dendritic structures
observable by microscopy.146,147
Figure 2.6 Schematic representation of the photo-controlled supramolecular polymerisa-
tion of branching CD-based polymers. Adapted from references.16
CB[n]s have only recently seen applications in the formation of supramolecular poly-
mers. They are highly valuable to investigate due to their extremely high binding constants
(Ka ∼ 107−15 M−1), and solubility in water.1 However, direct functionalisation of CB[n]s
with guest molecules is considerably harder than for the host molecules discussed so far,
as the only routes available require intensive synthesis and purification with low yields.
Recent efforts have been made to optimise CB[n] functionalisation, but have not been
applied to supramolecular polymerisations yet.1,148 Therefore, the available supramolecu-
lar polymerisation routes lie with combining multitopic guest molecules with CB[8], as
this host can encapsulate two guest molecules. However, the solubility of neat CB[8] is
limited (< 1 mM), thus limiting the scope due to the strong concentration dependence
on polymerisation.14 This solubility can be increased substantially on complexation of a
suitable guest molecule, but remains below ca. 20 mM.1
Cyclisation is quite prevalent in CB[8]-based polymers, for example simpleAB monomers
in combination with CB[8] form intramolecular (1:1) or dimeric (2:2) complexes rather
than polymeric structures.149 Thus, CB[8] mediated polymerisations are primarily by
a ring-chain mechanism, with optimisation of the monomers required to promote poly-
50
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
merisation.14 A simple A2 monomer based on naphthalene guests with a suitably rigid
aliphatic bicyclic linker can result in polymers at dilute concentrations due to the high
Ka ∼ 1011−12 of pi− pi stacking within CB[8] with no cyclisation observed.133 In contrast,
if significantly longer and flexible linkers are used, such as an octaethylene glycol in com-
bination with a tripeptide (Ka = 3.7 × 106 M−1), cyclisation can be efficiently avoided
through entropic considerations giving rise to supramolecular polymers.150 By utilising
an ABBA monomer, the rigidity of CB[8] can be exploited to promote polymerisation.
Zhang et al designed a monomer with viologen (A) and anthracene (B) guests as shown in
Fig. 2.7, which successfully polymerised with no cyclisation observed.17 Analogous photo-
responsive polymers replacing the anthracene for azobenzene have been developed, where
the photo-isomerisation of azobenzene can reversibly depolymerise the structure.65,151
Figure 2.7 Supramolecular polymerisation with CB[8] based upon an ABBA monomer,
thus inhibiting cyclisation and intramolecular complexation. Adapted from references.14,17
Branching topologies have only recently been investigated for CB[8]-based supra-
molecular polymers. A single monomer methodology has been investigated by Zhang,
where an A3 architecture is used.
152 The A unit was designed to form a homoternary
(2:1) complex with CB[8], and has been achieved with either naphthalene or azastilbene
guests.134,152 The latter case allowed photo-crosslinking within the CB[8] cavity by a [2+2]
cycloaddition to result in a covalent branched polymer.152 However, the DP and extent
of intra-chain cyclisation were not investigated, with primary characterisation carried out
by DOSY NMR and UV-vis studies. Extended 2D and 3D supramolecular organic frame-
works have been prepared in solution utilising tritopic and tetratopic guests in combination
with CB[8] via homoternary complexation.153–155 These frameworks showed excellent ab-
sorption of various small molecules and showed application as soluble, porous materials.
Following on from the discussed theory behind branching polymerisations established
by Flory,91 monomers for CB[8]-based host-guest polymers based upon an A3 or an A2−B3
architecture should result in a cross-linked network and gelation at a critical conversion
point. Thus far, this has not been achieved. Herein, the fabrication of a branching
supramolecular polymer based on an A2 − B3 architecture has been achieved in water.
51
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Solubility limitations were reached before gelation was observed on the bulk scale. Sub-
sequent confinement into emulsions allowed control over the self-assembly of the polymers
into a macroscopic gel, reflecting the theory outlined by Flory applied to supramolecular
systems for the first time.91
2.2 Results and Discussion
2.2.1 Supramolecular Polymerisation
Figure 2.8 (a) Chemical structures and schematic representations of CB[8], and the
ditopic (A2) and tritopic (B3) monomers used in this work. (b) Overview of the proposed
assembly process; A2 is first complexed with CB[8] (2 eq.), and then mixed with B3
(0.67 eq.) to form the polymer. This polymer can be disassembled in response to either (i)
reversible photo-isomerisation of the azobenzene with UVA light (left), or (ii) introduction
of a competitive guest 1-adamantylamine hydrochloride (ADA) (right).
In order to achieve branching supramolecular polymerisation that could lead to a
cross-linked network, a ditopic monomer (A2) and tritopic monomer (B3) architecture
was used, where A and B are guests for CB[8] that can bind in a heteroternary fashion.
A heteroternary binding mechanism is highly desired, as this will prevent self-association
of A or B functionalities. As discussed in the introduction, if A and B stoichiometries
are kept at 1:1, then gelation should be achieved at ρ = 0.71. The water-soluble guests
synthesised are shown in Fig. 2.8a, based upon a ditopic viologen derivative (A2) and a
52
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
tritopic azobenzene derivative (B3).
Viologen is electron-deficient and binds efficiently as the first guest in CB[8], followed
by the azobenzene moiety, yielding a heteroternary complex.64,65 The binding constant
of the initial viologen-CB[8] complex has been reported to be Ka = 5.0 × 105 M−1,64,65
with sequential binding to the azobenzene imidazolium salt derivative reported to be
Ka = 3.5× 104 M−1, yielding an overall Ka of 1.8× 1010 M−2.64,65 This ternary complex
is only formed when the azobenzene is present in its dominant E isomer. The proposed
assembly process is outlined in Fig. 2.8b.
The spacers between guest moieties on monomers A2 and B3 were selected for optimal
supramolecular polymerisation and water solubility. In particular, they are not com-
pletely rigid but have a small degree of flexibility, as this should suppress the formation of
intra-chain cross-links, and the effect of long chain entanglement observed with oligomeric
linkers.11,152,156–158 The distance between reactive A and B groups after disassociation of
A2 with B3 should be far apart enough to prevent cyclisation.
Figure 2.9 Stacked 1H NMR spectra in D2O of A2 (a) upon addition of 1 equiv. (b) and
2 equiv. (c) of CB[8]. The complexation proceeds as previously reported.18
The monomers were first characterised by 1H NMR spectroscopy, before combining
with CB[8] to induce polymerisation. The binding of A2 to 1 and 2 equiv. of CB[8] in
D2O was shown to be consistent with literature, with 2 CB[8] molecules bound to the
terminal viologen derivatives (Fig. 2.9) via an intermediate state where CB[8] binds to
the central ethylene glycol unit.18 The 1H NMR spectrum of B3 in D2O and d6-DMSO
is shown in Fig. 2.10. In d6-DMSO the molecule is well solvated, with extra peaks
53
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 2.10 Stacked 1H NMR spectra in D2O (a) and d6-DMSO (b) of B3. The aromatic
region in D2O is characteristic of when pi − pi stacking interactions are occurring, this is
not the case when in d6-DMSO as this solvates the molecule well. R represents further
azobenzene imidazolium groups.
relating to the Z isomer present. In more polar D2O, however, self-association of the
azobenzene imidazolium groups was observed, which is clear from the overlap of peaks
in the aromatic region. A similar structure without the imidazolium salt present has
been reported for the assembly of dendritic rotaxanes in D2O with CB[8] exhibiting no
self-association in the NMR spectra, suggesting the azobenzene imidazolium salts are the
cause of this effect, rather than the formation of BTA columnar stacks.28,124,125 In order
to dissolve B3 effectively in water, it was dissolved in a minimal amount of d6-DMSO
and thereafter diluted with D2O (resultant solutions 0.7 v/v % d6-DMSO in D2O). The
assembly process was designed such that the A2-CB[8] (2:1) complex was pre-assembled
in water, before introduction of B3.
Upon combining a solution of A2-CB[8] (2:1) with a solution of B3 to give a final
composition of A2:B3:CB[8] = 1.5:1:3 molar ratio ([A2] = 0.55 mM), the formation of a
polymer was suggested in the 1H NMR spectrum, as shown in Fig. 2.11c. The significant
broadening and loss of intensity of all the peaks suggested some degree of polymerisation
had taken place, along with typical peak shifts for host-guest binding.64 Of particular
interest is the broadening observed for the CB[8] peaks, which has been observed in other
CB[8]-based polymerisations.14,133,134 Gelation was not observed at this concentration,
implying that either the conversion of the polymer is not enough (ρ = 0.71 for gelation), or
that intramolecular cyclisation was dominant. If the concentration is increased further (up
to [A2] = 1.10 mM), the solution became turbid due to a loss of water solubility, as shown in
Fig. 2.12. This is likely a result of intra-chain cyclisation resulting in particulate structures
forming, along with the well-established limited solubility of CB[8]. This means that the
polymerisation is hindered at these concentrations below the gelation point, concurrent
with the reported literature.14,133,134
54
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 2.11 1H NMR spectra in D2O of (a) B3, (b) A2-CB[8] (1:2), (c) A2-B3-CB[8]
(1.5:1:3). Concentrations in all cases were A2 = 0.55 mM, B3 = 0.37 mM, CB[8] = 1.11
mM.
2.2.2 Stimuli-Responsiveness
The response of these assemblies in solution was probed in response to different external
stimuli. Firstly, the reversible photo-responsiveness was investigated for the azobenzene-
based B3 monomer, and the combined polymeric systems. The irreversible chemical re-
sponse of the polymer to a competing guest was also considered.
Experiments were carried out to characterise the photo-physical properties of B3 as
shown in Fig. 2.13. From integration of the 1H NMR spectrum in d6-DMSO, it was
calculated that the isomeric E -Z ratio of B3 under equilibrium at room temperature is
80 % (E )-B3. Upon exposure to UVA light (λmax = 360 nm, 1 h) a photostationary state
of 17 % (E )-B3 could be reached. Continued exposure to UVA light did not drive the
isomerisation any further. Upon heating this solution to 80 ◦C, > 99 % conversion to the
(E )-B3 isomer was achieved, which then relaxed back to its equilibrium state over 12 h
at room temperature. As the NMR spectrum was complex in water, analogous reversible
photoisomerisation in water was confirmed by UV-vis spectroscopy. This is shown in Fig.
2.14.
With this information in hand, the combined supramolecular polymer was illuminated
with UVA light in order to photoisomerise the (E )-B3 to (Z )-B3 and consequently dis-
rupt the heteroternary complex, as tracked by 1H NMR spectroscopy (Fig. 2.15). Upon
exposure to UVA, the polymer solution was observed to lose its characteristic orange col-
55
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 2.12 A photo to illustrate the loss in water solubility occurring upon assembly of
the supramolecular polymer at higher concentrations. Before combination at equivalent
concentrations, the separate components remain completely dissolved. Concentrations are
as follows: A2 1.10 mM, B3 0.73 mM, CB[8] 2.2 mM. Concentrations used in
1H NMR
measurements were half of this, at the limit before the solution became turbid.
our from the azobenzene, attributed to precipitation and sedimentation of the resultant
(Z )-B3-CB[8] complex. This is reflected in the spectrum (Fig. 2.15b), where a significant
drop in intensity is observed, and peaks corresponding to B3 have mostly disappeared.
Furthermore, the peaks for A2 recover their chemical shifts for that of the uncomplexed
molecule, evidencing no interaction with CB[8]. The (Z )-B3-CB[8] complex is formed due
to the proximity of the cationic imidazolium salts present on the B3 monomer favouring
binding to the carbonyl groups of the CB[8], as observed previously.64 Notably, precip-
itation does not occur in the absence of CB[8] after UVA exposure. Disassembly of the
polymer is reversible, with isomerisation back to (E )-B3 achievable upon heating. Heat-
ing the (Z )-B3-CB[8] dispersion to 80
◦C resulted in the dissolution of the precipitate
and reformation of the polymer (Fig. 2.15c). UV-vis spectroscopy was also performed to
follow this isomerisation, shown in Fig. 2.16, showing similar trends as to that of the B3
monomer alone in solution (Fig. 2.14).
The ability of the polymer to respond to a chemical stimulus was also demonstrated,
as shown schematically in Fig. 2.8b. Titrating an increasing amount of the competitive
guest 1-adamantylamine hydrochloride (ADA) resulted in the chemical shifts of the peaks
associated with the A2 and B3 monomers in their
1H NMR spectrum returning to their
uncomplexed state as shown in Fig. 2.17. In concert, the CB[8] peaks shift and sharpen
as they complex with ADA, with the appearance of the complexed ADA peaks around
56
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 2.13 Stacked 1H NMR spectra in d6-DMSO of B3 in response to UVA light and
heat treatment. (a) State immediately after dissolution of solid B3; (b) after exposure to
UVA light for 1 h; (c) after heating at 80 ◦C for 6 h; (d) after leaving to reach equilibrium
state in ambient conditions for 12 h. The proportion of the E and Z isomers were cal-
culated from comparing integration of the aromatic peaks corresponding to each isomer,
and that of the CH2 at ca. 5.5 ppm. Blue circles represent peaks that change for the E
isomer, and red squares for the Z isomer.
Figure 2.14 UV-vis measurements taken of the B3 monomer in water (0.7 v/v% DMSO).
As is characteristic for azobenzene photoisomerisations, the n to pi∗ transition at 324 nm
is altered dramatically upon UVA-triggered isomerisation, while the pi to pi∗ transition
absorbance at 426 nm increases. After heating, the photoisomerisation can be completely
reversed.19
1.5 ppm. The solution is visibly free of precipitates. These results confirm the complete
disassembly of the polymer upon addition of ADA by displacement of the guests within
CB[8].
57
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 2.15 1H NMR spectra in D2O of (a) A2-B3-CB[8] branched supramolecular poly-
mer in its equilibrium state; (b) after 1 h of UVA exposure; (c) after subsequent heating
at 80 ◦C for 6 h. Note, upon heating the azobenzene is >99 % E isomer, higher than that
originally present in (a).
Figure 2.16 UV-vis measurements taken of the A2-B3-CB[8] polymer. A similar trend
in photoisomerisation is observed as for the B3 monomer alone. At room temperature in
the dark, a small amount of recovery of the equilibrium is observed after 45 mins. Here
a blue LED light exposure for 10 mins is sufficient to drive the recovery of the E isomer
due to the low concentration used ([A2] = 55.0 µM).
58
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 2.17 1H NMR spectra in D2O of A2-B3-CB[8] polymer (a) upon titration with
0.5 (b), and 1 (c) equiv. of ADA relative to CB[8].
59
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
2.2.3 Gelation at Liquid-Liquid Interfaces
Whilst the formation of a branched supramolecular polymer in dilute solutions has been
shown, an increase in concentration results in precipitation as opposed to gelation. This is
attributed to a decrease in solubility, which prevents formation of extended architectures.
In order to circumvent precipitation and drive formation of an extended 3D cross-linked
network, a liquid-liquid interface was used as a template to drive self-assembly from dilute
aqueous solution.
Microfluidic emulsions make ideal candidates for studying self-assembly processes in a
controlled environment. There have been many advances in recent years in their applic-
ation across the sciences, from studying single cells, to controlling chemical reactions in
liquid microreactors.79,159
Microfluidic Emulsions
In the following experiments, PDMS microfluidic chips were used that have a flow-focussing
junction, where the shearing force of two immiscible liquid flows, and stabilisation of
the droplet interface with surfactants, allows the formation of monodisperse droplets
(Fig. 2.19a).79 Interfacial gelation was explored in the resultant emulsions, using aqueous
droplets as templates in a continuous phase of inert perfluorocarbon oil (Fluorinert FC-
40, 3M) containing a neutral perfluorinated triblock polymer surfactant (Sphere Fluidics)
with charged perfluoropolyether dopants. This provided a unique environment that al-
lowed the polymer to undergo a gelation phase transition from dilute solution to form an
elastic, self-healing interfacial gel.
Figure 2.18 A schematic outline of the assembly process from dilute solution to an interfa-
cial gel. At dilute solution there are no precipitates, but there is also no extended polymer
network. After emulsification and subsequent electrostatic attraction and accumulation to
the liquid-liquid interface, the concentration and density of the polymer rapidly increases
over ca. 2 s, leading to inter-chain cross-linking and gelation (Fig. 2.19b).20,21
This methodology has been previously employed in the assembly of supramolecular
polymer microcapsules from CB[8] cross-linked linear polymer networks (< 5 mol% guest
for CB[8] present on the polymer backbone).20,21,71 It has recently been shown that after
evaporative concentration of the aqueous droplet in ambient conditions, the interfacial
polymer-CB[8] network undergoes a phase transition from a viscous fluid to a physically
cross-linked gel, reflecting gelation at a critical concentration on the bulk scale.20 This was
60
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
easily studied through optical microscopy, as after a gelation at the interface had occurred
the droplet loses its spherical shape and ‘buckles’, forming wrinkles along the droplet
interface. This is because a solid gel phase is now present at the interface, and thus
there is no longer high surface energy requiring a spherical shape to reduce the interaction
surface area.20
Figure 2.19 Transmission optical micrographs of: (a) an example flow-focussing junction
resulting in monodisperse microdroplets, typically with flow rates of 150 and 100 µL h−1 for
the oil and water flows respectively; (b) top panel shows aqueous microdroplets containing
A2-B3-CB[8] and oil phase containing 2 wt.% neutral triblock and 1 wt.% RFCOOH, and
bottom panel is at higher magnification; (c) the same as in (b) but with 1 wt.% RFNH2.
In this experiment, the concentration dependence of supramolecular polymerisation
(i.e. the adaptability of supramolecular polymers) was exploited, with a schematic over-
view of the assembly process shown in Fig. 2.18. By loading aqueous droplets with a
dilute solution of A2-B3-CB[8] ([A2] = 0.55 mM), we can then control self-assembly at
the microdroplet interface by electrostatic accumulation,21 as A2 and B3 monomers bear
multiple cationic charges. This results in a rapid increase in concentration at the inter-
face directly from solution, as diffusion on the micrometre length scales is on the order
of a few seconds. This electrostatic attraction was achieved through the introduction of
61
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
a carboxylic acid-terminated perfluoropolyether (RFCOOH, Krytox 157-FSL), and repul-
sion with a synthesised amine-terminated derivative (RFNH2). The full structures and
synthetic procedures are outlined in section 2.4.1.
Here, we have achieved a similar result from supramolecular polymers constructed
entirely from small molecules. As a result of the rapid (ca. 2 s)21 assembly of the polymer
at the interface due to electrostatic accumulation, the polymer is spatially confined at a
high density, stabilised by the droplet surface charge. This results in inter-chain cross-links
readily forming, giving an elastic and self-healing cross-linked gel at the interface without
the need for evaporative concentration.
Figure 2.20 Transmission optical micrographs of emulsions on glass slides containing:
(a) aqueous droplets containing A2-B3 and no CB[8]; (b) aqueous droplets containing A2-
B3-CB[7]. The continuous phase was 2 wt.% neutral triblock and 1 wt.% RFCOOH in
FC-40.
This is illustrated in Fig. 2.19b, where droplets have undergone a phase transition to
form ‘buckled’ droplets in the presence of RFCOOH, indicating the formation of an elastic
gel.20 If a positively-charged dopant is employed (RFNH2), assembly at the interface is
disfavoured and the droplets evaporate resulting in a microparticle structure upon com-
plete loss of water (Fig. 2.19c). During this process, precipitation occurs as observed on
the bulk scale at higher concentrations. Initially these precipitates are free-flowing inside
the droplet, and then halt motion as all of the water evaporates. Fig. 2.19b highlights
how directed assembly to the interface prevents precipitation from occurring.
If CB[8] is absent from the droplet, this buckling transition does not occur and instead
the droplets simply evaporate until the high charge density destabilises the system, and
the droplets burst, shown in Fig. 2.20a. This is likely a result of the highly charged species
preferentially wetting the hydrophilic glass surface. Similarly, if CB[7] is used, which will
only accommodate one viologen guest due to its smaller cavity volume, then the droplets
again destabilise upon evaporation. In this case, the droplets decrease in size until nearly
all water has evaporated before destabilising; this is likely a result of some of the charged
62
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 2.21 Transmission optical micrographs of emulsions deposited onto a glass slide
containing: (a) aqueous droplets containing 500 kDa dextran and (b) aqueous droplets
containing A2-B3-CB[8] and 500 kDa dextran. Both continuous phases comprised FC-
40 containing 2 wt.% neutral triblock and 1 wt.% RFCOOH. Upon addition of PFOH
complete coalescence is observed over a few seconds in case (a), and in case (b) the network
stabilises the droplet interface to coalescence.
species being bound to CB[7].
The interfacial gel was then applied in the stabilisation of droplets after the removal
of their stabilising surfactants. Perfluorooctanol (PFOH) was used as a de-emulsifier.
PFOH is a very poor fluorosurfactant in comparison to the neutral triblock in use, and
upon addition of excess PFOH droplets rapidly coalesce within seconds (Fig. 2.21a).
However, when PFOH is added to droplets that have the interfacial gel present, the gel
inhibits coalescence, stabilising the droplet in the absence of effective fluorosurfactants
(Fig. 2.21b). The network is not acting as a surfactant as there is no distinct amphiphilic
structure present in the 3D cross-linked structure, but is acting as a soft matter barrier,
allowing the droplet to become non-uniform in shape (i.e. non-spherical). Note that in
order to prevent rapid evaporation occurring after removal of surfactants, 500 kDa dextran
polymer (1 mg mL−1) was added to the aqueous phase to increase the viscosity.
Pendant Droplets
In order to study the interfacial gel in an environment where evaporative flux is not a dis-
ruptive factor, measurements were carried out in a pendant (hanging) droplet geometry.
Pendant droplet measurements have been shown to be ideal for observing interfacial phe-
nomena and calculating material properties.20,22,160
63
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
As the perfluorocarbon oil is denser than water, the liquid-liquid interface was inverted
relative to the microfluidic experiments, with a hanging perfluorocarbon droplet in a con-
tinuous aqueous phase. The concentration of the polymer was lower than that used in the
microfluidic environment (diluted to [A2] = 0.55 µM). This configuration also required a
lower concentration of RFCOOH to drive interfacial assembly (0.001 wt.% versus 1 wt.%)
and was carried out in the absence of the neutral triblock surfactant, as shearing flows
and droplet stabilisation was no longer required.
The interfacial tension of the droplet interface (γ) can be directly measured by shape
analysis and fitting to the Young-Laplace equations.22,160 Typically this is routine and
carried out via in-built instrumental software. The γ measured for a droplet of RFCOOH
in FC-40, suspended in water was 46 ± 2 mN m−1.
Figure 2.22 Plots showing how γ changes with equilibration time for different exper-
iments. RFCOOH was in FC-40 at 0.001 wt.%, RFNH2 was in FC-40 at 0.01 wt.%,
A2-B3-CB[8] was diluted 20x from the microfluidic stock solution to [A2] = 0.55 µM.
ADA was added in 10x stoichiometric excess to CB[8] to facilitate rapid displacement.
When the aqueous phase contains dilute A2-B3-CB[8], γ decreases over 45 mins re-
flecting electrostatic accumulation (and therefore reduction of the interfacial free energy)
of the polymer to the interface limited by the rate of diffusion, reaching an equilibrium
value of 34 ± 1 mN m−1, as shown in Fig. 2.22. Upon reduction of the droplet volume,
buckling can be clearly observed at the interface in Fig. 2.23a due to compressive elastic
stresses. This indicates that the equilibrium pendant droplet had a solid film at its in-
terface - as in the microfluidic droplet environment.22 Once the gel has been formed,
the droplet can be expanded beyond its initial volume and then contracted to reach the
same cross-linked state with no equilibration time (Fig. 2.23b). This remarkable result
highlights the elasticity of the cross-linked film. In addition, if the buckled droplet (i.e.
cross-linked film under compressive stress) is observed for a minute at a constant volume,
the membrane will smooth out (i.e. relax) to recover a more spheroidal shape and thus
release the compressive stresses. This is a direct visualisation of the release of stress in
supramolecular polymer networks via bond breaking and forming rather than reptation
as in covalent networks.
64
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 2.23 Transmission optical micrographs of the pendant droplet during pendant
droplet measurements of continuous aqueous phase containing A2-B3-CB[8], and oil
droplet of 0.001 wt.% RFCOOH in FC-40. From left to right, (a) an equilibrated 5 µL
droplet pumped in and exhibiting buckling; (b) the same droplet after pumping out to 8.5
µL to stretch the interfacial gel followed by immediate pumping in with no equilibration
time.
Figure 2.24 Transmission micrographs of the pendant droplet during pendant droplet
measurements at equilibrium (top) and then after subsequent pumping in (bottom). (a)
Aqueous A2-B3 and droplet of RFCOOH in FC-40; (b) aqueous A2-B3-CB[8] and droplet
of RFNH2 in FC-40; (c) aqueous A2-B3-CB[8] and 10x excess of ADA and droplet of
RFCOOH in FC-40; (d) aqueous A2-B3-CB[8] after 1 h UVA irradiation and droplet of
RFCOOH in FC-40.
Control measurements containing only A2-B3 exhibited a lower γ than that of pure
water (25 ± 1 mN m−1, Fig. 2.22), confirming that the electrostatic-based accumulation
of monomers at the interface occurs without CB[8], but as there is no CB[8] present a
cross-linked polymer network cannot form and no buckling transition is observed (Fig.
2.24a). Similarly, a control with the opposing charge of fluorosurfactant (RFNH2) did not
show evidence of buckling upon reduction of the droplet volume (Fig. 2.24b) as there
is no preference for accumulation of the polymer to the droplet interface, mimicking the
microfluidic droplet experiments. Furthermore, as the cationic charge inhibits the initial
assembly at the interface, the equilibrium γ was higher at 42 ± 1 mN m−1, similar to the
‘water-only’ value vide supra.
65
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Upon addition of ADA as a chemical stimulus, which was previously shown to break
apart the network in NMR experiments, buckling was not observed after a reduction in
volume (Fig. 2.24c), illustrating how the network retains its chemical responsiveness as
a gel at the interface. This occurs as the ADA outcompetes the guest molecules present,
breaking apart the network into its constituent small molecules. the IFT observed was
lower however, as the ADA is present in excess as its HCl salt and so will accumulate at
the interface and stabilise the interfacial surface energy.
Figure 2.25 Transmission micrographs of the pendant droplet during pendant droplet
measurements after 45 min equilibration and subsequent pumping in at different concen-
trations of aqueous A2-B3-CB[8]. Here buckling can be observed at [A2] = 0.55 µM (a),
0.055 µM (b), and 0.0055 µM (c). It can be clearly observed that there is a lower density of
wrinkling in (b), implying a thinner crosslinked network had been formed, and no network
is formed at (c).
In contrast, photo-triggered disassembly did not prevent interfacial gelation. The dilute
A2-B3-CB[8] solution was exposed to UVA light as for the NMR experiments, within a
cuvette in a UV photoreactor. The cuvette was then transferred to the pendant droplet
instrumentation where an oil droplet was suspended into it in the dark. Pumping in of the
oil was carried out after 45 mins, as for the other experiments, whereupon buckling was
observed (Fig. 2.24d). This was a result of incomplete conversion of the B3 azobenzene
moieties to their Z isomers, even with CB[8] present. It was theorised that the low
concentration of remaining B3 in the E isomer was still sufficient to accumulate at the
pendant droplet interface and form a cross-linked film due to the massively increased
concentration factor at the interface when compared to bulk. Evidence for this is provided
in Fig. 2.25, where buckling of A2-B3-CB[8] was still observed after a further order of
magnitude of dilution.
The shape fitting analysis used to calculate γ is only valid for liquid systems, i.e. it
cannot be used to quantify buckled droplets where the presence of a solid film provides
elastic stresses. Recently, the group of Kierfeld has released an open access C/C++ soft-
ware to allow shape fitting of buckled droplets, OpenCapsule, in order to extract various
material properties (c.f. interfacial rheology).20,22,160 This is achieved through relating
the shape and wrinkling characteristics observed to shape equations derived from non-
linear membrane-shell theory via Hookean elasticity, allowing calculation of many useful
properties such as the elastic modulus, Poisson ratio, stress distribution and film thickness.
The processed and fitted images are shown in Fig. 2.26 for repeated compressions
with aqueous A2-B3-CB[8]. In Fig. 2.26a, the oil droplet after 45 mins equilibration is
shown and was used as a reference for calculations on buckled droplets. Fig. 2.26b shows
the shape fitting carried out on buckled droplets after pumping in and compressing the
66
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 2.26 Shape fitting of pendant droplet experiments carried out with the Open-
Capsule software.22 (a) The reference droplet before droplet compression after 45 mins of
equilibration; (b) calculated shape fits based upon images of subsequent droplet compres-
sion leading to buckling; (c) repeated compression after expanding the buckled droplet to
8.5 µL volume with no equilibration time as in Fig. 2.23b.
interface. Equivalent shape fitting after droplet expansion and compression from 8.5 µL
(c) are also shown. A summary table of the calculated data is shown in Table 2.1.
Elastic Analysis
Compression Series (b) Compression Series (c)
1 2 3 1 2 3
2D Poisson’s Ratio, ν2D 0.51 0.52 0.52 0.50 0.49 -0.94
Area Compression Modulus, K2D / mN m
−1 126 109 101 88 84 423
2D Young’s Modulus, Y2D /mN m
−1 122 104 97 88 86 1645
Bending Modulus, EB /Nm 5.9E-15 2.7E-14 4.4E-14 9.3E-15 1.0E-13 5.4E-14
Average Wrinkle Wavelength, λ /µm 134 216 267 148 310 337
Layer Thickness, T /nm 529 1215 1610 798 2719 872
Fit Error /pixels 0.5 0.6 0.9 0.6 0.8 31.7
Table 2.1 A summary table of the calculated material properties of the images shown in
Fig. 2.26 using the OpenCapsule software.22
ν2D describes the lateral contraction of the interfacial film upon stretching. In contrast
to the 3D Poisson’s ratio, ν2D can lie between values of −1 < 0 < 1. Y2D describes the
interfacial film’s resistance to stretching, and is related to the 3D Young’s modulus (Y3D)
and T by the relation Y2D = Y3DT . K2D can then be defined as K2D = Y2D/2(1−ν2D), and
is a measure of the films resistance to compression. Bending modulus (EB) can be derived
by taking into account the wavelength of wrinkling of the solid buckled film, along with
the thickness of the film, and is a tendency of the material to resist bending.20,22,160,161
The observed ν2D for the initial compression, and after stretching and compression,
remain nearly the same around 0.5 implying the material is not very compressible. During
both compressions, K2D decreases implying the material becomes less resistant to com-
67
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
pression as buckling proceeds. This is likely because the formation of wrinkles allows points
for stress relaxation to occur. A similar trend is observed for Y2D, as ν2D being almost 0.5,
K2D ∼ Y2D. EB increases with further compression in both cases due to an increase in the
film thickness (EB ∝ T 3). The film thickness substantially increases in both compressions
as would be expected, reflecting the changes in material properties, beginning at over 500
nm after the initial buckling step has been observed. Likely the cross-linked film before
compression was smaller than this value in line with previous work,20 but this would be
dependent on several variables, such as concentration of the polymer materials and of the
charged surfactants. Direct measurement of this relaxed cross-linked film thickness in-situ
could not be achieved.
The magnitude of the elastic constants are similar to those reported previously for phys-
ically cross-linked polymer films, where host-guest chemistry linked two polymer chains
together as opposed to being composed of only small branching molecules.20 Differences
here show a more isotropic material as observed from ν2D being 0.5, and a film that has
over double the K2D (and therefore Y2D), and at least an order of magnitude higher EB.
This implies a stronger material is present, however, the supramolecular network here was
at least double in thickness so direct comparison is difficult.
Overall, the values for the second compression after stretching the cross-linked network
and immediately compressing it almost recover their initial properties, and follow the same
trends upon compression. A poor fit was observed in one case (fit error = 31.7 pixels) for
compression series (c), likely a result of departure from axisymmetric geometries which
are required for accurate shape fitting.22 It is also worth noting that this set of fits have
been carried out with simple Hookean elasticity, which likely is not completely accurate
when considering physically cross-linked polymer networks due to the tendency towards
viscoelasticity. Future experiments can focus upon changes in properties with regards to
the rate of compression to gain further insight into more complex material properties.22
2.3 Conclusions and Future Outlook
This Chapter has shown the potential to drive supramolecular polymer network forma-
tion from small molecules. By utilising liquid-liquid interfaces as templates to control
and drive self-assembly processes, the poor observed solubility of CB[8]-based branching
supramolecular polymers in the bulk could be inhibited. This allowed efficient inter-chain
cross-linking to occur, which led to a macroscopic gel made entirely from non-covalent
interactions between small molecules. This interfacial gel showed many favourable prop-
erties expected of supramolecular polymer networks, such as being elastic, self-healing, and
responsive to various stimuli. Interfacial rheology has also been utilised to understand ma-
terial properties that were otherwise unobtainable. Furthermore, the application of liquid
interfaces and emulsions as ideal candidates to directly observe and control self-assembly
processes has been demonstrated. This research platform is ideal for future investigations
in host-guest self-assembly phenomena, due to low consumption of reagents and the ability
to truly visualise the molecular processes taking place.
There are many exciting future directions for investigations to continue. The develop-
68
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
ment of an efficient understanding on how architectural changes in monomer can result in
extended supramolecular polymer networks in bulk solution is required. As computational
modelling of non-covalent interactions becomes less intensive when dealing with multiple
solvated molecules, this will become more accessible to assess likelihood of different mech-
anisms such as ring-chain polymerisations. Potential applications of small molecule-based
interfacial gels should also be thoroughly investigated with regards to encapsulation of
active materials, and in droplet surface engineering as demonstrated, which is of great
interest in industrial applications.
2.4 Experimental
2.4.1 Materials and Methods
All chemicals were purchased at the highest purity available from Sigma-Aldrich unless
otherwise specified. Perfluorocarbon solvents FC-72, Novec7100, Novec7500 and Fluor-
inert FC-40 were obtained from 3M.
Instrumentation
1H NMR (400 MHz) spectra were recorded using a Bruker Avance QNP 400. Chemical
Shifts are recorded in ppm in CDCl3, D2O and d6-DMSO with internal references set to δ
7.26 ppm, 4.79 ppm, and 2.50 ppm respectively. COSY 2D NMR experiments were carried
out where appropriate to aid in assigning peaks. ATR FT-IR spectroscopy was performed
using a PerkinElmer Spectrum 100 series FT-IR spectrometer equipped with a universal
ATR sampling accessory. UV-Vis spectra were recorded on a Varian Cary 4000 UV-Vis
spectrophotometer in aqueous solutions with 1 nm resolution at 25 ◦C. Photoirradiation
was carried out using a LZC-ORG photoreactor from Luzchem Research Inc. equipped
with UVA lamps centred at 360 nm.
Droplets were imaged using a Vision research Phantom Miro ex4-M fast camera, at-
tached to an Olympus IX71 inverted microscope (10x - 64x objectives). Microfluidic
devices were fabricated from PDMS by soft lithography, where the network was designed
in silico (AutoCAD), printed as a negative photo-mask, and transferred onto a silicon wafer
spin-coated with SU-8 photoresist via UV photolithography to form a mould. PDMS and
the cross-linker (Sylgard 184 elastomer kit, Dow Corning) in a 10:1 ratio were poured into
the mould and allowed to stand overnight at 70 ◦C. The PDMS layer was removed and
inlets and outlets were imprinted with a biopsy punch (1 mm). The imprinted PDMS and
a glass substrate were exposed to an oxygen plasma for 8 s, then pressed together to seal
the channels. To render the channels fluorophilic they were immediately flushed with a
0.5 v/v% solution of trichloro(1H,1H,2H,2H -perfluorooctyl)silane (Alfa Aesar) in FC-40
and subsequently cured at 120 ◦C overnight.
Monodisperse water-in-oil droplet emulsions were generated with a fluorophilic flow-
focussing microfluidic channel. The diameter of the junction was 60 or 80 µm with a
channel depth of 50 or 75 µm respectively. To generate droplets, the continuous oil phase
and the discrete aqueous phase were injected with 1 mL syringes (NORM-JECT) into the
69
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
microfluidic device via two syringe pumps (PHD 2000, Harvard Apparatus) and tubing
(PORTEX, polyethene 1.09 OD, 0.38 ID) with flow rates typically of 150 and 100 µL
h−1, respectively. At the intersection, shear forces and surfactant self-assembly caused
the formation of aqueous droplets in oil. The continuous phase comprised of the per-
fluorinated oil FC-40, with 2 wt.% neutral triblock copolymer fluorosurfactant (XL171,
Sphere Fluidics). To this was added 1 wt.% of Krytox 157-FSL (RFCOOH, a carboxylic
acid-terminated perfluoropolyether, Dupont) or RFNH2. The dispersed phase consisted
of an aqueous solution of monomers and CB[8], and dissolved 500 kDa dextran at 1 mg
mL−1 in some cases. For a typical experiment, a concentration of 0.55 mM A2, 0.37 mM
B3, and 1.10 mM CB[8] was used. Once formed, droplets were output onto the surface of
a glass slide into a reservoir of FC-40 oil (50 µL).
Pendant droplet measurements were performed using a commercial instrument (First
Ten Angstroms, FTA1000) and shape fitting for interfacial tension measurements was
carried out by the in-built software. The higher density oil-phase (FC-40, 0.001 wt.%
RFCOOH or 0.01 wt.% RFNH2) was hung from a 22 gauge needle (0.7176 mm OD, 0.4143
mm ID) and allowed to equilibrate with the continuous lighter aqueous phase for 45 mins.
For interfacial compression the droplet volumes were reduced from 5 µL until buckling was
observed at 0.1 µL s−1. ADA was added until in excess after 45 mins equilibration had
already taken place, and the hanging drop was allowed to equilibrate for a further 45 mins
before measurement. UVA exposure of the aqueous solution was carried out similarly to
the NMR measurements for 1 h before introduction of the oil droplet. Each measurement
was subject to 3 repetitions.
A2 Synthesis
A2 was synthesised according to literature procedures.
18 Subsequent ion exchange to the
more water-soluble tetrachloride salt was carried out. The tetrabromide salt was dissolved
in minimum H2O, to which a saturated aqueous solution of NH4PF6 was added dropwise
until no further precipitation took place. This was filtered, washed with H2O and dried
under vacuum. this tetrahexafluorophosphate salt was dissolved in minimum ACN, to
which a saturated ACN solution of tetrabutylammonium chloride was added until no
further precipitation took place. This was filtered, washed with ACN and dried under
vacuum to give the product as an orange solid.
B3 Synthesis
(4-(phenyldiazenyl)phenyl)methanol
(4-(phenyldiazenyl)phenyl)methanol was prepared according to previously reported liter-
ature procedures.162
70
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
1-(4-(phenyldiazenyl)benzyl)-1H -imidazole (1)
(4-(phenyldiazenyl)phenyl)methanol (3.04 g, 14.3 mmol) was dissolved in anhydrous NMP
(36 mL) under a flow of N2, to which was added CDI (3.02 g, 18.6 mmol). The reaction
mixture was heated and stirred at 150 ◦C for 3 h. After cooling, DCM was added (50 mL)
and extraction with water (25 mL x2) was carried out, followed by brine (25 mL). The
organic phases were combined and DCM was added until a final volume of 150 mL. Wa-
ter (100 mL) was added, followed by dropwise addition of aqueous HCl (5 M) until phase
transfer to the aqueous phase occurred. The phases were separated, and the organic phase
was extracted with further aqueous HCl (2 M, 10 mL x2). After combining the aqueous
phases, saturated NaHCO3 was added until reaching pH 7. The product precipitate was
filtered and dried to purity as an orange amorphous solid (0.98 g, 26%). TLC (SiO2)
DCM:acetone 1:1, Rf = 0.10. δH (CDCl3, 400 MHz) 7.89-7.93 (4H, m), 7.46-7.55 (3H,
m), 7.29 (2H, d, 8.6 Hz), 7.60 (1H, s), 7.13 (1H, t), 6.94 (1H, t), 5.21 (2H, s) ppm. δC
(CDCl3, 100 MHz) 152.5 (ArC), 152.4 (ArC), 138.9 (ArC), 137.5 (ImCH), 131.3 (ArCH),
130.1 (ArCH, 2C), 129.1 (ArCH, 2C), 127.9 (ImCH), 123.4 (ArCH, 2C), 122.9 (ArCH, 2C),
119.3 (ImCH), 68.0 (CH2) ppm. HRMS calculated for [M + H]
+: C16H15N4 263.1297,
found 263.1262. FT-IR νmax/cm
−1 3114brd, 3064brd, 2942brd, 1605, 1582, 1505, 1442.
3,3’,3”-(((benzene-1,3,5-tricarbonyl)tris(azanediyl))tris(propane-3,1-diyl))tris(1-
(4-(phenyldiazenyl)benzyl)-1H -imidazol-3-ium) bromide (3)
N1,N3,N5-tris(3-bromopropyl)benzene-1,3,5-tricarboxamide was synthesised as pre-
viously reported (2, 0.30 g, 0.53 mmol)28 and was dissolved in ACN (10 mL) and treated
with 1 (0.83 g, 3.2 mmol) which was previously dissolved in ACN (2 mL). The reaction
mixture was heated under reflux for 7 days until a viscous precipitate was formed. The
reaction mixture was cooled at room temperature and excess solvent was decanted off.
The precipitate was suspended in deionized water (500 mL) and stirred overnight. The
suspension obtained was filtered, and the liquid phase was freeze dried to yield the desired
71
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
compound 3 as an orange amorphous solid (85 mg, 12%). λmax(H2O with 0.7 v/v% of
DMSO)/nm 332 and 426. δH (d6-DMSO, 400 MHz, E -isomer) 9.60 (3H, s), 9.15 (3H,
m), 8.68 (3H, s), 7.92-7.80 (21H, m), 7.60-7.28 (12H, m), 5.50 (6H, s), 4.32 (6H, m), 3.36
(6H, m, under H2O solvent peak detected by COSY), 2.15 (6H, m). δC (d6-DMSO, 400
MHz, E -isomer) 165.5 (CO), 151.8, 137.9 (ArC), 136.8, 134.5 (ArCH), 131.9 (ArCH, 2C),
129.5 (ArCH, 2C), 129.4, 123.0 (ArCH, 2C), 122.6 (ArCH, 2C), 51.5 (CH2), 47.0 (CH2),
36.1 (CH2), 29.4 (CH2). HRMS calculated for [M -HBr]
2+: C66H65O3N15 557.7692, found
557.7695. FT-IR νmax/cm
−1 3403brd, 2922brd, 2852brd, 1729, 1654, 1542, 1448.
CB[n] Synthesis
CB[8] and CB[7] were prepared according to previously reported literature procedures.48
RFNH2 Synthesis
The synthesis of RFNH2 was adapted from literature procedures from microwave chem-
istry to conventional synthetic techniques.21 Krytox 157-FSL was purified before use by
dissolution in FC-72 and solvent extraction with THF (x3) followed by removal of solvent
in vacuo. FT-IR νmax/cm
−1 1779 (CO).
Krytox-Cl
Krytox 157-FSL (6.78 g, 3.4 mmol) was dissolved in Novec7100 (8 mL) under N2. Oxalyl
chloride (0.92 mL, 10.9 mmol) was added by syringe followed by 1 drop of catalytic an-
hydrous DMF. The reaction was stirred vigorously at room temperature for 16 h. The
reaction was quenched by removal of solvent and starting material in vacuo. The residue
was dissolved in FC-72 (10 mL) and extracted from anhydrous THF (5 mL x3) followed by
removal of solvent in vacuo to give a colourless oil Krytox-Cl (5.95 g, 88%). νmax/cm
−1
1809 (CO).
RFNH2
Ethylene diamine (9.0 mL, 134.6 mmol) was heated to 60 ◦C under N2. Krytox-Cl (5.05
g, 2.5 mmol) in Novec7500 (7 mL) was added by syringe over 30 mins. The reaction was
heated to 80 ◦C and stirred vigorously for 24 h. After cooling, MeOH (10 mL) was added
and the mixture stirred for 1 h. The solvents were removed in vacuo, and the residue was
dissolved in FC-72 (15 mL). The dispersion was filtered, and then extracted with THF (10
mL x3). The perfluorocarbon phases were combined and dried in vacuo to give RFNH2
as a slightly yellow viscous oil (4.85 g, 96%). δH (5 v/v% CDCl3 in perfluoroctane, 400
72
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
MHz) 7.38 (1H, br), 3.67-3.83 (2H, br), 3.51 (1H, br), 3.01 (1H, t), 1.13 (2H, br). FT-IR
νmax/cm
−1 3344brd (NH), 1705 (CO), 1539 (NH).
2.5 Acknowledgements
Dr. Aniello Palma and Dr. Richard M. Parker are acknowledged for all their help and
useful discussions throughout the project. Dr. Ziyi Yu is acknowledged for creating
the Si master chips of microfluidic devices by photolithographic methods. Dr. Villads
E. Johansen is also acknowledged for his kind help in setting up the pendant droplet
elastometry C/C++ software in Linux.
73
Chapter 3
Cooperative Intramolecular
Host-Guest Complexes
This work has been compiled into a manuscript for publication:
A. S. Groombridge, M. Olesin´ska, G. Wu, O. A. Scherman, 2017, in preparation.
3.1 Introduction
As introduced in Chapter 2, the fabrication of hyperbranched aqueous supramolecular
polymers remains a difficult challenge with CB[n] host-guest chemistry. Truly hyper-
branched polymers with no possibility of inter-chain cross-linking require a single monomer
(e.g. ABn) that displays distinct anisotropy in its reactivity.
11,90,91 Significant achieve-
ments to this goal have been made with some of the host macrocycles discussed so far,
by covalent attachment of guests to the macrocycle itself. Examples include crown eth-
ers,24,163,164 pillararenes,25 and CDs.23,165–167 as shown in Fig. 3.1.
Figure 3.1 Example structures of AB2 monomers for constructing hyperbranched poly-
mers. Adapted from references.23–25
The lack of accessible chemistry to functionalise CB[n]s with guest molecules means a
two component methodology must be employed.1,148 By taking advantage of the stepwise
and selective heteroternary (1:1:1) complexation possible with the larger CB[8], one can
74
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
envisage a suitable ABn monomer with two different guest moieties on it. Upon intro-
duction of CB[8], intermolecular heteroternary complexation can take place resulting in a
polymer. As discussed previously in Chapter 2, an architecture that is too small and rigid
will likely result in dimerisation with no further intermolecular interactions, and one that
is larger and very flexible will likely cyclise as oligomeric structures.26 There also exists
the possibility of intramolecular host-guest complexation.
Figure 3.2 The formation of an intramolecular heteroternary complex between naphthyl
and viologen units. Only the 1:1 complex was observed. Adapted from references.26
The formation of intramolecular host-guest complexes with CB[8] was first investigated
in 2002, where a linear molecule bearing naphthyl and viologen guests contained many
flexible units.26,149 As shown in Fig. 3.2, upon addition of 1 equiv. of CB[8], either a
back-folded intramolecular complex or a linear intermolecular complex could potentially
be formed. After analysing the 1H NMR and DOSY in detail, it was concluded that the
intramolecular complex was dominant. The intramolecular complex was then applied as
a chemical sensor, whereupon introduction of a stronger electron acceptor than viologen
the molecule would alter its conformation to a linear state and show a large colour change
related to changing donor-acceptor interactions. Further attachement of this binding motif
to a metal centre allowed photo-induced radical cation formation to occur.168
Figure 3.3 Intramolecular complexes being formed with CB[8] in response to a reducing
agent, based upon viologen homodimerisation. Adapted from references.27
Reversible switches based upon redox chemistry have been constructed with two vi-
ologen guests, where after reduction in the presence of CB[8] the viologen stacks and
forms a homodimer.169 This allowed interconversion between a [2]pseudorotaxane and
an intramolecular folded complex in response to external stimuli, acting as a molecular
machine. Interconversion between heteroternary and homodimer folded intramolecular
75
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
complexes has been achieved with similar redox chemistries, based upon viologen and
naphthyl guests.170 The intramolecular folded structure has also been utilised in order to
synthesise new cyclic structures.171 The photo-induced cycloaddition of 2-naphthalene was
carried out within a CB[8] cavity, where each naphthalene was connected by an oligoeth-
elyene glycol. The resultant cyclic structure was not observed in the absence of CB[8].
The use of molecular tweezers based on dinaphthalenes (see Fig. 3.3) in combination with
viologen groups also lead to intramolecular complexes in response to reduction.27
Precise control over molecular architectures for host-guest binding is highly desir-
able.172 It is of interest to investigate novel host-guest binding motifs in great detail, due
to the potential discovery of new materials with important properties. For example the
recent investigation by Wu et al into reported 1:1 binding with CB[8] led to the discovery
of new binding architectures through relating enthalpy and entropy of binding events.31
Binding that displayed significantly higher enthalpy yet the same stoichiometry of 1:1 bin-
ary complexes were found to in fact be 2:2 quaternary complexes. This discovery allowed
retrospective explanations for previously observed enhanced molecular conductance,173
and allowed the design of further binding motifs with highly controlled photo-physical
properties. Non-covalent dendrimers have also been synthesised with CB[6] and CB[8],
highlighting the selectivity available by employing multiple CB[n] molecules.28 The second
generation dendrimer held 13 separate components together entirely through host-guest
interactions as shown in Fig. 3.4, allowing future design of higher order structures to be
considered. Hou et al demonstrated the varied host-guest architectures and pseudoro-
taxanes that were possible by combining multifunctional branching guest molecules with
CB[8].174 Steric hindrance due to the bulk of CB[8] was a dominant factor in some of the
structures, but could be overcome by formation of heteroternary complexes with higher
binding affinities.
Figure 3.4 A [10]pseudorotaxane dendrimer composed of 13 different molecules. The
structure was formed with high selectivity, employing different types of CB[n] in water.
Adapted from references.28
Other developments have focussed upon ‘self-sorting’ complex mixtures, a common
phenomenon in nature that is rare in synthesised systems.29,172,175–177 The mixtures will
spontaneously sort themselves into well-defined and separate supramolecular assemblies
by non-covalent interactions. These have applications in artificial regulatory mixtures,
76
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
and in unique solution-based biomimetic devices. Mukhopadhyay et al demonstrated a
‘social’ self-sorting mixture where 12 individual components based on aqueous host-guest
chemistry could selectively form their individual complexes.175 The ensemble was social
(between molecules) rather than narcissistic (within molecules) as each supramolecular
interaction was between other components of the mixture rather than self-associative.
The response of the ensemble to pH and concentration were reversible, whereas the effect
of temperature caused irreversible changes to the mixture. Liu et al then showed that
mixtures of different CB[n]s alone could form self-sorting ensembles in the presence of
various guests with different affinities for each host.176 Differences in binding of the same
guest with different CB[n]s varied by 3 to 7 orders of magnitude, allowing exquisite control
in designing self-sorting mixtures. Even so, examples of self-sorting ensembles composed
of different CB[n]s are still rare, with even less examples where multiple CB[n]s can self-
sort onto the same guest molecule backbone. Jiang et al combined the use of entropically
favourable back-folded intramolecular complexation with CB[7] and CB[8] to form a self-
sorting mixture of very similar components.29 This allowed the selective synthesis of a
variety of supramolecular constructs from mixtures of a number of similar components,
with an example shown in Fig. 3.5. Another example is self-sorting supramolecular
polymerisation, where attachment of CB[7] to the benzyl centre of a monomer gives a
rigid ‘supramonomer’.177 CB[8] is then introduced that binds to naphthyl terminal groups,
which form intermolecular homoternary complexes and thus polymers. Control in this case
was achieved by a 100 times difference in Ka for CB[7] to the terminal naphthyl groups,
and CB[8] to the central benzyl group.
Figure 3.5 A mass spectrum and molecular model of a linear guest molecule back-folded
on itself to form an intramolecular heteroternary complex with CB[8] and a binary complex
with CB[7]. Adapted from references.29
Herein, effective synthesis of small molecules with multiple guests for CB[7] and CB[8]
in an AB2 architecture was investigated. A balance between flexibility, molecular size, and
water solubility was taken into consideration. While intra- or intermolecular complexes
both had the potential to form, the more entropically favourable intramolecular complex
was the only observed interaction. This is a unique example of intramolecular complex-
ation not based on a back-folded molecule, as it displayed very high binding constants
and slow disassociation kinetics. Further investigations into the response of this molecule
77
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
to various external stimuli, and potential applications in generating self-sorting mixtures,
were investigated.
3.2 Results and Discussion
3.2.1 Synthetic Design
The syntheses of AB2 molecules focussed upon a compact small molecule maximising po-
tential water-solubility. This was a result of previous experiences in Chapter 2, where the
second hydrophobic guest for CB[8] in the B3 molecule and molecular stacking were limit-
ing factors in water-solubility.178 Furthermore, previously reported linear supramolecular
polymers based on CB[8] have allowed for very small spacers between guests. Even a crys-
tal structure was possible for a CB[8]-based supramolecular polymer with a single CH2
spacer between guests.65 This implies that a small spacer is sufficient to prevent unfavour-
able steric crowding between adjacent CB[n] molecules. As such, it was thought that a
rigid aromatic core could be avoided for an AB2 architecture and so avoid the introduction
of unnecessary hydrophobicity and competing pi−pi interactions or H -bonding with units
such as BTAs.94
Figure 3.6 Proposed schematic self-assembly of supramolecular hyperbranched polymers
from CB[8] and an AB2 molecule based on two viologens and one naphthyl guest. Here,
the ratio of AB2 to CB[8] would be 1:2.
In order to achieve this, an AB2 multitopic guest molecule utilising heteroternary
complexation with viologen derivatives and either a naphthyl or azobenzene second guest
moiety was synthesised. Such a molecule should maintain anisotropy in the binding mech-
anism, as in their native states viologen will bind to the CB[8] molecule first (typical
Ka = 1.1× 105 M−1)1 and will not bind a second viologen unless reduced, i.e. B cannot
bind to B without an external stimulus. The reduced viologen could then form a 2:1
78
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
homoternary complex with overall binding constant Ka = 2.0×107 M−2.1,19 Sequentially,
the hydrophobic second guest will bind to form a 1:1:1 heteroternary host-guest complex
(a bond between A and B, overall ca. Ka ∼ 109−10 M−2),1 thus forming a cross-link
between monomers. A cannot bind to A. A further CB[8] molecule will bind to the
extra viologen unit present, resulting in a branch point and leading to a hyperbranched
polymer with a significant proportion of CB[8] present as binary complexes. This means
the Ka for stepwise polymer formation will be the second binding event, being on the
order of 104−6 M−1, similar to that of commonly used H -bonding UPy in chloroform.12
Importantly, the hyperbranched architecture should result in no internal cyclisation or
inter-chain cross-linking leading to an entirely branched and globular structure. This is
shown schematically in Fig. 3.6, where the polymer formed has some limited flexibility
from its sp3 CH2 units (R) and so can lead to a 3D polymer.
Figure 3.7 The synthetic pathway to AB2 guest molecules based on two viologens and
either a naphthyl or azobenzene guest (NpVio2 or AzoVio2). (a) MsCl, DIPEA, DCM, r.t.
4 h; (b) DIPEA, diethanolamine, THF, reflux 3 days; (c) CBr4, PPh3, DCM, r.t. 16 h; (d)
MeI, DCM, r.t. 16 h; (e) Me-Bipy, DMF, 80 ◦ C 3 days; (f) H2O, NH4PF6; (g) acetone,
Bu4NCl.
A simple tertiary amine core was targeted for the proposed AB2 molecules, as this
will be highly water soluble and will easily provide an asymmetric substrate for synthesis.
In the first case, diethanolamine was used as the substrate, as shown in Fig. 3.7. The
synthesis was designed such that viologen moieties were attached in the final stage, as the
presence of organic salts complicates further synthetic modification. Therefore, the hy-
drophobic second guest for CB[8] was attached first through activation of its aryl alcohol
equivalent with a mesyl group, followed by nucleophilic substitution with the diethano-
lamine. Ideal attachment of viologen will result in a product that precipitates cleanly
upon formation due to its increased hydrophilicity and charge density, otherwise tedious
separation by reverse-phase chromatography and ion exchange will be necessary. There-
fore, converting the alcohols to electrophilic alkyl bromides by the Appel reaction was
required, as the formation of hydrophilic bromide organic salts will aid this precipitation
from polar organic reaction solvents. If more hydrophilic alkyl chloride is used, then the
loss in reactivity could prolong the length of the reactions, and may cause loss of solubility
upon formation of the doubly-charged intermediate as opposed to the desired tetra-charged
product. Subsequent attachment of the viologen proceeded cleanly in DMF as the pure
79
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
tetra-charged product, where the reasonably hydrophilic double-charged intermediate is
still soluble. When alternative less polar solvents such as acetonitrile were used, a mixed
double- and tetra-charged product precipitant was obtained. The full synthesis was com-
pleted with reasonable yields to the final desired molecules NpVio2 and AzoVio2 from
the aryl alcohols (overall 20-37 %). As a mixed iodide/bromide salt, the molecules were
soluble in water, but as maximum water-solubility was desired for host-guest chemistry in
concentrated aqueous solutions, ion exchanges were carried out to the much more soluble
tetrachloride salt. These proceeded quantitatively by simple precipitation procedures to
first the organic-soluble hexafluorophosphate salt, then the chloride salt.
For the photo-responsive azobenzene-based AB2 molecule, AzoVio2, the equilibrium
E/Z ratio can also be readily determined by 1H NMR in D2O as 85 % E (see Fig. 3.33).
The photo-physical properties of this molecule in response to illumination with UVA light
were studied in detail and are discussed in section 3.2.6.
Assuming a simple isodesmic pathway for polymerisation (see Chapter 2), theoretical
molecular masses and conversions (ρ) of the potential polymer can be calculated following
Eq. 3.1 in combination with the Carothers equation and are summarised in Table 3.1,
where the monomer molecular mass was calculated assuming two CB[8] molecules were
already associated.24 It becomes clear that the higher binding constant could potentially
give rise to high molecular weight polymers at low concentrations, as yet unachieved in
water.
DP =
1
1− ρ
Ka =
ρ
(1− ρ)(2[AB2]0 − ρ[AB2]0)
0 = Ka[AB2]0ρ
2 − (1− 3Ka[AB2]0)ρ+ 2Ka[AB2]0
(3.1)
[AB2]0 /mM ρ DP Mn /kDa
0.1 0.9902 1× 102 3.4× 105
1.0 0.9990 1× 103 3.4× 106
10.0 0.9999 1× 104 3.4× 107
100.0 1.0000 1× 105 3.4× 108
Table 3.1 A table showing the calculated ρ, DP , and polymer molecular weight (Mn)
for the NpVio2 molecule complexed with two CB[8] molecules (3352 g mol
−1), assuming
a Ka of 10
6.24
The conformation of these two AB2 molecules in water have the potential to self-
associate through weak donor-acceptor interactions between the electron-deficient viologen
and electron-rich naphthyl or azobenzene as in the CB[8] cavity, however 1H NMR spectra
in D2O show sharp and distinct peaks that do not have concentration-dependent chemical
shifts. This showed that the kinetics of self-association must be very fast, and that any
potential interaction was intramolecular and not intermolecular. In the following sections,
80
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
it became clear that intermolecular complexation and polymerisation was not observed,
with cooperative intramolecular binding being the dominating effect.
3.2.2 NMR Titrations with CB[7] and CB[8]
After having synthesised two variations on the AB2 architecture, NpVio2 and AzoVio2,
NMR spectroscopy, UV-vis spectrophotometry, and isothermal titration calorimetry (ITC)
were focussed upon to study the supramolecular assemblies formed with CB[7] and CB[8].
From a series of 1D and 2D experiments (1D 1H & 13C, 2D 1H-1H COSY, 1H-13C HSQC,
1H-13C HMBC, and 1H-D DOSY), the solution-phase structures of the assemblies were
derived. At 5 mM AB2 concentration in D2O, CB[7] or CB[8] solid was added and heated
at 70 ◦ C and ultra-sonicated to dissolve, then left for 2 hours to equilibrate before NMR
analysis. The volume was also kept constant at 0.7 mL, as DOSY experiments can be
sensitive to changes in concentration and volume.
NpVio2
Figure 3.8 Stacked 1H NMR spectra of NpVio2 in D2O (bottom spectrum), followed by
addition of 0.8 (middle spectrum) and then 2.0 equiv. of CB[8] (top spectrum). The peak
assignments in red show proton environments outside of the CB[8] cavity or forming a 1:1
binary complex with viologen, and those in blue are inside the cavity as a heteroternary
1:1:1 complex.
81
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Upon introducing up to 1 equiv. of CB[8], a new set of well-defined sharp peaks can be
observed in the 1H NMR spectrum, coinciding with the appearance of CB[8] peaks. This
is exemplified in Fig. 3.8, showing how the addition of 0.8 equiv. of CB[8] results in all
of the introduced CB[8] forming a host-guest complex with NpVio2, with excess NpVio2
unchanged in solution. Of particular note is the splitting of CB[8] protons q and r into
two doublets rather than one. This is a known phenomena that indicates the kinetics of
dissociation for the complex are slow (at least 0.5 s). This is because the protons at the top
of CB[8] are in a different local environment to those at the bottom, giving rise to different
chemical shifts if the binding constant is high and the CB[8] is ‘locked’ in place. CB[8] is
free to rotate about the complex, but cannot disassociate and re-associate quickly, which
would result in averaging of the peaks. This has been observed before in the literature for
systems with very high binding constants, for example in the formation of 2:2 quaternary
complexes.31
The chemical shifts of NpVio2 have also dramatically changed. If a peak shifts upfield
(i.e. to a lower ppm), this indicates it has become more shielded to the magnetic field.
If a peak shifts downfield then this implies it is more deshielded. With respect to CB[8]
binding, this means that if a proton is encapsulated within the cavity, then it will become
shielded and an upfield shift of around 1 ppm is expected.1 If a proton resides close to the
carbonyl portals of CB[8], then it will also become more deshielded due to proximity to
the C=O ring current.
As such, we can see in Fig. 3.8 that all of the protons corresponding to the naph-
thyl group in the molecule have shifted significantly upfield meaning they are complexed
within the CB[8] cavity. An example set of 2D NMR experiments are shown in section
3.4.2 to demonstrate the peak assignments for the complexation of 1 CB[8] molecule to
NpVio2, with subsequent data in the Supplementary Information, section A.1.1. As there
are two equivalent viologen environments present in NpVio2, when one CB[8] is complexed
asymmetry is introduced to the molecule. The peaks corresponding to the viologen com-
plexed with naphthyl show shifts upfield, and those uncomplexed show shifts downfield.
Due to the rigid planar structure of viologen it is possible to derive the precise location of
the CB[8] on the molecule by the degree of shifting observed for different environments,
provided its location is not dynamic. For example, peak a showed a downfield shift upon
complexation, whereas peak b showed a slight upfield shift. We can determine that the
CB[8] must therefore encapsulate protons in environment b, but environment a is just
outside the cavity and interacts with the carbonyl portals. Peaks d and c show very large
upfield shifts of 1.8 and 1.4 ppm, showing they are far inside the cavity. Peak e then has
a smaller upfield shift, and also displays some broadening; this can be good evidence of
dynamic binding in this region. Peaks f, g, and h reflect this hypothesis, by showing they
are also encapsulated in the cavity and have more dynamic binding, likely due to their
higher degrees of conformational freedom. The effect of downfield shifting on the unbound
viologen environments was seen to decay with proposed distance from the CB[8] carbonyl
portals, with the effect greatest for e and smallest for b.
After the addition of a further CB[8], the second viologen present in the molecule can
82
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.9 The proposed complex of NpVio2 with 2 CB[8] molecules forming an asym-
metric complex. The positions of CB[8] were derived from the 1H NMR experiments and
optimisation was carried out with MMFF94 molecular mechanics in Avogadro v1.2.
then be complexed to form a 1:1 binary complex. This was thought to be much more
dynamic, due to the lower expected binding constant for this interaction. The naphthyl
protons remain shifted upfield as they are still bound to CB[8] in a 1:1:1 complex with a
viologen. This is evidenced by peaks a - h for the bound species remaining upfield shifted
in similar positions, however in a more dynamic state. The previously unbound viologen
now binds dynamically to CB[8], but only from environments a - d, with e, f and g showing
downfield shifts. An additional CB[8] environment was also observed, but this extra peak
did not display any splitting. This means the second CB[8] has much faster kinetics for
binding to the free viologen and has freedom to disassociate and re-associate. Overall this
shows that intramolecular host-guest binding takes precedent with no clear evidence of
intermolecular association. Molecular mechanics calculations also provide evidence for the
proposed binding, shown in Fig. 3.9.
Similar experiments can be carried out with CB[7], which can only encapsulate one
guest molecule as opposed to two. Upon complexation with 1 CB[7], similar changes in
chemical shift are observed as for with CB[8] but with much broader peaks. This broadness
was attributed to a smaller overall binding affinity (typically Ka = 10
4−6 M−1),1 and
some motion of the CB[7] occurring along the molecule. This results in averaging the
chemical environment as it is in partial flux, but predominately in a single bound state.
The complexed viologen peaks c, d, e exhibit significant upfield shifts, more so for c and
d, allowing the derivation of the position of CB[7]. Peaks a and b show slight downfield
shifts, meaning these environments are just outside the cavity. In contrast to complexation
with CB[8], peaks f and g remain unchanged in their chemical shifts, likely a result of
the smaller CB[7] cavity size not allowing their partial complexation. A further result of
complexation is the remaining uncomplexed viologen became conformationally confined
adjacent to the free naphthyl group. As a result of the naphthyl group’s aromatic ring
current, the nearby protons e, f, c, d, h are relatively more shielded and show slight upfield
83
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.10 Stacked 1H NMR spectra of NpVio2 in D2O (bottom spectrum), followed
by addition of 1.0 (middle spectrum) and then 2.0 equiv. of CB[7] (top spectrum). The
peak assignments in red show proton environments outside of the CB[7] cavity, and those
in blue are inside the cavity.
shifts. Equally a downfield shift can be observed for protons g just outside the effect of
the ring current. Splitting of the broad CB[7] peak q was observed, but was not obvious
for peak r. This is a result of the protons facing towards the CB[7] experiencing more of
a local difference in chemical environment than those facing away between the top and
bottom of the CB[7]. The naphthyl peaks also demonstrate a partial downfield shift as
a result of proximity to the carbonyl portals of the complexed CB[7], more obvious for
peaks i, j, and k.
Upon complexation of a further CB[7] molecule, the spectrum became more dynamic
in its binding, with very broad peaks. Key information could still be elucidated, however,
such as the loss of asymmetry in the binding resulting in a single peak for the viologen
protons a - e that could easily be assigned. Peaks a and e were partially downfield shifted,
whereas b, c, and d were upfield shifted to different extents. This allowed the CB[7]
position to be derived as fully encapsulating protons c and d with partial encapsulation
of b. The naphthyl peaks l - o showed a further downfield shift as compared to when
complexing one CB[7], further supporting the proposed structure as this group would be
adjacent to the carbonyl portals of two CB[7] molecules. The sp3 protons f, g and h were
84
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
in highly dynamic states and could not be observed clearly. Overall this allows the model
in Fig. 3.11 to be proposed for the binding architecture.
Figure 3.11 The proposed complex of NpVio2 with 2 CB[7] molecules forming a sym-
metric complex. The positions of CB[7] are derived from the 1H NMR experiments and
optimisation was carried out with MMFF94 molecular mechanics in Avogadro v1.2.
Figure 3.12 1H NMR spectrum of NpVio2 in D2O with 1 equiv. each of CB[8] and of
CB[7]. The initially symmetric molecule is now asymmetric, so peaks corresponding to
complexation with CB[7] or CB[8] can be assigned with the prefix 7 or 8 respectively.
Sequential addition of 1 equiv. of CB[8] followed by 1 equiv. of CB[7] or vice versa
led to the formation of a self-sorting mixed complex. The CB[8] bound as previous to
form a 1:1:1 ternary complex with a viologen and naphthyl group, and the CB[7] bound
to the remaining viologen. The peak assignments are shown in Fig. 3.12. Similar upfield
shifts were observed for the naphthyl protons as for complexation with 1 CB[8]. Equally,
the largest shifts for the viologen moieties upon complexation of either CB[7] or CB[8]
were observed for protons c and d, evidencing this is the predominant location of the
85
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.13 The proposed complex of NpVio2 with CB[8] and CB[7] molecules form-
ing a self-sorting complex. The positions of CB[7] and CB[8] are derived from the 1H
NMR experiments and optimisation was carried out with MMFF94 molecular mechanics
in Avogadro v1.2.
CB cavities. The alkyl units f and g were only within the cavity for CB[8], showing
upfield shifts, whereas downfield shifts are observed with CB[7]. Protons h and 8g also
interestingly display two equilibrium states. Of further interest in this case is that only
peak 8q facing into the cavity shows splitting into two doublets as for complexation,
whereas 8r has averaged into a single doublet. A calculated model is shown in Fig. 3.13.
AzoVio2
Upon introduction of 1 CB[8] to AzoVio2, complexation of one viologen moiety and partial
encapsulation of the azobenzene was observed. Asymmetry in the molecule has again been
introduced, with two sets of peaks corresponding to each viologen moiety present. The
viologen aromatic peaks c, d, and e show a significant upfield shift of more than 1 ppm,
and b shows a smaller shift of 0.3 ppm. This shows the CB[8] cavity fully encapsulates c,
d, and e with partial encapsulation of b. In comparison to the binary complex between
NpVio2 and 1 CB[8], peak a also shows an upfield shift of almost 0.4 ppm. This was
thought to not be due to encapsulation in CB[8], instead to be due to the extended length
of the azobenzene, allowing an overlap of the aromatic ring current with protons a (ring k
- m), causing a shielding effect. Peaks f and g also show encapsulation within the cavity
and some broadening similar to NpVio2. The first aromatic ring of the azobenzene (i
and j ) shows large upfield shifts, with k showing downfield shifts and l and m remaining
unchanged. This implies the CB[8] can only encapsulate one ring of the azobenzene and
the N=N double bond. The CB[8] peaks q and r show splitting as for NpVio2, implying
the complex formed is static with slow disassociation kinetics.
With a further CB[8] molecule, a similar asymmetric intramolecular complex can be
observed as for NpVio2, with both a heteroternary complex and a binary complex present.
86
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.14 Stacked 1H NMR spectra of AzoVio2 in D2O (bottom spectrum), followed by
addition of 1.0 (middle spectrum) and then 2.0 equiv. of CB[8] (top spectrum). The peak
assignments in red show proton environments outside of the CB[8] cavity or forming a 1:1
binary complex with viologen, and those in blue are inside the cavity as a heteroternary
1:1:1 complex.
This is evidenced from the changes in peaks of the previously uncomplexed viologen, where
b, c, d (in red) show upfield shifts, and e, f, a have shifted downfield. Changes to the peaks
for the heteroternary complex are also observed, with most significantly the peak e (in
blue) is shifted much less upfield compared to before, likely a result of the adjacent second
CB[8] making complexation close to the core of the molecule less favourable. The second
CB[8] peaks q and r do not show splitting again showing how this second interaction is
less strong. This allowed the model in Fig. 3.15 to be proposed.
Upon introduction of a CB[7] host, one viologen was encapsulated in its cavity as
for NpVio2 however in this case the peaks were sharper. This implied there was a more
stable equilibrium state. The complexed viologen peaks c, d, e exhibit significant upfield
shifts being inside the cavity, with b and a showing downfield shifts. The peaks f and g
adjacent to CB[7] were unchanged in their chemical shift as for NpVio2. As a result of
complexation, the azobenzene is now conformationally confined adjacent to the remaining
viologen, allowing interactions of their ring currents to affect relative shielding from the
magnetic field. This was mostly observed for peaks a, b, d and g, which correlates with
87
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.15 The proposed complex of AzoVio2 with 2 CB[8] molecules forming an asym-
metric complex. The positions of CB[8] are derived from the 1H NMR experiments and
optimisation was carried out with MMFF94 molecular mechanics in Avogadro v1.2.
the azo bond separation between the two aromatic rings of the azobenzene. Splitting of
both CB[7] peaks q and r was observed, evidencing this was also a ‘locked’ complex. The
azobenzene protons remained almost unchanged after complexation, with only a slight
downfield shift observed for protons i, which are the closest to the bound CB[7] carbonyl
portals.
Upon introduction of a further CB[7], a complex spectrum was obtained. The spectrum
was less dynamic and more resolved than for the case with NpVio2 likely due to the more
extended azobenzene less affected by confinement between the two CB[7] molecules. The
dominant E isomer complex could be readily observed, with symmetry being recovered and
so a single set of peaks shown for the viologen protons a - g. Peaks a and e were downfield
shifted, whereas b, c, and d were upfield shifted to different extents. This allowed the CB[7]
position to be derived as fully encapsulating protons c and d with partial encapsulation
of b. The azobenzene protons i, j, and k showed significant downfield shifting being in
close proximity to two sets of CB[7] cabronyl portals, with l and m remaining relatively
unchanged. Peaks f and g showed significant downfield shifts also expected from proximity
to the portals, with h becoming too broad to observe.
A subset of peaks was also observed relating to the Z isomer in two different config-
urations. As the azobenzene moiety is not encapsulated within the cavity, evidenced by
its protons experiencing significant downfield shifting, it will be present in equilibrium
between the E and Z isomer. Two clear sets of peaks for the viologen protons a - e were
observed in similar regions as for the E isomer. This could be a result of the introduction
of asymmetry in the complex due to the presence of the Z isomer, causing steric crowding
on perhaps one side of the molecule. However, this would require a kinetically trapped
complex to prevent averaging of the peaks in the 1H NMR. A more likely reason would be
the presence of two equilibrium states with similar energy minima in competition. This
is supported by further experiments into the photo-isomerisation of these complexes in
88
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.16 Stacked 1H NMR spectra of AzoVio2 in D2O (bottom spectrum), followed by
addition of 1.0 (middle spectrum) and then 2.0 equiv. of CB[7] (top spectrum). The peak
assignments in red show proton environments outside of the CB[7] cavity, and those in
blue are inside the cavity. Two different complexes with the Z isomer could be observed,
denoted Zx and Zx’ where applicable.
section 3.2.6, where after UVA exposure to drive the formation of the Z isomer, a single Z
isomeric state dominates as opposed to an asymmetric complex. This allowed the model
in Fig. 3.17 to be proposed for the E isomer, with further details in section 3.2.6 with
regards to the Z isomer.
A mixed CB[7] and CB[8] complex was also prepared as for NpVio2, with spectrum
shown in Fig. 3.18. Here we can see similarities to binding 1 CB[8] or 1 CB[7] to AzoVio2.
For example, the largest upfield shifts for the azobenzene moiety were observed for protons
i and j, with k - m extended beyond the CB[8] cavity. Equally, the largest shifts for the
viologen moieties upon complexation of either CB[7] or CB[8] were observed for protons
c and d, evidencing this is the predominant location of the CB cavities. The alkyl units
f and g were only within the cavity for CB[8], showing upfield shifts, whereas downfield
shifts are observed with CB[7]. Similar to the case with NpVio2, splitting of the peaks q
and r was only observed for the CB[8] complex. A calculated model is shown in Fig. 3.19.
89
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.17 The proposed complex of AzoVio2 with 2 CB[7] molecules forming a symmet-
ric complex. The positions of CB[7] are derived from the 1H NMR experiments and optim-
isation was carried out with MMFF94 molecular mechanics in Avogadro v1.2. Azobenzene
being twisted or planar in polar solvents is still under debate.30
Figure 3.18 1H NMR spectrum of AzoVio2 in D2O with 1 equiv. of CB[8] and of CB[7].
The initially symmetric molecule is now asymmetric, so peaks corresponding to complex-
ation with CB[7] or CB[8] can be assigned with the prefix 7 or 8 respectively.
90
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.19 The proposed complex of AzoVio2 with CB[8] and CB[7] molecules form-
ing a self-sorting complex. The positions of CB[7] and CB[8] are derived from the 1H
NMR experiments and optimisation was carried out with MMFF94 molecular mechanics
in Avogadro v1.2.
91
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Inverse Architecture
It was hypothesised that if the asymmetric architecture was inverted such that there was
one hydrophilic viologen moiety and two hydrophobic naphthyl moieties present, it may
be possible to drive the formation of intermolecular supramolecular assemblies (structure
shown in Fig. 3.20). This was likely possible due to the presence of two hydrophobic
components, meaning that upon introduction of 1 equiv. of CB[8], self-organisation would
take place to conceal both of the naphthyl groups from water, i.e. inside CB[8] cavities.
The synthesis was carried out taking advantage of the methodology developed, and is
detailed in section 3.4.5.
Figure 3.20 1H NMR spectrum of Np2Vio in D2O with 1 equiv. of CB[8]. The initially
symmetric molecule is now asymmetric.
Upon introduction of CB[8], however, a similar type of complexation can be observed
as previous in Fig. 3.20, where the conformation is more favourable for the formation of
intramolecular complexes as compared to intermolecular. However, the NMR spectrum
shows high complexity due to the presence of two naphthyl groups which have very similar
proton and carbon chemical shift, making spectrum interpretation challenging. All of the
aromatic peaks for viologen show a marked upfield shift, with the largest shifts present for
c and d, and smaller for b and e. Peaks a and f show downfield shifts, so the CB[8] cavity
must be sat in the centre of the viologen mainly over protons c and d. Peaks h and g have
become too broad to enable assignment through 1H-1H or 1H-13C correlation spectra,
showing they have become highly dynamic in their environment. One of the naphthyl
groups has also been encapsulated in CB[8], introducing asymmetry in the complex. Peaks
(in blue) i, j, and k show smaller upfield shifts than l - o, implying that the centre of
the cavity sits more over the end of the naphthyl group. This is more obvious from
the molecular model proposed in Fig. 3.21. The unbound naphthyl group now shows
significant downfield shifting for peaks i, j, k, l, and o, more so than would be expected
92
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
for simply being in proximity to the carbonyl portals of CB[8]. Peak n shows a smaller
downfield shift, and peak m shows a very slight upfield shift of 0.05 ppm. This was thought
to be due to the unbound hydrophobic naphthyl having very unfavourable interactions with
the solvent, and so will be conformationally confined as close to the external surface of
the CB[8] as possible. This results in more of the protons being in quite close proximity
to the carbonyl portals, resulting in downfield shifts. Peak m could be displaying a slight
upfield shift due to being just outside the ring current effect of the carbonyl, similar to the
aromatic ring currents discussed previously, resulting in being more shielded. The CB[8]
protons q and r showed splitting as before, implying the CB[8] is kinetically trapped as
seen for the previous systems.
Figure 3.21 The proposed complexes of Np2Vio with 1 CB[8] molecule forming an asym-
metric complex. The position of CB[8] was derived from the 1H NMR experiments and
optimisation was carried out with MMFF94 molecular mechanics in Avogadro v1.2.
Overall this allowed the models in Fig. 3.21 to be proposed, where it is most likely
that the unbound naphthyl will be conformationally confined close to the external surface
of CB[8] due to its hydrophobicity.
With CB[7], more complex spectra are observed as shown in Fig. 3.22. Previously
upon slow introduction of CB[7] or CB[8] in steps, well defined peaks can be observed
for the newly formed complex. This means a simple two component mixture was present,
unbound free guest molecules and guest-CB[n] conjugates (as in Fig. 3.8 with 0.8 equiv.
CB[8]). In the case of CB[7] addition to Np2Vio, the spectrum broadened substantially
and lost much of its intensity. This implies that after introduction of a small amount of
CB[7], some of the viologen can be bound and then the two free naphthyl groups will
stack, possibly in an intermolecular fashion, into higher ordered structures. A suggested
model is shown in Fig. 3.23. In summary, the NMR spectroscopy carried out thus far
provides strong evidence for the formation of exclusively intramolecular complexes of the
AB2 and A2B guest molecules with CB[8]. The assembly pathway is shown in Fig. 3.24.
93
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.22 1H NMR spectrum of Np2Vio in D2O (bottom) with 0.2 (middle) and 1.0
equiv. (top) of CB[7].
Figure 3.23 The proposed complex of Np2Vio with 1 CB[7] molecule. Optimisation was
carried out with MMFF94 molecular mechanics in Avogadro v1.2..
94
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.24 An overview scheme of the proposed assembly process of the guest molecules
with CB[7] and CB[8] evidenced by NMR spectroscopy.
95
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
3.2.3 DOSY Experiments
Diffusion-ordered NMR spectroscopy (DOSY) is an excellent technique for use in host-
guest and supramolecular chemistry. It seeks to correlate peaks in the NMR spectrum with
their diffusion coefficient (D) in solution, and is often used in characterising supramolecular
systems.14,65 Therefore, if the disassociation kinetics of host-guest complexes are slow, then
individual diffusion coefficients for complex mixtures can be derived. This information can
then be used to show if there is a single complex present, or multiple. It can also be used
to show how large the complexes are, for example if it is a simple dimeric structure, or if
oligomeric or polymeric species are observed as these would give rise to very large changes
in diffusion. Molecular size can be derived by the Stokes-Einstein equation (Eq. 2.4),
assuming a spherical shape.65
Molecule Solo 1CB[8] 2CB[8] 1CB[7] 2CB[7] 1CB[7] + 1CB[8]
NpVio2 2.95 2.04 1.80 2.06 1.78 1.77
AzoVio2 2.78 1.98 1.65 2.03 1.60 1.76
Np2Vio 3.07 2.03 - - - -
CB[7] 2.63 - - - - -
CB[8] 2.84* - - - - -
Table 3.2 Collated DOSY results for the supramolecular assemblies formed in units of
10−10 m2 s−1. The diffusion coefficient (D) has been derived from the most intense peak
in the spectrum, with highest error being ±0.01. Each measurement was undertaken at
5 mM concentration, with the exception of *CB[8] being a filtered saturated solution c.a.
0.5 mM.
A summary of D obtained for each complex is shown in Table 3.2, and the calculated
diameter via the Stokes-Einstein equation is shown in Table 3.3. An example DOSY
spectrum is shown in Fig. 3.41. The coefficients shown were calculated by peak intens-
ity rather than integration (24 pulses, 32 scans), as overlapping peaks will give rise to
averaging over the two environments. The amount that D and diameters reduced was
similar for all the molecules studied, in-line with literature values for small CB[n]-based
host guest complexes.149,152,179 It should be noted that direct comparison between other
works is difficult, as D depends heavily on acquisition parameters and the conditions em-
ployed, however internal comparison is a highly valuable metric. This provides further
evidence that the complexation observed was completely intramolecular in nature, rather
than intermolecular.
Molecule Solo 1CB[8] 2CB[8] 1CB[7] 2CB[7] 1CB[7] + 1CB[8]
NpVio2 1.66 2.41 2.73 2.38 2.76 2.77
AzoVio2 1.77 2.48 2.97 2.42 3.07 2.79
Np2Vio 1.60 2.42 - - - -
CB[7] 1.87 - - - - -
CB[8] 1.73* - - - - -
Table 3.3 A table of hydrodynamic diameters (nm) of the complexes formed calculated
from D, derived from DOSY NMR experiments. Calculations were carried out with the
Stokes-Einstein equation, at 25 ◦ C.
96
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
3.2.4 UV-vis Spectrophotometry
UV-vis spectrophotometry is a useful technique for monitoring CB[n] host-guest binding
as this typically induces many changes in the electronic properties of the guests. Dilute
solutions of each AB2 molecule were titrated with CB[7] or CB[8] solutions, and their
spectra are shown in Fig. 3.25. The initial NpVio2 spectrum has λmax at 223 and 260
nm. Naphthyl groups typically have an intense absorption band around 220 nm and a
minor band around 280 nm, and here there is evidence of a shoulder peak being present
with NpVio2 at ca. 285 nm. Viologen has an intense absorption at 260 nm overlapping
with this.19 Upon addition of CB[8], the intramolecular heteroternary complex will first
be formed up to 1 equiv., followed by the binary complex up to 2 equiv.. This is reflected
in the significant drop in intensity for the naphthyl peak at 223 nm upon complexation,
and a slight red shift to λmax = 225 nm, followed by no more significant changes. The
viologen peak at 260 nm also decreases in intensity, first displaying a small blue shift to
259 nm with 1 equiv., then a red shift to 263 nm. These shifts reflect the donor-acceptor
stacking interactions between naphthyl and viologen with 1 equiv. of CB[8], followed by
binary complexation of viologen with a further CB[8] leading to averaging of the two
environments. As the viologen peak reduced in intensity upon both being encapsulated
within CB[8], the secondary band for naphthyl became more clear, with an isobestic point
observed at 295 nm.
Figure 3.25 UV-vis spectra of each AB2 molecule followed by titration with CB[8] or
CB[7].
97
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
With introduction of CB[7], the naphthyl peak does not initially change until 1 equiv.
of CB[7] has been added, whereupon a red shift of the band to 226 nm was observed with
a small reduction in intensity. This was thought to be due to confinement of the naphthyl
between the electronegative carbonyl portals of the two CB[7] molecules (see Fig. 3.11).
The viologen peak reduces in intensity as binary complexation occurs, initially red shifting
to 262 nm with 1 equiv. of CB[7], then blue shifting to 257 nm.
AzoVio2 showed λmax of 260 and 332 nm. The peak at 260 nm is typical of viologen,
similar to NpVio2. The peak at 333 nm corresponded to the pi − pi∗ transition of E -
azobenzene. Upon introduction of 1 and 2 equiv. of CB[8], the viologen peak decreased in
intensity with a continual red shift to 263 nm. The azobenzene peak decreased rapidly in
intensity up to 1 equiv. of CB[8] with a blue shift to 324 nm coinciding with intramolecular
complexation, followed by a slight red shift to 328 nm. This supports the formation of
intramolecular complexes due to the donor-acceptor interactions within the cavity. With
CB[7], the viologen peak decreased in intensity whilst first red shifting to 262 nm with
1 equiv. then blue shifting to 256 nm with a further equiv. similar as for NpVio2. The
azobenzene peak showed a small decrease in intensity but was otherwise unaltered.
Np2Vio showed λmax of 226 and 265 nm, corresponding to the naphthyl and viologen
groups respectively. There is also a shoulder peak present at 220 nm, likely due to the
existence of the two naphthyl groups in different conformations or through intramolecular
stacking between them. Upon introduction of 1 equiv. of CB[8], the intensity of the
naphthyl peak at 226 nm decreases and the shoulder peak is mostly lost as stacking
interactions disappear. The viologen peak decreases in intensity in a similar fashion, with
an isobestic point observed at 291 nm due to the secondary naphthyl band as for NpVio2.
With CB[7], as the viologen is encapsulated with up to 1 equiv. it decreases in intensity
and blue shifts to 263 nm. As evidenced by NMR experiments, introduction of CB[7]
results in stacking of naphthyl components as the viologen becomes encapsulated. The
main naphthyl peak red shifts to 232 nm, and the shoulder peak becomes much clearer,
also red shifting to 224 nm. This corresponds to enhanced stacking between the aromatic
naphthyl units.
3.2.5 Isothermal Titration Calorimetry
Isothermal titration calorimetry (ITC) is an often-used method to determine supramolecu-
lar binding events, binding stoichiometry, and Gibbs free energies of solution processes.42,180
A cell containing a substrate is titrated with a ligand, and the heat difference with a ref-
erence cell is recorded. Fitting of the resulting isotherm with various models allows the
direct determination of solution binding events. Although typically used in life sciences
for studying protein binding, it has seen a surge in application for studying host-guest
chemistry, often in conjunction with other techniques to gain a full understanding of the
kinetics and thermodynamics of host-guest interactions.1 For CB[n]-based supramolecular
chemistry, ITC was used to extract the effect of solvation on the unusually high binding
affinities for CB[8] complexes and resulted in the concept of ‘high-energy water’ as a key
driving force in their host-guest chemistries.181 It was more recently used to correlate
98
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
binding stoichiometries with the actual solution state geometries by looking closely at
enthalpy versus entropy for various reported host-guest complexes.31
Figure 3.26 A schematic diagram showing the proposed binding events observed in the
ITC measurements shown in Fig. 3.27 for NpVio2 and AzoVio2.
In this case, a solution of AB2 guest molecule (1 mM) was injected into a cell con-
taining CB[7] (61 µM) or CB[8] (51 µM) solution in multiple stepwise injections, with the
isotherms shown in Fig. 3.27. The concentrations of CB[7] and CB[8] were calibrated by a
separate titration with ADA.HCl, as CB[n] will contain some encapsulated H2O. NpVio2
and AzoVio2 showed similar isotherms with CB[8] and CB[7]. After the first few injec-
tions the guest molecule will be present at a low concentration in vast excess of CB[n].
This means that 2 CB[n] molecules will bind spontaneously, evidenced by the binding
stoichiometry observed to be n = 0.5 with both molecules for the first step in the isotherm
for CB[7] and CB[8]. However, there is a following step in each case at ca. n = 1.2.
This was thought to be a result of the first CB[n] binding being stronger than that of the
second, either due to steric crowding (CB[7]), or heteroternary being stronger than binary
complexation (CB[8]). This is shown schematically in Fig. 3.26, where K1CB[n] signifies
the first binding with CB[n], and K2CB[n] with the second. Similar multi-step ITC results
have been reported on investigating supramolecular polymerisations.177
Whilst fitting of the secondary step of the isotherm could not be achieved, fitting
of the first step was attempted with a two sites binding model, calculating Ka for each
site.180 With CB[8] the K1CB[8] and K2CB[8] were calculated to be > 1.0 × 1010 M−1
and 8.6 × 106 M−1 for NpVio2, and > 1.0 × 109 M−1 and 2.5 × 107 M−1 for AzoVio2,
respectively. However, due to ITC not able to measure Ka > 10
9 M−1, the results were
deemed inaccurate with CB[8]. With CB[7] the K1CB[7] and K2CB[7] were calculated being
5.1 × 105 M−1 and 5.3 × 105 M−1 for NpVio2, and 8.1 × 104 M−1 and 1.1 × 105 M−1 for
AzoVio2 respectively. However, the second step in the isotherm overlaps with the first.
These also gave a cumulative Ka > 10
9, and therefore were unreliable.
Np2Vio, in contrast, can bind just 1 CB[8] to form a heteroternary complex, with
n = 1.0 as shown; again measurement of Ka could not be achieved accurately. Using a
one site hetero-association model, a value of Ka = 1.4 × 106 M−1 was obtained with a
poor fitting curve as shown, however this cannot be the case as the addition of ADA.HCl
(Ka = 5.5× 108 M−1) did not displace the complex (see section 3.2.6). With CB[7] there
was strong evidence of an aggregation process occurring. Initial introduction of Np2Vio
is exothermic due to the binding between viologen and CB[7], however as the number of
Np2Vio molecules exceeds that of CB[7] an endothermic process was observed. As pre-
99
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.27 ITC isotherms for cells containing CB[7] (61 µM) or CB[8] (51 µM) titrated
with each AB2 (1 mM) molecule in H2O.
viously discussed, the addition of CB[7] conformationally restricts the naphthyl groups
in close proximity to each other leading to intra- and possibly intermolecular stacking.
This endothermic process is likely due to further addition of Np2Vio disrupting the inter-
molecular stacking.
With CB[8] systems, it appeared that K1CB[8] > 10
9 M−1, therefore this must be re-
duced to be measured by ITC. Through introduction of a competitive weaker binding guest
with the CB[8] in the cell, the effective Ka of the intramolecular complex will be reduced
100
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
and allow direct measurement in a displacement titration.42 This has been employed in
studying CB[n] binding for high affinity guests previously for ferrocene and adamantane
derivatives.57,182 Calculation of the Ka and ∆H for high affinity ligands can be carried
out with the relationship shown in Eq. 3.2, where Kexp and ∆Hexp are the experimentally
observed binding constant and enthalpy, [Comp] is the concentration of the competitive
guest, and KComp and ∆HComp are the binding constant and enthalpy for the competitive
guest. Here, ADA.HCl was employed in the cell as a competitive guest, which resulted in
simple isotherms of a single step, shown in Fig. 3.28.42,57,182 This is because in the presence
of competing guest, the second weaker binary complex between viologen and CB[8] can
not be formed. The ∆H of ADA.HCl to CB[8] binding has not been reported thus far as
NMR titrations have been used to calculate its Ka, so this needed to be calculated.
58,176
First, the full thermodynamic data for the complexation of methyl viologen into CB[8]
was calculated from a separate titration, similar with reported literature values.1 Then,
ADA.HCl (1 mM) was titrated into a cell containing methyl viologen (1 mM) and CB[8]
(51 µM), and fitted with a hetero-association model to determine Kexp and ∆Hexp. After
using Eq. 3.2, the thermodynamic data could be calculated as in Table 3.4, with a similar
Ka to that previously reported.
176 In a similar fashion, AB2 (1 mM) was titrated into the
cell containing ADA.HCl (1 mM) and CB[8] (51 µM), with the calculated data shown in
Table 3.4.
Ka = Kexp × [Comp]×KComp
∆H = ∆Hexp + ∆HComp
(3.2)
Figure 3.28 ITC isotherms for cells containing CB[8] (51 µM) and ADA.HCl (1 mM)
titrated with each AB2 (1 mM) molecule in H2O.
101
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Molecule K1CB[8] ∆H1CB[8] T∆S1CB[8] K1CB[7] ∆H1CB[7] T∆S1CB[7]
NpVio2 1.0× 1013 -91.8 -17.5 1.5× 106 -14.2 22.2
AzoVio2 1.6× 1012 -75.9 -6.3 2.7× 106 -22.3 14.4
Np2Vio 6.7× 1011 -81.9 -14.4 2.1× 106 -64.7 -28.7
Methyl viologen 4.4× 106 -27.1 10.8 - - -
ADA.HCl 5.5× 108 -45.9 52.4 - - -
Table 3.4 A table summarising the binding constants for CB[7] or CB[8] to each AB2
molecule studied with Ka in units of M
−1, ∆H and T∆S in units of kJ mol−1.
The values of K1CB[8] obtained for NpVio2 and Np2Vio are 100-1000 times higher than
those reported for the heteroternary complex between methyl viologen and 2-naphthol
(6.8× 1010 M−2).1 AzoVio2 also shows 1000 times stronger binding than for the separate
guest molecules (1.6 − 3.8 × 109 M−2).1,19,64 This is a direct consequence of the binding
being cooperative and intramolecular. As the CB[8] approaches the molecule, it will
encapsulate both the viologen and naphthyl or azobenzene simultaneously, making the
binding event cooperative rather than stepwise. Disassociation then requires CB[8] to
slide along the stacked guest molecules, which is highly unfavoured, resulting in a strong
equilibrium to the CB[8] being complexed. This is supported by the NMR experiments,
where the CB[8] protons at the top and bottom of the molecules can be resolved from one
another, demonstrating how the CB[8] is kinetically locked in place. The higher binding
for NpVio2 over Np2Vio can be justified when considering the increased water solubility
of viologen over naphthyl making the complex more favourable.
Figure 3.29 Thermodynamic data for CB[8]-based host guest complexation determined
by ITC. Homoternary 2:1 complexes (blue), heteroternary 1:1:1 complexes (red) and binary
1:1 or quaternary 2:2 complexes (green) are shown. Adapted from references.31
Upon considering the enthalpic and entropic contributions to intramolecular compl-
exation, further thermodynamic details of binding can be determined. As shown in Fig.
3.29, these contributions to binding show distinct trends depending on the binding mech-
anism.31 The values of ∆H1CB[8] for the cooperative intramolecular complexes studied are
102
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
between -76 and -92 kJ mol−1, and T∆S is between -6 and -18 kJ mol−1. On compar-
ison with the literature data shown, the cooperative intramolecular complexes are more
enthalpically favoured than most intermolecular heteroternary complexes, but display sim-
ilar entropic contributions. The enthalpic contribution for removing the conformationally
restricted water from the CB[8] cavity has been calculated to be 66 kJ mol−1,54 suggesting
the remaining enthalpy is due to enhanced donor acceptor interactions of the intramolecu-
lar complex, and favourable interactions with the rest of the AB2 molecule. The entropic
contribution is similar for NpVio2 and Np2Vio as the conformational changes to encapsu-
late the naphthyl and viologen within CB[8] will also be similar. With azobenzene there is
less of an entropic penalty, likely a result of the more extended aromatic structure resulting
in less confinement within the cavity.
CB[7] is more water soluble than CB[8], and so the titration can simply be inverted
with CB[7] in the syringe at a high concentration and AB2 molecules in the cell at a low
concentration, potentially allowing the use of a stepwise binding model to fit the two step
binding isotherm. The results are shown in Fig. 3.30. However, as there was significant
overlap between the first and second CB[7] binding isotherms, deconvolution of these via
the stepwise binding model could not be achieved for NpVio2 and AzoVio2. If the first
binding event is modelled separately by a hetero-association model, then K1CB[7] can be
calculated as in Table 3.4, however, it remained partially overlapped with the second
binding event. Np2Vio also showed competing interactions as CB[7] was titrated into
the cell, due to inducing stacking interactions, leading to an unreliable result. As CB[7]
reaches and surpasses 1 equiv., a hetero-association model can be applied, however with
a poor fit. The values of K1CB[7] and ∆H1CB[7] obtained are in line with those reported
for methyl viologen and CB[7] (104−7 M−1) for NpVio2 and AzoVio2.1 K2CB[7] could not
be calculated as the second step in these isotherms could not be modelled, but displayed
similar heats of injection implying an equivalent, but perhaps reduced, binding constant.
103
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.30 ITC isotherms for cells containing AB2 (51 µM) titrated with CB[7] (0.69
mM) in H2O.
3.2.6 Advanced Function
In this section, the response of these cooperative intramolecular complexes to various
stimuli, and the formation of higher ordered structures was investigated.
Photo-Physical Properties of AzoVio2 Complexes
The synthesised AzoVio2 molecule was inherently photo-responsive due to the presence of
an azobenzene group. It was thought that this would allow photo-physical control over
the molecule and complexes formed. In previous works with CB[8], the isomerisation from
the thermodynamically favoured linear E isomer to the bent Z isomer has been utilised
to alter the complexes present.19,64,65 Tian et al showed that a viologen and azobenzene
heteroternary complex could be disrupted by UVA exposure, with the phenolic azoben-
zene guests used not binding with CB[8] in their Z -isomeric state (Ka < 10
3 M−1).19 If
an azobenzene has a pendant cation present (pyridinium, imidazolium, or ammonium),
however, the Z isomer will bind stronger to CB[8] than viologen due to favourable space
filling and electrostatic interactions with the carbonyl portals.64,65 With AzoVio2, the iso-
merisation was followed upon complexation with CB[7] or CB[8], and profound differences
were observed.
The molecule and subsequent complexes with CB[7] or CB[8] were exposed to UVA
light centred at 350 nm as in Chapter 2. The conformation of molecules in solution and
supramolecular complexes could be determined from UV-vis spectrophotometer measure-
ments in conjunction with 1H NMR experiments.
By UV-vis, the AzoVio2 at a low concentration (19.2 µM) displays fast isomerisation
104
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.31 UV-vis spectra demonstrating the photo-physical properties of AzoVio2,
AzoVio2 with 2 equiv. CB[8], and AzoVio2 with 2 equiv. CB[7]. The initial photo-
stationary state is shown, followed by UVA exposure until the spectrum remains un-
changed, then heated at 70 ◦ C until the spectrum remains unchanged.
kinetics. After 10 mins of UVA exposure, the molecule had reached its photo-stationary
state, evidenced by no more change in the UV-vis trace upon further exposure. As shown
in Fig. 3.31, in response to UVA light the peak corresponding to the E isomer at 333 nm
reduced in intensity (pi−pi∗), and the peak corresponding to the Z isomer (n−pi∗) at 423
nm was enhanced.19 The peak corresponding to viologen was also observed to blue shift
from 260 nm to 257 nm upon the conformational change. This was likely a result of the
disruption of aromatic donor-acceptor interactions between E -azobenzene and viologen,
as the Z isomer will overlap less and be more polar. Upon heating, the isomerisation can
be reversed to drive formation of predominately the E isomer, which will relax back to
the initial equilibrium state after several hours.
In the presence of two CB[8] molecules, i.e. AzoVio2-2CB[8], this isomerisation is
severely inhibited. Even after 1 h of UVA exposure, the isomerisation could not be driven
further than observed in Fig. 3.31. Here, the peak of the E isomer at 327 nm (blue-
shifted due to complexation and enhanced donor-acceptor effects) was observed to decrease
in intensity only a small amount, with little observable change for the peak hidden at
423 nm. The viologen peak at 263 nm (red-shifted from complexation) was completely
unchanged. This implied that the inclusion of the azobenzene group within the cavity
altered the photo-isomerisation equilibrium due the Z isomer not being able to form a
heteroternary complex. This was likely due to the high Ka and slow disassociation kinetics
(K1CB[8] = 1.6×1012 M−1), which results in a cooperative binding mechanism rather than
sequential viologen and azobenzene association. However, upon heating to recover the E
isomer a significant change was observed in the spectrum. The 327 nm peak was recovered
well, but the 263 nm peak had loss some intensity, and there appeared to be more intensity
from the peak at 423 nm. It was thought that a small population of the AzoVio2-2CB[8]
complex had altered its structure, likely due to any Z isomer potentially forming a binary
complex with CB[8], with the free viologen cationic charge at its portal to stabilise it.
NMR investigations below attempted to resolve these speculations.
In contrast, in the presence of two CB[7] molecules the isomerisation proceeded as for
uncomplexed AzoVio2. A large reduction of similar magnitude was observed in the peak
at 332 nm (blue shifted 1 nm upon complexation), and the appearance of a peak at 424 nm
occurred. Therefore, no inhibition of the isomerisation was observed as the azobenzene was
not included within the CB[7] cavity. The viologen peak was also of great interest as an
105
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
enhancement of signal was observed along with a large blue shift to 253 nm. In addition,
the peak shape altered significantly. This implied that after isomerisation had occurred,
some of the viologen was displaced from the CB[7] cavity, potentially by some binding of
the Z isomer of azobenzene to CB[7] as has been observed in the literature with weaker
binding constants (Ka = 10
3−4 M−1), or by the more polar Z isomer gaining stability
adjacent to CB[7].183 The process did not appear to be completely reversible, however,
not fully recovering the initial equilibrium state, likely due to the Z isomer binding to
CB[7].
From 1H NMR experiments it was possible to directly monitor the proportion of the
E to Z isomer by peak integration, showing saturation after 4 h of UVA exposure. The
kinetics of the isomerisation were slower in this case due to the much higher concentra-
tion (5 mM) used for NMR experiments than for UV-vis measurements. An overview
kinetic plot is shown in Fig. 3.32, highlighting the differences in photo-physical properties
obtained by complexation with different equivalents of either CB[7] or CB[8].
Figure 3.32 A plot showing the kinetics of the photo-isomerisation of AzoVio2 present in
different complexes of CB[7] and CB[8]. Error bars have been calculated by the standard
deviation of the integration of 3 different peaks relating to the E and Z isomers within
the same spectrum.
For AzoVio2 in the absence of any CB[n] the equilibrium state was observed to be
85.3 ± 0.5 % E -AzoVio2, which could be driven to 70.4 ± 0.4 % Z -AzoVio2 after 4 h
UVA exposure. Upon heating, the process could be reversed to give 98.3 ± 1.5 % E -
AzoVio2, which then relaxed to the initial equilibrium state after several hours exhibiting
full reversibility. A representative NMR stack is shown in Fig. 3.33.
With CB[8], the AzoVio2-1CB[8] and AzoVio2-2CB[8] complexes could be followed
by 1H NMR throughout their photo-isomerisation in a similar fashion (Supplementary
Information). Before exposure to UVA light in both cases, almost none of the Z isomer
was observable by NMR due to the heteroternary complex of viologen and azobenzene
moieties driving the formation of the E isomer (2.0 ± 1.9 % and 0.7 ± 0.9 % for 1 and
2 CB[8] complexes respectively). The Z isomer cannot form a heteroternary complex due
106
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.33 Stacked 1H NMR spectra of AzoVio2 following its photo-isomerisation. From
the bottom the equilibrium state is shown, then predominately the Z isomer after 4 h UVA
irradiation, then predominately the E isomer after heating. Peaks labelled red correspond
to the E isomer, and those in blue to the Z isomer.
to its bulky conformation, but has the potential to form a binary complex with CB[8] that
could show competitive binding with a UV stimulus in some cases.64,65 After exposure to
UVA light, new peaks were observed. By measuring their integration, the isomerisation
was observed to be very hindered as compared to the case with no CB[8], giving values
for the Z isomer to be 7.5 ± 4.7 % and 9.2 ± 2.4 % for the 1CB[8] and 2CB[8] complexes
respectively after 4 h of UVA irradiation. Interestingly, upon heating the AzoVio2-2CB[8]
complex for extended periods the isomerisation could not be driven back far to the E
isomer as for the case without CB[8] (7.0 ± 2.3 % Z ). This was likely due to different
interactions of the Z isomer with CB[8] preventing rearrangement.
With the AzoVio2 and CB[7] complexes, small peaks were observed before UVA ex-
posure that were due to the Z isomer present in the equilibrium state, similar to that of
the molecule without any CB[7] (Supplementary Information). For AzoVio2-1CB[7] and
AzoVio2-2CB[7], the Z isomer was present at 13.2 ± 3.8 % and 19.7 ± 2.1 % respectively.
These values are similar to that without CB[7], with an enhancement observed for the
complex with 2 CB[7] molecules. This suggests that after complexation with 2 CB[7] mo-
lecules, the Z configuration becomes more favoured either in its conformationally restricted
environment, or by inclusion in the cavity. Upon UVA irradation, the photo-stationary
state is reached within 4 h as for the uncomplexed molecule (63.5 ± 7.9 and 63.9 ± 9.0 %
Z respectively). With 1 CB[7], the equilibrium state is recovered, however with 2 CB[7]
only partial recovery is observed (57.3 ± 4.7 % Z ). Overall it was implied that the Z -
azobenzene in the presence of 2CB[7]s is more stable, either a result of inclusion within
CB[7], or by the more polar Z configuration being stabilised by adjacent CB[7] molecules.
107
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Higher Order Complexes
Figure 3.34 A schematic showing the potential for the NpVio2-2CB[8] complex to bind
with another guest molecule to form multiple heteroternary complexes.
Figure 3.35 Stacked 1H NMR spectra of NpVio2-2CB[8] in D2O (bottom spectrum),
followed by addition of 1 equiv. of 2-naphthol (top spectrum). The peak assignments in
red show proton environments outside of the CB[8] cavity or binding with 2-naphthol, and
those in blue are inside the cavity as an intramolecular heteroternary 1:1:1 complex.
Figure 3.36 The DOSY NMR spectrum for the NpVio2-2CB[8] complex with 2-naphthol
(left), and the optimised model carried out with MMFF94 molecular mechanics in
Avogadro v1.2. (right).
After formation of the complex between NpVio2 and 2 CB[8] molecules, it was expected
that the binary complex between viologen and CB[8] could undergo further binding to a
108
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
suitable guest molecule to selectively give another heteroternary complex in a self-sorting
manner, as shown in Fig. 3.34. 2-naphthol was selected as a suitable second guest as it is
similar to the already present heteroternary complex, with a reported binding constant of
Ka = 6.2×105 M−1 to the viologen-CB[8] complex.1 After introduction of solid 2-naphthol,
the expected complex was formed resulting in a colour change from orange to deep red. The
corresponding 1H NMR spectrum stack shown in Fig. 3.35 shows significant differences
to evidence complexation. The peaks labelled in blue corresponding to the viologen-
naphthyl intramolecular complex remain almost completely unchanged upon addition of
2-naphthol to NpVio2-2CB[8]. The peaks labelled in red corresponding to the binary
viologen-CB[8] complex display many changes as the 2-naphthol is encapsulated forming
an intermolecular heteroternary complex. The peaks corresponding to 2-naphthol (t - z )
are upfield shifted and overlap with the existing naphthyl peaks; this shows that they are
also encapsulated within CB[8] as a heteroternary complex. The viologen peak b showed
a significant downfield shift of 0.6 ppm, becoming more similar to the intramolecular peak
b. Overall the CB[8] was concluded to fully encapsulate the 2-naphthol, and would be
centred over protons c and d. Shifts were also observed for the CB[8] peaks, with p - r
shifting upfield for the CB[8] complexing with 2-naphthol, likely a result of the added pi
ring current within the cavity.
DOSY NMR (Fig. 3.36) showed a single species was present with D = 1.78 × 10−10
m2 s−1 and diameter = 2.76 nm, in line with the previously discussed DOSY results. A
model is also shown in Fig. 3.36, allowing visualisation of the complex.
Chemical Responsiveness
The chemical responsiveness of these complexes was also explored through simple introduc-
tion of a guest molecule with a high binding affinity for CB[7] and CB[8], 1-adamantylamine
hydrochloride (ADA.HCl). The binding constant has been reported to be 4.2× 1012 M−1
for CB[7] and measured to be 5.5 × 108 M−1 for CB[8].1,58 The 1H NMR was followed
upon titration of this chemical stimulus into a solution of the complexes.
For the NpVio2-2CB[8] and AzoVio2-2CB[8] complexes as shown in Fig. 3.37 1 CB[8]
can be displaced from the complex, being the weaker binary complex between viologen and
CB[8]. The intramolecular heteroternary complex displayed stronger binding as previously
discussed, where it could not be displaced after heating to 70 ◦C for 24 h in 3 times
excess ADA.HCl. The same behaviour was observed for Np2Vio where no change was
observed upon addition of ADA.HCl. This provided further evidence for the high binding
constant as discussed with ITC measurements, but also provides the opportunity to control
supramolecular topology in complex aqueous mixtures, similar to self-sorting behaviour.175
In contrast, the AB2-CB[7] complexes could be completely disassembled in response to
ADA.HCl addition as shown in Fig. 3.38. This is because of the much stronger binding of
ADA.HCl to CB[7] than that of the AB2 molecules, resulting in irreversible displacement.
The self-sorting nature of this system was investigated by inverting the addition of
components. A 1:1 complex of ADA.HCl with CB[8] was pre-assembled, to which different
AB2 guests were added. The stacked spectra for each separate experiment are shown in
109
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.37 Stacked 1H NMR spectra of the addition of excess ADA.HCl to (a) Np2Vio-
CB[8], (b) AzoVio2-2CB[8], and (c) NpVio2-2CB[8]. ADA.HCl could only displace the
binary complexes and not the intramolecular heteroternary complexes. Schematics of the
complexes present have been included for each step, and peaks corresponding to ADA.HCl
have been highlighted with green squares.
Fig. 3.39. In each case it was clear that the addition of AB2 guest molecule completely
displaced the ADA.HCl-CB[8] complex, however only 1 CB[8] is associated with each
guest. Further addition of guest results in a mixture of AB2-CB[8], AB2, and ADA.HCl.
Therefore, in the presence of competing guest, the mixture will self-sort to form exclusively
the intramolecular complexes, with no evidence of the remaining viologen binding with
CB[8]. This data supports the ITC measurements undertaken, where ADA.HCl was used
as a competitive guest in order to calculate the binding constant for the intramolecular
complexation.
110
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.38 Stacked 1H NMR spectra of the addition of excess ADA.HCl to (a) Np2Vio-
2CB[7], (b) AzoVio2-2CB[7], and (c) NpVio2-2CB[7]. In each case the initial AB2-2CB[7]
complex was displaced resulting in an ADA.HCl-CB[7] complex and the unbound AB2
molecule. Schematics of the complexes present have been included for each step, and
peaks corresponding to ADA.HCl have been highlighted with green squares.
111
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.39 Stacked 1H NMR spectra of the addition of Np2Vio (a), AzoVio2 (b), NpVio2
(c) to the binary complex of ADA.HCl with CB[8] (d) in D2O. Schematics of the complexes
present have been included for each step, and peaks corresponding to ADA.HCl have been
highlighted with green squares.
112
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
3.2.7 Rigid Architectures
As these initial attempts into the synthesis of hyperbranched supramolecular polymers
have highlighted some of the intricacies involved, furtherAB2 architectures can be designed
that would promote intermolecular association over intramolecular. Thus far, a compact
molecular structure has been focussed upon to account for the lack of rigidity in NpVio2,
AzoVio2 and Np2Vio. Rigid cores have been used for some supramolecular polymers based
on CB[8] in the literature, either with an aromatic or bicyclic core, or in combination with
rigid alkyne components.133,134 As has been previously discussed, these were avoided in
the molecular designs used in order to promote water solubility. Therefore, attempts were
made at utilising a water soluble core, instead of simple benzyl-based cores with poor
solubility.
2,6-diaminopurine is a derivative of guanine, one of the main four nucleobases found
in DNA and RNA. Guanine itself is highly water-soluble, but does not contain simple
asymmetric 2:1 reactivity for the formation of an AB2 guest molecule. 2,6-diaminopurine
replaces the carbonyl group with a primary amine group, revealing asymmetric reactivity
and easier functionalisation. The secondary amine present within the rings is the most
reactive site present and so can be reacted with separately to that of the primary amines.
After derivitisation of this group, the primary amines can be reacted with to lead to the
ideal AB2 molecule. A proposed reaction scheme is shown in Fig. 3.40.
Figure 3.40 A proposed reaction pathway to an AB2 molecule with a rigid and water-
soluble core. The first step has been carried out successfully. Hydrophobic guests could
be coupled by either amide coupling to give an amide, or by reductive amination to give
a secondary amine.
113
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
2,6-diaminopurine is only readily soluble in water, with examples of reactions in the
literature reportedly carried out in DMF. Upon replicating these procedures, it was found
to be sparingly soluble in 80 ◦ C DMF, so the reactions were carried out as mixed disper-
sions. Initial reactions starting at the 5-membered ring were attempted taking inspiration
from the literature.184–186 A simple nucleophilic substitution with bromoethanol in the
presence of K2CO3 base was carried out. This reaction did not run to completion in
reasonable times (> 3 days), and the resultant mixture of 2,6-diaminopurine and product
eluted similarly by both normal and reverse phase column chromatography making puri-
fication difficult. A hydrophobic protected alcohol (TBDMS) was used instead as this
should drastically alter the polarity of the product, and allow reactions to occur from the
amines in subsequent steps without potential competing reactions from the alcohol. This
reaction proceeded well with yields of 51 % after normal phase column chromatography.
Upon attempted further reactions with the naphthyl/azobenzene aldehydes by reduct-
ive amination, guest acyl chlorides by amide coupling, or standard amide coupling proced-
ures difficult mixtures were obtained with only trace amounts of product being present.
This synthetic route could be carried further forwards in the future by careful optimisation
of the reaction conditions used.
3.3 Conclusions
Initial experiments into constructing supramolecular hyperbranched polymers revealed
that a small, compact AB2 guest molecule with some flexibility did not form intermolecu-
lar species, instead favouring intramolecular self-complexation. This, however, did lead to
the discovery of unique supramolecular host-guest complexes being generated. This new
type of cooperative heteroternary complexation will provide a useful functional handle in
the future for highly favourable and kinetically locked CB[8] complexes, already applied
here to self-sorting mixtures and in generating higher order complexes. The understand-
ing gained throughout this work has made a significant contribution to controlling supra-
molecular architectures with high binding affinities, and will allow future work to take
great advantage over these developments.
Other stimuli such as temperature, pH, or light could be further investigated in ap-
plying these intramolecular complexes to intelligent molecular switches. CB[8] could then
be ‘trapped’ or act as a bulky stopper group through formation of an intramolecular com-
plex. After significant changes to the guest molecules or the complex, the CB[8] or guest
molecules will then be released on demand to interact with their environment. This would
also release the CB[8] stopper from the end of the molecule, allowing access to e.g. a
polymer chain. NMR has been shown to be an essential tool in their characterisation, but
simple colour changes can also be followed due to the formation of different donor-acceptor
complexes, meaning these materials could also act as self-regulatory sensors.
Future efforts into the formation of supramolecular hyperbranched polymers with
CB[n] would be to first focus on a rigid, water-soluble core that can inhibit self-complexation
occurring as initially explored in section 3.2.7. While the synthetic procedure may be
complex, this pathway will lead to an AB2 type guest architecture that cannot form in-
114
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
tramolecular complexes due to the rigid core preventing molecular folding in this way.
Ideally, however the direct attachment of guest molecules to CB[n] would form polymers
from a single water-soluble component. The attachment of functional handles to CB[n]
is slowly becoming more available, with highest yielding efforts by C-H activation with
oxidation chemistry.1,148 Most likely, the use of functionalised CB[7] will allow the gener-
ation of highly water-soluble monomers that could polymerise with exceedingly high Ka
in linear, cross-linking, and hyperbranching architectures. Attachment of guests could
be achieved with thiol-ene or azide-alkyne click chemistry between guest substrate and
functional CB[7]. This would result in significant changes in the solution and solid state
properties of the polymers, at very low monomer concentrations. Once synthesised, these
new polymers will be of great interest to investigate for applications in stimuli-responsive
viscosity modifiers, and in small molecule encapsulation.
3.4 Experimental
3.4.1 Instrumentation
All chemicals were purchased from Sigma Aldrich and used as received unless otherwise
specified. UV-vis spectrophotometry, 400 MHz NMR, UV reactor, FTIR spectroscopy
were carried out as in Chapter 2. 500 MHz 1H NMR spectra were recorded on an Bruker
Avance 500 TCI Cryoprobe spectrometer. ITC experiments were carried out on a Malvern
MicroCal iTC200 at 298.15 K with 26 injections of 1.5 µL each with typical concentrations
of 51 µM and 61 µM for CB[8] and CB[7] in the cell, and 1 mM for the guest molecules
in the syringe. Raw data was processed and integrated with NITPIC (v1.2.0), fitted in
Sedphat (v12.1b), and visualised through GUSSI (v1.1.0).180,187
3.4.2 2D NMR
An example set of 2D NMR experiments is shown in Fig. 3.41 for the NpVio2-1CB[8]
complex to demonstrate how the peaks in the 1H NMR spectrum were typically assigned.
Other 2D spectra are shown in the Supplementary Information. Each peak in the NpVio2-
1CB[8] complex 1H NMR spectrum was correlated to a 13C peak by the HSQC experiment,
also noting which HSQC peaks were positive (CH or CH3) or negative (CH2). The
13C
shifts for each peak were then compared to that of uncomplexed NpVio2 to provide the first
evidence for their assignment. The COSY spectrum could then be employed to correlate
adjacent protons with each other, allowing further assignment of most of the peaks. The
HMBC spectrum was employed last to assign any remaining peaks that had not yet been
resolved, and to check over all the previous assignments made by the other techniques.
Finally, DOSY NMR confirmed the presence of a single supramolecular species diffusing
at a constant value.
115
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 3.41 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the NpVio2-1CB[8] complex. For COSY, HSQC and HMBC positive peaks are
shown in blue and negative peaks are shown in red.
116
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
3.4.3 NpVio2 Synthesis
Figure 3.42 Scheme of NpVio2 synthesis.
1-methyl-[4,4’-bipyridin]-1-ium iodide; Me-Bipy
The synthesis was adapted from the literature.188 4,4’-bipyridine (24.7 g, 0.16 mol) was
dissolved in DCM (500 mL). Iodomethane (28.37 g, 0.20 mol) was added in one portion,
and the reaction was left for 16 h. The product precipitant was filtered and washed with
further DCM (100 mL) to give the product as a yellow solid (37.10 g, 79 %). FT-IR
νmax/cm
−1 3078w, 3020br, 1648s, 1639s, 1602, 1579w, 1547s, 1524, 1498, 1467w, 1416s.
δH (D2O, 400 MHz) 8.91 (2H, d, 6.54 Hz), 8.78 (2H, d, 6.17 Hz), 8.39 (2H, d, 6.43 Hz),
7.92 (2H, d, 6.20 Hz), 4.45 (3H, s) ppm. δC (D2O, 100 MHz) 153.5 (ArC), 149.9 (ArCH),
145.6 (ArCH), 142.6 (ArC), 125.8 (ArCH), 122.5 (ArCH), 47.8 (CH3) ppm.
2,2’-((naphthalen-2-ylmethyl)azanediyl)bis(ethan-1-ol)
2-naphthalenemethanol (1.75 g, 11.1 mmol) was dissolved in anhydrous DCM (35 mL)
under N2 and cooled to 0
◦C by an ice bath. Methanesulfonyl chloride (1.03 mL, 13.3
mmol) was added, followed by dropwise addition of N,N -diisopropylethylamine (DIPEA,
3.30 mL, 12.7 mmol). The ice bath was removed, and the reaction stirred for 4 h. After
dilution with DCM (10 mL), the mixture was washed with saturated NH4Cl solution
twice, dried with anhydrous MgSO4, and the solvent was removed under reduced pressure
to give the crude activated alcohol. Diethanolamine (1.42 g, 13.5 mmol) was dissolved in
anhydrous THF (25 mL) under N2. The crude activated alcohol was dissolved in anhydrous
THF (5 mL), and was added by syringe to the amine followed by dropwise addition of
DIPEA (3.40 mL, 13.1 mmol). The mixture was refluxed at 70 ◦C for 3 days. Solvent
was removed under reduced pressure and purification was carried out by SiO2 column
chromatography in a mixture of DCM:MeOH 95:5 to give the product as a brown oil (1.30
g, 48 %). TLC (SiO2) Rf = 0.26 DCM:MeOH 95:5. FT-IR νmax/cm
−1 3342br, 3053,
2948br, 2880br, 2819br, 1634w, 1602s, 1509s, 1449br. δH (d6-DMSO, 400 MHz) 7.84-7.90
117
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
(3H, m), 7.81 (1H, s), 7.54 (1H, dd, 8.42, 1.51 Hz), 7.45-7.52 (2H, m), 4.38 (2H, t, 5.43
Hz), 3.80 (2H, s), 3.49 (4H, q, 6.03 Hz), 2.60 (4H, t, 6.36 Hz) ppm. δC (d6-DMSO, 100
MHz) 137.7 (ArC), 132.9 (ArC), 132.3 (ArC), 127.5 (ArCH, 3 overlapped), 127.3 (ArCH),
126.7 (ArCH), 125.9 (ArCH), 125.5 (ArCH), 59.2 (CH2, 2 overlapped), 56.4 (CH2) ppm.
HRMS calculated for [M + H]+: C15H20NO2
+ 246.1491, found 246.1491.
2-bromo-N -(2-bromoethyl)-N -(naphthalen-2-ylmethyl)ethan-1-amine
2,2’-((naphthalen-2-ylmethyl)azanediyl)bis(ethan-1-ol) (911 mg, 3.71 mmol) was dissolved
in anhydrous DCM (37 mL) and tetrabromomethane (2.71 g, 8.18 mmol) was added, fol-
lowed by triphenylphosphine (2.14 g, 8.16 mmol). The reaction was stirred at ambient
conditions under N2 for 16 h. The reaction mixture was washed with aqueous sat. NaHCO3
twice and brine once. The organic phase was dried with anhydrous MgSO4 before removal
of solvent by rotary evaporation. Purification was carried out by SiO2 column chromato-
graphy in a mixture of petroleum ether 40-60 and ethyl acetate 6:1 to give the product
as a brown oil that crystallised upon standing (1.28 g, 93 %). TLC (SiO2) Rf = 0.71
Pet:EtOAc 6:1. δH (CDCl3, 400 MHz) 7.79-7.86 (3H, m), 7.75 (1H, s), 7.44-7.54 (3H,
m), 3.89 (2H, s), 3.38 (4H, t, 7.25 Hz), 3.03 (4H, t, 7.24 Hz) ppm. δC (CDCl3, 100 MHz)
136.5 (ArC), 133.4 (ArC), 133.2 (ArC), 128.4 (ArCH), 127.9 (ArCH), 127.3 (ArCH), 126.9
(ArCH), 126.3 (ArCH), 125.9 (ArCH), 59.2 (CH2), 56.4 (CH2), 30.1 (CH2) ppm. HRMS
calculated for [M + H]+: C15H18NBr2
+ 369.9806, found 369.9812.
1’,1”’-(((naphthalen-2-ylmethyl)azanediyl)bis(ethane-2,1-diyl))bis(1-methyl[4,4’-
bipyridine]-1,1’-diium) tetrachloride; NpVio2
1-methyl-[4,4’-bipyridin]-1-ium iodide (4.14 g, 13.89 mmol) was dissolved in anhydrous
DMF (40 mL) under N2 at 80
◦C. 2-bromo-N -(2-bromoethyl)-N -(naphthalen-2-ylmethyl)-
ethan-1-amine (1.28 g, 3.47 mmol) in anhydrous DMF (4 mL) was added in one portion,
and the mixture stirred for 3 days at 80 ◦C. The product precipitant was filtered and
washed with further DMF (20 mL) and ACN (40 mL) to give a brown amorphous solid
(1.65 g). This mixed iodide/bromide salt was dissolved in the minimum volume of water,
to which a saturated aqueous solution of ammonium hexafluorophosphate was added until
no further precipitation occurred. The product was separated by filtration and dried. The
product as a hexafluorophosphate salt was then dissolved in acetone, to which a saturated
acetone solution of tetrabutylammonium chloride was added until no further precipitation
occurred. The product was filtered off and dried to give a brown amorphous solid (1.05
g, 44 %). FT-IR νmax/cm
−1 3364br, 3107w, 3028br, 2853w, 1637s, 1559, 1508, 1442.
λmax(H2O)/nm 223 and 260 (/dm
3 mol−1 cm−1 69844 and 41325). δH (D2O, 500 MHz)
9.06 (4H, d, 6.74 Hz), 8.99 (4H, d, 6.90 Hz), 8.38 (4H, d, 6.85 Hz), 8.28 (4H, d, 6.86 Hz),
7.68 (1H, d, 8.11 Hz), 7.58 (1H, d, 8.38 Hz), 7.51 (1H, d, 8.11 Hz), 7.45 (1H, s), 7.29
(1H, t, 7.51 Hz), 7.23 (1H, t, 7.50 Hz), 7.19 (1H, dd, 8.39, 1.50 Hz), 4.91 (4H, t, 6.31
Hz), 4.54 (3H, s), 3.82 (2H, s), 3.58 (4H, t, 6.33 Hz) ppm. δC (D2O, 125 MHz) 149.2
(VioC), 148.7 (VioC), 146.2 (VioCH), 145.4 (VioCH), 136.3 (NpC), 132.8 (NpC), 131.8
(NpC), 128.1 (NpCH), 128.0 (NpCH, 2 overlapped), 127.5 (NpCH), 127.4 (NpCH), 126.7
118
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
(NpCH), 126.4 (VioCH), 126.2 (VioCH, NpCH overlapped), 59.6 (CH2), 58.0 (CH2), 54.1
(CH2), 48.3 (CH3) ppm. HRMS calculated for [M - Cl]
+: C37H39N5
35Cl3
+ 658.2266,
found 658.2236.
3.4.4 AzoVio2 Synthesis
Figure 3.43 Scheme of AzoVio2 synthesis.
(4-(phenyldiazenyl)phenyl)methanol
The synthesis was carried out according to literature procedures.189 The product was
obtained as an orange crystalline solid (4.38 g, 51 %). TLC (SiO2) Rf = 0.30 Pet:EtOAc
3:1. FT-IR νmax/cm
−1 3310br, 3059w, 2915w, 2863w, 1659w, 1584, 1527w, 1504w, 1487w,
1462w, 1442. δH (CDCl3, 400 MHz) 7.90-7.94 (4H, m), 7.45-7.55 (5H, m), 4.79 (2H, s)
ppm. δC (CDCl3, 100 MHz) 152.8 (ArC), 152.3 (ArC), 144.0 (ArC), 131.1 (ArCH), 129.2
(ArCH), 127.6 (ArCH), 123.2 (ArCH), 123.0 (ArCH), 65.0 (CH2) ppm.
2,2’-((4-(phenyldiazenyl)benzyl)azanediyl)bis(ethan-1-ol)
The synthesis was carried out similarly to that in section 3.4.3. (4-(phenyldiazenyl)phen-
yl)methanol (1.50 g, 7.08 mmol) was dissolved in anhydrous DCM (50 mL) under N2 and
cooled to 0 ◦C by an ice bath. Methanesulfonyl chloride (0.66 mL, 8.53 mmol) was added,
followed by dropwise addition of DIPEA (2.15 mL, 8.29 mmol). The ice bath was removed,
and the reaction stirred for 4 h. After dilution with DCM (10 mL), the mixture was washed
with saturated NH4Cl solution twice, dried with anhydrous MgSO4, and the solvent was
removed under reduced pressure to give the crude activated alcohol. Diethanolamine (897
mg, 8.53 mmol) was dissolved in anhydrous THF (12 mL) under N2. The crude activated
alcohol was dissolved in anhydrous THF (12 mL), and was added by syringe to the amine
followed by dropwise addition of DIPEA (2.20 mL, 8.48 mmol). The mixture was refluxed
at 70 ◦C for 3 days. Solvent was removed under reduced pressure and purification was
119
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
carried out by SiO2 column chromatography in a mixture of DCM:MeOH 95:5 to give
the product as an orange oil that crystallised upon standing (1.72 g, 81 %). TLC (SiO2)
Rf = 0.33 DCM:MeOH 95:5. FT-IR νmax/cm
−1 3293br, 3200br, 2948s, 2871br, 2836w,
1656, 1601s, 1585w, 1548w, 1523w, 1497w, 1474br, 1440s. δH (d6-DMSO, 400 MHz) 7.82-
7.90 (4H, m), 7.53-7.62 (5H, m), 4.39 (2H, m, 5.42 Hz), 3.75 (2H, s), 3.48 (4H, q, 6.00
Hz), 2.57 (4H, t, 6.32 Hz) ppm. δC (d6-DMSO, 100 MHz) 152.0 (ArC), 150.9 (ArC),
144.2 (ArC), 131.3 (ArCH), 129.4 (ArCH, 2 overlapped), 122.4 (ArCH), 122.3 (ArCH),
59.2 (CH2), 58.7 (CH2), 56.5 (CH2) ppm. HRMS calculated for [M + H]
+: C17H22N3O2
+
300.1712, found 300.1715.
2-bromo-N -(2-bromoethyl)-N -(4-(phenyldiazenyl)benzyl)ethan-1-amine
The synthesis was carried out similarly to that in section 3.4.3. 2,2’-((4-(phenyldiazen-
yl)benzyl)azanediyl)bis(ethan-1-ol) (600 mg, 2.00 mmol) was dissolved in anhydrous DCM
(20 mL) and tetrabromomethane (1.46 g, 4.40 mmol) was added, followed by triphen-
ylphosphine (1.16 g, 4.42 mmol). The reaction was stirred at ambient conditions under
N2 for 16 h. The reaction mixture was washed with aqueous sat. NaHCO3 twice and
brine once. The organic phase was dried with anhydrous MgSO4 before removal of solvent
by rotary evaporation. Purification was carried out by SiO2 column chromatography in a
mixture of petroleum ether 40-60 and ethyl acetate 6:1 to give the product as an orange
oil that crystallised upon standing (715 mg, 84 %). TLC (SiO2) Rf = 0.62 Pet:EtOAc 6:1.
FT-IR νmax/cm
−1 3048w, 2969, 2950, 2920, 2871w, 2805br, 2735w, 1602, 1584, 1497w,
1483w, 1461s, 1441s. δH (CDCl3, 400 MHz) 7.87-7.94 (4H, m), 7.45-7.55 (5H, m), 3.82 (2H,
s), 3.38 (4H, t, 7.25 Hz), 3.01 (4H, t, 7.25 Hz) ppm. δC (CDCl3, 100 MHz) 152.8 (ArC),
152.3 (ArC), 142.2 (ArC), 131.1 (ArCH), 129.4 (ArCH), 129.2 (ArCH), 123.1 (ArCH),
123.0 (ArCH), 58.8 (CH2), 56.5 (CH2), 30.2 (CH2) ppm. HRMS calculated for [M + H]
+:
C17H20N3Br2
+ 424.0024, found 424.0007.
1’,1”’-(((4-(phenyldiazenyl)benzyl)azanediyl)bis(ethane-2,1-diyl))bis(1-methyl-
[4,4’-bipyridine]-1,1’-diium) tetrachloride; AzoVio2
The synthesis was carried out similarly to that in section 3.4.3. 1-methyl-[4,4’-bipyridin]-
1-ium iodide (856 mg, 2.87 mmol) was dissolved in anhydrous DMF (8 mL) under N2 at
80 ◦C. 2-bromo-N -(2-bromoethyl)-N -(4-(phenyldiazenyl)benzyl)ethan-1-amine (305 mg,
717 µmol) in anhydrous DMF (1 mL) was added in one portion, and the mixture stirred
for 3 days at 80 ◦C. The product precipitant was filtered and washed with further DMF
(20 mL) and ACN (40 mL) to give an orange amorphous solid (424 mg). This mixed
iodide/bromide salt was dissolved in the minimum volume of water, to which a saturated
aqueous solution of ammonium hexafluorophosphate was added until no further precip-
itation occurred. The product was separated by filtration and dried. The product as
a hexafluorophosphate salt was then dissolved in acetone, to which a saturated acetone
solution of tetrabutylammonium chloride was added until no further precipitation oc-
curred. The product was filtered off and dried to give an orange amorphous solid (271
mg, 54 %). FT-IR νmax/cm
−1 3366br, 3109w, 3000br, 2853w, 1638s, 1560, 1508, 1445.
120
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
λmax(H2O)/nm 260 and 332 (/dm
3 mol−1 cm−1 43638 and 19880). δH (D2O, 500 MHz)
9.04 (4H, d, 6.93 Hz), 8.90 (4H, d, 6.76 Hz), 8.48 (4H, d, 6.88 Hz), 8.44 (4H, d, 6.89 Hz),
7.65-7.73 (5H, m), 7.51 (2H, d, 8.33 Hz), 7.22 (2H, d, 8.33 Hz), 4.92 (4H, t, 6.31 Hz),
4.26 (6H, s), 3.76 (2H, s), 3.56 (4H, t, 6.32 Hz) ppm. δC (D2O, 125 MHz) 151.3 (ArC),
150.6 (ArC), 149.5 (ArC), 149.0 (ArC), 146.2 (ArCH), 145.6 (ArCH), 142.6 (ArC), 132.4
(ArCH), 130.9 (ArCH), 129.9 (ArCH), 126.5 (ArCH, 2 overlapped), 122.7 (ArCH), 122.2
(ArCH), 59.7 (CH2), 57.5 (CH2), 54.3 (CH2), 48.2 (CH3) ppm. HRMS calculated for [M
- Cl]+: C39H41N7
35Cl3
+ 712.2484, found 712.2455.
3.4.5 Np2Vio Synthesis
Figure 3.44 Scheme of Np2Vio synthesis.
2-(bis(naphthalen-2-ylmethyl)amino)ethan-1-ol
2-naphthalenemethanol (1.52 g, 9.61 mmol) was dissolved in anhydrous DCM (20 mL)
under N2 and cooled to 0
◦C by an ice bath. Methanesulfonyl chloride (0.90 mL, 11.6
mmol) was added, followed by dropwise addition of DIPEA (2.87 mL, 11.1 mmol). The
ice bath was removed, and the reaction stirred for 4 h. After dilution with DCM (10
mL), the mixture was washed with saturated NH4Cl solution twice, dried with anhydrous
MgSO4, and the solvent was removed under reduced pressure to give the crude activated
alcohol. Ethanolamine (350 mg, 5.73 mmol) was dissolved in anhydrous THF (10 mL)
under N2. The crude activated alcohol was dissolved in anhydrous THF (5 mL), and was
added by syringe to the amine followed by dropwise addition of DIPEA (1.29 mL, 4.97
mmol). The mixture was refluxed at 70 ◦C for 3 days. Solvent was removed under reduced
pressure and purification was carried out by SiO2 column chromatography in a mixture
of Pet:EtoAc 2:1 to give the product as a light yellow fluffy solid (543 mg, 33 %). TLC
(SiO2) Rf = 0.30 Pet:EtOAc 2:1. FT-IR νmax/cm
−1 3446br, 3052, 2973w, 2951s, 2931,
2891, 2872, 2825br, 1634w, 1601s, 1507s, 1480, 1447. δH (d6-DMSO, 400 MHz) 7.85-7.91
(8H, m), 7.59 (2H, dd, 8.43 and 1.38 Hz), 7.44-7.52 (4H, m), 4.40 (1H, t, 5.38 Hz), 3.80
(4H, s), 3.54 (2H, q, 6.19 Hz), 2.58 (2H, t, 6.61 Hz) ppm. δC (d6-DMSO, 100 MHz)
137.4 (ArC), 133.0 (ArC), 132.3 (ArC), 127.8 (ArCH), 127.6 (ArCH, 2 overlapped), 127.2
(ArCH), 127.0 (ArCH), 126.1 (ArCH), 125.6 (ArCH), 59.3 (CH2), 58.5 (CH2), 55.5 (CH2)
ppm. HRMS calculated for [M + H]+: C24H24NO
+ 342.1852, found 342.1851.
121
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
2-bromo-N,N -bis(naphthalen-2-ylmethyl)ethan-1-amine
2-(bis(naphthalen-2-ylmethyl)amino)ethan-1-ol (249 mg, 729 µmol) was dissolved in an-
hydrous DCM (10 mL) and tetrabromomethane (267 mg, 805 µmol) was added, followed
by triphenylphosphine (211 mg, 804 µmol). The reaction was stirred at ambient condi-
tions under N2 for 16 h. The reaction mixture was washed with aqueous sat. NaHCO3
twice and brine once. The organic phase was dried with anhydrous MgSO4 before removal
of solvent by rotary evaporation. Purification was carried out by SiO2 column chromato-
graphy in a mixture of petroleum ether 40-60 and ethyl acetate 6:1 to give the product as
a light yellow oil that crystallised upon standing (242 mg, 82 %). TLC (SiO2) Rf = 0.66
Pet:EtOAc 6:1. FT-IR νmax/cm
−1 3052, 2939, 2800, 2727w, 1632w, 1600s, 1540w, 1506s,
1469w, 1453w, 1439. δH (CDCl3, 500 MHz) 7.83 (6H, d, 8.23 Hz), 7.79 (2H, s), 7.59
(2H, dd, 8.45 and 1.58 Hz), 7.43-7.50 (4H, m), 3.86 (4H, s), 3.40 (2H, t, 7.30 Hz), 2.99
(2H, t, 7.30 Hz) ppm. δC (CDCl3, 125 MHz) 136.8 (ArC), 133.5 (ArC), 133.0 (ArC),
128.2 (ArCH), 127.8 (ArCH, 2 overlapped), 127.5 (ArCH), 127.2 (ArCH), 126.2 (ArCH),
125.8 (ArCH), 59.0 (CH2), 55.6 (CH2), 30.4 (CH2) ppm. HRMS calculated for [M + H]
+:
C24H23NBr
+ 404.1008, found 404.1007.
1-(2-(bis(naphthalen-2-ylmethyl)amino)ethyl)-1’-methyl-[4,4’-bipyridine]-1,1’-diium
tetrachloride; Np2Vio
1-methyl-[4,4’-bipyridin]-1-ium iodide (375 mg, 1.26 mmol) was dissolved in anhydrous
ACN (18 mL) under N2 at 80
◦C. 2-bromo-N,N -bis(naphthalen-2-ylmethyl)ethan-1-amine
(250 mg, 618 µmol) was added in one portion, and the mixture refluxed for 4 days at 80
◦C. The product precipitant was filtered and washed with further ACN (40 mL) to give
an orange amorphous solid (265 mg). This mixed iodide/bromide salt was dissolved in
the minimum volume of 80 ◦C water, to which a saturated aqueous solution of ammonium
hexafluorophosphate was added until no further precipitation occurred. The product
was separated by filtration and dried. The product as a hexafluorophosphate salt was
then dissolved in acetone, to which a saturated acetone solution of tetrabutylammonium
chloride was added until no further precipitation occurred. The product was filtered off
and dried to give a light yellow amorphous solid (155 mg, 44 %). FT-IR νmax/cm
−1
3351br, 3225w, 3025br, 2832, 2806, 1638s, 1616w, 1598w, 1561, 1508s, 1480w, 1448s.
λmax(H2O)/nm 226 and 265 (/dm
3 mol−1 cm−1 113011 and 24309). δH (D2O, 500 MHz)
9.00 (2H, d, 6.78 Hz), 8.57 (2H, d, 6.80 Hz), 8.10 (2H, d, 6.78 Hz), 7.80 (2H, d, 6.73 Hz),
7.72 (4H, t, 7.11 Hz), 7.68 (2H, d, 8.40 Hz), 7.52 (2H, s), 7.37-7.44 (4H, m), 7.25 (2H,
dd, 8.40 and 1.26 Hz), 4.54 (3H, s), 4.46 (2H, t, 4.72 Hz), 3.70 (4H, s), 3.26 (2H, t, 4.69
Hz) ppm. δC (D2O, 125 MHz) 147.9 (ArC), 147.8 (ArC), 146.1 (ArCH), 144.9 (ArCH),
136.8 (ArC), 132.8 (ArC), 132.1 (ArC), 128.0 (ArCH), 127.9 (ArCH), 127.8 (ArCH), 127.5
(ArCH, 2 overlapped), 126.5 (ArCH), 126.1 (ArCH), 126.0 (ArCH), 124.9 (ArCH), 60.9
(CH2), 58.7 (CH2), 53.9 (CH2), 48.3 (CH3) ppm. HRMS calculated for [M - 2Cl]
2+:
C35H33N3
2+ 247.6337, found 247.6335.
122
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
3.4.6 Towards Rigid Asymmetric Guests
9-(2-((tert-butyldimethylsilyl)oxy)ethyl)-9H -purine-2,6-diamine
2,6-diaminopurine (574 mg, 3.82 mmol) and anhydrous K2CO3 (577 mg, 4.18 mmol)
were dispersed in anhydrous DMF (73 mL) under N2 and heated at 80
◦ C for 2 h.
2-(bromoethoxy)-tert-butyldimethylsilane (1.00 g, 4.19 mmol) was added by syringe and
the mixture was stirred at 80 ◦ C under N2 for 2 days. The reaction mixture was filtered,
and the solvent was removed in vacuo. Purification was carried out by SiO2 column chro-
matography in a mixture of DCM:MeOH 95:5 to give the product as a light yellow solid
(574 mg, 51 %). TLC (SiO2) Rf = 0.29 DCM:MeOH 95:5. FT-IR νmax/cm
−1 3408,
3367br, 3297br, 3126br, 3028w, 2955, 2931, 2887w, 2857, 1656s, 1598s, 1523, 1473, 1435s.
δH (d6-DMSO, 500 MHz) 7.60 (1H, s), 6.60 (2H, s), 5.73 (2H, s), 4.04 (2H, t, 5.28 Hz),
3.84 (2H, t, 5.11 Hz), 0.78 (9H, s), -0.13 (6H, s) ppm. δC (d6-DMSO, 125 MHz) 160.2
(ArC), 156.1 (ArC), 151.7 (ArC), 138.1 (ArCH), 113.2 (ArC), 60.9 (CH2), 44.9 (CH2),
25.7 (CH3), 17.8 (C), -5.7 (CH3) ppm. HRMS calculated for [M + H]
+: C13H25N6OSi
+
309.1854, found 309.1852.
3.5 Acknowledgements
Magdalena Olesin´ska is acknowledged for conducting the ITC measurements and providing
many useful discussions throughout this project. Dr. Guanglu Wu is also acknowledged
for many discussions regarding NMR and ITC experiments.
123
Chapter 4
Multiscale Self-Assembly of
Semi-Conducting and Plasmonic
Nanocrystals in Water
This work has been compiled into a manuscript for publication:
A. S. Groombridge,* Kamil Sokolowski,* Steven J. Barrow,* Felix Deschler, Annabel
Mikosch, Lissa F. L. Eyre, Hanyang Zhao, Junyang Huang, Jeremy J. Baumberg, Chris
Abell, Oren A. Scherman, 2017, in preparation.
4.1 Introduction
4.1.1 Nanocrystals
Confining metals and semiconductors to length scales on the order of hundreds to thou-
sands of atoms, the nanoscale, gives rise to fascinating properties not observed in the
bulk material. The simplest effect is the orders of magnitude increase in surface area of
the materials. For heterogeneous catalysis this is ideal, as the potential interaction of
reactive substrates with the catalytic surface will also increase orders of magnitude, thus
vastly speeding reaction times.32,190 Furthermore, due to the confined size and shape of
the ‘nanocrystals’ (NCs), less thermodynamically stable crystal facets become exposed on
the surface. These are typically more reactive than the most stable facets seen in bulk,
even allowing catalysis of reactions not at all observed in bulk. Other properties are a
result of entering the realm of quantum mechanical effects,191 and are typically material
dependent.
Metallic gold NCs have been in use for more than 2000 years.190 This so-called ‘soluble’
gold was used in ancient times as a dye due to its strong red colour, and was widely used
as a medicine in the Middle Ages, still seeing use today. By the mid-20th century the
synthesis and structure of colloidal gold was well-established, helped by the advent of
electron microscopes and atomic force microscopy. The first syntheses of simple spherical
AuNCs remain relevant today, and were performed by the reduction of a dissolved gold
124
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
salt resulting in its atomic aggregation in the presence of a stabilising capping ligand.
The archetypical examples of water-based synthesis are the Turkevitch and Frens citrate
method.192,193 Seed-growth methods and a catalogue of potential metal salts, reductants
and ligands have allowed the synthesis of a wide variety of different metallic NCs of any
shape or size desirable, both in water and organic solvents.32,190
Localised surface plasmon resonance is the predominant cause of the unique properties
of certain metallic NCs.190,191 Typical metals that result in plasmonic properties are highly
conductive in bulk with high electron mobilities, such as Ag, Au, Al or Cu. The collective
oscillation of free electrons in metals is commonly referred to as a ‘plasmon’. A plasmon
can be described as a cloud of electrons that is displaced from its equilibrium position
around a fixed lattice of positively charged ions. Plasmons can not exist in bulk metals,
but can exist on their surfaces where interaction with light becomes possible. On a metal
surface, the electric field of electromagnetic radiation (i.e. light) can excite a plasmon
resulting in a surface plasmon polariton that can propagate along the metal surface. This
phenomenon has been capitalised on for the surface plasmon resonance (SPR) technique
that is widely used in medical diagnostics.194 In SPR, a thin layer of metal is coupled to
a dielectric such as air or water, and irradiated with light corresponding to the energy
required to induce a resonant oscillation of the surface plasmon. The resultant SPR is
very sensitive to changes in its local environment. As the surface chemistry of the metal
is altered, e.g. through the adsorption of biomolecules, the SPR spectrum will change
allowing quantification of interaction processes.190,191,194
In metallic NCs, the size of the NC is much smaller than the incident photon wavelength.
Therefore, the plasmonic excitations cannot propagate and instead favour the formation
of a localised surface plasmon over the whole NC volume. As the entire electron cloud
of the NC is displaced upon the formation of the plasmon around the fixed NC lattice, a
dipole is formed. This has a restoring force that pulls the electron cloud back to the lattice
thus causing oscillation of the collective electron density. As a result of the well-defined
oscillation over the limited NC size, only incident photons with energies that precisely
couple to this localised surface plasmon resonance (LSPR) have any effect. A very strong
absorption of light of a specific wavelength is thus observed. This allows precise control
over the energy of the LSPR by simply varying the NC size, to great effect. Gold NCs
(AuNCs) are the archetypical example used to demonstrate these properties, as they have
been used since ancient times for their strong red colour, a result of absorption of light of
ca. λ = 525 nm corresponding to the LSPR of 20 nm AuNCs.190,191
Shape also plays an important role over LSPR properties. For perfectly spherical
plasmonic NCs, the LSPR will occur across the whole NC volume. If asymmetry is intro-
duced, for example by making nanorods or nanocubes, it is at the corners and edges that
the electromagnetic local field from the LSPR is at its greatest.191 The ability to ‘focus’
far field radiation (i.e. visible to NIR light) to sub-wavelength dimensions is unique to
plasmonic systems. This is of particular interest with regards to sensing applications such
as surface-enhanced Raman spectroscopy (SERS), where the strong local field enhance-
ment can result in an increase in Raman signal by a factor of 1010.191 Recently, this has
125
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
led to the detection of single molecules in the enhanced field between a AuNC and a Au
surface.195
NCs made from semi-conducting materials (quantum dots, QDs) are a relative recent
discovery when compared to plasmonic NCs. One of the first reported example of their syn-
thesis was in the 1980s by Ekimov et al where they were grown by phase decomposition
from a supersaturated glass-semiconductor solid.196 The glass acted as a dielectric me-
dium in which the QDs were dispersed, which is essential for them to retain their unique
properties. The QDs dispersed in solution that are widely used today were discovered
soon afterwards, with the first ligand-capped solution phase synthesis being described
in 1988 for CdSe.197 This synthesis draws a parallel to that established for metallic NCs.
Ligand-stabilised QDs may then exist as a means of electronic passivation.198 A calculated
structure of a ligand-capped CdSe QD is shown in Fig. 4.1 to illustrate the crystalline
nature of the core and the arrangement of ligands forming a dynamic external shell.32
Figure 4.1 The calculated structure of a 5 nm PbS QD stabilised by oleic acid. Adapted
from references.32
In contrast to metallic NCs that exhibit plasmonic properties, QDs instead demonstrate
well-defined quantum confinement effects on their electronic band structure, as shown in
Fig. 4.2.33,198 Here the band structure has finite energy levels, much like those in atoms or
molecules. The smallest QDs draw a parallel to large molecules in their molecular orbital
structures, with increasing size leading to the development of the valence and conduction
band structure eventually reaching that observed on the bulk scale. This results in a larger
band gap of the material than in the bulk, with exquisite control obtainable with altering
their size.
Since their discovery, QDs have seen an explosion of interest in varied applications. As
their band gap can be precisely controlled, their narrow fluorescence spectra can also be
controlled. This has given rise to applications in biological imaging,199 and in multiplexed
optical coding of biomolecules.200 The field of photovoltaic devices has also benefited
greatly from their use, seeing use in LEDs,32,201 and in solar cells for light harvesting.32,202
Due to their facile generation of photo-excited states with easily accessible light irradiation,
they have also seen wide use in photocatalysis.203
126
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.2 A schematic diagram showing the relation between bulk semiconductor band
structure to that of quantum dots. Eg is the band gap energy, and the cluster relates to
a QD. Adapted from references.33
4.1.2 Self-Assembly of Nanocrystals
The field of NC synthesis with a variety of different materials is now well-established,
and their use widespread across the sciences with many tales of commercial success in
diverse fields such as optoelectronics and catalysis.32,190 Of continuing significant interest
is the ability to controllably assemble these NCs into aggregated structures that retain or
enhance their individual properties. In this way, one can classify NCs as building blocks,
or ‘atom equivalents’, with the aim to construct NC-based ‘molecules’.198,204 This has
led to a field of research based upon the coupling of NCs together into higher ordered
structures by a variety of different methods.
Bare NCs without any capping ligands are very unstable, and will simply aggregate
and precipitate from solution to minimise their high surface energies resulting from free,
dangling bonds.32 The stabilising ligand provides an energy barrier to aggregation and
precipitation. Broadly, ligands can either be steric or electrostatic in nature, providing
an unfavourable repulsive interaction if NCs are in close proximity in a good solvent. As
shown in Fig. 4.3, sterically stabilised NCs have long chain ligands that provide a thick
entropic barrier to the aggregation of NCs. Electrostatically stabilised ligands are stable
due to the electrostatic repulsion of like charges from their electric double layer and are
modelled well by Derjaguin-Laudau-Verwey-Overbeck (DLVO) theory.32 Typically, NCs
will be synthesised with relatively dynamic ligands that are in constant exchange with
bulk excess ligands, e.g. citrate-capped AuNCs. This allows further ligand exchange to
occur if an excess of competing ligand is introduced, allowing conversion of the NC to a
desired functionality. With the case of citrate-AuNCs, this would be the introduction of
functionalised thiols, which form a strong covalent bond with the surface. Changing ligand
not only alters the external functionality of the NCs, e.g. to prevent protein adhesion,
but will also alter the electronic structure of the central NCs depending on how electron
donating or withdrawing the metal-ligand bond is.32
The strength of metal-ligand interactions can be further explored with regards to hard-
soft acid-base theory.205 In the theory that has been adapted to inorganic materials from
127
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.3 A schematic diagram showing NCs stabilised by (top) steric hindrance or
(bottom) electric double layer repulsion. Adapted from references.32
coordination chemistry, a ‘hard’ Lewis acid-base pair is based upon two electron processes
or electrostatic interactions, and a ‘soft’ pair is based upon one electron processes or
covalent interactions. These are typically strong, whereas interaction between a hard
Lewis base and a soft Lewis acid will give a weak interaction. A good example is the
comparison of ligand binding with AuNCs, which is a soft Lewis acid. The charged citrate
is a hard Lewis base, and so binds weak and dynamically via electrostatic attraction to
Au centres, whereas thiols are soft Lewis bases, and bind very strongly with covalent
chemistry.32,205
Homogeneous Self-Assemblies
The prominent examples of higher-ordered superstructures of NCs synthesised are ‘su-
perlattices’, colloidal crystals made entirely from individual NCs. A summary of all the
construction variables possible is shown in Fig. 4.4.34 These were first synthesised by al-
tering the ligands on CdSe QDs by solution exchange, and then using standard molecular
crystallisation techniques based on solvent affinity. The solvent was gradually altered from
a favourable mixture that solvated the ligands, to a less favourable solvent by evaporation
of the more volatile components, resulting in macroscale colloidal crystallisation.206 In this
method, the NCs can be thought of as ‘atoms’ being crystallised by van der Waals inter-
actions, and many variations thereof have since been reported expanding into mixed NC
superlattices.34 Interactions of neighbouring NCs can also be tuned by having insulating
or conductive ligands. This allows the properties of the superlattice to retain that of the
individual NCs, or to allow overlap of neighbouring electronic states. Limitations do exist
in this technique with the types of superlattices possible due to the all-attractive nature
of the inter-particle interactions occurring, resulting in mainly close-packed lattices until
NC shape and electrostatic charge is dramatically altered.207,208
Electrostatic colloidal crystals have been constructed allowing a departure from close-
packed structures, drawing an analogy to simple ionic salt crystals.208 In this case AgNCs
and AuNCs were designed with thiol-capped oppositely charged ligands bearing either a
128
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.4 An overview of NC composition and their self-assembly pathways to generate
superlattices (colloidal crystals). Adapted from references.34
carboxylic acid or a quaternary amine, resulting in spharelite structures. The assembly
required some different considerations compared to simple ions due to the large NC size
and multiple charges present. The charge screening length played an important role in
the resulting structures, with a small degree of charge screening promoting crystallisation
over simple flocculation.
The construction of rationally designed superlattices is still under intense investigation,
with current technologies relying on a combination of complementary DNA association
and shape control.209–212 The DNA-driven crystallisation, however, requires several days
of cooling a heated (i.e. denatured) mixture of DNA-functionalised NCs.211 This has been
exploited at great length with NCs for controllable assemblies, with the current state of
the art in superlattices to give complex clathrate type colloidal crystals through control
over DNA sequences, NCs composition, and NCs shapes and dimensions.212
DNA is conveniently attached to AuNC surfaces by standard coupling chemistry of
DNA strands to ligands or directly to Au surfaces with thiol derivitisation, allowing the
precise assembly of AuNCs on their highly specific hybridisation. The first examples were
reported in 1996, just one year after the development of superlattices.209,210 In both of
these pioneering works, oligonucleotides were coupled to AuNCs, which were then com-
bined with a linking DNA duplex. This allowed the controlled formation of dimers and tri-
mers of AuNCs with the expected interparticle distances,209 and the reversible formation of
AuNC aggregates by thermal denaturation.210 To date, DNA assembly techniques remain
the primary method for complex nanostructure assemblies, drawing analogies with DNA
origami, even allowing coupling of different NCs to DNA origami templates.34,204,212,213
129
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
As the surface coordination chemistry of QDs and NCs from other materials is well estab-
lished, DNA-based assemblies have been developed for almost any potential NC mater-
ial.34,204,214
Smaller aggregates of NCs are also of great interest for fundamental studies and in the
generation of advanced materials. The simplest modification to allow these assemblies is
the modification of NC ligands to be multivalent. This can be achieved with a multitopic
ligand, allowing cross-linking of NCs.215–217 Flexibility is an issue, where self-association is
a direct competitor to interparticle cross-links if the linker is too flexible, so only relatively
rigid systems have been exploited successfully thus far, primarily in making cross-linked
films. As discussed above, DNA is seeing prominent use for these types of applications.
Figure 4.5 The evolution of new electronic states upon QD aggregation. (a) A schematic
of two CdSe NCs (top) and their electronic states (bottom) separated by an energy barrier
of ∆E and interparticle distance ∆x. (b) Schematic showing the effect of ∆x on the
energetic states. Adapted from references.34
The controlled coupling of homogeneous NCs on a small scale is of great interest for
coupling their quantised energetic states. With metal NCs this is the coupling of their
LSPR states depending on interparticle distance, which increases the volume that the
LSPR can delocalise to giving significant red shifts in their absorbance.36,218 It further
expands the possibilities of focussing light to a sub-wavelength scale. At the precise
junction between particles at interparticle distances larger than those allowing quantum
tunnelling, this can lead to the creation of plasmonic ‘hot spots’ with exceedingly high
electric field strengths.36,219 With QDs this allows further delocalisation of their band
130
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
structure dependent on ligand and interparticle distances, tuning their energy levels as
shown in Fig. 4.5.34
Hybrid Self-Assemblies
The next step in the development of advanced materials is in coupling metallic and semi-
conducting NCs together, first achieved back in 1999 by DNA-based assembly.214 Since
then, hybrid superlattices and more complex DNA-based assemblies have been widely
reported.34,207,220
By combining QDs with plasmonic NCs, a host of non-additive properties have been
observed, as shown in Fig. 4.6 for superlattices.34,220 Of particular interest is hot-electron
transfer, whereby plasmonic NCs can allow generation of excited states in coupled QDs.220,221
As the Fermi level of metallic NCs typically sits between the valence and conduction band
of QDs, hot-electron transfer can be achieved by transfer of light-induced excited states.221
The effect is similar to that seen in dye-sensitised solar cells. Opposite pathways can also
be observed, where the excitation of QD electrons can then transfer to the metallic NC;
the control of these effects is achieved by the light used for excitation, and the energy levels
of the component NCs.222 This then has further ramifications for hybrid photocatalytic
systems through precise control of the manipulation of surface chemistries and electron
excitation. Truly, the coupling of the available diverse library of nanostructures will bring
in a new generation of advanced technologies.34
Figure 4.6 The variety of potential non-additive effects from the direct coupling metallic
and semi-conducting NCs in superlattices. (a,b) Control over luminescence properties,
either enhancing of quenching; (c,d) emergence of new magnetic properties of hybrid ma-
terials; (e,f) conductivity enhancements in mixed lattices; (g,h) enhancement in catalytic
activity or the ordered lattice. Adapted from references.34
On a scale of a few NCs, direct measurement of physical properties has also been
performed.34,220,221,223,224 A detailed study on QD exciton and plasmon interactions was
carried out on biotin/streptavidin-coupled CdSe/ZnS QDs and AuNCs by Cohen-Hoshen
131
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
and coworkers.224 This coupling allowed the precise determination of the AuNC plasmonic
effect on the QD absorption, separating it from changes in the emission. An enhancement
of the emission of QDs was observed when the excitation wavelength overlapped with the
LSPR band for the AuNC, and was in-line with the calculated enhancement dependent
on interparticle distance. Polarisation effects were also observed relative to the QD-AuNC
axis, as the LSPR shape will be dependent on the angle of laser excitation thus having
a direct effect on localised QD adducts. Further studies on the exciton-plasmon coupling
and properties that can affect it such as particle size and exciton-plasmon energy difference
are ongoing, with good agreements to quantum models.223
The key limitation to many of the described NC assembly pathways described for both
homogeneous and hybrid systems, is that their synthesis remains a non-trivial process.
Precisely ordered NC assemblies of almost any composition can be achieved by DNA-based
assembly as discussed thus far. DNA that can be coupled to surfaces is also becoming relat-
ively inexpensive.34 The issue lies in the assembly process being very time-consuming, and
the resultant structures not being very robust.211,212 Assembly pathways with multitopic
ligands between NCs can also require multi-step synthesis with slow solvent evaporation
or ligand exchange, at least overnight.34,217 More simple superlattices also require intens-
ive optimisation of conditions to gradually destabilise the NCs to promote van der Waals
packing, or balance electrostatic packing with these larger structures.34,206–208
Furthermore, being able to control the formation of solubilised aggregates versus bulk
precipitate aggregates/crystals is also highly desirable. These solubilised aggregates are on
the order of 10s to 100s of nanometres in diameter termed ‘supraparticles’, but assembly is
difficult to achieve without losing the properties of the component NCs or resulting in bulk
precipitation. Xia et al undertook a landmark study in the synthesis of QD supraparticles
from polydisperse NCs, and their coupling to plasmonic NCs in core-shell geometries.35
This was done by a self-limiting self-assembly process, whereby a polydisperse mixture
of NCs were synthesised and assembled simultaneously in-situ at 80 ◦C, with reaction
time-dependent NC and resultant supraparticle size. Representative supraparticles are
shown in Fig. 4.7. The pathway was described as first the nucleation and growth of some
NCs, followed by elongated aggregates of these NCs that later formed more isometric
clusters eventually resulting in compact and uniform supraparticles. For the formation of
hybrid systems the synthesis and assembly was similar, with pre-synthesised AuNCs being
introduced for stepwise growth of NCs on. Whilst the reported synthesis resulted in well
characterised supraparticles by a delicate balance of electrostatic and van der Waals forces,
the synthesis method was quite specific depending on the ligand and reduction pathways
used, and was also quite time-consuming. Furthermore, the electronic properties of the
supraparticles appeared to be hindered with no reported luminescence data.
More recently, studies have been carried out with different materials to follow the
supraparticle assembly mechanism, such as by confining the supraparticle assembly process
to microfluidic devices in order to get snapshots of the growth mechanism.225 Fu and
coworkers synthesised their Au supraparticles through reducing further Au precursor onto
seed AuNCs.35 This resulted in polydisperse NCs, giving a similar assembly pathway to
132
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
other reported works. The use of a microfluidic chip showed the growth mechanism to
be the formation of Au nanoplates, which rolled up to form a core upon which further
nanoarms nucleated. A key area for future development in supraparticle synthesis would
be to construct them from already synthesised NCs that are monodisperse, allowing access
to more desirable properties of the resultant structures.
Figure 4.7 (a,d) SEM, (b,e) TEM, and (c,f) size distribution of CdSe supraparticle
samples obtained at different reaction times: CdSe-20 (a-c) 20mins, and CdSe-30 (d-f)
120 mins. Adapted from references.35
Cucurbit[n]uril as a Molecular Junction
As previously discussed at length, CB[n] is a water-soluble, barrel-shaped, rigid macro-
cyclic host molecule that has been used for a host of applications.1 As it contains multiple
carbonyl portals at its top and bottom, it has seen recent emergence as a multitopic lig-
and in aggregation of water-soluble citrate-capped AuNCs, maintaining an interparticle
distance of 0.91 nm.36,195,218,226–228
First efforts were reported in 2008, investigating the formation of self-assembled mono-
layers of CB[n] on Au surfaces.228 Characterisation of the interaction between the Au
surface with CB[n] has proven challenging with primary evidence from FT-IR and XPS
experiments, but the suggested binding mechanism can be justified with regards to the
chelate effect of multiple electronegative carbonyls out-competing the weaker Au-citrate
interactions.32,227,228 An entropic release of H -bonded water molecules from the carbonyl
portals and AuNC surface may also play a significant role.227
This CB[n]-induced aggregation allowed access to the resultant coupled LSPR hotspots
between the particles at an ideal fixed distance of 0.91 nm, which have been capitalised on
for surface-enhanced Raman spectroscopy (SERS).36,195,218 Typically coupling NCs with
established techniques such as DNA methods can result in mixed interparticle lengths due
to some flexibility and thus complicate SERS results, but as CB[n] is rigid this distance
133
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
can be uniform across all the AuNC aggregate. This is exemplified in Fig. 4.8.36 As the
CB[n] is also Raman active, this allows facile internal referencing for in-situ diagnostics,
and in collecting reliable and comparable data.36,195 Another advantage for the use of
CB[n] for SERS is through utilising the versatile host-guest chemistry available. This
allows complete localisation of desired analytes to exactly the coupled LSPR hot-spots,
even allowing the SERS detection of single molecules.195
Figure 4.8 TEM images of the aggregated products in the diffusion-limited colloidal
growth regime, highlighting the uniform 0.9 nm interparticle distance. Adapted from
references.36
Quantification of analytes in biologically-relevant mixtures by SERS was also possible
due to the localisation of analytes to the hot spot.226 The aggregation process can be
monitored as it takes place over several minutes to hours, allowing direct studying by
coupling microfluidic droplets with dark field spectroscopy.218 This allowed access to early
stage aggregation of AuNCs within milliseconds of mixing, furthering understanding of
the mechanism for this process. Whilst the reported works have focussed on coupling
plasmonic AuNCs together, no work has yet looked at the interaction between QDs and
CB[n], and the potential formation of hybrid couplings.
In the following sections, investigations into the use of CB[7] as a rigid molecular ‘glue’
between semiconductor and metallic NCs were carried out across the length scales.
4.2 Results and Discussion
In the search for an efficient molecular junction between NCs, one should focus on rigid,
stable, and chemically inert systems. The water-soluble macrocycle CB[7] is ideal for this
application, as it is aprotic and unreactive, has a well-defined height of 0.91 nm, and does
134
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
not decompose until reaching high temperatures.1 In addition, the desired molecular junc-
tion should interact strongly with the NC surface, allowing strong inter-particle adhesion
and facile displacement of stabilising ligands. CB[7] contains a ring of 7 electronegative
carbonyl portals on its top and bottom, which should interact strongly with NC surfaces
that can be stabilised with electronegative ligands such as carboxylates. It will also take
up less surface area per surface-ligand interaction, typically a problem for charged lig-
ands due to repulsion, as the carbonyls are covalently held close together in the CB[7]
ring structure. Similar to the additive effects in H -bonding to form strong intermolecular
interactions, the additive chelate effect of 14 of these NC-CB[7]-NC surface interactions
should result in a very strong inter-particle bond.32,227
It was hypothesised that the interaction between CB[7] and QDs would be stronger
than the interaction with AuNCs reported thus far. This hypothesis was based upon the
hard-soft Lewis acid-base theory that has been frequently utilised in describing NC-ligand
interactions.32,205 In this theory, CB[7] can be considered a hard Lewis base ligand, due to
the presence of multiple strongly electronegative carbonyl portal groups that will interact
with the NC surface. AuNC surfaces are known to be soft Lewis acids, and so while the
interaction of CB[7] with AuNCs will be stronger than with citrate due to the chelate
effect, it is still not an optimal interaction.227 In contrast, semiconductor NC surfaces are
typically treated as hard Lewis acids, and so it was thought the assembly of QDs with
the Lewis basic CB[7] would be much stronger, especially when in combination with the
chelate effect.
As described in the following sections, it was shown that semiconductor NCs can be
assembled into porous fractal aggregates instantaneously in water under ambient condi-
tions with CB[7] acting as a molecular ‘glue’. Hybrid aggregates of QDs and metallic NCs
were also easily constructed. Fabrication across the nano-, micro- and bulk scales were
achieved with simple assembly pathways.
4.2.1 Bulk Aggregation of InP/ZnS QDs and AuNCs
Spherical citrate-stabilised AuNCs synthesised by the Turkevich method were used as
plasmonic NCs,192 and heavy metal-free InP/ZnS core-shell quantum dots (QDs) were
adopted as a model system because of their well-established optical properties.229 The QDs
were synthesised in a one pot approach in organic solvent with a long chain carboxylic
acid ligand to give a zinc-blende crystal structure adapting procedures from literature,
followed by an adapted ligand exchange and phase transfer reaction to the highly water
soluble 3-mercaptopropionate ligand (MPA).229 The QDs could then be stored as a free-
flowing powder and redispersed in water, with the core-shell structure making the QDs
air-stable. The NCs were monodisperse (Fig. 4.9 b-d) with hydrodynamic diameters (Dh)
of 13.0 ± 5.4 nm (AuNCs) and 3.6 ± 0.8 nm (QDs) obtained from the dynamic light
scattering (DLS) measurements (Fig. 4.9d). TEM measurements provided similar size
distributions of 12.1 ± 0.9 nm and 2.6 ± 0.5 nm (n = 100) respectively (Fig. 4.9c).
Both types of NCs were stable dispersions when kept in separate aqueous solutions,
as well as when mixed together in different ratios showing no precipitation or changes
135
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.9 (a) The crystal structure of CB[7]. (b) Schematic representation of the water-
soluble AuNCs and InP/ZnS QDs. (c) TEM micrographs of AuNCs and QDs. (d) DLS
measurements of AuNCs and QDs by volume percentage. (e) Overview of the sequential
assembly process of the bulk aggregated precipitant from aqueous dispersions of AuNCs
and QDs in different orders of addition. (f) Photographs under visible and UVA light
illumination of the bulk precipitant, showing colours dependent on their assembly process.
(g) STEM micrographs showing the nanoscale structure of the aggregates with AuNCs
easily resolvable from the QDs. (h) TEM micrographs showing the nanoscale structure of
the aggregates.
in their optical properties after several weeks. Solutions of absorbance λ470 = 0.50 and
λ520 = 1.86 for QDs and AuNCs, respectively, corresponded to a concentration of QDs of
2.3 × 1015 mL−1 and of AuNCs 5.7 × 1012 mL−1. As the first exciton absorption band
was centred at 470 nm (see Fig. 4.10), the size of the InP core could be inferred to be ca.
2.4 nm after comparison to literature studies due to the highly size-dependent electronic
properties of QDs.230 This is in agreement with the DLS and TEM characterisation shown
in Fig. 4.9, where the ZnS layer was ca. 0.2 - 1.0 nm thick (i.e. 1-2 layers of ZnS). By
then taking literature values for the molar absorption coefficient at 350 nm for 2.4 nm QDs
(350 = 4×105 L mol−1 cm−1), the concentration of QDs can be calculated as in Eq. 4.1 by
the Beer-Lambert Law.230,231 [QD] was the concentration of QDs, A350 was the absorption
at 350 nm (1.53), and l was the path length (1 cm). The absorption at 350 nm was
136
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
used as this was far from the absorption onset due to quantum confinement, and particle
number was calculated via the Avogadro number. The AuNC particle concentration was
determined following reported procedures by taking the absorbance below the LSPR band
at 450 nm (1.32) relating the interband absorption to determine the concentration of Au
atoms,232–234 and the size of particles.232
[QD] =
A350l
350
(4.1)
Figure 4.10 (a) The UV-vis spectra of a aqueous solutions of QDs, AuNCs and a 1:1
mixture of both. (b) The steady-state fluorescence spectrum of QDs with λex = 360 nm.
In a model experiment (Fig. 4.9e), upon addition of 1.0 mL of CB[7] aqueous solution
(0.87 mM, excess of ca. 100 CB[7] molecules to 1 QD) to 1 mL of QDs (A470 = 0.50), a
yellow precipitate is observed within a few seconds, which sediments after 5 minutes. If a
mixture of QDs and AuNCs is used where AuNCs are the minor component (1000 QDs
to 1 AuNC; A470 = 0.50; A520 = 1.86), a red precipitate is observed. Alternatively, if
CB[7] is first premixed with AuNCs under the same conditions for 1 min, followed by the
addition of QDs, a purple precipitate is observed due to the initial slower aggregation of
the AuNCs. Bulk aggregation and precipitation from solution was readily visible, and was
followed with photography in Fig. 4.11 under bright white light and UVA illumination.
In all cases the quantitative precipitation of all the present NCs was observed, and their
fluorescence activity was retained (with shifts in emission wavelength and intensity).
Monitoring of the aggregation process by photography suggested that if CB[7] is added
in the final step then AuNCs remain unaggregated (no red shift of the surface plasmon
resonance band at λ = 520 nm), allowing the distinct red colour to be preserved in the final
precipitate. Conversely, if CB[7] is added to AuNCs first a red shift in absorbance occurs
to λ = 650 nm to give an intense purple precipitate; this is characteristic of aggregation
of AuNCs by CB[7] through the formation of small strings and clusters.36,218
The STEM/TEM images in Fig. 4.9 g and h highlight the structural differences in the
precipitate formed depending on the assembly process. The red precipitate shows single
AuNCs dispersed randomly in a matrix of QDs aggregated with CB[7]. A meso- and
macro-porous fractal structure is obtained due to the isotropic nature of CB[7]-based ag-
137
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
F
ig
u
re
4
.1
1
P
h
otos
o
f
th
e
C
B
[7]
in
d
u
ced
a
ggregation
of
Q
D
s
an
d
A
u
N
C
s
in
b
righ
t
w
h
ite
ligh
t,
an
d
u
n
d
er
U
V
A
illu
m
in
ation
.
(a,b
)
Q
D
s
&
C
B
[7]:
from
left
to
righ
t,
a
n
in
itial
su
sp
en
sion
of
Q
D
s
(1
m
L
)
h
as
ex
cess
C
B
[7]
ad
d
ed
(0.5
m
L
)
an
d
th
e
aggregation
lead
in
g
to
p
recip
itation
w
as
follow
ed
in
b
o
th
w
h
ite
ligh
t
(a
)
an
d
U
V
A
lig
h
t
(b
).
(c,d
)
A
u
N
C
s,
Q
D
s
&
C
B
[7]:
an
in
itial
A
u
N
C
su
sp
en
sion
(0.5
m
L
)
is
m
ix
ed
w
ith
a
Q
D
su
sp
en
sion
(0.5
m
L
),
fo
llow
ed
b
y
ex
cess
C
B
[7
]
(0.5
m
L
).
(e,f)
A
u
N
C
s,
C
B
[7],
&
Q
D
s:
A
u
N
C
s
w
ere
p
rem
ix
ed
w
ith
C
B
[7],
follow
ed
b
y
ad
d
ition
of
Q
D
s.
138
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
gregation of NCs, with pores ca. 100 nm wide. The purple precipitate contains aggregated
strings and clusters of AuNCs trapped within the QD-CB[7] matrix. This set of simple
experiments demonstrates the interaction of these InP/ZnS QDs with CB[7] is preferable
to that of the AuNCs and is significantly faster, supporting the hypothesised high affinity
of CB[7] to QD interfaces.32,205 They also highlight the ability to directly couple QDs to
AuNCs with constant interparticle distances.
The high affinity of CB[7] to the QD surface was shown by 1H NMR titrations in Fig.
4.12. A solution of CB[7] was added stepwise to an NMR tube containing QDs (500 µL,
Abs470 = 0.50), with a spectrum taken after each addition. Upon addition of CB[7], the
broad peaks corresponding to the surface bound MPA ligands lose their intensity, and a
broad peak corresponding to surface bound CB[7] appears at 5.7 ppm. After addition of
ca. 30 CB[7] molecules per QD (60 nmol), a signal coalescence point was observed. It
was thought that at this point the QD interface was saturated with CB[7], and further
addition dispersed in solution.
Figure 4.12 1H NMR spectra in D2O of (a) QDs and of titration of CB[7] into QDs in
nmol quantities (b).
1H NMR experiments can further probe the exact composition of the precipitates to
determine information about the interaction between CB[7] and the NCs. A QD-CB[7]
precipitate was prepared from 0.5 mL of each in D2O as for the experiments in Fig. 4.11,
corresponding to 1.15 × 1015 QDs. The QD-CB[7] precipitate was washed with D2O 4
times to remove excess CB[7], salts, and ligands. The QDs of the precipitate could then
be dissolved in 10 µL of concentrated DCl (20 v/v% in D2O), and the ratio of solution
phase CB[7]:MPA could be calculated by comparing their integration to an added internal
standard of pyridine (1.0 µmol) as shown in Fig. 4.13. As the aggregation process is
quantitative, and the initial QD concentration is known, it was possible to calculate the
absolute ratio of QD:CB[7]:MPA in the precipitate to be 1:25:65.
This is reflected in experiments where the amount of CB[7] is substantially reduced.
If over 50 CB[7] molecules are introduced per QD, then precipitation proceeds instantly
as in Fig. 4.11. If 25-50 CB[7] molecules are introduced per QD, then a precipitant is
formed much slower, taking 24 h. However, if less than 25 CB[7] molecules are introduced,
then no precipitation was observed over week-long time scales, implying the formation of
water-stable aggregates, or ‘supraparticles’, was occurring.
139
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.13 1H NMR spectrum of the QD-CB[7] precipitant after dissolution of the QDs
with DCl, and addition of a pyridine reference (1 µmol). This allowed calculation of the
ratio of QD:CB[7]:MPA in the precipitant to be calculated as 1:25:65.
4.2.2 Supraparticles of InP/ZnS QDs and AuNCs
Figure 4.14 DLS kinetics data showing the two stage growth mechanism for CB[7]-
based aggregation of QDs (a,b) and AuNC-QDs (c,d) plotted on a normal and semi-log
scale. Fits have been calculated for the second stage of growth with an Ostwald ripening
mechanism.37
Bulk precipitates were achieved through the addition of excess quantities of CB[7]. How-
ever, the size of the resultant aggregates can be tuned towards the nanoscale to give
supraparticles if the concentration of CB[7] is controlled and minimised. In Fig. 4.14,
140
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
the aggregation process upon limiting quantities of CB[7] was followed by in-situ DLS
measurements for QD-CB[7] and AuNC-QD-CB[7] aggregation. The supraparticles size
could be controlled in the region of 6-50 nm and 15-35 nm respectively, with precipita-
tion occurring for larger sizes. Time-dependent DLS studies reveal there are two stages
in the assembly formation. Initially there is a very fast growth upon addition of CB[7].
For example, upon addition of 3.1 CB[7] molecules per QD (Abs470 = 0.50), Dh increases
rapidly to 10.0 ± 2.0 nm in the first 5 mins. This indicated very fast formation of small
clusters limited by the amount of reagent, likely in a range of configurations. The growth
then slows resulting in assemblies of 15.0 ± 3.0 nm after 24 h. This second stage of growth
can be fitted to an Ostwald ripening growth mechanism as reported for classical aggreg-
ation of NCs from ionic species, optimised by a least squares fitting method (details in
Supplementary Information, section A.2.1).37 The coarsening mechanism here is unclear
(i.e. migration of individual NCs, or migration of small clusters?), and warrants a further
in-depth physical investigation. This is shown schematically in Fig. 4.15.
Figure 4.15 A schematic diagram of the two stage mechanism of aggregation observed
for QD-CB[7] and AuNC-QD-CB[7].
In concert with control over size with increasing quantities of CB[7], the growth of the
supraparticles also accelerates as shown in Fig. 4.14. Systems with a higher concentration
of CB[7] reach the maximum Dh (50 nm and 35 nm for QD-CB[7] and AuNC-QD-CB[7],
respectively) for stable supraparticles, beyond which they lose stability and begin to pre-
cipitate from solution. This is consistent with the NMR experiments previously described
in Fig. 4.12. The lower stability for the AuNC-QD-CB[7] supraparticles could be a result
of several effects, such as being a complex mixture of different ligands, having a higher
overall mass, or the larger AuNCs acting as nucleation centres for accelerating aggreg-
ation. Systems containing below 27.9 or 7.2 CB[7] for QD-CB[7] and AuNC-QD-CB[7]
141
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
respectively, stay water-soluble for weeks.
The Ostwald ripening second stage of supraparticle growth suggests that larger aggreg-
ates grew at the expense of the smaller ones. Once the stability of the supraparticles is
lost upon sufficient amount of CB[7], then all of the material precipitates with no NCs left
in the supernatant. This suggests that the smaller aggregates take part in stabilising the
larger assemblies, screening interactions between them and acting as particulate surfact-
ants as has been observed for ionic supralattices.208 This is supported by the exponential
acceleration of Dh of the supraparticles with increasing amount of CB[7] (Fig. 4.14).
Figure 4.16 TEM micrographs of supraparticles made from QD-CB[7] (a,b,c) and AuNC-
QD-CB[7] (d,e,f) at low concentrations of CB[7] (a,b,d,e) and higher concentrations (c,f).
TEM images of different stages in this assembly process revealed morphological inform-
ation about the supraparticles (Fig. 4.16), by taking aliquots of the assemblies after 5 h of
incubation and drop casting them onto TEM grids. In the case of QD-CB[7] supraparticles
with the smallest amounts of CB[7] as in Fig. 4.16a (Dh = 11 nm by DLS), roughly spher-
ical clusters of QDs were observed, with diameters in agreement with DLS measurements.
The AuNC-QD-CB[7] supraparticles display a similar spherically-packed structure of QDs
that are glued to a single AuNC in Fig. 4.16d (Dh = 32 nm by DLS). The crystal lattices
of different QDs were resolvable in some of these clusters as shown in Fig. 4.16 b and e,
showing the QDs remained as distinct nanocrystals. Once the concentration of CB[7] is
raised (Fig. 4.16 c and f), more branched and elongated structures are present, reflecting
the fractal nature of the bulk precipitants previously observed. At higher concentrations,
this is followed by supraparticle aggregation to an extended network that precipitates as
in Fig. 4.9.
142
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
4.2.3 Microstructures of InP/ZnS QDs and AuNCs
In order for nanoparticulate composite materials to be capitalised on for solution-based
applications, access to alternative processable architectures is desirable. Micrometre-sized
particles are ideal due to their easy monitoring and removal as compared to nanoparticles
(e.g. by filtering or centrifugation), and have been widely used across the sciences as drug
delivery vehicles,235 versatile barcodes,236 and in complex syntheses.79,237 Microfluidic
emulsions containing droplets of water-in-oil present an ideal template for controlled self-
assembly as utilised in Chapter 2, wherein the chemical environment can be precisely con-
trolled. Emulsions have been widely used in the preparation of microparticles by various
processes, such as solvent removal, UV-initiated polymerisations, or by self-assembly.77,78
By confining the QD aggregation process to water droplets generated by flow-focussing
droplet microfluidics, composite microstructures were assembled by subsequent solvent
evaporation. The dried microstructures could then be studied in the solid phase, or redis-
persed into an all aqueous environment.
Control over the droplet chemical environment was achieved through through sur-
factant control as in Chapter 2. The three surfactants in use were based on the same
perfluoropolyether backbone being either a commercial inert triblock copolymer, a COOH-
terminated (RFCOOH), or a NH2-terminated surfactant (RFNH2). This allowed control of
the properties of the droplet interface and potential electrostatically-directed assembly,21
but also over the pH of the droplet. Only three aqueous assembly materials were used, be-
ing AuNCs, InP/ZnS QDs, and CB[7], resulting in 18 different combinations of materials
and surfactants in the following sets of experiments.
Optical Microscopy Investigations
A PDMS microfluidic chip was used as described in Chapter 2 for those systems without
CB[7] included in flow (i.e. a single inlet chip, Fig. 4.17a). When CB[7] was included a
multiple inlet chip was used instead shown in Fig. 4.17c, as this allowed the segregation of
the instantaneous CB[7] assembly process to the point at which droplets are made, thus
preventing clogging from solid precipitants.
After generation of the emulsion, it was output into a pool of perfluorocarbon FC-40
oil (50 µL) on a glass slide. As the aqueous phase is less dense, the droplets rise to the
top of the oil phase whereupon gradual evaporation of water can take place under ambient
conditions. As FC-40 has a high boiling point of 165 ◦C, it evaporates much later than
the water phase. This has many effects on the droplet environment, causing concentrating
of the materials encapsulated within. Different combinations of materials with RFCOOH
present were first studied by optical microscopy in the absence and presence of CB[7].
In Fig. 4.18a, a AuNC dispersion was emulsified. Upon significant evaporation of the
aqueous phase, aggregation leading to visible precipitation of the AuNCs was observed.
The concentrating effect of evaporation in combination with a reduction in overall pH
from the carboxylic acid surfactant (as discussed in section 4.2.3) destabilised the AuNCs
causing this precipitation. The dried precipitant did not maintain a coherent structure
upon dispersion in water. The dried aggregates also did not exhibit any LSPR, implying
143
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.17 (a) A single inlet flow-focussing droplet generation junction and (b) the
resultant emulsion. (c) A double inlet flow-focussing droplet generation junction and (e)
the resultant emulsion. Both emulsions were generated with overall flow rates 150/100 µL
h−1 of oil and water respectively.
Figure 4.18 Optical micrographs of aqueous droplets containing (a) AuNCs, (b) QDs,
or (c) AuNCs and QDs dispersed in FC-40 containing 2 wt.% neutral triblock and 1
wt.% RFCOOH. From left to right the evaporation of the emulsion is imaged, followed by
dispersal of the washed and dried structures into an all aqueous environment.
the metal surfaces were in contact with each other in their aggregation. Therefore, upon
acidic destabilisation the AuNCs likely aggregated and precipitated due to attractive inter-
particle van der Waals interactions.
In Fig. 4.18b the QDs assembled at the droplet interface and aggregated, causing
144
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
the characteristic buckling or wrinkling effect indicative of a phase transition to a solid
film at the interface (as discussed at length in Chapter 2).20 This was a result of the
carboxylate ligands on the QDs becoming protonated by RFCOOH at the interface and
likely causing the formation of a hydrogen-bonded network of ligands, and thus a solid
film. This was evidenced by the fluorescence still observed in this network, and supporting
bulk measurements carried out by altering pH shown in Fig. 4.19. Under acidic conditions
in bulk, a precipitant was formed that could be reversibly redispersed upon raising the
pH.
Figure 4.19 Photographs of solutions of QDs at pH7 and pH11. At pH 2 the QDs
precipitate due to their ligands becoming protonated. If base is added to this precipitant,
the QDs could be redispersed immediately.
AuNCs in contrast irreversibly aggregate and precipitate upon acidification, as the
citrate-AuNC bond is weaker and so upon protonation is displaced easily allowing van
der Waals driven aggregation to occur. This QD microstructure can be defined as a
colloidosome, an elastic shell of colloidal particles that can encapsulate material with the
same solvent within and outside.238 In this case the cross-linking force holding the particles
in place was likely the hydrogen-bonded network of ligands maintaining a small distance
between the QDs, allowing them to retain their nanoscale properties.
In Fig. 4.18c, both QDs and AuNCs are included within the droplets. The result is
additive as the two NCs and their ligand systems do not appear to interact. Therefore, first
aggregation of the AuNCs and their precipitation from solution was observed, followed by
the formation of the non-covalently cross-linked QD shell at the interface. Upon attempted
redispersion of the dried core-shell microstructure in water, however, the QDs completely
redispersed and the AuNC aggregates were free to diffuse. It was thought that the excess
amounts of free trisodium citrate within the dried structures could have resulted in a large
osmotic shock upon introduction of deionised water causing rupture. It could equally
be a result of these excess ligands penetrating the QDs’ MPA ligand shell, weakening the
bonding force between QDs similar to a charge screening effect, allowing simple dissolution
of the QDs into water with little energy barrier to prevent this.
Upon introduction of a flow of excess CB[7], the microstructures obtained changed
dramatically due to the various aggregation processes taking place inside the droplets.
In Fig. 4.20a, the introduction of CB[7] to AuNCs inside the droplet resulted in the
slow aggregation of AuNCs as previously reported in bulk solution.36,218 An unexpected
wrinkling of the droplet interface can be observed after some evaporation, however in this
case it was thought to be due to the adsorption of nanoscale AuNC-CB[7] aggregates to
the interface, similar to as observed with a Pickering emulsion. The colour has also shifted
145
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.20 Optical micrographs of aqueous droplets containing (a) AuNCs and CB[7],
(b) QDs and CB[7], or (c) AuNCs, QDs and CB[7] dispersed in FC-40 containing 2 wt.%
neutral triblock and 1 wt.% RFCOOH. From left to right the evaporation of the emulsion
is imaged, followed by dispersal of the washed and dried structures into an all aqueous
environment.
to dark blue, indicative of AuNC aggregation taking place.36 After further evaporation,
the wrinkled droplet features have disappeared, and the microstructure contracts to give
a dense dark blue microparticle, as opposed to staying as an elastic thin film of aggregates
at the interface. This further implies the wrinkling effect observed was transient and based
on weak non-covalent interactions. Upon rehydration however, these microparticles did
not appear to be very stable, with some expanding slightly as water permeates through,
and others dispersing back into water as aggregates.
In Fig. 4.20 b and c when QDs and CB[7] are present in the droplet, very similar
self-assembly pathways were observed, mimicking that observed in bulk experiments. The
aggregation of QDs with excess CB[7] was instantaneous, and was complete upon initial
observation of the droplets. When AuNCs are also included, these are simply trapped
as single NCs in the QD-CB[7] matrix, giving a red colour to the precipitate from their
individual LSPR. These aggregates appeared to be low density and highly porous in nature
in the optical micrographs, thought to be a result of their fractal aggregation observed in
bulk TEM measurements (Fig. 4.9g). This also supports a reagent-limited aggregation
process due to the stochastic aggregation, as each collision of QDs leads to a CB[7]-induced
cross-link.36 Further evaporation of the droplets only serves to confine these aggregates
and compress them into dense, smooth microparticles. Here, potential competing effects of
RFCOOH appear to be completely outcompeted by the rapid CB[7] aggregation of QDs.
Upon rehydration, a slight expansion and some cracking of the structures were observed.
146
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
The bulk aggregates have been shown to be porous, and the microparticles will contain
excess salts and ligands that were present in the NC suspensions or have been displaced
by CB[7]. This means that upon introduction of deionised water there will be a large
osmotic shock as these hydrophilic salts are dissolved, causing an outwards pressure from
inside the microparticles resulting in cracking. Further drying and rehydration steps do
not show this and are discussed in section 4.2.3.
Figure 4.21 Optical micrographs of aqueous droplets containing (a) AuNCs, (b) QDs, or
(c) AuNCs and QDs dispersed in FC-40 containing 2 wt.% neutral triblock surfactant and
1 wt.% RFNH2. From left to right the evaporation of the emulsion is imaged, followed by
dispersal of the washed and dried structures into an all aqueous environment.
Upon using RFNH2, entirely different microstructures can be observed. Images of the
systems without CB[7] present are shown in Fig. 4.21. With AuNCs in Fig. 4.21a, a
wrinkled shell can be observed to form only after almost all of the droplet has evaporated.
This implied a large concentrating effect was required before assembly at the interface was
favourable enough to lead to a solid phase transition. The reason for a phase transition
in this case was either a result of electrostatic attraction between the negatively-charged
citrate-coated AuNCs, or could alternatively be due to bonding between the amine of
the surfactant with the gold surface.239 The dried AuNC colloidosome structures were
stable to rehydration exhibiting a red colour, which implies there was still a reasonable
inter-particle distance (i.e. > 5 nm).
The QDs are unaffected by both the amine functional group, and by any potential in-
creases in pH of the droplet as shown in Fig. 4.21b. The QDs remain dispersed and form a
simple yellow-green solid bead of dried QDs after full evaporation. These beads completely
redissolve in water. However, when both AuNCs and QDs are included, while the QDs
were unchanged in their behaviour the aggregation and precipitation of the AuNCs was
147
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
observed similar to as in Fig. 4.18a. This could be a result of the AuNCs only forming
a shell in Fig. 4.21a when almost all the evaporation had occurred, meaning that due to
large excess quantity of the smaller QDs present they completely dwarf this effect. Then,
due to being in a highly concentrated environment, the larger AuNC simply aggregate
and precipitate. Upon addition of water the QDs redisperse, and the AuNC aggregates
gradually diffuse away.
Figure 4.22 Optical micrographs of aqueous droplets containing (a) AuNCs and CB[7],
(b) QDs and CB[7], or (c) AuNCs, QDs and CB[7] dispersed in FC-40 containing 2 wt.%
neutral triblock surfactant and 1 wt.% RFNH2. From left to right the evaporation of the
emulsion is imaged, followed by dispersal of the washed and dried structures into an all
aqueous environment.
Upon inclusion of CB[7] to droplets covered with RFNH2, the CB[7] induced aggreg-
ation process takes precedent over the final observed microstructures. For the AuNCs in
Fig. 4.22a, the presence of excess CB[7] has caused aggregation, leading to a dark blue
droplet. These aggregates assembled at the interface as in Fig. 4.21a. Once dried and
rehydrated, a water-stable shell can be observed, composed of AuNCs aggregated with
CB[7]. Whilst a phase transition to a solid film at the interface was not readily observed,
the dried structures show some evidence of being a collapsed thin shell rather than a
smooth microparticle, implying a shell was formed in the final stages of droplet evapora-
tion. This is in contrast to the AuNC-CB[7] system constructed with RFCOOH in Fig.
4.20a, where AuNC-CB[7] microparticles were not particularly stable in water.
Both droplets containing QDs, CB[7], and RFNH2 behaved similarly upon droplet
evaporation. The aggregation of QDs again takes precedent, being readily observable
with AuNCs due to their strong red colour in Fig. 4.22c. However, in contrast to using
RFCOOH, a preference of these aggregates to adhere to the interface was observed after
148
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
some droplet evaporation. The accumulation at the interface must be due to an attractive
interaction between RFNH2 and the aggregates. Since basic media do not affect these
aggregates in the bulk, and QDs have been shown not to interact with the amine group in
the absence of CB[7] (Fig. 4.21b), the aggregates must carry a partially negative charge.
Therefore, the attractive interaction was most likely electrostatic in nature. This gave
rise to quite large wrinkled droplets that contracted upon further evaporation to give non-
uniform dried aggregates. They were stable to being dispersed in water, similar to the
other QD-CB[7] systems.
Figure 4.23 Optical micrographs of aqueous droplets containing (a) AuNCs, (b) QDs,
or (c) AuNCs and QDs dispersed in FC-40 containing 2 wt.% neutral triblock surfactant.
From left to right the evaporation of the emulsion is imaged, followed by dispersal of the
washed and dried structures into an all aqueous environment.
A commercial neutral triblock surfactant can also be used for comparison of these self-
assembly processes. The commercial polymeric triblock was thought to contain two blocks
of perfluoropolyether, and one block of a polyethyleneglycol (PEG) derivative.77 In Fig.
4.23a, AuNCs were observed to form dark blue droplets upon evaporation, followed by a
wrinkled phase change at the interface. PEG is known to adsorb strongly to AuNCs,240
and with this surfactant the same process has been observed. The dark blue colour implies
aggregation has also occurred, which is likely a result of either displacement of citrate
ligands or increased concentration causing instability. In contrast, the QDs in Fig. 4.23b
showed no preference for the interface and formed microparticles that redissolved in water.
Upon combining AuNCs and QDs in Fig. 4.23c, the QDs did not assemble at the
droplet interface as before. The AuNCs are vastly outnumbered in the droplets by the
QDs, and so are shielded from their previous interaction with the droplet interface with
the droplets retaining their red colour from individual LSPR. Instead they begin to visibly
149
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
precipitate from solution after almost all the water has evaporated, followed by redispersion
in water.
Figure 4.24 Optical micrographs of aqueous droplets containing (a) AuNCs and CB[7],
(b) QDs and CB[7], or (c) AuNCs, QDs and CB[7] dispersed in FC-40 containing 2 wt.%
neutral triblock surfactant. From left to right the evaporation of the emulsion is imaged,
followed by dispersal of the washed and dried structures into an all aqueous environment.
Upon combining the AuNC with CB[7] with the neutral surfactant, the aggregation
process is very clearly observed with no preference for the droplet interface (Fig. 4.24a).
As the aggregation is clearly complete before droplet evaporation has occurred, further
evaporation simply confines these aggregates to a smaller volume, until a point at which
they cause the droplet to destabilise and the aggregate spreads across the glass surface.
Droplet destabilisation was likely due to the AuNC-CB[7] aggregates being too rigid to be
further compressed into a microparticle structure.
In Fig. 4.24 b and c, the aggregation of QDs and CB[7] was the main driving force for
the final microstructures observed, similar to as observed for RFCOOH in Fig. 4.20b and
c. Aggregation has occurred instantaneously in both cases with and without the AuNCs,
with further droplet evaporation serving to confine these aggregates and compress them
into a dense microparticle. This microparticle then demonstrates water stability as before.
Monitoring of pH and Control Over Assemblies
Accurate pH monitoring inside of emulsion droplets is not an easy task. Typically it will be
measured indirectly for example by introduction of a component that will be luminescent
in response to small changes in pH.241 To qualitatively monitor the internal pH of the
droplets used, the pH indicator methyl red was introduced as the aqueous phase as a
filtered saturated solution. Methyl red has a pKa of 4.34, and serves as an ideal indicator
150
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.25 Photographs of the pH indicator methyl red as aqueous solutions in neutral
and acidic water. Photographs are also shown of the emulsion generated from microfluidics
using either the neutral triblock surfactant (2 wt.%), or the neutral triblock and RFCOOH
(1 wt.%).
for mildly acidic conditions, changing from yellow to pink/red between pH 6.2 and 4.4.
When using RFCOOH an acidic emulsion was observed as seen in Fig. 4.25 where a
slightly pink/red emulsion was obtained. For comparison, emulsions prepared by using
only the neutral surfactant show a yellow colour. The oil layer can also be seen to take on
some of the colour of the indicator when using RFCOOH; this is due to it pulling the small
molecule indicator into the oil by electrostatic interactions as has recently been discussed
in the literature.88,242 Optical microscopy was attempted to follow these colour changes
throughout the evaporation process, but the small volumes imaged resulted in no visible
colour. This supports our proposed mechanism for the formation of AuNC aggregates
and QD colloidosomes based upon the droplets and their interfaces being acidic or basic
dependent on surfactants used.
Hydration of QD-CB[7] Microparticles
Redispersion of the QD-CB[7] microparticles in water was of great interest to show the
applicability of the microparticles to generate stable dispersions for solution-phase applic-
ations. It was shown in Fig. 4.26 that the first hydration altered the structure of the
microparticle by formation of several cracks on the particle, and an expansion in size. For
QDs and CB[7] the microparticles expanded from a size of 21.4 ± 0.6 µm to 24.7 ± 0.3 µm,
an increase of 16 %. For the composite particles with AuNC also included, they increased
from a size of 24.0 ± 0.6 µm to 30.5 ± 0.4 µm, an increase of 27 %. This expansion in
size and the formation of visible cracks was a result of an osmotic shock occurring to the
microparticles. The initial microparticles dried from the emulsion still contained excess
ligands from the NC suspensions (trisodium citrate or sodium 3-mercaptopropionate), in
addition to any salt from their synthesis. Most likely, upon immersion in water, these hy-
drophilic ligands and salts drive the water into the microparticle and are dissolved causing
this osmotic shock. This theory is supported by the larger expansion observed for the
composite microparticles with QDs, AuNCs, and CB[7] (27 %), as these would contain
significantly more excess ligands and salt from the AuNC suspension.
Upon evaporation of the water, compression of the particle is observed due to the loss of
water, pulling most of the cracks back together by meniscus forces. This resulted in smaller
151
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.26 Optical micrographs showing the repeated hydration and dehydration of the
dried microparticles. Water (100 µL) was added by micropipette and allowed to evaporate
under ambient conditions. The microparticles were composed of either (a) QDs and CB[7],
or (b) QDs, AuNCs, and CB[7]. They were generated with the FC-40 oil phase containing
2 wt.% neutral triblock and 1 wt.% RFCOOH.
microparticles of 18.8 ± 0.3 µm and 21.5 ± 0.4 µm for the QD-CB[7] and QD-AuNC-CB[7]
microparticles respectively. As there are no longer excess ligands and salts present, further
hydration and drying does not alter the structure, with only a small observed expansion
in the microparticle size of 6 % and 11% for each microparticle, respectively.
Imaging by SEM in Fig. 4.27 after repeated hydration and dehydration steps illustrates
the cracked structure that persists throughout the structures. However, it appears the
cracks do not penetrate deep into the microparticles, likely because after drying they have
mostly sealed in a self-healing manner.
Utilising back-scattered electron (BSE) imaging in Fig. 4.28 on the cracks also allows
the visualisation of the position of individual AuNCs in the matrix of QDs and CB[7]
for the composite structures. They were observed to be non-aggregated AuNCs with a
152
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.27 SEM images of the microparticles after five hydration and dehydration cycles.
(a) microparticles of QDs and CB[7] and (b) a close-up of the highlighted crack; (c) a
microparticle of QDs, AuNCs and CB[7] with a (d) close-up of the highlighted crack.
Figure 4.28 BSE images of the (a) QD-CB[7] and (b) QD-AuNC-CB[7] microparticles
with a beam power of 10 kV.
stochastic arrangement, which is in agreement with the deep red colour the microparticles
show. For comparison, the QD-CB[7] microparticle displays uniform BSE intensity in Fig.
4.28a.
Lastly it is quite clear from these SEM images that the pores of around 100 nm in
size observed from the bulk aggregates by TEM were not present. This is because the low
density QD-CB[7] aggregate observed in the droplet has been substantially compressed,
forming a more solid cross-linked network.
SEM and Energy-Dispersive X-Ray Analysis
Further SEM imaging was carried out for some of the microstructures synthesised, followed
by elemental mapping by Energy-Dispersive X-Ray (EDX) spectroscopy. A summary
figure of the SEM images is shown in Fig. 4.29. Due to the higher resolution obtainable
153
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
with the SEM as compared to optical microscopy, more in depth visualisation of the
dried microstructures was possible. With AuNC and RFCOOH, an aggregated AuNC
precipitate was observed, along with some salt crystals. For QDs and RFCOOH, a flat
and very thin shell of QDs was observed which has cracked likely in response to the high
vacuum. For the AuNCs and QDs with RFCOOH, an aggregate of AuNCs was observed
within the thin shell of QDs.
For CB[7] and RFCOOH, the three different combinations of materials give rise to very
similar looking microparticles. The main difference is the AuNC-CB[7] particles, which
appear to have a rougher surface topology, likely a result of the AuNCs being significantly
larger than the QDs.
For AuNCs with RFNH2 a very thin shell was formed, which also encapsulates a salt
crystal. The QDs and RFNH2 show a rough microparticle of unaggregated, precipitated
QDs. The AuNC-QD microparticle has lost some of its uniform shape, and as discussed
previously the QDs far outnumber the AuNCs and therefore do not form the thin shell
observed when only AuNCs are present.
For AuNC-CB[7] with RFNH2, an apparent solid microparticle was observed with con-
cave dimples, which from optical microscopy experiments was shown to be a thick shell of
AuNC-CB[7] aggregates. The QD-CB[7] and QD-AuNC-CB[7] microstructures formed are
very similar, with the QD-CB[7] aggregation driving the formation of the final structures.
As observed with the optical microscopy there was affinity of the QD-CB[7] aggregates
to the droplet interface, which was likely a result of electrostatic interactions. This has
resulted in a thick shell of QD-CB[7], which upon drying has significantly crumpled.
Figure 4.29 A summary table showing SEM images for various microstructures.
154
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.30 A summary table showing close-up SEM images of the surface of various
microstructures.
Close-up images of the dried microstructures revealed further topological information
such as thickness of colloidosome shells, and surface roughness. The QD and QD-AuNC
colloidosomes (with RFCOOH) exhibited a rough, crumpled surface composed of spher-
ical substructures ca. 50 nm in diameter. The colloidosome thickness was also demon-
strated to be below ca. 100 nm through observing the thickness of folds in the dried
shell, showing how this is a multilayer shell of thickness at least 30 QDs. The surface
of the CB[7]-aggregated microparticles with RFCOOH were much smoother, with surface
features ranging from ca. 100 - 250 nm in size. A rough surface was observed for the mi-
crostructures generated with RFNH2 without CB[7], and smoother surfaces with CB[7].
The QD-CB[7] and QD-AuNC-CB[7] microstructures formed with RFNH2 showed a high
surface area, having a very crumpled surface with folds c.a. 150 nm in width.
Elemental mapping was obtained for each of the imaged microstructures by EDX
spectroscopy. Most of the microstructures showed a uniform distribution of the elements
present as expected, so a few representative maps have been selected in Fig. 4.31. Fig.
4.31a shows an AuNC precipitant formed in the presence of RFCOOH. Elemental mapping
illustrates the formation of localised salt crystals that were likely NaCl. NaCl is formed in
the AuNC synthesis from the reduction of HAuCl4 with trisodium citrate. The distribution
of sodium was not localised to these crystals as it acted as a counter-ion to any free citrate.
Fig. 4.31b shows a very thin AuNC colloidosome, where the mapping of Au was difficult
due to the low amounts present, resulting in signal just above the level of background
noise. The distribution of Na and Cl support the formation of a NaCl crystal, with Na also
155
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.31 Selected SEM images and their corresponding EDX elemental maps of rel-
evant elements.
distributed across the structure associating to the citrate. This image also highlights how
the intensity of the Au signal is quite weak in comparison to other elements, exemplified
in the spectra shown in Fig. 4.32.
Fig. 4.31c shows the QD colloidosome encapsulating an aggregate of AuNCs, which
is reflected in the EDX mapping. The shell obtained was thicker than in Fig. 4.31b
implying a multilayer film. The mapping showed more intense signal at the edge of the
colloidosome, and over the AuNC aggregate; this is due to the shell collapsing on itself
from the spherical droplet shape upon evaporation. The distribution of Au was localised
to the visible aggregate structures within the colloidosome.
Fig. 4.31 d and e show microparticles of AuNC-QD-CB[7] with different surfact-
ants in their synthesis. The instantaneous CB[7]-induced aggregation of QDs results in a
stochastic distribution of all the elements present, only observing differences in topology.
EDX spectra were collected for each microstructure, shown in Fig. 4.32. The elemental
compositions obtained were mostly as expected: those containing AuNCs had Au, Na, Cl,
O and those containing QDs had In, P, Zn, S, O. A few unexpected elements were also
observed, for example a peak corresponding to Si at 1.74 eV in spectra (a, d, f) which
does not overlap with any other elements. This may be a result of some transfer of SiO2
from the glass in sample preparation, or an artefact from the instrument. In addition, a
156
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.32 Extracts of EDX spectra obtained for each microstructure imaged by SEM,
with assigned peaks corresponding to each element observed.
peak for Cl was observed in spectra for QDs and CB[7] (h and k), implying that the CB[7]
still contained trace amounts of HCl from its synthesis. The final point of interest from
these spectra was that the Au peak at 2.12 eV was typically quite weak, and overlapped
significantly with the peaks for P and S. This means quantification of the presence of Au
could not be achieved in cases where AuNC were known to be present making the EDX
spectra a good, but not definitive, source of extra evidence for the composition of the dried
microstructures.
Bulk Preparation of Microparticles
The microfluidic emulsion (1 mL) could also be collected into a vial containing a FC-
40 reservoir (250 µL), and the aqueous phase removed by evaporation in a 60 ◦C oven
overnight to give the QD-CB[7] microparticles as a suspension in oil as shown in Fig.
4.33a. Re-dispersion into water was carried out by allowing the oil to evaporate from
the vial in the oven, then washing away surfactant with HFE-7100 several times, and
pipetting 50 µL of water directly onto the microstructures. The microparticles retained
similar structures as for those previously generated in ideal evaporation conditions, and
retained their fluorescent properties (Fig. 4.33c).
157
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.33 (a) Optical micrographs of the QD-CB[7] microparticles generated as a
suspension in FC-40. (b) Optical micrographs and (c) fluorescence imaging of a dried and
washed microparticle before and after addition of water.
Fluorescence Microscopy
As the QDs were inherently fluorescent, they could be tracked within the droplets by
fluorescence microscopy. This proved to be very useful in following the self-assembly
processes taking place.
Figure 4.34 Fluorescence micrographs of a glass slide, and of the QD suspension on the
glass slide. Blue light (λ = 460 - 490 nm band-pass) was used for excitation, and a FITC
filter (λ = 515 - 550 nm) was used in detection.
First, as in Fig. 4.34, the ability to observe fluorescence with the excitation source
and filters available was confirmed by taking images of the fluorescence of a glass slide
before and after introduction of a drop of the QD suspension. The QDs showed λmax of
470 nm, which lies at the edge of the LED excitation source used. The AuNCs were not
fluorescent.
The self-assembly of microstructures containing the fluorescent QDs was followed for
droplets generated with RFCOOH. There was a short (2 s) time delay between the bright
field and fluorescence image. In Fig. 4.35 a and b containing QDs, the initial emulsion
droplets were observed to have a uniform fluorescence implying there was a considerable
amount of QDs present in the droplet that were not confined to the oil-water interface.
At a critical point in the evaporation, droplet wrinkling occurs as previously discussed,
and this is reflected in the fluorescence images where the QDs are now entirely localised
in this interfacial solid film. After drying and redispersal in water, the QD fluorescence
remained localised to the colloidosomes observed optically, implying that these remain as
158
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.35 Optical and fluorescence micrographs following the self-assembly process of
(a,b) QDs and (c,d) QDs and AuNCs within droplets upon drying, and their subsequent
rehydration. The FC-40 oil phase contained 2 wt.% neutral triblock and 1 wt.% RFCOOH.
stable structures. After inclusion of the AuNCs in Fig. 4.35 c and d the QDs behaved in
the same manner in their evaporation to dried structures. The pH-induced precipitation
of AuNCs were distinctly not fluorescent. Upon rehydration, the AuNC aggregates diffuse
into the water, and the QDs appear to redissolve leaving a small amount of residual QDs
adhered to the glass surface.
With CB[7] present (Fig. 4.36), the QD aggregation process was confirmed to be com-
plete and inclusive of all present QDs as soon as droplets were observed. This is shown
in Fig. 4.36 b and d where fluorescence of the QDs is confined to the aggregates within
the droplets. Further evaporation serves to confine the QD-CB[7] aggregate into a smaller
volume, until it reaches a point where it fills the entire droplet. At this point the droplet
loses its spherical shape and further evaporation compresses the QD-CB[7] aggregates into
dense microparticles. Addition of water showed the expansion of microparticles that has
159
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure 4.36 Optical and fluorescence micrographs following the self-assembly process of
(a,b) QDs and CB[7], and (c,d) AuNCs, QDs and CB[7] within droplets upon drying, and
their subsequent rehydration. The FC-40 oil phase contained 2 wt.% neutral triblock and
1 wt.% RFCOOH.
been previously discussed, with the fluorescent QDs remaining within the microparticle
structures, evidencing their stability. The same process was observed for droplets contain-
ing QDs and AuNCs.
This retention of QD fluorescence allows us to conclude that the CB[7]-induced aggreg-
ation and compression into microparticles has not significantly altered the QD activity.
4.2.4 Photoluminescence
The steady state photoluminescence (PL) of the QDs has been shown in Fig. 4.10. Pre-
liminary experiments into the effect of CB[7] addition to QDs and AuNC-QDs on the
PL and its quantum yield (PLQE) are shown in Fig. 4.37, with calculated PLQE values
shown in Table 4.1. The total volume was 300 µL, and the concentrations of AuNCs (10x
160
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
concentrated, 5.7× 1013 mL−1) and QDs (4x concentrated, 9.2× 1015 mL−1) were higher
than that used previously in order to increase the signal to noise ratio. For the addition of
CB[7] to QDs, a red shift was observed in the PL band from 539 nm to 554 nm, indicating
a lower energy excited state is present. The PLQE also increased by 1.6 times, which
is quite unusual for aggregated QD systems typically observing quenching rather than
enhancement. These preliminary results suggested that the excited state excitons were
delocalised across the aggregated supraparticles, thus resulting in a red shift, however the
enhancement of PL requires further physical investigations. When the QDs were mixed
with AuNCs, the PL was substantially quenched and PLQE was close to 0 %. Upon
addition of CB[7], a similar red shift in PL and an enhancement in PLQE was observed,
however direct evidence of energy transfer between the NCs requires further investigations.
Figure 4.37 The PL spectra for the QDs before and after addition of small amounts of
CB[7], and the same for the AuNC-QD mixed dispersions (left). A zoomed in spectrum
of just the AuNC-QD aggregates (right) is also shown.
QD : AuNC Ratio CB[7] molecules per QD PLQE /%
1 : 0 0.0 3.08
1 : 0 3.5 3.21
1 : 0 8.7 3.41
1 : 0 17.4 4.92
1 : 0.006 0.0 0.00
1 : 0.006 1.7 0.00
1 : 0.006 4.4 0.02
1 : 0.006 8.7 0.04
Table 4.1 A table showing the calculated PLQE for QD-CB[7] and AuNC-QD-CB[7]
aggregates.
4.3 Conclusions and Future Directions
In conclusion, the increase in scope of CB[n]s as a molecular junction between NCs has
been demonstrated across the length scales. This allowed the instantaneous and control-
lable aggregation of QDs in water, which up until now has been a time-consuming laborious
process. The resultant assemblies retained their nanoscale properties, such as LSPR and
fluorescence. Control over assembly was achieved at the nanoscale through generation of
161
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
water-stable supraparticles of controllable size; supraparticle research is currently in early
stages and these structures are an important contribution to generating unique materials.
Control on the macroscale was achieved either in bulk through generation of highly porous
solid frameworks of stochastic fractal arrangements, or by confinement in emulsions to gen-
erate compressed high density microparticles. QD colloidosomes were also synthesised for
the first time, and warrant further investigations with regards to their microencapsulation
properties.
There are multiple potential directions for this research to continue. Many of the
phenomena warrant more in-depth investigations, and a range of different NC materials
should be further investigated to confirm the universal applicability of the aggregation to
other semi-conducting hard Lewis acidic NCs. A full spectroscopic study of the aggreg-
ates at their different length scales is in order, to assess how coupling NCs with CB[7]
alters the electronic properties of the individual NCs and to investigate potential electron
transfer processes. Applications worth pursuing from this research mainly lie with their
optoelectronic properties, seeing application in photovoltaic devices, in photocatalysis,
and in sensing.
4.4 Experimental
4.4.1 Materials and Instrumentation
All reagents were purchased from Sigma Aldrich at the highest purity available and used
without further purification, unless otherwise specified.
500 MHz 1H NMR spectra were recorded on an Bruker Avance 500 TCI Cryoprobe
spectrometer. Chemical shifts are recorded in ppm in D2O with internal reference set to
the solvent peak at 4.79 ppm. UV-vis spectra were recorded as in Chapter 2. Dynamic
light scattering (DLS) was carried out on a Malvern Zetasizer Nano ZS90 instrument
fitted with a He-Ne laser (λ = 663 nm) at 25 ◦C. Dh was calculated according to the
Stokes-Einstein equation. Steady state fluorescence was recorded on a Varian Cary Eclipse
Spectrophotometer with a 1 cm path length quartz cuvette.
Transmission electron microscopy (TEM) was carried out on a FEI Philips Tecnai
20. Samples were prepared on holey carbon grids by pipetting 1 µL of desired aqueous
solution and allowing it to evaporate under ambient conditions (drop-casting). SEM was
carried out on a FEI magellan 400L FE-SEM, and on a TESCAN MIRA3 FEG-SEM.
EDX measurements were carried out with an Oxford Instruments Aztec Energy X-maxN
80 EDS system coupled with either SEM. Carbon coating was carried out on an EMITECH
950X from graphite.
4.4.2 Nanocrystal Preparation and Characterisation
Citrate-stabilised AuNCs were synthesised by the Turkevich method.192,233 A solution of
HAuCl4 · 3 H2O (12.5 mg, 0.032 mmol) in 45 mL water was heated to reflux for 30 min.
Under vigorous stirring a solution of trisodium citrate dihydrate (56 mg, 0.19 mmol) in
water (5 mL) was added quickly at 100 ◦C, and cooled to 80 ◦C for 2 h. Then the solution
162
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
was cooled to 0 ◦C in an ice bath. A clear red solution was obtained and stored at 4 ◦C for
further use. The AuNCs were characterised by UV-vis spectrophotometry (λmax = 520
nm, 5.7× 1012 NC per mL), and by TEM and DLS (see Fig. 4.9).
Water-soluble InP/ZnS core-shell NCs (QDs) were prepared by ligand exchange from
their organic solvent soluble counterparts and transferred to a water environment after
ligand exchange. Organic-solvent soluble InP/ZnS core-shell NCs coated with long chain
carboxylate ligands were prepared following a modified literature procedure.229 1-octa-
decene and 1-octylamine were degassed under vacuum for 12 h at temperature (100 ◦C
and 25 ◦C respectively) and purged with N2 for 1 h prior to use.
For the growth of the InP core, In(OAc)3 (486 mg, 1.6 mM), myristic acid (1404 mg,
616 mM) and 1-octadecene (20 mL) were loaded into a three-neck flask. The mixture was
heated up to 190 ◦C under N2 to dissolve the solid components and then the clear solution
was cooled down to 100 ◦C and kept under vacuum for 1 h under vigorous stirring to further
degass. The mixture was heated up to 175 ◦C under N2 and a mixture of P(TMS)3 (200
mg, 0.8 mM) and 1-octylamine (1240 mg, 9.6 mM) in 1-octadecene (6 mL, prepared in a
glove box) was injected into the hot reaction system. The growth of InP NCs was carried
out within 10 minutes, after which time the the reaction system was cooled down to 150
◦C.
For the subsequent growth of the ZnS shell, zinc stearate (0.1 M in 1-octadecene) and
sulfur (0.1 M in 1-octadecene) precursors (5 mL each, preheated to 150 ◦C) were added
to the reaction flask with the InP NCs, waiting for 10 min between each injection at 150
◦C. After that, the temperature was increased to 240 ◦C for 30 mins to allow the growth
of ZnS shell. Then the reaction was cooled down to 150 ◦C for a second injection of the
ZnS precursors as before (7 mL each) and again heated up to 240 ◦C for 30 mins. The
reaction was then cooled down to room temperature.
For purification, 30 mL of toluene was added to the cooled reaction mixture and the
NCs were precipitated by the addition of acetone and the product was centrifuged (10000
rpm, 10 min). The supernatant was discarded and the NCs were re-dispersed in a small
amount of toluene and again precipitated by the addition of acetone, centrifuged and
isolated. The procedure was repeated 4 times. The purified product was dried under
vacuum and kept at room temperature under N2 in the dark.
For ligand exchange, 200 mg of InP/ZnS core-shell NCs coated with long chain carboxy-
late ligands were dissolved in 80 mL of chloroform and added to a vigorously stirred
mixture of 3-MPA (10 mL) and saturated aqueous Na2CO3 (24 mL) in water (46 mL).
The prepared system was stirred vigorously for at least 1 h until the organic phase was
colourless and the aqueous phase became yellow and highly luminescent under UV-light
exposure. The aqueous phase was collected and the water-soluble crystals were purified
and isolated by a procedure of (repeated 4 times): precipitation (upon the addition of
acetone), centrifugation (10000 rpm, 10 min), and re-dispersion of the precipitant in water.
The collected solid product QDs were dried under vacuum and kept at room temperature
under N2 in the dark. The QDs were characterised by UV-vis spectrophotometry (λmax =
470 nm, 4.6× 1015 NC per mg), and by TEM and DLS (see Fig. 4.9).
163
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
4.4.3 Bulk Aggregates and Supraparticles
The bulk aggregates were formed by taking stock solutions of AuNCs and QDs (Abs520 =
1.86 and Abs470 = 0.50 respectively) in water, and adding varying amounts of aqueous
CB[7] solutions. For hybrid assemblies, AuNC solution was added to solid QDs to the
desired absorbency. In this way aggregates could be formed from QDs and CB[7], and
with AuNCs, QDs and CB[7].
Supraparticles were assembled by preparing solutions of QDs and/or AuNCs as above,
and adding CB[7] in µL additions.
4.4.4 Microfluidic Droplet Generation and Analysis
The experimental set-up was similar to that used in Chapter 2.
Optical microscopy was carried out using a Vision Research Phantom Miro EX-4 fast
camera with colour interpolation, mounted to an Olympus IX-71 inverted microscope
(10x - 64x objectives). Fluorescence micrographs were obtained under illumination from
a coolLED pE-300white (blue waveband, 450 mW) lamp and imaged with an Olympus
IX-81 inverted microscope (Prior proscan II automated stage) mounted with an Andor
iXonEM+ DU 897 EMCCD camera, controlled via a PC running custom LabVIEW 2013
software.
Microfluidic devices were fabricated as in Chapter 2. The diameter of the single-inlet
junction (Fig. 4.17a) was 60 µm with a channel depth of 50 µm. For the double inlet
chip (Fig. 4.17b), the junction was 200 µm with a depth of 80 µm. Flow rates of the
oil and aqueous phases were typically 150 and 100 µL h−1, respectively for the single
inlet chip, and 150/50/50 µL h−1 for the double inlet chip. The double inlet chip was
used to segregate CB[7] from the particle suspensions until immediately prior to droplet
formation.
The continuous phase comprised of FC-40, with 2 wt.% neutral triblock copolymer flu-
orosurfactant (XL171, Sphere Fluidics) and 1 wt.% of either Krytox 157-FSL (RFCOOH,
a carboxylic acid-terminated perfluoropolyether, Dupont) or synthesised amine derivative
RFNH2 (see Chapter 2 for synthesis). The dispersed phase consisted of an aqueous solu-
tion of various combinations of InP/ZnS QDs, AuNCs, and 0.67 mM CB[7]. Once formed,
droplets were output onto the surface of a glass slide into an FC-40 oil reservoir (20 µL)
for observation. Evaporation of the aqueous phase was carried out under ambient condi-
tions to form dried microstructures as a suspension in FC-40. After further evaporation of
the oil to dryness, residual surfactant was removed by thorough washing with low boiling
point solvent HFE7100 (3M).
Samples were prepared for SEM imaging by transferring dried and washed microstruc-
tures prepared on a glass slide by stamping with sticky conductive carbon tape and then
affixing to an SEM stub. Samples were then coated with a thin film of conductive carbon.
164
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
4.4.5 Photoluminescence
Solutions were prepared in a 1 mm path length cuvette of total volume 300 µL. Cuvettes
were placed in an integrating sphere and photo-excited using a 405 nm continuous-wave
laser. The laser and the emission spectra were measured and quantified using a calib-
rated Andor iDus DU490A InGaAs detector. Photoluminescence quantum efficiency was
calculated as per literature procedures.243
4.5 Acknowledgements
Dr. Kamil Sokolowski is acknowledged for adapting the synthesis of water-soluble InP/ZnS
QDs from literature procedures and carrying out all TEM/STEM imaging. Dr. Steven J.
Barrow is acknowledged for synthesising the AuNCs, and helped with some of the SEM and
EDX measurements. Research on the bulk precipitants and supraparticles were conducted
in collaboration with Dr. Kamil Sokolowski. PL measurements and PLQE calculations
were carried out by Annabel Mikosch, Lissa F. L. Eyre and Dr. Felix Deschler in the
Cavendish Laboratory. Fitting of the DLS kinetics data with an Ostwald ripening model
was carried out by Junyang Huang.
165
Chapter 5
Conclusions
The future of sustainable, macroscale materials will lie with a ‘bottom-up’ self-assembly
approach.244 After clever design of building blocks, complex functional materials will spon-
taneously construct themselves under the right conditions. This thesis has made various
contributions in the application of CB[n] host-guest chemistry to generate new materials
by controlling their self-assembly. Unique to CB[n] systems is the ease with which this
complex functionality can be incorporated into materials in water. Aqueous media is typ-
ically difficult to control self-assembly processes in due to extensive H -bonding and its
high polarity; for example, built-in response to external stimuli can be included with ease.
Interesting properties of self-assembled materials constructed by supramolecular chemistry
have also been observed with the CB[n] systems studied in this thesis, such as making ma-
terials that can self-heal, or be self-diagnostic. It was concluded that there remains much
potential to be explored in this exciting area of science in following studies. A highly
interdisciplinary and collaborative approach to research was employed, which enabled the
projects undertaken to broaden their scope beyond the realms of traditional chemistry.
The field of supramolecular polymers and their networks is beginning to reach maturity,
with various successful start-up companies (e.g. Aqdot Ltd. or SupraPolix B.V.) devel-
oping industrially relevant formulations for a wide variety of applications, however those
based upon small molecules and host-guest chemistry are rare. In Chapter 2, branching
supramolecular polymers were designed and synthesised from CB[8] and multiple guest-
functionalised monomers with a branching point. It was concluded that limitations exist in
driving the supramolecular polymerisation to high conversions, which would lead to inter-
chain cross-linking and gelation, instead observing a loss of solubility of the polymers due
to intra-chain cyclisation and poor solubility. However, by templating the self-assembly
process within water-in-oil emulsions the polymerisation could be confined to the water-oil
interface and thus be driven to high conversions, resulting in a macroscopic cross-linked
network. The interfacial gel was shown to be self-healing and elastic as expected for a non-
covalent supramolecular network. This research showed the potential for CB[n] host-guest
chemistry to allow macroscopic synthesis of cross-linked networks from small molecules,
however it also highlighted many of the challenges involved in replicating this in bulk
media. Future research can focus on controlling the molecular architecture of branching
guest monomers to enhance water solubility, and prevent intra-chain cyclisation. This will
166
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
result in a range of self-assembled soft materials that will be highly responsive to external
stimuli, and exhibit facile processing.
Molecular self-assembly with a single multi-guest monomer and CB[n] was further
investigated in Chapter 3 to attempt the synthesis of hyperbranched supramolecular poly-
mers, a class of dendritic globular polymers that do not form inter-chain cross-links. Inter-
molecular complexation was not observed however, with exclusively intramolecular com-
plexes being formed. This was due to the molecular architecture used being a molecular
‘hinge’ for intramolecular complexation, making it a highly cooperative process as yet
unobserved for CB[n]s. As this intramolecular complex displayed much stronger thermo-
dynamic binding behaviour and slower disassociation kinetics than normally expected, it
could then be applied in the formation of multicomponent self-sorting complex mixtures
and in controlling the interactions of the ensembles with external stimuli. These dynamic
assemblies have potential future applications in self-regulatory diagnostic materials, of
particular interest with CB[n] chemistry as significant optical changes could be designed
in response to changes occurring in the complex. Future applications could also lie in
the use of this intramolecular complex as an end stopper for macromolecules, forming
pseudorotaxane structures and producing dynamic molecular machines.
Control in the self-assembly of inorganic NCs has also been achieved in Chapter 4
through the use of the CB[7] macrocycle. Assembly of NCs into higher-ordered aggregates
and even bulk materials has been highly desirable, but has been a complex or time-
consuming process so far, highly specific to the materials in use. The assembly process
discovered and described in this thesis allowed instantaneous assembly of QDs from water
into highly porous aggregates across the length scales and was proposed to be universal.
The aggregates retained and even showed enhancement of their nanoscale properties such
as photoluminescence, and could be combined with metallic NCs to form hybrid composite
materials. The scope of this assembly process for future research is very broad, with fur-
ther experiments into fully characterising their electronic and optical properties currently
underway. Furthermore, only one type of QD and metallic NC have been investigated thus
far, and it is suggested that the assembly process is universal across semi-conducting NC
materials. Applications of these assemblies also warrant further research, such as into the
potential for energy transfer between different types of NCs, and in developing photocata-
lytic substrates taking advantage of their extremely high surface area and porous nature.
Some investigations have been shown so far into making these materials into processable
and recyclable microparticles, which should allow their easy application to solution-based
applications.
The use of simple droplet microfluidics for studying self-assembly processes, and in the
generation of new materials has been a key focal point of this thesis. Control over the
droplet environment has enabled modulation of the degree of supramolecular polymerisa-
tion, and thus over molecular self-assembly processes, with results rapidly visible with a
microscope. The number of droplets typically imaged in an experiment will number from
the hundreds to thousands, allowing reproducibility of processes to be investigated with
ease. Furthermore, confinement of hybrid organic-inorganic assemblies within droplets has
167
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
allowed the generation of unique processable materials, but also furthered the understand-
ing of their self-assembly process. It was concluded that as PDMS microfluidic chips are
now widely accessible to researchers, it is envisaged that their use will spread across the
sciences to study various phenomena in these highly controllable environments.
168
Bibliography
[1] S. J. Barrow, S. Kasera, M. J. Rowland, J. Del Barrio and O. A. Scherman, Chem.
Rev., 2015, 115, 12320–12406.
[2] M. J. Rowland, M. Atgie, D. Hoogland and O. A. Scherman, Biomacromolecules,
2015, 16, 2436–2443.
[3] J. Zhang, R. J. Coulston, S. T. Jones, J. Geng, O. A. Scherman and C. Abell, Science
(80-. )., 2012, 335, 690–694.
[4] K. Holmberg, B. Jonsson, B. Kronberg and B. Londman, Surfactants and Polymers
in Aqueous Solution, John Wiley & Sons, Chichester, UK, 2nd edn, 2003.
[5] L. Y. Chu, A. S. Utada, R. K. Shah, J. W. Kim and D. A. Weitz, Angew. Chemie -
Int. Ed., 2007, 46, 8970–8974.
[6] G. M. Whitesides, Nature, 2006, 442, 368–73.
[7] F. K. Balagadde´, L. You, C. L. Hansen, F. H. Arnold and S. R. Quake, Science (80-.
)., 2005, 309, 137–140.
[8] P. Zhu and L. Wang, Lab Chip, 2016, 17, 34–75.
[9] C. Holtze, A. C. Rowat, J. J. Agresti, J. B. Hutchison, F. E. Angile`, C. H. J. Schmitz,
S. Ko¨ster, H. Duan, K. J. Humphry, R. A. Scanga, J. S. Johnson, D. Pisignano and
D. A. Weitz, Lab Chip, 2008, 8, 1632–1639.
[10] O. Wagner, J. Thiele, M. Weinhart, L. Mazutis, D. A. Weitz, W. T. S. Huck and
R. Haag, Lab Chip, 2016, 16, 65–9.
[11] C. Gao and D. Yan, Prog. Polym. Sci., 2004, 29, 183–275.
[12] F. De Greef, M. Smulders, M. Wolffs, A. Schenning, R. Sijbesma and E. Meijers,
Chem. Rev., 2009, 109, 5687–5754.
[13] T. F. A. D. Greef and E. W. Meijer, Nature, 2008, 453, 171–173.
[14] L. Yang, X. Tan, Z. Wang and X. Zhang, Chem. Rev., 2015, 115, 7196–7239.
[15] D.-S. Guo, K. Chen, H.-Q. Zhang and Y. Liu, Chem. - An Asian J., 2009, 4, 436–445.
[16] R. Dong, Y. Liu, Y. Zhou, D. Yan and X. Zhu, Polym. Chem., 2011, 2, 2771–2774.
169
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
[17] Y. Liu, Y. Yu, J. Gao, Z. Wang and X. Zhang, Angew. Chemie Int. Ed., 2010, 122,
6726–6729.
[18] S. Deroo, U. Rauwald, C. V. Robinson and O. A. Scherman, Chem. Commun., 2009,
644–646.
[19] F. Tian, D. Jiao, F. Biedermann and O. A. Scherman, Nat. Commun., 2012, 3, 1207.
[20] A. R. Salmon, R. M. Parker, A. S. Groombridge, A. Maestro, R. J. Coulston, J. Hege-
mann, J. Kierfield, P. Cicuta, O. A. Scherman and C. Abell, Langmuir, 2016, 32,
10987–10994.
[21] R. M. Parker, J. Zhang, Y. Zheng, R. J. Coulston, C. A. Smith, A. R. Salmon, Z. Yu,
O. A. Scherman and C. Abell, Adv. Funct. Mater., 2015, 25, 4091–4100.
[22] J. Hegemann and J. Kierfeld, J. Colloid Interface Sci., 2017, submitted.
[23] Y. Bai, X.-d. Fan, W. Tian, T.-t. Liu, H. Yao, Z. Yang, H.-T. Zhang and W.-B.
Zhang, Polym. Chem., 2015, 6, 732–737.
[24] F. Huang and H. W. Gibson, J. Am. Chem. Soc., 2004, 126, 14738–14739.
[25] X. Wang, H. Deng, J. Li, K. Zheng, X. Jia and C. Li, Macromol. Rapid Commun.,
2013, 34, 1856–1862.
[26] J. W. Lee, K. Kim, S. Choi, H. Ko, S. Sakamoto, K. Tamaguchi and K. Kim, Chem.
Commun., 2002, 0, 2692–2693.
[27] C. Gao, S. Silvi, X. Ma, H. Tian, A. Credi and M. Venturi, Chem. - A Eur. J., 2012,
18, 16911–16921.
[28] S. Y. Kim, Y. H. Ko, J. W. Lee, S. Sakamoto, K. Yamaguchi and K. Kim, Chem. -
An Asian J., 2007, 2, 747–754.
[29] W. Jiang, Q. Wang, I. Linder, F. Klautzsch and C. A. Schalley, Chem. - A Eur. J.,
2011, 17, 2344–2348.
[30] L. Briquet, D. P. Vercauteren, E. A. Perpe`te and D. Jacquemin, Chem. Phys. Lett.,
2006, 417, 190–195.
[31] G. Wu, M. Olesin´ska, Y. Wu, D. Matak-Vinkovic and O. A. Scherman, J. Am.
Chem. Soc., 2017, 139, 3202–3208.
[32] M. A. Boles, D. Ling, T. Hyeon and D. V. Talapin, Nat. Mater., 2016, 15, 364–364.
[33] M. G. Bawendi, M. L. Steigerwald and L. E. Brus, Annu. Rev. Phys. Chem., 1990,
41, 477–496.
[34] M. A. Boles, M. Engel and D. V. Talapin, Chem. Rev., 2016, 116, 11220–11289.
170
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
[35] Y. Xia, T. D. Nguyen, M. Yang, B. Lee, A. Santos, P. Podsiadlo, Z. Tang, S. C.
Glotzer and N. A. Kotov, Nat. Nanotechnol., 2012, 7, 479–479.
[36] R. W. Taylor, T. C. Lee, O. A. Scherman, R. Esteban, J. Aizpurua, F. M. Huang,
J. J. Baumberg and S. Mahajan, ACS Nano, 2011, 5, 3878–3887.
[37] F. Huang, H. Zhang and J. F. Banfieldt, Nano Lett., 2003, 3, 373–378.
[38] J. W. Steed and J. L. Atwood, Supramolecular Chemistry Second Edition, John
Wiley & Sons, Chichester, UK, 2nd edn, 2009.
[39] J. M. Lehn, Angew. Chemie Int. Ed., 1988, 27, 89–112.
[40] J.-P. Sauvage, Angew. Chemie - Int. Ed., 2017, 56, 1–15.
[41] F. M. Menger, Proc. Natl. Acad. Sci., 2002, 99, 4818–4822.
[42] A. Velazquez-Campoy and E. Freire, Nat. Protoc., 2006, 1, 186–191.
[43] C. J. Pedersen, J. Am. Chem. Soc., 1967, 89, 7017–7036.
[44] B. Dietrich, J. Lehn and J. Sauvage, Tetrahedron Lett., 1969, 10, 2889–2892.
[45] C. D. Gutsche, B. Dhawan, H. N. Kwang and M. Ramamurthi, J. Am. Chem. Soc.,
1981, 103, 3782–3792.
[46] T. Ogoshi, S. Kanai, S. Fujinami, T.-a. Yamagishi and Y. Nakamoto, J. Am. Chem.
Soc., 2008, 130, 5022–5023.
[47] J. Szejtli, Chem. Rev., 1998, 98, 1743–1753.
[48] J. Kim, I. Jung, S. Kim, E. Lee, J. Kang, S. Sakamoto, K. Yamaguchi and K. Kim,
J. Am. Chem. Soc., 2000, 122, 540–541.
[49] W. A. Freeman, W. L. Mock and N.-Y. Shih, J. Am. Chem. Soc., 1981, 17, 7367–68.
[50] J. Lee, S. Samal, N. Selvapalam, H. Kim and K. Kim, Acc. Chem. Res., 2003, 36,
621–630.
[51] E. Masson, X. Ling, R. Joseph, L. Kyeremeh-Mensah and X. Lu, RSC Adv., 2012,
2, 1213–1247.
[52] A. Day, A. P. Arnold, R. J. Blanch and B. Snushall, J. Org. Chem., 2001, 66,
8094–8100.
[53] S. Mecozzi and J. Rebek, Chem. - A Eur. J., 1998, 4, 1016–1022.
[54] F. Biedermann, V. D. Uzunova, O. A. Scherman, W. M. Nau and A. De Simone, J.
Am. Chem. Soc., 2012, 134, 15318–15323.
[55] F. Biedermann, M. Vendruscolo, O. A. Scherman, A. De Simone and W. M. Nau,
J. Am. Chem. Soc., 2013, 135, 14879–14888.
171
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
[56] M. Florea and W. M. Nau, Angew. Chemie - Int. Ed., 2011, 50, 9338–9342.
[57] M. V. Rekharsky, T. Mori, C. Yang, Y. H. Ko, N. Selvapalam, H. Kim, D. Sobrans-
ingh, A. E. Kaifer, S. Liu, L. Isaacs, W. Chen, S. Moghaddam, M. K. Gilson, K. Kim
and Y. Inoue, Proc. Natl. Acad. Sci. USA, 2007, 104, 20737–20742.
[58] D. Sigwalt, M. Sˇekutor, L. Cao, P. Y. Zavalij, J. Hostasˇ, H. Ajani, P. Hobza, K. Mlin-
aric´-Majerski, R. Glaser and L. Isaacs, J. Am. Chem. Soc., 2017, 139, 3249–3258.
[59] F. Biedermann and O. A. Scherman, J. Phys. Chem. B, 2012, 116, 2842–2849.
[60] E. A. Appel, F. Biedermann, U. Rauwald, S. T. Jones, J. M. Zayed and O. A.
Scherman, J. Am. Chem. Soc., 2010, 14251–14260.
[61] M. J. Rowland, E. a. Appel, R. J. Coulston and O. A. Scherman, J. Mater. Chem.
B, 2013, 1, 2904.
[62] E. A. Appel, X. J. Loh, S. T. Jones, F. Biedermann, C. A. Dreiss and O. A. Scher-
man, J. Am. Chem. Soc., 2012, 134, 11767–11773.
[63] F. Tian, D. Jiao, F. Biedermann and O. A. Scherman, Nat. Commun., 2012, 3, 1207.
[64] C. S. Y. Tan, J. del Barrio, J. Liu and O. A. Scherman, Polym. Chem., 2015, 6,
7652–7657.
[65] J. del Barrio, P. N. Horton, D. Lairez, G. O. Lloyd, C. Toprakcioglu and O. A.
Scherman, J. Am. Chem. Soc., 2013, 135, 11760–11763.
[66] K. Kim, W. S. Jeon, J.-k. Kang, J. W. Lee, S. Y. Jon, T. Kim and K. Kim, Angew.
Chemie - Int. Ed., 2003, 42, 2293–2296.
[67] S. Angelos, Y.-W. Yang, K. Patel, J. F. Stoddart and J. I. Zink, Angew. Chemie -
Int. Ed., 2008, 47, 2222–2226.
[68] C. Hu, Y. Zheng, Z. Yu, C. Abell and O. A. Scherman, Chem. Commun., 2015, 51,
4858–4860.
[69] G. Stephenson, R. M. Parker, Y. Lan, Z. Yu, O. A. Scherman and C. Abell, Chem.
Commun., 2014, 50, 7048–51.
[70] Y. Zheng, Z. Yu, R. M. Parker, Y. Wu, C. Abell and O. A. Scherman, Nat. Commun.,
2014, 5, 1–9.
[71] Z. Yu, J. Zhang, R. J. Coulston, R. M. Parker, F. Biedermann, X. Liu, O. A.
Scherman and C. Abell, Chem. Sci., 2015, 6, 4929–4933.
[72] Z. Yu, Y. Lan, R. M. Parker, W. Zhang, X. Deng, O. A. Scherman and C. Abell,
Polym. Chem., 2016, 7, 5996–6002.
[73] J.-C. Baret, Lab Chip, 2012, 12, 422–433.
172
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
[74] B. P. Binks, Curr. Opin. Colloid Interface Sci., 2002, 7, 21–41.
[75] D. J. Holt, R. J. Payne, W. Y. Chow and C. Abell, J. Colloid Interface Sci., 2010,
350, 205–211.
[76] A. Huebner, S. Sharma, M. Srisa-Art, F. Hollfelder, J. B. Edel and A. J. DeMello,
Lab Chip, 2008, 8, 1244.
[77] L. Shang, Y. Cheng and Y. Zhao, Chem. Rev., 2017, 117, 7964–8040.
[78] A. B. Theberge, F. Courtois, Y. Schaerli, M. Fischlechner, C. Abell, F. Hollfelder
and W. T. S. Huck, Angew. Chemie - Int. Ed., 2010, 49, 5846–5868.
[79] A. J. DeMello, Nature, 2006, 442, 394–402.
[80] T. M. Squires and S. R. Quake, Rev. Mod. Phys., 2005, 77, 977–1026.
[81] J. C. Mcdonald, D. C. Duffy, J. R. Anderson, D. T. Chiu, H. Wu, O. J. A. Schueller
and G. M. Whitesides, Electrophoresis, 2000, 21, 27–40.
[82] H. Becker and L. E. Locascio, Talanta, 2002, 56, 267–287.
[83] D. J. Harrison, K. Fluri, K. Seiler, Z. Fan, C. S. Effenhauser and A. Manz, Science
(80-. )., 1993, 261, 895–897.
[84] Y. Xia and G. M. Whitesides, Angew. Chemie Int. Ed., 1998, 37, 550–575.
[85] D. Qin, Y. Xia and G. M. Whitesides, Nat. Protoc., 2010, 5, 491–502.
[86] T. Thorsen, R. W. Roberts, F. H. Arnold and S. R. Quake, Phys. Rev. Lett., 2001,
86, 4163–4166.
[87] D. J. Holt, R. J. Payne and C. Abell, J. Fluor. Chem., 2010, 131, 398–407.
[88] P. Gruner, B. Riechers, L. A. Chacon Orellana, Q. Brosseau, F. Maes, T. Beneyton,
D. Pekin and J. C. Baret, Curr. Opin. Colloid Interface Sci., 2015, 20, 183–191.
[89] G. Odian, Principles of Polymerization, John Wiley & Sons, New Jersey, USA, 4th
edn, 2004.
[90] M. Jikei and M. Kakimoto, Prog. Polym. Sci., 2001, 26, 1233–1285.
[91] P. J. Flory, J. Am. Chem. Soc., 1952, 74, 2718–2723.
[92] C. J. Hawker and J. M. J. Fre´chet, J. Am. Chem. Soc., 1990, 112, 7638–7647.
[93] L. Brunsveld, B. J. B. Folmer, E. W. Meijer and R. P. Sijbesma, Chem. Rev., 2001,
101, 4071–4097.
[94] E. Krieg, M. M. C. Bastings, P. Besenius and B. Rybtchinski, Chem. Rev., 2016,
116, 2414–2477.
[95] T. Aida, E. W. Meijer and S. I. Stupp, Science (80-. )., 2012, 335, 813–817.
173
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
[96] T. Park and S. C. Zimmerman, J. Am. Chem. Soc., 2006, 128, 13986–13987.
[97] W. H. Carothers, Chem. Rev., 1931, 8, 353–426.
[98] S. A. Schmid, R. Abbel, A. P. H. Schenning, E. W. Meijer, R. P. Sijbesma and L. M.
Herz, J. Am. Chem. Soc., 2009, 131, 17696–17704.
[99] C.-C. Chen and E. E. Dormidontova, Macromolecules, 2004, 37, 3905–3917.
[100] S. Lahiri, J. L. Thompson and J. S. Moore, J. Am. Chem. Soc., 2000, 122, 11315–
11319.
[101] M. Kastler, W. Pisula, D. Wasserfallen, T. Pakula and K. Mu, J. Am. Chem. Soc.,
2005, 127, 4286–4296.
[102] R. P. Sijbesma, F. H. Beijer, L. Brunsveld, B. J. B. Folmer, J. H. K. K. Hirschberg,
R. F. M. Lange, J. K. L. Lowe and E. W. Meijer, Science (80-. )., 1997, 278,
1601–1604.
[103] C. T. Seto and G. M. Whitesides, J. Am. Chem. Soc., 1990, 112, 6409–6411.
[104] R. K. Castellano, D. M. Rudkevich and J. Rebek Jr., Proc. Natl. Acad. Sci. USA,
1997, 94, 7132–7137.
[105] C. Schmuck, Tetrahedron, 2001, 57, 3063–3067.
[106] U. Velten and M. Rehahn, Chem. Commun., 1996, 0, 2639–2640.
[107] A. T. ten Cate and R. P. Sijbesma, Macromol. Rapid Commun., 2002, 23, 1094–
1112.
[108] P. R. Ashton, I. Baxter, S. J. Cantrill, M. C. T. Fyfe, P. T. Glink, J. F. Stoddart,
A. J. P. White and D. J. Williams, Angew. Chemie - Int. Ed., 1998, 37, 1294–1297.
[109] N. Yamaguchi, D. S. Nagvekar and H. W. Gibson, Angew. Chemie - Int. Ed., 1998,
37, 2361–2364.
[110] H. W. Gibson, N. Yamaguchi and J. W. Jones, J. Am. Chem. Soc., 2003, 125,
3522–3533.
[111] K. Ohga, Y. Takashima, H. Takahashi, Y. Kawaguchi, H. Yamaguchi and A. Harada,
Macromolecules, 2005, 38, 5897–5904.
[112] J. D. Hartgerink, J. R. Granja, R. A. Milligan and M. R. Ghadiri, J. Am. Chem.
Soc., 1996, 118, 43–50.
[113] S. Boileau, L. Bouteiller, F. Laupretre and F. Lortie, New J. Chem., 2000, 24,
845–848.
[114] A. Arnaud, J. Belleney, F. Boue´, L. Bouteiller, G. Carrot and V. Wintgens, Angew.
Chemie Int. Ed., 2004, 43, 1718–1721.
174
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
[115] C. Fouquey, J. Lehn and A. Levelut, Adv. Mater., 1990, 2, 254–257.
[116] B. A. Blight, C. A. Hunter, D. A. Leigh, H. McNab and P. I. T. Thomson, Nat.
Chem., 2011, 3, 246–250.
[117] G. B. W. L. Ligthart, H. Ohkawa, R. P. Sijbesma and E. W. Meijer, J. Am. Chem.
Soc., 2005, 127, 810–811.
[118] R. F. M. Lange, M. V. Gurp and E. W. Meijer, J. Polym. Sci. Part A, 1999, 37,
3657–3670.
[119] D. Markovitsi, H. Bengs and H. Ringsdorf, J. Chem. Soc Faraday Trans., 1992, 88,
1275–1279.
[120] Y. Yasuda, E. Iishi, H. Inada and Y. Shirota, Chem. Lett., 1996, 25, 575–576.
[121] S. Bonazzi, M. Capobianco, M. M. De Morais, A. Garbesi, G. Gottarelli, P. Mariani,
M. G. Ponzi Bossi, G. P. Spada and L. Tondelli, J. Am. Chem. Soc., 1991, 113,
5809–5816.
[122] U. Michelsen and C. A. Hunter, Angew. Chem. Int. Ed. Engl., 2000, 39, 764–767.
[123] A. K. Miller, Z. Li, K. A. Streletzky, A. M. Jamieson and S. J. Rowan, Polym.
Chem., 2012, 3, 3132.
[124] C. M. A. Leenders, L. Albertazzi, T. Mes, M. M. E. Koenigs, A. R. A. Palmans and
E. W. Meijer, Chem. Commun., 2013, 49, 1963–1965.
[125] J. Leenders, M. B. Baker, I. Pijpers, R. Lafleur, L. Albertazzi, A. R. A. Palmans
and E. W. Meijer, Soft Matter, 2016, 12, 2887–2893.
[126] J. Baram, E. Shirman, N. Ben-Shitrit, A. Ustinov, H. Weissman, I. Pinkas, S. G.
Wolf and B. Rybtchinski, J. Am. Chem. Soc., 2008, 130, 14966–14967.
[127] E. Krieg, E. Shirman, H. Weissman, E. Shimoni, S. G. Wolf, I. Pinkas and B. Rybtch-
inski, J. Am. Chem. Soc., 2009, 131, 14365–14373.
[128] P. Cordier, F. Tournilhac, C. Soulie´-Ziakovic and L. Leibler, Nature, 2008, 451,
977–80.
[129] Y. Yamamoto, T. Fukushima, Y. Suna, N. Ishii, A. Saeki, S. Seki, S. Tagawa,
M. Taniguchi, T. Kawai and T. Aida, Science (80-. )., 2006, 314, 1761–1764.
[130] R. Abbel, C. Grenier, M. J. Pouderoijen, J. W. Stouwdam, P. E. L. G. Lecle, R. P.
Sijbesma, E. W. Meijer and A. P. H. J. Schenning, J. Am. Chem. Soc., 2009, 131,
833–843.
[131] G. A. Silva, C. Czeisler, K. L. Niece, E. Beniash, D. A. Harrington, J. A. Kessler
and S. I. Stupp, Science (80-. )., 2004, 303, 1352–1355.
[132] P. J. Wyatt, Anal. Chim. Acta, 1993, 272, 1–40.
175
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
[133] Y. Liu, R. Fang, X. Tan, Z. Wang and X. Zhang, Chem. - A Eur. J., 2012, 18,
15650–15654.
[134] R. Fang, Y. Liu, Z. Wang and X. Zhang, Polym. Chem., 2013, 4, 900–903.
[135] N. Yamaguchi and H. W. Gibson, Angew. Chemie Int. Ed., 1999, 38, 143–147.
[136] B. Yu, S. Guo, L. He and W. Bu, Chem. Commun., 2013, 49, 3333–3335.
[137] G. Gattuso, A. Notti, A. Pappalardo, M. F. Parisi, I. Pisagatti, S. Pappalardo,
D. Garozzo, A. Messina, Y. Cohen and S. Slovak, J. Org. Chem., 2008, 73, 7280–
7289.
[138] S. Pappalardo, V. Villari, S. Slovak, Y. Cohen and M. F. Parisi, Chem. - A Eur. J.,
2007, 13, 8164–8173.
[139] Z. Zhang, Y. Luo, J. Chen, S. Dong, Y. Yu, Z. Ma and F. Huang, Angew. Chemie
Int. Ed., 2011, 123, 1433–1437.
[140] N. L. Strutt, H. Zhang, M. A. Giesener and J. F. Stoddart, Chem. Commun., 2012,
48, 1647–1649.
[141] C. Li, K. Han, J. Li, Y. Zhang, W. Chen, Y. Yu and X. Jia, Chem. - A Eur. J.,
2013, 19, 11892–11897.
[142] A. Harada, Y. Kawaguchi and T. Hoshino, J. Incl. Phenom., 2001, 41, 115–121.
[143] M. Miyauchi, Y. Kawaguchi and A. Harada, J. Incl. Phenom. Macrocycl. Chem.,
2004, 50, 57–62.
[144] Y. Hasegawa, M. Miyauchi, Y. Takashima, H. Yamaguchi and A. Harada, Macro-
molecules, 2005, 38, 3724–3730.
[145] Q. Zhang, D.-H. Qu, J. Wu, X. Ma, Q. Wang and H. Tian, Langmuir, 2013, 29,
5345–5350.
[146] E. Alvarez-parrilla, P. R. Cabrer, W. Al-soufi, F. Meijide, E. R. Nu´nez and J. V.
Tato, Angew. Chemie Int. Ed., 2000, 39, 2856–2858.
[147] V. H. Soto Tellini, A. Jover, J. Carrazana Garc´ıa, L. Galantini, F. Meijide and
J. Va´zquez Tato, J. Am. Chem. Soc., 2006, 128, 5728–5734.
[148] J. A. McCune, E. Rosta and O. A. Scherman, Org. Biomol. Chem., 2017, 15, 998–
1005.
[149] Y. H. Ko, K. Kim, E. Kim and K. Kim, Supramol. Chem., 2007, 19, 287–293.
[150] X. Tan, L. Yang, Y. Liu, Z. Huang, H. Yang, Z. Wang and X. Zhang, Polym. Chem.,
2013, 4, 5378–5381.
[151] L. Yang, Y. Bai, X. Tan, Z. Wang and X. Zhang, ACS Macro Lett., 2015, 4, 611–615.
176
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
[152] H. Yang, Z. Ma, Z. Wang and X. Zhang, Polym. Chem., 2014, 5, 1471–1476.
[153] J. Tian, T.-Y. Zhou, S.-C. Zhang, S. Aloni, M. V. Altoe, S.-H. Xie, H. Wang, D.-W.
Zhang, X. Zhao, Y. Liu and Z.-T. Li, Nat. Commun., 2014, 5, 1–11.
[154] L. Zhang, T.-Y. Zhou, J. Tian, H. Wang, D.-W. Zhang, X. Zhao, Y. Liu and Z.-T.
Li, Polym. Chem., 2014, 5, 4715–4721.
[155] K. D. Zhang, J. Tian, D. Hanifi, Y. Zhang, A. C. H. Sue, T. Y. Zhou, L. Zhang,
X. Zhao, Y. Liu and Z. T. Li, J. Am. Chem. Soc., 2013, 135, 17913–17918.
[156] H. Li, X. Fan, X. Shang, M. Qi, H. Zhang and W. Tian, Polym. Chem., 2016, 7,
4322–4325.
[157] H. Li, X. Fan, W. Tian, H. Zhang, W. Zhang and Z. Yang, Chem. Commun., 2014,
50, 14666–14669.
[158] X. Yan, T. R. Cook, J. B. Pollock, P. Wei, Y. Zhang, Y. Yu, F. Huang and P. J.
Stang, J. Am. Chem. Soc., 2014, 136, 4460–4463.
[159] X. C. I. Solvas and A. DeMello, Chem. Commun., 2011, 47, 1936–1942.
[160] S. Knoche, D. Vella, E. Aumaitre, P. Degen, H. Rehage, P. Cicuta and J. Kierfeld,
Langmuir, 2013, 29, 12463–12471.
[161] I. Jasiuk, J. Chen and M. F. Thorpe, Appl. Mech. Rev., 1994, 47, S18–S28.
[162] P. Fata´s, E. Longo, F. Rastrelli, M. Crisma, C. Toniolo, A. I. Jime´nez, C. Cativiela
and A. Moretto, Chem. - A Eur. J., 2011, 17, 12606–12611.
[163] B. Yu, B. Wang, S. Guo, Q. Zhang, X. Zheng, H. Lei, W. Liu, W. Bu, Y. Zhang and
X. Chen, Chem. - A Eur. J., 2013, 19, 4922–4930.
[164] L. Li, X. Zheng, B. Yu, L. He, J. Zhang, H. Liu, Y. Cong and W. Bu, Polym. Chem.,
2016, 7, 287–291.
[165] W. Li, J. Qu, J. Du, K. Ren, Y. Wang, J. Sun and Q. Hu, Chem. Commun., 2014,
50, 9584–7.
[166] Z. Ge, H. Liu, Y. Zhang and S. Liu, Macromol. Rapid Commun., 2011, 32, 68–73.
[167] H. T. Zhang, X. D. Fan, W. Tian, R. T. Suo, Z. Yang, Y. Bai and W. B. Zhang,
Chem. - A Eur. J., 2015, 21, 5000–5008.
[168] D. Zou, S. Andersson, R. Zhang, S. Sun, B. Akermark and L. Sun, J. Org. Chem.,
2008, 73, 3775–3783.
[169] W. S. Jeon, A. Y. Ziganshina, J. W. Lee, Y. H. Ko, J.-K. Kang, C. Lee and K. Kim,
Angew. Chemie - Int. Ed., 2003, 42, 4097–4100.
177
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
[170] J. W. Lee, I. Hwang, S. Jeon, H. Ko, S. Sakamoto, K. Yamaguchi and K. Kim,
Chem. - An Asian J., 2008, 3, 1277–1283.
[171] X.-L. Wu, L. Luo, L. Lei, G.-H. Liao, L.-Z. Wu and C.-H. Tung, J. Org. Chem.,
2008, 73, 491–494.
[172] Z. Huang, K. Qin, G. Deng, G. Wu, Y. Bai, J. F. Xu, Z. Wang, Z. Yu, O. A.
Scherman and X. Zhang, Langmuir, 2016, 32, 12352–12360.
[173] W. Zhang, S. Gan, A. Vezzoli, R. J. Davidson, D. C. Milan, K. V. Luzyanin, S. J.
Higgins, R. J. Nichols, A. Beeby, P. J. Low, B. Li and L. Niu, ACS Nano, 2016, 10,
5212–5220.
[174] Y. Tan, S. Hou, H. Chen and H. Ma, Chem. - An Asian J., 2017, 12, 476–483.
[175] P. Mukhopadhyay, A. Wu and L. Isaacs, J. Org. Chem., 2004, 69, 6157–6164.
[176] S. Liu, C. Ruspic, P. Mukhopadhyay, S. Chakrabarti, P. Y. Zavalij and L. Isaacs, J.
Am. Chem. Soc., 2005, 127, 15959.
[177] Z. Huang, L. Yang, Y. Liu, Z. Wang, O. A. Scherman and X. Zhang, Angew. Chemie
- Int. Ed., 2014, 53, 5351–5355.
[178] A. S. Groombridge, A. Palma, R. M. Parker, C. Abell and O. A. Scherman, Chem.
Sci., 2017, 8, 1350–1355.
[179] M. Ramaekers, S. P. W. Wijnands, J. L. J. Van Dongen, L. Brunsveld and P. Y. W.
Dankers, Chem. Commun., 2015, 51, 3147–3150.
[180] C. A. Brautigam, H. Zhao, C. Vargas, S. Keller and P. Schuck, Nat. Protoc., 2016,
11, 882–894.
[181] U. Rauwald, F. Biedermann, S. Deroo, C. V. Robinson and O. A. Scherman, J.
Phys. Chem. B, 2010, 114, 8606–8615.
[182] S. Moghaddam, C. Yang, M. Rekharsky, Y. H. Ko, K. Kim, Y. Inoue and M. K.
Gilson, J. Am. Chem. Soc., 2011, 133, 3570–3581.
[183] J. Wu and L. Isaacs, Chem. - A Eur. J., 2009, 15, 11675–11680.
[184] N. Bilbao, V. Va´zquez-Gonza´lez, M. T. Aranda and D. Gonza´lez-Rodr´ıguez,
European J. Org. Chem., 2015, 2015, 7160–7175.
[185] O. Bande, R. Abu El Asrar, D. Braddick, S. Dumbre, V. Pezo, G. Schepers, V. B.
Pinheiro, E. Lescrinier, P. Holliger, P. Marlie`re and P. Herdewijn, Chem. - A Eur.
J., 2015, 21, 5009–5022.
[186] B. Mohapatra and S. Verma, Cryst. Growth Des., 2016, 16, 696–704.
[187] S. Keller, C. Vargas, H. Zhao, G. Piszczek, C. A. Brautigam and P. Schuck, Anal.
Chem., 2012, 84, 5066–5073.
178
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
[188] F. Tian, N. Cheng, N. Nouvel, J. Geng and O. A. Scherman, Langmuir, 2010, 26,
5323–5328.
[189] P. Stawski, M. Sumser and D. Trauner, Angew. Chemie - Int. Ed., 2012, 51, 5748–
5751.
[190] M. C. M. Daniel and D. Astruc, Chem. Rev., 2004, 104, 293–346.
[191] V. Amendola, R. Pilot, M. Frasconi, O. M. Marago and M. Antonia lati, J. Phys.
Condens. Matter, 2017, 29, 1–48.
[192] J. Turkevich, P. C. Stevenson and J. Hillier, Discuss. Faraday Soc., 1951, 11, 55–75.
[193] G. Frens, Nat. Phys. Sci., 1973, 241, 20–22.
[194] J. Homola, S. S. Yee and G. Gauglitz, Sensors Actuators B Chem., 1999, 54, 3–15.
[195] R. Chikkaraddy, B. de Nijs, F. Benz, S. J. Barrow, O. A. Scherman, E. Rosta,
A. Demetriadou, P. Fox, O. Hess and J. J. Baumberg, Nature, 2016, 535, 127–130.
[196] A. I. Ekimov, A. L. Efros and A. A. Onushchenko, Solid State Commun., 1985, 56,
921–924.
[197] M. L. Steigerwald, A. P. Alivisatos, J. M. Gibson, T. D. Harris, R. Kortan, A. J.
Muller, A. M. Thayer, T. M. Duncan, D. C. Douglass and L. E. Brus, J. Am. Chem.
Soc., 1988, 110, 3046–3050.
[198] A. P. Alivisatos, Science (80-. )., 1996, 271, 933–937.
[199] M. Bruchez Jr., M. Moronne, P. Gin, S. Weiss and A. P. Alivisatos, Science (80-.
)., 1998, 281, 2013–2016.
[200] M. Han, X. Gao, J. Z. Su and S. Nie, Nat. Biotechnol., 2001, 19, 631–635.
[201] V. L. Colvin, M. C. Schlamp and A. P. Alivisatos, Nature, 1994, 370, 354–357.
[202] C. Chuang, P. R. Brown, V. Bulovic´ and M. Bawendi, Nat. Mater., 2014, 13, 796.
[203] Z. Kang, C. H. A. Tsang, N.-b. Wong, Z. Zhang and S.-t. Lee, J. Am. Chem. Soc.,
2007, 129, 12090–12091.
[204] R. J. Macfarlane, M. N. O’Brien, S. H. Petrosko and C. A. Mirkin, Angew. Chemie
- Int. Ed., 2013, 52, 5688–5698.
[205] R. G. Pearson, Inorg. Chem., 1988, 27, 734–740.
[206] C. B. Murray, C. R. Kagan and M. G. Bawendi, Science (80-. )., 1995, 270, 1335–
1338.
[207] E. V. Shevchenko, D. V. Talapin, N. A. Kotov, S. O’Brien and C. B. Murray, Nature,
2006, 439, 55–59.
179
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
[208] A. M. Kalsin, M. Fialkowski, M. Paszewski, S. K. Smoukov, K. J. M. Bishop and
B. A. Grzybowski, Science (80-. )., 2006, 312, 420–424.
[209] A. P. Alivisatos, K. P. Johnsson, X. Peng, T. E. Wilson, C. J. Loweth, M. P. Bruchez
Jr and P. G. Schultz, Nature, 1996, 382, 609–611.
[210] C. A. Mirkin, R. L. Letsinger, R. C. Mucic and J. J. Storhoff, Nature, 1996, 382,
607–609.
[211] E. Auyeung, T. I. N. G. Li, A. J. Senesi, A. L. Schmucker, B. C. Pals, M. O. de la
Cruz and C. A. Mirkin, Nature, 2013, 505, 73–77.
[212] H. Lin, S. Lee, L. Sun, M. Spellings, M. Engel, S. C. Glotzer and C. A. Mirkin,
Science (80-. )., 2017, 355, 931–935.
[213] B. Ding, Z. Deng, H. Yan, S. Cabrini and R. N. Zuckermann, J. Am. Chem. Soc.,
2010, 132, 1–16.
[214] G. P. Mitchell, C. A. Mirkin and R. L. Letsinger, J. Am. Chem. Soc., 1999, 121,
8122–8123.
[215] D. Lawless, S. Kapoor and D. Meisel, J. Phys. Chem., 1995, 99, 10329–10335.
[216] M. Brust, D. J. Schiffrin, D. Bethell and C. J. Kiely, Adv. Mater., 1995, 7, 795–797.
[217] A. Andre´, D. Zherebetskyy, D. Hanifi, B. He, M. Samadi Khoshkhoo, M. Jankowski,
T. Chasse´, L. W. Wang, F. Schreiber, A. Salleo, Y. Liu and M. Scheele, Chem.
Mater., 2015, 27, 8105–8115.
[218] A. R. Salmon, R. Esteban, R. W. Taylor, J. T. Hugall, C. A. Smith, G. Whyte, O. A.
Scherman, J. Aizpurua, C. Abell and J. J. Baumberg, Small, 2016, 12, 1788–1796.
[219] K. L. Wustholz, A. I. Henry, J. M. McMahon, R. G. Freeman, N. Valley, M. E.
Piotti, M. J. Natan, G. C. Schatz and R. P. V. Duyne, J. Am. Chem. Soc., 2010,
132, 10903–10910.
[220] R. Jiang, B. Li, C. Fang and J. Wang, Adv. Mater., 2014, 26, 5274–5309.
[221] K. Wu, J. Chen, J. R. Mcbride and T. Lian, Science (80-. )., 2015, 349, 632–635.
[222] A. O. Govorov, G. W. Bryant, W. Zhang, T. Skeini, J. Lee, N. A. Kotov, J. M.
Slocik and R. R. Naik, Nano Lett., 2006, 6, 984–994.
[223] C. Strelow, T. S. Theuerholz, C. Schmidtke, M. Richter, J. P. Merkl, H. Kloust,
Z. Ye, H. Weller, T. F. Heinz, A. Knorr and H. Lange, Nano Lett., 2016, 16, 4811–
4818.
[224] E. Cohen-Hoshen, G. W. Bryant, I. Pinkas, J. Sperling and I. Bar-Joseph, Nano
Lett., 2012, 12, 4260–4264.
180
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
[225] Q. Fu, Y. Sheng, H. Tang, Z. Zhu, M. Ruan, W. Xu and Y. Zhu, ACS Nano, 2015,
9, 172–179.
[226] S. Kasera, Z. Walsh, J. del Barrio and O. A. Scherman, Supramol. Chem., 2014, 26,
1–7.
[227] T.-C. Lee and O. A. Scherman, Chem. Commun., 2010, 46, 2438–2440.
[228] Q. An, G. Li, C. Tao, Y. Li, Y. Wu and W. Zhang, Chem. Commun., 2008, 1989–91.
[229] R. Xie, D. Battaglia and X. Peng, J. Am. Chem. Soc., 2007, 129, 15432–15433.
[230] S. Adam, D. V. Talapin, H. Borchert, A. Lobo, C. Mcginley, A. R. B. de Castro,
M. Haase, H. Weller and T. Mo¨ller, J. Chem. Phys., 2005, 123, 084706.
[231] D. V. Talapin, N. Gaponik, H. Borchert, A. L. Rogach, M. Haase and H. Weller, J.
Phys. Chem. B, 2002, 106, 12659–12663.
[232] W. Haiss, N. T. K. Thanh, J. Aveyard and D. G. Fernig, Anal. Chem., 2007, 79,
4215–4221.
[233] J. Kimling, M. Maier, B. Okenve, V. Kotaidis, H. Ballot and A. Plech, J. Phys.
Chem. B, 2006, 110, 15700–15707.
[234] J. Rodriguez-Ferna´ndez, J. Pe´rez-Juste, F. J. Garc´ıa de Abajo and L. M. Liz-Marza´n,
Langmuir, 2006, 22, 7007–7010.
[235] S. A. Agnihotri, N. N. Mallikarjuna and T. M. Aminabhavi, J. Control. Release,
2004, 100, 5–28.
[236] J. Lee, P. W. Bisso, R. L. Srinivas, J. J. Kim, A. J. Swiston and P. S. Doyle, Nat.
Mater., 2014, 13, 524–529.
[237] C. M. Doherty, D. Buso, A. J. Hill, S. Furukawa, S. Kitagawa and P. Falcaro, Acc.
Chem. Res., 2014, 47, 396–405.
[238] A. D. Dinsmore, M. F. Hsu, M. G. Nikolaides, M. Marquez, A. R. Bausch and D. A.
Weitz, Science (80-. )., 2002, 298, 1006–1009.
[239] H. Hiramatsu and F. E. Osterloh, Chem. Mater., 2004, 16, 2509–2511.
[240] T. Sakai and P. Alexandridis, Langmuir, 2005, 170, 8019–8025.
[241] J. Ehgartner, M. Strobl, J. M. Bolivar, D. Rabl, M. Rothbauer, P. Ertl, S. M. Borisov
and T. Mayr, Anal. Chem., 2016, 88, 9796–9804.
[242] P. Gruner, B. Riechers, B. Semin, J. Lim, A. Johnston, K. Short and J. Baret, Nat.
Commun., 2016, 7, 1–9.
[243] J. C. de Mello, H. F. Wittmannn and R. Friend, Adv. Mater., 1997, 9, 230–232.
[244] G. M. Whitesides and B. Grzybowski, Science (80-. )., 2002, 295, 2418–2422.
181
182
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Appendix A
Supplementary Information
A.1 Supplementary Data for Chapter 3
A.1.1 2D NMR Spectra
NpVio2
Figure A.1 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for NpVio2. For COSY, HSQC and HMBC positive peaks are shown in blue and
negative peaks are shown in red. 183
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure A.2 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the NpVio2-2CB[8] complex. For COSY, HSQC and HMBC positive peaks
are shown in blue and negative peaks are shown in red.
184
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure A.3 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the NpVio2-1CB[7] complex. For COSY, HSQC and HMBC positive peaks
are shown in blue and negative peaks are shown in red.
185
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure A.4 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the NpVio2-2CB[7] complex. For COSY, HSQC and HMBC positive peaks
are shown in blue and negative peaks are shown in red.
186
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure A.5 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the NpVio2-1CB[8]-1CB[7] complex. For COSY, HSQC and HMBC positive
peaks are shown in blue and negative peaks are shown in red.
187
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure A.6 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the NpVio2-2CB[8] and 2-naphthol complex. For COSY, HSQC and
HMBC positive peaks are shown in blue and negative peaks are shown in red.
188
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
AzoVio2
Figure A.7 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for AzoVio2. For COSY, HSQC and HMBC positive peaks are shown in blue
and negative peaks are shown in red.
189
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure A.8 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the AzoVio2-1CB[8] complex. For COSY, HSQC and HMBC positive peaks
are shown in blue and negative peaks are shown in red.
190
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure A.9 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the AzoVio2-2CB[8] complex. For COSY, HSQC and HMBC positive peaks
are shown in blue and negative peaks are shown in red.
191
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure A.10 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the AzoVio2-1CB[7] complex. For COSY, HSQC and HMBC positive peaks
are shown in blue and negative peaks are shown in red.
192
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure A.11 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the AzoVio2-2CB[7] complex. For COSY, HSQC and HMBC positive peaks
are shown in blue and negative peaks are shown in red.
193
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure A.12 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the AzoVio2-1CB[8]-1CB[7] complex. For COSY, HSQC and HMBC pos-
itive peaks are shown in blue and negative peaks are shown in red.
194
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Np2Vio
Figure A.13 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for Np2Vio. For COSY, HSQC and HMBC positive peaks are shown in blue and
negative peaks are shown in red.
195
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure A.14 1H-1H COSY, 1H-13C HSQC, 1H-13C HMBC, and 1H-D DOSY 2D NMR
spectra for the Np2Vio-CB[8] complex. For COSY, HSQC and HMBC positive peaks
are shown in blue and negative peaks are shown in red.
196
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
A.1.2 ITC Titrations
Figure A.15 ITC titrations of methyl viologen (1 mM) into CB[8] (51 µM), and a com-
petitive binding titration of ADA.HCl (1 mM) into methyl viologen-CB[8] (1 mM and 51
µM).
197
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
A.1.3 AzoVio2 Photo-isomerisation
1H NMR Stacks
Figure A.16 Stacked 1H NMR spectra of the AzoVio2-1CB[8] complex following its
photo-isomerisation. From the bottom the equilibrium state is shown, then predominately
the Z isomer after 4 h UVA irradiation, then predominately the E isomer after heating.
Peaks labelled blue correspond to the Z isomer.
198
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure A.17 Stacked 1H NMR spectra of the AzoVio2-2CB[8] complex following its
photo-isomerisation. From the bottom the equilibrium state is shown, then predominately
the Z isomer after 4 h UVA irradiation, then predominately the E isomer after heating.
Peaks labelled red correspond to the E isomer, and those in blue to the Z isomer.
199
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure A.18 Stacked 1H NMR spectra of the AzoVio2-1CB[7] complex following its
photo-isomerisation. From the bottom the equilibrium state is shown, then predominately
the Z isomer after 4 h UVA irradiation, then predominately the E isomer after heating.
Peaks labelled red correspond to the E isomer, and those in blue to the Z isomer.
200
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
Figure A.19 Stacked 1H NMR spectra of the AzoVio2-2CB[7] complex following its
photo-isomerisation. From the bottom the equilibrium state is shown, then predominately
the Z isomer after 4 h UVA irradiation, then predominately the E isomer after heating.
Peaks labelled red correspond to the E isomer, and those in blue to the Z isomer.
201
Aqueous Self-Assembly with Cucurbit[n]urils: From Solution to Emulsion
A.2 Supplementary Information for Chapter 4
A.2.1 Ostwald Ripening Fitting
The equation used for least squares fitting of the second stage of QD-CB[7] and AuNC-
QD-CB[7] supraparticle growth is shown in Eq. A.1, where D is the average particle
diameter, D0 is the average particle diameter at t = 0, t is time, and k is a temperature-
dependent material constant appropriate for the value of n.37 The fitting parameters
used are shown in Table A.1. The value of n is indicative of different types of Ostwald
growth mechanisms. Here it is clear that different mechanisms for growth were observed
at different concentrations of CB[7], likely a result of the propensity for single particles
to migrate at low CB[7] concentrations, versus small aggregates migrating at high CB[7]
concentrations. Time-resolved small angle X-ray scattering techniques would be ideal to
deconvolute this mechanism.
D −D0 = kt1/n (A.1)
QD:CB[7]:AuNC Ratio 1/n k t /min D0 /nm
1:3.1:0 0.10 1.78 5 3.31
1:6.3:0 0.14 1.98 5 3.88
1:8.7:0 0.41 0.28 2 7.59
1:20.9:0 0.53 0.16 2 9.75
1:24.4:0 0.47 0.44 5 10.22
1:27.9:0 0.47 0.53 2 11.72
1:31.4:0 0.84 0.07 20 19.05
1:3.6:0.006 0.00 10.14 55 2.85
1:7.2:0.006 0.76 0.07 15 15.83
1:8.6:0.006 0.55 1.20 2 9.34
1:10.8:0.006 0.64 1.61 5 6.89
Table A.1 A table of the fitting parameters used in an Ostwald ripening model to fit
DLS kinetics data of supraparticle growth.
202